cs.RO / 1 / 2606.28345
Auditing LLM-Governed Social Robots with Culture-Specific Moral Gradients
审计由大型语言模型（LLM）管理的具有文化特定道德梯度的社会机器人
Abstract
LLM-governed social robots increasingly decide who receives real-world assistance first. As prioritization norms vary across cultures by age, status, and group size, failure to calibrate pluralistically can scale into unequal access. Yet LLM moral audits remain English-centered, rarely test embodied contexts, leaving pluralistic calibration as an urgent diagnostic gap amid intensifying LLM-robot deployment. We introduce a gradient-based audit framework for multilingual evaluation of LLM moral trade-off behavior against cultural preference gradients. Grounded in nine cross-domain social robotics reviews (>8,000 papers), we derive symmetry-controlled scenarios across care, education, and services, translating the Moral Machine Experiment's "whom to spare" into "whom to assist first" dilemmas with preserved identity trade-offs (many vs. few; young vs. old; higher vs. lower status). We audit four LLMs across four country-language pairs in four prompting regimes (57,600 decisions), benchmarked against country-specific MME preference gradients. Ordinal concordance tests whether models differentiate cultural contexts; a governance typology maps vulnerabilities in gradient differentiation, directional tendency, and deliberation. We find persistent, culturally asymmetric gradient tracking failures that prompting alone cannot reliably correct: quality calibration is nearly twice as strong for Western-language decisions as for Chinese and Japanese; high determinism in majority-first trade-offs often erases cross-cultural gradients; partial sensitivity to age- and status-based norms risks sidelining minorities. Prompting effects are uneven; only contrastive exemplars yield consistent gains, while reasoning-only prompts can worsen tracking. Our results motivate multilingual, pluralistic audits as an LLM-robot pre-deployment gate and suggest model factors are a more robust lever than prompting alone.
Chinese Translation
由大型语言模型（LLM）管理的社会机器人日益决定谁优先获得现实世界的帮助。由于不同文化在年龄、地位和群体规模方面的优先级规范各异，未能进行多元化校准可能导致不平等的获取。然而，LLM的道德审计仍然以英语为中心，鲜有测试具身情境，使得多元化校准在日益增强的LLM-机器人部署中成为一个紧迫的诊断空白。我们提出了一种基于梯度的审计框架，用于对比文化偏好梯度，进行多语言的LLM道德权衡行为评估。基于九个跨领域的社会机器人文献综述（超过8000篇论文），我们推导出在关怀、教育和服务领域的对称控制场景，将道德机器实验中的“谁值得被放过”转化为“谁应优先获得帮助”的困境，并保留身份权衡（众多与少数；年轻与年老；高地位与低地位）。我们在四个国家-语言对中审计了四个LLM，采用四种提示机制（共57,600个决策），并与特定国家的MME偏好梯度进行了基准比较。序数一致性测试检验模型是否能区分文化背景；治理类型图谱映射梯度区分、方向倾向和深思熟虑中的脆弱性。我们发现，持续存在的文化不对称梯度追踪失败，单靠提示无法可靠纠正：西方语言决策的质量校准几乎是中文和日文的两倍；在多数优先权衡中，高决定性往往抹去跨文化梯度；对基于年龄和地位的规范的部分敏感性可能会使少数群体被边缘化。提示效果不均匀；只有对比示例能带来一致的收益，而仅依赖推理的提示可能会恶化追踪。我们的结果促使在LLM-机器人部署前进行多语言、多元化的审计，并建议模型因素比单纯的提示更为稳健的杠杆。
cs.RO / 2 / 2606.28384
A Query-Driven Communication-Efficient Digital Twins Design for Autonomous Driving
一种基于查询驱动的通信高效数字双胞胎设计用于自动驾驶
Abstract
Digital twins (DTs) have become a potential technology to perform risk-free simulation of physical entities for deterministic and high-reliability services in diverse scenarios such as autonomous driving and low-altitude economy. In the autonomous driving scenario, traditional DT methods that rely solely on vehicle's real-time state synchronization, however, might lead to unacceptable computing and communication consumption for construction of high-fidelity DT with redundant data. To address this issue, we first propose a query-driven DT architecture to enable the DT to actively request the desired environment data from vehicles based on its simulation result. Then, we formulate an optimization problem whose goal is to minimize autonomous driving position error while accounting for DT fidelity and communication constraints. We also design a cross-time-step progressive query mechanism to further improve communication efficiency. The simulation results show that our proposed method achieves a 24% reduction in planning position error compared to traditional methods, while reducing communication overhead by 40%.
Chinese Translation
数字双胞胎（Digital Twins, DTs）已成为在自动驾驶和低空经济等多种场景中进行无风险物理实体仿真的潜在技术，以提供确定性和高可靠性的服务。然而，在自动驾驶场景中，传统的仅依赖于车辆实时状态同步的DT方法可能导致高保真DT构建过程中由于冗余数据而产生不可接受的计算和通信消耗。为了解决这一问题，我们首先提出了一种查询驱动的DT架构，使DT能够根据其仿真结果主动请求车辆提供所需的环境数据。然后，我们构建了一个优化问题，其目标是在考虑DT保真度和通信约束的情况下，最小化自动驾驶位置误差。我们还设计了一种跨时间步的渐进查询机制，以进一步提高通信效率。仿真结果表明，与传统方法相比，我们提出的方法在规划位置误差上减少了24%，同时通信开销减少了40%。
cs.RO / 3 / 2606.28385
RoboGaze: Evaluating Robot World Models via Structured Vision-Language Analysis
RoboGaze：通过结构化视觉-语言分析评估机器人世界模型
Abstract
Recent advances in robot world models enable synthetic video generation for embodied prediction and planning. However, evaluating these videos is challenging: visually realistic outputs often violate physical laws, temporal consistency, or task logic, while conventional metrics and monolithic Vision-Language Model (VLM) judges fail to generalize or provide precise diagnostic value. We present RoboGaze, a training-free, multi-agent VLM framework that provides structured, interpretable evaluation for generated robot-manipulation videos. Given a task instruction and video, RoboGaze operates via a three-stage pipeline: task-scene grounding, dimension-specific specialist routing, and critic-based verification. It outputs temporally localized glitch reports categorized under a novel 6-dimension, 30-type robotics-specific taxonomy. To benchmark RoboGaze, we introduce a human-validated dataset of 382 clips spanning simulated and real-world multi-view manipulation. Evaluating eight open-source and proprietary VLM backbones, RoboGaze dramatically outperforms zero-shot baselines, improving description-F1 by up to +43 points and temporal alignment (F1 x IoU) by up to +37 points, closing approximately 85% of the gap to the human ceiling. Furthermore, its critic verifier mitigates the "cry-wolf" false-positive flaw of standard VLMs, lifting clean-clip accuracy from under 25% to over 80%. RoboGaze offers a scalable, highly interpretable diagnostic tool for the rigorous evaluation of robot world models.
Chinese Translation
近期机器人世界模型的进展使得能够生成用于具身预测和规划的合成视频。然而，评估这些视频具有挑战性：视觉上逼真的输出往往违反物理法则、时间一致性或任务逻辑，而传统指标和单一的视觉-语言模型（VLM）评估者无法进行有效的推广或提供精确的诊断价值。我们提出了RoboGaze，一个无训练的多智能体VLM框架，为生成的机器人操作视频提供结构化、可解释的评估。给定任务指令和视频，RoboGaze通过三阶段流程操作：任务场景定位、维度特定专家路由和基于评审的验证。它输出时间局部化的故障报告，按照一种新颖的6维、30类机器人特定分类法进行分类。为了基准测试RoboGaze，我们引入了一个经过人工验证的数据集，包含382个片段，涵盖模拟和真实世界的多视角操作。在评估八个开源和专有的VLM骨干网络时，RoboGaze显著超越零-shot基线，将描述-F1提高了最多43分，时间对齐（F1 x IoU）提高了最多37分，缩小了与人类上限约85%的差距。此外，其评审验证器减轻了标准VLM的“狼来了”假阳性缺陷，将干净片段的准确率从25%以下提升至80%以上。RoboGaze为机器人世界模型的严格评估提供了一个可扩展、高度可解释的诊断工具。
cs.RO / 4 / 2606.28455
Event-Conditioned Diagnostics of Kinematic, Contact, and Object-Permanence Fields in Passive Object-State World Models
被动物体状态世界模型中运动、接触和物体持久性场的事件条件诊断
Abstract
World models can predict future physical states, but prediction accuracy alone does not explain how physical information is organized and used inside their latent dynamics. We introduce a controlled diagnostic protocol for studying event-conditioned latent physical structure in passive object-state world models. The protocol tests whether hidden representations encode event-regime information, whether event contexts reweight non-exclusive physical field readouts, and whether field-aligned representational components have functional consequences for prediction. Using a balanced controlled-generator dataset with free-motion, collision, and occlusion events, we evaluate recurrent, attention-based, and latent state-space transition models under a fixed-horizon forecasting setup. The models learn useful predictive dynamics and their hidden states support reliable event-regime readout. Event contexts systematically reweight kinematic, contact, and object-permanence field readouts: free motion is kinematic-dominant, collision combines kinematic and contact structure, and occlusion combines motion-related and object-permanence structure. Time-aligned and directional-consistency analyses further show phase-related shifts in field emphasis. Finally, fixed-horizon projection causal field effect (CFE) shows that suppressing field-aligned directions can degrade event-relevant prediction, with strongest evidence for contact-aligned structure in collision-contact windows and more qualified evidence for object-permanence-aligned structure in hard-occlusion hidden windows. These results support event-conditioned organization and fixed-horizon functional sensitivity of latent physical fields, while not implying explicit physical modules, isolated causal circuits, or context-invariant sliding-window generalization.
Chinese Translation
世界模型能够预测未来的物理状态，但仅凭预测准确性并不能解释物理信息在其潜在动态中的组织和使用方式。我们提出了一种受控的诊断协议，用于研究被动物体状态世界模型中的事件条件潜在物理结构。该协议测试隐藏表示是否编码事件模式信息，事件上下文是否重新加权非独占物理场的读出，以及与场对齐的表示组件在预测中的功能后果。通过使用包含自由运动、碰撞和遮挡事件的平衡控制生成数据集，我们在固定时间范围的预测设置下评估了递归、基于注意力的和潜在状态空间转移模型。这些模型学习了有用的预测动态，其隐藏状态支持可靠的事件模式读出。事件上下文系统性地重新加权运动、接触和物体持久性场的读出：自由运动以运动为主，碰撞结合了运动和接触结构，而遮挡结合了与运动相关和物体持久性结构。时间对齐和方向一致性分析进一步显示了场强调的相位相关变化。最后，固定时间范围投影因果场效应（CFE）表明，抑制与场对齐的方向可能会降低事件相关预测的准确性，碰撞接触窗口中对接触对齐结构的证据最为强烈，而在硬遮挡隐藏窗口中对物体持久性对齐结构的证据则更为有限。这些结果支持事件条件组织和潜在物理场的固定时间范围功能敏感性，同时并不暗示明确的物理模块、孤立的因果电路或上下文不变的滑动窗口泛化。
cs.RO / 5 / 2606.28469
When May I Help You? On The Effect of Proactivity on Group Human-Robot Collaboration
我可以帮您吗？关于主动性对群体人机协作的影响
Abstract
Robot initiative is a central challenge in multi-party human-robot collaboration. A robot that contributes without being addressed may provide timely support, but it may also disrupt coordination, divide attention, or interrupt turn-taking; a robot that waits to be addressed may preserve human control, but it may also miss opportunities to assist. We investigate this design challenge in a collaborative escape room in which pairs of participants work with a humanoid robot under either a reactive interaction model, where the robot responds only when addressed, or a proactive model, where it listens continuously, contributes autonomously, and periodically re-initiates interaction. We evaluate both models using puzzle-solving performance, interaction frequency, and participant ratings on the Godspeed and RoSAS scales. The proactive model substantially increases interaction frequency, whereas the reactive model shows a descriptively higher overall success rate (92.86% vs. 71.42%). The strongest differences emerge when prior experience and personality are taken into account: participants with LLM experience solve the early puzzles faster in the reactive condition, and participants with prior robot experience show modified evaluations of proactive and reactive interaction as do introverted participants. These findings demonstrate that the effects of robot initiative are simultaneously shaped by users' prior experience, personality traits and more generally by the needs of the group.
Chinese Translation
机器人主动性是多方人机协作中的一个核心挑战。一个在未被提及时就主动贡献的机器人可能提供及时的支持，但也可能干扰协调、分散注意力或打断轮流发言；而一个等待被提及的机器人则可能保持人类的控制，但也可能错失提供帮助的机会。我们在一个协作逃脱室中研究这一设计挑战，在该环境中，参与者成对与一个类人机器人合作，采用反应式交互模型（即机器人仅在被提及时作出反应）或主动式模型（即机器人持续倾听、自动贡献并定期重新发起交互）。我们通过解谜表现、交互频率以及参与者在Godspeed和RoSAS量表上的评分来评估这两种模型。主动式模型显著提高了交互频率，而反应式模型则显示出更高的整体成功率（92.86%对比71.42%）。当考虑到先前经验和个性时，最明显的差异显现出来：具有大型语言模型（LLM）经验的参与者在反应条件下更快地解决早期难题，而具有机器人经验的参与者对主动和反应交互的评估有所不同，内向型参与者也是如此。这些发现表明，机器人主动性的影响同时受到用户先前经验、个性特征以及更广泛的群体需求的影响。
cs.RO / 6 / 2606.28475
Improvement of Robot's Simultaneous Localization and Mapping Using an Effective Transformation to Achieve Linear Model
通过有效变换实现线性模型的机器人同时定位与地图构建的改进
Abstract
Nowadays mobile robots have wide engineering applications. Simultaneous localization and mapping (SLAM) is an important task of these robots. The major and common algorithms used for this task are based on extended Kalman filter (EKF). One of the main problems in EKF-based SLAM is its divergence. The nonlinearity of motion and observation models and linearization error are the main reasons for the divergence. There have been some efforts to address this problem with limited success. In this paper, by applying a simple compass and using an effective transformation, we transform the non-linear state space model into a linear model. Then, by applying the original KF to this model, we reach a new method, which is called LMKF SLAM. We show that the LMKF SLAM is significantly superior to the state-of-the-art methods, especially EKF-based SLAMs, both in accuracy, convergence, and computational complexity. The proposed method is also more stable with respect to the uncertainty of sensors values and changes in system parameters. Experimental results verify these points.
Chinese Translation
如今，移动机器人在工程应用中具有广泛的应用。与此同时定位与地图构建（SLAM）是这些机器人面临的一项重要任务。用于此任务的主要和常见算法基于扩展卡尔曼滤波器（EKF）。EKF基础的SLAM面临的一个主要问题是其发散性。运动和观测模型的非线性以及线性化误差是导致发散的主要原因。虽然已有一些努力试图解决这个问题，但成功有限。在本文中，通过应用简单的指南针并使用有效的变换，我们将非线性状态空间模型转化为线性模型。然后，通过将原始卡尔曼滤波器（KF）应用于该模型，我们提出了一种新方法，称为LMKF SLAM。我们展示了LMKF SLAM在准确性、收敛性和计算复杂性方面显著优于最先进的方法，尤其是基于EKF的SLAM。此外，所提出的方法在传感器值的不确定性和系统参数变化方面也更为稳定。实验结果验证了这些观点。
cs.RO / 7 / 2606.28476
FADA: Few-Shot Domain Adaptation via Dynamics Alignment for Humanoid Control
FADA：通过动态对齐实现的人形控制少样本领域适应
Abstract
High-precision humanoid control is limited by target-domain dynamics mismatch, where the same control objective can induce different realized motions under changes in terrain, payload, or actuator response. Existing methods either pursue zero-shot transfer through domain randomization or in-context adaptation without target-domain specialization, or require heavy adaptation pipelines that leverage target-domain data, such as model calibration, residual learning, or policy retraining. In this paper, we present FADA (Few-Shot Domain Adaptation via Dynamics Alignment), a three-stage Planner-Inverse Dynamics Model (Planner-IDM) framework for few-shot adaptation in humanoid control. FADA first trains an oracle policy with privileged information and then distills the oracle behavior into a deployable Planner-IDM student through DAgger. At deployment, FADA freezes the planner and finetunes only the IDM using approximately 2 minutes of target-domain rollouts with standard supervised learning. Rather than requiring optimal demonstrations or rewards, FADA uses the paired actions and observations that are observed during these rollouts as supervision, aligning the IDM's action generation with target-domain dynamics. Experiments show that FADA outperforms both in-context and end-to-end adaptation baselines, improving task performance under dynamics shifts and enabling real humanoid robots to execute diverse high-precision whole-body tasks. Implementation details and qualitative hardware rollout videos are available at https://lecar-lab.github.io/FADA-humanoid/.
Chinese Translation
高精度的人形控制受到目标领域动态不匹配的限制，同一控制目标在地形、负载或执行器响应变化下可能导致不同的实际运动。现有方法要么通过领域随机化追求零-shot 转移，或在没有目标领域专门化的情况下进行上下文适应，要么需要重型适应流程，这些流程利用目标领域数据，如模型校准、残差学习或策略再训练。本文提出了 FADA（通过动态对齐实现的少样本领域适应），这是一个用于人形控制的三阶段规划-逆动态模型（Planner-IDM）框架。FADA 首先训练一个具有特权信息的神谕策略，然后通过 DAgger 将神谕行为提炼为可部署的 Planner-IDM 学生。在部署时，FADA 冻结规划器，仅使用约 2 分钟的目标领域回放数据通过标准监督学习微调 IDM。FADA 不需要最佳演示或奖励，而是使用在这些回放过程中观察到的配对动作和观察作为监督，将 IDM 的动作生成与目标领域动态对齐。实验表明，FADA 在动态变化下的任务表现优于上下文适应和端到端适应基线，使真实的人形机器人能够执行多样的高精度全身任务。实现细节和定性硬件回放视频可在 https://lecar-lab.github.io/FADA-humanoid/ 获取。
cs.RO / 8 / 2606.28529
The Speedup Paradox: Rethinking Inference Speed-Quality Trade-off in Embodied Tasks
加速悖论：重新思考具身任务中的推理速度与质量的权衡
Abstract
Embodied foundation models have recently been widely used to improve robot generalization and task success rates. Previous works apply lossy efficient-inference techniques such as quantization, pruning, and asynchronous inference, accepting small action quality degradation in exchange for lower per-step computation cost and inter-action latency. However, unlike traditional static ML tasks, embodied tasks involve repeated interaction with the environment, and task-level performance is determined not only by per-step cost, but also by closed-loop effects unique to embodied execution, which remain insufficiently characterized in current efficient-inference studies. In this work, we propose TISED (\underline{T}ask-level \underline{I}nference \underline{S}peedup \underline{E}ffect \underline{D}ecomposition), an analytical framework that unifies diverse lossy inference optimization techniques and decomposes their effects on static and dynamic tasks, and uncovers some paradoxical effects on task-level performance: (1) on \textit{static tasks}, optimization sometimes can lengthen end-to-end per-task completion time even as per-step latency drops; (2) on \textit{dynamic tasks}, moderate lossy optimization can raise task success rate even above the baseline; and (3) the monotonicity and sweet-spot location of both effects can shift with hardware configuration. Together, our findings provide a new perspective on adapting inference optimization techniques to embodied tasks.
Chinese Translation
具身基础模型最近被广泛应用于提高机器人泛化能力和任务成功率。以往的研究采用了如量化、剪枝和异步推理等有损高效推理技术，接受小幅的动作质量下降，以换取更低的每步计算成本和交互延迟。然而，与传统的静态机器学习任务不同，具身任务涉及与环境的重复交互，任务级性能不仅由每步成本决定，还受到具身执行特有的闭环效应的影响，而这些效应在当前的高效推理研究中尚未得到充分表征。在本研究中，我们提出了TISED（任务级推理加速效应分解），这是一个统一多种有损推理优化技术的分析框架，分解其对静态和动态任务的影响，并揭示了一些对任务级性能的悖论效应：（1）在静态任务中，优化有时会延长每个任务的端到端完成时间，即使每步延迟下降；（2）在动态任务中，适度的有损优化可以使任务成功率甚至超过基线；（3）这两种效应的单调性和最佳点位置可能会随着硬件配置的变化而改变。我们的发现为将推理优化技术适应于具身任务提供了新的视角。
cs.RO / 9 / 2606.28555
Robotic Arm-Based Spectral Sensing for Strawberry Positioning and Non-Destructive Sweetness Measurement
基于机器人手臂的光谱传感技术用于草莓定位和非破坏性甜度测量
Abstract
Accurate assessment of sweetness is essential for quality control in agriculture, yet conventional methods rely on destructive sampling and are difficult to scale. This thesis presents a robotic arm-based spectral sensing system for strawberry detection, localization, approach, and non-destructive sweetness estimation. The system integrates perception, calibration, and robotic control in a closed-loop pipeline. A YOLOv11s detector is adopted for real-time strawberry detection, while RGB-ToF calibration and mask-to-depth alignment are used to obtain geometrically consistent target localization. A custom eye-in-hand hand-eye calibration workflow is developed to estimate the rigid transform between gripper_link and cam_front, enabling reliable transformation of fruit targets into the robot base frame. Based on these estimates, the robot executes a waypoint-based search and an incremental closed-loop approach strategy to position the sensor at optimal working distance for sweetness sensing. Experimental results show strong end-to-end performance (88.10% success over 42 trials), with robust detection (95.24%) and successful approach execution once a target is detected (100% conditional success). Hand-eye calibration comparisons indicate that although Andreff yields the smallest translation norm in single-run results, the Park method provides better cross-sample consistency and therefore more stable downstream robot behavior. The residual failures are concentrated in the sensing stage, especially valid-region extraction for sweetness estimation under difficult depth/reflectance conditions. Overall, this work demonstrates the feasibility of integrating RGB-ToF perception, robotic manipulation, and non-destructive sensing for practical strawberry quality assessment, and provides a scalable baseline for future integration of learning-based policies such as Vision-Language-Action models.
Chinese Translation
甜度的准确评估对于农业质量控制至关重要，但传统方法依赖于破坏性取样，且难以规模化。本文提出了一种基于机器人手臂的光谱传感系统，用于草莓的检测、定位、接近和非破坏性甜度估计。该系统在闭环管道中集成了感知、校准和机器人控制。采用YOLOv11s检测器进行实时草莓检测，同时使用RGB-ToF校准和掩膜到深度的对齐方法，以获得几何一致的目标定位。开发了一种定制的眼在手（eye-in-hand）手眼校准工作流程，以估计夹持器（gripper_link）与前置摄像头（cam_front）之间的刚性变换，从而实现水果目标到机器人基坐标系的可靠转换。基于这些估计，机器人执行基于航点的搜索和增量闭环接近策略，以将传感器定位在最佳工作距离进行甜度传感。实验结果显示出强大的端到端性能（42次试验中成功率为88.10%），具有稳健的检测能力（95.24%），并且一旦检测到目标，接近执行的成功率为100%。手眼校准比较表明，尽管Andreff方法在单次运行结果中产生了最小的平移范数，但Park方法提供了更好的跨样本一致性，因此使得下游机器人行为更加稳定。残余失败主要集中在传感阶段，尤其是在困难的深度/反射条件下进行甜度估计时的有效区域提取。总体而言，本研究展示了将RGB-ToF感知、机器人操作和非破坏性传感整合用于实际草莓质量评估的可行性，并为未来整合基于学习的策略（如视觉-语言-动作模型）提供了可扩展的基准。
cs.RO / 10 / 2606.28592
Embodiment Meets Environment: Toward Context-Aware, Safe Physical Caregiving Robots
体现与环境的结合：面向上下文感知的安全物理护理机器人
Abstract
Physical caregiving robots need to assist different users with different tasks in diverse environments, and they come in many embodiments. While substantial progress has been made on individual caregiving tasks, most existing systems remain tightly coupled to specific environments and robot embodiments, and often do not explicitly model or constrain interactions around people, despite humans being special agents in the environment. This motivates a focus on adapting to context that emerges from the joint interaction between the environment and the robot's embodiment. We propose $E^2$-CARE, a framework that enables context-aware adaptation by representing primitive caregiving skills as interaction templates whose execution is reshaped online. $E^2$-CARE represents the environment, the robot, and the human within a unified 3D dynamic scene graph that models these interaction contexts explicitly, and synthesizes task-specific constraints to govern how each skill is executed. By enforcing these constraints at runtime, the same skill templates can be reused zero-shot and safely across diverse environments and robot embodiments. We evaluate $E^2$-CARE across four activities of daily living in hundreds of simulated household environments, including assistive home settings, and across diverse robot embodiments, and validate it through user studies on two caregiving tasks with two robots in various real-world environments. Results demonstrate consistent and successful adaptation across these environments and embodiments. Website: https://emprise.cs.cornell.edu/e2care
Chinese Translation
物理护理机器人需要在多样化的环境中协助不同的用户完成不同的任务，并且它们有多种体现形式。尽管在个别护理任务上取得了显著进展，但大多数现有系统仍然与特定环境和机器人体现紧密耦合，且往往没有明确建模或约束围绕人类的交互，尽管人类在环境中是特殊的主体。这促使我们关注来自环境与机器人体现之间联合交互所产生的上下文适应。我们提出了 $E^2$-CARE 框架，通过将原始护理技能表示为交互模板，使其执行能够在线重塑，从而实现上下文感知的适应。$E^2$-CARE 在一个统一的三维动态场景图中表示环境、机器人和人类，明确建模这些交互上下文，并合成特定任务的约束来管理每项技能的执行。通过在运行时强制执行这些约束，相同的技能模板可以在不同的环境和机器人体现中零-shot 安全重用。我们在数百个模拟家庭环境中评估 $E^2$-CARE 的四项日常活动，包括辅助家庭环境，并在多样化的机器人体现中进行验证，通过用户研究在不同真实环境中对两个机器人进行两项护理任务的验证。结果表明，在这些环境和体现中实现了一致且成功的适应。网站：https://emprise.cs.cornell.edu/e2care
cs.RO / 11 / 2606.28607
Fast and Accurate Outlier-Aware LiDAR Super-Resolution for SLAM Applications
用于SLAM应用的快速准确的异常值感知LiDAR超分辨率
Abstract
This work tackles the challenge of enhancing low-resolution LiDAR sensors for SLAM applications through a novel Deep Unrolling-based Super-Resolution (SR) model. We integrate an outlier removal module to ensure structural integrity while maintaining real-time performance. By leveraging a model-based optimization approach, our method efficiently reconstructs high-resolution point clouds while minimizing computational overhead. The proposed SR model is evaluated within a LiDAR SLAM framework, demonstrating significant improvements in pose estimation accuracy and efficiency compared to state-of-the-art SR methods.
Chinese Translation
本研究针对通过一种新颖的基于深度展开的超分辨率（SR）模型来增强低分辨率LiDAR传感器在SLAM应用中的挑战。我们集成了一个异常值移除模块，以确保结构完整性，同时保持实时性能。通过利用基于模型的优化方法，我们的方法有效地重建高分辨率点云，同时最小化计算开销。所提出的SR模型在LiDAR SLAM框架内进行了评估，与最先进的SR方法相比，显示出在姿态估计精度和效率方面的显著提升。
cs.RO / 12 / 2606.28637
PinNet: Keypoint-Aware Learned Local Descriptors with Geometric Embedding for Loop Closure in LiDAR SLAM
PinNet：具有几何嵌入的关键点感知学习局部描述子在LiDAR SLAM中的回环闭合
Abstract
Loop closure is essential to reduce drift and build globally consistent maps in large-scale environments. However, reliable loop closure with only geometric information from, e.g., a LiDAR sensor, remains challenging due to the difficulty of constructing discriminative geometric features. We present PinNet, a neural network that produces local geometric descriptors from point clouds for place recognition and scanto-scan registration. PinNet incorporates a neural network that generates keypoints and their corresponding descriptors, together with a plane-based geometric self-attention module that models inter-keypoint spatial relationships to enhance descriptor discriminability for loop-closure detection and point-cloud registration. The approach is comprehensively evaluated on multiple datasets collected with different LiDAR sensors. Experimental results demonstrate strong place-recognition performance, precise relative pose estimation, and successful single-shot localization in different environments.
Chinese Translation
回环闭合对于减少漂移和在大规模环境中构建全局一致的地图至关重要。然而，仅依靠来自LiDAR传感器等的几何信息进行可靠的回环闭合仍然具有挑战性，因为构建具有区分性的几何特征非常困难。我们提出了PinNet，这是一种神经网络，可以从点云中生成局部几何描述子，用于位置识别和扫描到扫描的配准。PinNet结合了一个生成关键点及其对应描述子的神经网络，以及一个基于平面的几何自注意力模块，该模块建模关键点之间的空间关系，以增强描述子在回环闭合检测和点云配准中的区分能力。该方法在多个使用不同LiDAR传感器收集的数据集上进行了全面评估。实验结果表明，该方法在位置识别性能、精确的相对姿态估计和在不同环境中的单次定位方面表现出色。
cs.RO / 13 / 2606.28712
J-LAW: Joint Localization and Actionable World Modeling via Coupled Latent Factor Graphs
J-LAW：通过耦合潜在因子图进行联合定位和可操作世界建模
Abstract
Classical SLAM estimates metric poses and a geometric map but produces no actionable predictive model for planning. Action-conditioned world models learn compact latent dynamics for planning but ignore global metric consistency and accumulate drift under open-loop rollout. We argue these are two views of the same estimation problem and propose J-LAW (Joint Localization and Actionable World Modeling) in this letter: a coupled factor graph that jointly optimizes metric object poses, latent world states, and latent landmark embeddings. The bridge is a pose-conditioned latent encoder and a learned pose--latent coupling factor, so that better localization improves the world model and vice versa. We cast observation, action-conditioned prediction, metric odometry, pose--latent coupling, latent loop closure, and latent landmark observation as probabilistic factors in a single MAP objective. Real-data experiments on PushT and WildGS show that coupled graph correction substantially reduces latent prediction RMSE and endpoint drift relative to open-loop rollout, while latent loop closure improves global trajectory consistency. J-LAW yields a map that is simultaneously metric (poses) and actionable (latent landmarks for planning).
Chinese Translation
经典的SLAM（同步定位与地图构建）估计度量姿态和几何地图，但未能生成可用于规划的可操作预测模型。基于动作的世界模型学习紧凑的潜在动态以进行规划，但忽视了全局度量一致性，并在开放环路展开下累积漂移。我们认为这两者是同一估计问题的两个视角，并在本文中提出J-LAW（联合定位和可操作世界建模）：一个耦合因子图，联合优化度量物体姿态、潜在世界状态和潜在地标嵌入。其桥梁是一个基于姿态的潜在编码器和一个学习的姿态-潜在耦合因子，从而使得更好的定位改善世界模型，反之亦然。我们将观察、基于动作的预测、度量里程计、姿态-潜在耦合、潜在回环闭合和潜在地标观察视为单一MAP目标中的概率因子。在PushT和WildGS上的真实数据实验表明，耦合图校正显著降低了潜在预测的均方根误差（RMSE）和相对于开放环路展开的终点漂移，而潜在回环闭合提高了全局轨迹一致性。J-LAW生成的地图同时是度量的（姿态）和可操作的（用于规划的潜在地标）。
cs.RO / 14 / 2606.28720
CubifyGS: Object-Centric 3D Gaussian Splatting for Lifelong Dynamic Scene Maintenance
CubifyGS：面向物体的3D高斯喷溅在终身动态场景维护中的应用
Abstract
Lifelong scene mapping under rigid object rearrangement remains a fundamental challenge in robotics. While 3D Gaussian Splatting (3DGS) enables high-fidelity modeling, primitive-level updates often cause persistent ghosting and slow recovery. We propose CubifyGS, an object-level mapping framework that shifts dynamic maintenance from passive re-optimization to active asset management. CubifyGS models movable instances as reusable Gaussian assets, detects object appearance and disappearance, and updates maps through asset retrieval, rigid transformation, and explicit pruning rather than reconstruction from scratch. To address geometric voids and local photometric mismatch after such edits, we further propose an event-triggered adaptive optimization strategy that focuses computation on affected regions. We validate our approach on a newly constructed high-fidelity dynamic benchmark, demonstrating that CubifyGS improves artifact suppression and maintenance efficiency over representative reproducible baselines in the evaluated object-rearrangement setting.
Chinese Translation
在刚性物体重新排列下，终身场景映射仍然是机器人技术中的一项基本挑战。尽管3D高斯喷溅（3DGS）能够实现高保真建模，但原始级别的更新常常导致持续的鬼影现象和缓慢的恢复。我们提出了CubifyGS，一个物体级别的映射框架，将动态维护从被动的重新优化转变为主动的资产管理。CubifyGS将可移动实例建模为可重用的高斯资产，检测物体的出现和消失，并通过资产检索、刚性变换和显式修剪而非从头重建来更新地图。为了解决此类编辑后出现的几何空洞和局部光度不匹配问题，我们进一步提出了一种事件触发的自适应优化策略，重点计算受影响区域。我们在新构建的高保真动态基准上验证了我们的方法，结果表明，CubifyGS在评估的物体重新排列设置中，相较于代表性的可重复基线，改善了伪影抑制和维护效率。
cs.RO / 15 / 2606.28746
He3-Seeker: Robotic Information Planning for Lunar Helium-3 Distribution Mapping
He3-Seeker：月球氦-3分布映射的机器人信息规划
Abstract
Lunar helium-3 is a highly valuable strategic resource, pivotal to the advancement of both deep-space exploration and space mining. Existing lunar helium-3 exploration methodologies rely primarily on indirect measurements via remote sensing, which are often characterized by limited precision, low reliability, and insufficient spatial resolution. In this paper, we introduce He3-Seeker, an active robotic exploration method for helium-3 distribution mapping. First, we provide a formal definition of the active helium-3 exploration problem. Subsequently, we developed the He3-Seeker framework, which is conceptually based on multi-point drilling, sampling, and in situ analysis. In particular, we use robotic information planning (RIP) to guide autonomous robot navigation and active sensing. Additionally, to thoroughly evaluate the proposed algorithm, we introduce a reliable method for generating reference data of lunar helium-3 distribution based on low-resolution orbital remote sensing measurements. Simulation experiments verify that He3-Seeker achieves both rapid and high-fidelity mapping of helium-3 distribution, providing a reliable solution for resource exploration tasks. Our code and simulation environment will be publicly accessible at https://github.com/OpenSpace-Lab/He3-Seeker.
Chinese Translation
月球氦-3是一种极具价值的战略资源，对于深空探索和太空采矿的推进至关重要。现有的月球氦-3探索方法主要依赖于通过遥感进行的间接测量，这些测量通常具有精度有限、可靠性低和空间分辨率不足的特点。本文介绍了He3-Seeker，一种用于氦-3分布映射的主动机器人探索方法。首先，我们对主动氦-3探索问题进行了正式定义。随后，我们开发了He3-Seeker框架，该框架在概念上基于多点钻探、采样和原位分析。特别地，我们使用机器人信息规划（Robotic Information Planning, RIP）来指导自主机器人导航和主动传感。此外，为了全面评估所提出的算法，我们引入了一种基于低分辨率轨道遥感测量生成月球氦-3分布参考数据的可靠方法。仿真实验验证了He3-Seeker能够快速且高保真地绘制氦-3分布，为资源探索任务提供了可靠的解决方案。我们的代码和仿真环境将公开访问，网址为 https://github.com/OpenSpace-Lab/He3-Seeker。
cs.RO / 16 / 2606.28760
Vision-Language Models for Deployable Social Robot Navigation: Bridging Semantic Reasoning and Low-Level Control
可部署社交机器人导航的视觉-语言模型：桥接语义推理与低级控制
Abstract
Social robot navigation (SRN) requires more than geometric path planning; it demands understanding human intentions, social norms, and contextual cues to generate socially compliant behaviors. Although classical navigation methods provide reliable metric planning and collision avoidance, they often lack the semantic reasoning capabilities necessary for operation in complex human-centered environments. Recent advances in Vision-Language Models (VLMs) have opened new opportunities for SRN by enabling high-level VLM understanding, commonsense reasoning, and natural language interaction. However, a fundamental challenge remains: how to integrate VLMs into real-time, safety-critical navigation systems and reliably translate their high-level reasoning into grounded navigation actions. In this survey, we present a unified perspective of VLM-based SRN and organize existing approaches into three interconnected components: high-level VLM reasoning, low-level planning and control, and intermediate mechanisms that bridge reasoning and action. Based on this perspective, we propose a structured roadmap for coupling VLMs with navigation systems, covering semantic reasoning, evaluators, spatial grounding, intermediate representations, and control modules. The roadmap highlights both the strengths of VLMs and the necessity of hybrid architectures for practical deployment. We further review representative datasets and evaluation platforms developed for SRN. Finally, we discuss key open challenges. This survey aims to provide a foundation for building reliable, socially compliant, and deployable VLM-enabled navigation systems.
Chinese Translation
社交机器人导航（SRN）不仅仅需要几何路径规划；它还要求理解人类意图、社会规范和上下文线索，以生成符合社会规范的行为。尽管经典导航方法提供了可靠的度量规划和碰撞避免，但它们往往缺乏在复杂人本环境中操作所需的语义推理能力。近期视觉-语言模型（VLMs）的进展为SRN开辟了新的机会，使其能够实现高层次的VLM理解、常识推理和自然语言交互。然而，一个根本性挑战依然存在：如何将VLM集成到实时的安全关键导航系统中，并可靠地将其高层次推理转化为具体的导航行动。在本次综述中，我们提出了基于VLM的SRN的统一视角，并将现有方法组织为三个相互关联的组成部分：高层次的VLM推理、低层次的规划与控制，以及桥接推理与行动的中介机制。基于这一视角，我们提出了一个结构化的路线图，以将VLM与导航系统结合，涵盖语义推理、评估器、空间定位、中介表示和控制模块。该路线图突出了VLM的优势以及实际部署中混合架构的必要性。我们进一步回顾了为SRN开发的代表性数据集和评估平台。最后，我们讨论了关键的开放挑战。本次综述旨在为构建可靠、符合社会规范且可部署的VLM支持的导航系统提供基础。
cs.RO / 17 / 2606.28805
Physics Models for Sim-to-Real Transfer in Professional-Level Robot Table Tennis
专业级机器人乒乓球中的模拟到现实转移的物理模型
Abstract
At competitive speeds and spins, a table tennis ball follows complex, counterintuitive trajectories that a robot must track and precisely counter within fractions of a second. Training a reinforcement learning policy capable of these skills is prohibitively expensive and dangerous in the real world, making high-fidelity simulation essential. Transferability of such policies, however, critically depends on how faithfully the simulation captures real-world dynamics--a requirement made even more stringent by the adversarial nature of the game, where any regime in which a model fails to approximate reality becomes an exploitable weakness for the opponent. Prior state-of-the-art in robot table tennis generally focuses on a limited range of velocities and spins and fails to capture the richness of ball behaviors encountered in professional-level play. In this work, we present physics models for the aerodynamic ball flight, for the contact dynamics between the ball and the table, as well as between the ball and the racket that accurately capture the ball behavior over a vast range of speeds and spins relevant to the game. Specifically, we model drag and Magnus force coefficients as functions of Reynolds number and spin ratio in the aerodynamics equations. For the table contact model we model effects of ball buckling on the coefficient of restitution and incorporate residuals into the instantaneous point-contact models. For the racket contact model we introduce a residual neural network component to complement coefficients related to normal and tangential coefficients of restitution as well as torsional spin damping. The resulting models were used for the first real-world robot table tennis AI agent capable of competing against professional players, to train reinforcement learning policies.
Chinese Translation
在竞争性的速度和旋转下，乒乓球遵循复杂且反直觉的轨迹，机器人必须在几分之一秒内跟踪并精确反击。训练能够掌握这些技能的强化学习策略在现实世界中代价高昂且危险，因此高保真模拟至关重要。然而，这些策略的可转移性在很大程度上依赖于模拟对现实世界动态的忠实捕捉——这一要求在游戏的对抗性质下变得更加严格，因为任何模型未能逼近现实的情况都可能成为对手可以利用的弱点。以往的机器人乒乓球研究通常集中在有限的速度和旋转范围内，未能捕捉到专业级比赛中球的丰富行为。在本研究中，我们提出了用于气动球飞行的物理模型、球与球台之间的接触动力学模型，以及球与球拍之间的接触模型，这些模型准确捕捉了与比赛相关的广泛速度和旋转下的球行为。具体而言，我们将阻力和马格努斯力系数建模为雷诺数和旋转比的函数。在球台接触模型中，我们对球的屈曲效应对恢复系数的影响进行了建模，并将残余效应纳入瞬时点接触模型中。在球拍接触模型中，我们引入了残差神经网络组件，以补充与法向和切向恢复系数以及扭转旋转阻尼相关的系数。所得到的模型被用于首个能够与职业选手竞争的现实世界机器人乒乓球人工智能代理，以训练强化学习策略。
cs.RO / 18 / 2606.28813
Human2Any: Human-to-Robot Transfer via Constraint-Aware Compositional Planning
Human2Any：通过约束感知组合规划实现人机转移
Abstract
Human videos are a scalable source of supervision for robot manipulation, as they are abundant and naturally capture rich object interactions. However, transferring human demonstrations to robots remains challenging due to embodiment mismatch, scene variation, and robot-specific feasibility constraints. We present Human2Any, a framework for learning reusable object-centric interaction priors from human videos without requiring real-world robot demonstrations in the target task contexts. Human2Any represents manipulation through object-object interaction motion, capturing task-relevant scene changes while abstracting away embodiment-specific details. It composes learned interaction priors with robot-side feasibility reasoning and motion planning, allowing the same human-derived knowledge to adapt to different embodiments, scene geometries, and task contexts. We validate Human2Any across diverse manipulation settings, including real-world experiments on a Franka tabletop setup and an RBY-1 humanoid mobile robot, demonstrating robust interaction-centric manipulation without real-world robot training data. Project website: https://human2any.github.io/.
Chinese Translation
人类视频是机器人操作的一种可扩展的监督来源，因为它们丰富且自然地捕捉了对象交互。然而，由于体现不匹配、场景变化和机器人特定的可行性约束，将人类演示转移到机器人上仍然具有挑战性。我们提出了Human2Any，一个从人类视频中学习可重用的以对象为中心的交互先验的框架，无需在目标任务上下文中进行真实世界的机器人演示。Human2Any通过对象间交互运动来表示操作，捕捉与任务相关的场景变化，同时抽象掉体现特定的细节。它将学习到的交互先验与机器人端的可行性推理和运动规划相结合，使得相同的人类衍生知识能够适应不同的体现、场景几何和任务上下文。我们在多种操作设置中验证了Human2Any，包括在Franka桌面设置和RBY-1类人移动机器人上的真实世界实验，展示了在没有真实世界机器人训练数据的情况下，稳健的以交互为中心的操作。项目网站：https://human2any.github.io/
cs.RO / 19 / 2606.28827
LNN-Fly: Continuous-Time UAV Navigation for Robust Obstacle Avoidance under Timing Mismatch
LNN-Fly：针对时序不匹配的鲁棒障碍规避的连续时间无人机导航
Abstract
End-to-end unmanned aerial vehicle (UAV) navigation can achieve impressive agility in simulation, yet its obstacle-avoidance behavior often degrades after deployment because the policy must tolerate simulator mismatch, sensing irregularity, and variable-rate control. These effects are especially dangerous in cluttered environments, where stale observations or short control irregularities can directly lead to collisions. We present LNN-Fly, a deployment-oriented continuous-time navigation policy for LiDAR-based UAV obstacle avoidance. The policy combines a dynamic-programming-inspired structured recurrent update, explicit conditioning on the elapsed control interval {\Delta}t, and an input-driven adaptive forgetting gate that refreshes stale latent state near hazards while preserving consistency during sustained maneuvers. It is trained with differentiable rollouts that incorporate deployment-relevant sensing and timing perturbations. In simulation, LNN-Fly improves obstacle-avoidance performance in the tested settings and shows better tolerance to reduced control frequency, sparse observations, and control-period jitter. It also transfers zero-shot from a simplified differentiable simulator to a physical quadrotor. In indoor cross-frequency real-world tests, the system achieves 100% success over 20 flights, while policy inference has a median latency of 0.514 ms on a desktop graphics processing unit (GPU) and about 2.5 ms on the onboard central processing unit (CPU), with onboard P95 latency below 30 ms.
Chinese Translation
端到端无人机（UAV）导航在仿真中可以实现令人印象深刻的灵活性，然而其障碍规避行为在部署后往往会下降，因为该策略必须容忍模拟器不匹配、传感器不规则性和可变速率控制。这些影响在杂乱环境中尤其危险，在这些环境中，过时的观测或短暂的控制不规则性可能直接导致碰撞。我们提出了LNN-Fly，一种面向部署的基于激光雷达（LiDAR）的无人机障碍规避的连续时间导航策略。该策略结合了受动态规划启发的结构化递归更新、对经过控制间隔{ ext{Δ}t}的显式条件以及一个输入驱动的自适应遗忘门，该遗忘门在接近危险时刷新过时的潜在状态，同时在持续机动期间保持一致性。它通过可微分的回放进行训练，回放中包含与部署相关的传感和时序扰动。在仿真中，LNN-Fly在测试环境中提高了障碍规避性能，并表现出对降低控制频率、稀疏观测和控制周期抖动的更好容忍度。它还可以从简化的可微分模拟器零-shot迁移到物理四旋翼。在室内跨频率的真实世界测试中，该系统在20次飞行中实现了100%的成功率，而策略推理在桌面图形处理单元（GPU）上的中位延迟为0.514毫秒，在机载中央处理单元（CPU）上的延迟约为2.5毫秒，机载P95延迟低于30毫秒。
cs.RO / 20 / 2606.28899
You Only Touch Once: 6-DoF Object Pose Estimation from Single Tactile Contact
一次触摸即可：基于单次触觉接触的六自由度物体姿态估计
Abstract
Accurate 6-DoF object pose estimation is fundamental to robotic manipulation, yet vision-based methods often fail under occlusion, poor lighting, and reflective or transparent surfaces. We present YOTO, a tactile-only pose estimation system that recovers the full 6-DoF object pose from a single pair of simultaneous contacts, without requiring contact history. YOTO represents each tactile contact as a local 3D point cloud and localizes it on the object surface through a coarse-to-fine network. The two localized contacts, together with the calibrated sensor poses, are then fed to a closed-form normal-aware SVD solver that recovers the full 6-DoF object pose in one step. To reduce real-data requirements, the localization network is pretrained on virtual tactile patches sampled from the object model and fine-tuned with a small number of real contacts. We further show that YOTO can operate on object models reconstructed from consumer-grade mobile scans, and quantify the gap relative to CAD-based models. Experiments on four geometrically diverse objects demonstrate accurate tactile contact localization and pose estimation, outperforming vision-based and geometric baselines, especially when visual perception is unreliable. Code, trained models, and the real GelSight dataset will be released upon publication.
Chinese Translation
准确的六自由度物体姿态估计是机器人操作的基础，但基于视觉的方法在遮挡、光照不足以及反射或透明表面下常常失效。我们提出了YOTO，一个仅基于触觉的姿态估计系统，它能够从一对同时接触中恢复完整的六自由度物体姿态，而无需接触历史。YOTO将每个触觉接触表示为局部三维点云，并通过粗到细的网络在物体表面进行定位。然后，将这两个定位的接触点与校准的传感器姿态一起输入到一个闭式形式的法线感知奇异值分解(SVD)求解器中，以一步恢复完整的六自由度物体姿态。为了减少对真实数据的需求，定位网络在从物体模型中采样的虚拟触觉贴片上进行预训练，并通过少量真实接触进行微调。我们进一步展示了YOTO可以在从消费级移动扫描重建的物体模型上运行，并量化与基于CAD模型的差距。在四个几何形状各异的物体上的实验表明，触觉接触定位和姿态估计准确，超越了基于视觉和几何的基线，尤其是在视觉感知不可靠的情况下。代码、训练模型和真实的GelSight数据集将在发表后发布。
cs.RO / 21 / 2606.28951
Cross-Session 3D LiDAR and Camera Fusion for Robust Localization of Unmanned Aerial Vehicles in GPS-Denied Environments
跨会话3D激光雷达与相机融合用于GPS信号缺失环境中无人机的稳健定位
Abstract
Accurate localization of unmanned aerial vehicles (UAVs) is essential for applications such as structural health monitoring, especially in environments where Global Positioning System (GPS) signals are denied or unreliable, like indoor spaces, tunnels, urban canyons, or areas beneath large structures. To address this challenge, we propose Cross-Fusion, a novel method for real-time UAV localization that integrates data from a 3D Light Detection and Ranging (LiDAR) and a monocular camera. A key contribution is its cross-session fusion strategy, which integrates visual and geometric information collected from multiple agents during routine baseline surveys to improve localization consistency and map completeness. The system employs LiDAR-based odometry for motion tracking and image-based feature matching via a single red-green-blue (RGB) camera to correct drift and improve accuracy. Unlike visual-inertial systems, Cross-Fusion maintains a simple sensor setup and avoids the complexity of stereo or global shutter configurations. Experimental results demonstrate that Cross-Fusion achieves localization accuracy comparable to GPS-based methods and performs reliably in challenging feature-sparse environments.
Chinese Translation
无人机（UAV）的准确定位对于结构健康监测等应用至关重要，尤其是在全球定位系统（GPS）信号被拒绝或不可靠的环境中，如室内空间、隧道、城市峡谷或大型结构下方区域。为了解决这一挑战，我们提出了一种名为Cross-Fusion的新方法，用于实时无人机定位，该方法整合了3D光学测距（LiDAR）和单目相机的数据。其关键贡献在于跨会话融合策略，该策略整合了在常规基线调查中多个代理收集的视觉和几何信息，以提高定位一致性和地图完整性。该系统采用基于LiDAR的里程计进行运动跟踪，并通过单个红绿蓝（RGB）相机进行图像特征匹配，以纠正漂移并提高准确性。与视觉惯性系统不同，Cross-Fusion保持了简单的传感器设置，避免了立体或全局快门配置的复杂性。实验结果表明，Cross-Fusion实现了与基于GPS的方法相当的定位精度，并在特征稀疏的挑战性环境中表现出可靠性。
cs.RO / 22 / 2606.28995
HJ-SafeDMP: Hamilton-Jacobi Reachability-Guided Dynamic Movement Primitives for Provably Safe Robot Motion
HJ-SafeDMP：基于哈密顿-雅可比可达性指导的动态运动原语以确保机器人运动的可证明安全性
Abstract
Robots deployed in safety-critical environments must execute motions that are simultaneously robust to disturbances and provably safe from collisions. Dynamic Movement Primitives (DMPs) offer inherent stability, temporal flexibility, and efficient trajectory generalization from single demonstrations, but they lack formal safety certificates. Conversely, Hamilton-Jacobi (HJ) Reachability analysis provides a principled framework for computing worst-case safety margins and forward-invariant safe sets, but classical grid-based methods suffer from the curse of dimensionality and are impractical for real-time control. This paper introduces HJ-SafeDMP, a framework that integrates DMPs with learned HJ Reachability-based safety value functions to achieve provably safe, robust, and computationally efficient robot motion. We learn a Control Barrier Value Function (CBVF) from offline demonstration data using a model-free, finite-difference HJ recursion and deploy it as a real-time safety filter via a closed-form control law that modulates the DMP output. Unlike optimization-based CBF-QP approaches, our method achieves safety filtering without online quadratic program solves, preserving the computational efficiency of DMPs. We further incorporate an expectile-based offline learning objective that avoids querying out-of-distribution actions, and a conformal prediction calibration step that provides finite-sample probabilistic safety coverage. Experimental evaluation on a 7-DOF robot manipulator demonstrates that HJ-SafeDMP achieves formal safety guarantees with orders-of-magnitude faster execution than optimization-based baselines, while maintaining the robustness and adaptability of DMPs for human-robot interaction.
Chinese Translation
部署在安全关键环境中的机器人必须执行既能抵御干扰又能确保碰撞安全的运动。动态运动原语（Dynamic Movement Primitives, DMPs）提供了固有的稳定性、时间灵活性以及从单一演示中高效的轨迹泛化，但缺乏正式的安全证书。相反，哈密顿-雅可比（Hamilton-Jacobi, HJ）可达性分析提供了一个原则性框架，用于计算最坏情况下的安全边际和前向不变安全集，但经典的基于网格的方法受到维度诅咒的影响，难以在实时控制中应用。本文介绍了HJ-SafeDMP，一个将DMP与基于学习的HJ可达性安全值函数相结合的框架，以实现可证明安全、稳健且计算高效的机器人运动。我们使用无模型的有限差分HJ递归从离线演示数据中学习控制障碍值函数（Control Barrier Value Function, CBVF），并通过一个封闭形式的控制律将其作为实时安全过滤器，调节DMP输出。与基于优化的CBF-QP方法不同，我们的方法在不进行在线二次规划求解的情况下实现了安全过滤，保持了DMP的计算效率。我们进一步结合了一种基于期望值的离线学习目标，避免查询分布外的动作，以及一个符合预测的校准步骤，提供有限样本的概率安全覆盖。在一个7自由度的机器人操纵器上的实验评估表明，HJ-SafeDMP实现了正式的安全保证，其执行速度比基于优化的基线快几个数量级，同时保持了DMP在人与机器人交互中的稳健性和适应性。
cs.RO / 23 / 2606.29028
Keypose Exploration: Efficient Automatic Trajectory Labelling and Cross-Embodiment Policy Transfer
关键姿态探索：高效自动轨迹标注与跨体态策略转移
Abstract
Keypose-based manipulation decomposes tasks into critical waypoints to simplify policy learning for long-horizon tasks, but existing approaches rely on task-specific heuristics or manual annotation to extract keyposes from demonstrations. We present an automatic trajectory labelling pipeline for grasp-related tasks. This pipeline combines vision-language models (VLMs) for semantic event detection with classical trajectory analysis for precise temporal alignment, requiring VLM inference only on one single demo among repeating ones per task. Using the labelled data, we train a keypose-guided Diffusion Policy (DP) that exploits keypose conditioning to intervene demonstration distributions. We explore the possibility to apply this property for cross-embodiment transfer: candidate keyposes are sampled and filtered via a reachability map, steering the policy toward kinematically feasible keyposes for the target robot. As a preliminary feasibility study, experiments on two robomimic tasks show that the labelled data produces policies matching a standard DP baseline, and that reachability-filtered keypose conditioning may benefit zero-shot transfer on the multimodal insertion task when feasible candidates are available.
Chinese Translation
基于关键姿态的操作将任务分解为关键的路径点，以简化长时间跨度任务的策略学习，但现有方法依赖于特定任务的启发式方法或手动注释来从演示中提取关键姿态。我们提出了一种用于抓取相关任务的自动轨迹标注管道。该管道结合了视觉-语言模型（VLMs）进行语义事件检测与经典轨迹分析以实现精确的时间对齐，仅需在每个任务的重复演示中对单个演示进行VLM推理。利用标注的数据，我们训练了一种基于关键姿态的扩散策略（Diffusion Policy, DP），该策略利用关键姿态条件来干预演示分布。我们探索了将这一特性应用于跨体态转移的可能性：候选关键姿态通过可达性图进行采样和过滤，引导策略朝向目标机器人可运动的关键姿态。作为初步可行性研究，在两个仿机器人任务上的实验表明，标注的数据生成的策略与标准DP基线相匹配，并且在可行候选可用时，可达性过滤的关键姿态条件可能有利于多模态插入任务的零-shot 转移。
cs.RO / 24 / 2606.29065
A Unified Framework for Multi-Contact Path Planning in the Rolling Robot Systems
滚动机器人系统中的多接触路径规划统一框架
Abstract
Rolling motion planning is challenging because rolling contact imposes nonholonomic constraints and the configuration evolves on a curved manifold. The problem becomes substantially harder in multi-contact settings, where multiple bodies roll without slip and the contact states are coupled. This paper presents a new framework for multi-contact path planning in spherical rolling robotics under no-slip constraints. We first derive a compact kinematic model for multi-sphere rolling using Montana's contact-coordinate formulation, where each contact is represented by a stacked five-state vector. Building on this model, we construct a Voronoi-based roadmap directly on the spherical contact manifold, incorporating spherical-cap obstacles and mutual-exclusion regions via on-manifold collision checking, and refine discrete graph paths using manifold-consistent log-exp smoothing. The resulting smoothed surface paths are then lifted to admissible multi-contact rolling motions through the derived Montana kinematics and validated via forward simulation. We further evaluate feasibility and path quality versus trajectory smoothness, Voronoi seed density, and computation time. The proposed framework provides a foundation for extending the method to non-spherical geometries, time-varying obstacle environments, and experimental validation on physical rolling robotic platforms.
Chinese Translation
滚动运动规划具有挑战性，因为滚动接触施加了非完整约束，并且配置在曲面流形上演变。在多接触环境中，问题变得更加复杂，其中多个物体在无滑动的情况下滚动，且接触状态相互耦合。本文提出了一种新的框架，用于在无滑动约束下的球形滚动机器人中的多接触路径规划。我们首先使用Montana的接触坐标公式推导出多球滚动的紧凑运动学模型，其中每个接触由一个堆叠的五状态向量表示。在此模型的基础上，我们直接在球形接触流形上构建了一个基于Voronoi的路线图，通过流形上的碰撞检测结合球冠障碍物和互斥区域，并使用流形一致的对数-指数平滑来优化离散图路径。然后，将得到的平滑表面路径通过推导的Montana运动学提升为可接受的多接触滚动运动，并通过前向仿真进行验证。我们进一步评估了路径的可行性和质量与轨迹平滑度、Voronoi种子密度和计算时间之间的关系。所提出的框架为将该方法扩展到非球形几何、时间变化障碍环境以及在物理滚动机器人平台上的实验验证提供了基础。
cs.RO / 25 / 2606.29089
TAP-VLA: Tactile Annotation Prompting for Vision Language Action Models
TAP-VLA：用于视觉语言动作模型的触觉注释提示
Abstract
Vision-Language-Action (VLA) models demonstrate impressive reasoning over visual, semantic, and spatial task variations by leveraging large-scale vision and language pre-training. They remain, however, largely blind to contact forces, which seldom manifest clearly in visual feedback but are central to contact-rich manipulation. Tactile sensing measures these forces directly, but integrating it into VLAs is difficult: tactile data is absent from the large-scale corpora used to pre-train VLAs, so adding it as a new input modality induces a distribution shift that erodes the very pre-training that makes VLAs effective. We propose Tactile Annotation Prompting for Vision-Language-Action models (TAP-VLA), a simple framework that supplies tactile feedback through visual augmentation rather than architectural change. TAP-VLA extracts shear fields from visuo-tactile sensors and overlays them as spatially-grounded vectors onto the multi-view RGB images the policy already consumes, yielding a clear, interpretable tactile cue in the VLA's native observation space. Because the architecture is untouched, the approach requires no tactile pre-training, adds negligible compute, and stays close to the pre-training distribution. Across four contact-rich tasks, TAP-VLA succeeds on 78% of trials, compared to under 50% for vision-only fine-tuning and alternative tactile-fusion baselines -- including tasks where the baselines perform no better than chance.
Chinese Translation
视觉-语言-动作（VLA）模型通过利用大规模的视觉和语言预训练，在视觉、语义和空间任务变体上展现出令人印象深刻的推理能力。然而，它们在接触力方面仍然基本上是盲目的，这些力在视觉反馈中很少清晰显现，但对接触丰富的操作至关重要。触觉传感能够直接测量这些力，但将其整合到VLA中是困难的：触觉数据在用于预训练VLA的大规模语料库中缺失，因此将其作为新的输入模态添加会导致分布转移，从而削弱了使VLA有效的预训练。我们提出了用于视觉-语言-动作模型的触觉注释提示（TAP-VLA），这是一个简单的框架，通过视觉增强而非架构变更提供触觉反馈。TAP-VLA 从视-触觉传感器中提取剪切场，并将其作为空间定位的向量叠加到策略已经使用的多视角RGB图像上，从而在VLA的原生观察空间中产生清晰、可解释的触觉线索。由于架构未受影响，该方法不需要触觉预训练，增加的计算量微不足道，并且保持接近预训练分布。在四个接触丰富的任务中，TAP-VLA在78%的试验中成功，而仅有50%以下的视觉单一微调和替代触觉融合基线成功，包括基线在某些任务中的表现甚至不如随机猜测。
cs.RO / 26 / 2606.29115
When Stopping Fails: Rethinking Minimal Risk Conditions through Human-Interactive Autonomous Driving for Safe Transportation Systems
当停止失败时：通过人机交互自主驾驶重新思考最小风险条件以实现安全交通系统
Abstract
Autonomous vehicles (AVs) are increasingly deployed in urban environments, yet their safety frameworks remain primarily designed around collision avoidance and minimal risk condition (MRC) behaviors such as slowing or stopping when uncertainty arises. Although effective in reducing immediate crash risk, real-world deployments indicate that stopping alone does not guarantee safe integration into human-governed roadway systems. Incidents reported by municipalities and public records show that AV fallback behaviors can obstruct traffic, interfere with emergency response operations, and create accessibility challenges for passengers and pedestrians. This paper presents an analysis of publicly documented incidents involving AV stopping behavior and human-AV interaction failures. We categorize these incidents according to limitations in perception, planning, and control within current AV architectures. Using this taxonomy, we identify key gaps in existing safety paradigms, particularly the lack of mechanisms for interpreting human authority, responding to multimodal instructions, and adapting to dynamic, socially regulated traffic conditions. We then review emerging research directions that support human-interactive perception, language-grounded and accessibility-aware planning, and assisted control through remote guidance and teleoperation. The analysis highlights the need to augment current AV safety frameworks with capabilities that enable cooperative interaction with human agents and infrastructure. These findings suggest that reliable urban deployment of AVs requires moving beyond passive fallback strategies toward human-interactive autonomy.
Chinese Translation
自主车辆（AVs）在城市环境中的部署日益增加，但其安全框架主要围绕碰撞避免和最小风险条件（MRC）行为设计，例如在不确定性出现时减速或停车。尽管在降低即时碰撞风险方面有效，现实世界的部署表明，仅靠停车并不能保证安全地融入人类管理的道路系统。市政当局和公共记录报告的事件显示，AV的后备行为可能会阻碍交通，干扰紧急响应操作，并给乘客和行人带来无障碍挑战。本文分析了涉及AV停车行为和人机交互失败的公开记录事件。我们根据当前AV架构中感知、规划和控制的局限性对这些事件进行了分类。利用这一分类法，我们识别出现有安全范式中的关键缺口，特别是缺乏解释人类权威、响应多模态指令以及适应动态社会调节交通条件的机制。随后，我们回顾了支持人机交互感知、基于语言的规划和关注无障碍的辅助控制的前沿研究方向。分析强调了增强当前AV安全框架的必要性，以实现与人类代理和基础设施的协作互动。这些发现表明，可靠的城市AV部署需要超越被动的后备策略，朝着人机交互自主性发展。
cs.RO / 27 / 2606.29123
On the Identifiability of Aided Inertial Navigation Under Measurement Delays: A Geometric Approach
测量延迟下辅助惯性导航的可识别性：一种几何方法
Abstract
In aided inertial navigation, measurements from different sensors are often subject to unknown relative time delays. Consider a single aiding sensor whose measurements have an unknown but constant delay relative to the inertial-measurement data stream. We study the identifiability of the delay and the initial navigation state that parameterizes the trajectory. Identifiability depends on both the temporal structure of the aiding measurements and the form of the trajectory itself. Our geometric analysis shows that, for a larger class of uninformative (i.e., degenerate) trajectories than has previously been reported, the delayed measurement model admits a continuous symmetry that prevents unique delay-and-state recovery.
Chinese Translation
在辅助惯性导航中，来自不同传感器的测量通常受到未知的相对时间延迟的影响。考虑一个单一的辅助传感器，其测量相对于惯性测量数据流具有未知但恒定的延迟。我们研究了延迟和参数化轨迹的初始导航状态的可识别性。可识别性依赖于辅助测量的时间结构以及轨迹本身的形式。我们的几何分析表明，对于比以往报道的更大类的不具信息性（即退化）轨迹，延迟测量模型承认一种连续对称性，这阻碍了延迟和状态的唯一恢复。
cs.RO / 28 / 2606.29165
Multi-Contact Force Estimation for Continuum Robots via Gaussian-Parameterized Factor Graphs
基于高斯参数化因子图的连续机器人多接触力估计
Abstract
Continuum robots offer key advantages in navigating unstructured environments, but their safe operation requires accurate estimation of the external contact forces acting anywhere along the robot body. Estimating these forces at unknown locations is an ill-conditioned problem, particularly for multiple contacts. We propose a unified shape and force estimation framework formulated on a factor graph. By incorporating a Gaussian mixture force parameterization into a discretized probabilistic Cosserat rod model, we reduce the dimensionality of the unknown external forces and mitigate the ill-conditioning of node-wise force estimation. The framework fuses strain, tendon tension, and pose measurements to simultaneously estimate the robot's shape and external forces while accounting for modeling and sensor uncertainties. Numerical simulations demonstrate that the proposed method outperforms existing methods in terms of force location and magnitude estimation for both single and multi-contact scenarios. We further present a progressive variant that introduces basis functions on demand to estimate contact forces sequentially during a simulated confined-navigation task.
Chinese Translation
连续机器人在非结构化环境中具有重要优势，但其安全操作需要准确估计作用于机器人身体任意位置的外部接触力。在未知位置估计这些力是一个病态问题，尤其是在多接触情况下。我们提出了一个统一的形状和力估计框架，该框架基于因子图进行构建。通过将高斯混合力参数化引入离散化的概率Cosserat杆模型，我们降低了未知外部力的维度，并减轻了节点级力估计的病态性。该框架融合了应变、腱张力和姿态测量，以同时估计机器人的形状和外部力，同时考虑建模和传感器的不确定性。数值仿真表明，所提出的方法在单接触和多接触场景中，在力的位置和大小估计方面优于现有方法。我们进一步提出了一种渐进变体，该变体根据需要引入基函数，以在模拟的受限导航任务中顺序估计接触力。
cs.RO / 29 / 2606.29173
TacGen: Touch Is a Necessary Dimension of Physical-World Representation -- Addressing Tactile Data Scarcity with Scalable Vision-to-Touch Alignment and Generation
TacGen：触觉是物理世界表征的必要维度——通过可扩展的视觉-触觉对齐与生成解决触觉数据稀缺问题
Ye, Wanghao, Das, Aarosh, Chen, Sihan, Wang, Yiting, Tian, Bowei, Sun, Guoheng, He, Shwai, Shen, Zheyu, Wang, Ziyao, He, Yexiao, Liu, Zhaoyi, Liu, Meng, Zhang, Yuning, Feng, Meng, Wang, Ziyi, Dai, Yilong, Dong, Yifei, Peng, Siyuan, Duan, Zhenle, Liu, Joshua, Xiong, Lang, Li, Ang
Abstract
Touch resolves the physical-property ambiguity left by vision: exploratory contact recovers shape, texture, compliance, and material, and visuo-haptic object representations converge in ventral visual cortex. We ask whether representation learning can reproduce this grounding. TacGen mitigates the tactile-data scarcity bottleneck by combining pre-specified V+T contrastive alignment with a latent-space residual-MLP V->T generator that synthesizes tactile latents from RGB for tactile-data scaling. With matched DINOv2 backbones, splits, and probes, V+T improves matched V-only on mass (Delta R^2=+0.570), density (Delta acc=+0.067), hardness (+0.117), and uncertainty-banded force labels (Delta R^2=+0.281); all CIs exclude zero. The same representation lifts matched-capacity TACTO manipulation 0.246->0.979 while V-only capacity scaling accounts for only 4.5% of the gap, preserving 95.5%. The generator reaches cross-seed +0.589, with real tactile +0.585 inside the seed interval; the architecture comparison shows a 13pp downstream gap between reconstruction quality and representation utility. Across five-seed SSVTP/TVL reproductions, YCB-Sight transfer, three-backbone checks, permutation/random-feature controls, hash-verified manifests, and measured-force validation checks, the evidence supports the claim that touch supplies a necessary physical evidence channel for representations of contact-dependent properties.
Chinese Translation
触觉解决了视觉所留下的物理属性模糊性：探索性接触恢复形状、纹理、顺应性和材料，视觉-触觉物体表征在腹侧视觉皮层中汇聚。我们探讨表征学习是否能重现这种基础。TacGen通过结合预先指定的视觉-触觉（V+T）对比对齐与潜在空间残差多层感知机（MLP）视觉到触觉（V->T）生成器，缓解了触觉数据稀缺的瓶颈，该生成器从RGB合成触觉潜变量以实现触觉数据的扩展。在匹配的DINOv2主干网络、分割和探针下，V+T在质量（Delta R^2=+0.570）、密度（Delta acc=+0.067）、硬度（+0.117）和不确定性带的力标签（Delta R^2=+0.281）上优于匹配的仅视觉（V-only），所有置信区间均不包含零。同样的表征提升了匹配容量的TACTO操作从0.246提升至0.979，而仅视觉容量扩展仅占差距的4.5%，保留了95.5%。生成器在跨种子实验中达到+0.589，在种子区间内的真实触觉为+0.585；架构比较显示重建质量与表征效用之间存在13个百分点的下游差距。在五个种子的SSVTP/TVL重现、YCB-Sight转移、三种主干检查、排列/随机特征控制、哈希验证清单和测量力验证检查中，证据支持触觉为接触依赖属性的表征提供必要的物理证据通道的主张。
cs.RO / 30 / 2606.29192
Empowering a Single-Frequency GNSS Receiver to Achieve High-Precision Positioning with Relative Observations
赋能单频GNSS接收机实现高精度定位的相对观测
Abstract
Global Navigation Satellite System (GNSS) navigation is widely used to provide absolute, outdoor positioning in field robotics. Advances in Real-Time Kinematic (RTK) technology can achieve centimeter-level accuracy, facilitating autonomous navigation tasks. However, the cost and extra infrastructure used for RTK still hinder the application and more cost-effective solutions are desired. In this letter, we present a novel tightly-coupled state estimation framework that achieves high-precision localization by using low-cost, mass-market single-frequency GNSS receivers with any relative motion sensors (e.g., wheel encoder, camera, LiDAR). We propose a sliding-window factor graph that integrates generic relative motion with global epoch-to-anchor constraints derived from continuous carrier phase tracking. To eliminate the reliance on physical base stations, we introduce a virtual anchor mechanism: upon the initial observation of a satellite, its state is locked as a virtual reference to establish global epoch-to-anchor constraints. By substituting multi-frequency hardware redundancy with single-frequency multi-modal kinematic priors and a robust cycle-slip recovery technique, our approach ensures carrier-phase integrity on cheap receivers. Extensive real-world experiments on heterogeneous low-cost sensor suites validate that our method improves the accuracy of a single-frequency receiver from several meters to decimeter-level precision across diverse environments, providing an accurate, cost-effective and reliable alternative for autonomous navigation.
Chinese Translation
全球导航卫星系统（GNSS）导航广泛应用于提供绝对的户外定位，尤其在领域机器人中。实时动态（RTK）技术的进步可以实现厘米级的精度，从而促进自主导航任务的执行。然而，RTK所需的成本和额外基础设施仍然阻碍了其应用，因此亟需更具成本效益的解决方案。在本文中，我们提出了一种新颖的紧耦合状态估计框架，通过使用低成本、大众市场的单频GNSS接收机与任何相对运动传感器（如轮编码器、摄像头、激光雷达）结合，实现高精度定位。我们提出了一种滑动窗口因子图，整合了通用的相对运动与基于连续载波相位跟踪推导的全球时刻到锚点约束。为了消除对物理基站的依赖，我们引入了一种虚拟锚点机制：在首次观测到卫星时，其状态被锁定为虚拟参考，以建立全球时刻到锚点的约束。通过用单频多模态运动先验和稳健的周期滑移恢复技术替代多频硬件冗余，我们的方法确保了廉价接收机上的载波相位完整性。在异构低成本传感器组合上的广泛实地实验验证了我们的方法能够将单频接收机的精度从几米提升到分米级精度，提供了一种准确、经济且可靠的自主导航替代方案。
cs.RO / 31 / 2606.29201
Behavior Uncloning: Distilling Mode Redirection into Policy Weights without Inference-Time Steering
行为去克隆：在无推理时间引导的情况下将模式重定向提炼为策略权重
Abstract
Behavior-cloned policies often learn multiple behavior modes from demonstration datasets, including modes that are unsafe or otherwise undesired at deployment. For example, a policy trained on diverse handover demonstrations may learn to pass a knife blade-first. Standard remedies such as data curation and inference-time steering either require access to the original demonstrations for full retraining or add substantial inference-time overhead. To address this gap, we propose MoRE(Mode Redirection), which redirects policy rollouts toward desired behavior modes through a short "uncloning" step. Specifically, MoRE distills the redirection signal from a temporary mode classifier into the policy weights to steer behavior. A retain loss balances this edit by preserving desired-mode competence, allowing the standalone policy to suppress unwanted modes with zero inference-time overhead. Across eight simulated and real-world tasks, MoRE improves the average deployment success rate (SR) by 44 percentage points over the original mixed-mode policy. Among all compared adaptation and steering baselines, MoRE achieves the strongest SR and approaches the filtered-data retraining reference, while preserving task competence and inference speed. MoRE also generalizes across robot policy backbones, including Diffusion Policy and the Pi0.5 VLA, diverse task categories, and real-world deployments.
Chinese Translation
行为克隆策略通常从示范数据集中学习多种行为模式，包括在部署时不安全或不希望出现的模式。例如，针对多样化交接示范训练的策略可能会学习到刀具先刀刃的传递方式。标准的解决方案，如数据整理和推理时间引导，要么需要访问原始示范以进行全面重训练，要么会增加显著的推理时间开销。为了解决这一问题，我们提出了MoRE（模式重定向），它通过一个短暂的“去克隆”步骤将策略执行重定向到期望的行为模式。具体而言，MoRE将来自临时模式分类器的重定向信号提炼为策略权重，以引导行为。保持损失通过保留期望模式的能力来平衡这一编辑，使得独立策略能够以零推理时间开销抑制不希望的模式。在八个模拟和真实世界任务中，MoRE将平均部署成功率（SR）提高了44个百分点，相较于原始的混合模式策略。在所有比较的适应和引导基准中，MoRE实现了最强的SR，并接近过滤数据重训练的参考，同时保持任务能力和推理速度。MoRE还在机器人策略骨干（包括Diffusion Policy和Pi0.5 VLA）、多样化任务类别和真实世界部署中具有良好的泛化能力。
cs.RO / 32 / 2606.29209
AnyBody: Free-Form Whole-Body Humanoid Control from Arbitrary Keypoint Guidance
AnyBody：基于任意关键点引导的自由形式全身类人控制
Abstract
We present AnyBody, a unified whole-body humanoid controller driven by an arbitrary subset of body keypoints chosen at deploy time. Prior physics-based trackers either rely on expensive full-body motion capture and error-prone trajectory retargeting, which bottleneck scalable data collection and policy learning, or decompose upper- and lower-body control into separate hierarchical representations, sacrificing the coordinated whole-body motions that loco-manipulation requires. We close this gap by learning a single latent motion representation that any keypoint subset can address. To achieve this, we first train a privileged teacher tracker on a large unstructured motion corpus and distill it online into a deterministic encoder-decoder student whose latent space is a unit sphere. We then train a transformer keypoint encoder that admits any subset of body keypoints through masked self-attention, aligning it to the privileged latent. Additionally, we treat the frozen decoder as a motor prior and specialize downstream tasks with a lightweight residual corrector in the latent space. We demonstrate the effectiveness of AnyBody by tracking large-scale human motions from arbitrary keypoint subsets, free-form control, flexibly teleoperating, and learning downstream behaviors including locomotion, in-air writing, and obstacle-reach.
Chinese Translation
我们提出了AnyBody，一个统一的全身类人控制器，能够根据在部署时选择的任意子集的身体关键点进行驱动。以往的基于物理的跟踪器要么依赖于昂贵的全身动作捕捉和容易出错的轨迹重定向，这限制了可扩展的数据收集和策略学习，要么将上半身和下半身控制分解为单独的层次表示，从而牺牲了步态操控所需的协调全身运动。我们通过学习一个单一的潜在运动表示来填补这一空白，任何关键点子集都可以利用该表示。为此，我们首先在一个大型非结构化运动语料库上训练一个特权教师跟踪器，并将其在线提炼为一个确定性的编码器-解码器学生，其潜在空间为单位球面。然后，我们训练一个变换器关键点编码器，通过掩蔽自注意力机制接纳任意子集的身体关键点，并将其与特权潜在空间对齐。此外，我们将冻结的解码器视为运动先验，并在潜在空间中使用轻量级残差校正器专门化下游任务。我们通过从任意关键点子集跟踪大规模人类动作、自由形式控制、灵活遥操作以及学习下游行为（包括步态、空中书写和障碍物到达）来展示AnyBody的有效性。
cs.RO / 33 / 2606.29222
CORE Planner: Contextual-memory Oriented Reinforcement-learning in Unknown Environments for Robot Navigation
CORE 规划器：面向上下文记忆的强化学习在未知环境中的机器人导航
Abstract
Autonomous navigation in unknown environments requires a robot to efficiently reach a predefined goal while exploring without prior maps. Although progress has been made in this area, most existing works still rely on traditional planning methods with hand-crafted rules, while learning-based methods often suffer from limited environmental memory and challenges in simulation-to-real (sim-to-real) transfer. To overcome these limitations, we propose a Contextual-memory Oriented Reinforcement-learning (CORE) planner for robot navigation in unknown environments. The proposed CORE planner effectively combines the core advantages of traditional and learning-based methods. Specifically, our method uses a sparse visibility graph for structured environment representation, reducing the computational overhead of dense grid maps, and employs a Transformer network to achieve a holistic environmental understanding, thereby significantly improving navigation efficiency. Moreover, we introduce a visibility graph-based graph sparsification method and a contextual memory mechanism, which alleviates local optima and enhances computational performance in large-scale scenes. Finally, our approach achieves zero-shot sim-to-real transfer after training solely on image-based environments, requiring no fine-tuning. Experimental results show that CORE Planner consistently outperforms state-of-the-art methods, including the traditional FAR Planner and all learning-based baselines, across representative environments, reducing travel distance by 13\% over traditional FAR Planner and by up to 48\% relative to learning-based baselines, with larger gains observed in more complex environments. In real-world scenarios, CORE successfully navigates without human intervention, showcasing zero-shot sim-to-real transfer. Code is available at https://github.com/BBD00/core_planner.
Chinese Translation
在未知环境中实现自主导航要求机器人在没有先前地图的情况下高效到达预定义目标，同时进行探索。尽管在这一领域已有所进展，但大多数现有工作仍依赖于传统规划方法和手工制定的规则，而基于学习的方法往往面临环境记忆有限和模拟到现实（sim-to-real）迁移的挑战。为克服这些局限性，我们提出了一种面向上下文记忆的强化学习（CORE）规划器，用于在未知环境中的机器人导航。所提出的CORE规划器有效结合了传统方法和基于学习的方法的核心优势。具体而言，我们的方法使用稀疏可见性图进行结构化环境表示，减少了密集网格地图的计算开销，并采用Transformer网络实现整体环境理解，从而显著提高导航效率。此外，我们引入了一种基于可见性图的图稀疏化方法和上下文记忆机制，缓解了局部最优问题，并增强了在大规模场景中的计算性能。最后，我们的方法在仅基于图像的环境中训练后实现了零样本的模拟到现实迁移，无需微调。实验结果表明，CORE规划器在代表性环境中始终优于最先进的方法，包括传统的FAR规划器和所有基于学习的基线，相较于传统的FAR规划器减少了13%的行程距离，相较于基于学习的基线减少了多达48%的行程距离，在更复杂的环境中观察到更大的增益。在现实场景中，CORE成功实现无人干预导航，展示了零样本的模拟到现实迁移。代码可在 https://github.com/BBD00/core_planner 获取。
cs.RO / 34 / 2606.29237
MoPe: Motion Permanence for Robust Monocular Gaussian Mapping in Dynamic Environments
MoPe：动态环境中稳健单目高斯映射的运动持久性
Abstract
Robust robot autonomy depends on scene representations that remain stable enough to support localization, navigation, and downstream decision making in dynamic environments. Monocular Gaussian Splatting SLAM provides high-fidelity mapping, but current uncertainty-aware methods still treat dynamic regions largely as per-frame observations. This makes the representation effectively memoryless: when a pedestrian slows, pauses, or reappears after occlusion, the current frame may look static, allowing dynamic content to be absorbed into the map and leaving persistent ghosting artifacts. We argue that this failure reflects a representation-level mismatch. Dynamic-ness is not an instantaneous appearance property, but a temporal property defined by motion history. Building on this view, we introduce Motion Permanence: the principle that an object's dynamic identity should persist over time rather than be re-decided from each frame independently. We realize this principle in MoPe, a memory-aware uncertainty filter for monocular Gaussian mapping. MoPe propagates the historical dynamic posterior through geometry-consistent SE(3) warping and fuses it with current-frame evidence using bounded Bayesian log-odds updates. The resulting persistent posterior guides tracking, mapping, dynamic-aware Gaussian insertion, and Gaussian-level post-cleanup. On Wild-SLAM, Bonn, and TUM sequences, MoPe improves tracking robustness and reduces residual ghosting, with the strongest gains on dynamic-human scenes that most directly violate the memoryless assumption. These results show that maintaining temporal dynamic state inside the scene representation is a practical step toward more reliable representation-centric autonomy in changing real-world environments.
Chinese Translation
稳健的机器人自主性依赖于能够在动态环境中支持定位、导航和后续决策的稳定场景表示。单目高斯点云SLAM提供了高保真度的映射，但当前的考虑不确定性的算法仍然将动态区域视为逐帧观测。这使得表示在实际操作中变得无记忆：当行人减速、停顿或在遮挡后重新出现时，当前帧可能看起来是静态的，从而使动态内容被吸收到地图中，留下持久的虚影伪影。我们认为这种失败反映了表示层面的不匹配。动态性并不是瞬时的外观属性，而是由运动历史定义的时间属性。在此基础上，我们引入了运动持久性（Motion Permanence）：一个物体的动态身份应该随时间持续，而不是独立于每一帧重新决定。我们在MoPe中实现了这一原则，MoPe是一个具有记忆意识的不确定性滤波器，用于单目高斯映射。MoPe通过几何一致的SE(3)变形传播历史动态后验，并使用有界贝叶斯对数几率更新将其与当前帧证据融合。最终得到的持久后验指导跟踪、映射、动态感知的高斯插入和高斯级别的后处理。在Wild-SLAM、Bonn和TUM序列上，MoPe提高了跟踪的稳健性并减少了残余虚影，在最直接违反无记忆假设的动态人类场景中获得了最大的提升。这些结果表明，在场景表示中保持时间动态状态是朝着在变化的现实世界环境中实现更可靠的以表示为中心的自主性迈出的实际一步。
cs.RO / 35 / 2606.29259
PL-LIT: A LiDAR-Inertial-Thermal SLAM Using Point-Line Features and Thermographic Mapping
PL-LIT：一种基于点线特征和热成像映射的激光雷达-惯性-热SLAM
Abstract
Thermal imaging is resilient to adverse conditions, such as intense illumination, low-light operation, and fog, and can therefore mitigate odometry degradation when visible-spectrum imagery becomes unreliable. Nevertheless, most thermal cameras employ automatic gain control (AGC), and thermal images often present low global contrast despite containing informative edge structures. These characteristics undermine brightness constancy and cause conventional optical flow tracking-based odometry pipelines that fundamentally rely on the brightness constancy assumption across consecutive frames. To address these issues, we propose a general LiDAR-Inertial-Thermal SLAM system that accommodates both visible-light and thermal cameras. PL-LIT combines an online photometric calibration module with a deep neural network for point-line feature extraction, enabling more stable and repeatable thermal tracking. For state estimation, we design a tightly coupled LiDAR-Inertial-Thermal formulation within an Error-State Iterated Kalman Filter (ESIKF). We further introduce a line-feature constraint scheme ensuring the reliability of geometric constraints across varying thermal appearances. In addition, PL-LIT builds a probabilistic thermal-intensity voxel map, which supports real-time thermal anomaly detection. Extensive experiments demonstrate that PL-LIT exhibits generality and robustness in visible-light environments, achieves state-of-the-art performance on long-range thermal infrared datasets, and provides practical safety inspection functionality based on thermographic mapping.
Chinese Translation
热成像在不利条件下具有较强的适应性，例如强光照、低光照操作和雾霾，因此可以在可见光图像变得不可靠时减轻里程计的衰退。然而，大多数热成像相机采用自动增益控制（AGC），热图像通常呈现低全局对比度，尽管包含有信息的边缘结构。这些特性削弱了亮度恒定性，并导致传统的基于光流跟踪的里程计管道在连续帧之间基本依赖于亮度恒定性假设。为了解决这些问题，我们提出了一种通用的激光雷达-惯性-热SLAM系统，能够同时支持可见光和热成像相机。PL-LIT结合了在线光度标定模块和用于点线特征提取的深度神经网络，从而实现更稳定和可重复的热跟踪。对于状态估计，我们在误差状态迭代卡尔曼滤波器（ESIKF）中设计了紧耦合的激光雷达-惯性-热模型。此外，我们引入了一种线特征约束方案，以确保在不同热外观下几何约束的可靠性。此外，PL-LIT构建了一个概率热强度体素地图，支持实时热异常检测。大量实验表明，PL-LIT在可见光环境中表现出通用性和鲁棒性，在长距离热红外数据集上实现了最先进的性能，并基于热成像映射提供了实用的安全检查功能。
cs.RO / 36 / 2606.29358
LAMP: Long-Horizon Adaptive Manipulation Planning for Multi-Robot Collaboration in Cluttered Space
LAMP：用于拥挤空间中多机器人协作的长时间适应性操作规划
Abstract
Multi-robot manipulation requires jointly reasoning about contact formations, robot motions under coupled dynamics, and collision avoidance. Systematically searching over this large space is difficult and becomes increasingly intractable as the number of robots grows, the task horizon lengthens, or the scene becomes more cluttered. Existing approaches therefore either learn to solve the problem end-to-end via reinforcement learning or restrict planning to a simpler surrogate problem, such as planning object motions while learning short-horizon contact primitives. However, neither paradigm scales to the problem instances we target: longhorizon multi-robot manipulation in extremely dense environments. In this paper, we propose a Long-horizon Adaptive Manipulation Planning (LAMP) framework with two planners that enable tractable search over the full coupled space by combining a learned generative manipulation model: a LAMPA* planner that systematically searches over the coupled objectrobot space, and LAMP-Lazy: a lazy planner that enables real-time replanning through deferred evaluation. Experiments in challenging simulated environments demonstrate that our approach solves complex long-horizon tasks in highly cluttered environments that prior methods cannot handle.
Chinese Translation
多机器人操作需要共同推理接触形态、在耦合动力学下的机器人运动以及避免碰撞。系统地在这个庞大的空间中进行搜索是困难的，随着机器人数量的增加、任务时间跨度的延长或场景的复杂化，这一问题变得愈加不可处理。因此，现有的方法要么通过强化学习学习端到端解决问题，要么将规划限制在一个更简单的替代问题上，例如在学习短时间接触原语的同时规划物体运动。然而，这两种范式都无法扩展到我们所针对的问题实例：在极其密集的环境中进行长时间的多机器人操作。在本文中，我们提出了一个长时间适应性操作规划（LAMP）框架，包含两个规划器，通过结合一个学习的生成操作模型，使得在完整的耦合空间中进行可处理的搜索成为可能：一个系统地在耦合物体-机器人空间中搜索的 LAMPA* 规划器，以及一个通过延迟评估实现实时重新规划的懒惰规划器 LAMP-Lazy。我们在具有挑战性的模拟环境中的实验表明，我们的方法能够解决以往方法无法处理的高度拥挤环境中的复杂长时间任务。
cs.RO / 37 / 2606.29372
SPACE: Swarm Pheromone Fields for Adaptive Collision-Aware Exploration
SPACE：用于自适应碰撞感知探索的群体信息素场
Abstract
Massive robot swarms can explore unknown environments quickly, but adding robots eventually stops helping. Doorways and dense traffic create congestion, increasing inter-robot contacts and reducing the value of each additional robot. We study this safety-efficiency tradeoff for ground swarms of tens to hundreds of robots. We present SPACE, Swarm Pheromone Fields for Adaptive Collision-Aware Exploration. Inspired by ant foraging, SPACE maintains a shared environmental field with an attractive frontier pheromone, a repellent explore pheromone, and a fast robot-density field. Coordination is decentralized and mediated through this field. We evaluate SPACE on real building floorplans, namely sixteen home layouts from the HouseExpo dataset and eight campus floors from the KTH dataset, with swarms of up to two hundred and fifty-six robots. SPACE lies on the empirical Pareto frontier. It attains the lowest inter-robot contact rate at every congested swarm size, four to seventeen times fewer than a greedy nearest-frontier planner, while keeping coverage time within about two percent of that near time-optimal planner. The results indicate that, at this scale, coordination mainly improves safety rather than coverage time.
Chinese Translation
大规模机器人群可以快速探索未知环境，但增加机器人数量最终会导致效果递减。门口和密集交通会造成拥堵，增加机器人之间的接触，降低每增加一个机器人的价值。我们研究了这种安全性与效率的权衡，针对由数十到数百个机器人组成的地面群体。我们提出了SPACE（Swarm Pheromone Fields for Adaptive Collision-Aware Exploration），该方法受到蚂蚁觅食的启发，维护一个共享的环境场，其中包含吸引前沿的信息素、排斥探索的信息素和快速的机器人密度场。协调是去中心化的，通过这个场进行调节。我们在真实的建筑平面图上评估了SPACE，具体包括来自HouseExpo数据集的十六个家庭布局和来自KTH数据集的八个校园楼层，机器人数量最多可达二百五十六个。SPACE位于经验上的帕累托前沿。在每个拥堵的群体规模下，它达到了最低的机器人间接触率，比贪婪的最近前沿规划器低四到十七倍，同时保持覆盖时间在接近时间最优规划器的约百分之二以内。结果表明，在这个规模下，协调主要改善了安全性，而不是覆盖时间。
cs.RO / 38 / 2606.29460
Understanding LLM Intervention Explanations in Multi-Party Human-Robot Interaction
理解多方人机交互中的大型语言模型干预解释
Abstract
Large Language Models (LLMs) are increasingly embedded in social robots to support natural group interactions, yet their role in complex multi-party settings remains underexplored. In particular, it is unclear how LLM-driven robots decide when and why to intervene in group conversations. This paper investigates the intervention explanations generated by an LLM-based orchestrator in a multi-party interaction involving three human participants and two robots. We conducted a between-subjects study with 24 groups (66 university students), comparing a homogeneous condition (two robots with the same role, i.e., a mover) and a heterogeneous condition (two robots with different roles, i.e., a mover and an opposer). At each conversational turn, the LLM orchestrator decided whether to intervene and generated a textual explanation of its decision. We performed a thematic analysis of 610 intervention explanations, identifying five recurring themes. Results show that explanations are facilitation-oriented, emphasizing agreement, participation, and interaction flow. While patterns remain stable across conditions, role differentiation emerges: the mover supports coordination, whereas the opposer drives goal-oriented interventions. These findings contribute to explainable AI by characterizing how LLM-driven systems justify intervention decisions in real-time, multi-party human-robot interaction.
Chinese Translation
大型语言模型（LLMs）越来越多地嵌入社交机器人中，以支持自然的群体互动，但它们在复杂的多方环境中的角色仍然未被充分探索。特别是，目前尚不清楚基于LLM的机器人如何决定何时以及为何在群体对话中进行干预。本文研究了在涉及三名人类参与者和两台机器人的多方互动中，由LLM驱动的协调者生成的干预解释。我们进行了一个被试间研究，共有24组（66名大学生），比较了同质条件（两台角色相同的机器人，即移动者）和异质条件（两台角色不同的机器人，即移动者和反对者）。在每个对话轮次中，LLM协调者决定是否干预，并生成其决策的文本解释。我们对610个干预解释进行了主题分析，识别出五个反复出现的主题。结果表明，解释以促进为导向，强调一致性、参与和互动流。尽管模式在不同条件下保持稳定，但角色差异显现：移动者支持协调，而反对者推动以目标为导向的干预。这些发现通过描述LLM驱动系统如何在实时多方人机交互中为干预决策提供解释，为可解释人工智能做出了贡献。
cs.RO / 39 / 2606.29469
MTD-Map: Single-Stage Long-Term LiDAR Map Maintenance Framework via Mixture Transition Distribution
MTD-Map：基于混合转移分布的单阶段长期激光雷达地图维护框架
Abstract
While robust map maintenance has advanced significantly, existing studies have focused on specific tasks, especially dynamic object removal or change detection. In this paper, we take a holistic view of the map maintenance problem and propose MTD-Map, a single-stage framework that handles both dynamic object removal and change detection without separate task-specific modules. MTD-Map employs an explicit representation that compactly encodes the direction and duration of occupancy transitions through Mixture Transition Distribution (MTD) modeling. We develop a recursive MTD formulation that encodes historical occupancy patterns into an augmented state to capture high-order temporal dependencies. Furthermore, a stability-driven adaptive strategy balances noise suppression with the preservation of quasi-static structures. Extensive experiments verify that MTD-Map robustly removes dynamic objects and achieves competitive change detection performance, subsequently reducing computational costs. Our project page is available at: https://taeyoung96.github.io/mtd_map/.
Chinese Translation
尽管稳健的地图维护技术已取得显著进展，但现有研究主要集中于特定任务，尤其是动态物体移除或变化检测。本文从整体上看待地图维护问题，提出了MTD-Map，一个单阶段框架，能够在没有单独任务特定模块的情况下同时处理动态物体移除和变化检测。MTD-Map采用显式表示，通过混合转移分布（Mixture Transition Distribution, MTD）建模紧凑地编码占用转移的方向和持续时间。我们开发了一种递归MTD公式，将历史占用模式编码到增强状态中，以捕捉高阶时间依赖性。此外，一种基于稳定性的自适应策略在噪声抑制与准静态结构的保留之间取得平衡。大量实验验证了MTD-Map在动态物体移除方面的稳健性，并实现了具有竞争力的变化检测性能，从而降低了计算成本。我们的项目页面可访问： https://taeyoung96.github.io/mtd_map/。
cs.RO / 40 / 2606.29501
Learning Transferable Dynamics Priors from Action to World Modeling
从动作到世界建模的可转移动态先验学习
Abstract
We study action-conditioned world modeling as a scalable way to learn transferable dynamics priors for robot learning. By pretraining a model to predict how actions drive visual scene evolution, the resulting world model captures reusable interaction dynamics beyond appearance-level video generation. Concretely, we pretrain a multi-view interactive base diffusion world model, A2World, on large-scale robot manipulation data with real action annotations. We validate the learned dynamics priors from two complementary perspectives. First, we adapt A2World into a task- or scene-specialized real-world simulator, A2World-sim, whose long-horizon rollouts support simulator-based policy evaluation and scalable what-if analysis by replacing real-robot rollouts with world model rollouts. Second, starting from the same pretrained weights, we adapt A2World into a video-action joint prediction model, A2World-policy, that predicts actions under visual and instruction conditioning. Experiments across simulation benchmarks and real-robot settings demonstrate that action-conditioned world model pretraining yields transferable dynamics priors that benefit both simulator-centric and policy-centric robot learning.
Chinese Translation
我们研究了以动作为条件的世界建模，作为一种可扩展的方法来学习可转移的动态先验，以促进机器人学习。通过预训练一个模型来预测动作如何驱动视觉场景的演变，所得到的世界模型捕捉了超越外观层面的视频生成的可重用交互动态。具体而言，我们在大规模机器人操作数据上，使用真实动作注释预训练了一个多视角交互基础扩散世界模型A2World。我们从两个互补的角度验证所学习的动态先验。首先，我们将A2World适配为一个任务或场景专用的真实世界模拟器A2World-sim，其长时间跨度的模拟支持基于模拟器的策略评估和可扩展的假设分析，通过用世界模型的模拟替换真实机器人模拟。其次，从相同的预训练权重出发，我们将A2World适配为一个视频-动作联合预测模型A2World-policy，该模型在视觉和指令条件下预测动作。跨越模拟基准和真实机器人设置的实验表明，以动作为条件的世界模型预训练产生了可转移的动态先验，这对以模拟器为中心和以策略为中心的机器人学习均有益。
cs.RO / 41 / 2606.29517
CORE: Common Outcome Regularities from Action-Free Visual Demonstrations for Robot Manipulation
CORE：从无动作视觉演示中提取的共同结果规律用于机器人操控
Abstract
Robot imitation learning often relies on costly robot demonstrations, while abundant action-free visual demonstrations, such as human videos, are difficult to use because they lack robot-executable actions and suffer from embodiment gaps. We propose CORE, a policy learning framework that extracts Common Outcome Regularities from visual demonstrations. Rather than transferring explicit actions across embodiments, CORE exploits a key observation: although successful trajectories for the same task can be diverse, their terminal states often share stable object configurations, spatial relations, and contact constraints. CORE first trains a terminal outcome encoder with contrastive and auxiliary temporal objectives, then aggregates successful terminal embeddings into visual goal prototypes, and finally injects these prototypes as global goal conditions into robot policies. Compared with language instructions, visual goal prototypes provide more concrete geometric and physical constraints for task completion. Across Meta-World, RoboTwin 2.0, and real-world manipulation, CORE improves the average success rate of the corresponding policy backbones by up to +3.9, +11.1, and +17.0 percentage points, respectively, and outperforms text-conditioned variants under the evaluated settings.
Chinese Translation
机器人模仿学习通常依赖于昂贵的机器人演示，而丰富的无动作视觉演示（如人类视频）由于缺乏可执行的机器人动作且存在体现差距，难以使用。我们提出了CORE，一个从视觉演示中提取共同结果规律的策略学习框架。CORE并不是在不同体现之间转移明确的动作，而是利用一个关键观察：尽管同一任务的成功轨迹可能多样化，但它们的终态通常共享稳定的物体配置、空间关系和接触约束。CORE首先通过对比和辅助时间目标训练终态编码器，然后将成功的终态嵌入聚合成视觉目标原型，最后将这些原型作为全局目标条件注入到机器人策略中。与语言指令相比，视觉目标原型为任务完成提供了更具体的几何和物理约束。在Meta-World、RoboTwin 2.0和现实世界操控中，CORE分别将相应策略骨干的平均成功率提高了最多+3.9、+11.1和+17.0个百分点，并在评估设置下优于文本条件变体。
cs.RO / 42 / 2606.29566
Analyzing Uncertainty in the Spatial Representation of the Kinematic Bicycle Model
分析运动自行车模型空间表示中的不确定性
Abstract
Locating a vehicle and determining its orientation in an uncertain environment is a critical challenge in autonomous vehicle navigation and path planning. To address these challenges, a vehicle estimates its pose while depending on sensor data that offer noisy measurements. These uncertainties in pose quantities are expressed mathematically as a covariance matrix. The real-time computation of the covariance matrix is critical because of the non-linearity involved in the kinematic model. The challenge is thus to evaluate the evolution of the covariance matrix of a vehicle's discretized stochastic kinematics. The purpose of this study is to obtain a near-accurate evolution of the covariance matrix of the rear-wheel bicycle kinematic model under uncertainties in wheel displacement and steering angle. We used Taylor's series to linearize the nonlinear trigonometric functions and provided closed-form expectations of random variables with the required accuracy. Our analytical findings are in good agreement with those obtained from Monte-Carlo simulations. Our contribution is probably the first detailed closed-form presentation of the covariance matrix constituents of the vehicle under evaluation, which were previously reported either incorrectly or incompletely. These findings aid in identifying the potential and constraints of the discretized kinematic model as well as its stochastic analysis. The techniques presented here are useful for the simultaneous localization and odometry self-calibration of certain mobile robots and autonomous vehicles.
Chinese Translation
在不确定环境中定位车辆并确定其朝向是自动驾驶车辆导航和路径规划中的一个关键挑战。为了解决这些挑战，车辆在依赖提供噪声测量的传感器数据的同时估计其位姿。这些位姿量的不确定性在数学上表示为协方差矩阵。由于运动模型中涉及的非线性，协方差矩阵的实时计算至关重要。因此，挑战在于评估车辆离散化随机运动学的协方差矩阵的演变。本研究的目的是在轮子位移和转向角的不确定性下，获得后轮自行车运动模型协方差矩阵的近似准确演变。我们使用泰勒级数对非线性三角函数进行线性化，并提供了随机变量的封闭形式期望，确保所需的准确性。我们的分析结果与蒙特卡洛模拟获得的结果相吻合。我们的贡献可能是首次详细呈现评估车辆的协方差矩阵成分的封闭形式，这些成分之前的报告要么不准确，要么不完整。这些发现有助于识别离散化运动模型的潜力和限制，以及其随机分析。这里提出的技术对于某些移动机器人和自动驾驶车辆的同时定位与里程计自校准非常有用。
cs.RO / 43 / 2606.29570
Hierarchical Policy Learning via Spectral Decomposition
通过谱分解进行层次化策略学习
Abstract
In this paper, we identify a semantic decomposition in robot action sequences, separating task-level motion intent from execution-level refinements. By analyzing actions in the spectral domain using the discrete cosine transform (DCT), we observe that low-frequency components capture global motion trajectories, while high-frequency components encode precise timing, alignment, and contact behaviors. Motivated by this structure, we propose Causal Spectral Policy (CSP), which models action generation as a causal coarse-to-fine process: coarse motion is predicted from observation and language, and fine corrections are generated conditionally on the realized trajectory. Across simulation and real-world evaluations, CSP consistently outperforms strong baselines on precision-sensitive manipulation tasks. Additionally, we propose human-inspired teleoperation noise injection as a data augmentation method, under which our approach demonstrates strong robustness to noisy demonstrations.
Chinese Translation
在本文中，我们识别出机器人动作序列中的语义分解，将任务级运动意图与执行级细化相分离。通过使用离散余弦变换（DCT）在谱域中分析动作，我们观察到低频成分捕捉全局运动轨迹，而高频成分则编码精确的时序、对齐和接触行为。基于这一结构，我们提出了因果谱策略（Causal Spectral Policy, CSP），该策略将动作生成建模为一种因果的粗到细过程：粗略运动从观察和语言中预测，而细微修正则基于已实现的轨迹进行条件生成。在模拟和现实世界评估中，CSP在对精度敏感的操作任务上始终优于强基线。此外，我们提出了受人类启发的遥操作噪声注入作为数据增强方法，在这种情况下，我们的方法表现出对噪声演示的强鲁棒性。
cs.RO / 44 / 2606.29673
Privacy-Preserving Decentralized Cooperative Localization with Range-Only Measurements: A Convex Optimization Based Approach
基于范围测量的隐私保护去中心化协作定位：一种基于凸优化的方法
Abstract
Cooperative localization using range-based measurements is critical for multi-robot systems operating in GPS-denied and unstructured environments. However, traditional cooperative approaches require sharing explicit spatial coordinates across the network, presenting a severe security vulnerability in privacy-sensitive missions. While recent literature has explored privacy-preserving alternatives, these methods typically rely on accuracy-degrading noise injection or computationally prohibitive cryptographic protocols. To overcome these limitations, we propose a novel, natively privacy-preserving Decentralized Cooperative Localization (DCL) framework based on convex optimization. Discarding probabilistic noise models, we assume strictly bounded measurement noise and formulate the localization problem via Semi-Definite Programming (SDP) to compute a Maximum-Volume Inscribed Ellipsoid (MVE). Our approach introduces novel intersection-plane constraints derived from landmark measurements to significantly tighten individual spatial bounds. To incorporate inter-robot range measurements securely, we uniquely decompose coupling constraints into localized Linear Matrix Inequalities (LMIs). Agents achieve fleet-wide spatial consensus by iteratively exchanging only abstract dual variables, completely avoiding the transmission of explicit primal position estimates. Extensive 3D Monte Carlo simulations demonstrate that our DCL framework outperforms existing SDP-based localization method in accuracy, while guaranteeing operational privacy and maintaining highly scalable, parallelizable computation.
Chinese Translation
基于范围测量的协作定位对于在GPS无法使用和非结构化环境中操作的多机器人系统至关重要。然而，传统的协作方法需要在网络中共享明确的空间坐标，这在隐私敏感的任务中存在严重的安全漏洞。尽管近期文献探讨了隐私保护的替代方案，但这些方法通常依赖于降低精度的噪声注入或计算上不可承受的加密协议。为克服这些限制，我们提出了一种新颖的、原生隐私保护的去中心化协作定位（DCL）框架，基于凸优化。我们摒弃了概率噪声模型，假设测量噪声严格有界，并通过半正定规划（SDP）来构造定位问题，以计算最大体积内切椭球（MVE）。我们的方法引入了基于地标测量的新的交叉平面约束，以显著收紧个体空间边界。为了安全地整合机器人间的范围测量，我们独特地将耦合约束分解为局部线性矩阵不等式（LMI）。代理通过迭代交换仅抽象的对偶变量，实现全队的空间共识，完全避免了显式原始位置估计的传输。广泛的三维蒙特卡洛仿真表明，我们的DCL框架在精度上优于现有的基于SDP的定位方法，同时保证了操作隐私，并保持高度可扩展和可并行计算。
cs.RO / 45 / 2606.29677
Lateral String Stability for Vehicle Platoons
车辆编队的横向串稳定性
Abstract
Connected and automated vehicle (CAV) platooning promises gains in energy efficiency and traffic throughput and, most critically, in safety. These safety benefits hinge on string stability, which determines how disturbances propagate along a platoon. While longitudinal string stability is well studied, lateral string stability, which governs the propagation of path-tracking errors that can lead to unsafe deviations from the intended path, remains underexplored. Its importance is increasing as autonomous vehicles rely more heavily on onboard sensing and map-free navigation, where sensor occlusion and dense formations amplify safety risks. This paper presents a new framework for lateral string stability that directly addresses safety-critical path-relative tracking errors and enables consistent comparison across vehicles following the same road geometry. Central to this framework is an arc-length (Eulerian) viewpoint, a departure from traditional analyses, that clarifies how tracking errors at a given point on the path propagate from one vehicle to the next. A formal definition of lateral string stability is introduced along with two control strategies: an onboard-sensing-only controller and a novel learn-from-predecessor approach utilizing vehicle-to-vehicle (V2V) communication. We show that onboard sensing alone cannot guarantee attenuation of path-tracking errors, imposing a fundamental safety limitation, whereas V2V communication enables true error attenuation.
Chinese Translation
连接和自动化车辆（CAV）编队承诺在能源效率、交通通行能力以及最重要的安全性方面带来收益。这些安全益处依赖于串稳定性，它决定了干扰在编队中的传播方式。尽管纵向串稳定性已被广泛研究，但横向串稳定性，即控制路径跟踪误差的传播，这些误差可能导致偏离预定路径的安全风险，仍然未得到充分探索。随着自主车辆越来越依赖车载传感器和无地图导航，其重要性日益增加，因为传感器遮挡和密集编队会放大安全风险。本文提出了一种新的横向串稳定性框架，直接解决与安全相关的路径相对跟踪误差，并能够在遵循相同道路几何形状的车辆之间进行一致的比较。该框架的核心是弧长（欧拉）视角，这一视角不同于传统分析，明确了在路径上某一点的跟踪误差如何从一辆车传播到另一辆车。我们引入了横向串稳定性的正式定义，并提出了两种控制策略：仅依赖车载传感的控制器和一种利用车对车（V2V）通信的创新“向前学习”方法。我们展示了仅依靠车载传感无法保证路径跟踪误差的衰减，从而施加了基本的安全限制，而V2V通信则能够实现真正的误差衰减。
cs.RO / 46 / 2606.29731
Real-Time Compliance and Position Control of a Hyper-redundant Soft Robotic Arm
超冗余软体机器人手臂的实时合规性与位置控制
Abstract
Robots working in unstructured or partially unobservable environments must combine accurate motion with physical compliance that can passively correct contact misalignment. Soft robots provide this compliance but have struggled to precisely control their tip compliance and position. This paper presents a robot architecture designed around that control problem: a 7-link arm whose six articulated joints provide twelve independently driven revolute axes, each actuated by an antagonistic pair of pneumatic muscles, so that every axis can simultaneously change its angle and linearly adjust its stiffness. The rigid articulated backbone makes the tip compliance and position of the arm predictable enough to be commanded quantitatively in real time. The robot employs a unified iterative inverse-kinematics and inverse-compliance controller to achieve simultaneous, quantitative control of both compliance and position. The task-space compliance and kinematics models and the control law are derived and verified on both the physical arm and a matched simulation. Simulation is then used to study how the same framework extends to other arm morphologies. Finally, the arm demonstrates tasks that have been difficult for both rigid and soft arms: rejecting disturbances while writing on a moving whiteboard, and passively correcting hidden misalignment during a key-insertion and drawer-opening task. That these tasks succeed under so straightforward a controller is evidence for the advantage of this algorithm-informed structural design.
Chinese Translation
在非结构化或部分不可观察的环境中工作的机器人必须将精确的运动与能够被动纠正接触不对准的物理合规性相结合。软体机器人提供了这种合规性，但在精确控制其末端合规性和位置方面面临挑战。本文提出了一种围绕这一控制问题设计的机器人架构：一个7节臂，其六个关节提供十二个独立驱动的旋转轴，每个轴由一对对抗性的气动肌肉驱动，从而使每个轴能够同时改变其角度并线性调整其刚度。刚性关节骨架使得手臂的末端合规性和位置足够可预测，可以实时定量指令。该机器人采用统一的迭代逆运动学和逆合规性控制器，以实现合规性和位置的同时定量控制。任务空间合规性和运动学模型以及控制律在物理手臂和匹配的仿真中得到了推导和验证。随后，仿真用于研究相同框架如何扩展到其他手臂形态。最后，该手臂展示了一些对刚性和软体手臂都具有挑战性的任务：在移动白板上书写时抵抗干扰，以及在钥匙插入和抽屉打开任务中被动纠正隐藏的不对准。这些任务在如此简单的控制器下成功完成，证明了这种算法驱动的结构设计的优势。
cs.RO / 47 / 2606.29738
MyGO-Splat: Multi-Objective Closed-Loop Geometric Feedback for RGB-Only Gaussian SLAM
MyGO-Splat：用于仅RGB的高斯SLAM的多目标闭环几何反馈
Abstract
Real-time monocular Simultaneous Localization and Mapping (SLAM) fundamentally suffers from scale ambiguity and a lack of geometric self-correction. While 3D Gaussian Splatting (3DGS) enables high-fidelity rendering, existing RGB-only systems remain open-loop because depth priors are injected into mapping but refined geometry cannot effectively regulate tracking drift. We present MyGO-Splat, a closed-loop Gaussian SLAM framework that analytically rasterizes Gaussian primitives into pixel-wise depth and surface normals, allowing the map to actively supervise camera pose optimization. To bridge monocular priors and scale consistency, our framework introduces scale-aware adaptive alignment that projects foundation-model depth estimates into the globally optimized Gaussian space, forming a self-correcting cycle for scale feedback. Extensive evaluations show that this closed-loop design improves scale stability and appearance-geometry consistency, achieving performance comparable to RGB-D methods while using only monocular input.
Chinese Translation
实时单目同时定位与地图构建（SLAM）在本质上受到尺度模糊和几何自我校正缺乏的影响。尽管3D高斯点云（3DGS）能够实现高保真渲染，但现有的仅RGB系统仍然是开放式的，因为深度先验被注入到地图构建中，但精炼的几何体无法有效调节跟踪漂移。我们提出了MyGO-Splat，一个闭环高斯SLAM框架，能够将高斯原语解析为逐像素深度和表面法线，从而使地图能够主动监督相机姿态优化。为了弥合单目先验与尺度一致性，我们的框架引入了尺度感知自适应对齐，将基础模型的深度估计投影到全局优化的高斯空间中，形成一个自我校正的尺度反馈循环。大量评估表明，这种闭环设计改善了尺度稳定性和外观-几何一致性，尽管仅使用单目输入，其性能可与RGB-D方法相媲美。
cs.RO / 48 / 2606.29755
Multi-UAV Formation Cooperative Obstacle Avoidance and Adaptive Shape Deformation Control in Complex Environments Based on BI-APF-RRT and Affine Transformation
基于 BI-APF-RRT 和仿射变换的复杂环境中多无人机编队协同避障与自适应形状变形控制
Abstract
Aiming at the problem that obstacle avoidance flexibility and formation integrity are difficult to coexist in multi-UAV formation motion in complex obstacle environments , and that the traditional artificial potential field (APF) method easily falls into local optima, a cooperative obstacle avoidance algorithm for multi-UAV formations integrating BI-APF-RRT and affine transformation is proposed. First, abandoning the traditional APF centroid path planning method , a goal-biased Bidirectional Artificial Potential Field method RRT (BI-APF-RRT) algorithm is adopted to conduct global collision-free path planning for the centroid of the leader formation. By introducing an improved artificial potential field and cubic B-spline interpolation, the smoothness and rapid convergence of the global path are ensured. Secondly, using the generated global path as the guiding trajectory for the formation's centroid , combined with an affine transformation matrix (including non-uniform scaling and rotation) , the formation can adaptively deform based on the distance to obstacles while moving along the optimal path. Finally, the followers track the leaders through a distributed control law , enabling the entire formation to safely cross complex obstacle areas without disassembling.
Chinese Translation
针对多无人机编队在复杂障碍环境中避障灵活性与编队完整性难以共存的问题，以及传统人工势场（APF）方法容易陷入局部最优的缺陷，提出了一种集成 BI-APF-RRT 和仿射变换的多无人机编队协同避障算法。首先，摒弃传统的 APF 重心路径规划方法，采用目标偏置的双向人工势场方法 RRT（BI-APF-RRT）算法，对领队编队的重心进行全局无碰撞路径规划。通过引入改进的人工势场和三次 B 样条插值，确保了全局路径的平滑性和快速收敛性。其次，利用生成的全局路径作为编队重心的指导轨迹，结合仿射变换矩阵（包括非均匀缩放和旋转），编队可以在沿最优路径移动的同时，根据与障碍物的距离自适应变形。最后，跟随者通过分布式控制律跟踪领队，使整个编队能够安全穿越复杂障碍区域而不发生解散。
cs.RO / 49 / 2606.29757
Cross-Spectral Stereo Inertial Odometry
跨光谱立体惯性测距
Abstract
Standard stereo VIO focuses exclusively on the benefit of metric scale via single-spectrum baselines, often overlooking the risks of spectral redundancy. This structural limitation leads to correlated failures, where both sensors simultaneously fail in degraded environments that affect their shared spectrum. Leveraging a cross-spectral system presents a complementary solution to this issue, yet the significant appearance gap between modalities renders standard matching ineffective. Existing deep learning-based matchers, while effective, introduce inference latencies that violate real-time constraints. To bridge this gap, we present an asynchronous real-time cross-spectral visual-thermal-inertial (VTI) system that temporally decouples high-latency deep matching from high-rate state estimation. Our architecture incorporates a spectral-aware weighting scheme that dynamically balances modality reliance based on photometric entropy and thermal noise, ensuring robustness against both abrupt lighting changes and thermal artifacts. Furthermore, we introduce a seamless handling mechanism for thermal Non-uniformity Correction (NUC) to maintain tracking continuity. Extensive experiments across diverse scenarios confirm that our system overcomes spectral redundancy, yielding superior accuracy in nominal daylight while ensuring robustness in visually degraded environments. We will open source our code and data: https://github.com/seungsang07/cross-spectral-stereo-inertial-odometry
Chinese Translation
标准立体视觉惯性测距（VIO）专注于通过单光谱基线获得度量尺度的好处，往往忽视了光谱冗余带来的风险。这一结构性限制导致了相关故障，即在影响其共享光谱的恶劣环境中，两个传感器同时失效。利用跨光谱系统提供了一种互补的解决方案，但不同模态之间显著的外观差异使得标准匹配无效。现有的基于深度学习的匹配器虽然有效，但引入的推理延迟违反了实时约束。为了解决这一问题，我们提出了一种异步实时跨光谱视觉-热成像-惯性（VTI）系统，该系统在时间上将高延迟深度匹配与高频率状态估计解耦。我们的架构结合了一种光谱感知加权方案，根据光度熵和热噪声动态平衡模态依赖性，确保对突发光照变化和热伪影的鲁棒性。此外，我们引入了一种无缝处理热非均匀性校正（NUC）的机制，以保持跟踪的连续性。在多种场景下进行的广泛实验确认我们的系统克服了光谱冗余，在正常日光下实现了更高的准确性，同时在视觉退化环境中确保了鲁棒性。我们将开源我们的代码和数据： https://github.com/seungsang07/cross-spectral-stereo-inertial-odometry
cs.RO / 50 / 2606.29766
Trajectory Optimization for Collision-Aware Redundant Robotic Multi-Axis Additive Manufacturing by Constrained Gradient Projection
基于约束梯度投影的碰撞感知冗余机器人多轴增材制造轨迹优化
Abstract
Redundant robotic multi-axis additive manufacturing (MAAM) enables support-free and conformal fabrication, but trajectory optimization for long-horizon paths remains challenging under strict deposition-position constraints and time-varying collision constraints. This work proposes a computational framework for collision-aware trajectory optimization in redundant robotic MAAM. We first formulate nozzle-workpiece relative kinematics using a relative Jacobian, and develop a differentiable SDF-based collision model that captures fabrication-induced geometry evolution and provides optimization gradients. The deposition position is then enforced as a hard waypoint-wise equality constraint through iterative projection onto the self-motion manifold, with the loss gradient restricted to the corresponding tangent space. Experiments on an 8-DOF robotic MAAM platform with diverse long-horizon support-free and conformal toolpaths show that our method maintains a mean nozzle-position error below 10{\mu}m, reduces maximum joint jerk by up to $77.6\%$, and eliminates all sampled collision and orientation violations. Compared with the SQP-based baseline, it achieves up to a 10.2x speedup and improved convergence. Physical fabrication experiments further verify that the resulting smooth, collision-free trajectories enable successful printing of complex geometries with fewer visible deposition artifacts.
Chinese Translation
冗余机器人多轴增材制造（MAAM）实现了无支撑和符合形状的制造，但在严格的沉积位置约束和时变碰撞约束下，长时间路径的轨迹优化仍然具有挑战性。本研究提出了一种用于冗余机器人MAAM的碰撞感知轨迹优化的计算框架。我们首先使用相对雅可比矩阵对喷嘴与工件之间的相对运动学进行公式化，并开发了一种可微分的基于SDF（Signed Distance Function）的碰撞模型，该模型捕捉了制造引起的几何演变并提供优化梯度。然后，通过迭代投影到自运动流形上，将沉积位置强制作为硬性航点等式约束，同时将损失梯度限制在相应的切空间内。在一个具有多样化长时间无支撑和符合形状工具路径的8自由度机器人MAAM平台上的实验表明，我们的方法保持了平均喷嘴位置误差低于10μm，最大关节抖动减少了高达77.6%，并消除了所有采样的碰撞和方向违规情况。与基于SQP（Sequential Quadratic Programming）的基线相比，它实现了高达10.2倍的加速和改进的收敛性。物理制造实验进一步验证了所得到的平滑、无碰撞轨迹能够成功打印复杂几何形状，并减少可见的沉积伪影。
cs.RO / 51 / 2606.29774
Analytic Concept-Centric Memory for Agentic Embodied Manipulation
面向代理体操控的分析性概念中心记忆
Abstract
Long-horizon embodied manipulation requires agents to remember persistent objects, track changing scene states, and reuse prior interaction knowledge. However, existing agent memories are often stored as unstructured histories or embedding-based records, making it difficult to retrieve manipulation-relevant object parts, physical states, action effects, and executable skills. We propose an analytic concept-centric memory framework for agentic embodied manipulation. Our memory organizes experience around structured analytic concepts, where objects are represented by semantic parts, parametric templates, grounded poses, affordances, and manipulation states. It further connects object and scene memories with transition memory for action-induced state changes and skill memory for template-grounded and policy-grounded execution. At runtime, the agent performs structured coarse-to-fine retrieval to identify relevant objects, states, transitions, and skills, supporting state-consistent reasoning and skill reuse. Experiments on memory-dependent manipulation, articulated-object generalization, real-world memory evaluation, and ablations show that our approach improves task completion, retrieval accuracy, object re-identification, and cross-object skill generalization over unstructured and embedding-based memory baselines.
Chinese Translation
长时间的体操控要求代理能够记住持久的物体、跟踪变化的场景状态，并重用先前的交互知识。然而，现有的代理记忆通常以非结构化的历史记录或基于嵌入的记录存储，这使得检索与操控相关的物体部分、物理状态、动作效果和可执行技能变得困难。我们提出了一种面向代理体操控的分析性概念中心记忆框架。我们的记忆围绕结构化的分析性概念组织经验，其中物体由语义部分、参数模板、基础姿态、可用性和操控状态表示。它进一步将物体和场景记忆与过渡记忆连接，以处理由动作引起的状态变化，以及与模板基础和策略基础执行相关的技能记忆。在运行时，代理执行结构化的粗到细检索，以识别相关的物体、状态、过渡和技能，支持状态一致的推理和技能重用。在依赖记忆的操控、关节物体泛化、现实世界记忆评估和消融实验中的实验表明，我们的方法在任务完成、检索准确性、物体重新识别和跨物体技能泛化方面优于非结构化和基于嵌入的记忆基线。
cs.RO / 52 / 2606.29783
FalconTrack: Photorealistic Auto-Labeled Perception and Physics-Aware Vision-Based Aerial Tracking
FalconTrack：基于视觉的逼真自动标注感知与物理感知的空中跟踪
Abstract
Vision-based aerial tracking is critical in GPS-denied environments. Reliable perception for tracking depends on large-scale labeled data, yet most photorealistic datasets rely on heavy manual annotation and are time-consuming to produce. We present FalconTrack, a unified perception-and-tracking framework that (i) leverages a photorealistic editable simulator for automated label generation and (ii) combines multi-head perception with physics-aware tracking for zero-shot sim-to-real transfer. FalconTrack provides an automated labeling pipeline in a Gaussian Splatting simulator that isolates target Gaussians from short object videos and composites them with randomized backgrounds to generate RGB, mask, class, and 6-DoF pose labels, producing about 10k labeled images in under 20 minutes. Using this dataset, we train a multi-head perception module with staged learning and reprojection consistency, and fuse its outputs with class-conditioned dynamics priors in an EKF for tracking. Our perception model outperforms two baselines and reaches 96-100% class accuracy in zero-shot sim-to-real transfer on three geometrically diverse objects and two environments, while maintaining consistent performance in unseen simulated and real scenes. In real hardware closed-loop visual tracking, the onboard system runs at about 25 Hz and achieves 100% success in sim-to-real F1-tenth and gate tracking in five trajectories across two environments, while a mask-centered vision baseline drops to 60% success on F1-tenth during fast out-of-view scenarios.
Chinese Translation
基于视觉的空中跟踪在GPS无法使用的环境中至关重要。可靠的跟踪感知依赖于大规模标注数据，然而大多数逼真数据集依赖于繁重的手动标注，且制作过程耗时。我们提出了FalconTrack，一个统一的感知与跟踪框架，它（i）利用一个可编辑的逼真模拟器进行自动标签生成，以及（ii）结合多头感知与物理感知跟踪，实现零-shot的模拟到现实转移。FalconTrack在一个高斯喷溅模拟器中提供了一个自动标注管道，该管道从短视频中隔离目标高斯，并将其与随机背景合成，以生成RGB、掩码、类别和6自由度姿态标签，在不到20分钟的时间内生成约10,000张标注图像。利用该数据集，我们训练了一个多头感知模块，采用分阶段学习和重投影一致性，并将其输出与EKF中的类别条件动态先验融合以进行跟踪。我们的感知模型在三个几何多样的物体和两个环境中超越了两个基线，并在零-shot的模拟到现实转移中达到了96-100%的类别准确率，同时在未见过的模拟和真实场景中保持了一致的性能。在真实硬件闭环视觉跟踪中，机载系统以约25 Hz的频率运行，在两个环境中的五条轨迹上实现了100%的模拟到现实F1-十分之一和门跟踪成功，而以掩码为中心的视觉基线在快速视野外场景中成功率降至60%。
cs.RO / 53 / 2606.29825
Data-Driven Modeling and Control for Tethered Space Systems with Koopman-Informed Graphs
基于数据驱动的系留空间系统建模与控制：以库普曼图为基础
Abstract
Modeling tethered space systems is critical for advanced orbital operations. Flexible components such as tethers and space nets are integral to these systems but present significant control challenges due to their high dimensional, strongly coupled, and nonlinear dynamics. While data driven methods offer alternative modeling approaches, they frequently struggle with long term predictive stability and spatial generalization. To address this, we propose the Koopman Graph Dynamics (KGD) framework to learn the structural dynamics by integrating the global linear evolution of the Koopman operator with the local topological priors of Graph Neural Networks. Building upon this representation, we develop a KGD based Model Predictive Control strategy for tethered space systems. Subsequently, the ground experiments on flexible tether and space net demonstrate the high precision modeling capabilities of the proposed method. Crucially, the framework exhibits exceptional capacity for spatial transfer without retraining. Models trained exclusively on small configurations successfully predict and control significantly larger, unseen physical scales. Furthermore, the orbit simulations within a physics engine verify the effectiveness of the proposed approach for tethered space systems.
Chinese Translation
系留空间系统的建模对于先进的轨道操作至关重要。柔性组件如系绳和空间网是这些系统的核心，但由于其高维度、强耦合和非线性动态，带来了显著的控制挑战。尽管数据驱动方法提供了替代的建模途径，但它们在长期预测稳定性和空间泛化方面常常面临困难。为了解决这一问题，我们提出了库普曼图动态（Koopman Graph Dynamics, KGD）框架，通过将库普曼算子的全局线性演化与图神经网络的局部拓扑先验相结合，来学习结构动态。在此基础上，我们开发了一种基于KGD的模型预测控制策略，专门用于系留空间系统。随后，针对柔性系绳和空间网的地面实验展示了所提方法的高精度建模能力。重要的是，该框架在不重新训练的情况下展现出卓越的空间迁移能力。仅在小规模配置上训练的模型能够成功预测和控制显著更大且未见过的物理尺度。此外，在物理引擎中的轨道仿真验证了所提方法在系留空间系统中的有效性。
cs.RO / 54 / 2606.29834
STEAM: Self-Supervised Temporal Ensemble Advantage Modeling for Real-World Robot Learning
STEAM：用于现实世界机器人学习的自监督时间集成优势建模
Abstract
Real-world robot learning increasingly relies on heterogeneous data, but demonstrations and rollouts often mix useful progress with stalls, corrections, and suboptimal behavior. Effective policy learning therefore requires frame-level advantages that distinguish reliable local progress from failures and regressions. We propose Self-supervised Temporal Ensemble Advantage Modeling (STEAM), a label-free method that learns such advantages from expert demonstrations. STEAM trains an ensemble of temporal-offset predictors on frame pairs within expert trajectories, using the normalized temporal offset between two frames as a self-supervised signal. Each predictor maps a frame pair to a distribution over temporal offsets, which is converted into a scalar advantage. STEAM then takes the minimum advantage across the ensemble to score mixed-quality rollout data conservatively. Across real-world bimanual towel folding, chip checkout, cola restocking, and single-arm pick-and-place tasks, STEAM identifies stalls, failures, and recoveries. When combined with CFGRL, STEAM further improves policy success rate by 59%, 54.3%, 23% and 16.2% over baselines, respectively.
Chinese Translation
现实世界的机器人学习越来越依赖于异构数据，但演示和回放往往将有用的进展与停滞、修正和次优行为混合在一起。因此，有效的策略学习需要帧级优势，以区分可靠的局部进展与失败和退步。我们提出了自监督时间集成优势建模（STEAM），这是一种无标签的方法，从专家演示中学习这种优势。STEAM在专家轨迹中的帧对上训练一组时间偏移预测器，使用两个帧之间的归一化时间偏移作为自监督信号。每个预测器将帧对映射到时间偏移的分布，该分布被转换为标量优势。然后，STEAM在整个集成中取最小优势，以保守地对混合质量的回放数据进行评分。在现实世界的双手毛巾折叠、芯片结账、可乐补货和单臂抓取放置任务中，STEAM能够识别停滞、失败和恢复。当与CFGRL结合时，STEAM进一步提高了策略成功率，分别比基线提高了59%、54.3%、23%和16.2%。
cs.RO / 55 / 2606.29846
Legible Shared Autonomy: Implicit Communication of Robot Belief through Motion
可读的共享自主性：通过运动隐式传达机器人信念
Abstract
Shared autonomy systems combine user input with autonomous assistance to help users with motor impairments control robot arms to perform everyday manipulation tasks, by inferring user goals and providing appropriate guidance. However, the robot's internal beliefs about user goals cannot be observed by users. Traditional shared autonomy systems provide assistance along efficient shortest paths toward inferred goals, but when multiple objects lie in similar directions, such assistive motion remains ambiguous and fails to reveal the specific goal identified by the robot. This creates two critical problems. First, when the robot correctly infers the goal, users continue controlling because they cannot perceive understanding from ambiguous assistive motion, wasting effort when autonomous completion would suffice. Second, when the robot misunderstands intent, users cannot quickly detect errors until assistive motion diverges significantly, requiring substantial corrective input. We address this by introducing legible motion into shared autonomy, where robot actions must both advance toward the goal and clearly reveal which goal has been inferred, enabling users to understand the robot's beliefs and adjust control accordingly. The robot modulates communication strength through confidence-aware adaptive authority allocation by providing assertive legible assistive actions when confident while increasing user authority when uncertain, transforming shared autonomy into transparent bidirectional collaboration. User studies including simulation and physical experiments with a six-degree-of-freedom robot arm demonstrate that legible shared autonomy significantly improves users' understanding of robot beliefs and reduces user control effort compared to standard shared autonomy.
Chinese Translation
共享自主系统结合用户输入与自主辅助，帮助运动障碍用户控制机器人手臂执行日常操作任务，通过推断用户目标并提供适当的指导。然而，用户无法观察到机器人对用户目标的内部信念。传统的共享自主系统沿着推断目标的高效最短路径提供辅助，但当多个物体位于相似方向时，这种辅助运动仍然模糊，无法揭示机器人所识别的具体目标。这造成了两个关键问题。首先，当机器人正确推断出目标时，由于用户无法从模糊的辅助运动中感知理解，用户仍然继续控制，浪费了努力，而自主完成就足够了。其次，当机器人误解意图时，用户无法快速检测错误，直到辅助运动显著偏离，这需要大量的纠正输入。我们通过在共享自主中引入可读运动来解决这一问题，其中机器人的动作必须既朝向目标推进，又清晰地揭示所推断的目标，使用户能够理解机器人的信念并相应调整控制。机器人通过自信感知的自适应权威分配调节沟通强度，在自信时提供果断的可读辅助动作，而在不确定时增加用户的权威，将共享自主转变为透明的双向协作。包括仿真和物理实验在内的用户研究表明，可读的共享自主显著提高了用户对机器人信念的理解，并减少了与标准共享自主相比的用户控制努力。
cs.RO / 56 / 2606.29851
TACO: A Test and Check Framework for Robust Pose Graph Optimization
TACO：一种用于鲁棒姿态图优化的测试与检查框架
Abstract
Pose Graph Optimization (PGO) is one of the most widely adopted approaches for solving Simultaneous Localization and Mapping (SLAM) problems. However, PGO approaches are particularly sensitive to outliers, which can substantially degrade the quality of the estimated trajectories. These outliers arise from incorrect place recognition associations caused by perceptual aliasing in the environment. In this paper, we present TACO (short for Test And Check Optimization), a robust optimization framework designed to filter out outliers from PGO systems. Rather than explicitly modeling measurements as inliers or outliers, TACO finds an approximation to the maximally consistent set of measurements incrementally through two complementary components: (i) The test component, namely the Incremental Probabilistic Consensus (IPC) algorithm, evaluates the consistency of each incoming loop closure online. (ii) The check component dubbed Switchable Outlier Sanitization leverages the existing Switchable Constraints to periodically sanitize any inconsistent measurements from the consistent set that IPC may have mistakenly included. We evaluate TACO on 2D SLAM and 3D Visual SLAM datasets against several state-of-the-art methods. The results show robustness comparable to state-of-the-art offline methods while preserving the computational efficiency required for online deployment, achieving a success rate above 90% in 2D and 83% in 3D across outlier rates up to 50%, with mean convergence times of approximately 45 ms and 100 ms, respectively. We release an open-source implementation of our method with this paper.
Chinese Translation
姿态图优化（PGO）是解决同时定位与地图构建（SLAM）问题的最广泛采用的方法之一。然而，PGO 方法对异常值特别敏感，这可能会显著降低估计轨迹的质量。这些异常值源于由于环境中的感知混淆而导致的不正确位置识别关联。本文提出了 TACO（测试与检查优化的缩写），这是一个旨在从 PGO 系统中过滤异常值的鲁棒优化框架。TACO 通过两个互补组件逐步找到最大一致性测量集的近似值，而不是像以往那样明确将测量建模为内点或外点：（i）测试组件，即增量概率共识（IPC）算法，在线评估每个输入环闭合的一致性；（ii）检查组件，称为可切换异常值清理，利用现有的可切换约束定期清理 IPC 可能错误包含的一致集中的任何不一致测量。我们在 2D SLAM 和 3D 视觉 SLAM 数据集上评估了 TACO，并与几种最先进的方法进行了比较。结果显示，TACO 的鲁棒性可与最先进的离线方法相媲美，同时保持在线部署所需的计算效率，在异常值率高达 50% 的情况下，2D 和 3D 的成功率分别超过 90% 和 83%，平均收敛时间约为 45 毫秒和 100 毫秒。我们随本文发布了我们方法的开源实现。
cs.RO / 57 / 2606.29868
Normalizing Flow-Enhanced Message Passing for Multirobot Collaborative Localization
增强正则化流的消息传递用于多机器人协作定位
Abstract
Accurate, robust, and adaptive localization is essential for various robotic operations. This paper proposes a new message passing (MP) algorithm for realizing collaborative localization in a distributed manner. The algorithm unifies Gaussian belief propagation (GBP) and mean-field (MF) approximation, where GBP preserves dependencies among robot states, and MF enables estimation of noise statistics. To effectively handle non-conjugate terms from nonlinear measurement models, the algorithm adopts a parametric formulation in which these terms are treated by gradient estimators. Beyond linearization and sampling, we further design a normalizing flow (NF)-based gradient estimator, enabling learnable sampling. End-to-end training tunes NF parameters according to the behavior of MP, improving the overall estimation performance. To support estimation of practical robotic states that involve rotations, the method is then extended to Lie group state spaces. Finally, the method is applied to multirobot localization task fusing odometry, global navigation satellite system (GNSS) measurements, and inter-robot ultra wideband (UWB) ranging. Simulations and experiments on autonomous surface vehicles (ASVs) demonstrate its improved accuracy, robustness, and adaptability.
Chinese Translation
准确、稳健和自适应的定位对于各种机器人操作至关重要。本文提出了一种新的消息传递（MP）算法，以分布式方式实现协作定位。该算法统一了高斯信念传播（GBP）和均值场（MF）近似，其中GBP保留了机器人状态之间的依赖关系，而MF则使得噪声统计的估计成为可能。为了有效处理来自非共轭项的非线性测量模型，算法采用了一种参数化形式，将这些项通过梯度估计器进行处理。除了线性化和采样之外，我们进一步设计了一种基于正则化流（NF）的梯度估计器，实现可学习的采样。端到端训练根据MP的行为调整NF参数，从而提高整体估计性能。为了支持涉及旋转的实际机器人状态的估计，该方法进一步扩展到李群状态空间。最后，该方法应用于多机器人定位任务，融合了里程计、全球导航卫星系统（GNSS）测量和机器人间超宽带（UWB）测距。对自主水面车辆（ASVs）的仿真和实验表明其在准确性、稳健性和适应性方面的提升。
cs.RO / 58 / 2606.29875
AUSLUN: A Fixed-Hover UAV--USV System for GNSS-Denied Maritime Search and Navigation
AUSLUN：一种用于GNSS拒绝环境下海洋搜索与导航的固定悬停无人机-无人水面艇系统
Abstract
Global navigation satellite system (GNSS) denial can prevent an unmanned surface vehicle (USV) from both finding a distant vessel and maintaining a globally referenced approach. This paper presents AUSLUN (Automatic UAV Search, Localization, and USV Navigation), a fixed-hover aerial-surface system that uses a coastal unmanned aerial vehicle (UAV), which estimates its own pose through visual-inertial odometry (VIO), as a long-range sensing and navigation anchor. The central design shifts sensing motion from UAV translation to a zoom pod and closes the loop through three coupled elements: polygon-aware annular pod scanning, modality-aware bearing-range localization, and target-relative USV guidance with visual-loss recovery. The same gated recursive estimator uses laser range for the non-cooperative target and datalink range for the cooperative USV. Search-planning simulations show that the adaptive yaw bounds reduce scan time and redundant coverage relative to a matched fixed-sector scan, and GPS-referenced field data show that the gated recursive estimator outperforms non-recursive baselines in localization accuracy. An integrated maritime mission further demonstrates the complete search-to-navigation sequence, including a deliberately triggered visual-loss recovery. These results establish the feasibility and operating boundary of fixed-hover UAV assistance for stationary-target approach in coastal GNSS-denied environments. The source code and a video demonstration are publicly available at https://github.com/xirhxq/pod_search and https://youtu.be/S-5RkJs35JI.
Chinese Translation
全球导航卫星系统（GNSS）拒绝可能会阻碍无人水面艇（USV）找到远处的船只并维持全球参考的接近。本文提出了AUSLUN（自动无人机搜索、定位与USV导航），这是一种固定悬停的空中-水面系统，利用沿海无人机（UAV）作为远程传感和导航锚点，该无人机通过视觉惯性里程计（VIO）估计自身姿态。中心设计将传感运动从无人机平移转移到变焦舱，并通过三个耦合元素闭合回路：多边形感知的环形舱扫描、模态感知的方位-距离定位，以及具有视觉丢失恢复的目标相对USV引导。同样的门控递归估计器使用激光测距用于非合作目标，使用数据链路测距用于合作USV。搜索规划模拟表明，自适应偏航边界相较于匹配的固定扇区扫描减少了扫描时间和冗余覆盖，而GPS参考的实地数据表明，门控递归估计器在定位精度上优于非递归基线。一个集成的海洋任务进一步展示了完整的搜索到导航序列，包括故意触发的视觉丢失恢复。这些结果确立了在沿海GNSS拒绝环境中固定悬停无人机辅助静态目标接近的可行性和操作边界。源代码和视频演示可在 https://github.com/xirhxq/pod_search 和 https://youtu.be/S-5RkJs35JI 上公开获取。
cs.RO / 59 / 2606.29892
Trust Your Instincts: Confidence-Driven Test-Time RL for Vision-Language-Action Models
相信你的直觉：基于信心的测试时强化学习用于视觉-语言-动作模型
Abstract
Reinforcement learning (RL) has become indispensable for pushing Vision-Language-Action Models (VLAs) beyond static imitation learning. However, existing RL methods typically require external environmental feedback, relying on predefined success signals to guide policy updates. In this work, we show that VLA models possess useful internal evaluative capabilities: in discrete-action VLAs, trajectories with higher generation confidence are significantly more likely to succeed. Based on this observation, we introduce T^2VLA (Test-time VLA), an architecture-agnostic test-time RL framework that enables VLA models to achieve self-bootstrapping policy improvement. Instead of relying on external rewards, T^2VLA leverages trajectory-level similarity to high-confidence expert demonstrations as an intrinsic reward signal. In addition, we propose a Confidence-Driven Dual Expert Bootstrapping mechanism, which dynamically balances a Local Pseudo-Expert for exploration and a Global Expert Pool for training stability. Extensive experiments on the LIBERO and RoboTwin benchmarks show that T^2VLA consistently outperforms supervised baselines and approaches oracle RL performance with ground-truth rewards, achieving effective improvement without external reward feedback. Furthermore, T^2VLA adapts to distinct VLA paradigms, including both OpenVLA-OFT and the pi series.
Chinese Translation
强化学习（RL）已成为推动视觉-语言-动作模型（VLA）超越静态模仿学习的不可或缺的工具。然而，现有的RL方法通常需要外部环境反馈，依赖预定义的成功信号来指导策略更新。在本研究中，我们展示了VLA模型具有有用的内部评估能力：在离散动作的VLA中，生成信心更高的轨迹成功的可能性显著更大。基于这一观察，我们引入了T^2VLA（测试时VLA），一种与架构无关的测试时强化学习框架，使VLA模型能够实现自我引导的策略改进。T^2VLA不依赖外部奖励，而是利用与高信心专家演示的轨迹级相似性作为内在奖励信号。此外，我们提出了一种基于信心的双专家自引导机制，动态平衡局部伪专家用于探索和全球专家池用于训练稳定性。在LIBERO和RoboTwin基准上的大量实验表明，T^2VLA始终优于监督基线，并在真实奖励下接近oracle RL性能，实现了在没有外部奖励反馈的情况下的有效改进。此外，T^2VLA能够适应不同的VLA范式，包括OpenVLA-OFT和π系列。
cs.RO / 60 / 2606.29898
Critical Interval MSE: Toward Reliable Offline Validation for Robot Manipulation Policies
关键区间均方误差：迈向机器人操控策略的可靠离线验证
Abstract
Real-world evaluation is the gold standard for robot policies because it tests them against the physical conditions and deployment challenges they are ultimately designed to handle. However, real-world evaluation is also the bottleneck for iterating on robot policies: it is costly, difficult to reproduce, and often too sparse to reliably compare nearby model variants. A straightforward proxy for performance is validation loss on expert demonstrations, but this proxy is often poorly correlated with real-world performance. In this paper, we introduce Critical Interval MSE (CI-MSE), an intuitively simple yet effective offline validation metric. CI-MSE restricts error computation to task-critical segments and pairs it with simple action-alignment procedures that better match rollout-time behavior. Across simulation and real-world experiments, CI-MSE yields a stronger correlation between validation error and rollout performance than raw MSE. Across a wide range of policy checkpoints, CI-MSE achieves a Spearman's rank correlation of $-0.87$, much closer to the ideal value of $-1$ than raw MSE's $-0.61$, demonstrating a significant improvement. We show through sensitivity analysis that our metric is robust to a wide range of hyperparameters. We further study the effectiveness of CI-MSE under evaluation distribution shifts and suggest design boundaries when using this metric. In summary, this paper provides a simple and reliable offline validation tool for accelerating policy iteration. Project webpage: https://ci-mse.github.io/
Chinese Translation
现实世界的评估是机器人策略的金标准，因为它将策略置于最终设计要应对的物理条件和部署挑战之下进行测试。然而，现实世界的评估也是迭代机器人策略的瓶颈：它成本高、难以重现，并且通常稀疏到无法可靠比较相近的模型变体。专家演示的验证损失是性能的一个直接代理，但这个代理通常与现实世界的性能相关性较差。在本文中，我们引入了关键区间均方误差（Critical Interval MSE, CI-MSE），这是一种直观简单但有效的离线验证指标。CI-MSE将误差计算限制在任务关键段，并与简单的动作对齐程序相结合，更好地匹配了滚动时间行为。在模拟和现实世界实验中，CI-MSE在验证误差与滚动性能之间产生了比原始均方误差（MSE）更强的相关性。在广泛的策略检查点中，CI-MSE达到了$-0.87$的斯皮尔曼等级相关系数，远比原始MSE的$-0.61$接近理想值$-1$，显示出显著的改进。我们通过敏感性分析表明，我们的指标对广泛的超参数具有鲁棒性。我们进一步研究了CI-MSE在评估分布变化下的有效性，并在使用该指标时建议设计边界。总之，本文提供了一种简单可靠的离线验证工具，以加速策略迭代。项目网页：https://ci-mse.github.io/
cs.RO / 61 / 2606.29908
Pondering the Way: Spatial-perceiving World Action Model for Embodied Navigation
思考路径：用于具身导航的空间感知世界行动模型
Abstract
Existing world model-based planners for visual navigation typically follow a verification-centric paradigm, decoupling goal intent from trajectory synthesis. This approach suffers from candidate dependence, heavy computational overhead, and inconsistencies between sampled actions and predicted visuals. To address these issues, we propose SWAM (Spatial-perceiving World Action Model), a task-centric joint observation-action generation framework. Given start and goal RGB observations, SWAM performs single-pass inference to simultaneously generate intermediate RGB-D sequences and corresponding action trajectories, promoting goal-consistent trajectory generation and improved spatial feasibility. While SWAM leverages depth pseudo-labels during training to internalize spatial priors, it requires only monocular RGB input at inference time. We further introduce a visual-guided action refinement module and a trajectory-scale regularization loss to enforce fine-grained alignment between motion and visual cues while stabilizing predictions across varying distances. Extensive experiments show that SWAM significantly outperforms state-of-the-art two-stage planners in success rate, trajectory accuracy, and inference efficiency, while demonstrating robust zero-shot generalization to unseen environments.
Chinese Translation
现有的基于世界模型的视觉导航规划器通常遵循以验证为中心的范式，将目标意图与轨迹合成解耦。这种方法存在候选依赖、计算开销大以及采样动作与预测视觉之间的不一致性等问题。为了解决这些问题，我们提出了SWAM（空间感知世界行动模型），这是一种以任务为中心的联合观察-行动生成框架。在给定起始和目标RGB观察的情况下，SWAM进行单次推理，同时生成中间RGB-D序列和相应的行动轨迹，从而促进目标一致的轨迹生成和提高空间可行性。虽然SWAM在训练过程中利用深度伪标签来内化空间先验，但在推理时仅需要单目RGB输入。我们进一步引入了一个视觉引导的行动精细化模块和一个轨迹尺度正则化损失，以强制运动与视觉线索之间的细粒度对齐，同时在不同距离下稳定预测。大量实验表明，SWAM在成功率、轨迹准确性和推理效率上显著优于最先进的两阶段规划器，同时在未见环境中表现出强大的零-shot泛化能力。
cs.RO / 62 / 2606.29910
Sphere-VIO: Fast and Robust Visual-Inertial Odometry via Unified Spherical Representation for Heterogeneous Multi-Camera Systems
Sphere-VIO：通过统一球面表示实现异构多摄像头系统的快速且稳健的视觉惯性里程计
Abstract
Multi-camera visual-inertial odometry (VIO) overcomes the inherent limitations of pure visual systems by expanding the field of view. However, existing algorithms are typically tailored for fixed camera setups and lack unified compatibility with heterogeneous multi-camera systems. Meanwhile, due to the absence of a unified cross-camera representation and association mechanism, current methods struggle to achieve a balance among robust cross-camera feature tracking, stable depth estimation, and reliable real-time performance. To address these issues, we present Sphere-VIO, a lightweight filter-based VIO framework with unified spherical representation for heterogeneous multi-camera systems. Specifically, we first propose a Unified Spherical Panorama Model (USPM) that supports all standard camera models and enables bidirectional fast mapping between multi-camera images and a shared spherical space without sequential stitching, simplifying cross-camera feature management and improving triangulation efficiency. Second, we design a parallel-accelerated depth-guided semi-direct tracking pipeline, namely Hierarchical Omnidirectional Feature Alignment (HOFA), with global spherical constraints for robust cross-camera matching, and fuse multi-camera depth observations into a standard depth filter for stable initialization. Finally, we develop a multi-camera-adapted ESKF backend that employs spherical bearing residuals and Schur complement marginalization to minimize computational overhead, enabling accurate real-time state estimation on resource-constrained devices. Extensive experiments on public benchmarks and a custom omnidirectional dataset show that Sphere-VIO achieves superior trade-offs between accuracy, robustness, efficiency, and cross-camera generality.
Chinese Translation
多摄像头视觉惯性里程计（VIO）通过扩展视野克服了纯视觉系统的固有局限性。然而，现有算法通常针对固定摄像头设置进行优化，缺乏与异构多摄像头系统的统一兼容性。同时，由于缺乏统一的跨摄像头表示和关联机制，当前方法在实现稳健的跨摄像头特征跟踪、稳定的深度估计和可靠的实时性能之间难以取得平衡。为了解决这些问题，我们提出了Sphere-VIO，一种基于轻量级滤波器的VIO框架，具有针对异构多摄像头系统的统一球面表示。具体而言，我们首先提出了一种统一球面全景模型（USPM），该模型支持所有标准摄像头模型，并能够在多摄像头图像和共享球面空间之间进行双向快速映射，而无需顺序拼接，从而简化跨摄像头特征管理并提高三角测量效率。其次，我们设计了一种并行加速的深度引导半直接跟踪管道，即分层全向特征对齐（HOFA），该管道具有全球球面约束以实现稳健的跨摄像头匹配，并将多摄像头深度观测融合到标准深度滤波器中以实现稳定初始化。最后，我们开发了一种适应多摄像头的扩展卡尔曼滤波（ESKF）后端，利用球面方向残差和Schur补充边际化来最小化计算开销，从而在资源受限设备上实现准确的实时状态估计。在公共基准和自定义全向数据集上的大量实验表明，Sphere-VIO在准确性、稳健性、效率和跨摄像头通用性之间实现了优越的权衡。
cs.RO / 63 / 2606.29917
Flying to Image-Specified Objects: 3D Quadrotor Navigation via Cross-Graph Memory and Viewpoint Planning
飞向图像指定对象：通过跨图记忆和视点规划实现的三维四旋翼导航
Abstract
Instance-Specific Image-Goal Navigation (InstanceImageNav) requires a robot to navigate toward the exact object instance depicted in a query image. Extending this task to quadrotors is challenging due to continuous 3D control, limited field of view (FOV), and safety constraints, which make successful navigation highly dependent on selecting informative viewpoints. We propose a hierarchical navigation framework for quadrotor InstanceImageNav that separates high-level decision making from low-level motion execution. Instead of navigating directly to spatial locations, the system generates viewpoint-aware action nodes around frontier regions and potential target objects, enabling the robot to explore while maintaining informative viewpoints for detecting the target instance. A lightweight semantic memory maintains object-level and observation-level context, allowing semantic cues to propagate to candidate action nodes for decision making. A learning-based policy selects the most promising action node, and a trajectory planner generates dynamically feasible 3D flight paths for safe execution. Experiments in simulation demonstrate consistent improvements over strong baselines, and real-world quadrotor flights validate the practicality and robustness of the proposed framework.
Chinese Translation
实例特定图像目标导航（InstanceImageNav）要求机器人导航至查询图像中所描绘的确切对象实例。将这一任务扩展到四旋翼上具有挑战性，因为连续的三维控制、有限的视场（FOV）和安全约束使得成功导航高度依赖于选择信息丰富的视点。我们提出了一种四旋翼实例图像导航的分层导航框架，将高层决策与低层运动执行分开。该系统并非直接导航至空间位置，而是在边界区域和潜在目标对象周围生成视点感知的动作节点，使机器人能够在探索的同时保持信息丰富的视点以检测目标实例。一个轻量级的语义记忆维护对象级和观察级上下文，允许语义线索传播到候选动作节点以进行决策。基于学习的策略选择最有前景的动作节点，而轨迹规划器生成动态可行的三维飞行路径以确保安全执行。模拟实验表明，相较于强基线方法有持续的改进，而现实世界的四旋翼飞行验证了所提框架的实用性和鲁棒性。
cs.RO / 64 / 2606.29934
RoamFlow: Reinforcement-Aligned One-Step Action MeanFlow Policy for Image-Goal Navigation
RoamFlow：强化对齐的一步动作均流策略用于图像目标导航
Abstract
Image-goal navigation is a key challenge in embodied robotics, where an agent must reach a target specified solely by a goal image. While existing reinforcement learning approaches map perceptual observations directly to actions, they struggle to model long-horizon dependencies, often leading to suboptimal trajectories. To address this limitation, we propose RoamFlow, a generative navigation framework that leverages MeanFlow to predict the average velocity field for trajectory synthesis, enabling efficient few-step generation and reducing inference latency. We further adopt a two-stage training strategy that combines expert imitation for stable initialization with reinforcement learning for task-specific policy refinement. Extensive experiments in both Habitat simulation and real-world robotic platforms demonstrate that RoamFlow achieves efficient inference while maintaining strong navigation performance under real-time constraints.
Chinese Translation
图像目标导航是具身机器人领域的一项关键挑战，其中代理必须到达仅由目标图像指定的目标。尽管现有的强化学习方法直接将感知观察映射到动作，但它们在建模长时间依赖关系方面存在困难，常常导致次优轨迹。为了解决这一局限性，我们提出了RoamFlow，一种生成导航框架，利用均流（MeanFlow）预测轨迹合成的平均速度场，从而实现高效的少步生成并减少推理延迟。我们进一步采用了两阶段训练策略，将专家模仿用于稳定初始化与强化学习结合，以进行任务特定的策略优化。在Habitat仿真和真实机器人平台上的大量实验表明，RoamFlow在保持强大导航性能的同时，实现了高效推理，满足实时约束。
cs.RO / 65 / 2606.29936
OpenSPM: An Environment-Transferable Robotic Key Spatial Pose Memory and Closed-Loop High-Frequency Flow-Matching Action Generation Model
OpenSPM：一种可环境转移的机器人关键空间姿态记忆与闭环高频流匹配动作生成模型
Abstract
Open-environment tabletop robotic manipulation requires systems to possess semantic understanding, precise geometric pose estimation, and high-frequency action generation. While end-to-end vision-language-action (VLA) models excel at semantic generalization, they often lack explicit geometric constraints for fine-grained tasks and require costly training. To bridge the gap between high-level semantics and low-level physical execution, we propose OpenSPM, an open environment spatial persistent memory framework consisting of spatial pose memory and flow-matching action generation model. OpenSPM first leverages semantically conditioned 3D perception and Kalman filtering to track continuous 6D poses. It then extracts key spatial poses from human demonstrations, keeping them as transferable, object-centric spatial persistent memory entries. During inference, OpenSPM retrieves relevant memory entries in terms of natural language instructions, transfers the spatial poses to new scenes using SE(3) transformations, and generates high-frequency action chunks via a lightweight conditional flow-matching model. Combined with real-time proprioceptive state feedback and terminal residual correction, the system effectively suppresses trajectory error accumulation. Evaluated on ten LIBERO-GOAL tasks, OpenSPM achieves an 85.6% success rate and an equivalent control frequency of 1033.3 Hz, while requiring minimal inference AI computing power. Extensive ablations illustrate that structured spatial persistent memory and closed-loop residual correction play a crucial role in reliable, high-frequency robotic manipulation.
Chinese Translation
开放环境下的桌面机器人操作要求系统具备语义理解、精确的几何姿态估计和高频动作生成。尽管端到端的视觉-语言-动作（VLA）模型在语义泛化方面表现优异，但它们通常缺乏针对细粒度任务的显式几何约束，并且训练成本高昂。为了弥合高层语义与低层物理执行之间的差距，我们提出了OpenSPM，这是一种开放环境空间持久记忆框架，包含空间姿态记忆和流匹配动作生成模型。OpenSPM首先利用语义条件下的三维感知和卡尔曼滤波来跟踪连续的6D姿态。然后，它从人类演示中提取关键空间姿态，将其作为可转移的以物体为中心的空间持久记忆条目。在推理过程中，OpenSPM根据自然语言指令检索相关的记忆条目，使用SE(3)变换将空间姿态转移到新场景，并通过轻量级条件流匹配模型生成高频动作块。结合实时的本体状态反馈和终端残差校正，该系统有效抑制了轨迹误差的累积。在十个LIBERO-GOAL任务上的评估中，OpenSPM达到了85.6%的成功率和1033.3 Hz的等效控制频率，同时所需的推理AI计算能力极小。大量消融实验表明，结构化的空间持久记忆和闭环残差校正在可靠的高频机器人操作中发挥了至关重要的作用。
cs.RO / 66 / 2606.29937
REPAIR-Bench: A Benchmark for Robot Error Perception And Interaction Recovery
REPAIR-Bench：机器人错误感知与交互恢复的基准测试
Abstract
Understanding how users perceive and respond to robot failures is essential for building robust and trustworthy robot systems. Prior work, however, (i) often treats failures as independent events, (ii) emphasizes binary failure detection, (iii) with rule-based recovery modeling. We present REPAIR-Bench, built on 214 interaction trials from 41 participants, the benchmark spans four induced failure types and provides synchronized facial action units, head pose, speech transcripts, and post-interaction affect and recovery reports. The benchmark spans three novel evaluation tasks that jointly capture the lifecycle of failure in human-robot interaction (HRI): (i) failure detection over inter-dependent interaction sessions, modeling longitudinal user adaptation across repeated failures; (ii) visual failure-type classification beyond binary success/failure formulations; and (iii) user-centered recovery prediction, inferring users' preferred recovery strategies from interaction context rather than relying on manually designed or rule-based strategies. In baseline experiments, hierarchical recurrent modeling improved failure detection over a single-session model (strict F1: 0.80 vs. 0.68), achieved a failure localization mean signed error of -0.51 s, median absolute error of 2.97 s and, for recovery prediction, a QLoRA-tuned Mistral-7B reached Hit@5=0.76 and F1@5=0.32. REPAIR-Bench provides both the HRI and Medical HRI communities with a standardized framework for (1) evaluating robot failures and (2) building transparent, adaptive, and trustworthy recovery systems.
Chinese Translation
理解用户如何感知和应对机器人故障对于构建稳健和可信赖的机器人系统至关重要。然而，先前的研究往往（i）将故障视为独立事件，（ii）强调二元故障检测，（iii）采用基于规则的恢复建模。我们提出了REPAIR-Bench，该基准基于41名参与者的214次交互试验，涵盖四种诱发的故障类型，并提供同步的面部动作单元、头部姿态、语音转录以及交互后的情感和恢复报告。该基准包含三个新颖的评估任务，联合捕捉人机交互（HRI）中故障的生命周期：（i）在相互依赖的交互会话中进行故障检测，建模用户在重复故障中的长期适应；（ii）超越二元成功/失败形式的视觉故障类型分类；以及（iii）以用户为中心的恢复预测，从交互上下文推断用户的首选恢复策略，而不是依赖手动设计或基于规则的策略。在基线实验中，层次递归建模提高了单会话模型的故障检测（严格F1: 0.80 vs. 0.68），实现了故障定位的平均签名误差为-0.51秒，中位绝对误差为2.97秒；在恢复预测方面，经过QLoRA调优的Mistral-7B达到了Hit@5=0.76和F1@5=0.32。REPAIR-Bench为HRI和医疗HRI社区提供了一个标准化框架，以（1）评估机器人故障和（2）构建透明、适应性强且可信赖的恢复系统。
cs.RO / 67 / 2606.29940
WARP: Whole-Body Retargeting for Learning from Offline Human Demonstrations
WARP：基于离线人类示范的全身重定向学习
Abstract
Direct transfer from human demonstration to learnable robot action is a crucial step towards scalable whole-body mobile manipulation. While human data scales better than mobile teleoperation, it requires overcoming significant embodiment gaps. Existing retargeting methods yield imprecise or inconsistent solutions, causing action multi-modality that prevents supervised policies from reliably converging. We present Whole-body-Aware Retargeting from human Pose (WARP), an offline pipeline that explicitly models embodiment differences to extract precise, unique whole-body actions. WARP leverages a closed-form Shoulder-Elbow-Wrist (SEW) geometric solver for exact end-effector tracking while preserving whole-body structural intent. Paired with lazy mobile-base control, it extracts accurate, consistent robot trajectories. Evaluations show WARP provides highly reliable data for open-loop real-world replay. To our knowledge, WARP is the first framework to achieve zero-shot whole-body mobile manipulation directly from offline human demonstrations, eliminating the need for human-in-the-loop teleoperation action data. More details on https://warp-retarget.github.io/
Chinese Translation
从人类示范直接转移到可学习的机器人动作是实现可扩展全身移动操作的重要一步。尽管人类数据在规模上优于移动遥控，但这需要克服显著的体现差距。现有的重定向方法产生不精确或不一致的解决方案，导致动作多模态性，从而阻碍监督策略的可靠收敛。我们提出了基于人类姿态的全身感知重定向（Whole-body-Aware Retargeting from human Pose，WARP），这是一种离线管道，明确建模体现差异，以提取精确、独特的全身动作。WARP利用封闭形式的肩肘腕（Shoulder-Elbow-Wrist，SEW）几何求解器进行精确的末端执行器跟踪，同时保持全身结构意图。结合懒惰的移动基座控制，它提取准确、一致的机器人轨迹。评估结果表明，WARP为开放环节的真实世界重放提供了高度可靠的数据。据我们所知，WARP是第一个能够直接从离线人类示范中实现零样本全身移动操作的框架，消除了对人类参与的遥控动作数据的需求。更多详细信息请访问 https://warp-retarget.github.io/
cs.RO / 68 / 2606.29941
Seeing Touch from Motion: A Unified Modality-Aware Visuo-Tactile Policy with Tactile Motion Correlation
从运动中感知触觉：一种统一的模态感知视觉-触觉策略与触觉运动相关性
Abstract
Visuo-Tactile policies leveraging optical tactile sensors have shown great promise in contact-rich manipulation. These sensors achieve high spatial resolution and multi-dimensional force sensing by utilizing an internal camera to monitor the deformation of their elastic gel surface, thereby indirectly inferring tactile cues. Despite their advantages, extracting fine-grained contact states necessary for contact-rich manipulation remains an open challenge. Existing methods typically use either raw images or cumulative motion fields to represent tactile cues. However, both are prone to perception ambiguity. Raw tactile images mainly capture appearance changes, while cumulative motion fields only reflect the aggregate gel deformation. Consequently, distinct fine-grained contact states can exhibit highly similar patterns, making it difficult to explicitly distinguish subtle contact variations. To address this issue, we explore the dynamic priors of tactile motion and discover that the correlation between transient and cumulative motion can explicitly distinguish fine-grained contact states. Based on this insight, we propose a motion-aware tactile representation to facilitate contact-rich manipulation. Beyond tactile representation, effective fusion of tactile and visual modalities is also critical. Most existing fusion methods either directly concatenate features from each modality or train modality-specific networks separately and fuse their outputs. However, these strategies struggle to simultaneously model cross-modal interactions and preserve modality-specific characteristics. In this work, we take advantage of the Mixture-of-Transformers architecture and propose a unified modality-aware visuo-tactile policy that captures cross-modal complementarity while maintaining modality-specific properties.
Chinese Translation
利用光学触觉传感器的视觉-触觉策略在接触丰富的操作中展现出巨大的潜力。这些传感器通过内部相机监测其弹性胶表面的变形，实现高空间分辨率和多维力感知，从而间接推断触觉线索。尽管具有这些优势，提取进行接触丰富操作所需的细粒度接触状态仍然是一个未解决的挑战。现有方法通常使用原始图像或累积运动场来表示触觉线索。然而，这两者都容易导致感知模糊。原始触觉图像主要捕捉外观变化，而累积运动场仅反映胶体的整体变形。因此，不同的细粒度接触状态可能表现出高度相似的模式，使得明确区分微妙的接触变化变得困难。为了解决这个问题，我们探索了触觉运动的动态先验，发现瞬态运动与累积运动之间的相关性可以明确区分细粒度接触状态。基于这一见解，我们提出了一种运动感知的触觉表示，以促进接触丰富的操作。除了触觉表示，触觉和视觉模态的有效融合也至关重要。大多数现有的融合方法要么直接连接来自每个模态的特征，要么分别训练模态特定的网络并融合其输出。然而，这些策略在同时建模跨模态交互和保持模态特定特性方面存在困难。在本研究中，我们利用混合变换器架构，提出了一种统一的模态感知视觉-触觉策略，能够捕捉跨模态互补性，同时保持模态特定属性。
cs.RO / 69 / 2606.29948
Heterogeneous Tactile Transformer
异构触觉变换器
Abstract
Tactile sensors are inherently heterogeneous: a model trained on one sensor cannot be directly used on another, which limits learning contact-rich manipulation policies from diverse tactile data at scale. To bridge this gap, we propose the Heterogeneous Tactile Transformer (HTT), a framework that learns shared tactile representations across heterogeneous sensors. HTT consists of sensor-specific encoders and a shared transformer trunk, and is pretrained with per-modality masked reconstruction together with cross-modal alignment between paired sensors. Pretraining uses our novel Heterogeneous Paired Tactile (HPT) dataset, containing 1.6M synchronized paired frames across four vision- and array-based tactile sensors. Across distinct tactile perception and real-world manipulation tasks, HTT is shown to learn transferable representations that adapt to new tasks and previously unseen sensors. Dataset, code, and model checkpoints will be released upon publication at https://jxbi1010.github.io/htt-gh-page/.
Chinese Translation
触觉传感器本质上是异构的：在一个传感器上训练的模型不能直接用于另一个，这限制了从多样化触觉数据中大规模学习丰富接触操作策略的能力。为了解决这一问题，我们提出了异构触觉变换器（Heterogeneous Tactile Transformer，HTT），这是一个跨异构传感器学习共享触觉表征的框架。HTT由特定于传感器的编码器和一个共享的变换器主干组成，并通过每种模态的掩蔽重建以及配对传感器之间的跨模态对齐进行预训练。预训练使用我们新颖的异构配对触觉（Heterogeneous Paired Tactile，HPT）数据集，该数据集包含来自四种基于视觉和阵列的触觉传感器的160万帧同步配对数据。在不同的触觉感知和现实世界操作任务中，HTT被证明能够学习可迁移的表征，适应新任务和以前未见过的传感器。数据集、代码和模型检查点将在出版时发布，网址为 https://jxbi1010.github.io/htt-gh-page/。
cs.RO / 70 / 2606.30101
SIR: Structured Image Representations for Explainable Robot Learning
SIR：用于可解释机器人学习的结构化图像表示
Abstract
Existing robot policies based on learned visual embeddings lack explicit structure and are sensitive to visual distractions. Thus, the representations that drive their behaviour are often opaque, making their decision-making process difficult to interpret. To address this, we introduce Structured Image Representations (SIR), a method that leverages Scene Graphs (SGs) as an intermediate representation for robot policy learning. Our approach first constructs a fully connected graph, using image-derived features as initial node representations. Then, a module learns to sparsify this graph end-to-end, creating a task-relevant sub-graph that is passed to the action generation model. This process makes our model intrinsically explainable. Evaluations on RoboCasa show that our sparse graph policies outperform image-based baselines on average with 19.5% vs 14.81% success rate. Most importantly, we show that the learned sparse graphs are a powerful tool for model analysis. By analysing when the model's sub-graph deviates from human expectation, such as by including distractor nodes or omitting key objects, we successfully uncover dataset biases, including spurious correlations and positional biases. https://github.com/intuitive-robots/SIR_Model
Chinese Translation
现有基于学习的视觉嵌入的机器人策略缺乏明确的结构，并且对视觉干扰敏感。因此，驱动其行为的表示通常是不透明的，使得其决策过程难以解释。为了解决这个问题，我们提出了结构化图像表示（Structured Image Representations，SIR），这是一种利用场景图（Scene Graphs，SGs）作为机器人策略学习的中间表示的方法。我们的方法首先构建一个完全连接的图，使用图像派生特征作为初始节点表示。然后，一个模块学习以端到端的方式稀疏化该图，创建一个与任务相关的子图，并将其传递给动作生成模型。这个过程使我们的模型在本质上是可解释的。在RoboCasa上的评估表明，我们的稀疏图策略在成功率上平均优于基于图像的基线，成功率为19.5%对比14.81%。最重要的是，我们展示了学习到的稀疏图是模型分析的强大工具。通过分析模型的子图何时偏离人类期望，例如通过包含干扰节点或省略关键对象，我们成功揭示了数据集偏差，包括虚假相关性和位置偏差。
cs.RO / 71 / 2606.30109
TacEvo: Self-Evolving Architecture Discovery for Robotic Tactile Perception via LLM-Driven Quality-Diversity Search
TacEvo：通过LLM驱动的质量-多样性搜索实现机器人触觉感知的自我进化架构发现
Abstract
Vision-based tactile sensing converts contact-induced surface deformation into images, enabling robots to infer contact forces and fine surface textures that are not accessible through conventional vision alone. However, tactile images are sensor- and physics-specific, so effective architectures often require expert intuition and extensive manual iteration. Existing neural architecture search (NAS) pipelines can reduce this burden, but they are often computationally expensive and restricted to hand-designed search spaces, which limits architectural novelty and diversity. We introduce TacEvo, a self-evolving architecture discovery framework that improves network designs from downstream feedback. TacEvo uses an LLM to generate code-level mutations and crossovers, and a MAP-Elites quality-diversity loop that preserves diverse elite architectures while preferentially reusing prompts that consistently yield improvements. Exploration is guided by two behavioural descriptors, Architectural Diversity and Efficiency Ratio, which encourage coverage across structural variations and compute-size trade-offs. On ViTacTip force regression and grating classification, TacEvo achieves high autonomous generation reliability (96.0%/94.5% trainable) and improves best validation fitness over 20 generations by 56.1%/96.1%. In a 20-seed post-search high-fidelity evaluation, TacEvo matches the expert baseline on force prediction and outperforms it on fine-grained grating classification. These results suggest that LLM-driven self-evolving search constitutes a practical paradigm for AI-assisted scientific discovery in specialised robotic sensing.
Chinese Translation
基于视觉的触觉感知将接触引起的表面变形转换为图像，使机器人能够推断接触力和细微的表面纹理，这些信息是传统视觉无法获取的。然而，触觉图像是特定于传感器和物理的，因此有效的架构往往需要专家的直觉和大量的手动迭代。现有的神经架构搜索（NAS）流程可以减轻这一负担，但通常计算成本高且受限于手工设计的搜索空间，这限制了架构的新颖性和多样性。我们提出了TacEvo，一种自我进化的架构发现框架，通过下游反馈改进网络设计。TacEvo使用LLM生成代码级的变异和交叉，并采用MAP-Elites质量-多样性循环，保留多样化的精英架构，同时优先重用那些持续产生改进的提示。探索由两个行为描述符引导，即架构多样性和效率比，这鼓励在结构变体和计算规模权衡之间的覆盖。在ViTacTip力回归和光栅分类任务中，TacEvo实现了高达96.0%/94.5%的自主生成可靠性，并在20代中将最佳验证适应度提高了56.1%/96.1%。在20个种子的后搜索高保真评估中，TacEvo在力预测上与专家基线相匹配，并在细粒度光栅分类上超越了专家。这些结果表明，LLM驱动的自我进化搜索构成了AI辅助科学发现在专业机器人感知中的一种实用范式。
cs.RO / 72 / 2606.30111
Automating the Design of Embodied AgentArchitectures
自动化设计具身代理架构
Abstract
Embodied agents are typically built as hand-designed compositions of perception, memory, planning, and action modules. This modularity exposes a large architectural design space, but current systems still rely on researcher intuition to choose where information is stored, how observations are processed, and how model calls are connected. Agent Architecture Search (AAS) automates such design for text-domain agents, but has not been systematically evaluated on perceptual embodied agents through simulator rollouts. We study this transfer. We introduce AgentCanvas, a typed-graph runtime that hosts embodied executors as editable node-and-wire programs with simulator-aware execution and episode-level logs, and KDLoop, a coding-agent search procedure that cycles through proposal, critique, experiment, and distillation, with triggered reflection after stalls. We evaluate three AAS variants across four embodied executors spanning vision-language navigation, embodied question answering, and language-conditioned manipulation. The resulting 3x4 matrix shows that architecture-level search can produce deployable and directional success-rate gains on embodied tasks, while one apparent high-scoring candidate is rejected as leak-bearing. At the same time, the experiments expose constraints that are muted in text-domain AAS: optimization signals can be masked by rollout noise, search can become trapped in local edit basins, and episode-level credit assignment only partially emerges even when detailed logs are available. These results characterize both the promise and the current limits of automated architecture search for embodied agents.
Chinese Translation
具身代理通常被构建为感知、记忆、规划和行动模块的手工设计组合。这种模块化暴露了一个广泛的架构设计空间，但当前系统仍依赖研究者的直觉来选择信息存储位置、观察处理方式以及模型调用的连接方式。代理架构搜索（Agent Architecture Search, AAS）自动化了文本领域代理的设计，但尚未通过模拟器回合对感知具身代理进行系统评估。我们研究了这种转移。我们引入了AgentCanvas，一个类型图运行时，作为可编辑的节点和连接程序托管具身执行器，具有模拟器感知的执行和回合级日志，以及KDLoop，一个编码代理搜索程序，通过提案、批评、实验和提炼循环进行，并在停滞后触发反思。我们评估了三种AAS变体，涵盖了四个具身执行器，涉及视觉-语言导航、具身问答和语言条件下的操作。结果的3x4矩阵表明，架构级搜索能够在具身任务上产生可部署和方向性的成功率提升，而一个明显高得分的候选者被拒绝为存在泄漏。同时，实验揭示了在文本领域AAS中被掩盖的约束：优化信号可能被回合噪声掩盖，搜索可能陷入局部编辑盆地，即使在详细日志可用时，回合级信用分配也只部分出现。这些结果表征了自动化架构搜索在具身代理中的潜力和当前限制。
cs.RO / 73 / 2606.30113
SA-VLA: State-aware tokenizer for improving Vision-Language-Action Models' performance
SA-VLA：一种状态感知的标记器，以提高视觉-语言-动作模型的性能
Abstract
Discrete action tokenization provides a compact interface for autoregressive VLA policies, but accurately recovering continuous robot actions from discrete codes remains challenging. Existing tokenizers typically map each discrete code to a fixed continuous action prototype, ignoring the robot's current proprioceptive state. This limitation is particularly pronounced in manipulation, where the same action token may require different continuous controls under different joint configurations, object poses, and contact conditions. We therefore propose SA-VLA, a state-aware action tokenizer that conditions action decoding on robot state. We study two state-injection mechanisms for VQ-based action tokenization: cross-attention between state and action features, and a lightweight state adapter that predicts action-wise modulation factors for state-conditioned action modulation and reconstruction. The adapter formulation expands the effective support of a finite codebook by allowing each discrete token to represent a family of state-dependent continuous actions, while preserving the efficiency and compatibility of discrete action modeling. Integrated into an LLM-based VLA policy, SA-VLA supports both autoregressive and parallel action-token decoding with minimal changes to the model interface. On 12 RoboTwin manipulation tasks, SA-VLA improves the average success rate from 0.29 to 0.56 over the strongest tokenizer baseline. In zero-shot sim-to-real experiments on three real-world tasks, it further improves average success from 0.15 to 0.33 over the strongest tokenizer baseline. These results demonstrate that state-conditioned action decoding is a simple and effective mechanism for reducing the compression gap in discrete VLA policies.
Chinese Translation
离散动作标记化为自回归的视觉-语言-动作（VLA）策略提供了一个紧凑的接口，但从离散代码准确恢复连续机器人动作仍然具有挑战性。现有的标记器通常将每个离散代码映射到一个固定的连续动作原型，忽略了机器人的当前本体状态。这一局限性在操作中尤为明显，因为在不同的关节配置、物体姿态和接触条件下，相同的动作标记可能需要不同的连续控制。因此，我们提出了SA-VLA，一种状态感知的动作标记器，它在动作解码时考虑机器人的状态。我们研究了两种基于矢量量化（VQ）的动作标记化的状态注入机制：状态与动作特征之间的交叉注意力，以及一个轻量级的状态适配器，该适配器预测用于状态条件下动作调制和重建的动作级调制因子。适配器的形式扩展了有限代码本的有效支持，允许每个离散标记表示一系列依赖于状态的连续动作，同时保持离散动作建模的效率和兼容性。集成到基于大型语言模型（LLM）的VLA策略中，SA-VLA支持自回归和并行动作标记解码，对模型接口的更改最小。在12个RoboTwin操作任务中，SA-VLA将平均成功率从0.29提高到0.56，超越了最强标记器基线。在三个真实世界任务的零样本仿真到现实实验中，它进一步将平均成功率从0.15提高到0.33，超越了最强标记器基线。这些结果表明，状态条件下的动作解码是一种简单而有效的机制，用于缩小离散VLA策略中的压缩差距。
cs.RO / 74 / 2606.30151
AERIS: Aerial-Edge Role-Driven Intelligence at Runtime via Orchestrated Language-Model Swarm
AERIS：通过协调的语言模型群在运行时实现的空中边缘角色驱动智能
Abstract
Integrating large language models into robotic systems holds promise for enhancing autonomy, yet practical deployment remains constrained by strict heartbeat-constrained scheduling and limited computational power. We propose AERIS: an edge deployment framework for aerial platforms. It organizes dedicated small language models combined with lightweight perception and control modules into roles that can be instantiated at runtime, and dynamically rebinds them across different executors as resources change, thereby pushing intelligent capabilities to the edge. AERIS achieves long-horizon instruction decomposition through an attention-subgoal alignment mechanism, which involves annotating the currently active instruction step in messages, thereby progressively approaching long-term objectives. We evaluate AERIS on a high-fidelity UAV Vision-and-Language Navigation benchmark. Under a heartbeat-timed execution mechanism, AERIS maintains a stable perception-decision-control loop between a low-frequency planner and a high-frequency controller, supporting real-time closed-loop operation. We further validate its deployability through two real-world experiments focused on planning and fast response. A demonstration video is provided in the supplementary materials.
Chinese Translation
将大型语言模型集成到机器人系统中有望增强自主性，但实际部署仍受限于严格的心跳约束调度和有限的计算能力。我们提出了AERIS：一个用于空中平台的边缘部署框架。它将专用的小型语言模型与轻量级感知和控制模块组合成可以在运行时实例化的角色，并在资源变化时动态重新绑定这些角色，从而将智能能力推向边缘。AERIS通过注意力-子目标对齐机制实现长期指令分解，该机制涉及在消息中标注当前活动的指令步骤，从而逐步接近长期目标。我们在一个高保真度的无人机视觉与语言导航基准上评估了AERIS。在心跳定时执行机制下，AERIS在低频率规划器和高频率控制器之间维持稳定的感知-决策-控制循环，支持实时闭环操作。我们通过两个专注于规划和快速响应的真实世界实验进一步验证了其可部署性。补充材料中提供了演示视频。
cs.RO / 75 / 2606.30166
Self-supervised Geometry Reasoning for LiDAR Simultaneous Localization and Mapping
自监督几何推理用于激光雷达同时定位与地图构建
Abstract
LiDAR simultaneous localization and mapping (SLAM) relies on local geometric quantities such as covariances, correspondences, and surface structures. However, most existing pipelines rely on hand-crafted estimates of local geometry and use them as fixed inputs to LiDAR SLAM, which can make the estimated local geometry noisy and unstable in sparse regions of a point cloud or when using low-resolution LiDAR. To address this issue, this paper introduces a self-supervised framework that learns an explicit symbolic representation of local geometry and uses it to improve LiDAR SLAM recursively. Specifically, each point is represented as a Gaussian distribution, allowing local geometry to be described by a covariance. Without dense geometry labels or ground-truth poses, the framework learns by maximizing the likelihood of local geometry, with self-supervision derived from consistency relations over symbolic geometric representations, including predicted covariances, correspondences, and trajectory from SLAM. The learned geometry is then fed back into LiDAR SLAM, forming a reciprocal loop in which improved geometry enhances localization and mapping, and improved localization provides cleaner supervision for subsequent geometry reasoning. This framework is backend-agnostic and can be plugged into existing LiDAR SLAM pipelines without architectural changes. Experiments on KITTI under varying LiDAR resolutions show that the proposed method improves both odometry and global registration.
Chinese Translation
激光雷达同时定位与地图构建（SLAM）依赖于局部几何量，如协方差、对应关系和表面结构。然而，大多数现有的流程依赖于手工制作的局部几何估计，并将其作为固定输入用于激光雷达SLAM，这可能导致在点云稀疏区域或使用低分辨率激光雷达时，估计的局部几何变得嘈杂和不稳定。为了解决这个问题，本文提出了一种自监督框架，该框架学习局部几何的显式符号表示，并利用它递归地改善激光雷达SLAM。具体而言，每个点被表示为一个高斯分布，从而允许用协方差来描述局部几何。在没有稠密几何标签或真实姿态的情况下，该框架通过最大化局部几何的似然性进行学习，自监督来自于符号几何表示的一致性关系，包括预测的协方差、对应关系和来自SLAM的轨迹。学习到的几何随后反馈到激光雷达SLAM中，形成一个互惠循环，其中改善的几何增强了定位和地图构建，而改善的定位为后续的几何推理提供了更干净的监督。该框架与后端无关，可以无须架构更改地插入现有的激光雷达SLAM流程中。在不同激光雷达分辨率下的KITTI实验表明，所提出的方法改善了里程计和全局配准。
cs.RO / 76 / 2606.30243
KYON: Semi-Modular Wheel-Legged Quadruped With Agile Bimanual Capability
KYON：具有灵活双手能力的半模块化轮腿四足机器人
Abstract
This paper presents KYON, a hybrid wheel-legged quadruped robot equipped with a bimanual upper body for loco-manipulation tasks. The platform features a semi-modular design with a reconfigurable lower legs, enabling both wheeled and legged locomotion depending on the environment. A design approach that places actuators in the base and uses transmission mechanisms reduces distal inertia, improving agility and dynamic performance. The robot integrates a whole-body control framework together with a reinforcement learning based policy to handle nonlinear dynamics and enhance robustness to disturbances for the execution of locomotion and manipulation tasks, independently. Experimental results demonstrate effective dynamic locomotion and bimanual manipulation, validating the platform's capability to operate in complex and unstructured scenarios.
Chinese Translation
本文介绍了KYON，一种混合轮腿四足机器人，配备了用于运动操控任务的双手上半身。该平台采用半模块化设计，具有可重构的下肢，能够根据环境实现轮式和腿式运动。将执行器放置在基座并使用传动机制的设计方法减少了远端惯性，提高了敏捷性和动态性能。该机器人集成了全身控制框架以及基于强化学习的策略，以处理非线性动态并增强对干扰的鲁棒性，从而独立执行运动和操控任务。实验结果证明了有效的动态运动和双手操控，验证了该平台在复杂和非结构化场景下的操作能力。
cs.RO / 77 / 2606.30268
ConCent: Contact-Centric Real-to-Sim-to-Real Learning from One Demonstration
ConCent：基于接触的从真实到仿真再到真实的单次演示学习
Abstract
Sim-to-real policy transfer -- deploying policies trained in simulation in the real world -- is a promising paradigm for scaling robot manipulation without large-scale real-world data. However, transferring simulation-trained policies remains challenging due to discrepancies in contact dynamics -- particularly in contact-rich tasks where subtle differences can alter task outcomes entirely. Because interaction between the manipulated object and the environment is mediated through contact, task success depends on accurately reproducing task-relevant contacts. Accordingly, in manipulation, contact-centric fidelity -- reproducing both the contact event sequence (when, where, and how contacts occur) and the local contact dynamics (how forces and motions evolve at each contact) -- is a necessary condition for task success. Based on this insight, we propose a contact-centric real-to-sim-to-real RL framework that uses task-relevant contact event sequences extracted from real demonstrations as the learning objective. We approximate objects as groups of primitives and optimize their contact geometry in simulation so that the resulting local contact dynamics explain the observed state transitions. The contact event sequence is automatically extracted by replaying the demonstration. This sequence serves as a structured reward signal, guiding the policy toward physically plausible contact regimes validated in reality and preventing exploitation of unrealistic simulator contacts. The signal is obtained automatically, requiring no per-task reward design. Experiments on contact-rich manipulation tasks demonstrate more stable and robust sim-to-real policy transfer compared to unconstrained RL baselines.
Chinese Translation
从仿真到真实的策略迁移——在现实世界中部署在仿真中训练的策略——是一个有前景的范式，可以在没有大规模真实数据的情况下扩展机器人操作。然而，由于接触动态的差异，迁移仿真训练的策略仍然具有挑战性，特别是在接触丰富的任务中，细微的差异可能会完全改变任务结果。由于被操作物体与环境之间的交互是通过接触介导的，因此任务的成功依赖于准确再现与任务相关的接触。相应地，在操作中，接触中心的保真度——再现接触事件序列（接触发生的时间、地点和方式）以及局部接触动态（每次接触时力和运动的演变）——是任务成功的必要条件。基于这一洞察，我们提出了一种接触中心的从真实到仿真再到真实的强化学习框架，该框架使用从真实演示中提取的任务相关接触事件序列作为学习目标。我们将物体近似为原始体的组合，并在仿真中优化它们的接触几何，以便生成的局部接触动态能够解释观察到的状态转变。接触事件序列通过重放演示自动提取。该序列作为结构化奖励信号，引导策略朝向在现实中验证的物理上合理的接触状态，并防止对不现实的仿真接触的利用。该信号是自动获得的，无需为每个任务设计奖励。在接触丰富的操作任务上的实验表明，与不受限制的强化学习基线相比，仿真到真实的策略迁移更为稳定和鲁棒。
cs.RO / 78 / 2606.30275
ActiveVital: Geometry-Aware Embodied Vital Signs Monitoring for Home Healthcare Robots
ActiveVital：面向家庭医疗机器人的几何感知体征监测
Abstract
Home robots require reliable vital signs monitoring to support long-term companionship and safety in daily environments, yet obtaining respiration and heart rate without physical contact remains challenging in unconstrained home settings. Millimeter-wave (mmWave) radar offers a promising solution due to its phase sensitivity to sub-millimeter motions. However, mmWave measurements are fundamentally constrained by observation geometry, since only the radial component of motion is observable. Consequently, arbitrary robot-human orientations often introduce angular misalignment that destabilizes vital signs estimation. To address this limitation, we reformulate vital signs monitoring from passive signal recovery to active geometric regulation. We propose ActiveVital, a vision-guided sensing framework that treats sensing geometry as an explicit control variable for robots. It localizes the chest anchor via visual keypoints and converts alignment errors into control commands. This steers the robot-mounted radar toward near-normal incidence to the thoracic surface, maximizing radial observability within a perception-action loop. A differential phase enhancement module further stabilizes signal extraction under motion. Experiments show that ActiveVital reduces respiration interval error from 0.87 s to 0.14 s and heart rate error from 13.59 bpm to 2.22 bpm, achieving accuracy comparable to controlled static sensing while remaining robust under unconstrained robot-human configurations.
Chinese Translation
家庭机器人需要可靠的生命体征监测，以支持长期陪伴和日常环境中的安全性，但在不受限制的家庭环境中获取无接触的呼吸和心率仍然具有挑战性。毫米波（mmWave）雷达由于其对亚毫米运动的相位敏感性，提供了一种有前景的解决方案。然而，mmWave测量在本质上受到观察几何的限制，因为只有运动的径向分量是可观察的。因此，任意的机器人-人类方向往往会引入角度失调，从而使生命体征估计不稳定。为了解决这一限制，我们将生命体征监测从被动信号恢复重新构建为主动几何调节。我们提出了ActiveVital，一个视觉引导的感知框架，将感知几何视为机器人的一个显式控制变量。它通过视觉关键点定位胸部锚点，并将对齐误差转换为控制命令。这使得机器人安装的雷达朝向胸部表面近乎法向入射，从而在感知-行动循环中最大化径向可观察性。差分相位增强模块进一步在运动下稳定信号提取。实验表明，ActiveVital将呼吸间隔误差从0.87秒降低到0.14秒，心率误差从13.59 bpm降低到2.22 bpm，达到了与受控静态感知相当的准确性，同时在不受限制的机器人-人类配置下保持了鲁棒性。
cs.RO / 79 / 2606.30290
X-Morph: Human Motion Priors for Scalable Robot Learning Across Morphologies
X-Morph：跨形态可扩展机器人学习的人类运动先验
Abstract
Recent progress in humanoid behavior models has been driven in large part by abundant human motion data, but comparable motion data is scarce for non-humanoid legged robots such as quadrupeds, hexapods, and quadruped manipulators. A promising alternative is to repurpose human motion across embodiments; however, direct retargeting often produces motions that are visually plausible yet physically inconsistent or difficult to track under robot dynamics. We present X-Morph, a human-motion-to-robot-behavior pipeline that converts human motion into deployable locomotion and loco-manipulation policies for diverse non-humanoid legged morphologies. A cross-morphology retargeting stage converts human motions into kinematically plausible, intent-preserving robot references, which are then tracked by a privileged RL policy and distilled into a causal student policy. We evaluate X-Morph on three morphologically distinct platforms: a quadruped, a hexapod, and a quadruped equipped with a manipulator. The resulting policies track diverse retargeted motions, generalize to unseen human motions, and support downstream use cases including video-based teleoperation, behavior-prior control, and text-conditioned motion generation. These results suggest that large-scale human motion can serve as a substrate for learning broad, reusable behavior priors beyond humanoid robots. Project page: https://maker-rat.github.io/morph/
Chinese Translation
近年来，人形行为模型的进展在很大程度上得益于丰富的人类运动数据，但对于非人形的腿部机器人（如四足机器人、六足机器人和四足操控器）而言，类似的运动数据却相对稀缺。一种有前景的替代方案是跨形态重用人类运动；然而，直接的重定向往往会产生视觉上合理但在物理上不一致或在机器人动态下难以跟踪的运动。我们提出了X-Morph，一个将人类运动转换为可部署的非人形腿部形态的运动和操控策略的管道。跨形态重定向阶段将人类运动转换为运动学上合理、意图保留的机器人参考，然后由特权强化学习（RL）策略进行跟踪，并提炼为因果学生策略。我们在三种形态上明显不同的平台上评估X-Morph：一个四足机器人、一个六足机器人和一个配备操控器的四足机器人。最终的策略能够跟踪多样的重定向运动，推广到未见过的人类运动，并支持包括基于视频的远程操作、行为先验控制和文本条件的运动生成等下游应用。这些结果表明，大规模的人类运动可以作为学习超越人形机器人的广泛、可重用行为先验的基础。项目页面：https://maker-rat.github.io/morph/
cs.RO / 80 / 2606.30293
CSAR: Containerized System Architecture for Robotics
CSAR：面向机器人技术的容器化系统架构
Abstract
Robotic applications increasingly rely on distributed computational infrastructures that combine embedded devices, edge servers, and cloud resources. This evolution, together with the collaborative nature of robotics projects, has made the development, integration, deployment, and long-term operation of robotic systems significantly more complex. In practice, multi-user robotics software teams face persistent challenges related to dependency isolation, compatibility, reproducibility, efficient sharing of specialized hardware, and deployment across heterogeneous environments. In this paper, we present CSAR (Containerized System Architecture for Robotics), a container-centric architectural framework designed specifically for robotics teams and the edge-cloud continuum. CSAR combines LXC/LXD-based system containerization, ROS 2/DDS-based communication, and a three-layer edge infrastructure to organize computation into hardware-affine, persistent execution environments that remain decoupled from the volatility of experimental workloads. Through its Infrastructure Core, Platform and Multi-User Orchestration, and Compute and Acceleration layers, CSAR provides strong isolation, controlled resource sharing, and topology-aware networking for distributed robotic applications. To demonstrate its validity, we describe a real deployment of CSAR in an academic robotics laboratory and evaluate it through representative use cases involving edge-offloaded 3D SLAM and GPU-accelerated semantic mapping. The results indicate that CSAR simplifies software integration, improves the utilization of shared computational resources, and facilitates safe prototyping, as well as reproducible and collaborative experimentation in robotics teams. The implementation described in this paper, including deployment templates, configuration files, and documentation, is available at https://github.com/goyoambrosio/CSAR.
Chinese Translation
机器人应用日益依赖于分布式计算基础设施，这些基础设施结合了嵌入式设备、边缘服务器和云资源。这一演变，加上机器人项目的协作性质，使得机器人系统的开发、集成、部署和长期运行变得更加复杂。在实践中，多用户机器人软件团队面临着与依赖隔离、兼容性、可重现性、高效共享专用硬件以及在异构环境中部署相关的持续挑战。本文提出了CSAR（Containerized System Architecture for Robotics），一个专为机器人团队和边缘-云连续体设计的以容器为中心的架构框架。CSAR结合了基于LXC/LXD的系统容器化、基于ROS 2/DDS的通信以及三层边缘基础设施，将计算组织成与硬件紧密相关的、持久的执行环境，这些环境与实验工作负载的波动保持解耦。通过其基础设施核心、平台与多用户编排、计算与加速层，CSAR为分布式机器人应用提供了强隔离、受控资源共享和拓扑感知网络。为了验证其有效性，我们描述了CSAR在一个学术机器人实验室的实际部署，并通过涉及边缘卸载的3D SLAM和GPU加速的语义映射的代表性用例进行了评估。结果表明，CSAR简化了软件集成，提高了共享计算资源的利用率，并促进了安全原型制作以及机器人团队中的可重现和协作实验。本文描述的实现，包括部署模板、配置文件和文档，可在 https://github.com/goyoambrosio/CSAR 获取。
cs.RO / 81 / 2606.30318
Chronos: A Physics-Informed Full-History Framework for Non-Markovian Long-Horizon Manipulation
Chronos：一种物理信息驱动的全历史框架用于非马尔可夫长时间操作
Abstract
General-purpose robot policies should be modeled as dynamical systems, yet many VLA and generative imitation policies still rely on present observations or short windows. This Markovian shortcut fails in memory-dependent manipulation: identical observations can demand different actions after different histories. We present Chronos, a physics-informed full-history framework for non-Markovian long-horizon manipulation. The key idea is to elevate observation history from auxiliary context to the latent state of the policy dynamics. At each physical control step, Chronos forms one state-representative token by fusing observation and proprioception, so the token sequence is aligned one-to-one with physical time. A selective state space model propagates this causal historical state, which conditions a multimodal coarse action prior through implicit maximum likelihood estimation (IMLE). This prior is then refined by a second-order Schrodinger-inspired bridge that predicts acceleration fields, yielding smoother and more physically grounded robot motion. Across 16 simulated tasks and 4 real-world experiments, Chronos is evaluated on precision insertion, general manipulation, and memory-dependent long-horizon control. On RMBench, where success requires remembering task phase, Chronos achieves 73.6% average success, outperforming Markovian VLA baseline pi0.5 by +62.4 percentage points, a 6.6x relative gain, while using 10x fewer parameters. It also surpasses the memory VLA Mem-0 by 22.8 points while using over 30x fewer parameters. In real-world dual-arm experiments using a single RGB camera, Chronos achieves 78% average success over four tasks, including 72% on the three memory-dependent tasks, whereas pi0.5 achieves 7% overall and 0% on the memory-dependent subset. These results suggest that history should not be treated as auxiliary context, but as the latent state of the manipulation policy.
Chinese Translation
通用机器人策略应被建模为动态系统，然而许多变分自适应（VLA）和生成模仿策略仍依赖于当前观察或短时间窗口。这种马尔可夫捷径在依赖记忆的操作中失效：相同的观察可能在不同的历史下要求不同的动作。我们提出了Chronos，一种物理信息驱动的全历史框架，用于非马尔可夫长时间操作。其关键思想是将观察历史从辅助上下文提升为策略动态的潜在状态。在每个物理控制步骤中，Chronos通过融合观察和本体感觉形成一个状态代表性标记，从而使标记序列与物理时间一一对应。一个选择性状态空间模型传播这一因果历史状态，通过隐式最大似然估计（IMLE）对多模态粗略动作先验进行条件化。然后，通过一个二阶薛定谔启发的桥梁对该先验进行精炼，以预测加速度场，从而实现更平滑且更符合物理规律的机器人运动。在16个模拟任务和4个真实世界实验中，Chronos在精确插入、一般操作和依赖记忆的长时间控制方面进行了评估。在RMBench上，成功需要记住任务阶段，Chronos实现了73.6%的平均成功率，超越了马尔可夫VLA基线pi0.5，提升了62.4个百分点，获得了6.6倍的相对增益，同时使用的参数减少了10倍。它还在使用超过30倍更少参数的情况下，超越了记忆VLA Mem-0 22.8个百分点。在使用单个RGB摄像头进行的真实双臂实验中，Chronos在四个任务上实现了78%的平均成功率，其中在三个依赖记忆的任务上达到了72%，而pi0.5的整体成功率为7%，在依赖记忆的子集上为0%。这些结果表明，历史不应被视为辅助上下文，而应作为操作策略的潜在状态。
cs.RO / 82 / 2606.30362
ReactiveBFM: Reactive Closed-Loop Motion Planning Towards Universal Humanoid Whole-Body Control
ReactiveBFM：面向通用类人机器人全身控制的反应式闭环运动规划
Abstract
While current Behavior Foundation Models (BFMs) provide robust control priors for humanoids, they only execute pre-defined reference motions. As a result, they are vulnerable to environmental shifts and incapable of reactive whole-body coordination. Naively cascading them with generative motion planners fails to achieve true reactivity, as inevitable tracking discrepancies induce fatal cumulative exposure bias. To bridge this gap, we propose ReactiveBFM, a real-time closed-loop planning-control framework. At its core, we effectively mitigate exposure bias via a scheduled prefix sampling curriculum, forcing the generative planner to actively learn error-recovery behaviors from imperfect physical states rather than ground-truth trajectories. Systematically, to reconcile the severe latency mismatch between auto-regressive planning and high-frequency tracking, we introduce an asynchronous replanning mechanism. Combined with trajectory chunking to temporally ensemble spatial references, our system guarantees spatio-temporally fluid execution without physical jitter. Deployed on the Unitree G1 humanoid, ReactiveBFM demonstrates unprecedented physical agility across a vast repertoire of text-conditioned closed-loop motions. Notably, ReactiveBFM achieves zero-shot moving target reaching, showcasing intricate whole-body coordination and on-the-fly replanning. In sim-to-sim benchmarking under severe perturbations, ReactiveBFM achieves a 93.1% success rate, significantly outperforming cascaded open-loop baselines by 28.6%.
Chinese Translation
尽管当前的行为基础模型（BFMs）为类人机器人提供了稳健的控制先验，但它们仅执行预定义的参考动作。因此，它们对环境变化非常脆弱，无法实现反应式的全身协调。简单地将它们与生成运动规划器级联，无法实现真正的反应性，因为不可避免的跟踪差异会导致致命的累积暴露偏差。为了解决这一问题，我们提出了ReactiveBFM，一个实时闭环规划控制框架。其核心是通过调度前缀采样课程有效减轻暴露偏差，迫使生成规划器主动从不完美的物理状态中学习错误恢复行为，而不是从真实轨迹中学习。系统性地，为了调和自回归规划与高频跟踪之间严重的延迟不匹配，我们引入了一种异步重规划机制。结合轨迹分块以时间上集成空间参考，我们的系统确保了无物理抖动的时空流畅执行。在Unitree G1类人机器人上部署的ReactiveBFM展示了前所未有的物理灵活性，涵盖了广泛的文本条件闭环动作。值得注意的是，ReactiveBFM实现了零样本移动目标到达，展示了复杂的全身协调和即时重规划。在严重扰动下的仿真对仿真基准测试中，ReactiveBFM实现了93.1%的成功率，显著优于级联的开放循环基线，提升幅度达到28.6%。
cs.RO / 83 / 2606.30367
FutureNav: Unified World-Action Modeling for Vision-and-Language Navigation
FutureNav：用于视觉与语言导航的统一世界-动作建模
Abstract
Vision-and-language navigation (VLN) in continuous environments requires an agent to ground instructions in egocentric observations while maintaining spatial understanding across long action sequences. Recent navigation foundation models have shown strong progress by scaling vision-language models, but they often learn navigation primarily as direct action generation, without explicitly modeling world states or predicting their future evolution. We introduce FutureNav, a VLM-based unified world-action modeling framework for vision-and-language navigation. Specifically, FutureNav jointly encodes text, visual, and spatial features and feeds them into the LLM, and optimizes four objectives for simultaneous world and action modeling: an action policy objective for navigation action prediction, inverse and forward dynamics objectives for modeling state transitions, and a future generation objective for predicting future spatial states. This unified architecture strengthens action prediction while explicitly modeling the world, without sacrificing inference speed. Extensive experiments show that, with only a 4B-scale backbone, FutureNav achieves state-of-the-art performance on multiple VLN benchmarks and substantially outperforms prior VLN methods, paving the way toward future world-action models for VLN. We will release the code and models to support future research.
Chinese Translation
在连续环境中的视觉与语言导航（VLN）要求代理能够将指令与自我中心的观察相结合，同时在长时间动作序列中保持空间理解。近期的导航基础模型通过扩展视觉-语言模型取得了显著进展，但它们通常将导航主要视为直接的动作生成，而没有明确建模世界状态或预测其未来演变。我们提出了FutureNav，一个基于视觉-语言模型（VLM）的统一世界-动作建模框架，用于视觉与语言导航。具体而言，FutureNav联合编码文本、视觉和空间特征，并将其输入到大型语言模型（LLM）中，同时优化四个目标以实现世界和动作的同步建模：用于导航动作预测的动作策略目标、用于建模状态转移的逆向和前向动力学目标，以及用于预测未来空间状态的未来生成目标。这种统一架构在明确建模世界的同时增强了动作预测能力，而不牺牲推理速度。大量实验表明，FutureNav在仅使用4B规模的基础模型的情况下，在多个VLN基准上达到了最先进的性能，并显著超越了之前的VLN方法，为未来的VLN世界-动作模型铺平了道路。我们将发布代码和模型以支持未来的研究。
cs.RO / 84 / 2606.30404
HUMEMBR: Learning Human Routines for Predictive Embodied Navigation
HUMEMBR：学习人类日常行为以实现预测性具身导航
Abstract
Understanding and navigating human-centered environments over extended periods of time while considering human behavior and routines remains a fundamental challenge in robotics. In real-world settings, robots may be asked to locate a specific individual, predict where that person is likely to be, or estimate when they typically leave a building. Addressing such queries requires reasoning over extensive histories of observations and capturing long-term behavioral patterns. To this end, we introduce Human-Centered Memory for Embodied Robots (HUMEMBR), a system designed for embodied question answering and routine-conditioned navigation. HUMEMBR integrates a continuous memory construction process with a parallel retrieval and querying mechanism, enabling the system to accumulate structured representations of human routines while supporting interactive, user-driven queries. Our experimental results indicate that HUMEMBR improves long-horizon reasoning about human behavior relative to full-context LLM baselines, while using substantially fewer tokens. Furthermore, we deploy HUMEMBR on a physical robot in two distinct environments, showing its ability to handle diverse queries and navigation tasks under real-world conditions.
Chinese Translation
理解和导航以人为中心的环境，并在考虑人类行为和日常活动的情况下，长时间保持这一能力，仍然是机器人技术中的一项基本挑战。在现实世界中，机器人可能被要求定位特定个体，预测该个体可能出现的位置，或估计他们通常何时离开建筑物。解决这些问题需要对大量观察历史进行推理，并捕捉长期行为模式。为此，我们提出了面向具身机器人的人类中心记忆系统（HUMEMBR），该系统旨在实现具身问答和基于日常行为的导航。HUMEMBR将连续的记忆构建过程与并行检索和查询机制相结合，使系统能够积累人类日常行为的结构化表示，同时支持互动的用户驱动查询。我们的实验结果表明，HUMEMBR在关于人类行为的长远推理方面相较于全上下文的LLM基线有显著提升，同时使用的标记数量大幅减少。此外，我们在两个不同环境中将HUMEMBR部署在物理机器人上，展示了其在现实条件下处理多样化查询和导航任务的能力。
cs.RO / 85 / 2606.30456
Vision-Language-Action Models: Experimental Insights from a Real-World UR5 Platform
视觉-语言-动作模型：来自真实世界UR5平台的实验见解
Abstract
This project investigates whether recent Vision-Language-Action (VLA) models can be transferred from controlled research benchmarks to a real-world robotic platform, specifically a UR5e manipulator, in a reproducible and operationally meaningful manner. The work integrates real-robot data acquisition, dataset engineering (compatible with the RLDS format), and the fine-tuning and deployment of OpenVLA and OpenVLA-OFT models, with systematic validation of action representations and control interfaces. The project resulted in several foundational assets: (i) a complete real-robot data acquisition pipeline, (ii) a dataset conversion workflow aligned with RLDS standards, (iii) an initial fine-tuning and inference infrastructure for VLA models, and (iv) a structured set of experimental observations grounded in real-robot trials. These elements collectively establish a reproducible framework for evaluating learning-based manipulation systems beyond simulation. Empirically, the experiments reveal a consistent gap between promising offline indicators and unstable closed-loop behavior on the physical system: this gap cannot be attributed solely to model limitations, it is strongly influenced by action semantics, coordinate frame conventions, temporal alignment between modalities, image preprocessing consistency, and dataset coverage and quality. These observations lead to a key interpretation: the successful deployment of VLA systems in real-world settings depends less on incremental improvements in model capacity and more on precise control of the entire data-model-control pipeline. The project reframes VLA-based robotics from a primarily model-centric challenge to a system-level problem; it highlights the difficulty of running robust task execution on the real robot and provides a clear, experimentally grounded understanding of the conditions required for reliable deployment.
Chinese Translation
本项目探讨了近期的视觉-语言-动作（VLA）模型是否能够以可重复和具有操作意义的方式，从受控的研究基准转移到真实世界的机器人平台，特别是UR5e操纵器。该工作整合了真实机器人数据采集、数据集工程（与RLDS格式兼容）以及OpenVLA和OpenVLA-OFT模型的微调和部署，同时系统地验证了动作表示和控制接口。该项目产生了几个基础资产：（i）完整的真实机器人数据采集管道，（ii）与RLDS标准对齐的数据集转换工作流，（iii）VLA模型的初步微调和推理基础设施，以及（iv）基于真实机器人试验的结构化实验观察集。这些元素共同建立了一个可重复的框架，用于评估超越仿真的基于学习的操控系统。从实证上看，实验揭示了有希望的离线指标与物理系统上不稳定的闭环行为之间存在一致的差距：这一差距不能仅归因于模型的局限性，而是受到动作语义、坐标框架约定、模态之间的时间对齐、图像预处理一致性以及数据集覆盖和质量的强烈影响。这些观察导致了一个关键的解释：VLA系统在真实世界环境中的成功部署与模型能力的渐进改进关系不大，而更多依赖于对整个数据-模型-控制管道的精确控制。该项目将基于VLA的机器人技术从一个主要以模型为中心的挑战重新框定为一个系统级问题；它突显了在真实机器人上进行稳健任务执行的困难，并提供了对可靠部署所需条件的清晰、实验基础的理解。
cs.RO / 86 / 2606.30457
Behavior Prompting Policy: Demonstrations as Prompts for Manipulation
行为提示策略：作为操作提示的示范
Abstract
We study behavior prompting, a paradigm that enables robots to perform new tasks at inference time given a single human demonstration, which we call a behavior prompt. To enable this capability, we present contributions in algorithm, data, and evaluation. For algorithm, we introduce Behavior Prompting Policy (BPP), an in-context visuomotor architecture that translates the behavior prompt and the current observation into robot actions. For data, we identify that task diversity is the primary driver of the prompting capability and introduce iPhUMI, a handheld manipulation interface for collecting diverse training data. For evaluation, we introduce DrawAnything and LIBERO-Gen to evaluate test-time adaptation to unseen drawing and tabletop manipulation tasks. We also demonstrate that iPhUMI serves as a practical interface for specifying behavior prompts at test time, enabling a human to command a robot via a single demonstration to complete known tasks or to define new robot capabilities. Altogether, behavior prompting provides a flexible and scalable way to teach robots new skills without the need for expensive fine-tuning. Our project website is located at https://behavior-prompting.github.io/ .
Chinese Translation
我们研究了行为提示，这是一种使机器人能够在推理时根据单个人工示范执行新任务的范式，我们称之为行为提示。为了实现这一能力，我们在算法、数据和评估方面做出了贡献。在算法方面，我们提出了行为提示策略（Behavior Prompting Policy, BPP），这是一种上下文视觉运动架构，它将行为提示和当前观察转换为机器人动作。在数据方面，我们识别出任务多样性是提示能力的主要驱动因素，并引入了iPhUMI，这是一种用于收集多样化训练数据的手持操作接口。在评估方面，我们引入了DrawAnything和LIBERO-Gen，以评估在未见过的绘图和桌面操作任务上的测试时适应性。我们还展示了iPhUMI作为在测试时指定行为提示的实用接口，使人类能够通过单个示范命令机器人完成已知任务或定义新的机器人能力。总体而言，行为提示提供了一种灵活且可扩展的方式来教会机器人新技能，而无需昂贵的微调。我们的项目网站位于 https://behavior-prompting.github.io/ 。
cs.RO / 87 / 2606.30474
Grasp-Oriented Non-Prehensile Manipulation via Learning a Graspability Field
面向抓取的非抓取操作通过学习抓取性场
Abstract
Non-prehensile manipulation is often used as a preparatory step for robotic grasping, yet existing approaches typically require a predefined target object pose. In practice, however, objects admit multiple graspable configurations and the desired pose is not known in advance. We reformulate non-prehensile manipulation for grasping as optimizing an object centric graspability objective rather than reaching a specific pose. We construct a graspable set from synthesized grasps and define a graspability field that measures how suitable an object configuration is for successful grasp execution. The scalar measure provides a dense learning signal for reinforcement learning and determines when to terminate manipulation. This yields a closed-loop manipulation-to-grasp pipeline driven by a single policy. Experiments in simulation and on a real robot show that the policy reliably reconfigures objects into graspable states and transitions to grasping without external planners or manually specified stopping conditions. The predicted graspability distance correlates with real world grasp success, which indicates that the learned representation captures grasp feasibility of object configurations.
Chinese Translation
非抓取操作通常作为机器人抓取的准备步骤，但现有方法通常需要预定义的目标物体姿态。然而，在实际操作中，物体可以具有多种可抓取的配置，并且所需的姿态并不事先已知。我们将非抓取操作重新表述为优化以物体为中心的抓取性目标，而不是达到特定的姿态。我们从合成的抓取中构建一个可抓取集合，并定义一个抓取性场，衡量物体配置在成功执行抓取方面的适宜性。这个标量度量为强化学习提供了密集的学习信号，并决定何时终止操作。这产生了一个由单一策略驱动的闭环操作到抓取的流程。在仿真和真实机器人上的实验表明，该策略可靠地将物体重新配置为可抓取状态，并在没有外部规划器或手动指定停止条件的情况下过渡到抓取。预测的抓取性距离与现实世界中的抓取成功率相关，这表明学习到的表示捕捉了物体配置的抓取可行性。
cs.RO / 88 / 2606.30537
Learning from Mistakes: Rollout-Retrieval Lifelong Policy Learning for Autonomous Driving
从错误中学习：用于自主驾驶的回滚-检索终身策略学习
Abstract
Autonomous driving policies should be able to improve continually as deployment exposes them to increasingly diverse and long-tail traffic situations. However, most learning-based policies are trained or fine-tuned on expert demonstrations and then rely largely on generalization to handle challenging closed-loop scenarios, lacking an explicit mechanism to correct and retain the mistakes exposed in these scenarios. This paper studies autonomous driving policy improvement from a lifelong learning perspective: Can a pretrained policy improve continually by accumulating corrective knowledge derived from its own mistakes, while retaining previously acquired driving competence? To answer this question, we propose Rollout-Retrieval Lifelong Policy Learning (R$^2$LPL), a policy learning framework that retrieves corrective targets from recoverable policy-induced mistakes and retains the resulting knowledge through lifelong policy learning. R^2LPL addresses a key bottleneck in continual policy improvement: closed-loop mistakes reveal where the policy is weak, but do not directly specify what the policy should learn. By filtering recoverable mistake-related states and retrieving feasible corrective targets, R$^2$LPL turns sparse failure evidence into compact supervised knowledge for stable and sample-efficient policy improvement. We evaluate R$^2$LPL on large-scale closed-loop nuPlan benchmarks. With only a few rollout and continual-learning cycles, R$^2$LPL elevates a learning-based planner with moderate initial performance to state-of-the-art performance across the evaluated benchmarks, especially on the challenging and long-tail Test14-hard split. These results demonstrate the effectiveness of R$^2$LPL in converting recoverable closed-loop mistakes into corrective knowledge for sustained policy improvement.
Chinese Translation
自主驾驶策略应能够随着部署而不断改进，因为部署使其接触到越来越多样化和长尾的交通情况。然而，大多数基于学习的策略是在专家演示上进行训练或微调的，然后在处理具有挑战性的闭环场景时主要依赖于泛化，缺乏明确的机制来纠正和保留在这些场景中暴露出的错误。本文从终身学习的角度研究自主驾驶策略的改进：一个预训练的策略能否通过积累源自自身错误的纠正知识而不断改进，同时保留先前获得的驾驶能力？为了解答这个问题，我们提出了回滚-检索终身策略学习（R$^2$LPL），这是一个从可恢复的策略引发的错误中检索纠正目标并通过终身策略学习保留所得到知识的策略学习框架。R$^2$LPL解决了持续策略改进中的一个关键瓶颈：闭环错误揭示了策略的薄弱之处，但并未直接指定策略应学习什么。通过过滤可恢复的错误相关状态并检索可行的纠正目标，R$^2$LPL将稀疏的失败证据转化为紧凑的监督知识，以实现稳定且样本高效的策略改进。我们在大规模闭环nuPlan基准上评估了R$^2$LPL。仅通过少量的回滚和持续学习周期，R$^2$LPL将一个初始性能适中的基于学习的规划器提升至评估基准中的最先进性能，尤其是在具有挑战性和长尾的Test14-hard分割上。这些结果证明了R$^2$LPL在将可恢复的闭环错误转化为持续策略改进的纠正知识方面的有效性。
cs.RO / 89 / 2606.30552
Training Vision-Language-Action Models with Dense Embodied Chain-of-Thought Supervision
使用密集的具身思维链监督训练视觉-语言-动作模型
Abstract
Cross-embodiment transfer in vision-language-action (VLA) models remains challenging because low-level state and action spaces differ fundamentally across robot platforms. We observe that the high-level cognitive process underlying manipulation, including scene perception, object identification, task planning, and sub-task decomposition, is largely shared across embodiments. Based on this observation, we present ZR-0, a 2.6 billion parameter end-to-end VLA model that uses dense Embodied Chain-of-Thought (ECoT) supervision to align cross-embodiment representations within the vision-language model (VLM). ZR-0 adopts a dual-stream architecture: a pre-trained VLM (System 2) generates structured ECoT reasoning during training, while a Diffusion Transformer-based action expert (System 1) produces continuous action chunks via flow matching. The two components are coupled through cross-attention, with an attention mask that restricts the action expert to input prompt features only, enabling ECoT generation to be entirely skipped at inference without any performance loss. ZR-0 is pre-trained on ProcCorpus-60M, a large-scale dataset comprising approximately 60 million frames (approximately 1,000 hours) from over 400K trajectories, with dense ECoT annotations covering 96.8% of all frames. We evaluate ZR-0 on three simulation benchmarks spanning single-arm (LIBERO), bimanual (RoboTwin 2.0), and humanoid (RoboCasa GR-1 Tabletop) embodiments, as well as real-world experiments on the xArm platform, demonstrating strong performance across all settings. Code and model checkpoints are available at https://github.com/RUCKBReasoning/ZR-0.
Chinese Translation
视觉-语言-动作（VLA）模型中的跨具身转移仍然面临挑战，因为不同机器人平台的低级状态和动作空间在根本上存在差异。我们观察到，操控背后的高级认知过程，包括场景感知、物体识别、任务规划和子任务分解，在不同具身之间在很大程度上是共享的。基于这一观察，我们提出了ZR-0，一个具有26亿参数的端到端VLA模型，它利用密集的具身思维链（ECoT）监督来对齐视觉-语言模型（VLM）中的跨具身表示。ZR-0采用双流架构：一个预训练的VLM（系统2）在训练过程中生成结构化的ECoT推理，而一个基于扩散变换器的动作专家（系统1）通过流匹配生成连续的动作片段。这两个组件通过交叉注意力相耦合，注意力掩码限制动作专家仅输入提示特征，从而使得在推理时可以完全跳过ECoT生成而不损失性能。ZR-0在ProcCorpus-60M上进行预训练，这是一个大规模数据集，包含来自超过40万条轨迹的约6000万帧（约1000小时），并且密集的ECoT注释覆盖了96.8%的所有帧。我们在三个模拟基准上评估ZR-0，涵盖单臂（LIBERO）、双手（RoboTwin 2.0）和类人（RoboCasa GR-1桌面）具身，以及在xArm平台上的真实世界实验，展示了在所有设置下的强大性能。代码和模型检查点可在https://github.com/RUCKBReasoning/ZR-0获取。
cs.RO / 90 / 2606.30575
MOAR Planner: Multi-Objective and Adaptive Risk-Aware Path Planning for Infrastructure Inspection with a UAV
MOAR 规划器：基于多目标和自适应风险意识的无人机基础设施检查路径规划
Abstract
The problem of autonomous navigation for UAV inspection remains challenging as it requires effectively navigating in close proximity to obstacles, while accounting for dynamic risk factors such as weather conditions, communication reliability, and battery autonomy. This paper introduces the MOAR path planner which addresses the complexities of evolving risks during missions. It offers real-time trajectory adaptation while concurrently optimizing safety, time, and energy. The planner employs a risk-aware cost function that integrates pre-computed cost maps, the new concepts of damage and insertion costs, and an adaptive speed planning framework. With that, the optimal path is searched in a graph using a discrete representation of the state and action spaces. The method is evaluated through simulations and real-world flight tests. The results show the capability to generate real-time trajectories spanning a broad range of evaluation metrics: around 90% of the range occupied by popular algorithms. The proposed framework contributes by enabling UAVs to navigate more autonomously and reliably in critical missions.
Chinese Translation
无人机检查的自主导航问题仍然具有挑战性，因为它需要在靠近障碍物的情况下有效导航，同时考虑动态风险因素，如天气条件、通信可靠性和电池续航能力。本文介绍了 MOAR 路径规划器，该规划器解决了任务中不断变化的风险复杂性。它提供实时轨迹适应，同时优化安全性、时间和能量。该规划器采用风险意识成本函数，整合了预先计算的成本地图、损害和插入成本的新概念，以及自适应速度规划框架。通过在图中使用状态和动作空间的离散表示，搜索最优路径。该方法通过仿真和实际飞行测试进行评估。结果表明，该方法能够生成涵盖广泛评估指标的实时轨迹：约占流行算法的 90% 范围。所提出的框架通过使无人机在关键任务中更自主和可靠地导航做出了贡献。
cs.RO / 91 / 2606.30581
Realtime Wind Estimation using Low Cost Quadrotor Uncrewed Aerial Vehicles
基于低成本四旋翼无人机的实时风速估计
Abstract
In environmental monitoring as well as emergency response applications such as wildfires, wind velocity measurement is essential. Quadrotor UAVs have become popular platforms for wind velocity estimation due to their maneuverability, compact size, and cost-effectiveness. Numerous studies use the Extended Kalman Filter (EKF) to estimate the wind velocity based on the quadrotor dynamic model. However, most of them use hovering quadrotors only for wind estimation, others use a near-linear trajectory to estimate near-constant velocities. Furthermore, EKF performance is constrained by its reliance on linearized approximations of the nonlinear quadrotor dynamics around current states, limiting accuracy in highly nonlinear scenarios, including windy conditions. This study proposes the use of an Unscented Kalman Filter (UKF), a nonlinear estimator to provide accurate wind estimations while maintaining the trajectory of the quadrotor UAV. The quadrotor is modeled on the Special Euclidean group SE(3) and the approach is evaluated through numerical simulations using a geometric controller to maintain quadrotor flight paths. The results indicate that as the nonlinearity of the simulation increases, the UKF consistently outperforms the EKF. This demonstrates the potential of the UKF as a reliable estimator for highly nonlinear scenarios, capable of maintaining the trajectory with minimal deviation while providing accurate wind velocity estimations.
Chinese Translation
在环境监测以及野火等紧急响应应用中，风速测量至关重要。由于其机动性、紧凑的体积和成本效益，四旋翼无人机（Quadrotor UAV）已成为风速估计的热门平台。许多研究利用扩展卡尔曼滤波器（Extended Kalman Filter, EKF）基于四旋翼动态模型来估计风速。然而，它们大多数仅使用悬停的四旋翼进行风速估计，其他研究则使用近线性轨迹来估计近乎恒定的风速。此外，EKF的性能受到其依赖于当前状态下非线性四旋翼动态的线性化近似的限制，这在高度非线性的场景（包括多风条件）中限制了其准确性。本研究提出使用无迹卡尔曼滤波器（Unscented Kalman Filter, UKF），一种非线性估计器，以在保持四旋翼无人机轨迹的同时提供准确的风速估计。四旋翼在特殊欧几里得群SE(3)上建模，并通过数值仿真评估该方法，使用几何控制器保持四旋翼飞行路径。结果表明，随着仿真的非线性增加，UKF始终优于EKF。这表明UKF在高度非线性场景中作为可靠估计器的潜力，能够在提供准确风速估计的同时，保持轨迹偏差最小。
cs.RO / 92 / 2606.30613
Sequential Planning via Anchored Robotic Keypoints
通过锚定机器人关键点进行顺序规划
Abstract
We present Sequential Planning via Anchored Robotic Keypoints, SPARK, a training-free neurosymbolic manipulation system that reaches 43.7% on six LIBERO-PRO position \& task cells, more than doubling CaP-Agent0 and Vision-Language-Action (VLA) baselines. CaP-Agent0, a multi-turn code-generation agent, achieves 18.2% by re-querying an LLM at every turn, but its restart-from-scratch solution proves costly against minor policy failures. Perception is the layer that fails most under position and task changes so SPARK spends its computation there. A single Gemini call composes the plan as a typed behavior tree (BT) of composable primitives, each already containing the low-level control (motion, grasping, depth geometry) a code-generation agent would otherwise regenerate on every trial. The rest of the budget goes to perception: a second Gemini call proposes three alternative text prompts per object, SAM3 evaluates each, and we keep the prompt$\to$label pair with the most confident detection and a recovery loop then retries a failed primitive against freshly detected objects, with no new LLM call. The alternative prompts add +27.7 points on the spatial suite and +10.0 on the object suite, with the recovery loop adding +5.0 overall. SPARK runs the same primitives on three robot families (UR10e, Franka FR3, bimanual Franka) across nine unique tasks at twenty trials each, averaging 68%. Since the detector, planner, and controller modules sit behind the typed plan, they swap independently without training, and each primitive's checkable post-condition traces a failure to the corresponding module or a kinematic limit. Every trial logs a verified, labeled trajectory, so a training-free planner that already beats VLAs can supply the data those policies need without teleoperation. Project page: https://cwru-aism.github.io/spark-page/
Chinese Translation
我们提出了通过锚定机器人关键点进行顺序规划（Sequential Planning via Anchored Robotic Keypoints，SPARK），这是一种无需训练的神经符号操作系统，在六个LIBERO-PRO位置与任务单元上达到了43.7%的成功率，超过了CaP-Agent0和视觉-语言-动作（Vision-Language-Action，VLA）基线的两倍以上。CaP-Agent0是一个多轮代码生成代理，通过在每轮重新查询大型语言模型（LLM）实现了18.2%的成功率，但其从头开始重启的解决方案在面对小的策略失败时成本高昂。感知是最容易在位置和任务变化下失败的层，因此SPARK将其计算资源集中在此。一次Gemini调用将计划构建为一个类型化的行为树（Behavior Tree，BT），其中的可组合原语已经包含了代码生成代理在每次试验中需要重新生成的低级控制（运动、抓取、深度几何）。其余的预算用于感知：第二次Gemini调用为每个对象提出三个替代文本提示，SAM3对每个进行评估，我们保留检测最为自信的提示$ o$标签对，并通过恢复循环对新检测到的对象重新尝试失败的原语，而无需新的LLM调用。替代提示在空间套件上增加了27.7分，在对象套件上增加了10.0分，恢复循环整体增加了5.0分。SPARK在三种机器人系列（UR10e、Franka FR3、双手Franka）上运行相同的原语，涵盖九个独特任务，每个任务进行二十次试验，平均成功率为68%。由于检测器、规划器和控制器模块位于类型化计划之后，它们可以独立交换而无需训练，每个原语的可检查后置条件将失败追溯到相应的模块或运动学限制。每次试验都会记录经过验证的标记轨迹，因此一个已经超过VLA的无训练规划器可以提供这些策略所需的数据，而无需远程操作。项目页面：https://cwru-aism.github.io/spark-page/
cs.RO / 93 / 2606.30632
GROW$^2$: Grounding Which and Where for Robot Tool Use
GROW$^2$: 机器人工具使用中的对象选择与定位
Abstract
Can the robot use a plate to cut a cake if no knife is available? Tool use greatly expands robot capabilities, but to use tools creatively beyond their intended functions, the robot faces the challenge of $\textit{open-world affordance grounding}$: select an open-category object to act as a tool and localize its specific region of action. To this end, we introduce GROW$^2$ (GROunding Which and Where), which leverages object parts as a natural abstraction to split the grounding process hierarchically into semantic and geometric levels, thus bypassing the need for data-heavy, end-to-end training. Semantically, GROW$^2$ harnesses the commonsense reasoning of Vision-Language Models (VLMs) to parse a natural-language task instruction, select a suitable object as the tool, and identify task-relevant parts on the tool and the target object. Geometrically, vision foundation models then ground the selected parts into precise 3D regions from a single RGB-D image. Experiments on established benchmarks show that GROW$^2$ outperforms state-of-the-art baselines on affordance prediction benchmarks. Further, it achieves zero-shot generalization over open-category objects and outperforms baselines in both simulated and real-world robot tool use experiments.
Chinese Translation
如果没有刀，机器人能否使用盘子切蛋糕？工具使用极大地扩展了机器人的能力，但为了超越工具的预期功能创造性地使用工具，机器人面临着$ extit{开放世界赋能基础}$的挑战：选择一个开放类别的对象作为工具，并定位其具体的操作区域。为此，我们提出了GROW$^2$（GROunding Which and Where），该方法利用对象部件作为自然抽象，将赋能过程分层为语义和几何两个层面，从而避免了对数据密集型的端到端训练的需求。在语义层面，GROW$^2$利用视觉-语言模型（Vision-Language Models, VLMs）的常识推理来解析自然语言任务指令，选择合适的对象作为工具，并识别工具和目标对象上的任务相关部件。在几何层面，视觉基础模型将选定的部件从单个RGB-D图像中定位到精确的3D区域。基于已建立的基准测试的实验表明，GROW$^2$在赋能预测基准上超越了最先进的基线。此外，它在开放类别对象上实现了零-shot泛化，并在模拟和真实世界的机器人工具使用实验中均优于基线。
cs.RO / 94 / 2606.30645
VLK: Learning Humanoid Loco-Manipulation from Synthetic Interactions in Reconstructed Scenes
VLK：从重建场景中的合成交互学习类人运动操控
Abstract
Perception-based humanoid loco-manipulation requires connecting egocentric observations and task instructions to whole-body motion. Learning this mapping requires synchronized egocentric images, language commands, and robot-compatible kinematic trajectories, yet no existing data source provides this complete tuple at scale. We address this bottleneck by generating vision-language-kinematics (VLK) supervision synthetically in reconstructed scenes. Our pipeline leverages 3D Gaussian Splatting to reconstruct metric-scale indoor environments, synthesizes navigation and object-interaction trajectories using privileged scene information, and renders paired egocentric observations after the fact. We produce 48,000 paired trajectories with no human intervention and train a VLK policy that predicts short-horizon whole-body kinematic trajectories. A whole-body tracker converts these predictions into actions on the physical humanoid. We evaluate on the physical Unitree G1 performing navigation and single-object transport, demonstrating that synthesized interactions in reconstructed scenes provide effective supervision for sim-to-real perception-based humanoid loco-manipulation. Project Website: https://vision-language-kinematics.github.io/
Chinese Translation
基于感知的类人运动操控需要将自我中心的观察与任务指令连接到全身运动。学习这种映射需要同步的自我中心图像、语言指令和机器人兼容的运动轨迹，但现有的数据源并未在规模上提供这一完整的元组。我们通过在重建场景中合成生成视觉-语言-运动学（VLK）监督来解决这一瓶颈。我们的流程利用3D高斯点云重建度量尺度的室内环境，使用特权场景信息合成导航和物体交互轨迹，并在事后渲染配对的自我中心观察。我们生成了48,000条配对轨迹，无需人工干预，并训练了一个VLK策略，该策略预测短时间范围内的全身运动轨迹。一个全身跟踪器将这些预测转换为物理类人的动作。我们在物理Unitree G1上进行评估，执行导航和单一物体运输，证明在重建场景中合成的交互为基于感知的类人运动操控提供了有效的监督。项目网站：https://vision-language-kinematics.github.io/
cs.CV / 1 / 2606.28371
GeoISF: Instance Semantic Forest Inspired Large-Scale Cross-View Geo-Localization via Ground LiDAR-to-Satellite Image
GeoISF：基于地面激光雷达到卫星图像的实例语义森林启发的大规模跨视角地理定位
Abstract
The problem of localization on a large-scale satellite image given a frame of query ground view point clouds remains challenging. Existing LiDAR-to-image cross-view localization methods struggle in large-scale scenarios due to limited semantic alignment and the modality gap between point clouds and satellite images. This paper introduces the large-scale LiDAR-to-image geo-localization pipeline called GeoISF. GeoISF introduces an instance semantic forest constructed using WordNet, which enhances temporal semantic representation and discriminative power by integrating semantic trees from multiple frames. By leveraging environmental semantic representation as a shared medium, GeoISF effectively bridges the modality gap and improves semantic matching accuracy. Extensive experiments demonstrate the superior performance of GeoISF in large-scale cross-view localization, 13.22 times better than the parallel LiDAR-to-image method in the R@10 metric on the KITTI dataset. The proposed method addresses the existing gap in large-scale LiDAR-to-image cross-view localization, offering a robust solution to the computational and accuracy challenges inherent in such scenarios. We will release the code as an open-source resource available online for the broader research community.
Chinese Translation
在给定一帧查询地面视点云的情况下，基于大规模卫星图像的定位问题仍然具有挑战性。现有的激光雷达到图像的跨视角定位方法在大规模场景中面临语义对齐有限和点云与卫星图像之间的模态差距等问题。本文提出了一种名为GeoISF的大规模激光雷达到图像的地理定位管道。GeoISF引入了一个使用WordNet构建的实例语义森林，通过整合来自多个帧的语义树，增强了时间语义表示和区分能力。通过利用环境语义表示作为共享媒介，GeoISF有效地弥合了模态差距，提高了语义匹配的准确性。大量实验表明，GeoISF在大规模跨视角定位中的表现优越，在KITTI数据集的R@10指标上比并行的激光雷达到图像方法提高了13.22倍。所提出的方法解决了大规模激光雷达到图像跨视角定位中的现有差距，为此类场景固有的计算和准确性挑战提供了稳健的解决方案。我们将把代码作为开放源代码资源在线发布，供更广泛的研究社区使用。
cs.CV / 2 / 2606.28377
Memory-Augmented LSTM Autoencoder for Unsupervised Activity Recognition with IMU Sensor Fusion
基于记忆增强的LSTM自编码器的无监督活动识别与IMU传感器融合
Abstract
HAR using Inertial Measurement Unit (IMU) sensors is vital for healthcare monitoring and rehabilitation. Despite deep learning advancements, major challenges remain: reliance on labeled data, multi-sensor fusion complexity, and the limited ability of unsupervised methods to capture spatiotemporal dependencies. These issues are pronounced in real-world scenarios with noisy data, overlapping activities, and missing labels. We propose a fully unsupervised spatiotemporal feature fusion framework using a memory-augmented autoencoder. It enhances activity representations via short temporal windows of multi-sensor IMU data, enabling real-time applications. Our framework extracts hierarchical static features via a Stacked Autoencoder, fusing them within and across sensors. A sequence-to-sequence LSTM Autoencoder then temporally refines these features, incorporating historical motion patterns without labels. We analyze key hyperparameters to identify configurations that maximize feature separability under short-window constraints. Evaluated on DaLiAc and PAMAP2 using realistic inter-class window segmentation, our method achieves 96.6% and 98.4% accuracy, respectively, surpassing supervised baselines and unsupervised approaches. Our method improves feature separability by up to 9% despite shorter temporal windows. While our realistic inter-class segmentation reduces accuracy by ~7%, it was intentionally adopted to better reflect real-world activity transitions and practical relevance.
Chinese Translation
使用惯性测量单元（IMU）传感器进行人类活动识别（HAR）对于健康监测和康复至关重要。尽管深度学习取得了进展，但仍面临主要挑战：对标记数据的依赖、多传感器融合的复杂性，以及无监督方法捕捉时空依赖性的能力有限。这些问题在具有噪声数据、重叠活动和缺失标签的真实场景中尤为明显。我们提出了一种完全无监督的时空特征融合框架，采用记忆增强的自编码器。该框架通过多传感器IMU数据的短时间窗口增强活动表示，支持实时应用。我们的框架通过堆叠自编码器提取层次静态特征，并在传感器内部和之间进行融合。接着，序列到序列的LSTM自编码器在时间上细化这些特征，结合历史运动模式而无需标签。我们分析了关键超参数，以识别在短窗口约束下最大化特征可分离性的配置。在DaLiAc和PAMAP2数据集上进行评估，采用现实的类间窗口分割，我们的方法分别达到了96.6%和98.4%的准确率，超越了监督基线和无监督方法。尽管我们的短时间窗口导致特征可分离性提高了最多9%，但现实的类间分割将准确率降低了约7%，这一点是故意采用的，以更好地反映真实世界的活动过渡和实际相关性。
cs.CV / 3 / 2606.28383
Zero-Label Driving Scenario Complexity Detection via Joint Embedding Predictive Architecture
通过联合嵌入预测架构进行零标签驾驶场景复杂性检测
Abstract
Identifying complex and safety-critical driving scenarios in large unlabelled datasets is an important but expensive problem. Existing approaches rely on human annotators, supervised classifiers, or carefully engineered rule sets, all of which require substantial prior knowledge about what constitutes a difficult scenario. We ask whether a model can discover scenario complexity on its own, with no labels at any stage. We train a minimal Joint Embedding Predictive Architecture (JEPA) on structured agent state data from the nuPlan mini dataset and use the temporal prediction error as a zero-shot complexity score. Without access to any ground-truth labels during training or evaluation setup, the model assigns significantly higher scores to scenarios involving unprotected turns, crosswalk interactions, and pedestrian proximity, and significantly lower scores to lane-following and stationary-traffic scenarios. We validate this finding through four ablation experiments that isolate the source of the signal, and through a downstream anomaly detection evaluation that achieves Average Precision of 0.512 against a 0.436 chance baseline. The results show that temporal prediction error in a self-supervised latent world model is a practical proxy for driving scenario complexity.
Chinese Translation
在大规模未标记数据集中识别复杂且安全关键的驾驶场景是一个重要但代价高昂的问题。现有方法依赖于人工标注者、监督分类器或精心设计的规则集，这些方法都需要对什么构成困难场景有大量的先验知识。我们探讨模型是否能够在没有任何标签的情况下自主发现场景复杂性。我们在来自nuPlan迷你数据集的结构化代理状态数据上训练了一个最小的联合嵌入预测架构（Joint Embedding Predictive Architecture, JEPA），并使用时间预测误差作为零-shot复杂性评分。在训练或评估过程中没有访问任何真实标签的情况下，该模型对涉及无保护转弯、斑马线交互和行人接近的场景赋予显著更高的评分，而对车道跟随和静止交通场景则赋予显著更低的评分。我们通过四个消融实验验证了这一发现，这些实验隔离了信号的来源，并通过下游异常检测评估实现了0.512的平均精度，相较于0.436的随机基线。结果表明，自监督潜在世界模型中的时间预测误差是驾驶场景复杂性的一个实用代理。
cs.CV / 4 / 2606.28386
Data Provenance for Image Auto-Regressive Generation
图像自回归生成的数据来源
Abstract
Image autoregressive models (IARs) have recently demonstrated remarkable capabilities in visual content generation, achieving photorealistic quality and rapid synthesis through the next-token prediction paradigm adapted from large language models. As these models become widely accessible, robust data provenance is required to reliably trace IAR-generated images to the source model that synthesized them. This is critical to prevent the spread of misinformation, detect fraud, and attribute harmful content. We find that although IAR-generated images often appear visually identical to real images, their generation process introduces characteristic patterns in their outputs, which serves as a reliable provenance signal for the generated images. Leveraging this, we present a post-hoc framework that enables the robust detection of such patterns for provenance tracing. Notably, our framework does not require modifications of the generative process or outputs. Thereby, it is applicable in contexts where prior watermarking methods cannot be used, such as for generated content that is already published without additional marks and for models that do not integrate watermarking. We demonstrate the effectiveness of our approach across a wide range of IARs, highlighting its high potential for robust data provenance tracing in autoregressive image generation.
Chinese Translation
图像自回归模型（IARs）最近在视觉内容生成方面展现了显著的能力，通过从大型语言模型改编的下一个标记预测范式，达到了照片级真实感的质量和快速合成。随着这些模型的广泛可用，强大的数据来源追踪变得至关重要，以可靠地追溯IAR生成的图像到合成它们的源模型。这对于防止虚假信息的传播、检测欺诈行为以及归属有害内容至关重要。我们发现，尽管IAR生成的图像在视觉上往往与真实图像相同，但其生成过程在输出中引入了特征模式，这为生成图像提供了可靠的数据来源信号。基于此，我们提出了一个事后框架，能够稳健地检测这些模式以进行来源追踪。值得注意的是，我们的框架不需要对生成过程或输出进行修改。因此，它适用于那些先前水印方法无法使用的场景，例如已经发布且没有额外标记的生成内容，以及不集成水印的模型。我们在广泛的IAR模型中展示了我们方法的有效性，突显了其在自回归图像生成中进行稳健数据来源追踪的高潜力。
cs.CV / 5 / 2606.28389
SoccerNet 2026 Player-Centric Ball Action Spotting: Per-Player Attention with Agreement-Based Ensembling
SoccerNet 2026 以球员为中心的球类动作识别：基于一致性的每位球员注意力与集成
Abstract
We present our submission to the SoccerNet 2026 Player-Centric Ball Action Spotting challenge, which uses a two-stage pipeline: a Track-Aware Action Detector (TAAD) produces per-player action logits from broadcast video, and a Denoising Sequence Transduction (DST) transformer converts game-state features and TAAD logits into structured event sequences. We improve the TAAD with a temporal transformer that adds cross-frame context, alongside several training fixes. For the DST stage, we introduce a two-stage per-player attention mechanism operating on game-state features, and show that a spatial-first attention ordering (cross-player attention before temporal attention) improves validation Macro-F1 by 1.87%. To exploit architectural diversity, we train four model variants and combine them with a Weighted Event Fusion ensemble that applies agreement filtering to suppress single-model false positives while preserving recall, plus a dedicated exception for the rare tackle class. Our final system improves the challenge Macro-F1 from a baseline of 48.6 to 58.94.
Chinese Translation
我们提出了针对 SoccerNet 2026 以球员为中心的球类动作识别挑战的提交方案，该方案采用了一个两阶段的流程：一个轨迹感知动作检测器（Track-Aware Action Detector, TAAD）从广播视频中生成每位球员的动作 logits，而去噪序列转导（Denoising Sequence Transduction, DST）变换器则将比赛状态特征和 TAAD logits 转换为结构化事件序列。我们通过引入一个时间变换器来增强 TAAD，该变换器增加了跨帧上下文，同时进行了一些训练修正。在 DST 阶段，我们引入了一个两阶段的每位球员注意力机制，该机制在比赛状态特征上运行，并展示了空间优先的注意力顺序（跨球员注意力优先于时间注意力）使验证的 Macro-F1 提高了 1.87%。为了利用架构的多样性，我们训练了四个模型变体，并通过加权事件融合集成（Weighted Event Fusion ensemble）将它们结合在一起，该集成应用了一致性过滤以抑制单模型的假阳性，同时保持召回率，并为稀有的铲球类别提供了专门的例外。我们的最终系统将挑战的 Macro-F1 从基线的 48.6 提高到 58.94。
cs.CV / 6 / 2606.28390
Automated Quality Assessment of Geospatial Vector Data: A GeoAI Approach using Spatial Representation Learning
地理空间矢量数据的自动化质量评估：基于空间表征学习的GeoAI方法
Abstract
Geospatial vector data quality is a foundational research topic in GIS, yet classic rule-based quality assessment algorithms often struggle with diverse urban morphologies and massive data volumes. Recently, Geospatial Artificial Intelligence (GeoAI) shows promising potential for automating geospatial analysis, while its application to native vector data remains largely underexplored. To fill this research gap, we proposed Topo4Vec, an automated GeoAI framework, designed for scalable vector data quality assessment via advanced Spatial Representation Learning (SRL). Specifically, Topo4Vec relax the labor-intensive manual annotation process via topological error simulation, such as overlapping polygons and street network connectivity errors e.g., overshoots and undershoots. Then, it leverages state-of-the-art SRL approaches to encode complex, native vector geometries (e.g., polylines and polygons) into a latent space where topological errors are isolated from valid ones. A systematic performance evaluation across three study areas (Los Angeles, Munich, and Singapore) demonstrates the effectiveness and robustness of Topo4Vec, achieving a peak accuracy of 0.99 for detecting overlapping building footprints and 0.60 for overshoots and undershoots in street networks. Moreover, lessons learned from Topo4Vec shed a promising light into a scalable and autonomous GeoAI approach for large-scale vector data consistency and quality monitoring within the fast-growing geospatial data ecosystems. The code and data used in the paper are made openly available in https://figshare.com/s/612148eeb4bccadbd715.
Chinese Translation
地理空间矢量数据质量是地理信息系统（GIS）中的基础研究主题，但经典的基于规则的质量评估算法在多样化的城市形态和海量数据量面前常常显得力不从心。近年来，地理空间人工智能（GeoAI）在自动化地理空间分析方面展现出良好的潜力，但其在原生矢量数据上的应用仍然未被充分探索。为填补这一研究空白，我们提出了Topo4Vec，一个自动化的GeoAI框架，旨在通过先进的空间表征学习（SRL）实现可扩展的矢量数据质量评估。具体而言，Topo4Vec通过拓扑错误模拟（如重叠多边形和街道网络连通性错误，例如超出和不足）来放宽劳动密集型的手动标注过程。然后，它利用最先进的SRL方法将复杂的原生矢量几何（如折线和多边形）编码到一个潜在空间中，在该空间中，拓扑错误与有效数据被隔离开来。对三个研究区域（洛杉矶、慕尼黑和新加坡）进行的系统性能评估表明，Topo4Vec的有效性和鲁棒性，在检测重叠建筑轮廓时达到最高准确率0.99，在街道网络中的超出和不足检测中达到0.60。此外，从Topo4Vec中获得的经验教训为在快速增长的地理空间数据生态系统中实现大规模矢量数据一致性和质量监测的可扩展和自主的GeoAI方法提供了良好的启示。本文中使用的代码和数据已在https://figshare.com/s/612148eeb4bccadbd715上公开。
cs.CV / 7 / 2606.28391
Few-class Fidelity: Evaluating Explanations of Real-conditions CNN classifiers with Optimized Perturbations
少类保真度：使用优化扰动评估真实条件下CNN分类器的解释
Abstract
The wide use of Convolutional Neural Networks (CNN) in numerous domains and real-world classification applications is justified by their high precision and automation speed, helping users concentrate on higher-expertise tasks. To better understand the models and avoid bias during deployment, eXplainable Artificial Intelligence (XAI) techniques can be used after training. But as the list of XAI solutions expand, comparisons between them diverge, and consensus over their evaluation cannot be reached. This paper proposes a variation of Fidelity-based XAI metrics, with a focus on real-conditions applications, where the number of classes is often low. The approach generates in-distribution, uncertainty-provoking perturbations, to ensure proper measurement of the XAI methods faithfulness. As demonstration of the evaluation framework usefulness, it is compared with human-centric object localization and segmentation metrics. Once applied to both medical and natural imaging applications, it highlights the intricate correlation between domain, data curation, and XAI solution choices in order to validate training of a new CNN model.
Chinese Translation
卷积神经网络（CNN）在众多领域和现实世界分类应用中的广泛使用，得益于其高精度和自动化速度，帮助用户专注于更高专业性的任务。为了更好地理解模型并避免在部署过程中产生偏见，可以在训练后使用可解释人工智能（XAI）技术。然而，随着XAI解决方案的不断增加，它们之间的比较逐渐分歧，且在评估方面无法达成共识。本文提出了一种基于保真度的XAI指标变体，重点关注真实条件下的应用，其中类别数量通常较少。该方法生成分布内、引发不确定性的扰动，以确保对XAI方法的忠实性进行适当测量。作为评估框架有效性的示范，它与以人为中心的物体定位和分割指标进行了比较。应用于医学和自然影像应用后，强调了领域、数据整理和XAI解决方案选择之间复杂的相关性，以验证新CNN模型的训练。
cs.CV / 8 / 2606.28392
RADIANT-PET: Reasoning-Augmented PET/CT Lesion Segmentation with Large Language Models and Reinforcement Learning
RADIANT-PET：基于推理增强的大语言模型和强化学习的PET/CT病灶分割
Abstract
Accurate lesion segmentation in PET/CT is critical for oncology, yet remains challenging because physiologic tracer uptake and artifacts can mimic malignant signal. We present RADIANT-PET, a reasoning-augmented framework that couples a high-sensitivity voxel-level segmentation model with lesion-level large language model (LLM) adjudication. Candidate uptake regions are generated with a deliberately permissive segmentation stage, then converted into structured textual descriptions that summarize uptake intensity, morphology, and regional and global anatomical context. An LLM classifies each candidate as true lesion vs. false positive, optionally leveraging the radiology report as additional clinical context. To strengthen lesion-level reasoning, we further optimize a local LLM via reinforcement learning using Group Relative Policy Optimization, rewarding correct lesion classification and anatomically concordant site assignment. Across AutoPET and an OSU test cohort, RADIANT-PET consistently outperforms strong image-only baselines, with the largest improvements observed when radiology reports are provided. Overall, these results demonstrate that LLM-based lesion-level reasoning adds a novel reasoning layer beyond conventional segmentation, suppressing physiologic false positives and aligning voxel-level predictions with clinical interpretation. The project repository is available at: https://github.com/jwang-580/RADIANT-PET.
Chinese Translation
在PET/CT中，准确的病灶分割对肿瘤学至关重要，但由于生理示踪剂摄取和伪影可能模仿恶性信号，这一任务仍然具有挑战性。我们提出了RADIANT-PET，这是一个推理增强框架，将高灵敏度的体素级分割模型与病灶级的大语言模型（LLM）裁决相结合。候选摄取区域通过一个故意宽松的分割阶段生成，然后转换为结构化的文本描述，概述摄取强度、形态以及区域和全局解剖背景。LLM将每个候选区域分类为真实病灶或假阳性，选项上可以利用放射学报告作为额外的临床背景。为了增强病灶级推理，我们进一步通过强化学习优化本地LLM，采用群体相对策略优化（Group Relative Policy Optimization），奖励正确的病灶分类和解剖一致的部位分配。在AutoPET和OSU测试队列中，RADIANT-PET始终优于强大的仅图像基线，当提供放射学报告时，观察到最大的改进。总体而言，这些结果表明，基于LLM的病灶级推理在传统分割之外增加了一个新的推理层，抑制生理假阳性，并将体素级预测与临床解释对齐。项目代码库可在以下网址获取：https://github.com/jwang-580/RADIANT-PET。
cs.CV / 9 / 2606.28393
Transition-Aware best-of-N sampling for Longitudinal Chest X-ray Reports
考虑过渡的最佳N采样方法用于纵向胸部X光报告
Abstract
In longitudinal clinical practice, every chest X-ray is read in the context of the patients prior exam, and much of what the radiologist communicates is the change from one visit to the next. To the best of our knowledge, we present the first training-free best-of-N sampling scheme for pre-trained chest X-ray report generators that is explicitly aware of this longitudinal prior to current transition. We call it transition-aware best-of-N sampling, each report is split into sentences and embedded into an unordered set in Rd; each (prior, current) pair is reduced to a fixed-dim directional vector via a set-to-set distance designed to encode the change between the two sets; and candidates are scored by cosine distance from their candidate transition vector to a cached bank of ground-truth training transition vectors, aggregated as min or kNN. We instantiate the framework with four directional set distances (mean-shift, novelty residual, directed-Hausdorff anchor, and cost-weighted optimal transport) and evaluate on a multi-visit AP-PA cohort, running inference under three prompts on three vision-language generators. Transition-aware best-of-N outperforms random selection across the board, with the largest relative gains on the Impression section.
Chinese Translation
在纵向临床实践中，每次胸部X光检查都是在患者先前检查的背景下进行解读的，放射科医师所传达的许多信息都是从一次就诊到下次就诊的变化。据我们所知，我们提出了一种首个不需要训练的最佳N采样方案，专为预训练的胸部X光报告生成器设计，明确考虑了当前过渡的纵向先验。我们称之为考虑过渡的最佳N采样，每个报告被拆分为句子并嵌入到Rd中的无序集合中；每个（先前，当前）对通过一种旨在编码两个集合之间变化的集合到集合距离被简化为固定维度的方向向量；候选者的评分通过其候选过渡向量与缓存的真实训练过渡向量库之间的余弦距离来进行，聚合为最小值或k近邻。我们用四种方向集合距离（均值漂移、新颖性残差、定向Hausdorff锚点和成本加权最优运输）实例化该框架，并在一个多次就诊的AP-PA队列上进行评估，在三个视觉-语言生成器上运行三个提示的推理。考虑过渡的最佳N在各个方面都优于随机选择，在印象部分获得了最大的相对增益。
cs.CV / 10 / 2606.28394
GPU-Accelerated Inverse Structural Anastylosis from Block Collapse Dynamics
基于GPU加速的块崩塌动力学逆结构重组
Abstract
The physical anastylosis of collapsed architectural monuments -- the meticulous reassembly of fallen stone elements into their original structural configuration -- represents one of the most intellectually demanding challenges in conservation science. Traditional approaches depend heavily on expert archaeologist judgement and manual block-by-block correspondence, a process that is both labour-intensive and inherently subjective. Inspired by the combinatorial complexity of this problem as manifested in the game of Jenga, we present Jenga Inverse Predictor , a GPU-accelerated deep learning framework that addresses structural anastylosis as an inverse prediction task. Given an image of a collapsed block assembly, JIP-2 reconstructs the most probable prior tower configuration by: (1) implementing a complete rigid-body physics engine with OBB/SAT collision detection and a Projected Gauss-Seidel (PGS) contact solver accelerated with Numba JIT and CuPy CUDA; (2) applying the analytical force thresholds of Ziglar (CMU, 2006) -- F_app = 3*mu_s*m*g (Y-axis, torque-free) and F_app = 4*mu_s*m*g (X-axis, torque risk) -- over three friction levels (mu_s in {0.25, 0.40, 0.60}) across 450 simulated episodes; (3) training a dual-stream ResNet-18 that injects a friction one-hot vector and jointly predicts block removal count, per-position removal probabilities, centre-of-mass imbalance, and Ziglar torque risk; and (4) generating a smooth 3-D video of the block-by-block reverse reconstruction. We discuss implications for computer-assisted anastylosis at the UNESCO Maya site of Uxmal, Yucatan, and provide a detailed technical description of the full pipeline, architecture, and loss formulation.
Chinese Translation
倒塌建筑遗迹的物理重组——将倒下的石材元素精确地重新组装成其原始结构配置——是保护科学中最具智力挑战性的任务之一。传统方法在很大程度上依赖于专家考古学家的判断和逐块的手动对应，这一过程既劳动密集又固有主观性。受到这一问题在游戏Jenga中表现出的组合复杂性的启发，我们提出了Jenga Inverse Predictor（JIP），这是一个基于GPU加速的深度学习框架，将结构重组视为一个逆预测任务。给定一个倒塌块体组合的图像，JIP-2通过以下方式重建最可能的先前塔结构配置：(1) 实现一个完整的刚体物理引擎，具备OBB/SAT碰撞检测和使用Numba JIT及CuPy CUDA加速的投影高斯-赛德尔（PGS）接触求解器；(2) 在450个模拟实验中，针对三个摩擦水平（μ_s为{0.25, 0.40, 0.60}），应用Ziglar（CMU, 2006）的分析力阈值——F_app = 3*μ_s*m*g（Y轴，无扭矩）和F_app = 4*μ_s*m*g（X轴，扭矩风险）；(3) 训练一个双流ResNet-18网络，注入摩擦一热向量，并共同预测块移除数量、每个位置的移除概率、质心不平衡和Ziglar扭矩风险；(4) 生成块逐块逆重建的平滑3D视频。我们讨论了计算机辅助重组在尤卡坦州乌克斯马尔（Uxmal）联合国教科文组织玛雅遗址的应用，并提供了完整流程、架构和损失公式的详细技术描述。
cs.CV / 11 / 2606.28395
JASPR: Joint Spatial Representation learning of histology and spatial genomics for improved virtual genomic screening and clinical prognostication
JASPR：组织学与空间基因组学的联合空间表征学习，以改善虚拟基因筛选和临床预后评估
Abstract
Recent studies have shown that spatial properties of tumors are critical for understanding disease biology and predicting patient outcomes. These spatial properties are increasingly uncovered through complementary modalities: spatial transcriptomics (ST) captures spatially-resolved molecular states, while hematoxylin and eosin-stained whole slide images (HE) reveal tissue morphology. While approaches are emerging to fuse these modalities, effective methods that learn not only joint representations but also incorporate spatial context across modalities are lacking. Here, we present JASPR (Joint Spatial Representation learning), a self-supervised deep learning framework that integrates HE images and ST data through a cross-modal reconstruction objective that incorporates spatial context within HE images and ST profiles. It employs shared modules to capture universal spatial properties across modalities, while modality-specific experts encode features unique to morphological and genomic data. We train and validate JASPR on breast cancer datasets, demonstrating that its learned joint representation substantially improves HE-based prediction of 9,248 genes and provides prognostic value for breast cancer outcomes.
Chinese Translation
近期研究表明，肿瘤的空间特性对于理解疾病生物学和预测患者预后至关重要。这些空间特性越来越多地通过互补的方式被揭示：空间转录组学（Spatial Transcriptomics, ST）捕捉空间分辨的分子状态，而苏木精-伊红染色的全切片图像（Hematoxylin and Eosin, HE）则揭示了组织形态。尽管已经出现了一些融合这些模式的方法，但有效的学习联合表征并在不同模式间融入空间上下文的方法仍然缺乏。在此，我们提出了JASPR（联合空间表征学习），这是一种自监督深度学习框架，通过跨模态重建目标整合HE图像和ST数据，并在HE图像和ST特征中融入空间上下文。它采用共享模块捕捉跨模式的普遍空间特性，同时模态特定专家编码与形态和基因组数据独特相关的特征。我们在乳腺癌数据集上训练和验证了JASPR，结果表明其学习的联合表征显著改善了基于HE的9,248个基因预测，并为乳腺癌预后提供了价值。
cs.CV / 12 / 2606.28396
RadarTwin: Scene-Specific mmWave Radar Simulation and Learning for Mobile Indoor Perception
RadarTwin：针对移动室内感知的场景特定毫米波雷达仿真与学习
Abstract
Millimeter-wave (mmWave) radar perception is limited by data scarcity: models trained on existing radar datasets fail to generalize to new objects, environments, and sensing trajectories. We present RadarTwin, a framework for generating deployment-specific radar training data before real data collection. Given a 3D reconstruction of a target space (phone LiDAR, robot-mounted sensing, or RGB-to-3D), RadarTwin uses a vision-language model to infer radar-relevant surface materials and a physics-based ray tracer to synthesize raw frequency-modulated continuous-wave (FMCW) radar measurements with multi-bounce propagation. To study what transfers from simulation to reality, we collect a paired real-simulated dataset spanning household objects, material classes, distances, rotations, translations, and mobile sensing trajectories. We show that simulated and real radar share the same object-discriminative shape and material features, and that modeling the environment's multipath is essential to matching real measurements. A representation trained on simulation alone recognizes real objects at 2.5 times chance with no real radar labels, and a few labeled examples raise this to 95.3% on a 12-way recognition task. RadarTwin enables training radar perception for a new space before any real radar data is collected there.
Chinese Translation
毫米波（mmWave）雷达感知受到数据稀缺的限制：在现有雷达数据集上训练的模型无法推广到新的物体、环境和感知轨迹。我们提出了RadarTwin，一个在真实数据收集之前生成特定部署雷达训练数据的框架。给定目标空间的3D重建（手机LiDAR、机器人搭载传感器或RGB到3D），RadarTwin利用视觉-语言模型推断与雷达相关的表面材料，并使用基于物理的光线追踪器合成具有多次反射传播的原始频率调制连续波（FMCW）雷达测量数据。为了研究从仿真到现实的转移，我们收集了一组配对的真实-仿真数据集，涵盖家庭物体、材料类别、距离、旋转、平移和移动感知轨迹。我们展示了仿真和真实雷达共享相同的物体区分形状和材料特征，并且建模环境的多路径效应对于匹配真实测量至关重要。仅在仿真上训练的表示在没有真实雷达标签的情况下以2.5倍的机会识别真实物体，而少量标记示例将这一比例提高到95.3%在一个12类识别任务中。RadarTwin使得在任何真实雷达数据收集之前为新空间训练雷达感知成为可能。
cs.CV / 13 / 2606.28397
CLOSER-VLN: Closed-Loop Self-Verified Retrieval-Augmented Reasoning for Aerial Vision-Language Navigation
CLOSER-VLN：用于空中视觉-语言导航的闭环自验证检索增强推理
Abstract
Vision-language navigation (VLN) has recently advanced with large language and multimodal models, enabling agents to follow natural-language instructions in unseen environments without training a task-specific navigation policy. However, most existing VLN methods relying on large models still adopt an open-loop decision-execution approach, where candidate actions are generated from instructions and observations but are rarely verified or corrected before execution. This causes critical issues in aerial VLN, where minor errors in intermediate actions may quickly accumulate into large trajectory deviations and lead to target loss. To address this issue, we propose Closed-loop Self-verified Retrieval-augmented Reasoning (CLOSER), a training-policy-free method that sequentially performs action reasoning, reliability verification, targeted retrieval, and action correction in a closed-loop manner before executing concrete actions. We instantiate the CLOSER in aerial VLN tasks and develop a CLOSER-VLN framework, which is composed of three components: a hierarchical reasoner for generating candidate actions based on available information, a multidimensional action verifier for assessing the reliability of actions generated by the reasoner, and a verification-triggered multimodal retriever for retrieving targeted exemplars from a memory bank only when verification fails. We conduct experimental evaluations on the CityNav benchmark, where CLOSER-VLN achieves 32.01% SR and 21.28% SPL on the test-unseen split, confirming the effectiveness of closed-loop reasoning.
Chinese Translation
视觉-语言导航（VLN）最近随着大型语言和多模态模型的发展而取得了进展，使得智能体能够在未见环境中遵循自然语言指令，而无需训练特定任务的导航策略。然而，大多数现有的依赖大型模型的VLN方法仍然采用开放式决策执行方法，其中候选动作是根据指令和观察生成的，但在执行之前很少进行验证或修正。这在空中VLN中造成了严重问题，因为中间动作中的小错误可能迅速累积成大的轨迹偏差并导致目标丢失。为了解决这个问题，我们提出了闭环自验证检索增强推理（CLOSER），这是一种无训练策略的方法，它在执行具体动作之前，以闭环方式顺序执行动作推理、可靠性验证、目标检索和动作修正。我们在空中VLN任务中实例化了CLOSER，并开发了CLOSER-VLN框架，该框架由三个组件组成：一个基于可用信息生成候选动作的层次推理器，一个评估推理器生成的动作可靠性的多维动作验证器，以及一个在验证失败时仅从记忆库中检索目标示例的验证触发多模态检索器。我们在CityNav基准上进行了实验评估，CLOSER-VLN在测试未见分割上达到了32.01%的成功率（SR）和21.28%的成功路径长度（SPL），验证了闭环推理的有效性。
cs.CV / 14 / 2606.28398
Semantic-Aware Generative Image Transmission for Resource-Constrained Visual IoT Systems
面向资源受限视觉物联网系统的语义感知生成图像传输
Abstract
Resource-constrained visual Internet of Things (IoT) systems, such as edge cameras, unmanned sensing platforms, industrial inspection nodes, and remote monitoring sensors, often need to transmit task-relevant visual evidence over low-rate wireless links to an edge/cloud service. Existing image communication methods usually compress or transmit complete global representations, leaving limited room to exploit receiver-side generative restoration. This paper proposes a semantic-aware generative image transmission framework for edge-assisted visual IoT. The image captured by an IoT visual sensor is encoded into a discrete token grid by a VQ encoder. At the IoT transmitter or nearby gateway, token recoverability, estimated from prediction entropy and local structure complexity, is fused with semantic importance obtained from instance segmentation and category-aware scoring. A spatial dispersal sampler then selects the tokens to be transmitted under a bitrate budget. The transmitter sends only the quantization indices of kept tokens and a binary mask map, while the edge/cloud receiver recovers masked tokens through MaskGIT with Halton sequence scheduling. Experiments on Kodak and VisDrone scenes under AWGN and Rayleigh channels show that the proposed method provides a flexible bitrate-quality tradeoff for narrowband visual IoT links. At 0.074 bpp, it uses 44.6% of the transmitted bits of the 0.167-bpp DeepJSCC/WITT reference while achieving 29.9 dB PSNR. A pseudo-GT downstream detection study on Kodak further shows that semantic-aware masking preserves task-relevant objects better than random masking at both 30% and 50% mask ratios.
Chinese Translation
资源受限的视觉物联网（IoT）系统，如边缘摄像头、无人感知平台、工业检测节点和远程监测传感器，通常需要通过低速无线链路将与任务相关的视觉证据传输到边缘/云服务。现有的图像通信方法通常压缩或传输完整的全局表示，导致在接收端生成恢复的空间有限。本文提出了一种面向边缘辅助视觉物联网的语义感知生成图像传输框架。由物联网视觉传感器捕获的图像通过向量量化（VQ）编码器编码为离散的标记网格。在物联网发射器或附近的网关中，从预测熵和局部结构复杂度估计的标记可恢复性与通过实例分割和类别感知评分获得的语义重要性相融合。然后，空间分散采样器在比特率预算下选择要传输的标记。发射器仅发送保留标记的量化索引和二进制掩码图，而边缘/云接收器通过使用Halton序列调度的MaskGIT恢复被掩盖的标记。在AWGN和Rayleigh信道下对Kodak和VisDrone场景的实验表明，所提出的方法为窄带视觉物联网链路提供了灵活的比特率-质量权衡。在0.074 bpp时，它使用了0.167-bpp DeepJSCC/WITT参考的44.6%传输比特，同时实现了29.9 dB的峰值信噪比（PSNR）。在Kodak上的伪GT下游检测研究进一步表明，语义感知掩蔽在30%和50%掩蔽比率下比随机掩蔽更好地保留了与任务相关的对象。
cs.CV / 15 / 2606.28399
Meta-learning as a principle for human-like visual representations
元学习作为类人视觉表征的原则
Abstract
The structure of human visual representations underpins our capacity for adaptive behaviour. While pretrained neural networks model human visual representations with unprecedented success, a large discrepancy remains. We propose one reason: these networks optimise a single fixed objective, whereas human representations must support open-ended tasks. We hypothesise this flexibility arises from meta-learning (learning to learn), a pressure shaping representations to acquire new tasks from few observations. To test this, we train a sequence model, without any supervision from human data, across thousands of semantically rich tasks mapping images to high-level concepts. Compared to their pretrained base encoders, meta-learned representations better predict human similarity judgements, semantic rule learning, and high-level visual cortex. Behavioural gains depend on disentangled, high-level task distributions, while brain alignment is driven primarily by the learning-to-learn pressure. Our results suggest the flexibility of human visual representations reflects the functional demand to learn new semantic relationships on the fly.
Chinese Translation
人类视觉表征的结构支撑了我们适应性行为的能力。尽管预训练的神经网络在模拟人类视觉表征方面取得了前所未有的成功，但仍存在较大差距。我们提出一个原因：这些网络优化的是单一固定目标，而人类表征必须支持开放式任务。我们假设这种灵活性源于元学习（meta-learning，学习如何学习），这种压力促使表征从少量观察中获取新任务。为了验证这一点，我们在数千个语义丰富的任务上训练了一个序列模型，这些任务将图像映射到高级概念，而没有任何来自人类数据的监督。与其预训练的基础编码器相比，元学习的表征更好地预测了人类的相似性判断、语义规则学习和高级视觉皮层。行为上的提升依赖于解耦的高级任务分布，而大脑对齐主要受学习如何学习的压力驱动。我们的结果表明，人类视觉表征的灵活性反映了在飞速变化的环境中学习新语义关系的功能需求。
cs.CV / 16 / 2606.28401
Vision-driven Preference Synthesis for Mitigating Hallucinations in VLMs
基于视觉驱动的偏好合成以减轻视觉语言模型中的幻觉现象
Abstract
Vision-Language Models (VLMs) have shown strong performance in visual understanding, yet they still suffer from hallucinations, generating content that is not grounded in the image. Preference alignment is a promising approach to improve visual faithfulness, but its success depends heavily on how preference pairs are constructed. Existing methods exhibit two key limitations; (a) intervention-based methods often introduce significant deviation from the policy distribution, and (b) sampling-based methods often underuse visual information during the construction. In this paper, we propose ViPSy (Vision-driven Preference Synthesis), a framework for constructing preference data that are both policy-aligned and visually grounded. Our framework consists of two stages; in the first stage, ViPSy derives a visual cue from recurring object-level content across semantically aligned image variants, so preference construction can rely on visual information rather than language priors. In the second stage, ViPSy conditions the policy's own rollouts on this cue, allowing candidates to be guided by visually grounded content while staying close to the policy's response distribution. The resulting candidates remain close to the policy's response distribution while better leveraging visual information from the image. Experiments show that the resulting VLM, preference-aligned with ViPSy-constructed preference pairs, achieves a new state-of-the-art in hallucination mitigation. Compared with the previous state-of-the-art method, it reduces hallucination rates on AMBER and Object HalBench by 35.7% and 24.5%, respectively. The resulting model further improves on general visual grounding benchmarks, e.g., MMStar, MMVP, and CV-Bench, while also yielding gains in semantic segmentation and ImageNet linear probing, underscoring the effectiveness of our framework in enhancing the model's visual capabilities.
Chinese Translation
视觉语言模型（VLMs）在视觉理解方面表现出色，但仍然存在幻觉现象，生成与图像无关的内容。偏好对齐是一种有前景的方法，可以提高视觉的真实性，但其成功在很大程度上依赖于偏好对的构建方式。现有方法存在两个主要局限性：（a）基于干预的方法往往会引入与策略分布的显著偏差；（b）基于采样的方法在构建过程中往往未能充分利用视觉信息。本文提出了ViPSy（基于视觉驱动的偏好合成），这是一个构建既与策略对齐又与视觉相结合的偏好数据的框架。我们的框架分为两个阶段；在第一阶段，ViPSy从语义对齐的图像变体中提取重复的对象级内容作为视觉线索，从而使偏好构建可以依赖于视觉信息而非语言先验。在第二阶段，ViPSy基于该线索对策略的自身回滚进行条件化，使候选者能够在保持接近策略响应分布的同时，受到视觉基础内容的引导。最终的候选者在保持接近策略响应分布的同时，更好地利用了图像中的视觉信息。实验表明，使用ViPSy构建的偏好对进行偏好对齐的VLM在减轻幻觉方面达到了新的最先进水平。与之前的最先进方法相比，它在AMBER和Object HalBench上的幻觉率分别降低了35.7%和24.5%。所得到的模型在一般视觉基础基准（如MMStar、MMVP和CV-Bench）上进一步提高，同时在语义分割和ImageNet线性探测中也取得了进展，突显了我们框架在增强模型视觉能力方面的有效性。
cs.CV / 17 / 2606.28402
DCSNet: Multiscale Feature Aggregation for Small Medical Object Segmentation with Detection-guided Hierarchical Cropping
DCSNet：基于检测引导的分层裁剪的小型医学物体分割的多尺度特征聚合
Abstract
Small object segmentation in medical imaging is primarily hindered by class imbalance and inherent boundary complexity. Consequently, conventional global networks frequently fail to detect sparse targets or suffer from severe edge degradation. To overcome these limitations, we propose the Detection-guided Cropping Segmentation Network (DCSNet), an end-to-end framework that transforms global dense prediction into a localized refinement process. This framework integrates two core components, namely Detection-guided Hierarchical Cropping (DGHC) and Multiscale Feature Aggregation (MSFA). The DGHC module leverages region proposals to dynamically extract object-centric features, effdataectively filtering out massive background interference to mitigate class imbalance. Subsequently, the MSFA module operates strictly within these purified regions, synergizing a Transformer encoder with a pixel-adaptive fusion strategy. This mechanism dynamically aggregates multiscale features to capture both semantic context and fine-grained details for sharp boundary delineation. Extensive experiments across three diverse medical datasets demonstrate that DCSNet significantly outperforms existing state-of-the-art methods, yielding substantial improvements in boundary precision and offering a highly robust solution for clinical micro-lesion segmentation.
Chinese Translation
医学影像中的小物体分割主要受到类别不平衡和固有边界复杂性的阻碍。因此，传统的全局网络常常无法检测稀疏目标或遭遇严重的边缘退化。为克服这些限制，我们提出了检测引导裁剪分割网络（DCSNet），这是一个端到端的框架，将全局密集预测转化为局部精细化过程。该框架集成了两个核心组件，即检测引导分层裁剪（DGHC）和多尺度特征聚合（MSFA）。DGHC模块利用区域提议动态提取以物体为中心的特征，有效过滤掉大量背景干扰，以减轻类别不平衡。随后，MSFA模块严格在这些净化区域内操作，将Transformer编码器与像素自适应融合策略协同工作。该机制动态聚合多尺度特征，以捕捉语义上下文和细粒度细节，从而实现清晰的边界划分。在三个不同的医学数据集上进行的广泛实验表明，DCSNet显著优于现有的最先进方法，在边界精度方面取得了显著提升，并为临床微病变分割提供了一种高度稳健的解决方案。
cs.CV / 18 / 2606.28405
Enhancing Layer Interaction Using Key-Correlated Layer Attention
通过关键相关层注意力增强层间交互
Abstract
Recent advances in network architecture design have introduced layer attention to enhance inter-layer interactions. In such frameworks, each layer queries all preceding layers to establish cross-layer connections. However, layer attention results in quadratic computational complexity with respect to network depth. To mitigate this issue, prior works have proposed Recurrent Layer Attention (RLA) and linear attention mechanisms, which suffer from static information updates and limited long-range cross-layer dependency modeling. To overcome these limitations, we propose Key-Correlated Layer Attention (KCLA), inspired by our observation that Key representations in layer attention exhibit high cosine similarity. KCLA achieves linear computational complexity while preserving dynamic information updates, directly derived from the foundational definition of layer attention. Furthermore, KCLA maintains long-range cross-layer connections and features a fixed spatial complexity, independent of network depth. Empirical evaluations demonstrate that KCLA delivers good performance across diverse tasks, including image recognition, object detection, and medical image segmentation. The code is publicly available at https://github.com/bgx666/KCLA.
Chinese Translation
近期在网络架构设计方面的进展引入了层注意力，以增强层间交互。在此类框架中，每一层都会查询所有前面的层，以建立跨层连接。然而，层注意力在网络深度方面会导致二次计算复杂度。为了解决这一问题，之前的研究提出了递归层注意力（Recurrent Layer Attention, RLA）和线性注意力机制，这些方法在静态信息更新和有限的长距离跨层依赖建模方面存在不足。为了克服这些限制，我们提出了关键相关层注意力（Key-Correlated Layer Attention, KCLA），这一方法的灵感来源于我们观察到层注意力中的关键表示具有较高的余弦相似性。KCLA 实现了线性计算复杂度，同时保留了动态信息更新，直接源自层注意力的基础定义。此外，KCLA 维持了长距离的跨层连接，并具有固定的空间复杂度，与网络深度无关。实证评估表明，KCLA 在图像识别、目标检测和医学图像分割等多种任务中表现良好。代码已公开发布在 https://github.com/bgx666/KCLA。
cs.CV / 19 / 2606.28410
RSGPNet: Geometric Prompting for Remote Sensing Open-Vocabulary Semantic Segmentation
RSGPNet：用于遥感开放词汇语义分割的几何提示
Abstract
Open-vocabulary semantic segmentation (OVSS) enables text-guided segmentation of unseen objects, breaking fixed-class limitations to achieve open-world understanding. However, existing OVSS methods primarily focus on modifying the CLIP attention mechanism, which still suffers from unstable local segmentation for remote sensing (RS) domain. To address these limitations, we propose RSGPNet, a training-free geometric prompting framework for RS OVSS that refines segmentation by leveraging object geometric areas and consistency constraints. Specifically, RSGPNet comprises three core modules: a Text-guided Coarse Mask module (TCM), a Geometric Re-prompting Module (GRP), and a Coarse-to-fine Consistency Verification Mechanism (CVM). TCM utilizes text prompts and the input image to construct initial coarse segmentation masks. GRP then converts these coarse masks into geometric box prompts, feeding them back into the segmentation model to generate refined masks. Finally, CVM employs consistency computation to prevent prompting from reinforcing erroneous regions. They allow the model to improve segmentation accuracy in complex areas, such as category boundaries. Extensive experiments on RS datasets demonstrate that RSGPNet significantly outperforms state-of-the-art methods across both quantitative and qualitative metrics while exhibiting excellent interpretability. The code is released at \href{https://github.com/wangshanwen001/RSGPNet}{https://github.com/wangshanwen001/RSGPNet}.
Chinese Translation
开放词汇语义分割（OVSS）实现了对未见物体的文本引导分割，打破了固定类别的限制，以实现开放世界理解。然而，现有的OVSS方法主要集中在修改CLIP注意机制上，这在遥感（RS）领域仍然存在局部分割不稳定的问题。为了解决这些局限性，我们提出了RSGPNet，一种无需训练的几何提示框架，用于RS OVSS，通过利用物体几何区域和一致性约束来优化分割。具体而言，RSGPNet包含三个核心模块：文本引导粗略掩膜模块（TCM）、几何重新提示模块（GRP）和粗到细一致性验证机制（CVM）。TCM利用文本提示和输入图像构建初始粗略分割掩膜。然后，GRP将这些粗略掩膜转换为几何框提示，并将其反馈到分割模型中以生成精细掩膜。最后，CVM采用一致性计算来防止提示强化错误区域。这使得模型能够在复杂区域（如类别边界）中提高分割准确性。在RS数据集上的大量实验表明，RSGPNet在定量和定性指标上显著超越了最先进的方法，同时展现出优异的可解释性。代码已发布在 https://github.com/wangshanwen001/RSGPNet。
cs.CV / 20 / 2606.28416
AEGIS: A Semantic GAN and Evidential Learning Frameworkfor Robust Adversarial Detection in Vision Sensors
AEGIS：一种语义生成对抗网络和证据学习框架，用于视觉传感器中的鲁棒对抗检测
Abstract
Deep neural networks (DNNs) have shown outstanding performance in visual recognition tasks within vision sensor networks; however, they are still vulnerable to adversarial manipulations and imperceptible perturbations that can lead to erroneous predictions. To address that, this paper presents AEGIS, a semantic aware and uncertainty guided adversarial detection framework designed for robust image classification in vision sensors pipelines. At its core, a SemantiGAN module functions as a multi class semantic discriminator, identifying and filtering visually inconsistent adversarial inputs before they propagate further in the pipeline. For inputs that pass this stage, a stochastic augmentation process generates test time variations, from which handcrafted instability metrics FlipScore, Prediction Inconsistency, Layerwise Cosine Similarity (early and mid layers), and Entropy are computed. These features are aggregated into a compact five dimensional vector and processed by an Evidential Deep Learning (EDL) classifier, which models output evidence using a Dirichlet distribution to yield both class predictions and calibrated uncertainty estimates. Evaluations on the Tiny ImageNet dataset across six categories clean, FGSM, PGD, patch based, functional, and geometric attacks demonstrate the effectiveness of AEGIS. The proposed framework achieves an AUROC of 92.1\%, an AUPRC of 90.2\%, and an accuracy of 90.7\%, outperforming conventional softmax-based detectors in terms of detection performance, robustness, interpretability, and uncertainty calibration.
Chinese Translation
深度神经网络（DNNs）在视觉传感器网络中的视觉识别任务中表现出色；然而，它们仍然容易受到对抗性操控和不可察觉的扰动的影响，这可能导致错误的预测。为了解决这个问题，本文提出了AEGIS，一个语义感知和不确定性引导的对抗检测框架，旨在增强视觉传感器管道中的图像分类鲁棒性。其核心是一个SemantiGAN模块，作为多类语义鉴别器，在对抗输入进一步传播之前识别并过滤视觉上不一致的对抗样本。对于通过此阶段的输入，随机增强过程生成测试时的变化，并计算手工设计的不稳定性指标，包括FlipScore、预测不一致性、逐层余弦相似度（早期和中间层）以及熵。这些特征被聚合成一个紧凑的五维向量，并由证据深度学习（Evidential Deep Learning, EDL）分类器处理，该分类器使用Dirichlet分布建模输出证据，以产生类别预测和校准的不确定性估计。在Tiny ImageNet数据集上对六类攻击（干净样本、FGSM、PGD、基于补丁的、功能性和几何攻击）进行的评估表明AEGIS的有效性。所提出的框架在检测性能、鲁棒性、可解释性和不确定性校准方面超越了传统的基于softmax的检测器，达到了92.1%的AUROC、90.2%的AUPRC和90.7%的准确率。
cs.CV / 21 / 2606.28417
DiffRGD: An Inference-Time Diffusion Guidance Through Riemannian Gradient Descent
DiffRGD：一种通过黎曼梯度下降的推理时扩散引导
Abstract
Recently, diffusion models have been widely adopted in generative modeling and have served as foundational models for many image generation tasks. To control the generation without costly re-training or fine-tuning, many works seek inference-time guidance methods to steer the latent via a differentiable objective at inference time. However, these methods cannot effectively preserve the original Gaussian distribution because they introduce distributional drift, thereby degrading the sample quality. To address this gap, we propose DiffRGD, a distribution-aware guidance framework that explicitly preserves the latent Gaussian structure. DiffRGD formulates each sampling step as a constrained optimization problem on a spherical manifold induced by the latent Gaussian distribution, and solves it efficiently via Riemannian Gradient Descent (RGD). DiffRGD is a plug-and-play method that can be seamlessly integrated into any pre-trained diffusion model. Extensive experiments demonstrate that DiffRGD outperforms previous methods in most image restoration and conditional generation tasks. Our codebase is available at https://github.com/jwliao1209/DiffRGD.
Chinese Translation
近年来，扩散模型在生成建模中得到了广泛应用，并成为许多图像生成任务的基础模型。为了在不进行昂贵的重新训练或微调的情况下控制生成，许多研究寻求在推理时通过可微分目标引导潜在变量的方法。然而，这些方法无法有效保持原始的高斯分布，因为它们引入了分布漂移，从而降低了样本质量。为了解决这一问题，我们提出了DiffRGD，一种显式保持潜在高斯结构的分布感知引导框架。DiffRGD将每个采样步骤公式化为在由潜在高斯分布诱导的球面流形上的约束优化问题，并通过黎曼梯度下降（Riemannian Gradient Descent, RGD）高效求解。DiffRGD是一种即插即用的方法，可以无缝集成到任何预训练的扩散模型中。大量实验表明，DiffRGD在大多数图像恢复和条件生成任务中优于之前的方法。我们的代码库可在 https://github.com/jwliao1209/DiffRGD 获取。
cs.CV / 22 / 2606.28419
MedDiffuseMix: Preserving Diagnostic Evidence with Saliency-Aware Diffusion Medical Image Data Augmentatio
MedDiffuseMix：基于显著性引导的扩散医学图像数据增强以保留诊断证据
Abstract
Limited data availability, class imbalance, and domain variability remain major barriers to reliable medical image classification. Conventional augmentation can improve training diversity but may distort diagnostically informative structures, whereas unconstrained generative augmentation may introduce label-inconsistent content. This paper proposes MedDiffuseMix, a saliency-guided diffusion mixing framework for controlled medical image augmentation. The method uses classifier-derived saliency maps to separate high-saliency diagnostic regions from low-saliency background areas and applies diffusion-guided mixing mainly to regions with lower diagnostic importance. Adaptive mixing, Gaussian boundary blending, and a saliency-preservation constraint reduce semantic distortion and reject or attenuate samples that shift model attention away from clinically relevant evidence. The framework is evaluated on four public benchmarks: the Radiological Society of North America pneumonia chest radiography dataset, Musculoskeletal Radiographs, PatchCamelyon, and the Breast Cancer Histopathological Image Classification dataset. Experiments with convolutional and transformer-based classifiers show that MedDiffuseMix improves accuracy, F1-score, and area under the receiver operating characteristic curve compared with standard augmentation, Mixup, GenMix, SaliencyMix, and diffusion-based augmentation baselines. Ablation studies confirm the importance of saliency guidance, adaptive region mixing, and smooth boundary blending. Visual attribution analysis further indicates that MedDiffuseMix better preserves diagnostically salient regions. These results suggest that saliency-guided diffusion mixing is an effective augmentation strategy for limited-data medical image classification.
Chinese Translation
有限的数据可用性、类别不平衡和领域变异性仍然是可靠医学图像分类的主要障碍。传统的增强方法可以改善训练多样性，但可能会扭曲具有诊断信息的结构，而不受限制的生成增强可能会引入标签不一致的内容。本文提出了MedDiffuseMix，一种基于显著性引导的扩散混合框架，用于控制医学图像增强。该方法利用分类器生成的显著性图将高显著性诊断区域与低显著性背景区域分离，并主要对低诊断重要性的区域应用扩散引导混合。自适应混合、高斯边界融合和显著性保留约束减少了语义扭曲，并拒绝或减弱那些使模型注意力偏离临床相关证据的样本。该框架在四个公共基准上进行了评估：北美放射学会肺炎胸部X光数据集、肌肉骨骼X光片、PatchCamelyon和乳腺癌组织病理图像分类数据集。与标准增强、Mixup、GenMix、SaliencyMix和基于扩散的增强基线相比，使用卷积和基于变换器的分类器的实验表明，MedDiffuseMix提高了准确性、F1分数和接收者操作特征曲线下面积。消融研究确认了显著性引导、自适应区域混合和光滑边界融合的重要性。视觉归因分析进一步表明，MedDiffuseMix更好地保留了具有诊断意义的显著区域。这些结果表明，显著性引导的扩散混合是一种有效的有限数据医学图像分类增强策略。
cs.CV / 23 / 2606.28421
JuZhou 1.0 Technical Report: The First Edge-Native Text-to-Image Foundation Model Trained Entirely on China-Developed AI Accelerators
JuZhou 1.0 技术报告：首个完全基于中国开发的 AI 加速器训练的边缘原生文本到图像基础模型
Chen, Ce, Wang, Congrui, Li, Yonglin, Wan, Zhenchen, Geng, Mingyang, Xiao, Junhao, Xing, Zhengpeng, Hu, Yaqing, Wu, Yao, Qu, Zhaoyang, Lan, Long, Liu, Xinwang, Peng, Yingqi, Li, Shijia, Zhang, Zufeng, Ma, Chen, Zhou, Jingjing, Wang, Xingyu, Lu, Qilin, Jiang, Bin, Sun, Qilin, Gu, Shanzhi, Jin, Yaoguang, Liu, Tongliang, Ma, Kede, Peng, Yifan
Abstract
Text-to-image (T2I) diffusion models typically require substantial computational resources and cloud infrastructure, posing significant challenges for edge deployment in terms of latency, cost, and user privacy. We present JuZhou 1.0, an ultra-lightweight T2I foundation model designed for fully offline, on-device execution. JuZhou 1.0 achieves its efficiency through four key designs: (1) a compact image-generation backbone consisting of a 0.385B-parameter denoising U-Net and a 1.90M-parameter distilled decoder, totaling approximately 0.387B parameters; (2) Rectified Flow training combined with DMD2 distillation, reducing inference to 4 sampling steps; (3) Chinese semantic alignment trained on 9M curated image-text pairs, enabling direct Chinese prompting without external translation at inference time; and (4) a training and distillation pipeline completed on domestically developed Sugon K100 AI accelerators without relying on NVIDIA GPUs for training or distillation. Despite its compact scale, the 28-step base model of JuZhou 1.0 achieves an overall GenEval score of 0.69, outperforming published baselines including SDXL (2.6B, 0.55), SD3-Medium (2B, 0.62), and IF-XL (4.3B, 0.61). We further validate the full poetry-to-image pipeline on Android and the core CLIP-U-Net-VAE generation branch on iOS. On a smartphone powered by the Snapdragon 8 Elite Gen 5 Mobile Platform, the 4-step U-Net denoising branch runs in approximately 1.6 seconds, while the full Android poetry-to-image pipeline takes 4.5 seconds with on-device prompt refinement on Xiaomi 17 Pro Max. These results position JuZhou 1.0 as a practical approach to mobile text-to-image generation and provide a concrete reference for Chinese-native generation, domestic-compute training, and fully offline on-device deployment after one-time installation.
Chinese Translation
文本到图像（T2I）扩散模型通常需要大量计算资源和云基础设施，这在延迟、成本和用户隐私方面为边缘部署带来了重大挑战。我们提出了 JuZhou 1.0，这是一种超轻量级的 T2I 基础模型，旨在实现完全离线的设备端执行。JuZhou 1.0 通过四个关键设计实现了其高效性：（1）一个紧凑的图像生成主干，由 0.385B 参数的去噪 U-Net 和 1.90M 参数的蒸馏解码器组成，总参数量约为 0.387B；（2）结合 DMD2 蒸馏的校正流训练，将推理步骤减少到 4 步；（3）在 9M 精心挑选的图像-文本对上训练的中文语义对齐，使得在推理时可以直接使用中文提示而无需外部翻译；（4）在国内开发的曙光 K100 AI 加速器上完成的训练和蒸馏流程，无需依赖 NVIDIA GPU 进行训练或蒸馏。尽管规模紧凑，JuZhou 1.0 的 28 步基础模型整体 GenEval 得分为 0.69，超越了已发布的基准，包括 SDXL（2.6B，0.55）、SD3-Medium（2B，0.62）和 IF-XL（4.3B，0.61）。我们进一步在 Android 上验证了完整的诗歌到图像管道，并在 iOS 上验证了核心 CLIP-U-Net-VAE 生成分支。在搭载 Snapdragon 8 Elite Gen 5 移动平台的智能手机上，4 步 U-Net 去噪分支的运行时间约为 1.6 秒，而完整的 Android 诗歌到图像管道在小米 17 Pro Max 上的设备端提示优化后耗时 4.5 秒。这些结果使 JuZhou 1.0 成为移动文本到图像生成的实用方法，并为中文本土生成、国内计算训练和完全离线的设备端部署提供了具体参考。
cs.CV / 24 / 2606.28516
CLEAR-MoE: Shared-Basis Expert Extraction from Frozen Vision Transformers via Calibration-Driven Layer Selection
CLEAR-MoE：通过校准驱动的层选择从冻结的视觉变换器中提取共享基础专家
Abstract
We present CLEAR-MoE, a four-phase post-training pipeline that converts a frozen pretrained Vision Transformer (ViT) into a sparse Mixture-of-Experts (MoE) model without updating backbone weights. The pipeline (i) scores feed-forward network (FFN) layers by sparsity, clusterability, and output sensitivity; (ii) decomposes selected layers into a shared low-rank SVD basis and per-cluster residual experts using k-means clustering; (iii) trains lightweight routers supervised by cluster labels; and (iv) dispatches tokens through pluggable CUDA backends. On Imagenette with DeiT-Small, CLEAR-MoE retains 99.9% of the dense model's accuracy (86.70 +/- 0.02% versus 86.73%). Extensive ablation studies reveal a consistent empirical finding: the shared SVD basis is the primary factor responsible for preserving accuracy. Random routing, learned routing, and three different router architectures produce nearly identical performance, with accuracy varying by at most 0.06 percentage points (86.62%-86.68%). Accuracy also remains stable across different SVD ranks, expert counts (2-8), calibration set sizes (50-500), and random seeds. This behavior generalizes across five ViT backbones (DeiT-Tiny, DeiT-Small, DeiT-Base, ViT-Small, and ViT-Base), covering models from 5.7M to 86.6M parameters, with accuracy differences <= 0.10 percentage points from their dense counterparts. On a GTX 960 GPU, routing and scatter-gather overhead make the CLEAR-MoE FFN 1.3-1.7x slower than the dense implementation. A dispatch microbenchmark further shows that routing is an order of magnitude more memory-bound than expert matrix multiplications, identifying fused dispatch kernels as a promising direction for future optimization.
Chinese Translation
我们提出了CLEAR-MoE，这是一种四阶段的后训练流程，将冻结的预训练视觉变换器（ViT）转换为稀疏的专家混合模型（MoE），而无需更新主干权重。该流程（i）根据稀疏性、聚类能力和输出敏感性对前馈网络（FFN）层进行评分；（ii）使用k均值聚类将选定层分解为共享的低秩SVD基础和每个聚类的残差专家；（iii）训练由聚类标签监督的轻量级路由器；（iv）通过可插拔的CUDA后端分发令牌。在使用DeiT-Small的Imagenette数据集上，CLEAR-MoE保留了密集模型99.9%的准确率（86.70 +/- 0.02%对比86.73%）。大量的消融研究揭示了一个一致的经验发现：共享的SVD基础是保持准确率的主要因素。随机路由、学习路由和三种不同的路由器架构产生了几乎相同的性能，准确率最多变化0.06个百分点（86.62%-86.68%）。在不同的SVD秩、专家数量（2-8）、校准集大小（50-500）和随机种子下，准确率也保持稳定。这种行为在五个ViT主干（DeiT-Tiny、DeiT-Small、DeiT-Base、ViT-Small和ViT-Base）中具有普遍性，涵盖了从5.7M到86.6M参数的模型，与其密集对应模型的准确率差异不超过0.10个百分点。在GTX 960 GPU上，路由和散点聚集的开销使得CLEAR-MoE的FFN比密集实现慢1.3-1.7倍。进一步的分发微基准测试显示，路由在内存绑定方面比专家矩阵乘法高一个数量级，识别出融合分发内核作为未来优化的有前景方向。
cs.CV / 25 / 2606.28520
Detecting Clinical Hallucinations in LVLMs via Counterfactual Visual Grounding Uncertainty
通过反事实视觉基础不确定性检测大型视觉语言模型中的临床幻觉
Abstract
Large vision-language models (LVLMs) are increasingly used for clinical image understanding, yet they remain vulnerable to \emph{hallucinations}--producing textual findings or attributes not supported by the image. We present a vision-traceable hallucination detection framework that audits arbitrary LVLM responses via visual evidence grounding, requiring neither modification nor internal access to the hidden states of LVLMs. Given an LVLM response, we extract visually verifiable entities and use a medical-domain-adapted Qwen-VL grounding verifier to localize each entity on the input image. To enhance the robustness of our detection method, we introduce a counterfactual entity perturbation method and estimate visual evidence uncertainty by contrasting factual and counterfactual grounding results. Specifically, we compute an entity-level uncertainty score from the positive confidence, counterfactual confidence, and their grounding overlap for binary hallucination decision-making. Experiments on multiple medical imaging modalities and LVLM backbones demonstrate that our method consistently improves hallucination detection performance over recent baselines, while providing interpretable localization evidence and strong cross-model transferability. Code and dataset are available at https://github.com/Agentic-CliniAI/CounterVHD.
Chinese Translation
大型视觉语言模型（LVLMs）在临床图像理解中越来越多地被使用，但它们仍然容易受到 extit{幻觉}的影响——生成与图像不符的文本发现或属性。我们提出了一种可追溯视觉的幻觉检测框架，通过视觉证据基础审核任意LVLM响应，无需修改或内部访问LVLM的隐藏状态。给定LVLM的响应，我们提取可视验证的实体，并使用经过医学领域适配的Qwen-VL基础验证器在输入图像上定位每个实体。为了增强我们检测方法的鲁棒性，我们引入了一种反事实实体扰动方法，并通过对比事实和反事实基础结果来估计视觉证据的不确定性。具体而言，我们从正向置信度、反事实置信度及其基础重叠计算实体级不确定性评分，以进行二元幻觉决策。对多种医学成像模式和LVLM骨干网络的实验表明，我们的方法在幻觉检测性能上始终优于最近的基线，同时提供可解释的定位证据和强大的跨模型迁移能力。代码和数据集可在https://github.com/Agentic-CliniAI/CounterVHD获取。
cs.CV / 26 / 2606.28537
MammoFlow: Multiview Mammogram Synthesis with Anatomically Consistent Flow Matching
MammoFlow：具有解剖一致性流匹配的多视角乳腺X光图像合成
Abstract
Multiview mammography relies on paired craniocaudal (CC) and mediolateral oblique (MLO) views to provide complementary projections of a 3D breast volume, enabling precise anomaly localization. However, acquiring high-quality, balanced datasets remains challenging for deep learning applications. We propose a novel method to synthesize multiview mammograms by leveraging the inherent geometric relationship between CC and MLO views. To enforce an implicit 3D consistency prior during generation, we develop an alignment module that searches a 2D affine transformation subspace to establish optimal anatomical correspondence. Leveraging this alignment, we introduce a pixel-space self-consistency loss based on the Earth Mover's Distance (EMD) between the 1D anteroposterior (AP) axis tissue distributions of the generated images. Integrated into a pretrained flow matching model, MammoFlow forces synthesized pairs to share physically plausible tissue distributions from the chest wall to the nipple. To our knowledge, this is the first work to guide multiview mammogram generation using implicit geometric tissue correspondence. Our method demonstrates superior image quality, passes expert radiologist evaluation, and generates physically consistent pairs that improve downstream classification AUC by 5%. Code is available at https://github.com/XYPB/MammoFlow
Chinese Translation
多视角乳腺摄影依赖于成对的头尾视图（CC）和内外斜视图（MLO）来提供三维乳腺体积的互补投影，从而实现精确的异常定位。然而，获取高质量、平衡的数据集对于深度学习应用仍然具有挑战性。我们提出了一种新颖的方法，通过利用CC和MLO视图之间固有的几何关系来合成多视角乳腺X光图像。为了在生成过程中强制执行隐式的三维一致性先验，我们开发了一个对齐模块，该模块在二维仿射变换子空间中搜索，以建立最佳的解剖对应关系。利用这种对齐，我们引入了一种基于地球搬运工距离（EMD）的像素空间自一致性损失，该损失基于生成图像的1D前后轴（AP）组织分布。集成到一个预训练的流匹配模型中，MammoFlow强制合成的图像对在胸壁到乳头之间共享物理上合理的组织分布。据我们所知，这是首次通过隐式几何组织对应关系指导多视角乳腺X光图像生成。我们的方法展示了优越的图像质量，经过专家放射科医师评估，并生成物理一致的图像对，使下游分类的AUC提高了5%。代码可在 https://github.com/XYPB/MammoFlow 获取。
cs.CV / 27 / 2606.28551
DataComp-VLM: Improved Open Datasets for Vision-Language Models
DataComp-VLM：改进视觉-语言模型的开放数据集
Farina, Matteo, Udandarao, Vishaal, Nguyen, Thao, Kuzucu, Selim, Böther, Maximilian, Hochlehnert, Andreas, Ghosh, Adhiraj, Nezhurina, Marianna, Roth, Karsten, Struber, Joschka, Zhang, Yuhui, Dziadzio, Sebastian, Sui, Elaine, Jahagirdar, Soumya, Ghosh, Dhruba, Hammoud, Hasan, De Min, Thomas, Caldarella, Simone, Mirza, Jehanzeb, Keh, Sedrick, Cherti, Mehdi, Kuehne, Hilde, Schiele, Bernt, Yeung-Levy, Serena, Naeem, Muhammad Ferjad, Tombari, Federico, Klimovic, Ana, Ricci, Elisa, Bethge, Matthias, Oh, Sewoong, Prabhu, Ameya, Tonioni, Alessio, Jitsev, Jenia, Mancini, Massimiliano, Schmidt, Ludwig, Parthasarathy, Nikhil
Abstract
Building performant Vision-Language Models (VLMs) requires carefully curating large-scale training datasets, yet the community lacks systematic benchmarks for evaluating such curation strategies. We introduce DataComp for VLMs (DCVLM), a benchmark for controlled data-centric experiments to improve VLM training. As part of DCVLM, we collect 160 datasets spanning four data types -- image-caption pairs, multimodal interleaved documents, text-only, and instruction-tuning data -- into a corpus of 6T multimodal tokens. DCVLM allows participants to test curation strategies (filtering, mixing, formatting, sampling) across 1B-8B models and 6.25B-200B token budgets. Models are then evaluated on a carefully selected suite of up to 52 downstream benchmarks across 9 domains. We conduct extensive experiments on DCVLM and find that data mixing, not filtering, is key to a high-quality training dataset: instruction-heavy mixtures scale better than caption-heavy ones, with gains widening at larger scales. The resulting dataset, DCVLM-Baseline, enables training an 8B VLM to 63.6% accuracy on our 33-task core suite with 200B training tokens. Compared to FineVision, the state-of-the-art open VLM training dataset, this represents an improvement of +5.4pp. DCVLM and all accompanying artifacts will be made publicly available at https://www.datacomp.ai/dcvlm/.
Chinese Translation
构建高性能的视觉-语言模型（VLMs）需要精心策划大规模的训练数据集，但目前社区缺乏系统的基准来评估这种策划策略。我们推出了针对 VLM 的 DataComp（DCVLM），这是一个用于控制数据中心实验以改善 VLM 训练的基准。作为 DCVLM 的一部分，我们收集了 160 个数据集，涵盖四种数据类型——图像-标题对、多模态交错文档、仅文本和指令调优数据——形成一个包含 6T 多模态标记的语料库。DCVLM 允许参与者在 1B-8B 模型和 6.25B-200B 标记预算下测试策划策略（过滤、混合、格式化、采样）。然后，模型在经过精心挑选的最多 52 个下游基准测试中进行评估，涵盖 9 个领域。我们在 DCVLM 上进行了广泛的实验，发现数据混合而非过滤是高质量训练数据集的关键：以指令为主的混合比以标题为主的混合扩展得更好，且在更大规模下收益更为显著。最终生成的数据集 DCVLM-Baseline，使得在我们的 33 项核心任务上，训练一个 8B VLM 达到 63.6% 的准确率，使用 200B 的训练标记。与最先进的开放 VLM 训练数据集 FineVision 相比，这代表着 +5.4 个百分点的提升。DCVLM 及所有相关文献将公开发布于 https://www.datacomp.ai/dcvlm/。
cs.CV / 28 / 2606.28568
KM-Speaker: Keypoint-Based Style Control for High-Quality Speech-Driven 3D Facial Animation and Dialogue Localization
KM-Speaker：基于关键点的风格控制用于高质量语音驱动的3D面部动画和对话定位
Abstract
Speech-driven 3D facial animation methods face significant challenges in simultaneously achieving high-fidelity motion and precise artistic control at production quality. Existing controllable models typically learn global style control by relying on large-scale, low-quality \emph{in-the-wild} datasets that compromise overall animation realism. Furthermore, these frameworks often lack the fine-grained temporal precision required for demanding tasks such as dialogue localization (e.g., dubbing), where matching specific facial expressions is as critical as lip synchronization. We present KM-Speaker (Keypoint-Matching Speaker), a novel keypoint-conditioned flow-based generative framework that provides both global style guidance and frame-level temporal control from reference performances. We propose a disentanglement strategy that separates audio-driven lip motion from keypoint-driven upper-face dynamics, together with a global style context preservation mechanism to ensure coherent full-face expressiveness. KM-Speaker advances example-based 3D facial animation by achieving high-fidelity motion and flexible controllability in a data-constrained setting, consistently outperforming state-of-the-art methods in lip-sync accuracy, style adherence, and expressive temporal control.
Chinese Translation
语音驱动的3D面部动画方法在同时实现高保真运动和精确艺术控制方面面临重大挑战，尤其是在制作质量上。现有的可控模型通常依赖于大规模、低质量的 extit{in-the-wild}数据集来学习全局风格控制，这影响了整体动画的真实感。此外，这些框架往往缺乏在对话定位等高要求任务中所需的细粒度时间精度（例如，配音），在这些任务中，匹配特定的面部表情与唇部同步同样重要。我们提出了KM-Speaker（关键点匹配说话者），这是一种新颖的基于关键点条件的流式生成框架，能够提供来自参考表演的全局风格指导和帧级时间控制。我们提出了一种解耦策略，将音频驱动的唇部运动与关键点驱动的上半脸动态分离，同时引入全局风格上下文保持机制，以确保面部表情的连贯性。KM-Speaker通过在数据受限的环境中实现高保真运动和灵活的可控性，推动了基于示例的3D面部动画的发展，在唇部同步精度、风格遵循和表现性时间控制方面始终优于最先进的方法。
cs.CV / 29 / 2606.28570
Digitizing Coaching Intelligence: An Agentic Framework for Holistic Athlete Profiling using VLM and RAG
数字化教练智能：基于代理的整体运动员画像框架，使用VLM和RAG
Abstract
Athlete assessment is a critical process for tracking physical progress and identifying elite talent. However, during mass recruitment drives, traditional methods rely on manual observation, which is inherently subjective and unscalable, or basic computer vision (CV) systems limited to quantitative repetition counting. These standard approaches lack the "coaching intelligence" required to evaluate qualitative physiological markers such as form degradation, spinal articulation, and fatigue. This paper presents a novel, LLM-based hybrid agentic framework for automated, holistic athlete profiling that strictly aligns with the Sports Authority of India (SAI) assessment protocols. Orchestrated via LangGraph, our dual-pipeline architecture synthesizes the geometric precision of CV (MediaPipe) for kinematic tracking with the semantic reasoning of Vision-Language Models (Llama-4-scout). To overcome the latency and token constraints associated with multimodal video processing, we introduce a 3 X 3 "Smart Grid" temporal chunking strategy, reducing computational overhead by over 88% while preserving critical temporal continuity. To ensure data integrity and mitigate hallucination, the framework pioneers an autonomous "LLM-as-a-Judge" self-correction loop that cross-references quantitative and qualitative metrics before persistence. Finally, we implement a dual-persistence Retrieval-Augmented Generation (RAG) pipeline utilizing a vector search engine (ChromaDB). This enables coaches to bypass rigid SQL databases and perform complex semantic queries (e.g., "Identify athletes with high endurance but poor core rigidity") using natural language. Experimental results demonstrate that this multi-agent approach significantly bridges the gap between raw biometric tracking and actionable coaching insights, offering a scalable, objective solution for national talent identification.
Chinese Translation
运动员评估是跟踪身体进展和识别精英人才的关键过程。然而，在大规模招聘活动中，传统方法依赖于人工观察，这本质上是主观且不可扩展的，或者依赖于仅限于定量重复计数的基本计算机视觉（CV）系统。这些标准方法缺乏评估形式退化、脊柱关节活动和疲劳等定性生理指标所需的“教练智能”。本文提出了一种新颖的基于大语言模型（LLM）的混合代理框架，用于自动化的整体运动员画像，严格遵循印度体育管理局（SAI）的评估协议。通过LangGraph协调，我们的双管道架构将计算机视觉（MediaPipe）在运动学跟踪中的几何精度与视觉-语言模型（Llama-4-scout）的语义推理相结合。为了克服与多模态视频处理相关的延迟和令牌限制，我们引入了一种3 X 3“智能网格”时间切块策略，将计算开销减少超过88%，同时保持关键的时间连续性。为了确保数据完整性并减轻幻觉现象，该框架开创了一个自主的“LLM作为裁判”自我校正循环，在持久化之前交叉引用定量和定性指标。最后，我们实施了一个双持久化的增强检索生成（RAG）管道，利用向量搜索引擎（ChromaDB）。这使得教练能够绕过僵化的SQL数据库，使用自然语言执行复杂的语义查询（例如，“识别耐力高但核心刚性差的运动员”）。实验结果表明，这种多代理方法显著缩小了原始生物特征跟踪与可操作教练洞察之间的差距，为国家人才识别提供了一种可扩展、客观的解决方案。
cs.CV / 30 / 2606.28581
SatSplat: Geometrically-Accurate Gaussian Splatting for Satellite Imagery
SatSplat：用于卫星图像的几何精确高斯点云渲染
Abstract
High-resolution satellite imagery demands 3D reconstruction methods that deliver both speed and geometric accuracy. Recent adaptations of 3D Gaussian Splatting (3DGS) to satellite imagery demonstrate strong efficiency, but reconstruction quality often degrades under diverse illumination across multi-date, high-altitude acquisitions (with small intersection angles), limiting applicability to remote sensing and vision tasks. We present SatSplat, the first framework to adapt 2D Gaussian Splatting (2DGS) to satellite photogrammetry, with online camera adjustment. We approximate satellite cameras with an affine model and learn a minimal delta parameterization for in-splat camera refinement from dense observations. The formulation is implemented with a 2DGS scene representation. To handle time-varying shadows and illumination changes, we integrate geometric shadow mapping and per-camera color correction during training. Across the evaluated DFC2019 and IARPA2016 benchmark sites, SatSplat achieves strong geometric accuracy while significantly outperforming prior 3DGS-based baselines. On our processed DFC2019 benchmark, SatSplat reduces mean absolute error by 11.93% and peak video memory by 31% relative to the previous state of the art. Our approach enables large-scale digital surface modeling with practical computational efficiency. The project page is available at https://gdaosu.github.io/satsplat/.
Chinese Translation
高分辨率卫星图像需要提供速度和几何精度的3D重建方法。最近将3D高斯点云渲染（3DGS）应用于卫星图像的改进展示了良好的效率，但在多日期、高海拔采集（交叉角度小）下，重建质量往往会因不同的光照条件而下降，从而限制了其在遥感和视觉任务中的应用。我们提出了SatSplat，这是第一个将2D高斯点云渲染（2DGS）适配于卫星摄影测量的框架，并实现了在线相机调整。我们用仿射模型近似卫星相机，并从密集观测中学习最小的增量参数化，以进行点云内相机的精细调整。该公式通过2DGS场景表示实现。为了处理随时间变化的阴影和光照变化，我们在训练过程中集成了几何阴影映射和每个相机的颜色校正。在评估的DFC2019和IARPA2016基准站点中，SatSplat实现了强大的几何精度，同时显著超越了之前基于3DGS的基线。在我们处理的DFC2019基准上，SatSplat将平均绝对误差降低了11.93%，并将峰值视频内存降低了31%，相较于之前的最先进技术。我们的方法使大规模数字表面建模具备了实用的计算效率。项目页面可访问：https://gdaosu.github.io/satsplat/
cs.CV / 31 / 2606.28593
Animation2Code: Evaluating Temporal Visual Reasoning in Video-to-Code Generation
Animation2Code：评估视频到代码生成中的时间视觉推理
Abstract
While recent vision-language models (VLMs) have achieved significant improvements on static visual-to-code tasks such as generating code for webpages, charts, or SVGs, it remains unclear whether they can recover temporal dynamics when motion is present. To this end, we introduce Animation2Code, a benchmark for evaluating temporal visual reasoning via reconstructing executable web animation code from videos. Animation2Code consists of 1,069 web animation videos with diverse visual appearances and motion patterns, paired with corresponding HTML/CSS/JavaScript implementations. We propose two human-aligned metrics, appearance similarity and temporal similarity, which allow us to disentangle visual fidelity from temporal alignment when comparing rendered animations against ground-truth samples. Benchmarking state-of-the-art VLMs on this dataset shows that current VLMs struggle to maintain temporal consistency in reconstruction, even when achieving high appearance similarity, including under finetuning and iterative refinement settings. Code and data are available at https://anya-ji.github.io/animation2code-website .
Chinese Translation
尽管最近的视觉-语言模型（VLMs）在静态视觉到代码任务上取得了显著进展，例如为网页、图表或SVG生成代码，但在运动存在时，它们是否能够恢复时间动态仍然不清楚。为此，我们提出了Animation2Code，这是一个通过从视频重建可执行的网页动画代码来评估时间视觉推理的基准。Animation2Code包含1,069个具有多样视觉外观和运动模式的网页动画视频，并配有相应的HTML/CSS/JavaScript实现。我们提出了两个与人类对齐的指标：外观相似性和时间相似性，这使我们能够在比较渲染动画与真实样本时，将视觉保真度与时间对齐分开。在该数据集上对最先进的VLM进行基准测试表明，当前的VLM在重建时难以保持时间一致性，即使在外观相似性高的情况下，包括在微调和迭代优化设置下。代码和数据可在 https://anya-ji.github.io/animation2code-website 获取。
cs.CV / 32 / 2606.28604
IMU-HOI: A Symbiotic Framework for Coherent Human-Object Interaction and Motion Capture via Contact-Conscious Inertial Fusion
IMU-HOI：一种用于一致的人体-物体交互和运动捕捉的共生框架，通过接触意识惯性融合
Abstract
Capturing full-body human motion with object interactions is crucial for AR/VR and robotics applications, yet it remains challenging for conventional vision-based methods due to occlusions and constrained capture volumes. Inertial measurement units (IMUs) offer a compelling alternative without line-of-sight requirements, but existing IMU-based motion capture assumes an isolated human and ignores object contacts and dynamics. To bridge this gap, we present IMU-HOI, a novel framework that jointly recovers full-body human pose and 6-DoF object trajectory from sparse IMUs on the body and object, explicitly modeling human-object interaction. Our approach first infers probabilistic hand-object contacts directly from IMU streams and uses them as a high-level signal to route between kinematic and inertial reasoning. These contact cues drive a three-stage fusion pipeline that refines human pose and root translation, and fuses hand-based forward kinematics with object-IMU integration for object motion, yielding coherent, drift-resilient trajectories for both human and object. Experiments on challenging human-object interaction scenarios demonstrate substantial accuracy gains over prior inertial motion capture methods. Moreover, IMU-HOI can be plugged into existing sparse-IMU mocap backbones with minimal changes, effectively extending the scope of purely inertial motion capture from isolated humans to full human-object interaction and joint motion estimation.
Chinese Translation
捕捉全身人体运动与物体交互对于增强现实/虚拟现实（AR/VR）和机器人应用至关重要，但由于遮挡和受限的捕捉体积，传统的基于视觉的方法仍然面临挑战。惯性测量单元（IMUs）提供了一种不依赖视线的有力替代方案，但现有的基于IMU的运动捕捉假设人类是孤立的，忽视了物体接触和动态。为了解决这一问题，我们提出了IMU-HOI，这是一种新颖的框架，能够从身体和物体上的稀疏IMU中联合恢复全身人体姿态和6自由度（6-DoF）物体轨迹，明确建模人体-物体交互。我们的方法首先直接从IMU数据流中推断出概率性的手-物体接触，并将其作为高层信号在运动学推理和惯性推理之间进行路由。这些接触线索驱动一个三阶段融合管道，精炼人体姿态和根部平移，并将基于手的前向运动学与物体-IMU集成融合，以获取物体运动，从而为人体和物体提供一致且抗漂移的轨迹。在具有挑战性的人体-物体交互场景中的实验表明，相较于先前的惯性运动捕捉方法，准确性有显著提升。此外，IMU-HOI可以在现有的稀疏IMU运动捕捉骨干中进行最小改动地集成，有效地将纯惯性运动捕捉的范围从孤立的人体扩展到完整的人体-物体交互和联合运动估计。
cs.CV / 33 / 2606.28622
Meshtryoshka: Differentiable Rendering of Real-World Scenes via Mesh Rasterization
Meshtryoshka：通过网格光栅化实现真实场景的可微渲染
Abstract
Differentiable rendering has emerged as a powerful approach for 3D reconstruction and novel view synthesis. State-of-the-art differentiable rendering methods combine a variety of custom representations of 3D geometry and appearance with specialized renderers. However, most downstream tasks in computer graphics rely on 3D meshes. While prior work has attempted differentiable rendering with mesh representations, these approaches are limited to object-centric scenes and fail to reconstruct large-scale, unbounded scenes. In this work, we introduce Meshtryoshka, a novel mesh differentiable rendering framework that combines an off-the-shelf triangle rasterizer with a 3D representation that consists of nested mesh shells which resemble a matryoshka doll. In every forward pass, the mesh shells are extracted anew from a 3D signed distance function via iso-surface extraction, and the opacities for each vertex are computed as a function of signed distance. Each mesh shell is then rasterized independently, and the final image is created via alpha compositing. Crucially, mesh vertex positions are updated only indirectly via gradients that flow through the opacity values into the signed distance function, and hence, our method is compatible with off-the-shelf mesh renderers that need not be differentiable with respect to vertex positions. On object-centric scenes, our method performs competitively with surface-based differentiable rendering techniques. Our differentiable mesh rendering method scales to unbounded, real-world 3D scenes, where it yields high-quality novel view synthesis results approaching those of state-of-the-art, non-mesh methods. Our method suggests that it may be possible to solve the differentiable rendering problem without relying on specialized renderers, only using conventional tools from the computer graphics toolbox.
Chinese Translation
可微渲染已成为三维重建和新视图合成的强大方法。最先进的可微渲染方法结合了多种自定义的三维几何和外观表示与专用渲染器。然而，计算机图形学中的大多数下游任务依赖于三维网格。尽管之前的工作尝试使用网格表示进行可微渲染，但这些方法仅限于以物体为中心的场景，无法重建大规模、无界的场景。在本研究中，我们提出了Meshtryoshka，一种新颖的网格可微渲染框架，它结合了现成的三角形光栅化器和由嵌套网格壳组成的三维表示，这些网格壳类似于俄罗斯套娃。在每次前向传播中，网格壳通过等值面提取从三维有符号距离函数中重新提取，并且每个顶点的透明度作为有符号距离的函数进行计算。然后，每个网格壳独立进行光栅化，最终图像通过α合成生成。重要的是，网格顶点位置仅通过流经透明度值到达有符号距离函数的梯度间接更新，因此，我们的方法与不需要对顶点位置进行可微分的现成网格渲染器兼容。在以物体为中心的场景中，我们的方法与基于表面的可微渲染技术表现出竞争力。我们的可微网格渲染方法扩展到无界的真实世界三维场景，在这些场景中，它产生的高质量新视图合成结果接近于最先进的非网格方法。我们的方法表明，可能无需依赖专用渲染器，仅使用计算机图形工具箱中的常规工具即可解决可微渲染问题。
cs.CV / 34 / 2606.28626
SIGNET: Motion-Level Knowledge Transfer for Cross-Language Sign Language Translation
SIGNET：用于跨语言手语翻译的运动级知识迁移
Abstract
Sign language translation (SLT) remains challenging due to its high spatio-temporal complexity, long sequences, and the need to model multiple articulators without relying on gloss annotations. Existing approaches are typically tailored to individual datasets or languages and struggle to scale, while overlooking the relationships between sign languages that could inform more effective cross-lingual transfer. We present \textbf{SIGNET}, a framework that enables motion-level knowledge transfer for cross-language sign language translation. Our key insight is that, although sign languages differ in grammar and lexicon, pretrained models capture motion-level visual patterns that can be reused across datasets and languages. \textbf{SIGNET} integrates multiple pretrained sign language backbones through an attention-based, hand-prior aggregation mechanism that guides a gated fusion network in dynamically selecting the most relevant experts. Comprehensive experiments on four benchmarks (How2Sign, Phoenix14T, CSL-Daily, and MeineDGS) demonstrate state-of-the-art translation performance, and \textbf{SIGNET} also surpasses prior methods on WLASL for sign language recognition.
Chinese Translation
手语翻译（SLT）因其高度的时空复杂性、长序列以及需要在不依赖于注释的情况下建模多个发音器而仍然面临挑战。现有的方法通常针对单个数据集或语言进行定制，难以扩展，同时忽视了手语之间的关系，这些关系可以为更有效的跨语言迁移提供信息。我们提出了 extbf{SIGNET}，一个用于跨语言手语翻译的运动级知识迁移框架。我们的关键见解是，尽管手语在语法和词汇上存在差异，但预训练模型捕捉到的运动级视觉模式可以在不同的数据集和语言之间重复使用。 extbf{SIGNET}通过基于注意力的手部优先聚合机制集成多个预训练的手语骨干网络，指导一个门控融合网络动态选择最相关的专家。在四个基准（How2Sign、Phoenix14T、CSL-Daily 和 MeineDGS）上的全面实验表明了其最先进的翻译性能， extbf{SIGNET}在手语识别的WLASL上也超越了之前的方法。
cs.CV / 35 / 2606.28630
Physics-Grounded Disentangled Flow Modeling for Brain Disease Progression Trajectory
基于物理的解耦流建模用于脑疾病进展轨迹
Abstract
Forecasting longitudinal brain lesion evolution is critical for disease monitoring and treatment planning. Existing approaches typically learn a direct mapping from a baseline image to a future observation, without explicitly modeling the physical mechanisms underlying the lesion progression. Such an entangled modeling of structural deformation and image intensity variation limits physical plausibility, model generalization, and interpretability. To address this, we propose PDF, a Physics-grounded Disentangled Flow matching framework for longitudinal brain disease forecasting. We explicitly decompose the longitudinal modeling of lesion growth into two processes, each learned by a dedicated flow matching network: morphology evolution, which captures lesion growth and structural deformation; and intensity evolution, which models signal changes driven by variations in lesion concentration. To enforce physics-grounded constraints, we introduce a PDE-regularized loss based on lesion growth dynamics, that enforces a diffusion-reaction-advection formulation for morphological evolution. Experiments on three public longitudinal datasets spanning diverse brain diseases demonstrate state-of-the-art performance, validating the effectiveness of the disentangled modeling framework and physics-grounded learning design. Code is publicly available at https://github.com/jhuldr/PDF.
Chinese Translation
预测纵向脑损伤演变对疾病监测和治疗规划至关重要。现有方法通常直接学习从基线图像到未来观测的映射，而未明确建模损伤进展背后的物理机制。这种结构变形与图像强度变化的纠缠建模限制了物理合理性、模型泛化能力和可解释性。为了解决这一问题，我们提出了PDF（Physics-grounded Disentangled Flow）匹配框架，用于纵向脑疾病预测。我们明确将损伤生长的纵向建模分解为两个过程，每个过程由专门的流匹配网络学习：形态演变，捕捉损伤生长和结构变形；以及强度演变，建模由损伤浓度变化驱动的信号变化。为了强制执行基于物理的约束，我们引入了一种基于损伤生长动态的PDE正则化损失，强制形态演变遵循扩散-反应-对流公式。在三个涵盖多种脑疾病的公共纵向数据集上的实验表明，该方法表现出最先进的性能，验证了解耦建模框架和基于物理学习设计的有效性。代码可在 https://github.com/jhuldr/PDF 获取。
cs.CV / 36 / 2606.28635
AEGIR: Modeling Area Emitters for Indoor Inverse Rendering using Gaussian Splatting
AEGIR：基于高斯点云建模室内区域发射体的逆向渲染
Abstract
Inverse rendering requires separating illumination from surface materials, which is highly ambiguous due to their tight coupling in observed images. While Gaussian Splatting is efficient for novel view synthesis, existing relightable methods approximate scene lighting using discrete point lights, global environment maps, or implicit representations. By ignoring the physical spatial extent of real-world emitters, these approaches produce incorrect light attenuation and unrealistic shadows. We present AEGIR (Area Emitters for Gaussian Inverse Rendering), a framework that explicitly models local area emitters within a relightable Gaussian Splatting representation. Joint optimization of emitters, materials, and geometry is challenging due to flexible emitter parameterization, which increases both the number of parameters and the ambiguity between illumination and materials. We address this by introducing a differentiable deferred rendering pipeline that integrates multiple importance sampling with targeted regularization. As a result, AEGIR accurately simulates local light transport and achieves more consistent decomposition. Experiments show that explicit area emitters improve illumination reconstruction and enhance downstream tasks, including novel view synthesis, controlled relighting, and virtual object insertion, particularly in scenes with complex local lighting.
Chinese Translation
逆向渲染需要将照明与表面材料分离，这在观察图像中由于它们的紧密耦合而高度模糊。虽然高斯点云在新视角合成中效率较高，但现有的可重光照方法使用离散点光源、全局环境光图或隐式表示来近似场景照明。由于忽略了现实世界发射体的物理空间范围，这些方法产生了不正确的光衰减和不现实的阴影。我们提出了AEGIR（区域发射体用于高斯逆向渲染），这是一个在可重光照的高斯点云表示中显式建模局部区域发射体的框架。由于发射体参数化的灵活性，发射体、材料和几何体的联合优化具有挑战性，这增加了参数数量并加大了照明与材料之间的模糊性。我们通过引入一个可微分的延迟渲染管道来解决这一问题，该管道将多重重要性采样与针对性正则化相结合。因此，AEGIR能够准确模拟局部光传输，并实现更一致的分解。实验表明，显式区域发射体改善了照明重建，并增强了下游任务，包括新视角合成、控制重光照和虚拟物体插入，特别是在具有复杂局部照明的场景中。
cs.CV / 37 / 2606.28643
Obliviate: Erasing Concepts from Autoregressive Image Generation Models
Obliviate：从自回归图像生成模型中消除概念
Abstract
The widespread adoption of generative AI models has intensified concerns about misuse, including the creation of unsafe or disturbing imagery. To mitigate such issues, several concept erasure approaches have been proposed to remove harmful content from multimodal generative models. Yet concept erasure for autoregressive image generation remains largely unexplored, despite the growing relevance of these models in recent trends toward unified multimodal architectures. In this work, we fill this gap by introducing Obliviate, a guidance-based concept erasure method for autoregressive image generation. Our method builds on three key design choices: KL-based supervision over visual token distributions, trajectory-level updates over full autoregressive rollouts, and aligned visual prefixes for stable target construction. We evaluate Obliviate on three state-of-the-art autoregressive text-to-image models, Liquid, Emu3-Gen, and Janus-Pro, covering the erasure of explicit content, graphic violence, and branded imagery. Obliviate consistently outperforms current alternatives, reducing nudity on the defensive RAB benchmark from 91.58 to 3.15 while preserving overall model utility.
Chinese Translation
生成性人工智能模型的广泛应用加剧了对其滥用的担忧，包括创建不安全或令人不安的图像。为了解决这些问题，已经提出了几种概念消除方法，以去除多模态生成模型中的有害内容。然而，尽管这些模型在最近统一多模态架构的趋势中日益重要，自回归图像生成的概念消除仍然基本未被探索。在本研究中，我们通过引入Obliviate填补了这一空白，Obliviate是一种基于引导的自回归图像生成概念消除方法。我们的方法基于三个关键设计选择：对视觉标记分布的KL（Kullback-Leibler）基监督、对完整自回归展开的轨迹级更新，以及用于稳定目标构建的对齐视觉前缀。我们在三种最先进的自回归文本到图像模型Liquid、Emu3-Gen和Janus-Pro上评估了Obliviate，涵盖了对显性内容、图形暴力和品牌图像的消除。Obliviate在性能上始终优于当前的替代方案，将防御性RAB基准上的裸体内容从91.58降低到3.15，同时保持整体模型的效用。
cs.CV / 38 / 2606.28654
FedLAS: Feature-Modulated Bidirectional Label Smoothing for Neural Network Calibration
FedLAS：用于神经网络校准的特征调制双向标签平滑
Abstract
Deep Neural Network (DNN) classifiers suffer from poor calibration when their softmax outputs (predictive confidence) deviate from the empirical likelihoods. This manifests itself as either overconfident incorrect predictions or under-confident correct predictions. Label smoothing (LS) enhances model calibration by introducing entropy regularization during training through redistributing probability mass from the ground-truth label to the remaining classes. LS, including Margin-based LS (MbLS), have restrictive assumptions: they rely on predefined, uniform smoothing rules and only tackle overconfidence. In reality, samples exhibit diverse characteristics, such as difficulty/ambiguity, that interact with the evolving nature of the model being trained. In training, samples may have various degrees of under- or overconfidence. To overcome this, a mechanism that identifies the specific confidence state of each sample and determines the appropriate degree of smoothing in each training step is needed, tailoring the adjustment to the individual sample. We propose FedLAS: Feature-Modulated Bidirectional Label Smoothing, a plug-and-play algorithm for label smoothing-based losses. In FedLAS, we introduce a Feature Norm-based Confidence Indicator (NCI) to control smoothing and a Bidirectional Calibration Gating (BCG) module to detect both over and under-confidence. Our algorithm can be integrated with LS and MbLS based losses when applied to standard DNNs, enhancing performance. Extensive experiments on standard and fine-grained high-resolution vision benchmarks show that FedLAS consistently improves calibration compared to modern baselines, reducing Expected Calibration Error (ECE) and Adaptive ECE while maintaining Top-1 accuracy. Code: github.com/nadarasarbahavan/FEDLAS
Chinese Translation
深度神经网络（DNN）分类器在其 softmax 输出（预测置信度）偏离经验似然时，往往表现出较差的校准。这种情况表现为过于自信的错误预测或缺乏自信的正确预测。标签平滑（LS）通过在训练过程中引入熵正则化，将概率质量从真实标签重新分配到其他类别，从而增强模型的校准。包括基于边际的标签平滑（MbLS）在内的 LS 具有限制性假设：它们依赖于预定义的均匀平滑规则，并仅处理过度自信的问题。实际上，样本表现出多样的特征，例如难度/模糊性，这些特征与正在训练的模型的不断演变相互作用。在训练过程中，样本可能表现出不同程度的过度自信或缺乏自信。为了解决这个问题，需要一种机制来识别每个样本的具体置信状态，并在每个训练步骤中确定适当的平滑程度，从而针对个别样本进行调整。我们提出了 FedLAS：特征调制双向标签平滑，这是一种基于标签平滑损失的即插即用算法。在 FedLAS 中，我们引入了一种基于特征范数的置信指标（NCI）来控制平滑，并使用双向校准门控（BCG）模块来检测过度自信和缺乏自信。我们的算法可以与 LS 和 MbLS 基于的损失集成应用于标准 DNN，从而提升性能。在标准和细粒度高分辨率视觉基准上的广泛实验表明，FedLAS 在校准方面始终优于现代基线，降低了期望校准误差（ECE）和自适应 ECE，同时保持了 Top-1 准确率。代码：github.com/nadarasarbahavan/FEDLAS
cs.CV / 39 / 2606.28656
SemDynReg: Semantics-Guided Deformation Regularization for Dynamic 3D Gaussian Splatting
SemDynReg：基于语义引导的动态3D高斯点云变形正则化
Abstract
Deformable 3D Gaussian Splatting (3DGS) has emerged as an efficient approach for rendering dynamic scenes in a wide range of 3D applications. However, existing deformation field-based approaches largely lack explicit object-level modeling, often resulting in inconsistent Gaussian deformations within individual objects and unwanted coupling between different objects. To address this limitation, we introduce a semantics-guided framework that enforces dynamic regularization at the object level, aiming to achieve spatially consistent object-wise deformation. Specifically, we first extract segmentation masks using the Segment Anything Model (SAM) and derive semantic features from input images. An object-ID map is then constructed via feature relevance matching with a predefined object dictionary. Guided by this object-ID map, we identify the pixel-wise top-k contributing Gaussians for each object and impose consistency regularization on their deformation parameters, including position, scale, and rotation. Unlike prior methods that learn deformation fields without explicit object-level constraints, our approach incorporates semantic cues to guide deformation behavior at the object level. Experimental results demonstrate that our semantics-aware regularization improves object-level deformation consistency and outperforms baseline methods in rendering quality, achieving higher PSNR and SSIM and lower LPIPS in dynamic 3DGS rendering. Our project page is available at https://dyn-reg-3dgs.github.io/.
Chinese Translation
可变形的3D高斯点云（3DGS）已成为在广泛的3D应用中渲染动态场景的高效方法。然而，现有的基于变形场的方法在对象级建模方面普遍缺乏明确性，常常导致单个对象内的高斯变形不一致以及不同对象之间的不必要耦合。为了解决这一局限性，我们提出了一种基于语义引导的框架，在对象级别上强制实施动态正则化，旨在实现空间上一致的对象变形。具体而言，我们首先使用Segment Anything Model（SAM）提取分割掩膜，并从输入图像中导出语义特征。然后，通过与预定义对象字典的特征相关性匹配构建对象ID映射。在该对象ID映射的指导下，我们识别每个对象的像素级前k个贡献高斯，并对其变形参数（包括位置、尺度和旋转）施加一致性正则化。与之前在没有明确对象级约束的情况下学习变形场的方法不同，我们的方法结合了语义线索，以指导对象级的变形行为。实验结果表明，我们的语义感知正则化提高了对象级变形的一致性，并在渲染质量上超越了基线方法，实现了动态3DGS渲染中更高的PSNR和SSIM以及更低的LPIPS。我们的项目页面可访问 https://dyn-reg-3dgs.github.io/。
cs.CV / 40 / 2606.28673
BackTranslation2.0 -- A Linguistically Motivated Metric to Assess Sign Language Production
BackTranslation2.0 -- 一种基于语言学的评估指标用于评估手语生成
Abstract
Sign Languages (SLs) are the primary means of communication for millions of deaf individuals, yet existing evaluation metrics for generated SL remain simplistic and poorly aligned with human judgements. We introduce BackTranslation2.0, a linguistically grounded evaluation metric for text-to-sign translation that moves beyond na\"ive backtranslation. Our approach adopts an agentic framework in which a deterministic pipeline orchestrates a suite of specialised tools to assess four scoring dimensions - grammatical correctness, phonological accuracy, motion fluency, and generation fidelity - aligned with human rater assessments. Tool outputs are not treated independently: a set of large language model (LLM)-based cross-referential comparison modules evaluates consistency across tools and checks outputs against linguistic expectations, enabling structured reasoning over grammatical, phonological, and motion-level evidence. Final dimension scores are computed through deterministic weighted formulas over validated tool outputs. To validate BackTranslation2.0, we introduce and evaluate on a British Sign Language (BSL) dataset rated in a human rater study across the same quality dimensions, following a protocol developed in collaboration between linguists and deaf experts, benchmarking against six baseline metrics. Our method demonstrates strong correlation with human judgements across all dimensions, providing a more comprehensive, interpretable, and linguistically principled evaluation framework for sign language production systems.
Chinese Translation
手语（SL）是数百万聋人沟通的主要方式，但现有的手语生成评估指标仍然过于简单，且与人类评判的对齐程度较差。我们提出了BackTranslation2.0，这是一种基于语言学的文本到手语翻译评估指标，超越了简单的反向翻译。我们的方法采用了一种代理框架，其中一个确定性的流程协调一系列专业工具，以评估四个评分维度——语法正确性、音韵准确性、动作流畅性和生成保真度，这些维度与人类评分者的评估相一致。工具输出并非独立处理：一组基于大型语言模型（LLM）的交叉参考比较模块评估工具之间的一致性，并根据语言学期望检查输出，从而在语法、音韵和动作层面的证据上进行结构化推理。最终维度得分通过对经过验证的工具输出应用确定性加权公式计算得出。为了验证BackTranslation2.0，我们引入并在一个英国手语（BSL）数据集上进行评估，该数据集在一项人类评分者研究中根据相同的质量维度进行评分，遵循在语言学家与聋人专家合作下开发的协议，并与六个基准指标进行比较。我们的方法在所有维度上与人类评判表现出强相关性，为手语生成系统提供了一个更全面、可解释且基于语言学原则的评估框架。
cs.CV / 41 / 2606.28676
Predicting Metastatic Risk from Primary Tissue Architecture via Distance-Aware Spatial Modeling
通过距离感知空间建模预测原发组织结构的转移风险
Abstract
Predicting the risk of distant metastasis from primary tumor tissue histology is a critical yet challenging task in computational pathology. Multiple Instance Learning (MIL) approaches can attend to subdomains in tumor regions that harbor features of metastatic cancer progression. However MIL models treat tissue patches as unordered bags, discarding the spatial layout that defines the metastatic potential. We propose that metastatic risk is inherently dictated by the geometric arrangement of the tumor microenvironment at the interface with tumor cells. Our model is designed to explicitly capture the spatial relationships between tumor cells, tumor associated fibroblasts and infiltrating lymphocytes. For this purpose, we propose Distance aware Tissue Modeling for Multiple Instance Learning(DTMf-MIL), a novel method that reinforces visual features with explicit spatial priors. By computing signed distance functions (SDF) relative to tissue phenotypes, our model learns to recognize structural signatures of metastatic risk. This geometric awareness translates directly to superior clinical performance as DTMf-MIL significantly outperforms state-of-the-art methods that ignore spatial layout on metastasis prediction from tissue in the primary tumor. We further validate our approach on public benchmarks, demonstrating that spatial awareness consistently improves diagnostic accuracy across diverse clinical tasks.
Chinese Translation
从原发肿瘤组织组织学中预测远处转移的风险是计算病理学中一项关键但具有挑战性的任务。多实例学习（Multiple Instance Learning, MIL）方法可以关注肿瘤区域中具有转移性癌症进展特征的子域。然而，MIL模型将组织块视为无序的集合，忽略了定义转移潜力的空间布局。我们提出转移风险本质上由肿瘤微环境与肿瘤细胞界面的几何排列决定。我们的模型旨在明确捕捉肿瘤细胞、肿瘤相关成纤维细胞和浸润淋巴细胞之间的空间关系。为此，我们提出了一种距离感知组织建模的多实例学习方法（Distance aware Tissue Modeling for Multiple Instance Learning, DTMf-MIL），这是一种通过显式空间先验增强视觉特征的新方法。通过计算相对于组织表型的有符号距离函数（Signed Distance Functions, SDF），我们的模型学习识别转移风险的结构特征。这种几何意识直接转化为卓越的临床表现，因为DTMf-MIL在从原发肿瘤组织中预测转移方面显著优于忽略空间布局的最先进方法。我们进一步在公共基准上验证了我们的方法，证明空间意识在多种临床任务中始终提高了诊断准确性。
cs.CV / 42 / 2606.28677
SATB-VR: Training Few-Step Video Restoration Diffusion Model using SNR-Aware Trajectory Blending
SATB-VR：使用SNR感知轨迹混合训练少步视频恢复扩散模型
Abstract
While diffusion models excel in video restoration, their reliance on extensive iterative steps limits efficiency. Conversely, aggressive single-step distillation often compromises fine texture recovery. To achieve an optimal balance, we present SATB-VR, a few-step paradigm that jump-starts the denoising process via an auxiliary predictor, explicitly bypassing early low signal-to-noise ratio (SNR) steps. However, naive joint training of the predictor and the denoiser inherently introduces a severe train-inference discrepancy. To resolve this, we propose the SNR-Aware Trajectory Blending (SATB) strategy. During the forward process, SATB constructs the noisy input by dynamically blending the predictor's output with the ground-truth trajectory based on the SNRs. This forces the denoiser to robustly compensate for initial prediction errors while smoothly converging to the clean data manifold. Furthermore, we introduce a Denoiser-Driven Consistency (DDC) loss, leveraging the concurrently updated denoiser as a dynamic evaluator to explicitly align internal features and boost predictor accuracy. Extensive experiments demonstrate that, under flexible few-step inference regimes (\eg, $\le 5$ steps), SATB-VR performs favorably against existing approaches on synthetic, real-world, and AIGC benchmarks.
Chinese Translation
尽管扩散模型在视频恢复中表现出色，但其对大量迭代步骤的依赖限制了效率。相反，激进的单步蒸馏往往会妨碍细致纹理的恢复。为了实现最佳平衡，我们提出了SATB-VR，一种通过辅助预测器启动去噪过程的少步范式，明确绕过早期低信噪比（SNR）步骤。然而，预测器和去噪器的简单联合训练本质上引入了严重的训练-推理不一致性。为了解决这个问题，我们提出了SNR感知轨迹混合（SATB）策略。在前向过程中，SATB通过根据SNR动态混合预测器的输出与真实轨迹来构建噪声输入。这迫使去噪器稳健地补偿初始预测误差，同时平滑地收敛到干净数据流形。此外，我们引入了一种去噪器驱动一致性（DDC）损失，利用同时更新的去噪器作为动态评估器，明确对齐内部特征并提升预测器的准确性。大量实验表明，在灵活的少步推理模式下（例如，$ ext{≤} 5$ 步），SATB-VR在合成、真实世界和AIGC基准测试中相较于现有方法表现优越。
cs.CV / 43 / 2606.28688
LogiCo: A Unified Framework for Logical and Structural Anomaly Detection
LogiCo：一个统一的逻辑与结构异常检测框架
Abstract
Current anomaly detection methods primarily focus on structural anomalies, while paying insufficient attention to anomalies that violate logical constraints. Conversely, top-performing logical anomaly detection approaches address this by modeling global semantic consistency, but perform poorly on subtle structural anomalies due to inadequate detection granularity. In this paper, we propose LogiCo, a unified framework for Logical and structural anomaly detection via Component-level feature reconstruction. Unlike existing methods that rely on explicit global semantic modeling, LogiCo employs a novel component-level feature reconstruction technique to capture inter-component logical constraints. Specifically, LogiCo maps pre-trained image features into a discrete component-level feature space and performs collaborative feature reconstruction at both component and patch levels, enabling it to effectively detect both logical and structural anomalies. Furthermore, to address the specific challenge of count-related logical anomalies, we integrate a segmentation-map discriminator that extends the model's capability to identify quantitative inconsistencies. LogiCo achieves state-of-the-art performance on both logical and structural anomaly detection across four benchmarks, including MVTec-LOCO, MVTec-AD, VisA, and Real-IAD, demonstrating its superiority and practical feasibility. The code is available at https://github.com/cnulab/LogiCo.
Chinese Translation
当前的异常检测方法主要集中于结构异常，而对违反逻辑约束的异常关注不足。相反，表现优异的逻辑异常检测方法通过建模全局语义一致性来解决这一问题，但由于检测粒度不足，在微妙的结构异常上表现不佳。本文提出了LogiCo，一个通过组件级特征重建进行逻辑与结构异常检测的统一框架。与依赖于显式全局语义建模的现有方法不同，LogiCo采用了一种新颖的组件级特征重建技术，以捕捉组件间的逻辑约束。具体而言，LogiCo将预训练的图像特征映射到离散的组件级特征空间，并在组件和补丁级别进行协同特征重建，从而有效检测逻辑和结构异常。此外，为了应对与计数相关的逻辑异常这一特定挑战，我们集成了一个分割图判别器，扩展了模型识别定量不一致性的能力。LogiCo在四个基准数据集（包括MVTec-LOCO、MVTec-AD、VisA和Real-IAD）上实现了逻辑和结构异常检测的最先进性能，展示了其优越性和实际可行性。代码可在 https://github.com/cnulab/LogiCo 获取。
cs.CV / 44 / 2606.28697
Mitigating Batch Effects in Histopathology via Language-Mediated Robust Embedding Generation
通过语言介导的稳健嵌入生成减轻组织病理学中的批次效应
Abstract
Pathology foundation models (PFMs) have demonstrated strong potential across clinical and scientific applications, yet their performance is often hindered by batch effects, which are non-biological variations across tissue source institutions (TSIs) that distort learned feature representations and impair generalization. Conventional mitigation strategies, such as stain normalization, offer limited success in addressing these high-dimensional, complex artifacts. We present GLMP (General-purpose LLM-Mediated Pathology model), a novel framework that generates robust numerical embeddings from histology image patches through an intermediate textual representation. By leveraging pretrained general-purpose multimodal large language models (MLLMs) and text encoders, GLMP effectively prioritizes biologically meaningful signals over TSI-specific artifacts, thereby improving cross-institutional generalization. To our knowledge, GLMP is the first pathology model to use text descriptions of histological features as an intermediate representation for generating numerical embeddings from histology images. Our results highlight the untapped potential of broad-domain, non-specialized MLLMs in computational pathology and introduce a new paradigm for building versatile, generalizable, and robust pathology models.
Chinese Translation
病理基础模型（PFMs）在临床和科学应用中展现出强大的潜力，但其性能常常受到批次效应的影响，批次效应是指来自组织来源机构（TSIs）之间的非生物学变异，这些变异扭曲了学习到的特征表示并损害了模型的泛化能力。传统的减轻策略，如染色标准化，在解决这些高维复杂伪影方面效果有限。我们提出了GLMP（通用LLM介导的病理模型），这是一个新颖的框架，通过中间文本表示从组织学图像块生成稳健的数值嵌入。通过利用预训练的通用多模态大型语言模型（MLLMs）和文本编码器，GLMP有效地优先考虑生物学上有意义的信号，而非TSI特定的伪影，从而改善跨机构的泛化能力。据我们所知，GLMP是第一个使用组织学特征文本描述作为中间表示来生成组织学图像数值嵌入的病理模型。我们的结果突显了广域非专业MLLMs在计算病理学中的未开发潜力，并引入了一种构建多功能、可泛化和稳健病理模型的新范式。
cs.CV / 45 / 2606.28724
CCRC: A Change-Aware Captioning and Reasoning Chain for Image Change Captioning and Segmentation
CCRC：一种基于变化感知的图像变化描述与推理链用于图像变化描述与分割
Abstract
Understanding and localizing subtle changes between paired images is critical for tasks such as surveillance and image editing. However, traditional Image Change Captioning (ICC) methods lack spatial grounding, limiting their precision. We introduce Image Change Captioning and Segmentation (ICCS), a new multimodal task that jointly requires structured change description and pixel-level localization. To address ICCS, we propose the Change-aware Captioning and Reasoning Chain (CCRC), a dual-chain framework that decouples semantic reasoning from spatial segmentation. The first chain, Chain-of-Change-Captioning (CCC), enhances fine-grained change perception via a visual fusion module based on Multi-Head Change-aware Attention inserted between the visual and language components of a Multimodal Large Language Model (MLLM). CCC also determines whether a change is segmentable. If not, it alone generates the caption. Otherwise, the second chain, Chain-of-Change-Segmenting (CCS), is activated, leveraging spatial priors from CCC and refining masks with a Change-aware Token Refiner for accurate boundary localization. We evaluate CCRC on both synthetic and real-world change detection benchmarks with pixel-level supervision. Experiments show CCRC achieves state-of-the-art performance.
Chinese Translation
理解和定位成对图像之间的细微变化对于监控和图像编辑等任务至关重要。然而，传统的图像变化描述（ICC）方法缺乏空间基础，限制了其精确性。我们提出了图像变化描述与分割（ICCS），这是一项新的多模态任务，要求同时进行结构化变化描述和像素级定位。为了解决ICCS问题，我们提出了变化感知描述与推理链（CCRC），这是一种双链框架，将语义推理与空间分割解耦。第一条链，变化描述链（CCC），通过一个基于多头变化感知注意力的视觉融合模块增强细粒度变化感知，该模块插入在多模态大型语言模型（MLLM）的视觉和语言组件之间。CCC还确定变化是否可分割。如果不可分割，它将单独生成描述。否则，第二条链，变化分割链（CCS），被激活，利用CCC的空间先验并通过变化感知令牌细化器精炼掩膜，以实现准确的边界定位。我们在合成和真实世界的变化检测基准上对CCRC进行了评估，并进行了像素级监督。实验表明，CCRC达到了最先进的性能。
cs.CV / 46 / 2606.28745
FreqOrtho-SR: Frequency-Guided Orthogonal Expert Learning for Real-World Image Super-Resolution
FreqOrtho-SR：基于频率引导的正交专家学习用于真实世界图像超分辨率
Abstract
Diffusion prior-based methods have shown impressive results in real-world image super-resolution (ISR), yet two key challenges persist: balancing pixel-level fidelity with semantic quality, and adapting to diverse degradations. Existing dual-branch approaches freeze the pixel module during semantic training, but the semantic branch can still expand capacity within the pixel subspace, precluding genuine perceptual improvement. Moreover, using a single static adapter cannot generalize across heterogeneous real-world corruptions. To address both issues, we propose FreqOrtho-SR, which comprises: $\textbf{Freq}$uency-guided Mixture of LoRA Experts (FreqMoE), it routes inputs to specialized experts via a non-parametric FFT-based degradation-feature extractor that encodes frequency-domain signatures, enabling stable and interpretable specialization across corruption types; and $\textbf{Ortho}$gonal Gradient Projection (OGP), which reframes the dual-objective optimization as a subspace-constrained problem: by extracting the pixel-fidelity subspace via SVD on combined expert weight deltas and projecting semantic gradients onto its null space, OGP guarantees orthogonality between the two objectives, enabling genuinely complementary learning without mutual interference. Experiments show that FreqOrtho-SR achieves competitive overall performance and a strong fidelity-perception trade-off across multiple benchmarks with efficient single-step inference. The source code of our method can be found at $\href{https://github.com/sonhm3029/FreqOrtho-SR}{\texttt{sonhm3029/FreqOrtho-SR}}$.
Chinese Translation
基于扩散先验的方法在真实世界图像超分辨率（ISR）中取得了令人印象深刻的结果，但仍然面临两个关键挑战：平衡像素级保真度与语义质量，以及适应多样化的退化现象。现有的双分支方法在语义训练期间冻结像素模块，但语义分支仍然可以在像素子空间中扩展能力，从而阻碍真正的感知改进。此外，使用单一静态适配器无法在异构真实世界损坏中进行泛化。为了解决这两个问题，我们提出了FreqOrtho-SR，其包括：$ extbf{Freq}$uency-guided Mixture of LoRA Experts (FreqMoE)，通过一个基于非参数FFT的退化特征提取器将输入路由到专业专家，该提取器编码频域特征，使得在不同退化类型之间实现稳定且可解释的专业化；以及$ extbf{Ortho}$gonal Gradient Projection (OGP)，它将双目标优化重新构造为一个子空间约束问题：通过对组合专家权重增量进行SVD提取像素保真度子空间，并将语义梯度投影到其零空间，OGP保证了两个目标之间的正交性，从而实现真正互补的学习而不互相干扰。实验表明，FreqOrtho-SR在多个基准测试中实现了具有竞争力的整体性能，并在保真度与感知之间达成了良好的平衡，且具有效率的单步推理。我们方法的源代码可以在$ exttt{https://github.com/sonhm3029/FreqOrtho-SR}$找到。
cs.CV / 47 / 2606.28757
A Physics-Grounded Benchmark for Multi-Agent Dynamics in World Models
基于物理的多智能体动态世界模型基准测试
Abstract
Generative world models hold immense promise as scalable simulators for autonomous systems, particularly for synthesizing rare but safety-critical multi-agent interactions, such as vehicle collisions. However, current evaluation paradigms index heavily on visual fidelity and semantic alignment, leaving a critical blind spot: they cannot reliably quantify whether generated dynamics actually obey the fundamental physical laws required for reliable simulation. Assessing this physical plausibility is inherently difficult due to a lack of physical metrics and the challenge of extracting metric-scale kinematics from uncalibrated video rollouts. To bridge this gap, we introduce CrashTwin, a physics-grounded evaluation framework designed to stress-test the physical trustworthiness of world models. CrashTwin couples a diverse dataset of multi-agent collision scenarios, comprising 25K controllable synthetic and 12K in-the-wild real-world collision sequences with a novel calibration-free reconstruction pipeline, enabling the recovery of 3D physical attributes directly from world model rollouts. We propose a diagnostic suite that systematically evaluates three dimensions: spatio-temporal consistency, momentum and kinetic energy conservation, and world-dynamics integrity. Extensive benchmarking of state-of-the-art models reveals a crucial insight: high perceptual quality frequently masks severe physical violations during complex interactions. By quantitatively exposing these failure modes, CrashTwin provides a vital diagnostic tool for developing physically grounded world models capable of reliable real-world simulation.
Chinese Translation
生成的世界模型作为自主系统的可扩展模拟器具有巨大的潜力，尤其是在合成稀有但对安全至关重要的多智能体交互（如车辆碰撞）方面。然而，目前的评估范式过于依赖视觉真实感和语义一致性，留下了一个关键的盲点：它们无法可靠地量化生成的动态是否实际遵循可靠模拟所需的基本物理定律。由于缺乏物理度量标准以及从未校准的视频回放中提取度量级运动学的挑战，评估这种物理合理性本质上是困难的。为了解决这一问题，我们引入了CrashTwin，一个基于物理的评估框架，旨在对世界模型的物理可信性进行压力测试。CrashTwin结合了一个多样化的多智能体碰撞场景数据集，包括25K可控的合成碰撞序列和12K真实世界的碰撞序列，以及一个新颖的无校准重建管道，使得能够直接从世界模型回放中恢复3D物理属性。我们提出了一套诊断工具，系统地评估三个维度：时空一致性、动量和动能守恒以及世界动态的完整性。对最先进模型的广泛基准测试揭示了一个重要的见解：高感知质量常常掩盖复杂交互中严重的物理违规行为。通过定量揭示这些失效模式，CrashTwin为开发能够可靠进行真实世界模拟的基于物理的世界模型提供了一个重要的诊断工具。
cs.CV / 48 / 2606.28758
X-Mind: Efficient Visual Chain-of-Thought via Predictive World Model for End-to-End Driving
X-Mind：通过预测世界模型实现高效的视觉思维链用于端到端驾驶
Abstract
Predicting future states is essential for autonomous agents, yet current Vision-Language-Action (VLA) models fundamentally lack this capability, relying instead on reactive perception-action mapping. While integrating Predictive World Models (PWMs) addresses this gap, existing approaches either incur prohibitive cascaded latency or act as shallow terminal tasks that fail to deeply embed forward-looking reasoning. To endow VLA models with this reasoning capability, we propose X-Mind. Rather than treating PWMs as an external auxiliary module, this framework internalizes them as the Visual Chain-of-Thought (Visual CoT). By enforcing a world rollout prior to action, the model is constrained to imagine future evolution first, yielding a driving policy that is robustly grounded in environmental dynamics and aware of the future consequences its actions will unfold. The challenge here is efficiency, and we tackle it on two fronts. First, we introduce a compact representation of visual thinking: an abstract sketch that fuses a Bird's-Eye-View (BEV) layout with abstract driving priors (e.g., navigation intents and traffic rules). Rather than rolling out dense future frames, the model reasons over this sketch as a mental canvas; aided by a Deep Compression Autoencoder (DC-AE), a 12-frame future rollout is reduced to merely 96 tokens, alleviating the long-context computational bottleneck. Second, to accelerate generation further, we propose a recurrent block diffusion scheme that unrolls the denoising steps across the layers of the large drive model, folding iterative refinement into the backbone's one forward pass. Trained and validated on large-scale real-world data, X-Mind achieves competitive end-to-end driving performance, which makes it a highly practical, low-latency solution that successfully deploys large-scale cognitive reasoning directly onto resource-constrained vehicle platforms.
Chinese Translation
预测未来状态对自主代理至关重要，但当前的视觉-语言-行动（VLA）模型在这一能力上存在根本性缺陷，依赖于反应式的感知-行动映射。虽然整合预测世界模型（PWM）可以填补这一空白，但现有方法要么导致高昂的级联延迟，要么作为浅层终端任务，未能深入嵌入前瞻性推理。为了赋予VLA模型这一推理能力，我们提出了X-Mind。该框架并不将PWM视为外部辅助模块，而是将其内化为视觉思维链（Visual Chain-of-Thought，Visual CoT）。通过在行动之前强制进行世界展开，模型被约束首先想象未来的发展，从而产生一个稳健地基于环境动态并意识到其行动将产生未来后果的驾驶策略。这里的挑战在于效率，我们从两个方面来解决。首先，我们引入了一种紧凑的视觉思维表示：一种将鸟瞰图（Bird's-Eye-View，BEV）布局与抽象驾驶先验（例如导航意图和交通规则）融合的抽象草图。模型不是展开密集的未来帧，而是将这一草图作为心理画布进行推理；借助深度压缩自编码器（Deep Compression Autoencoder，DC-AE），12帧的未来展开被简化为仅96个标记，缓解了长上下文计算瓶颈。其次，为了进一步加速生成，我们提出了一种递归块扩散方案，该方案在大型驾驶模型的层之间展开去噪步骤，将迭代细化折叠到主干的单次前向传递中。在大规模真实世界数据上进行训练和验证后，X-Mind实现了具有竞争力的端到端驾驶性能，使其成为一种高度实用、低延迟的解决方案，成功地将大规模认知推理直接部署到资源受限的车辆平台上。
cs.CV / 49 / 2606.28785
Stochastic Optimal Control Sampling for Diffusion Inverse Problems
扩散逆问题的随机最优控制采样
Abstract
Benefiting from the strong ability to capture data distributions, diffusion models have become powerful tools for solving image inverse problems. The key is to controllably steer the sampling trajectory toward the measurements while respecting the diffusion prior. In this work, we introduce Stochastic Optimal Control Sampling (SOCS), which models the denoising process as a dynamical system and injects control signals via SOC. Previous SOC-based approach addresses inverse problems by optimizing over the entire trajectory, which is computationally expensive. In contrast, we derive a closed-form control update and apply it at each sampling step, pulling the measurement-consistent clean prediction back onto the denoising flow. In SOCS, we can readily modulate the control strength to align with the diffusion model's native capabilities and thereby enhance perceptual quality. Our method is compatible with a variety of linear stochastic differential equation backbones. Extensive experiments across a broad spectrum of image inverse tasks demonstrate that SOCS achieves accurate measurement-aligned reconstructions with improved visual fidelity and stronger quantitative performance.
Chinese Translation
得益于强大的数据分布捕捉能力，扩散模型已成为解决图像逆问题的强大工具。关键在于可控地引导采样轨迹朝向测量值，同时尊重扩散先验。在本研究中，我们引入了随机最优控制采样（Stochastic Optimal Control Sampling, SOCS），将去噪过程建模为动态系统，并通过随机最优控制（SOC）注入控制信号。之前基于SOC的方法通过优化整个轨迹来解决逆问题，这在计算上是昂贵的。相比之下，我们推导出了一种闭式控制更新，并在每个采样步骤应用它，将与测量一致的干净预测拉回到去噪流中。在SOCS中，我们可以轻松调节控制强度，以与扩散模型的固有能力对齐，从而增强感知质量。我们的方法与多种线性随机微分方程框架兼容。广泛的实验涵盖了各种图像逆任务，证明SOCS能够实现准确的测量对齐重建，具有更好的视觉保真度和更强的定量性能。
cs.CV / 50 / 2606.28787
BREIT: A Framework for Brain Stroke Reconstruction using Multi-Frequency 3D EIT
BREIT：一种基于多频率三维电阻成像的脑卒中重建框架
Abstract
Multi-Frequency Electrical Impedance Tomography (MF-EIT) is a non-invasive, low-cost modality that reconstructs electrical property distributions from boundary voltages. For stroke imaging, progress in 3D deep-learning reconstruction is limited by the lack of large-scale datasets with paired ground-truth (GT) volumes and by non-standardized pipelines for data generation, simulation, and evaluation. We introduce BREIT, a modular framework for 3D MF-EIT stroke reconstruction providing: (i) a neuroimaging-to-EIT pipeline that converts CT/MRI into frequency-dependent GT admittivity volumes; (ii) a self-contained Python 3D Complete Electrode Model (CEM) forward solver for simulating MF-EIT voltages; and (iii) a 3D D-bar implementation supporting non-uniform electrode layouts. Building on BREIT, we propose dFNO-bar, which integrates Fourier Neural Operators into D-bar by learning a mapping from scattering data $t(\xi)$ to conductivity $\sigma(x){=}\Re\{\gamma\}$. We evaluate dFNO-bar against D-bar, Deep D-bar, and Gauss--Newton reconstructions on UCLH-matched synthetic data, and observe higher brain SSIM with comparable CC across noise settings. Code and data are publicly available at: https://github.com/djahiddj13/BREIT
Chinese Translation
多频率电阻成像（MF-EIT）是一种非侵入性、低成本的成像技术，能够从边界电压重建电属性分布。在脑卒中成像方面，3D深度学习重建的进展受到缺乏配对真实值（GT）体积的大规模数据集以及数据生成、模拟和评估的非标准化流程的限制。我们提出了BREIT，这是一个用于3D MF-EIT脑卒中重建的模块化框架，提供：（i）一个神经成像到EIT的流程，将CT/MRI转换为频率依赖的GT导纳体积；（ii）一个自包含的Python 3D完整电极模型（CEM）正向求解器，用于模拟MF-EIT电压；以及（iii）一个支持非均匀电极布局的3D D-bar实现。在BREIT的基础上，我们提出了dFNO-bar，它通过学习从散射数据$t(\xi)$到电导率$\sigma(x){=} ext{Re}\{ ext{γ}\}$的映射，将傅里叶神经算子集成到D-bar中。我们在UCLH匹配的合成数据上评估了dFNO-bar与D-bar、Deep D-bar和高斯-牛顿重建的性能，观察到在不同噪声设置下，脑部结构相似性指数（SSIM）更高，而相关系数（CC）相当。代码和数据可在以下网址公开获取：https://github.com/djahiddj13/BREIT
cs.CV / 51 / 2606.28792
Virtual Ring Try-On
虚拟戒指试戴
Abstract
This paper presents an innovative approach that enables the users to capture their hand and try the jewel ring on their hand. The user captures the image of the hand using the React Native base GUI of the mobile application and selects the ring that the user wants to try, and the output image will have the user's hand with the ring image. This approach is implemented using a combination of MediaPipe hand point detection and YOLO-V8 custom object detection. The hand image uploaded by the user first undergoes mediapipe hand point detection. It will give the hand points and a Region of Interest mask where the ring is going to be placed. Then the ring is passed through YOLO object detection, in which ring points are detected, and background is removed. After that, using vector algebra, the angular discrepancy between the finger's reference axis and the ring's principal axis is computed. Also, ring size is rescaled according to finger thickness, preserving the aspect ratio to maintain perceptual realism. Then the ring is placed on the hand image and the output image is generated and shown on the user screen.
Chinese Translation
本文提出了一种创新的方法，使用户能够捕捉手部图像并在手上试戴珠宝戒指。用户通过移动应用的基于 React Native 的图形用户界面捕捉手部图像，并选择想要试戴的戒指，输出图像将显示用户的手和戒指图像。该方法结合了 MediaPipe 手部关键点检测和 YOLO-V8 自定义物体检测。用户上传的手部图像首先经过 MediaPipe 手部关键点检测，生成手部关键点和戒指将要放置的兴趣区域掩码。然后，戒指通过 YOLO 物体检测进行处理，检测戒指的关键点并去除背景。接着，使用向量代数计算手指参考轴与戒指主轴之间的角度差异。同时，戒指的大小根据手指厚度进行重新缩放，保持纵横比以维持感知真实感。最后，戒指被放置在手部图像上，生成输出图像并显示在用户屏幕上。
cs.CV / 52 / 2606.28799
PSP: Harnessing Position and Shape Priors for Cross-Domain Few-Shot Medical Image Segmentation
PSP：利用位置和形状先验进行跨域少样本医学图像分割
Abstract
Few-Shot Medical Image Segmentation (FSMIS) offers a powerful solution to data scarcity but struggles to generalize across different imaging modalities. This performance collapse stems primarily from the drastic texture discrepancies between domains, which mislead models trained on source-specific intensity distributions. While existing methods attempt to align frequency or local texture features, they often fail to decouple semantic structure from domain-specific appearance. To address this, we identify a critical invariance: despite distinct imaging physics, the position and geometric shape of organs remain robustly consistent across modalities. Therefore, we propose a novel framework that harnesses Position and Shape Priors (PSP) for cross-domain FSMIS. Specifically, PSP first introduces a Position Coordinate Embedding (PCE) module to inject relative spatial coordinates for rapid organ localization. Subsequently, a Shape Prototype Modulation (SPM) module constructs domain-invariant structural prototypes via explicit shape priors, effectively filtering out texture noise. Furthermore, the Hybrid-Prototype Prediction (HPP) module adaptively calibrates the support prototype to the query feature distribution, mitigating feature misalignment. Extensive experiments on two public medical imaging datasets demonstrate that PSP significantly outperforms state-of-the-art methods.
Chinese Translation
少样本医学图像分割（FSMIS）为数据稀缺问题提供了强有力的解决方案，但在不同成像模态之间的泛化能力较弱。这种性能崩溃主要源于域之间显著的纹理差异，这误导了在特定源强度分布上训练的模型。尽管现有方法试图对齐频率或局部纹理特征，但它们往往未能将语义结构与特定域的外观解耦。为了解决这个问题，我们识别出一个关键的不变性：尽管成像物理特性不同，器官的位置和几何形状在各个模态之间仍然保持稳健的一致性。因此，我们提出了一种新颖的框架，利用位置和形状先验（PSP）进行跨域FSMIS。具体而言，PSP首先引入一个位置坐标嵌入（PCE）模块，以快速注入相对空间坐标进行器官定位。随后，形状原型调制（SPM）模块通过显式形状先验构建域不变的结构原型，有效过滤纹理噪声。此外，混合原型预测（HPP）模块自适应地将支持原型校准到查询特征分布，从而减轻特征不对齐的问题。在两个公共医学成像数据集上的大量实验表明，PSP显著优于最先进的方法。
cs.CV / 53 / 2606.28804
ViPSim: Collaborating Visual and Parameter Spaces for Consistent Long-Horizon Embodied World Models
ViPSim：协作视觉与参数空间以实现一致的长时间范围具身世界模型
Abstract
Embodied World Models (EWMs) have emerged as a scalable and risk-free paradigm for advancing embodied intelligence, enabling the safety-critical evaluation of Vision-Language-Action systems. However, their reliability as evaluation benchmarks and foundational simulators is often hindered by the representation gap between low-dimensional actions and high-dimensional video synthesis. This gap results in a lack of geometric correspondence, manifesting as accumulated trajectory drift and inconsistent robot-object interactions during long-horizon rollouts. To bridge this gap, we propose ViPSim, a framework that achieves consistent long-horizon generation through the synergistic collaboration of Visual and Parameter Spaces. We define the Visual Space as a domain of explicit spatial priors, integrating pixel-aligned projections of end-effector pose, camera perspectives, depth-informed scene geometry, and robotic morphological masks to provide dense structural grounding. Concurrently, the Parameter Space serves as a domain of numerical drivers, injecting raw action sequences and camera matrices to provide precise motion guidance. By unifying these two spaces, ViPSim ensures that the generated states are simultaneously anchored by geometric boundaries and steered by numerical commands. Extensive experiments demonstrate that ViPSim is backbone-agnostic and significantly enhances trajectory consistency. Notably, our approach exhibits emergent capabilities in generating complex interactions with deformable objects (e.g., cloth folding) and maintains robust performance in out-of-distribution and cross-embodiment scenarios, providing a high-fidelity foundation for the automated evaluation and predictive control of embodied agents.
Chinese Translation
具身世界模型（EWMs）作为一种可扩展且无风险的范式，已成为推进具身智能的重要工具，使得对视觉-语言-动作系统的安全关键评估成为可能。然而，它们作为评估基准和基础模拟器的可靠性常常受到低维动作与高维视频合成之间的表示差距的影响。该差距导致几何对应关系的缺失，表现为在长时间范围的展开过程中轨迹漂移的累积和机器人与物体交互的不一致性。为了解决这一问题，我们提出了ViPSim，一个通过视觉与参数空间的协同合作实现一致的长时间范围生成的框架。我们将视觉空间定义为显式空间先验的领域，整合了末端执行器姿态、相机视角、深度信息场景几何和机器人形态掩模的像素对齐投影，以提供密集的结构基础。同时，参数空间作为数值驱动的领域，注入原始动作序列和相机矩阵，以提供精确的运动指导。通过统一这两个空间，ViPSim确保生成的状态同时受到几何边界的约束和数值命令的引导。大量实验表明，ViPSim与骨干网络无关，并显著增强了轨迹一致性。值得注意的是，我们的方法在生成与可变形物体（例如，折叠布料）的复杂交互方面表现出新兴能力，并在分布外和跨具身场景中保持强大的性能，为具身代理的自动评估和预测控制提供了高保真基础。
cs.CV / 54 / 2606.28820
CoGS: Compositional Dynamic Human-Object Scenes Gaussian Splatting from Monocular Video
CoGS：基于单目视频的组合动态人-物场景高斯点云重建
Abstract
Reconstructing dynamic human--object interaction scenes from monocular video is difficult because the human, manipulated object, and background obey different motion models while sharing the same pixels. Existing dynamic radiance-field and Gaussian-splatting methods often entangle these components, causing object motion to leak into the human or static scene, and monocular human reconstruction remains underconstrained in regions that are rarely observed. We present CoGS, a compositional Gaussian-splatting framework for monocular human--object scene reconstruction. CoGS decomposes the video into three coordinated branches: an articulated human initialized from a complete canonical prior, a rigid object field driven by an estimated object trajectory, and a static scene field regularized by weak scene-only planar primitives when available. A six-stage optimization schedule first stabilizes the human and object independently, then fuses them with the scene under full-image supervision, visibility-aware human anchoring, object silhouette and motion constraints, and delayed scene regularization. This design keeps each component responsible for its own geometry and motion while allowing photometric evidence to correct the final composite. Experiments on HOSNeRF and NeuMan show that CoGS improves both human--object interaction reconstruction and in-the-wild human--scene rendering, achieving stronger fidelity and perceptual quality across full-frame and human-focused evaluations. Code will be released upon publication.
Chinese Translation
从单目视频重建动态人-物交互场景是困难的，因为人类、操控物体和背景遵循不同的运动模型，同时共享相同的像素。现有的动态辐射场和高斯点云方法通常将这些组件纠缠在一起，导致物体运动渗透到人类或静态场景中，而单目人类重建在很少观察到的区域仍然受到约束不足的影响。我们提出了CoGS，一个用于单目人-物场景重建的组合高斯点云框架。CoGS将视频分解为三个协调的分支：一个从完整的典型先验初始化的关节人类，一个由估计的物体轨迹驱动的刚性物体场，以及一个在可用时由弱场景平面原语正则化的静态场景场。六阶段优化计划首先独立稳定人类和物体，然后在全图监督、考虑可见性的人的锚定、物体轮廓和运动约束以及延迟场景正则化的条件下将它们与场景融合。该设计使每个组件对其自身的几何和运动负责，同时允许光度证据修正最终的合成结果。在HOSNeRF和NeuMan上的实验表明，CoGS在动态人-物交互重建和自然场景人类渲染方面均有所改善，在全帧和人类聚焦评估中实现了更强的保真度和感知质量。代码将在发表后发布。
cs.CV / 55 / 2606.28826
RefGlass-GS: A UAV-Enabled Fusion Framework for Photorealistic, Semantic and Interactive Digitization of Reflective Glass Facades via Gaussian Splatting
RefGlass-GS：一种基于无人机的融合框架，实现反射玻璃外立面的照片级真实感、语义和交互式数字化
Abstract
Existing digitization of buildings with reflective glass facades suffers from geometric reconstruction distortion, unrealistic view-dependent texture rendering, and difficulties in object-based semantic enhancement. Therefore, we propose RefGlass-GS, a fusion framework that enables end-to-end UAV-based photorealistic, semantic, and interactive digitization of reflective glass facades. The contributions include: (1) proposing an individual glass panel segmentation method based on maximum a posteriori estimation with structural regularities, robust to severe reflection and background interference; (2) formulating a UAV viewpoint planning optimization function that maximizes the coverage of view-dependent appearance for sufficient data capture; (3) developing an optimized Gaussian Splatting framework with a Reflection MLP, a novel deferred shading function, and two enhanced regularization terms for effective modeling of high-frequency near-field reflections; (4) introducing a standardized data organization paradigm for structuring GS-based representations into object-based models, facilitating interactive facility management on digital twin platforms. Experiments on real-world reflective glass facade scenes validate the effectiveness and superiority of the proposed method. Specifically, the glass panel segmentation achieves an improvement of 0.1927 in mIoU over SOTA methods, and only our method enables instance-level panel extraction. The UAV view planning improves novel view synthesis for reflective facades by 13.15 dB in PSNR compared to commercially used nap-of-the-object planning methods. The RefGlass-GS modeling outperforms SOTA Gaussian Splatting approaches for reflective scenes with an average improvement of 5.08 dB in PSNR.
Chinese Translation
现有的反射玻璃外立面建筑数字化面临几何重建失真、不真实的视依赖纹理渲染以及基于对象的语义增强困难等问题。因此，我们提出了RefGlass-GS，这是一种融合框架，能够实现基于无人机的反射玻璃外立面的端到端照片级真实感、语义和交互式数字化。其贡献包括：（1）提出了一种基于最大后验估计的单个玻璃面板分割方法，该方法利用结构规律性，能够抵抗严重的反射和背景干扰；（2）制定了一种无人机视角规划优化函数，以最大化视依赖外观的覆盖范围，从而确保数据捕获的充分性；（3）开发了一种优化的高斯点云框架，结合了反射多层感知机（Reflection MLP）、一种新颖的延迟阴影函数以及两个增强的正则化项，以有效建模高频近场反射；（4）引入了一种标准化的数据组织范式，将基于高斯点云的表示结构化为基于对象的模型，促进数字双胞胎平台上的交互式设施管理。在真实世界的反射玻璃外立面场景中的实验验证了所提方法的有效性和优越性。具体而言，玻璃面板分割在mIoU上比最先进的方法提高了0.1927，且仅我们的算法能够实现实例级面板提取。无人机视角规划在PSNR上比商业使用的物体表面规划方法提高了13.15 dB，改善了反射外立面的新视图合成。RefGlass-GS建模在反射场景中超越了最先进的高斯点云方法，PSNR平均提高了5.08 dB。
cs.CV / 56 / 2606.28828
Ground4D: Consistency-Aware 4D Reconstruction from Monocular Video
Ground4D：基于一致性意识的单目视频4D重建
Abstract
Learning a 4D scene representation from a single monocular video that supports dynamic novel-view synthesis while maintaining faithful geometry over time remains challenging. Dynamic Gaussian Splatting achieves strong rendering performance through photometric optimization, yet does not explicitly enforce multi-view geometric consistency. In contrast, 3D foundation models recover coherent scene geometry and camera motion, but their point-based outputs are not designed for photorealistic rendering. We propose Ground4D, a geometry-grounded framework built on two stages. First, we perform geometry initialization via 3D foundation models, leveraging VGGT in a training-free manner to reconstruct multi-view-consistent 3D geometry and camera poses from monocular video. The recovered geometry provides a structured and reliable initialization for dynamic Gaussian representations. Second, we conduct geometry-consistency-aware refinement via dynamic Gaussian Splatting, optimizing the representation through differentiable rendering while maintaining multi-view geometric consistency across both observed and synthesized viewpoints. Furthermore, Ground4D inherently models the continuous 4D dynamics of the scene, naturally supporting rendering at arbitrary timestamps. By integrating foundation-level geometric priors into dynamic Gaussian optimization, Ground4D achieves stronger reconstruction fidelity and rendering performance, underscoring the role of geometry-grounded constraints in robust 4D scene modeling.
Chinese Translation
从单个单目视频中学习4D场景表示，以支持动态新视图合成，同时保持时间上的几何真实性，仍然具有挑战性。动态高斯点云通过光度优化实现了强大的渲染性能，但并未明确强制执行多视图几何一致性。相比之下，3D基础模型恢复了一致的场景几何和相机运动，但其基于点的输出并未针对照片级真实感渲染进行设计。我们提出了Ground4D，这是一个基于几何的框架，分为两个阶段。首先，我们通过3D基础模型进行几何初始化，利用VGGT以无训练的方式从单目视频重建多视图一致的3D几何和相机姿态。恢复的几何为动态高斯表示提供了结构化且可靠的初始化。其次，我们通过动态高斯点云进行几何一致性意识的细化，通过可微渲染优化表示，同时在观察和合成视点之间保持多视图几何一致性。此外，Ground4D本质上建模了场景的连续4D动态，自然支持在任意时间戳下的渲染。通过将基础级几何先验整合到动态高斯优化中，Ground4D实现了更强的重建保真度和渲染性能，强调了几何约束在稳健的4D场景建模中的重要作用。
cs.CV / 57 / 2606.28840
DLGStream: Dynamic Language-embedded Guassian Splatting for Open-vocabulary Enabled Free-viewpoint Video Streaming
DLGStream：动态语言嵌入高斯溅射用于开放词汇的自由视角视频流媒体
Abstract
3D Gaussian Splatting~(3DGS) has emerged as a promising paradigm for reconstructing streamable free-viewpoint video~(FVV) from multi-view videos. However, 3DGS-based FVVs typically lack user interaction and editing capabilities, which diminishes the immersive experience. Recent research has integrated language features from CLIP into 3DGS via distillation, enabling open-vocabulary queries and supporting many downstream applications. Nevertheless, the stringent requirements of FVV, low frame size and high FPS, make current language Gaussian representations unsuitable for language-embedded FVV. In this paper, we propose DLGStream, a novel language-embedded FVV representation that streams time-varying language features alongside Gaussian attributes to support 4D environment interaction, scene editing, and spatial intelligence. Specifically, we propose a dual-opacity dynamic language Gaussian representation, which maintains two opacity attributes for color and language features to deal with performance degradation that occurs when colors and features are jointly optimized. Furthermore, we introduce an interpolation-based deformation field to reduce temporal redundancy. This deformation field can also be used for 4D frame interpolation, boosting FVV sequences from low to high FPS. Experimental results demonstrate that DLGStream achieves superior performance in both on open-vocabulary segmentation and reconstruction quality with an average frame size of merely 43 KB. The code is available on \href{https://github.com/kkkzh/DLGStream}{https://github.com/kkkzh/DLGStream}.
Chinese Translation
3D高斯溅射（3DGS）已成为从多视角视频重建可流式自由视角视频（FVV）的有前景的范式。然而，基于3DGS的FVV通常缺乏用户交互和编辑能力，这降低了沉浸体验。最近的研究通过蒸馏将CLIP中的语言特征集成到3DGS中，使得开放词汇查询成为可能，并支持许多下游应用。然而，FVV的严格要求（低帧大小和高帧率）使得当前的语言高斯表示不适合语言嵌入的FVV。本文提出了DLGStream，一种新颖的语言嵌入FVV表示，能够流式传输随时间变化的语言特征与高斯属性，以支持4D环境交互、场景编辑和空间智能。具体而言，我们提出了一种双不透明度动态语言高斯表示，维护颜色和语言特征的两个不透明度属性，以应对颜色和特征联合优化时出现的性能下降。此外，我们引入了一种基于插值的变形场，以减少时间冗余。该变形场还可用于4D帧插值，将FVV序列从低帧率提升至高帧率。实验结果表明，DLGStream在开放词汇分割和重建质量方面表现优越，平均帧大小仅为43 KB。代码可在 [https://github.com/kkkzh/DLGStream](https://github.com/kkkzh/DLGStream) 获取。
cs.CV / 58 / 2606.28845
Personalizing MLLMs via Reinforced Multimodal Reference Game
通过强化多模态参考游戏个性化多模态大型语言模型
Abstract
Personalizing Multimodal Large Language Models (MLLMs) aims to recognize users' unique concepts from visual data and provide personalized responses. Although prior work has shown the benefit of concept descriptions and reasoning for this task, MLLM descriptions often include information, such as state and context, that does not help and may in fact hinder the unique identification of the target concept among other visually similar items. Effective descriptions of personal concepts should instead be accurate, discriminative, and free of distracting details. To achieve such descriptions, we introduce Reinforced Reference Game (RRG), a learning framework that promotes discriminative descriptions through a novel reinforced multimodal reference game. The MLLM plays both the roles of speaker and listener in a contrastive game setting, whose goal is to effectively communicate discriminative information about a target concept. Our approach formulates a verifiable contrastive reward over hard positives (dissimilar views of the same concept) and hard negatives (visually similar but different concepts). Empirically, RRG achieves state-of-the-art across multiple tasks on three personalization benchmarks. RRG generalizes to unseen domains and outperforms existing methods based on concept descriptions and personalization-specific RL frameworks. We will release code and models in the project page.
Chinese Translation
个性化多模态大型语言模型（MLLMs）旨在从视觉数据中识别用户的独特概念并提供个性化响应。尽管之前的研究表明概念描述和推理对这一任务的益处，但MLLM的描述往往包含诸如状态和上下文等信息，这些信息并不有助于识别目标概念，反而可能妨碍在其他视觉上相似的项目中对其的独特识别。有效的个人概念描述应当是准确的、具有区分性的，并且不包含分散注意力的细节。为了实现这样的描述，我们引入了强化参考游戏（Reinforced Reference Game, RRG），这是一种通过新颖的强化多模态参考游戏促进区分性描述的学习框架。MLLM在对比游戏设置中同时扮演说话者和听者的角色，其目标是有效传达关于目标概念的区分性信息。我们的方法在难正样本（同一概念的不同视角）和难负样本（视觉上相似但不同的概念）上制定了可验证的对比奖励。实证结果表明，RRG在三个个性化基准测试的多个任务上达到了最先进的水平。RRG能够推广到未见领域，并且优于基于概念描述和个性化特定强化学习框架的现有方法。我们将在项目页面上发布代码和模型。
cs.CV / 59 / 2606.28859
EpiSAM: Character Segmentation in Challenging Stone Inscriptions
EpiSAM：在具有挑战性的石刻铭文中的字符分割
Abstract
Stone inscriptions are invaluable sources of historical and linguistic knowledge, yet their automated analysis remains a major challenge due to surface irregularities, erosion, and low visual contrast. Conventional document and handwriting analysis techniques fail to perform well in these scenarios. In this work, we propose character detection as a core strategy for robust inscription analysis. We introduce EpiSAM, a prompt-guided transformer framework for character segmentation in stone inscriptions. Rather than treating characters in isolation, EpiSAM employs a novel neighbor-aware strategy, explicitly predicting adjacent characters alongside the target. These contextual cues resolve boundary ambiguities, improving mask generation and enabling more accurate character segmentation. Furthermore, we expand an existing stone inscription dataset by adding dense polygonal annotations for characters, thereby enabling comprehensive research on Southeast Asian epigraphy. Experimental results show that EpiSAM achieves consistent improvements over existing baselines, while also exhibiting strong zero-shot generalization in challenging epigraphic scenarios.
Chinese Translation
石刻铭文是历史和语言知识的宝贵来源，但由于表面不规则、侵蚀和低视觉对比度，其自动化分析仍然面临重大挑战。传统的文档和手写分析技术在这些情况下表现不佳。在本研究中，我们提出将字符检测作为稳健铭文分析的核心策略。我们引入了EpiSAM，一个基于提示引导的变换器框架，用于石刻铭文中的字符分割。EpiSAM并不将字符孤立处理，而是采用了一种新颖的邻域感知策略，明确预测目标字符旁边的相邻字符。这些上下文线索解决了边界模糊问题，改善了掩膜生成，并实现了更准确的字符分割。此外，我们通过为字符添加密集的多边形注释，扩展了现有的石刻铭文数据集，从而促进了对东南亚铭文的全面研究。实验结果表明，EpiSAM在现有基准上实现了一致的改进，同时在具有挑战性的铭文场景中也表现出强大的零样本泛化能力。
cs.CV / 60 / 2606.28862
HKVLM: Faithful Reasoning Grounding by Binding Language Queries to a Frozen Detector
HKVLM：通过将语言查询绑定到冻结检测器实现忠实推理基础
Abstract
Many visual requests -- ``the object to open this bottle'', ``the person not wearing a helmet'' -- require reasoning, not just category matching. Pure open-vocabulary detectors need an explicit phrase; vision-language models (VLMs) can reason yet ``see but mis-speak'', attending to the right region but returning the wrong box or label. We argue this is a \emph{binding} failure: in coordinate-as-text VLMs localization passes through the autoregressive head, coupling it to language generation; in two-stage pipelines the model's intent is squeezed through a single class string. We present HKVLM, which removes localization from the language path. A frozen, language-aligned detector emits class-agnostic region proposals; a frozen language model encodes reasoning instructions as referential query embeddings; a lightweight \emph{alignment hook} binds queries to regions by contrastive retrieval and bipartite assignment in a shared embedding space. A perception-grounded faithfulness veto forbids naming an object that no region supports. Only the hook is trained, targeting small-data cold-start settings where monolithic VLM tuning struggles. We formalize a \emph{say-vs-see} decomposition separating localization error (SeeErr) from binding error (SayErr), and evaluate on RefCOCO/RefCOCO+/RefCOCOg and POPE. With frozen Grounding DINO and Qwen2.5-VL, training only the hook lifts grounding accuracy by $50$--$90\times$ over untrained cross-space matching; the faithfulness veto raises POPE accuracy from near-chance ($0.50$) to $0.66$--$0.76$ and reduces hallucination from ${\sim}0.99$ to $0.23$--$0.43$, with gains from $200$ expressions. Increasing proposals from $M{=}50$ to $M{=}300$ improves grounding by $19$--$24\%$ without retraining, confirming that residual error is perceptual (SeeErr) rather than binding (SayErr).
Chinese Translation
许多视觉请求——“打开这个瓶子的物体”，“没有戴头盔的人”——需要推理，而不仅仅是类别匹配。纯开放词汇检测器需要明确的短语；视觉语言模型（VLMs）能够推理，但却“看见却说错”，关注正确的区域却返回错误的框或标签。我们认为这是一个 extit{绑定}失败：在坐标作为文本的VLMs中，定位通过自回归头进行，导致其与语言生成耦合；在两阶段管道中，模型的意图被压缩为单一类别字符串。我们提出HKVLM，它将定位从语言路径中移除。一个冻结的、与语言对齐的检测器发出类别无关的区域提议；一个冻结的语言模型将推理指令编码为指代查询嵌入；一个轻量级的 extit{对齐钩子}通过对比检索和在共享嵌入空间中的二分分配将查询绑定到区域。一个基于感知的忠实性否决禁止命名没有区域支持的物体。只有钩子被训练，目标是小数据冷启动设置，在这些设置中，单一的VLM调优面临困难。我们形式化了一个 extit{说-看}分解，将定位误差（SeeErr）与绑定误差（SayErr）分开，并在RefCOCO/RefCOCO+/RefCOCOg和POPE上进行评估。使用冻结的Grounding DINO和Qwen2.5-VL，仅训练钩子使得基础准确率提高了$50$--$90 imes$，相较于未训练的跨空间匹配；忠实性否决将POPE准确率从接近偶然（$0.50$）提高到$0.66$--$0.76$，并将幻觉率从${ ilde{0.99}}$降低到$0.23$--$0.43$，其中获得了$200$个表达式的收益。将提议数量从$M{=}50$增加到$M{=}300$，在不重新训练的情况下提高了$19$--$24\%$的基础准确率，确认残余误差是感知性的（SeeErr），而非绑定性的（SayErr）。
cs.CV / 61 / 2606.28864
On Test-Time Scaling for Vision-Language Models
关于视觉-语言模型的测试时缩放
Abstract
Test-time scaling is a paradigm where large models use additional compute at inference to achieve better performance, without changing model weights. While it has been widely studied for Large Language Models (LLMs), its applicability to Large Vision-Language Models (LVLMs) remains less explored and analyzed, with limited analysis of whether, when, and to what extent these approaches transfer to LVLMs. In this work, we ask a simple but fundamental question: can conventional test-time scaling methods developed for LLMs be directly applied to LVLMs? We present the first comprehensive study of test-time scaling for LVLMs, spanning multiple models and model sizes, nine test-time scaling methods, and six diverse benchmarks. Our main findings is that 1) different from previous findings, small, well-performing models benefit the most from test-time scaling, enabling performance improvements of up to around 30\%, reaching large models performance, and often outperforming them, 2) LVLMs lose focus when given more compute than necessary, and 3) Visual information is encoded early in the reasoning chain, after which the chain is dominated by text-only reasoning and the contribution of image tokens drops significantly. Finally, we also provide a global and fine-grained analysis on the quality and information sufficiency of the reasoning chains produced. Overall, our findings and analysis provide practical guidance and insights into LVLMs and their deployment in research and industry.
Chinese Translation
测试时缩放是一种范式，其中大型模型在推理时使用额外的计算资源以实现更好的性能，而不改变模型权重。虽然这一方法已在大型语言模型（LLMs）中得到了广泛研究，但其在大型视觉-语言模型（LVLMs）中的适用性仍然较少被探讨和分析，关于这些方法是否、何时以及在多大程度上可以转移到LVLMs的分析也相对有限。在本研究中，我们提出了一个简单但根本性的问题：为LLMs开发的传统测试时缩放方法是否可以直接应用于LVLMs？我们首次对LVLMs的测试时缩放进行了全面研究，涵盖了多个模型和模型规模、九种测试时缩放方法以及六个不同的基准测试。我们的主要发现是：1）与之前的发现不同，表现良好的小型模型在测试时缩放中受益最大，性能提升可达约30%，达到大型模型的性能，并且往往超过它们；2）LVLMs在获得超过必要的计算资源时会失去焦点；3）视觉信息在推理链的早期被编码，之后推理链主要由文本推理主导，图像标记的贡献显著下降。最后，我们还对生成的推理链的质量和信息充分性进行了全局和细致的分析。总体而言，我们的发现和分析为LVLMs及其在研究和工业中的应用提供了实用的指导和见解。
cs.CV / 62 / 2606.28905
Projection-based coupling of infrared thermography and stereocorrelation-based digital image correlation
基于投影的红外热成像与基于立体相关的数字图像相关的耦合
Abstract
Full-field measurement techniques such as digital image correlation and infrared thermography are prevalent in experimental solid mechanics. Digital image correlation is used to analyze surface deformation, while infrared thermography quantifies surface temperature fields. However, sophisticated procedures are necessary to express both datasets in the same Lagrangian frame, especially when analyzing non-flat surfaces. In this study, we propose an external projection-based coupling that uses the pinhole camera model to relate two-dimensional temperature data measured by infrared thermography to three-dimensional point coordinates from stereocorrelation-based digital image correlation. Unlike existing multiview approaches, we utilize two independently calibrated industrial-grade systems and augment the experimental evaluation with the pinhole camera model. The projection matrix of the camera model is calibrated using a single image of a reference object. Through this projection, temperature fields are accurately represented at material points. Our method is particularly suited for, but not restricted to, curved surfaces and straightforward to embed in existing experimental protocols, as the image registration is kept as is. Additionally, we propose using radial basis functions as a global interpolation ansatz in both space and time to compute in-plane temperature gradients and even temperature rates on curved surfaces, thereby providing an extensive and information-rich full-field dataset.
Chinese Translation
全场测量技术，如数字图像相关和红外热成像，在实验固体力学中广泛应用。数字图像相关用于分析表面变形，而红外热成像则量化表面温度场。然而，在分析非平面表面时，需要复杂的程序将这两组数据表达在同一拉格朗日框架中。在本研究中，我们提出了一种基于外部投影的耦合方法，利用针孔相机模型将红外热成像测得的二维温度数据与基于立体相关的数字图像相关获得的三维点坐标关联起来。与现有的多视图方法不同，我们使用两个独立校准的工业级系统，并通过针孔相机模型增强实验评估。相机模型的投影矩阵通过对参考物体的单幅图像进行校准。通过这种投影，温度场在材料点上得到了准确的表示。我们的方法特别适用于但不限于曲面，并且易于嵌入现有的实验协议，因为图像配准保持不变。此外，我们建议在空间和时间上使用径向基函数作为全局插值假设，以计算曲面上的面内温度梯度甚至温度变化率，从而提供一个广泛且信息丰富的全场数据集。
cs.CV / 63 / 2606.28912
What Color is the Sky (for a non-human) ?
天空的颜色是什么（对于非人类而言）？
Abstract
The light of the daytime sky contains a mixture of many colors yet is perceived as blue by human observers. This is largely due to the particular response functions of the human cones. Under these response functions skylight and blue light are metamers: they yield the exact same excitation of the cones. In this paper we ask: is it possible to define the ``color'' of the sky for other visual systems? We present a simple computational method to determine monochromatic metamers to a given input light for arbitrary visual systems. Using published values on spectral sensitivity functions of various species, we use our method to determine the dominant wavelength of monochromatic metamers to skylight. For a wide range of species (bichromats, trichromats and tetrachormats) we find monochromatic metamers to skylight but the dominant wavelength of the metamer can vary drastically between species and be very different from the color perceived by humans.
Chinese Translation
白天天空的光线包含多种颜色的混合，但人类观察者却感知为蓝色。这主要是由于人类视锥细胞的特定反应函数。在这些反应函数下，天空光和蓝光是同质异构体（metamers）：它们对视锥细胞产生完全相同的激发。在本文中，我们提出了一个问题：是否可以为其他视觉系统定义天空的“颜色”？我们提出了一种简单的计算方法，以确定对任意视觉系统给定输入光的单色同质异构体。利用已发布的各种物种的光谱敏感性函数值，我们使用该方法确定天空光的单色同质异构体的主导波长。对于广泛的物种（双色视觉、三色视觉和四色视觉），我们发现存在天空光的单色同质异构体，但这些同质异构体的主导波长在不同物种之间可能有显著差异，并且与人类感知的颜色非常不同。
cs.CV / 64 / 2606.28920
ExACT: Exemplar-Driven Calibrated Refinement for Training-Free Visual Grounding in Remote Sensing Images
ExACT：基于示例驱动的标定精细化框架用于无训练的遥感图像视觉定位
Abstract
Remote sensing visual grounding (RSVG) aims to locate specific objects in high-resolution RS imagery using free-form natural language descriptions. While recent advances in multimodal large language models (MLLMs) show great potential for such open-vocabulary RSVG, their training-free adaptation is hindered by the modality gap between abstract linguistic semantics and fine-grained visual cues. In cluttered RS scenes, this gap inevitably causes severe localization drift. To bridge this gap, we propose Exemplar-driven Calibrated Refinement (ExACT), a novel training-free framework driven by a one-shot visual prompting mechanism to explicitly provide discriminative structural guidance for precise pixel-level localization. Specifically, we propose a Vision Exemplar-based Calibrator (VEC) that extracts fine-grained visual correspondences from the given exemplar to rectify the rough cross-modal priors from frozen MLLMs, effectively suppressing background artifacts and accurately outlining target boundaries. Subsequently, a Structure-Aware Refiner (SAR) employs an iterative merge-and-select clustering strategy to consolidate the calibrated priors into high-quality positive and negative geometric prompts. These prompts then guide the Segment Anything Model (SAM) to achieve precise pixel-level predictions. Extensive experiments confirm the superiority of ExACT over existing training-free and weakly-supervised methods.
Chinese Translation
遥感视觉定位（RSVG）旨在利用自由形式的自然语言描述在高分辨率遥感图像中定位特定物体。尽管最近在多模态大型语言模型（MLLMs）方面的进展显示出在开放词汇RSVG中的巨大潜力，但其无训练适应受到抽象语言语义与细粒度视觉线索之间的模态差距的阻碍。在杂乱的遥感场景中，这一差距不可避免地导致严重的定位漂移。为了解决这一问题，我们提出了基于示例驱动的标定精细化框架（ExACT），这是一种新颖的无训练框架，通过一次性视觉提示机制明确提供区分性结构指导，以实现精确的像素级定位。具体而言，我们提出了一种基于视觉示例的标定器（VEC），该标定器从给定示例中提取细粒度视觉对应关系，以修正来自冻结的MLLMs的粗略跨模态先验，有效抑制背景伪影并准确勾勒目标边界。随后，结构感知精细化器（SAR）采用迭代合并与选择聚类策略，将标定后的先验整合为高质量的正负几何提示。这些提示随后指导Segment Anything Model（SAM）实现精确的像素级预测。大量实验验证了ExACT在现有无训练和弱监督方法中的优越性。
cs.CV / 65 / 2606.28946
Character Recognition of Nepali Number Plate
尼泊尔车牌字符识别
Abstract
This paper presents a robust Automatic Number Plate Recognition (ANPR) system tailored for Nepali license plates written in Devanagari script. In this paper, a pipelined model was used that integrates YOLO-based models for license plate and character detection, followed by a CNN classifier trained on 34 Devanagari characters. Two publicly available data sets were used that incorporate diverse lighting, fonts, and structural variations. Data augmentation and additional training on embossed plates enhanced the generalizability of the model. The system achieved a recognition accuracy of up to 93\%, demonstrating strong performance under real-world conditions and providing a scalable solution for traffic management in Nepal. Code: https://github.com/Satyasakhadka/Nepali-NumberPlate-Character-Recognition
Chinese Translation
本文提出了一种针对使用天城文书写的尼泊尔车牌的强健自动车牌识别（ANPR）系统。本文采用了一种流水线模型，该模型集成了基于YOLO的车牌和字符检测模型，随后使用在34个天城文字符号上训练的卷积神经网络（CNN）分类器。使用了两个公开可用的数据集，这些数据集涵盖了多种光照、字体和结构变化。通过数据增强和对浮雕车牌的额外训练，提升了模型的泛化能力。该系统在实际条件下实现了高达93%的识别准确率，展示了其在尼泊尔交通管理中的强大性能和可扩展性。代码链接： https://github.com/Satyasakhadka/Nepali-NumberPlate-Character-Recognition
cs.CV / 66 / 2606.28971
Self-Evolving Agentic Image Restoration via Deliberate Planning and Intuitive Execution
自我演化的智能图像恢复：通过深思熟虑的规划与直观执行
Abstract
Real-world image restoration (IR) remains challenging due to complex and coupled degradations. While recent agentic IR frameworks leverage Large Language Models for flexible tool planning, they face two critical limitations. First, from a search scheme perspective, excessive reliance on greedy strategies fails to balance exploration and exploitation. Second, existing agentic systems underutilize information, exhibiting episodic amnesia. To address these challenges, we propose \textbf{Self-Evolving Agentic Image Restoration (SEAR)}, which formulates restoration as a sequential decision-making problem. Inspired by the dual-process theory, SEAR comprises an Intuitive Executor and a Deliberate Planner, respectively following the fast-thinking \textit{System 1} and slow-thinking \textit{System 2} principles. The Deliberate Planner employs Pruning-Aware Monte Carlo Tree Search for long-horizon reasoning, utilizing a hybrid no-reference reward and a Multimodal Large Language Model (MLLM)-based tournament to prevent metric exploitation. Complementarily, the Intuitive Executor leverages a self-evolving episodic memory indexed by degradation-aware state fingerprints. This mechanism distills expensive search trajectories into adaptive expertise, overcoming episodic amnesia while progressively amortizing cold-start exploration costs through memory reuse. Extensive experiments on synthetic and real-world benchmarks demonstrate its strong perceptual and quantitative performance.
Chinese Translation
现实世界中的图像恢复（IR）因复杂和耦合的退化而仍然具有挑战性。尽管最近的智能图像恢复框架利用大型语言模型进行灵活的工具规划，但它们面临两个关键限制。首先，从搜索方案的角度来看，过度依赖贪婪策略未能平衡探索与开发。其次，现有的智能系统未能充分利用信息，表现出情节性遗忘。为了解决这些挑战，我们提出了 extbf{自我演化的智能图像恢复（SEAR）}，将恢复过程形式化为一个序列决策问题。受到双重过程理论的启发，SEAR由直观执行者和深思熟虑的规划者组成，分别遵循快速思维的 extit{系统1}和慢速思维的 extit{系统2}原则。深思熟虑的规划者采用剪枝感知的蒙特卡洛树搜索进行长远推理，利用混合的无参考奖励和基于多模态大型语言模型（MLLM）的竞赛来防止指标的开发。作为补充，直观执行者利用一个自我演化的情节记忆，该记忆通过退化感知状态指纹进行索引。该机制将昂贵的搜索轨迹提炼为自适应的专业知识，克服情节性遗忘，同时通过记忆重用逐步摊销冷启动探索成本。在合成和现实世界基准上的广泛实验展示了其强大的感知和定量性能。
cs.CV / 67 / 2606.28980
Evidence-Based Text-Conditioned 3D CT Synthesis for Ovarian Cancer
基于证据的文本条件3D CT合成用于卵巢癌
Abstract
Ovarian cancer is frequently diagnosed at an advanced stage, making preoperative contrast-enhanced computed tomography (CT) central to staging and surgical planning; yet the scarcity of annotated imaging data, compounded by privacy regulations, limits the development of generalizable computational models in this domain. Text-conditioned 3D CT synthesis has shown promise, but existing pipelines depend on paired radiology reports and have been evaluated only on chest CT. We propose OvESyn (Ovarian Evidence-based Synthesis), a framework that constructs standardized Findings and Impression sections directly from CT-derived imaging descriptors and routine clinical metadata, without any original radiology report, and uses them to condition a latent diffusion model adapted to 493 high-grade serous ovarian carcinoma patients. This is the first text-conditioned 3D CT synthesis framework adapted to an abdomino-pelvic oncologic setting. A systematic ablation over two adaptation axes, vision-language encoder alignment and generator fine-tuning, identifies generator domain adaptation as the operative mechanism for crossing the domain gap and establishing the target anatomy: without it, synthesis remains anchored to the thoracic pretraining domain, with Precision and Recall collapsing to zero and FID2.5D exceeding 140, regardless of encoder alignment. Encoder alignment instead refines intensity and fine detail. The full OvESyn attains the best distributional and intensity fidelity (FID2.5D 29.35, Precision 0.671, Wasserstein-1 0.044), while the generator-only variant maximizes coverage (Recall 0.645), reflecting a fidelity/coverage trade-off governed by encoder adaptation. Requiring only automatic segmentations and routine preoperative metadata, OvESyn supports transferability to report-scarce settings and provides a foundation for synthetic cohort generation in abdomino-pelvic oncologic imaging.
Chinese Translation
卵巢癌常在晚期被诊断，因此术前增强对比计算机断层扫描（CT）在分期和手术规划中至关重要；然而，标注影像数据的稀缺，加上隐私法规的限制，限制了该领域可推广计算模型的发展。文本条件的3D CT合成显示出良好的前景，但现有的流程依赖于配对的放射学报告，并且仅在胸部CT上进行了评估。我们提出了OvESyn（卵巢基于证据的合成），这是一个框架，直接从CT衍生的影像描述符和常规临床元数据构建标准化的发现和印象部分，而无需任何原始放射学报告，并利用这些信息来条件化一个适应于493例高分化浆液性卵巢癌患者的潜在扩散模型。这是第一个适应于腹盆腔肿瘤环境的文本条件3D CT合成框架。通过对两个适应轴的系统消融，即视觉-语言编码器对齐和生成器微调，确定生成器领域适应是跨越领域差距和建立目标解剖结构的有效机制：没有它，合成仍然停留在胸部预训练领域，精确度和召回率降至零，FID2.5D超过140，无论编码器对齐如何。编码器对齐则精炼了强度和细节。完整的OvESyn实现了最佳的分布和强度保真度（FID2.5D 29.35，精确度0.671，Wasserstein-1 0.044），而仅生成器变体最大化了覆盖率（召回率0.645），反映了由编码器适应主导的保真度/覆盖率权衡。OvESyn仅需自动分割和常规术前元数据，支持在报告稀缺环境中的可转移性，并为腹盆腔肿瘤影像中的合成队列生成提供了基础。
cs.CV / 68 / 2606.28991
Learning from Acquisition: Metadata-driven Multimodal Pre-training for Cardiac MRI
从获取中学习：基于元数据驱动的多模态预训练用于心脏MRI
Abstract
Cardiac magnetic resonance imaging (CMR) routinely records structured acquisition metadata, yet most CMR foundation models rely primarily on image-only pre-training and leave this naturally available source of weak semantic supervision largely underexplored. We propose MetaCLIP-CMR, a metadata-driven framework based on Contrastive Language--Image Pre-training (CLIP), which converts imaging modality, anatomical view, scanner vendor, field strength, and scanner model into textual supervision for CMR representation learning. The pretrained image encoder is evaluated on imaging modality classification, cine view classification, and cardiac segmentation. MetaCLIP-CMR achieves 86.8% modality accuracy and 86.5% cine view accuracy, clearly outperforming ImageNet and masked reconstruction initialisations. For downstream cardiac segmentation, MetaCLIP-CMR consistently obtains the highest Dice score across the evaluated ACDC and M&Ms cine short-axis (SAX) settings under both full-data and 20% fine-tuning regimes. Compared with recent image-focused large-scale CMR pre-training models, MetaCLIP-CMR achieves comparable ACDC segmentation performance, while requiring less than 1% of their pre-training image scale. These results suggest that metadata learning offers a natural and easy-to-use strategy for transforming routinely recorded acquisition information into effective supervision for foundation-level CMR representation learning, highlighting the promise of metadata-driven multimodal pre-training.
Chinese Translation
心脏磁共振成像（CMR）常规记录结构化的获取元数据，但大多数CMR基础模型主要依赖于仅图像的预训练，未能充分利用这一自然可得的弱语义监督源。我们提出了MetaCLIP-CMR，这是一种基于对比语言-图像预训练（CLIP）的元数据驱动框架，将成像模态、解剖视图、扫描仪供应商、场强和扫描仪型号转换为心脏MRI表示学习的文本监督。预训练的图像编码器在成像模态分类、心动图视图分类和心脏分割上进行了评估。MetaCLIP-CMR在模态准确率上达到了86.8%，在心动图视图准确率上达到了86.5%，明显优于ImageNet和掩蔽重建初始化。在下游心脏分割任务中，MetaCLIP-CMR在评估的ACDC和M&Ms心动图短轴（SAX）设置下，在全数据和20%微调方案中始终获得最高的Dice分数。与最近的以图像为中心的大规模CMR预训练模型相比，MetaCLIP-CMR在ACDC分割性能上表现相当，同时其预训练图像规模不足1%。这些结果表明，元数据学习提供了一种自然且易于使用的策略，将常规记录的获取信息转化为有效的基础级CMR表示学习监督，突显了元数据驱动的多模态预训练的前景。
cs.CV / 69 / 2606.29004
SciFlow: Semantic Cross Interference for Self-Supervised Optical Flow Domain Generalization
SciFlow：自监督光流领域泛化的语义交叉干扰
Abstract
Motions of objects and scenes carry essential intelligence in video understanding, offering rich cues for interpreting dynamic settings and interactions. Due to the cost and scarcity of high-quality annotation or ground truth of pixel-wise optical flow, however, motion estimation models are typically trained in synthetic domains while deployed in real-world domains. Addressing synthetic-to-real domain generalization challenges has been crucial for developing practical solutions in diverse open-world use cases. This paper introduces SciFlow, a simple yet effective, network-agnostic, training-based approach that leverages self-supervised learning to generalize motion estimation across synthetic and open-world domains. Specifically, SciFlow imposes semantic interference from open-world images onto synthetic images during training, blending indomain features with cross-domain interference, which enables the network to adapt to the real-world domains. Additionally, SciFlow utilizes geometric consistency to ensure validity of the self-supervision. Our experiment results show that SciFlow not only significantly enhances model robustness amidst domain variations, but also remarkably enables synthetic-to-real domain generalization without requiring any ground truth in the open world.
Chinese Translation
物体和场景的运动在视频理解中承载着重要的智能，提供了丰富的线索以解释动态环境和交互。然而，由于高质量像素级光流的注释或真实值的成本和稀缺性，运动估计模型通常在合成领域进行训练，而在真实世界领域中部署。解决合成到真实领域的泛化挑战对于在多样化的开放世界应用中开发实用解决方案至关重要。本文介绍了SciFlow，这是一种简单而有效的、与网络无关的基于训练的方法，利用自监督学习来实现合成和开放世界领域之间的运动估计泛化。具体而言，SciFlow在训练过程中将开放世界图像的语义干扰施加到合成图像上，将领域内特征与跨领域干扰相结合，使网络能够适应真实世界领域。此外，SciFlow利用几何一致性来确保自监督的有效性。我们的实验结果表明，SciFlow不仅显著增强了模型在领域变化中的鲁棒性，而且在开放世界中无需任何真实值的情况下，显著实现了合成到真实领域的泛化。
cs.CV / 70 / 2606.29013
Mural: Transferring LLM knowledge to image generation via Mixture-of-Transformers
Mural：通过混合变换器将大型语言模型知识转移到图像生成
Abstract
Leveraging capabilities of large language models (LLMs) in text-to-image (T2I) synthesis is an important research direction. In this work we investigate whether the knowledge of a frozen LLM can be effectively utilized in T2I generation when trained exclusively on standard text-image pairs. We integrate a frozen, reasoning-capable LLM with a diffusion-based image generator via shared attention within the Mixture-of-Transformers (MoT) architecture. Our experiments span two critical questions: (1) what degree of the LLM's intrinsic knowledge remains accessible during T2I training, and (2) what novel capabilities emerge in the resulting system. Across established benchmarks, our models achieve strong performance among unified understanding-generation systems: 0.85 on GenEval, 86.75 on DPG-Bench, and 0.66 on WISE with inference-time reasoning, using only text-image data. Remarkably, we uncover emergent behaviors absent from training data, including cross-lingual image generation, color-guided composition, emoji / ASCII scene construction, and generation directed by world knowledge. These results demonstrate that pretrained LLM knowledge can guide image synthesis under standard text-to-image training paradigms, without interleaved multimodal signals or explicit reasoning supervision. Our findings open new avenues for harnessing frozen model capabilities in resource-constrained multimodal learning.
Chinese Translation
利用大型语言模型（LLMs）在文本到图像（T2I）合成中的能力是一个重要的研究方向。在本研究中，我们探讨了在仅使用标准文本-图像对进行训练时，冻结的LLM的知识是否可以有效地应用于T2I生成。我们通过在混合变换器（Mixture-of-Transformers, MoT）架构中共享注意力，将一个冻结的、具备推理能力的LLM与基于扩散的图像生成器集成。我们的实验涵盖了两个关键问题：（1）在T2I训练过程中，LLM的内在知识在多大程度上仍然可被访问，以及（2）在生成的系统中出现了哪些新能力。在已建立的基准测试中，我们的模型在统一理解-生成系统中表现出色：在GenEval上得分0.85，在DPG-Bench上得分86.75，在WISE上得分0.66，并且在推理时具备推理能力，仅使用文本-图像数据。值得注意的是，我们发现了一些训练数据中不存在的突现行为，包括跨语言图像生成、颜色引导的构图、表情符号/ASCII场景构建以及由世界知识引导的生成。这些结果表明，预训练的LLM知识可以在标准的文本到图像训练范式下指导图像合成，而无需交错的多模态信号或明确的推理监督。我们的发现为在资源受限的多模态学习中利用冻结模型的能力开辟了新的途径。
cs.CV / 71 / 2606.29020
Semantic-Aware, Physics-Informed, Geometry-Grounded Weather Video Synthesis
语义感知、物理驱动、几何基础的天气视频合成
Abstract
Weather synthesis aims to add weather effects to input videos while preserving scene identity, structure, and motion. The key limitation of existing methods is the lack of diversity in weather appearance and effective control over weather dynamics (e.g., temporal evolution and particle motion). Most approaches rely on text prompts, which are inherently underspecified and often fail to produce detailed weather characteristics. Additionally, general-purpose video editors optimized for clean and aesthetic outputs tend to suppress heavy weather phenomena, making dense particle effects difficult to generate. To address these, we propose a Semantic-Aware, Physics-Informed, and Geometry-Grounded framework that steers an off-the-shelf video editor to synthesize diverse global appearances and detailed particle dynamics. We factorize the synthesis into three conditional signals, so that each provides a distinct and stable source of guidance: semantics specifies what the weather should look like, dynamics governs how it evolves over time, and geometry determines where it should appear in the scene. Specifically, we introduce (1) semantic-aware appearance anchoring to establish the target appearance from scene semantics and user input; (2) physics-informed dynamic simulation to generate particle effects by simulating a Gaussian-represented particle field under gravity, wind, and turbulence; and (3) geometry-grounded video synthesis to align the simulated particles with target scene geometry and synthesize the final video. Experiments demonstrate that our method produces diverse, physically and visually realistic weather effects. Furthermore, we show that our synthesized data significantly improves the robustness of autonomous driving semantic segmentation under adverse weather conditions. Project page: https://jumponthemoon.github.io/w-crafter/.
Chinese Translation
天气合成旨在为输入视频添加天气效果，同时保持场景的身份、结构和运动。现有方法的主要限制在于天气外观缺乏多样性以及对天气动态（例如，时间演变和粒子运动）的有效控制。大多数方法依赖于文本提示，这些提示本质上是不够具体的，往往无法产生详细的天气特征。此外，针对干净和美观输出而优化的通用视频编辑器往往会抑制强烈的天气现象，使得生成密集的粒子效果变得困难。为了解决这些问题，我们提出了一种语义感知、物理驱动和几何基础的框架，该框架引导现成的视频编辑器合成多样化的全球外观和详细的粒子动态。我们将合成过程分解为三个条件信号，使每个信号提供独特且稳定的指导来源：语义指定天气应有的外观，动态控制其随时间的演变，几何确定其在场景中的出现位置。具体而言，我们引入了（1）语义感知的外观锚定，以根据场景语义和用户输入建立目标外观；（2）物理驱动的动态模拟，通过在重力、风和湍流下模拟高斯表示的粒子场来生成粒子效果；（3）几何基础的视频合成，以将模拟的粒子与目标场景几何对齐并合成最终视频。实验表明，我们的方法能够产生多样化、物理上和视觉上真实的天气效果。此外，我们还展示了合成的数据显著提高了在恶劣天气条件下自主驾驶语义分割的鲁棒性。项目页面：https://jumponthemoon.github.io/w-crafter/
cs.CV / 72 / 2606.29023
Efficient Spatio-Temporal Grounding with Multimodal Large Models via Second-Level Tracking and RL Verification
通过二级跟踪和强化学习验证实现高效的时空定位与多模态大模型
Abstract
Spatio-temporal grounding in long videos requires precise temporal localization and robust object tracking conditioned on natural-language queries. While recent vision-language models (VLMs) show strong reasoning ability, directly applying frame-by-frame inference to long sequences is computationally expensive and unstable. We propose a practical pipeline that shifts from frame-level to second-level tracking and performs cross-second smoothing to preserve continuity while reducing sequence length. To improve reasoning supervision, we synthesize chain-of-thought style trajectories using advanced multimodal models for temporal localization and target selection, and replace generated spatio-temporal coordinates with ground-truth annotations to avoid noisy supervision. We further optimize the policy with reinforcement learning using a verifier based on $t\_\mathrm{IoU}+mv\_\mathrm{IoU}$. Experiments across multiple FPS settings show that our method achieves a strong trade-off between efficiency and localization quality.
Chinese Translation
长视频中的时空定位需要精确的时间定位和基于自然语言查询的稳健对象跟踪。尽管近期的视觉语言模型（VLMs）展现出强大的推理能力，但将逐帧推理直接应用于长序列在计算上是昂贵且不稳定的。我们提出了一种实用的流程，将跟踪从帧级转移到二级跟踪，并执行跨秒平滑以保持连续性，同时减少序列长度。为了改善推理监督，我们利用先进的多模态模型合成链式思维风格的轨迹，用于时间定位和目标选择，并用真实注释替换生成的时空坐标，以避免噪声监督。我们进一步使用基于 $t ext{IoU}+mv ext{IoU}$ 的验证器，通过强化学习优化策略。多个FPS设置下的实验表明，我们的方法在效率和定位质量之间实现了良好的平衡。
cs.CV / 73 / 2606.29029
Adaptive Spectrum-Aware Feature Disentangled Network for Small Object Detection
自适应谱感知特征解耦网络用于小物体检测
Abstract
Small Object Detection (SOD) is a fundamental yet challenging problem in computer vision due to its limited spatial resolution and weak visual cues. Although recent approaches have achieved remarkable advances, the background distractors in different frequency spectra still degrade the performance. In this paper, we propose a novel small object detection framework termed SFDNet, which is capable of detecting small objects via efficient spectrum-aware feature disentanglement. Specifically, we propose an Adaptive Spectrum Disentanglement (ASD) module that decomposes backbone features into multiple complementary spectral components, aiming to construct discriminative object-relevant representations by discarding the background distractors for each component. Afterwards, to strengthen the semantic consistency of the similar objects in the same class, we propose a Class-Wise Prototype Distillation (CPD) procedure, which establishes class prototypes for the object instances and enforces the compact representation by efficient prototype distillation. Extensive experiments on multiple challenging benchmarks show that SFDNet outperforms existing state-of-the-art methods by a large margin. Our code is available at https://github.com/ManOfStory/SFDNet.
Chinese Translation
小物体检测（SOD）是计算机视觉中的一个基础但具有挑战性的问题，主要由于其有限的空间分辨率和弱视觉线索。尽管近期的方法取得了显著进展，但不同频谱中的背景干扰物仍然会降低性能。本文提出了一种新颖的小物体检测框架，称为SFDNet，该框架能够通过高效的谱感知特征解耦来检测小物体。具体而言，我们提出了一个自适应谱解耦（ASD）模块，该模块将主干特征分解为多个互补的谱成分，旨在通过丢弃每个成分的背景干扰物来构建具有区分性的物体相关表示。随后，为了增强同类相似物体的语义一致性，我们提出了一种类级原型蒸馏（CPD）过程，该过程为物体实例建立类原型，并通过高效的原型蒸馏强制紧凑表示。在多个具有挑战性的基准测试上的广泛实验表明，SFDNet在性能上大幅超越了现有的最先进方法。我们的代码可在 https://github.com/ManOfStory/SFDNet 获取。
cs.CV / 74 / 2606.29059
Flow Matching in Feature Space for Stochastic World Modeling
特征空间中的流匹配用于随机世界建模
Abstract
World modeling requires forecasting uncertain futures while preserving information useful for downstream perception. Existing visual world models often struggle to satisfy both goals: VAE-based stochastic models operate in low-dimensional reconstruction latents, which can limit perception performance, while deterministic predictors using strong pretrained features collapse multimodal futures into a single blurry mean. In this work, we propose FlowWM, a stochastic world model that performs flow matching directly within pretrained feature space (e.g., DINOv3). This is challenging because pretrained features are substantially high-dimensional, making standard diffusion recipes suboptimal. To address this, we investigate the design choices needed for feature-space flow matching and introduce a differentiable one-step projection mechanism that enables efficient training with temporal consistency and task-driven objectives. We evaluate FlowWM on two benchmarks: a synthetic benchmark for systematic evaluation of accuracy and diversity, and a real-world benchmark FuturePerception. FlowWM improves perception performance, mode coverage, and horizon robustness, validating our proposed design for stochastic world modeling in high-dimensional feature spaces.
Chinese Translation
世界建模需要在预测不确定的未来时保留对下游感知有用的信息。现有的视觉世界模型往往难以同时满足这两个目标：基于变分自编码器（VAE）的随机模型在低维重构潜变量中操作，这可能限制感知性能，而使用强预训练特征的确定性预测器则将多模态未来压缩为单一模糊均值。在本研究中，我们提出了FlowWM，一种在预训练特征空间（例如，DINOv3）中直接进行流匹配的随机世界模型。这一过程具有挑战性，因为预训练特征的维度相当高，使得标准扩散方法效果不佳。为了解决这个问题，我们研究了特征空间流匹配所需的设计选择，并引入了一种可微分的一步投影机制，使得在时间一致性和任务驱动目标下实现高效训练成为可能。我们在两个基准上评估了FlowWM：一个用于系统评估准确性和多样性的合成基准，以及一个真实世界基准FuturePerception。FlowWM提高了感知性能、模式覆盖率和视野鲁棒性，验证了我们在高维特征空间中进行随机世界建模的设计方案。
cs.CV / 75 / 2606.29093
From Fog Chamber to Aircraft Window: Pixel-Registered Imaging and Synthetic Fine-Tuning Enable Cross-Domain Defogging
从雾室到飞机窗：像素注册成像与合成微调实现跨领域去雾
Abstract
A deep defogging pipeline pretrained on controlled laboratory fog and fine-tuned with domain-randomized synthetic fog applied to clear outdoor scenes generalizes across a graded sequence of out-of-distribution settings with no target-domain training, from chamber-free free-flowing fog to iPhone video recorded through an aircraft cabin window in flight, an entirely unseen sensor, scene, and optical path. This directly addresses an open transfer limitation reported for real-world binocular defogging. Two design choices support the transfer. First, a single-camera fog imager photographs a flat-panel display through an artificial-fog enclosure with a fixed 114~mm scattering path, producing 5{,}495 pixel-aligned foggy/clear pairs. Exact registration permits a paired Laplacian ratio that predicts per-image restoration quality far better than single-image proxies (Spearman $\rho = 0.632$ versus $0.399$) and supports pixel-exact $L_1$ reconstruction training that avoids adversarial hallucination. Second, the fog-chamber checkpoint is fine-tuned on Mapillary Vistas crops overlaid with on-the-fly randomized synthetic fog spanning a broad range of strengths, spatial variations, airlights, and noise conditions. On a 552-image held-out split, a uniform comparison of 30 restoration backbones places NAFNet at the top (24.33~dB~/~0.7912~SSIM), with a compact alternative within 1.29~dB at 3\% of the parameter count, and a ResNet-50 classifier confirms that the restoration preserves semantic content rather than only pixel-level structure. On unpaired aircraft-window video, NIQE decreases from a mean of 6.22 to 4.97 after fine-tuning, with temporally stable output across full-motion sequences. The same backbone, under paired supervision, also reaches 20.71~dB~/~0.683~SSIM on a non-overlapping O-HAZE/NH-HAZE split (a transferability check rather than a competitive ranking).
Chinese Translation
一个在受控实验室雾中预训练的深度去雾管道，通过域随机化的合成雾进行微调，应用于清晰的户外场景，能够在没有目标领域训练的情况下，跨越一系列分布外设置进行泛化，从无室自由流动的雾到在飞行中通过飞机机舱窗户录制的iPhone视频，这是一种完全未见过的传感器、场景和光学路径。这直接解决了针对真实世界双目去雾的开放转移限制。两个设计选择支持这一转移。首先，一个单摄像头雾成像仪通过一个固定114毫米散射路径的人工雾封闭装置拍摄平面显示器，产生了5,495对像素对齐的雾/清晰图像。精确的配准允许使用配对拉普拉斯比率来预测每幅图像的恢复质量，远远优于单幅图像代理（Spearman $
ho = 0.632$ 对比 $0.399$），并支持像素精确的 $L_1$ 重建训练，避免对抗性幻觉。其次，雾室检查点在覆盖有随机合成雾的Mapillary Vistas裁剪图像上进行微调，合成雾的强度、空间变化、空气光和噪声条件范围广泛。在一个552幅图像的保留分割上，对30个恢复骨干网进行统一比较，NAFNet位居首位（24.33~dB~/~0.7912~SSIM），一个紧凑的替代方案在参数数量的3 ext{%}下与其相差1.29~dB，而ResNet-50分类器确认恢复保留了语义内容，而不仅仅是像素级结构。在未配对的飞机窗户视频上，经过微调后，NIQE从平均6.22降至4.97，且在全动态序列中输出稳定。相同的骨干网在配对监督下，在非重叠的O-HAZE/NH-HAZE分割上也达到了20.71~dB~/~0.683~SSIM（这是一个转移性检查，而非竞争排名）。
cs.CV / 76 / 2606.29095
HorizonRelight: Relighting Long-horizon Videos Consistently via Diffusion Transformers
HorizonRelight：通过扩散变换器一致性地重光照长时间视频
Abstract
Diffusion-based video relighting enables controllable relighting from a single input video, but modern video diffusion backbones are trained on short clips and applied to long-horizon videos through chunked sliding-window inference, often causing temporal discontinuities at chunk boundaries. We address this by reframing long-horizon relighting as \emph{temporally conditioned latent domain translation}. Our framework enforces cross-chunk continuity by propagating target-domain latents across boundaries and makes this behavior learnable using \emph{masked target-domain self-conditioning}, training the model to continue from temporally masked propagated context. We further introduce \emph{warm-start prompting} with a relit prompt anchor from a controllable generative model, which establishes the initial target-domain state and creates a general interface for prompt-based relighting. Experiments on in-the-wild long-horizon videos show markedly improved temporal consistency, with chunk-boundary artifacts largely reduced and unwanted appearance changes across chunks greatly suppressed.
Chinese Translation
基于扩散的视频重光照技术能够从单个输入视频中实现可控的重光照，但现代视频扩散骨干网络是在短片段上训练的，并通过分块滑动窗口推理应用于长时间视频，这往往导致片段边界处的时间不连续性。我们通过将长时间重光照重新构建为 extit{时间条件的潜在领域转换}来解决这个问题。我们的框架通过在边界之间传播目标领域的潜在向量来强制执行跨块连续性，并使用 extit{掩码目标领域自条件}使这一行为可学习，训练模型从时间掩码传播的上下文中继续。我们进一步引入 extit{热启动提示}，使用来自可控生成模型的重光照提示锚点，建立初始目标领域状态，并为基于提示的重光照创建一个通用接口。在实际长时间视频上的实验显示，时间一致性显著改善，片段边界伪影大幅减少，跨片段的不必要外观变化也得到了有效抑制。
cs.CV / 77 / 2606.29097
TrafficAlign: Aligning Large Language Models for Traffic Scenario Generation
TrafficAlign：为交通场景生成对齐大型语言模型
Abstract
Recent research has investigated the use of large language models (LLMs) to generate traffic scenarios for autonomous driving. However, pretrained LLMs often fail to align with real-world traffic distributions. In this work, we present TrafficAlign, an automated framework that synthesizes traffic scenarios based on real-world driving videos, performs data validation, and aligns LLMs with the synthesized scenarios. The evaluation shows that traffic scenarios generated by TrafficAlign are highly effective, revealing up to 10.8% more collisions on average across three autonomous driving models than state-of-the-art methods. Furthermore, fine-tuning these driving models with TrafficAlign-generated scenarios significantly reduced collision rates by 36.1% compared with the original models. A qualitative study using traffic datasets from six geographically diverse regions shows that TrafficAlign-generated scenarios exhibit strong alignment with corresponding traffic distributions in these regions.
Chinese Translation
近期研究探讨了使用大型语言模型（LLMs）生成用于自动驾驶的交通场景。然而，预训练的LLMs往往无法与现实世界的交通分布对齐。在本研究中，我们提出了TrafficAlign，一个基于真实驾驶视频合成交通场景的自动化框架，执行数据验证，并将LLMs与合成的场景对齐。评估结果表明，TrafficAlign生成的交通场景非常有效，在三种自动驾驶模型中，平均显示出比最先进的方法多出10.8%的碰撞。此外，使用TrafficAlign生成的场景对这些驾驶模型进行微调，与原始模型相比，碰撞率显著降低了36.1%。通过使用来自六个地理多样化区域的交通数据集进行的定性研究表明，TrafficAlign生成的场景与这些区域相应的交通分布表现出强烈的对齐性。
cs.CV / 78 / 2606.29102
BTI-Net: Bidirectional Decoder-Level Task Interaction via Uncertainty-Aware Gating for Multi-Task Medical Image Analysis
BTI-Net：通过不确定性感知门控实现双向解码器级任务交互的多任务医学图像分析
Abstract
Jointly learning to segment and classify medical images demands cross-task synergy, yet encoder-sharing architectures limit decoder reconstruction to task-private representations, permanently discarding the boundary cues and semantic priors each branch could supply to the other. This work introduces BTI-Net, which establishes bidirectional communication at every decoder level through two parallel pathways via Task Interaction Modules (TIM). Spatial boundary context is gated into the classification branch, while global semantic priors multiplicatively modulate the decoder, with refined features propagating progressively from coarse semantics to fine boundary detail across all four decoder resolutions. Since cross-task interaction is not equally reliable for every input, Uncertainty Proxy Attention (UPA) gates each TIM output per instance and per level using three signals that capture cross-task alignment, scene complexity, and prediction confidence, without external annotations or additional inference passes. Experiments on three medical benchmarks spanning ultrasound, dermoscopy, and brain MRI demonstrate consistent improvements in segmentation IoU and classification accuracy over both encoder-sharing and decoder-interaction baselines. Ablation confirms adaptive gating contributes +2.36 IoU over fixed bidirectional interaction, and classification accuracy improves by up to +2.26 points over the strongest multi-task baseline. UPA's uncertainty proxies serve as reliable single-pass task-failure signals without the overhead of stochastic sampling. Code: https://github.com/C-loud-Nine/BTI-Net_MTL
Chinese Translation
联合学习医学图像的分割和分类需要跨任务协同，然而共享编码器架构限制了解码器重建为任务私有表示，永久性地丢弃了每个分支可以为其他分支提供的边界线索和语义先验。本文介绍了BTI-Net，它通过任务交互模块（Task Interaction Modules, TIM）在每个解码器级别建立双向通信，形成两条并行路径。空间边界上下文被门控到分类分支中，而全局语义先验则以乘法方式调制解码器，经过四个解码器分辨率逐步传播的精细特征，从粗糙语义到细致边界细节。由于跨任务交互对每个输入的可靠性并不相同，不确定性代理注意力（Uncertainty Proxy Attention, UPA）根据三种信号对每个TIM输出进行门控，这些信号捕捉跨任务对齐、场景复杂性和预测置信度，而无需外部注释或额外的推理过程。在涵盖超声、皮肤镜和脑部MRI的三个医学基准测试上的实验表明，相较于共享编码器和解码器交互基线，分割的IoU和分类准确性均有持续改善。消融实验确认自适应门控相较于固定双向交互贡献了+2.36的IoU，分类准确性提高了最多+2.26分，相较于最强的多任务基线。UPA的不确定性代理作为可靠的单次任务失败信号，无需随机采样的开销。代码链接： https://github.com/C-loud-Nine/BTI-Net_MTL
cs.CV / 79 / 2606.29106
A Deep Multiscale Neural Network for Accurate Neurological Disorder Detection from MRI Scans and Real-Time Web Deployment
一种深度多尺度神经网络用于从MRI扫描中准确检测神经系统疾病并实现实时网络部署
Abstract
Neurological disorders involve diverse pathologies of the brain and nervous system, making early and accurate detection essential. While many deep CNNs have been developed for MRI-based classification of neurological disorders, most are optimized for binary tasks and often fail to capture the multi-class features needed to distinguish subtle anatomical differences across conditions. This study proposes the Enhanced Neurological Disorder Detection Network (End-Net) for multi-class MRI classification of neurological disorders. End-Net includes 24 convolutional layers, beginning with convolutional blocks followed by 21 optimized inception modules. These modules extract multiscale features via parallel 1 x 1, 3 x 3, and factorized 5 x 5 convolutional branches, along with max pooling, enabling the model to capture complementary texture, edge, shape, and contextual information. A global average pooling head, compact fully connected classifier, and dropout reduce parameters, limit overfitting, and improve robustness. End-Net was evaluated on the Multi-Class Neurological Disorder dataset, comprising MRI scans from patients with Alzheimer's disease, brain tumors, multiple sclerosis, and healthy controls. Severe class imbalance was addressed by augmenting minority classes with WGAN-GP and randomly undersampling the majority class. The results show that End-Net outperforms existing architectures in both accuracy and generalization. The model is also integrated into an online system for real-time web-based inference and accessibility.
Chinese Translation
神经系统疾病涉及大脑和神经系统的多种病理，早期和准确的检测至关重要。虽然已经开发了许多基于MRI的深度卷积神经网络（CNN）用于神经系统疾病的分类，但大多数网络是针对二分类任务进行优化的，往往无法捕捉到区分不同病症之间微妙解剖差异所需的多类特征。本研究提出了一种增强型神经系统疾病检测网络（Enhanced Neurological Disorder Detection Network，End-Net），用于神经系统疾病的多类MRI分类。End-Net包含24个卷积层，首先是卷积块，随后是21个优化的Inception模块。这些模块通过并行的1 x 1、3 x 3和分解的5 x 5卷积分支提取多尺度特征，并结合最大池化，使模型能够捕捉互补的纹理、边缘、形状和上下文信息。全局平均池化头、紧凑的全连接分类器和dropout技术减少了参数，限制了过拟合，并提高了模型的鲁棒性。End-Net在多类神经系统疾病数据集上进行了评估，该数据集包含阿尔茨海默病、脑肿瘤、多发性硬化症患者及健康对照者的MRI扫描。通过使用WGAN-GP增强少数类样本，并随机下采样多数类样本，解决了严重的类别不平衡问题。结果表明，End-Net在准确性和泛化能力上均优于现有架构。该模型还集成到一个在线系统中，实现实时基于网络的推断和可访问性。
cs.CV / 80 / 2606.29134
Beyond Backscatter: AlphaEarth Land-Cover Priors for Rapid SAR Flood Segmentation Across Foundation Backbones
超越后向散射：AlphaEarth土地覆盖先验用于快速SAR洪水分割的基础骨干网络
Abstract
Rapid flood mapping is critical for emergency response, yet optical imagery is often unusable during major flooding and single-temporal SAR is ambiguous, since new inundation, permanent water, and other smooth surfaces produce similar backscatter. This study evaluates whether stable land-context priors can improve post-event SAR flood segmentation when a registered, seasonally matched pre-event acquisition is unavailable. Using the CONUS (Continental United States) subset of ImpactMesh-Flood, we compare four backbones spanning distinct pretraining regimes-a from-scratch CNN UNet, an ImageNet-pretrained UNet, the SAR-pretrained TerraMind Vision Transformer, and the optical-satellite-pretrained DINOv3 Vision Transformer-in SAR-only, SAR+DEM, and SAR+AlphaEarth configurations under an identical fusion design, training protocol, and event-stratified split. Models are selected on a validation flood event and evaluated separately on two held-out events, Hurricane Florence and the Louisiana floods, with three-seed reporting for auxiliary configurations. Both auxiliary priors improve over the observed SAR-only baselines across all backbones and test events. AlphaEarth exceeds DEM on the harder Florence event for every backbone and achieves the best Florence IoU, while DEM is competitive on Louisiana and produces the best result there. The seed analysis reveals a trade-off: DEM is more stable across initializations, whereas AlphaEarth offers higher peak performance and higher recall on the harder event. Cross-event differences track flood-class prevalence and similarity to the training distribution, underscoring the need for per-event evaluation. We reframe single-temporal SAR flood segmentation as an alignment between radar observations and stable land-surface priors, where learned and physical context offer complementary pathways to more reliable rapid flood mapping.
Chinese Translation
快速洪水制图对于应急响应至关重要，但在重大洪水期间，光学影像往往无法使用，而单时相SAR数据则存在歧义，因为新淹没区、永久水体和其他光滑表面会产生相似的后向散射。本研究评估了在缺乏注册的、季节匹配的事前采集数据时，稳定的土地背景先验是否能改善事件后SAR洪水分割。我们使用ImpactMesh-Flood的CONUS（美国大陆）子集，比较了四种基础骨干网络，涵盖不同的预训练模式——从头开始训练的CNN UNet、在ImageNet上预训练的UNet、在SAR上预训练的TerraMind Vision Transformer，以及在光学卫星上预训练的DINOv3 Vision Transformer——在SAR-only、SAR+DEM和SAR+AlphaEarth配置下，采用相同的融合设计、训练协议和事件分层拆分。模型在一个验证洪水事件上进行选择，并在两个保留事件（飓风佛罗伦萨和路易斯安那洪水）上分别进行评估，辅助配置的报告采用三次随机种子。两个辅助先验在所有基础骨干网络和测试事件上均优于观察到的SAR-only基线。在更具挑战性的佛罗伦萨事件中，AlphaEarth在每个基础骨干网络上均超过了DEM，并达到了最佳的佛罗伦萨IoU，而DEM在路易斯安那事件中表现竞争，并在该事件中产生了最佳结果。种子分析揭示了一种权衡：DEM在初始化时更为稳定，而AlphaEarth在更具挑战性的事件中提供了更高的峰值性能和更高的召回率。跨事件的差异与洪水类别的流行程度及与训练分布的相似性相关，强调了逐事件评估的必要性。我们将单时相SAR洪水分割重新框定为雷达观测与稳定土地表面先验之间的对齐，其中学习到的和物理的背景提供了更可靠的快速洪水制图的互补路径。
cs.CV / 81 / 2606.29136
CMTFormer: Marrying Transformer with Hierarchical Information Interaction for RGB-Event Object Detection
CMTFormer：将变换器与层次信息交互结合用于RGB-事件目标检测
Abstract
Event cameras capture sparse brightness changes with high temporal resolution and high dynamic range, compensating for the deficiencies of the conventional RGB frames. However, previous multi-modal fusion techniques typically fail to handle the inherent heterogeneity between RGB frames and event streams, thus easily leading to noise amplification or redundant feature integration during cross-modal fusion. In this paper, we propose a Cross-Modal information inTeraction transFormer, coined as CMTFormer, which hierarchically integrates RGB and event information to achieve efficient and stable multimodal collaboration. Specifically, we design a shallow-to-deep information interaction scheme. In the shallow stage, we present the Shallow Alignment Module (SAM) to achieve an efficient fusion of RGB and event low-level features, which mitigates attribute disparities and prevents noisy information. In the middle stage, we devise the Cross-modal Enhancement Module (CEM) that utilizes texture and edge information to produce mutually reinforced middle-level features. In the deep stage, we present the Learnable Deep Fusion Module (LDFM) which performs high-level information aggregation through learnable weights, thus enabling the network to adaptively fuse RGB and event clues. A Spatial Prior Module is further designed to utilize global spatial information to enhance localization accuracy. Extensive experiments are conducted on two prevalent event-based object detection benchmarks, i.e., DSEC-Detection and PKU-DAVIS-SOD. Our CMTFormer consistently surpasses the detection counterparts in both uni-modal and multi-modal settings, strongly demonstrating the effectiveness of our paradigm. Codes will be available upon publication.
Chinese Translation
事件相机以高时间分辨率和高动态范围捕捉稀疏的亮度变化，弥补了传统RGB帧的不足。然而，以往的多模态融合技术通常无法处理RGB帧和事件流之间固有的异质性，从而在跨模态融合过程中容易导致噪声放大或冗余特征整合。本文提出了一种跨模态信息交互变换器，称为CMTFormer，旨在层次性地整合RGB和事件信息，以实现高效且稳定的多模态协作。具体而言，我们设计了一种从浅到深的信息交互方案。在浅层阶段，我们提出了浅层对齐模块（Shallow Alignment Module, SAM），以实现RGB和事件低级特征的高效融合，减轻属性差异并防止噪声信息。在中间阶段，我们设计了跨模态增强模块（Cross-modal Enhancement Module, CEM），利用纹理和边缘信息生成相互增强的中级特征。在深层阶段，我们提出了可学习深度融合模块（Learnable Deep Fusion Module, LDFM），通过可学习权重进行高级信息聚合，从而使网络能够自适应地融合RGB和事件线索。此外，还设计了空间先验模块（Spatial Prior Module），利用全局空间信息提高定位精度。在两个流行的基于事件的目标检测基准（即DSEC-Detection和PKU-DAVIS-SOD）上进行了广泛的实验。我们的CMTFormer在单模态和多模态设置中均持续超越检测对手，强有力地证明了我们范式的有效性。代码将在发表时提供。
cs.CV / 82 / 2606.29148
GPC: Large-Scale Generative Pretraining for Transferable Motor Control
GPC：用于可转移运动控制的大规模生成预训练
Abstract
Developing controllers capable of completing a wide range of tasks in a natural and life-like manner is a key challenge in enabling practical applications of physics-based character animation. In this work, we introduce Generative Pretrained Controllers (GPC), which leverage tokenization and next-token modeling to create general-purpose, reusable generative controllers from large-scale motion datasets. Our framework utilizes end-to-end reinforcement learning to jointly optimize a "motion vocabulary", modeled via Finite Scalar Quantization (FSQ), along with a corresponding control policy that can map the discrete codes to physics-based controls. After the "codebook" has been learned, the underlying structure of this large vocabulary is modeled by training a GPT-style autoregressive transformer, leading to a powerful generative controller that generates controls for a physically simulated character by performing next-token prediction. Once the generative controller has been trained, we propose a suite of adaptation techniques for finetuning the controller for new downstream tasks. Our proposed framework greatly simplifies the training process compared to previous tokenized methods, and achieves a 99.98% success rate in reproducing a vast corpus of motion clips. The generative controller exhibits a variety of natural emergent behaviors, such as responsive behaviors to perturbations and recovery behaviors after falling. This results in highly robust general purpose controllers for a variety of downstream applications.
Chinese Translation
开发能够以自然和逼真的方式完成广泛任务的控制器是实现基于物理的人物动画实际应用的关键挑战。在本研究中，我们引入了生成预训练控制器（Generative Pretrained Controllers，GPC），该方法利用标记化和下一个标记建模，从大规模运动数据集中创建通用的、可重用的生成控制器。我们的框架利用端到端的强化学习共同优化一个通过有限标量量化（Finite Scalar Quantization，FSQ）建模的“运动词汇”，以及一个相应的控制策略，该策略能够将离散代码映射到基于物理的控制。在“代码本”学习完成后，通过训练一个GPT风格的自回归变换器来建模这个大词汇的基础结构，从而生成一个强大的生成控制器，该控制器通过执行下一个标记预测为物理模拟的人物生成控制。一旦生成控制器训练完成，我们提出了一套适应技术，以微调控制器以适应新的下游任务。与之前的标记化方法相比，我们提出的框架大大简化了训练过程，并在再现大量运动片段方面达到了99.98%的成功率。生成控制器展现出多种自然的涌现行为，例如对扰动的响应行为和摔倒后的恢复行为。这使得生成控制器在各种下游应用中具有高度的鲁棒性。
cs.CV / 83 / 2606.29162
Spatially Localized Image Degradation Embeddings for Image Quality Assessment
用于图像质量评估的空间局部化图像退化嵌入
Abstract
Self-supervised learning (SSL) currently drives state-of-the-art performance in no-reference image quality assessment (NR-IQA). However, standard SSL pipelines uniformly apply synthetic distortions across the entire image field, which can limit their sensitivity to spatially localized and co-occurring degradations encountered in real-world content. In this work, we empirically expose this representational blind spot across existing state-of-the-art encoders, demonstrating their reduced sensitivity to spatially bounded image degradations. To bridge this gap, we introduce Spatial Localized Image Degradation Embeddings for Image Quality Assessment (SLIDE-IQA). SLIDE-IQA employs a dual-branch Vision Transformer framework that injects spatially bounded degradations into a contrastive pretraining objective. To handle the spatial complexity of these degradations, we introduce a Threshold-Bounded Exclusion Mechanism, a representational design choice that resolves structural conflicts arising from spatially localized distortions to ensure the latent space respects both degradation type and spatial scale. Finally, we show that SLIDE-IQA's synthetic-only pretraining significantly improves sensitivity to localized distortions, while achieving competitive performance on NR-IQA benchmarks against existing SSL NR-IQA models.
Chinese Translation
自监督学习（SSL）目前在无参考图像质量评估（NR-IQA）中推动了最先进的性能。然而，标准的SSL流程在整个图像区域均匀应用合成失真，这可能限制了它们对现实内容中空间局部和共现退化的敏感性。在本研究中，我们实证揭示了现有最先进编码器的这一表征盲点，展示了它们对空间限制图像退化的敏感性降低。为了解决这一问题，我们提出了用于图像质量评估的空间局部化图像退化嵌入（SLIDE-IQA）。SLIDE-IQA采用双分支视觉变换器框架，将空间限制的退化注入对比预训练目标中。为了处理这些退化的空间复杂性，我们引入了一种阈值限制排除机制，这是一种表征设计选择，解决了由空间局部失真引起的结构冲突，以确保潜在空间尊重退化类型和空间尺度。最后，我们展示了SLIDE-IQA的仅合成预训练显著提高了对局部失真的敏感性，同时在与现有SSL NR-IQA模型的NR-IQA基准测试中取得了竞争性能。
cs.CV / 84 / 2606.29167
Articulating then Matching: Zero-Shot Shape Matching for Uncurated Data
先表达再匹配：针对未整理数据的零样本形状匹配
Abstract
Finding dense correspondences between 3D shapes is a fundamental yet unresolved challenge, especially in real-world environments. These environments present severe challenges, including the lack of time and sufficient samples for training, the prevalence of uncurated extreme-high resolution data with topological distortions, and the need to handle diverse 3D representations. In this paper, we present ATM, a zero-shot framework that requires no correspondence-specific training and robustly addresses these issues at once through an articulate-then-match paradigm. Rather than relying on intrinsic geometric properties, we leverage powerful pretrained vision foundation models and parametric shape priors to estimate parametric shape models from multi-view renderings, and systematically ground these estimations via multi-view geometric consistency. By mapping diverse inputs into a shared canonical parametric space, we inherently establish robust coarse correspondences that bypass topological noise, which are then refined into precise dense mappings via spectral refinement. Operating purely on test-time optimized parametric reconstructions, ATM requires no correspondence training data, is naturally immune to connectivity artifacts, and seamlessly handles diverse 3D modalities, including meshes, point clouds, and 3D Gaussians. Extensive experiments demonstrate that our method achieves strong results on non-isometric benchmarks (average geodesic errors of 2.4-TOPKIDS, 3.8-SMAL), reducing errors by 73% and 37% respectively compared to the baseline URSSM. Furthermore, it exhibits unprecedented robustness on in-the-wild raw scans of up to 200k vertices per shape while maintaining near-constant computation time and consistent superior accuracy.
Chinese Translation
在3D形状之间寻找密集对应关系是一个基本但尚未解决的挑战，尤其是在现实世界环境中。这些环境带来了严重的挑战，包括缺乏足够的时间和样本进行训练、未整理的极高分辨率数据的拓扑失真普遍存在，以及需要处理多样化的3D表示。在本文中，我们提出了ATM，一个零样本框架，无需特定于对应关系的训练，并通过一种先表达后匹配的范式稳健地解决这些问题。我们不依赖内在几何属性，而是利用强大的预训练视觉基础模型和参数形状先验，从多视角渲染中估计参数形状模型，并通过多视角几何一致性系统地验证这些估计。通过将多样化输入映射到共享的规范参数空间，我们固有地建立了稳健的粗略对应关系，绕过了拓扑噪声，随后通过谱细化将其精炼为精确的密集映射。ATM完全基于测试时优化的参数重建，要求不需要对应训练数据，自然免疫于连通性伪影，并无缝处理多样化的3D模态，包括网格、点云和3D高斯分布。大量实验表明，我们的方法在非同构基准上取得了强劲的结果（平均测地误差为2.4-TOPKIDS，3.8-SMAL），与基线URSSM相比，误差分别减少了73%和37%。此外，它在每个形状高达20万个顶点的野外原始扫描上展现出前所未有的鲁棒性，同时保持近乎恒定的计算时间和一致的优越准确性。
cs.CV / 85 / 2606.29181
Anomaly Factory 3D: A Modular Framework for Diverse Pseudo-Anomaly Synthesis in Unsupervised 3D Anomaly Detection
异常工厂3D：用于无监督3D异常检测的多模块框架以合成多样的伪异常
Abstract
Detecting and localizing defects in 3D point clouds is challenging because abnormal samples are scarce and diverse, while training is often limited to normal data. We propose Anomaly Factory 3D (AF3AD), a modular framework that synthesizes diverse pseudo-anomalies from normal point clouds to expand the training data for unsupervised 3D anomaly detection methods that rely on pseudo-anomalies. AF3AD uses a center-conditioned parametric deformation model defined in local PCA frames, with kernel-controlled spatial falloff, anisotropy, directional gating, and normal/tangential displacement fields, enabling a broad set of geometric defect presets. We demonstrate its ease-of-use and effectiveness by integrating AF3AD with an offset-prediction detector and a reconstruction-based anomaly detection method, showing that AF3AD transfers across detection paradigms. Experiments on AnomalyShapeNet and Real3D-AD show consistent improvements in object- and point-level detection and localization, supported by ablations on preset groups and robustness under noise. AF3AD is designed as a standalone synthesis tool to facilitate adoption across different 3D anomaly detection paradigms. Code is available at github.com/vpc-ccg/AF3AD.
Chinese Translation
在3D点云中检测和定位缺陷具有挑战性，因为异常样本稀缺且多样，而训练通常仅限于正常数据。我们提出了异常工厂3D（Anomaly Factory 3D，AF3AD），这是一个模块化框架，能够从正常点云中合成多样的伪异常，以扩展依赖伪异常的无监督3D异常检测方法的训练数据。AF3AD使用在局部主成分分析（PCA）框架中定义的中心条件参数变形模型，结合核控制的空间衰减、各向异性、方向门控以及法向/切向位移场，从而实现广泛的几何缺陷预设。我们通过将AF3AD与偏移预测检测器和基于重建的异常检测方法集成，展示了其易用性和有效性，表明AF3AD能够在不同检测范式之间迁移。在AnomalyShapeNet和Real3D-AD上的实验显示，在物体级和点级检测与定位方面均有一致的提升，且通过对预设组的消融实验和在噪声下的鲁棒性支持了这一结果。AF3AD被设计为一个独立的合成工具，以促进在不同3D异常检测范式中的应用。代码可在github.com/vpc-ccg/AF3AD获取。
cs.CV / 86 / 2606.29198
DTI: Dynamic Trajectory Initialization for Generative Face Video Super-Resolution
DTI：生成面部视频超分辨率的动态轨迹初始化
Abstract
As the most perceptually powerful Face Video Super-Resolution (FVSR) method, existing works in Generative FVSR (GFVSR) mainly exploit the generative prior of pretrained diffusion models. However, viewed as full generation, they suffer from fixed sampling and expensive inference costs if without large-scale auxiliary training. Furthermore, an excessive pursuit of generic perceptual metrics often results in low fidelity. To address these issues, we present Dynamic Trajectory Initialization (DTI) paradigm for GFVSR, which reformulates GFVSR as an input-driven directional restoration. With a novel enhancement-and-injection conditioning mechanism for pretrained DiT backbone, fidelity of our model has been significantly improved without compromising perceptual quality. To dynamically set the starting sampling point, we propose a Discriminative Guide (DG) trained via objective Signal-to-Noise Ratio (SNR) alignment. With only minor model adaptation and fine-tuning, our method achieves a SOTA overall performance across diverse metrics and benchmarks. An analysis of relationship between actual comprehensive quality and common metrics is also conducted, which demonstrates the perception-distortion trade-off and that the LPIPS is the most convincing metric in our case.
Chinese Translation
作为感知能力最强的面部视频超分辨率（FVSR）方法，现有的生成性FVSR（GFVSR）工作主要利用预训练扩散模型的生成先验。然而，作为完全生成的方法，如果没有大规模辅助训练，它们会遭遇固定采样和高昂推理成本的问题。此外，过度追求通用感知指标往往导致低保真度。为了解决这些问题，我们提出了GFVSR的动态轨迹初始化（DTI）范式，将GFVSR重新表述为一种输入驱动的方向性恢复。通过为预训练的DiT主干网络引入一种新颖的增强与注入条件机制，我们的模型在不妥协感知质量的情况下显著提高了保真度。为了动态设置起始采样点，我们提出了一种通过目标信噪比（SNR）对齐训练的判别引导（DG）。仅需少量模型适应和微调，我们的方法在各种指标和基准测试中实现了最先进的整体性能。我们还分析了实际综合质量与常见指标之间的关系，证明了感知-失真权衡，并且在我们的案例中，LPIPS是最具说服力的指标。
cs.CV / 87 / 2606.29208
Zero-Gated Language-conditioned Human Motion Prediction
零门控语言条件下的人体运动预测
Abstract
Pose histories provide the core kinematic evidence for 3D human motion prediction, but they lack explicit high-level semantic guidance. This paper introduces ZGL, a lightweight language-conditioned predictor that uses captions of the observed motion as a semantic prior while preserving a strong motion backbone as the main source of dynamics. We render only the observed poses, generate a one-sentence description with a vision-language model, encode the caption with a frozen CLIP-L text tower, and project it into a small set of conditioning tokens. These tokens are injected into a DCT-based spatial-temporal Transformer by compact crossattention adapters with zero gates: each adapter output is multiplied by a learnable gate initialized to zero, so the full network is numerically identical to the pose-only baseline at initialization and can learn to use language only when it reduces prediction error. On Human3.6M, ZGL improves overall MPJPE over representative motion-prediction baselines in our comparison. Results on CMUMocap further show that compact caption conditioning transfers to a second benchmark and provides a practical semantic cue for 3D human motion prediction.
Chinese Translation
姿态历史提供了3D人体运动预测的核心运动学证据，但缺乏明确的高层语义指导。本文介绍了ZGL，一种轻量级的语言条件预测器，它利用观察到的运动的描述作为语义先验，同时保留强大的运动骨干作为动态的主要来源。我们仅渲染观察到的姿态，使用视觉-语言模型生成一句描述，利用冻结的CLIP-L文本塔对描述进行编码，并将其投影到一小组条件令牌中。这些令牌通过紧凑的交叉注意适配器注入到基于DCT的时空Transformer中，适配器的每个输出都乘以一个初始化为零的可学习门，因此在初始化时整个网络在数值上与仅基于姿态的基线相同，并且只有在减少预测误差时才能学习使用语言。在Human3.6M数据集上，ZGL在我们的比较中提高了整体MPJPE，相较于代表性的运动预测基线。CMUMocap上的结果进一步表明，紧凑的描述条件可以转移到第二个基准，并为3D人体运动预测提供了实用的语义线索。
cs.CV / 88 / 2606.29230
Again-Pose: Anchor-Guided Adaptive Inter-Frame Motion Cues Propagating for High-quality Human Pose Reconstruction
Again-Pose：基于锚点引导的自适应帧间运动线索传播用于高质量人类姿态重建
Abstract
Reconstructing continuous 3D human poses from unconstrained videos is challenging, especially in extreme motion scenarios involving severe motion blur and occlusion. Current state-of-the-art methods typically rely on implicit temporal attention to aggregate features across frames. However, under severe visual degradation, input features often suffer from collapse, rendering them indistinguishable from noise. In such cases, implicit aggregation fails to distinguish valid signals, leading to catastrophic reconstruction errors. To address this robustness gap, we propose a simple yet effective framework called Anchor-guided adaptive inter-frame motion cues propagating (Again-Pose), reformulating pose estimation in degraded frames as a motion-guided recovery task. Instead of blindly smoothing features, we explicitly identify high-quality Anchor Frames based on feature saliency and propagate reliable kinematic cues to "inpaint" the poses of degraded intermediate frames. Specifically, a Dual-path Motion-aware Module captures fine-grained inter-frame dynamics, while a Difference-weighted Fusion Module adaptively propagates these cues to suppress drift. Extensive experiments on standard benchmarks (Human3.6M, 3DPW, PoseTrack) and the challenging FineDiving dataset demonstrate that Again-Pose significantly outperforms state-of-the-art methods in robustness and stability, effectively recovering plausible poses where other methods fail.
Chinese Translation
从无约束视频中重建连续的三维人类姿态是一项具有挑战性的任务，尤其是在涉及严重运动模糊和遮挡的极端运动场景中。目前的最先进方法通常依赖于隐式时间注意力在帧间聚合特征。然而，在严重的视觉退化情况下，输入特征往往会崩溃，使其与噪声难以区分。在这种情况下，隐式聚合无法区分有效信号，导致灾难性的重建错误。为了解决这一鲁棒性缺口，我们提出了一个简单而有效的框架，称为基于锚点引导的自适应帧间运动线索传播（Again-Pose），将退化帧中的姿态估计重新表述为一个运动引导的恢复任务。我们不仅仅是盲目平滑特征，而是基于特征显著性明确识别高质量的锚点帧，并传播可靠的运动线索以“修复”退化中间帧的姿态。具体而言，双路径运动感知模块捕捉细粒度的帧间动态，而差异加权融合模块自适应地传播这些线索以抑制漂移。在标准基准（Human3.6M、3DPW、PoseTrack）和具有挑战性的FineDiving数据集上的大量实验表明，Again-Pose在鲁棒性和稳定性方面显著优于最先进的方法，有效恢复了其他方法失败的合理姿态。
cs.CV / 89 / 2606.29232
When Does Synthetic CT Transfer? A Label-Free Donor/Host Diagnostic for Medical Vision-Language Model Routing on Real Lung CT
合成CT何时转移？一种无标签的供体/宿主诊断方法用于真实肺部CT的医学视觉-语言模型路由
Abstract
A synthetic measurement of model competence is useful only if it survives the move to real data, yet the real labels that would verify it are exactly what medical imaging lacks. We ask whether transfer can be predicted in advance, label-free, and answer with a mechanism: on synthetic digital twins, competence that is donor-driven (a property of the transplanted nodule) survives the synthetic to real change of host, while host-driven competence (a property of the surrounding anatomy) need not. We test this on three lung CT vision-language tasks chosen to span that axis, across five public VLMs, four guidance conditions, and seven real datasets. The prediction holds in every case: presence and size orderings transfer (R2 >= 0.96), lobe does not; the split survives leave-source-out calibration, and the diagnostic names that boundary before any real label. TrialCouncil, a training-free council calibrated only on synthetic CT, confirms it by matching the best fixed model exactly where transfer is predicted. The contribution is not the router but the finding that transfer itself is predictable, label-free, from synthetic data alone.
Chinese Translation
合成模型能力的测量只有在其能够迁移到真实数据时才有用，然而，验证这一点的真实标签恰恰是医学影像所缺乏的。我们探讨是否可以在无标签的情况下提前预测转移，并提出一种机制：在合成数字双胞胎上，供体驱动的能力（即移植结节的特性）能够在宿主的合成到真实变化中存活，而宿主驱动的能力（即周围解剖结构的特性）则不一定能够存活。我们在三个肺部CT视觉-语言任务上进行测试，这些任务选择以覆盖该轴，涉及五个公共视觉-语言模型（VLMs）、四种指导条件和七个真实数据集。预测在每种情况下都成立：存在性和大小排序能够转移（R2 >= 0.96），而叶片则不能；这一分割在留源校准中得以保留，并且在任何真实标签之前的边界上都有诊断名称。TrialCouncil，一个仅在合成CT上进行校准的无训练委员会，通过在预测转移的地方精确匹配最佳固定模型来确认这一点。我们的贡献不在于路由器，而在于发现转移本身是可以从合成数据中无标签地预测的。
cs.CV / 90 / 2606.29255
Confidence-feedback-weighted graph matching network: online-offline laser-induced damage site matching under complex interference
置信反馈加权图匹配网络：复杂干扰下的在线-离线激光诱导损伤位置匹配
Abstract
Online inspection images of final optics in high-power laser facilities contain pseudo-damage sites that closely resemble true damage sites. Determining the authenticity of online-detected sites is therefore difficult and requires accurate matching to offline ground-truth sites. However, this matching remains highly challenging due to limited match-discriminative features, local geometric distortions, and numerous distractor sites. Existing matching models mainly suppress distractors implicitly through loss-function supervision. We propose a confidence-feedback-weighted graph matching network that requires only damage-site centroid coordinates as input. It estimates node matchability confidence from each round of matching scores and feeds it back as a reliability weight to guide subsequent edge-feature aggregation, thereby suppressing distractor propagation and enhancing cross-graph discriminability. Within this framework, a geometric consistency constraint calibrates spurious high-confidence matchability estimates, while a hard-example mining loss improves discrimination between structurally similar sites. Experiments on our Complex-Scene dataset show that the proposed method achieves a matching F1-score of 96.36$\%$ with robust and efficient performance.
Chinese Translation
高功率激光设施中最终光学元件的在线检测图像包含与真实损伤位置极为相似的伪损伤位置。因此，确定在线检测位置的真实性变得困难，并且需要与离线真实位置进行准确匹配。然而，由于匹配判别特征有限、局部几何失真以及众多干扰位置，这一匹配过程依然极具挑战性。现有的匹配模型主要通过损失函数监督隐式地抑制干扰项。我们提出了一种置信反馈加权图匹配网络，仅需损伤位置的质心坐标作为输入。该网络从每轮匹配得分中估计节点匹配置信度，并将其反馈作为可靠性权重，以指导后续的边特征聚合，从而抑制干扰项传播并增强跨图的判别能力。在此框架内，几何一致性约束校准虚假的高置信度匹配性估计，而困难样本挖掘损失则提高了结构相似位置之间的判别能力。在我们的复杂场景数据集上的实验表明，所提出的方法以稳健且高效的表现达到了96.36%的匹配F1分数。
cs.CV / 91 / 2606.29267
Enhancing Part-Level Point Grounding for Any Open-Source MLLMs
增强任何开源多模态大语言模型的部件级点定位
Abstract
Visual grounding aims to associate free-form textual queries with specific regions in an image. While recent Multimodal Large Language Models (MLLMs) have demonstrated promising capabilities in this domain, they primarily excel at object-level grounding and often struggle with part-level grounding-an essential requirement for fine-grained tasks such as robotic manipulation. In this work, we introduce a general approach that equips any open-source MLLMs with accurate 2D part-level point grounding, offering a more direct alternative to conventional grounding representations. Our method leverages the attention mechanisms inherently present in MLLMs. By synthesizing text-conditioned, grounding-aware queries within intermediate layers via the proposed Q-Synth Module, we capture target-relevant attention patterns and refine them with a lightweight Attention-to-Point Decoder, which converts these patterns into a point-centric heatmap for final prediction. Notably, all original MLLM parameters are frozen, ensuring full preservation of their pre-trained capabilities. Experiments show that our design consistently improves part-level grounding accuracy across datasets and can be seamlessly integrated into any open-source MLLMs.
Chinese Translation
视觉定位旨在将自由形式的文本查询与图像中的特定区域关联起来。尽管近期的多模态大语言模型（MLLMs）在这一领域展现了良好的能力，但它们主要在物体级定位方面表现出色，而在部件级定位方面常常面临挑战，这对于诸如机器人操作等细粒度任务至关重要。在本研究中，我们提出了一种通用方法，使任何开源的MLLMs具备准确的二维部件级点定位，提供了一种比传统定位表示更直接的替代方案。我们的方法利用了MLLMs中固有的注意力机制。通过在中间层合成文本条件的、关注定位的查询，利用所提出的Q-Synth模块，我们捕捉到与目标相关的注意力模式，并通过轻量级的Attention-to-Point解码器对其进行精炼，将这些模式转换为用于最终预测的点中心热图。值得注意的是，所有原始MLLM参数均被冻结，确保其预训练能力的完全保留。实验表明，我们的设计在各数据集上持续提高了部件级定位的准确性，并且可以无缝集成到任何开源MLLM中。
cs.CV / 92 / 2606.29282
ScaleErasure: Inference-Time Minimal Intervention for Precise Concept Erasure in Next-Scale Autoregressive Image Generation
ScaleErasure：在推理时进行最小干预以实现下一尺度自回归图像生成中的精确概念抹除
Abstract
Concept erasure aims to prevent image generative models from producing unsafe content while preserving their general generative capability. Meanwhile, next-scale autoregressive (AR) image generation has recently emerged as a new generative paradigm characterized by next-scale prediction, for which concept erasure remains largely unexplored. In this paradigm, semantic information is highly compressed at early scales, leading to severe entanglement between unsafe and unrelated semantics. In this paper, we propose ScaleErasure, an inference-time concept erasure method that performs minimal intervention. ScaleErasure precisely selects and guides predicted logits that are most relevant to the unsafe concept, thereby enabling effective erasure under severe semantic entanglement. Specifically, ScaleErasure performs two additional forward passes conditioned on the unsafe concept and the corresponding safe concept, and leverages their outputs to guide the target logits away from unsafe concepts toward safe concepts. To enable precise and minimal intervention, logits selection and guidance are conducted across three dimensions: scales, tokens, and bit channels. Experiments demonstrate that ScaleErasure outperforms adapted baselines in the next-scale AR paradigm, achieving more precise concept erasure while largely preserving general generative capability. The code is available at https://github.com/coziiizz/ScaleErasure.
Chinese Translation
概念抹除旨在防止图像生成模型产生不安全内容，同时保持其一般生成能力。与此同时，下一尺度自回归（AR）图像生成作为一种新的生成范式，最近崭露头角，其特征在于下一尺度预测，而在这一范式中，概念抹除仍然基本未被探索。在这一范式中，语义信息在早期尺度上高度压缩，导致不安全语义与无关语义之间严重纠缠。本文提出了ScaleErasure，一种在推理时进行最小干预的概念抹除方法。ScaleErasure精确选择并引导与不安全概念最相关的预测logits，从而在严重的语义纠缠下实现有效的抹除。具体而言，ScaleErasure在不安全概念和相应安全概念的条件下执行两个额外的前向传播，并利用它们的输出引导目标logits远离不安全概念朝向安全概念。为了实现精确和最小的干预，logits的选择和引导跨越三个维度进行：尺度、标记和比特通道。实验表明，ScaleErasure在下一尺度AR范式中优于适应性基线，实现了更精确的概念抹除，同时在很大程度上保持了总体生成能力。代码可在 https://github.com/coziiizz/ScaleErasure 获取。
cs.CV / 93 / 2606.29286
ASTAD: Asymmetric Style Transfer for Synthetic-to-Real Adaptation in Autonomous Driving
ASTAD：用于自主驾驶的合成到真实适应的非对称风格迁移
Abstract
Synthetic data mitigates the data scarcity problem in autonomous driving perception. However, the synthetic-to-real gap leads to performance degradation, hindering real-world model generalization. Although current methods leverage diffusion models for photorealistic style transfer to bridge this gap, they critically ignore a practical asymmetry: while synthetic data possesses perfect pixel-level annotations, real-world style reference images generally lack corresponding labels. Consequently, existing methods relying on symmetric semantic guidance suffer from either prohibitive annotation costs or severe semantic misalignment. To address this dilemma, we formally propose a novel task: Asymmetric Style Transfer for Autonomous Driving (ASTAD), which requires semantically consistent transfer using only labeled synthetic content and unlabeled real-world references. We further introduce the ASTModel, a training-free two-stage framework designed to bridge this domain gap under asymmetric constraints. ASTModel first extracts a coarse semantic prior from the unlabeled target, followed by dynamic prior refinement and class-consistent style injection during the denoising process. Extensive experiments demonstrate that ASTModel significantly outperforms existing methods in downstream perception utility and structural fidelity, while offering a 3.2$\times$ inference speedup. This work aligns synthetic-to-real adaptation with practical constraints, holding the potential to accelerate the scalable deployment of robust autonomous driving systems. Code: https://github.com/Dingyi-Yao/ASTAD.
Chinese Translation
合成数据缓解了自主驾驶感知中的数据稀缺问题。然而，合成到真实的差距导致性能下降，妨碍了模型在真实世界中的泛化。尽管当前方法利用扩散模型进行逼真的风格迁移以弥合这一差距，但它们严重忽视了一个实际的不对称性：合成数据具有完美的像素级注释，而真实世界的风格参考图像通常缺乏相应的标签。因此，依赖对称语义指导的现有方法面临着高昂的注释成本或严重的语义错位。为了解决这一困境，我们正式提出了一项新任务：自主驾驶的非对称风格迁移（ASTAD），该任务要求仅使用标记的合成内容和未标记的真实世界参考进行语义一致的迁移。我们进一步引入了ASTModel，一个无训练的两阶段框架，旨在在不对称约束下弥合这一领域差距。ASTModel首先从未标记的目标中提取粗略的语义先验，然后在去噪过程中进行动态先验细化和类别一致的风格注入。大量实验表明，ASTModel在下游感知效用和结构保真度方面显著优于现有方法，同时提供了3.2倍的推理加速。该工作将合成到真实的适应与实际约束对齐，具有加速强大自主驾驶系统可扩展部署的潜力。代码：https://github.com/Dingyi-Yao/ASTAD。
cs.CV / 94 / 2606.29301
Pointer-CAD v2: Plan-Then-Construct CAD Generation with Dimension-Aware Parametric Precision
Pointer-CAD v2：基于计划-再构建的维度感知参数化精度CAD生成
Abstract
Computer-aided design (CAD) plays a fundamental role in modern manufacturing by providing the high precision required for industrial production. Recent large language model based approaches formulate CAD generation as a sequence prediction problem and have achieved promising results. However, existing methods and evaluation protocols primarily emphasize visual similarity, while overlooking precise geometric parameters and correct metric scale. Small numerical deviations that are negligible at the shape-level may still violate industrial tolerance requirements, a problem further compounded by current autoregressive paradigms that utilize command sequence representations, aggressively quantize numerical parameters to ease LLM prediction. In this work, we present Pointer-CAD v2. Compared with v1 (arXiv:2603.04337), this version directly predicts continuous values, bypassing the need for quantized numerical parameters and thereby eliminating quantization errors. Specifically, we propose a unified framework that decouples parameter reasoning from geometric construction through a Plan-Then-Construct paradigm. Our method first produces a structured design plan with explicit metric scale parameters. These parameters are organized into a dictionary and directly referenced during sequence generation via a pointer mechanism, eliminating discretization errors and ensuring dimensionally consistent execution. In addition, we construct a new large-scale dataset with plan-level annotation and introduce three hierarchical geometry accuracy metrics to evaluate parametric fidelity at the vertex, edge, and face levels. Extensive experiments demonstrate that Pointer-CAD v2 consistently outperforms existing baselines and achieves substantial improvements in geometric accuracy, enabling reliable CAD generation for precision-critical engineering applications.
Chinese Translation
计算机辅助设计（CAD）在现代制造中发挥着基础性作用，提供了工业生产所需的高精度。最近基于大型语言模型的方法将CAD生成形式化为序列预测问题，并取得了令人鼓舞的成果。然而，现有的方法和评估协议主要强调视觉相似性，而忽视了精确的几何参数和正确的度量尺度。在形状层面上微不足道的小数值偏差仍可能违反工业公差要求，这一问题在当前利用命令序列表示的自回归范式中进一步加剧，这些范式为了简化大型语言模型的预测，激进地量化数值参数。在本研究中，我们提出了Pointer-CAD v2。与v1（arXiv:2603.04337）相比，该版本直接预测连续值，绕过了对量化数值参数的需求，从而消除了量化误差。具体而言，我们提出了一个统一框架，通过计划-再构建的范式将参数推理与几何构造解耦。我们的方法首先生成一个具有明确度量尺度参数的结构化设计计划。这些参数被组织成字典，并在序列生成过程中通过指针机制直接引用，消除了离散化误差，确保了维度一致的执行。此外，我们构建了一个新的大规模数据集，具有计划级注释，并引入了三个层次的几何精度指标，以评估在顶点、边和面级别上的参数保真度。大量实验表明，Pointer-CAD v2始终优于现有基线，并在几何精度上取得了显著改善，使得在对精度要求严格的工程应用中能够可靠地生成CAD。
cs.CV / 95 / 2606.29303
Occlusion-Robust Multi-Object Decoupling for Physics-Based Interaction
基于物理交互的遮挡鲁棒多物体解耦
Abstract
We propose a mask-free method for lossless multi-object 3D reconstruction from sparse and occluded real-world views, enabling physically plausible interaction via Material Point Method (MPM) simulation. Our key insight is that object coupling stems from occlusion and limited viewpoints, which we address by formulating multi-object decoupling as a sparse-view reconstruction problem. Using 3D Gaussian Splatting as base representation, we first obtain coarse instance partitions with a SAM2-trained segmentation field. Rather than relying on masks, we reconstruct fragmented geometries by leveraging a joint Score Distillation Sampling (SDS) process, which integrates reference-view supervision with novel-view synthesis guided by 2D and 3D diffusion priors to enforce both texture fidelity and 3D consistency. Furthermore, we incorporate geometry-aware priors such as intra-object and inter-object similarity to regularize geometric reasoning. Experimental results demonstrate that our method produces complete, simulation-ready 3D objects without requiring manual masks, enabling realistic dynamic interactions on both synthetic and real-world datasets.
Chinese Translation
我们提出了一种无掩膜的方法，用于从稀疏和遮挡的真实世界视图中进行无损多物体三维重建，从而实现通过材料点方法（Material Point Method, MPM）模拟的物理上合理的交互。我们的关键见解是，物体耦合源于遮挡和有限的视角，我们通过将多物体解耦表述为稀疏视图重建问题来解决这一问题。以三维高斯点云（3D Gaussian Splatting）作为基本表示，我们首先利用经过SAM2训练的分割场获得粗略的实例分区。我们不依赖掩膜，而是通过利用联合评分蒸馏采样（Score Distillation Sampling, SDS）过程重建碎片几何体，该过程将参考视图监督与由二维和三维扩散先验指导的新视图合成相结合，以确保纹理保真度和三维一致性。此外，我们结合了几何感知先验，如物体内部和物体之间的相似性，以规范几何推理。实验结果表明，我们的方法能够生成完整的、适合模拟的三维物体，而无需手动掩膜，从而在合成和真实世界数据集上实现逼真的动态交互。
cs.CV / 96 / 2606.29308
MirrorPPR: Exemplar-Based Portrait Photo Retouching
MirrorPPR：基于示例的肖像照片修饰
Abstract
While text-guided image editing has made remarkable progress, it remains limited in structural portrait retouching. Textual descriptions struggle to convey fine-grained changes to facial features and body proportions. To address this gap, we introduce Exemplar-Based Portrait Photo Retouching, where the model is given an exemplar pair and tasked with inferring and applying the same retouching operations to a new query image. Existing exemplar-based editing methods primarily focus on tasks with pronounced visual transformations. In contrast, structural portrait retouching involves extremely delicate and localized modifications, making accurate extraction and transfer of these edits challenging. To tackle this, we propose MirrorPPR, a novel framework designed to capture and transfer subtle structural retouching operations. Our method uses a Retouching Operation Extractor to capture the subtle differences from the exemplar pair. The extracted representations are then injected into a pre-trained Diffusion Transformer (DiT) through a connector and Low-Rank Adaptation (LoRA) modules. Furthermore, constructing perfectly aligned cross-identity training pairs is severely hindered by operation misalignment. To overcome this, we propose an advanced data self-augmentation paradigm that ensures strictly aligned retouching operations. To alleviate data scarcity and support this novel task, we introduce MirrorPPR47M, a large-scale dataset with over 47 million retouched pairs. By structuring the dataset into simulated and professional subsets, we enable progressive curriculum learning to smoothly optimize the network. Extensive experiments demonstrate that MirrorPPR significantly outperforms existing baselines in both retouching quality and identity preservation. The project page is available at https://sjtu-deng-lab.github.io/MirrorPPR.
Chinese Translation
尽管文本引导的图像编辑取得了显著进展，但在结构性肖像修饰方面仍然存在局限。文本描述难以传达对面部特征和身体比例的细微变化。为了解决这一问题，我们提出了基于示例的肖像照片修饰，其中模型接收一对示例，并被要求推断并将相同的修饰操作应用于新的查询图像。现有的基于示例的编辑方法主要集中在具有明显视觉变换的任务上。相比之下，结构性肖像修饰涉及极其细致和局部的修改，使得准确提取和转移这些编辑变得具有挑战性。为此，我们提出了MirrorPPR，这是一种新颖的框架，旨在捕捉和转移细微的结构修饰操作。我们的方法使用修饰操作提取器来捕捉示例对之间的细微差异。提取的表示随后通过连接器和低秩适应（Low-Rank Adaptation, LoRA）模块注入到预训练的扩散变换器（Diffusion Transformer, DiT）中。此外，构建完美对齐的跨身份训练对受到操作不对齐的严重阻碍。为了解决这个问题，我们提出了一种先进的数据自我增强范式，以确保严格对齐的修饰操作。为了缓解数据稀缺并支持这一新任务，我们引入了MirrorPPR47M，这是一个包含超过4700万对修饰图像的大规模数据集。通过将数据集结构化为模拟和专业子集，我们能够实现渐进式课程学习，以平滑优化网络。大量实验表明，MirrorPPR在修饰质量和身份保留方面显著优于现有基准。项目页面可访问 https://sjtu-deng-lab.github.io/MirrorPPR。
cs.CV / 97 / 2606.29314
D$^{2}$R$^{2}$OSR: Degradation-Disentangled Representation for Real-World Omnidirectional Image Super-Resolution
D$^{2}$R$^{2}$OSR：用于现实世界全景图像超分辨率的降解解耦表示
Abstract
With the growing demand for immersive visual experiences, high-quality omnidirectional images (ODIs) have become increasingly important. However, limitations in imaging devices and transmission bandwidth often lead to low-resolution ODIs, hindering the rendering of fine-grained 360{\deg} details, especially in the presence of real-world degradations and geometric distortions. Existing real-world super-resolution (Real-SR) methods are inadequate for ODIs, as their degradation models fail to account for the complex imaging pipeline involving fisheye capture and Equirectangular Projection (ERP), introducing severe aliasing and projection-specific distortions. To address these challenges, we propose D$^{2}$R$^{2}$OSR, a Degradation-Disentangled Representation framework for Real-world Omnidirectional image Super-Resolution. D$^{2}$R$^{2}$OSR explicitly models degradations arising from both fisheye imaging and ERP projection, guided by two key insights: (1) projection priors play a critical role in shaping real-world degradations, and (2) human perception in immersive environments is inherently viewpoint-centric. Accordingly, we introduce a Perspective Projection Representation (PPR) operating alongside the ERP branch to capture viewpoint-aware features, together with a Degradation-Specific Module (DSM) that jointly models ERP-induced geometric distortions and PPR-specific real-world degradations. Extensive experiments demonstrate that D$^{2}$R$^{2}$OSR achieves state-of-the-art performance and produces visually compelling, high-fidelity omnidirectional Real-SR results while maintaining favorable computational efficiency for low-resource deployment.
Chinese Translation
随着对沉浸式视觉体验需求的不断增长，高质量的全景图像（ODIs）变得愈发重要。然而，成像设备和传输带宽的限制常常导致低分辨率的全景图像，妨碍了细致的360°细节的渲染，尤其是在现实世界降解和几何失真的情况下。现有的现实世界超分辨率（Real-SR）方法对于全景图像而言并不充分，因为它们的降解模型未能考虑涉及鱼眼捕捉和等距投影（ERP）的复杂成像流程，从而引入了严重的混叠和特定投影的失真。为了解决这些挑战，我们提出了D$^{2}$R$^{2}$OSR，一个用于现实世界全景图像超分辨率的降解解耦表示框架。D$^{2}$R$^{2}$OSR明确建模了来自鱼眼成像和ERP投影的降解，基于两个关键见解： (1) 投影先验在塑造现实世界降解中起着关键作用，(2) 沉浸式环境中的人类感知本质上是视点中心的。因此，我们引入了一个与ERP分支并行工作的透视投影表示（PPR），以捕捉视点感知特征，并结合一个降解特定模块（DSM），共同建模ERP引起的几何失真和PPR特定的现实世界降解。大量实验表明，D$^{2}$R$^{2}$OSR实现了最先进的性能，并生成了视觉上引人注目、高保真的全景现实世界超分辨率结果，同时在低资源部署中保持了良好的计算效率。
cs.CV / 98 / 2606.29319
FDM-MFVT: Few-step Sampling Diffusion Model for Mask-Free Virtual Try-On
FDM-MFVT：一种用于无掩码虚拟试穿的少步采样扩散模型
Abstract
Image-based Virtual Try-On (IVTON) has greatly advanced through diffusion models, yet existing methods require many sampling steps and depend on masks with costly auxiliary networks. In addition, the absence of large-scale mask-free paired datasets further limits the development of mask-free IVTON. We propose FDM-MFVT, a few-step diffusion model for mask-free IVTON, integrating an Outfit-aware Noise Optimization Module (OANO) and an Instruction-driven Try-on Module (IDT) to enhance efficiency and flexibility.The OANO module initializes the alignment space with noise using the input image and only needs 6 steps to generate a higher-fidelity try-on image compared to 30 steps.The IDT module uses virtual try-on prompts and efficient adaptation to generate high-quality results from garment and person images alone. We further introduce MFVT, a 30,000-pair mask-free IVTON dataset. Experiments show that FDM-MFVT achieves superior quantitative and qualitative results with fewer inference steps than mask-based and mask-free baseline methods.
Chinese Translation
基于图像的虚拟试穿（IVTON）通过扩散模型得到了极大的发展，然而现有方法需要多个采样步骤，并依赖于昂贵的辅助网络进行掩码处理。此外，缺乏大规模的无掩码配对数据集进一步限制了无掩码IVTON的发展。我们提出了FDM-MFVT，一种用于无掩码IVTON的少步扩散模型，集成了服装感知噪声优化模块（OANO）和指令驱动试穿模块（IDT），以提高效率和灵活性。OANO模块使用输入图像初始化带噪声的对齐空间，仅需6个步骤即可生成比30个步骤更高保真度的试穿图像。IDT模块利用虚拟试穿提示和高效适配，从服装和人物图像中生成高质量结果。我们进一步引入了MFVT，一个包含30,000对的无掩码IVTON数据集。实验表明，FDM-MFVT在推理步骤更少的情况下，取得了优于基于掩码和无掩码基线方法的定量和定性结果。
cs.CV / 99 / 2606.29329
RAGA: Real Time Ray Traced Gaussian Shadow Casting for 3DGS Avatar-Scene Interaction
RAGA：实时光线追踪高斯阴影投射用于3D高斯点云（3DGS）头像-场景交互
Abstract
We study the problem of physically plausible shadow casting when animating 3D Gaussian Splatting (3DGS) avatars, either individually or in multi-avatar and object-interaction scenarios, within existing 3DGS scenes. In contrast to prior methods that rely on binary hit tests and mesh-based shadow casters, our method performs shadow computation entirely in Gaussian space, without requiring any mesh reconstruction. We introduce RAGA, a Ray-Traced Gaussian Shadow Casting formulation based on exact ray-Gaussian line integrals. For each occluding Gaussian, we integrate the opacity profile along the shadow ray and normalize by the theoretical maximum integral, producing a weight that captures how the ray traverses the occluder rather than merely whether an intersection occurred. To reduce temporal variance from clothing deformations in animated avatars, we further introduce an avatar proxy representation that stabilizes shadow casting while preserving visual fidelity. We implement RAGA using custom CUDA kernels integrated with the NVIDIA OptiX framework; as such, our shadow tracer runs at rates of about 50 FPS. We evaluate on single-avatar, multi-avatar, and avatar-object interaction scenarios across multiple datasets, demonstrating substantially improved shadow realism, temporal stability, and scene coherence. Our project page is available at https://miraymen.github.io/raga/.
Chinese Translation
我们研究了在现有3D高斯点云（3DGS）场景中，单独或在多头像和物体交互场景下为3D高斯点云（3DGS）头像动画时，物理上合理的阴影投射问题。与依赖于二元碰撞检测和基于网格的阴影投射方法不同，我们的方法完全在高斯空间中进行阴影计算，而无需任何网格重建。我们提出了RAGA，一种基于精确光线-高斯线积分的光线追踪高斯阴影投射公式。对于每个遮挡高斯体，我们沿着阴影光线积分不透明度曲线，并通过理论最大积分进行归一化，从而生成一个权重，该权重捕捉光线如何穿过遮挡物，而不仅仅是是否发生了交点。为了减少动画头像中服装变形带来的时间方差，我们进一步引入了一种头像代理表示，稳定阴影投射的同时保持视觉保真度。我们使用与NVIDIA OptiX框架集成的自定义CUDA内核实现了RAGA；因此，我们的阴影追踪器以约50帧每秒的速度运行。我们在单头像、多头像和头像-物体交互场景中对多个数据集进行了评估，展示了显著改善的阴影真实感、时间稳定性和场景一致性。我们的项目页面可在 https://miraymen.github.io/raga/ 查阅。
cs.CV / 100 / 2606.29333
HiReFF: High-Resolution Feedforward Human Reconstruction from Uncalibrated Sparse-View Video
HiReFF：基于未标定稀疏视角视频的高分辨率前馈人类重建
Abstract
Uncalibrated volumetric video streaming for human reconstruction is essential for holographic communication and AR/VR, yet remains challenging due to the need for temporal consistency and computational efficiency from sparse-view inputs. Existing methods rely on per-scene optimization or calibrated cameras, while recent feed-forward models are limited to low-resolution (0.5K) single-frame synthesis. We present HiReFF, a feed-forward method for 2K-resolution 360{\deg} human video reconstruction from uncalibrated sparse-view videos. Our framework decomposes the problem into two key tasks: foreground 3D Gaussian reconstruction from sparse-view videos (four views separated by 90{\deg}) and computationally efficient high-resolution synthesis. To enable the former, we propose Scale-synchronized Camera Calibration to resolve scale ambiguity for multi-view supervision, and Gaussian-wise Foreground Masking to reconstruct clean foregrounds by modulating Gaussian parameters. For efficient high-resolution synthesis, our High-resolution Side-tuning achieves 2K rendering by augmenting the Gaussian head with supplementary features while keeping the backbone at 0.5K, drastically reducing computational overhead. Experiments demonstrate that HiReFF significantly outperforms existing methods in high-resolution streaming volumetric video reconstruction. https://iridescentjiang.github.io/HiReFF
Chinese Translation
未标定的体积视频流对于人类重建在全息通信和增强现实/虚拟现实中至关重要，但由于需要从稀疏视角输入中实现时间一致性和计算效率，仍然面临挑战。现有方法依赖于每个场景的优化或标定相机，而最近的前馈模型仅限于低分辨率（0.5K）单帧合成。我们提出了HiReFF，一种从未标定稀疏视角视频中进行2K分辨率360°人类视频重建的前馈方法。我们的框架将问题分解为两个关键任务：从稀疏视角视频（四个视角相隔90°）中进行前景3D高斯重建，以及计算高效的高分辨率合成。为了实现前者，我们提出了尺度同步相机标定，以解决多视角监督中的尺度模糊问题，并通过调节高斯参数进行高斯前景遮罩，以重建干净的前景。为了实现高效的高分辨率合成，我们的高分辨率侧调节通过增强高斯头部的补充特征，同时保持主干在0.5K，从而实现2K渲染，显著减少计算开销。实验表明，HiReFF在高分辨率流体体积视频重建方面显著优于现有方法。
cs.CV / 101 / 2606.29334
Multi-scale Object-Aware Gaze Estimation via Geometric Reasoning
基于几何推理的多尺度对象感知注视估计
Abstract
Gaze target estimation aims to predict the semantic object an observer fixates upon within an image, a task deeply rooted in the object-oriented nature of human gaze. Observers tend to select a specific semantic entity as the attentional target, rather than responding randomly across arbitrary regions of the image. However, existing methods typically model this task as a direct mapping from global features to gaze heatmaps, essentially treating it as a pixel-level regression problem. This approach fails to explicitly represent the gazed object as a distinct entity, making it difficult to produce stable and semantically consistent predictions in complex scenes. To address this, we propose a two-stage gaze estimation framework guided by object semantics, reformulating gaze target estimation as a hierarchical reasoning process. Our method incorporates object-level representations during feature encoding to align image features with discrete semantic entities, then introduces multi-scale feature fusion and geometric constraints from head pose and gaze direction for fine-grained localization and object-level discrimination. Extensive experiments on GazeFollow, VideoAttentionTarget, ChildPlay, and GOO-Real demonstrate that our method achieves AUC of 0.961, 0.948, 0.987, and 0.977 respectively, delivering strong performance across all benchmarks while maintaining a compact parameter size of 7.1M.
Chinese Translation
注视目标估计旨在预测观察者在图像中注视的语义对象，这一任务深植于人类注视的对象导向特性。观察者倾向于选择特定的语义实体作为注意目标，而不是在图像的任意区域随机反应。然而，现有方法通常将这一任务建模为从全局特征到注视热图的直接映射，实质上将其视为一个像素级回归问题。这种方法未能明确将注视对象表示为一个独立的实体，使得在复杂场景中产生稳定且语义一致的预测变得困难。为了解决这个问题，我们提出了一种由对象语义引导的两阶段注视估计框架，将注视目标估计重新表述为一个分层推理过程。我们的方法在特征编码过程中结合了对象级表示，以将图像特征与离散的语义实体对齐，然后引入多尺度特征融合以及来自头部姿态和注视方向的几何约束，以实现细粒度定位和对象级区分。在 GazeFollow、VideoAttentionTarget、ChildPlay 和 GOO-Real 上的广泛实验表明，我们的方法分别达到了 0.961、0.948、0.987 和 0.977 的 AUC，在所有基准测试中表现出色，同时保持了 7.1M 的紧凑参数规模。
cs.CV / 102 / 2606.29337
W4A4 Quantization for Inference on Wan2.2-I2V-A14B
W4A4量化在Wan2.2-I2V-A14B上的推理
Abstract
We summarize our submission to Sub-Challenge 1: W4A4 Quantization for Inference (HiF4 / MXFP4) of the ICME 2026 Low-Bit-width Large-Model Quantization Challenge. The sub-challenge targets 4-bit weight and 4-bit activation inference on Wan-AI/Wan2.2-I2V-A14B under HiF4 or MXFP4 numerical formats. We adapt two complementary ideas from LLM quantization, MixQ-style mixed precision for sparse activation outliers and SmoothQuant-style per-channel smoothing, together with block-wise HiF4 packing for Wan2.2 feed-forward linear layers. Calibration on representative OpenS2V-5M batches identifies heavy-tailed activation channels; smoothing rebalances dynamic range before W4A4 rounding; and a dual-branch GEMM preserves outlier columns in higher precision while the bulk of channels use strict W4A4. On official VBench I2V metrics, our pipeline stays within 2-3.5 percent of FP16 on most quality axes and improves motion smoothness, outperforming a native HiFloat4 baseline that degrades roughly 5 percent relative to FP16 across all reported scores.
Chinese Translation
我们总结了对ICME 2026低比特宽大模型量化挑战的子挑战1：W4A4量化用于推理（HiF4 / MXFP4）的提交。该子挑战针对在HiF4或MXFP4数值格式下，Wan-AI/Wan2.2-I2V-A14B上的4位权重和4位激活推理。我们借鉴了LLM量化中的两个互补思路，MixQ风格的稀疏激活异常值混合精度和SmoothQuant风格的每通道平滑，同时结合了针对Wan2.2前馈线性层的块状HiF4打包。对代表性OpenS2V-5M批次的校准识别出重尾激活通道；平滑在W4A4取整之前重新平衡动态范围；而双分支GEMM在更高精度下保留异常值列，同时大部分通道使用严格的W4A4。在官方VBench I2V指标上，我们的管道在大多数质量维度上保持在FP16的2-3.5%以内，并改善了运动平滑性，优于原生HiFloat4基线，该基线在所有报告的分数中相对于FP16下降了约5%。
cs.CV / 103 / 2606.29350
Fast Enough to Act: Spatio-Temporal Visual Token Merging for Low-Latency Robotic VLMs and VLAs
足够快速以行动：低延迟机器人视觉语言模型和视觉语言动作模型的时空视觉标记合并
Abstract
Vision-language models and vision-language action models endow the robot with unprecedented capabilities. However, the input of video and high-resolution images yields a massive number of visual tokens, leading to extremely high inference latency and severely hindering the robot's real-time control. To break through this computational bottleneck, we propose ST-Merge, a plug-and-play, training-free framework that efficiently fuses redundant tokens directly during the visual encoding phase. By explicitly constructing 3D spatiotemporal coordinates, it employs a multi-queue parallel matching and weighted aggregation mechanism to achieve efficient and geometrically consistent fusion of redundant tokens across frames. In addition, we introduce a post-merge positional correction mechanism that effectively eliminates spatial deviation caused by merging by dynamically re-evaluating the rotational position code of the weighted centroid of the vision token, thereby ensuring the high-precision spatial awareness required for dexterous operation. In the Video Question Answering task on the mainstream VLM, Qwen2.5-VL, ST-Merge achieves a 2$\times$ inference speedup with only a tiny 1\% loss in precision. When deployed on the $\pi_{0.5}$ VLA policy, ST-Merge achieves an 8.3$\times$ speedup at 1024 $\times$ 1024 resolution and matches the baseline success rate at this high-resolution setting. At lower resolutions, it introduces a small drop in accuracy.
Chinese Translation
视觉语言模型和视觉语言动作模型赋予机器人前所未有的能力。然而，视频和高分辨率图像的输入产生了大量的视觉标记，导致极高的推理延迟，严重阻碍了机器人的实时控制。为了解决这一计算瓶颈，我们提出了ST-Merge，这是一种即插即用、无训练的框架，能够在视觉编码阶段高效地融合冗余标记。通过明确构建三维时空坐标，它采用多队列并行匹配和加权聚合机制，实现了跨帧冗余标记的高效且几何一致的融合。此外，我们引入了一种后合并位置校正机制，通过动态重新评估视觉标记加权质心的旋转位置编码，有效消除合并造成的空间偏差，从而确保灵巧操作所需的高精度空间感知。在主流视觉语言模型Qwen2.5-VL的视频问答任务中，ST-Merge实现了2倍的推理加速，仅损失了1%的精度。当部署在$ ext{π}_{0.5}$视觉语言动作策略上时，ST-Merge在1024 × 1024分辨率下实现了8.3倍的加速，并在该高分辨率设置下达到了基线成功率。在较低分辨率下，它引入了小幅的准确率下降。
cs.CV / 104 / 2606.29357
Dynamic Parsing and Updating Natural Language Specification using VLMs for Robust Vision-Language Tracking
基于视觉语言模型的动态解析与更新自然语言规范以实现稳健的视觉-语言跟踪
Abstract
Vision-language tracking guided by natural language specifications leverages high-level semantic cues of target objects to substantially boost tracking accuracy and robustness. Existing studies have verified that adaptively optimizing textual descriptions throughout the tracking process can effectively mitigate the semantic-visual mismatch induced by dynamic variations in target appearance, position, and other inherent attributes. Nevertheless, mainstream methods that directly generate textual information via sequence models or large language models inevitably suffer from inherent defects, including erroneous target updating, excessive background distraction, and pervasive hallucination artifacts. To address the aforementioned limitations, this paper proposes a novel language dependency parsing mechanism to precisely distill core tracking principal components, encompassing target objects, semantic concepts, and background contextual information. On this basis, we perform component-aware adaptive textual description updates by exploiting the powerful cross-modal understanding capability of the pre-trained vision-language model Qwen-VL. By integrating the proposed elaborately designed modules into the baseline framework, our method achieves consistent and superior tracking performance on multiple large-scale vision-language tracking benchmarks, including TNL2K, LaSOT, TNLLT, and OTB-LANG. The source code and pre-trained models will be released at https://github.com/Event-AHU/Open_VLTrack.
Chinese Translation
由自然语言规范指导的视觉-语言跟踪利用目标对象的高层语义线索，显著提高跟踪的准确性和稳健性。现有研究已验证，在跟踪过程中自适应优化文本描述可以有效缓解因目标外观、位置及其他固有属性的动态变化所引起的语义-视觉不匹配。然而，主流方法通过序列模型或大型语言模型直接生成文本信息，必然会遭遇固有缺陷，包括错误的目标更新、过度的背景干扰以及普遍存在的幻觉伪影。为了解决上述局限性，本文提出了一种新颖的语言依赖解析机制，以精确提取核心跟踪主成分，包括目标对象、语义概念和背景上下文信息。在此基础上，我们利用预训练的视觉-语言模型 Qwen-VL 强大的跨模态理解能力，执行组件感知的自适应文本描述更新。通过将所提出的精心设计模块整合到基线框架中，我们的方法在多个大型视觉-语言跟踪基准测试（包括 TNL2K、LaSOT、TNLLT 和 OTB-LANG）上实现了一致且优越的跟踪性能。源代码和预训练模型将发布在 https://github.com/Event-AHU/Open_VLTrack。
cs.CV / 105 / 2606.29360
SAFE-DiT: Semantics-Aware Fast-path Execution for High-Resolution Diffusion Transformers
SAFE-DiT：高分辨率扩散变换器的语义感知快速路径执行
Abstract
High-resolution Diffusion Transformer (DiT) inference contains substantial spatial redundancy, but many spatially adaptive implementations encode regional computation as attention masks, which can inadvertently move scaled dot-product attention (SDPA) away from FlashAttention fast paths. We identify this avoidable systems bottleneck as Mask-Induced Dispatch Tax (MIDT) and show that it grows with latent sequence length. We introduce SAFE-DiT, a training-free Semantics-Aware Fast-path Execution framework that separates exact mask elision from approximation-based spatial scheduling. SAFE-DiT removes only provenance-certified image self-attention masks that induce a row-wise constant shift in attention logits, preserves semantics-bearing masks such as text-padding masks, and realizes spatial adaptation through prompt-conditioned token partitioning, selective state updates with global context, and periodic context refresh. We call this acceleration-only configuration SAFE-Core and report sensitivity-weighted classifier-free guidance separately as SAFE-DiT+SW. On the evaluated PyTorch SDPA stack, redundant masks make long-sequence attention $4.1\times$ to $5.8\times$ slower than the mask-free path. On Lumina-Next, SAFE-DiT achieves $2.69\times$ end-to-end acceleration at $1024^2$ resolution and $5.09\times$ at $2560^2$, reduces peak memory at $2560^2$ from 94.1 to 27.9 GB, and enables $3072^2$ generation when dense inference runs out of memory. Paired metrics, component ablations, and a blinded human study support visual non-inferiority of SAFE-Core to the dense fast-path baseline, while SAFE-DiT+SW provides a separate prompt-alignment operating point without reintroducing spatial self-attention masks. Code is available at https://github.com/xuanhuayin/SAFE-DiT.
Chinese Translation
高分辨率扩散变换器（DiT）推理存在显著的空间冗余，但许多空间自适应实现将区域计算编码为注意力掩码，这可能无意中使得缩放点积注意力（SDPA）偏离FlashAttention快速路径。我们将这一可避免的系统瓶颈称为掩码引起的调度税（Mask-Induced Dispatch Tax，MIDT），并表明其随着潜在序列长度的增加而增长。我们提出了SAFE-DiT，一个无训练的语义感知快速路径执行框架，它将精确的掩码消除与基于近似的空间调度分开。SAFE-DiT仅移除那些经过来源认证的图像自注意力掩码，这些掩码在注意力逻辑中引起行常量偏移，同时保留具有语义意义的掩码，如文本填充掩码，并通过提示条件的令牌分区、选择性状态更新与全局上下文以及周期性上下文刷新实现空间自适应。我们将这种仅加速的配置称为SAFE-Core，并单独报告敏感性加权的无分类器引导作为SAFE-DiT+SW。在评估的PyTorch SDPA堆栈上，冗余掩码使得长序列注意力比无掩码路径慢$4.1 imes$到$5.8 imes$。在Lumina-Next上，SAFE-DiT在$1024^2$分辨率下实现了$2.69 imes$的端到端加速，在$2560^2$下实现了$5.09 imes$，并将$2560^2$时的峰值内存从94.1 GB降低到27.9 GB，同时在密集推理内存不足时支持$3072^2$的生成。配对指标、组件消融实验和盲人研究支持SAFE-Core在视觉上与密集快速路径基线的非劣性，而SAFE-DiT+SW提供了一个独立的提示对齐操作点，而不重新引入空间自注意力掩码。代码可在https://github.com/xuanhuayin/SAFE-DiT获取。
cs.CV / 106 / 2606.29374
L2D2-GS: Learning to Densify for Feedforward Dynamic Gaussian Scene Reconstruction
L2D2-GS：学习稠密化以进行前馈动态高斯场景重建
Abstract
High-fidelity reconstruction of dynamic urban environments is a cornerstone of autonomous driving simulation and large-scale world modeling. While 3D Gaussian Splatting (3DGS) has established a new standard for real-time rendering, its reliance on expensive per-scene optimization limits scalability. Conversely, recent feedforward methods that infer Gaussian parameters offer faster speed but face fundamental bottlenecks: they are memory-prohibitive at high resolutions and struggle to fuse dense multi-view observations consistently. This paper presents L2D2-GS, a unified framework that reformulates generalizable reconstruction not as a one-shot regression, but as a robust iterative process of optimization and densification. To resolve the ambiguity of supervision in primitive generation, we propose a self-supervised densification policy that derives explicit reward signals from global reconstruction gains to guide local densification. Furthermore, we mitigate irreversible early-stage artifacts through a geometric regularization mechanism, utilizing reparameterization to constrain the optimization manifold and prevent convergence to poor local optima. Extensive experiments on the PandaSet and Waymo datasets demonstrate that our method achieves state-of-the-art reconstruction fidelity and strong zero-shot generalization, while using fewer primitives than competing baselines.
Chinese Translation
高保真动态城市环境重建是自动驾驶仿真和大规模世界建模的基石。尽管3D高斯点云（3D Gaussian Splatting，3DGS）为实时渲染建立了新的标准，但其对每个场景的昂贵优化依赖限制了可扩展性。相反，最近的前馈方法通过推断高斯参数提供了更快的速度，但面临根本瓶颈：在高分辨率下，它们的内存需求过高，并且在一致融合稠密多视角观测方面存在困难。本文提出了L2D2-GS，一个统一框架，将可推广的重建重新表述为一个稳健的优化和稠密化迭代过程，而非一次性回归。为了解决原始生成中的监督模糊性，我们提出了一种自监督稠密化策略，从全局重建增益中推导出明确的奖励信号，以指导局部稠密化。此外，我们通过几何正则化机制减轻不可逆的早期阶段伪影，利用重新参数化约束优化流形，防止收敛到不良局部最优解。在PandaSet和Waymo数据集上的大量实验表明，我们的方法在重建保真度和强大的零-shot泛化方面达到了最先进的水平，同时使用的原始数量少于竞争基线。
cs.CV / 107 / 2606.29376
SAD-GS: Learning Reliable 3D Semantic Gaussian Fields via Dynamic Geo-Semantic Anchoring
SAD-GS：通过动态地理语义锚定学习可靠的3D语义高斯场
Abstract
Open-vocabulary 3D semantic Gaussian field learning relies on multi-view 2D supervision, whose semantic targets and spatial assignments are often unreliable. Across varying viewpoints, view-dependent features cause semantic identity drift, while propagated tracker masks introduce boundary leakage and identity switches. Directly optimizing against these unreliable 2D targets forces the 3D representation to absorb multi-view contradictions, leading to severe error accumulation. To resolve this limitation, we propose SAD-GS, a framework for learning reliable 3D semantic Gaussian fields via dynamic geo-semantic anchoring. Specifically, Semantic Anchor Distillation (SAD) distills per-view visual embeddings into consensus text anchors to establish a viewpoint-invariant semantic identity. Concurrently, the Geo-Semantic Feedback Loop (GSFL) leverages the evolving 3D field to actively filter tracker anomalies and refine spatial mask assignments via a conservative three-gate update rule. Extensive evaluations on LERF-OVS, 3D-OVS, and Mip-NeRF360 show that SAD-GS consistently achieves the best overall performance in both open-vocabulary localization and semantic segmentation. These comprehensive improvements validate the effectiveness and robustness of dynamic geo-semantic anchoring for reliable 3D semantic Gaussian field learning.
Chinese Translation
开放词汇的3D语义高斯场学习依赖于多视角的2D监督，而其语义目标和空间分配往往不可靠。在不同的视角下，视角依赖特征导致语义身份漂移，而传播的跟踪器掩码则引入边界泄漏和身份切换。直接针对这些不可靠的2D目标进行优化，迫使3D表示吸收多视角矛盾，导致严重的误差累积。为了解决这一限制，我们提出了SAD-GS，一个通过动态地理语义锚定学习可靠的3D语义高斯场的框架。具体而言，语义锚定蒸馏（Semantic Anchor Distillation, SAD）将每个视角的视觉嵌入提炼为共识文本锚，以建立视角不变的语义身份。同时，地理语义反馈循环（Geo-Semantic Feedback Loop, GSFL）利用不断演变的3D场，主动过滤跟踪器异常，并通过保守的三门更新规则细化空间掩码分配。在LERF-OVS、3D-OVS和Mip-NeRF360上的广泛评估表明，SAD-GS在开放词汇定位和语义分割中始终实现最佳整体性能。这些全面的改进验证了动态地理语义锚定在可靠的3D语义高斯场学习中的有效性和鲁棒性。
cs.CV / 108 / 2606.29379
DR-GS: Physically-Based Deformable and Relightable 2D Gaussians
DR-GS：基于物理的可变形和可重光照的二维高斯模型
Abstract
Gaussian splatting (GS) has garnered significant attention in VR/AR and digital content creation due to its explicit parameterization and efficient rendering capabilities. However, existing GS-based methods for deformable objects face two key limitations: (i) illumination is erroneously baked into textures, causing physically inconsistent responses under dynamic deformations and lighting changes; (ii) snapshot-based reconstruction restricts post-reconstruction material editing. To address these challenges, we propose Deformable and Relightable GS (DR-GS), a unified Gaussian framework that integrates physically-based inverse rendering, relighting, and deformation-aware manipulation. Through explicitly disentangling geometry, illumination, and material representations, DR-GS overcomes the limitations of static snapshots, resolving unrealistic appearance under varying conditions while enabling post-reconstruction parameter editing. Extensive experiments show that DR-GS achieves leading visual quality across static reconstruction, dynamic deformation, and relighting, reliably preserving reflections and specular highlights on glossy surfaces. It further establishes a fully decoupled geometry-illumination-material pipeline, enabling high-quality 3D asset creation and comprehensive post-editing.
Chinese Translation
高斯点云（Gaussian splatting, GS）因其明确的参数化和高效的渲染能力，在虚拟现实/增强现实（VR/AR）和数字内容创作中引起了广泛关注。然而，现有的基于GS的可变形物体方法面临两个主要限制：（i）光照错误地嵌入到纹理中，导致在动态变形和光照变化下的物理不一致响应；（ii）基于快照的重建限制了重建后材料的编辑。为了解决这些挑战，我们提出了可变形和可重光照的GS（Deformable and Relightable GS, DR-GS），这是一个统一的高斯框架，集成了基于物理的逆渲染、重光照和变形感知操作。通过明确地解耦几何、光照和材料表示，DR-GS克服了静态快照的局限性，解决了在不同条件下的不现实外观，同时支持重建后的参数编辑。大量实验表明，DR-GS在静态重建、动态变形和重光照方面实现了领先的视觉质量，可靠地保留了光滑表面的反射和高光。此外，它进一步建立了一个完全解耦的几何-光照-材料管道，支持高质量的3D资产创建和全面的后期编辑。
cs.CV / 109 / 2606.29384
Event-VLA: Action-Conditioned Event Fusion for Robust Vision-Language-Action Model
事件-VLA：基于动作条件的事件融合用于稳健的视觉-语言-动作模型
Abstract
Vision-Language-Action (VLA) models have become an important paradigm of embodied AI. However, existing VLA models typically assume well-lit and stable indoor settings, while real-world embodied manipulation may involve degraded RGB observations caused by illumination shifts, posing critical challenges for robust robotic manipulation. To address this gap, we propose \textbf{Event-VLA}, an event-enhanced VLA framework for generalizable manipulation across varying illumination conditions. We formulate VLA-based manipulation under degraded visibility as a practical robustness problem for RGB-centric policies, and introduce event streams as an illumination-robust, motion-sensitive complementary observation to improve robustness across visibility levels. Specifically, unlike conventional multimodal fusion that directly merges event features into the global semantic token space, Event-VLA injects event information through an action-query routing pathway. It uses learnable action queries to extract task-relevant semantics from the VLA reasoning process, and selectively aggregates event tokens via gated cross-attention to construct event-aware action representations. This design preserves the pretrained RGB-language semantic priors while effectively leveraging event information for robust action prediction. Experiments in simulation and real-world deployment show that Event-VLA maintains strong manipulation performance under normal lighting and improves success rates under low-light degradation and near-dark real-world settings.
Chinese Translation
视觉-语言-动作（VLA）模型已成为具身人工智能的重要范式。然而，现有的VLA模型通常假设在光照良好且稳定的室内环境中运行，而现实世界中的具身操控可能涉及由光照变化引起的降级RGB观测，这对稳健的机器人操控提出了关键挑战。为了解决这一问题，我们提出了 extbf{事件-VLA}，一个增强事件的VLA框架，旨在实现跨不同光照条件的可泛化操控。我们将基于VLA的操控在降级可见性下形式化为RGB中心策略的实际稳健性问题，并引入事件流作为一种光照稳健、运动敏感的补充观测，以提高在不同可见性水平下的稳健性。具体而言，与传统的多模态融合直接将事件特征合并到全局语义标记空间不同，事件-VLA通过动作查询路由路径注入事件信息。它使用可学习的动作查询从VLA推理过程中提取与任务相关的语义，并通过门控交叉注意机制选择性地聚合事件标记，以构建事件感知的动作表示。这一设计在有效利用事件信息进行稳健动作预测的同时，保留了预训练的RGB-语言语义先验。模拟和现实世界部署的实验表明，事件-VLA在正常光照下保持强大的操控性能，并在低光降级和近乎黑暗的现实世界环境中提高成功率。
cs.CV / 110 / 2606.29395
NaLA: A 3D Native LLM Layout Agent for High-quality 3D Scene Generation
NaLA：一种用于高质量3D场景生成的3D原生LLM布局代理
Abstract
Recently, Large Language Models (LLMs) have emerged as promising layout agents for 3D scene generation. Existing layout agents still suffer from implausible layout generation because most of them convert 3D assets and 3D layouts into textual descriptions as inputs and outputs, which involves severe information loss due to the modality gap between texts and 3D assets and 3D layouts. We propose NaLA, a native 3D LLM layout Agent for high-quality 3D scene generation by placing 3D assets in the scene. For the inputs, NaLA encodes 3D scene boundaries and 3D assets directly into the LLM, preserving fine-grained geometry and enabling explicit reasoning over relationships like collisions, surface supporting, and containment. To accurately output the positions and orientations of assets, NaLA adopts a coarse-to-fine prediction mechanism that first predicts discrete poses in an autoregressive manner and then refines the discrete poses with a continuous regression. Trained on diverse layout datasets, NaLA attains strong geometric perception and layout coherence. Experiments demonstrate that NaLA outperforms prior layout agents in both generation quality and inference efficiency, with comprehensive ablation studies to verify each component's effectiveness.
Chinese Translation
近年来，大型语言模型（LLMs）作为3D场景生成的有前景的布局代理而受到关注。现有的布局代理仍然存在不合理的布局生成问题，因为它们大多数将3D资产和3D布局转换为文本描述作为输入和输出，这导致由于文本与3D资产及3D布局之间的模态差异而严重的信息损失。我们提出了NaLA，一种用于高质量3D场景生成的原生3D LLM布局代理，通过在场景中放置3D资产来实现。对于输入，NaLA将3D场景边界和3D资产直接编码到LLM中，保留了细粒度的几何信息，并能够对碰撞、表面支撑和包含等关系进行明确推理。为了准确输出资产的位置和方向，NaLA采用了一种粗到细的预测机制，首先以自回归的方式预测离散姿态，然后通过连续回归对离散姿态进行细化。经过多样化布局数据集的训练，NaLA获得了强大的几何感知能力和布局一致性。实验表明，NaLA在生成质量和推理效率上均优于之前的布局代理，并通过全面的消融研究验证了每个组件的有效性。
cs.CV / 111 / 2606.29400
Learning to Adaptively Allocate Gaussians for Arbitrary-Scale Image Super-Resolution
学习自适应分配高斯函数以实现任意尺度图像超分辨率
Abstract
In computer graphics, visual content is continuously warped, zoomed and resampled. This occurs when engines upscale frames, users zoom into 3D scenes, or foveated VR applies varying scaling. Handling these transformations requires Arbitrary-Scale Super-Resolution (ASR). Traditional models, designed for fixed scales, typically predict at a lower integer scale (e.g., x4) and rely on sub-optimal interpolation for continuous resolutions, compromising quality. Furthermore, most methods process pixels uniformly. Since fine details are sparse, this creates overhead; efficiency dictates concentrating resources only where structural complexity demands it. While implicit models and Gaussian Splatting (GS) enable continuous representation, GS is advantageous due to adaptive densification. However, transitioning GS into a feed-forward model for ASR is non-trivial. Standard GS optimization needs high-resolution gradients to drive primitive growth, which are unavailable during inference. Thus, the network must autonomously predict GS densification from low-resolution inputs. To solve this, we propose QuADA-GS. After encoding inputs into a latent space, a Neural Routing Architecture evaluates local complexity to distribute a global budget, assigning specific upsampling factors to features to avoid redundant processing. Features are dynamically densified based on these factors, forming an irregular topology decoded into 2D Gaussian primitives. To coordinate features before decoding, we introduce Hierarchical Pointer Convolution. This non-grid operator achieves O(1) neighbor lookup complexity, facilitating efficient spatial communication and bypassing dense bottlenecks. Experiments show QuADA-GS achieves state-of-the-art ASR performance, maintaining low latency and a lean memory footprint.
Chinese Translation
在计算机图形学中，视觉内容不断被扭曲、缩放和重新采样。这种情况发生在引擎放大帧时、用户缩放3D场景时，或在注视点虚拟现实中应用不同的缩放时。处理这些变换需要任意尺度超分辨率（Arbitrary-Scale Super-Resolution, ASR）。传统模型设计用于固定尺度，通常在较低的整数尺度（例如，x4）下进行预测，并依赖次优插值来处理连续分辨率，从而影响质量。此外，大多数方法均匀处理像素。由于细节稀疏，这会造成开销；效率要求仅在结构复杂性需要的地方集中资源。尽管隐式模型和高斯溅射（Gaussian Splatting, GS）能够实现连续表示，但由于自适应密集化，GS更具优势。然而，将GS转变为ASR的前馈模型并非易事。标准GS优化需要高分辨率梯度来驱动原始体的生长，而在推理过程中这些梯度是不可用的。因此，网络必须自主预测低分辨率输入的GS密集化。为了解决这个问题，我们提出了QuADA-GS。在将输入编码到潜在空间后，神经路由架构评估局部复杂性以分配全局预算，为特征分配特定的上采样因子，以避免冗余处理。特征根据这些因子动态密集化，形成解码为2D高斯原语的不规则拓扑。为了在解码之前协调特征，我们引入了层次指针卷积。这种非网格算子实现了O(1)的邻居查找复杂度，促进了高效的空间通信，并绕过了密集瓶颈。实验表明，QuADA-GS在ASR性能上达到了最先进的水平，同时保持低延迟和精简的内存占用。
cs.CV / 112 / 2606.29414
FiRe: Frequency Reparameterization as a Preconditioner for Periodic Implicit Neural Representations
FiRe：频率重参数化作为周期性隐式神经表示的预条件器
Abstract
Periodic Implicit Neural Representations (INRs) such as SIREN and FINER assign every neuron, the same global frequency, spending the representational budget inefficiently when local signal content varies. We introduce FiRe (Frequency Reparameterization), that accelerates optimization by reparameterizing per-neuron frequency of periodic INRs without changing their underlying activation function. FiRe gives each neuron a bounded, input-dependent frequency via a separate low-rank gating path and is applicable to any periodic activation function. The gate acts as an implicit preconditioner that improves optimization conditioning at initialization via the Neural Tangent Kernel (NTK). This better-conditioned initialization makes optimization converge faster, and the high-frequency content of the reconstruction tracks the target more closely at a fixed computational budget. On 2D image fitting, FiRe increases PSNR over a parameter-matched baseline (up to +1 dB at short training budgets), with gains that vary with resolution and diminish at full convergence. We characterize how performance depends on resolution, rank, and training budget, and give an NTK account that predicts these trends.
Chinese Translation
周期性隐式神经表示（INRs），如 SIREN 和 FINER，为每个神经元分配相同的全局频率，当局部信号内容变化时，这种方法在表示预算上效率低下。我们提出了 FiRe（频率重参数化），通过对周期性 INRs 的每个神经元频率进行重参数化，而不改变其基础激活函数，从而加速优化。FiRe 通过一个单独的低秩门控路径为每个神经元提供一个有界的、依赖于输入的频率，并适用于任何周期性激活函数。该门控作为一种隐式预条件器，通过神经切线核（Neural Tangent Kernel，NTK）在初始化时改善优化条件。这种更好条件的初始化使得优化收敛更快，并且在固定计算预算下，重建的高频内容更紧密地跟踪目标。在二维图像拟合中，FiRe 在与参数匹配的基线相比，增加了峰值信噪比（PSNR）（在短训练预算下可达 +1 dB），增益随着分辨率变化，并在完全收敛时减小。我们描述了性能如何依赖于分辨率、秩和训练预算，并给出了一个 NTK 解释，以预测这些趋势。
cs.CV / 113 / 2606.29416
Can Machines Really See Objects in Images? A Study Based on Syntactic Distance and Visual Self-Referential Instances
机器真的能在图像中识别物体吗？基于句法距离和视觉自指实例的研究
Abstract
Can a vision model truly see an object, or does it only fit surface-level visual cues? Following Wittgenstein's view that the limits of language are the limits of the world, we view a model's recognition ability as bounded by the descriptive system it has learned. In current vision models, this system is often realized through learned feature representations that exploit local statistical cues. We therefore ask whether a model can still classify correctly when such local cues provide no stable basis for distinction. We formalize this question with syntactic distance, which measures class separability through the symmetry of the operations mapping one class to the other: positive distance exposes exploitable local features, whereas zero distance requires global semantics rather than local rules. We construct a visual self-referential task in maximum-variance binary noise: positive samples contain a closed square, while negative samples contain an otherwise identical square with one flipped boundary pixel. The two classes differ in global semantics but have zero syntactic distance, making local statistical shortcuts unreliable. Experiments on ResNets and Vision Transformers reveal a consistent phase-transition phenomenon, with accuracy collapsing to random guessing once the image scale crosses a critical point and does not recover within the tested range. Larger training sets and models only delay this collapse, while globally attentive ViTs reach it earlier. These results reveal a structural capability boundary of current architectures on global-concept tasks, suggesting that general intelligence may require creating new language, not reusing an existing one.
Chinese Translation
视觉模型真的能看见一个物体，还是仅仅适应表层的视觉线索？根据维特根斯坦的观点，语言的界限就是世界的界限，我们认为模型的识别能力受限于其所学习的描述系统。在当前的视觉模型中，这一系统通常通过利用局部统计线索的学习特征表示来实现。因此，我们询问当这些局部线索无法提供稳定的区分基础时，模型是否仍然能够正确分类。我们用句法距离来形式化这个问题，它通过映射一个类别到另一个类别的操作的对称性来测量类别的可分性：正距离揭示可利用的局部特征，而零距离则需要全局语义而非局部规则。我们构建了一个在最大方差二元噪声下的视觉自指任务：正样本包含一个封闭的正方形，而负样本则包含一个在边界像素翻转的情况下与之完全相同的正方形。这两个类别在全局语义上有所不同，但具有零句法距离，使得局部统计捷径不可靠。在 ResNets 和 Vision Transformers 上的实验揭示了一种一致的相变现象，当图像尺度超过临界点时，准确率迅速下降至随机猜测，并且在测试范围内无法恢复。更大的训练集和模型仅仅延迟了这种崩溃，而全局注意的 ViTs 则更早达到这一点。这些结果揭示了当前架构在全局概念任务上的结构能力边界，暗示一般智能可能需要创造新的语言，而不是重用现有的语言。
cs.CV / 114 / 2606.29417
Bit-ViP: Leveraging Bit-planes to Preserve Visual Privacy in Images through Obfuscation
Bit-ViP：利用位平面通过模糊处理保护图像的视觉隐私
Abstract
The unprecedented growth of computer vision applications, such as surveillance systems and social media, raises security and visual privacy concerns, especially when data is stored on cloud servers. Image obfuscation offers a way to preserve visual privacy while maintaining an adequate level of usability; thus, it has been a topic of great interest in recent years. However, prior obfuscation schemes are either vulnerable to malicious attacks, such as model inversion to reconstruct original images from obfuscated images, or generate non-trainable obfuscated images, making them unusable for achieving reasonable accuracy. This paper proposes a novel bit-plane-based image obfuscation scheme, {\em Bit-ViP}, to preserve visual privacy for image-based recognition tasks. The Bit-ViP scheme produces secure, usable images by incorporating an innovative end-to-end obfuscation function. While doing so, the obfuscated image would contain non-invertible noise (generated by Lorenz's chaotic system and differential privacy), making it hard for an adversary to reconstruct the original image. We conduct extensive experiments on two popular activity recognition datasets, namely UCF101 and HMDB51, to validate the effectiveness of Bit-ViP. In the face of attacks on reconstruction, pixel frequency, information entropy, and pixel inter-correlation, we present a rigorous security analysis demonstrating tangible improvements over existing schemes.
Chinese Translation
计算机视觉应用的前所未有增长，如监控系统和社交媒体，引发了安全和视觉隐私的担忧，尤其是在数据存储于云服务器时。图像模糊处理提供了一种在保持适当可用性水平的同时保护视觉隐私的方法，因此近年来备受关注。然而，先前的模糊处理方案要么容易受到恶意攻击，例如模型反演以从模糊图像重建原始图像，要么生成不可训练的模糊图像，使其无法用于实现合理的准确性。本文提出了一种新颖的基于位平面的图像模糊处理方案，{ extit{Bit-ViP}}，以保护图像识别任务中的视觉隐私。Bit-ViP方案通过结合创新的端到端模糊处理功能，生成安全且可用的图像。同时，模糊图像将包含不可逆的噪声（由洛伦兹混沌系统和差分隐私生成），使得对手难以重建原始图像。我们在两个流行的活动识别数据集上进行了广泛的实验，即UCF101和HMDB51，以验证Bit-ViP的有效性。在面对重建攻击、像素频率、信息熵和像素间相关性时，我们进行了严格的安全分析，展示了相较于现有方案的显著改进。
cs.CV / 115 / 2606.29428
Robust Zero-shot Anomaly Detection under Limited Auxiliary Anomaly Priors
在有限辅助异常先验下的鲁棒零-shot异常检测
Abstract
Zero-shot anomaly detection aims to identify defects in arbitrary novel domains; however, existing models assume that the auxiliary data contains a rich diversity of anomalies, neglecting the far more complex and unpredictable variations in real-world target domains. This study introduces DIVE, the first approach to investigate the scenario of limited auxiliary anomaly priors and resolve the resulting substantial performance degradation. Through a shallow-and-deep text embedding injection strategy during visual encoding, DIVE learns to abstract generic anomaly concepts shared across the auxiliary training domain and diverse target domains. Moreover, we propose a disentanglement mechanism to tackle the suboptimal alignment between visual embeddings entangled with object semantics and object-agnostic textual prompts. Experiments demonstrate that, under the setting of limited anomaly patterns in auxiliary data, DIVE outperforms SOTA baselines by up to 16.2% and 28.5% on two classification metrics, and 23.4%, 24.1%, and 47.0% on three segmentation metrics, in terms of average performance across twelve datasets. Furthermore, it maintains highly competitive performance when auxiliary data exhibits sufficient anomaly diversity.
Chinese Translation
零-shot异常检测旨在识别任意新领域中的缺陷；然而，现有模型假设辅助数据包含丰富多样的异常，忽视了现实目标领域中更为复杂和不可预测的变化。本研究提出了DIVE，这是首个探讨有限辅助异常先验场景并解决由此导致的显著性能下降的方法。通过在视觉编码过程中采用浅层和深层文本嵌入注入策略，DIVE学习在辅助训练领域和多样化目标领域中共享的通用异常概念。此外，我们提出了一种解耦机制，以解决与对象语义交织的视觉嵌入和与对象无关的文本提示之间的次优对齐问题。实验表明，在辅助数据中有限异常模式的设置下，DIVE在十二个数据集的平均性能上，分别在两个分类指标上超越最先进的基线模型（SOTA）高达16.2%和28.5%，在三个分割指标上超越23.4%、24.1%和47.0%。此外，当辅助数据表现出足够的异常多样性时，DIVE仍能保持高度竞争的性能。
cs.CV / 116 / 2606.29430
EvLIR: Learning Illumination Residuals from Ordered Events for Low-Light Image Enhancement
EvLIR：从有序事件中学习照明残差以增强低光图像
Abstract
Low-light image enhancement is severely ill-posed when the input frame contains missing structure, saturated noise, and weak local contrast. Event cameras provide asynchronous brightness-change observations with high temporal resolution, but prior works often treat voxel channels as an unordered or static feature stack before fusion, rather than explicitly modeling their within-window temporal evolution, weakening the temporal evidence that makes events useful. We propose EvLIR, a temporal-residual enhancement framework that learns illumination residuals from ordered events for low-light image enhancement. Given a low-light frame and its aligned event voxel, EvLIR preserves the ordered temporal bins of the event stream and introduces a Temporal Event Residual Module (TERM) to encode short-window event dynamics with a lightweight ConvGRU. The resulting temporal state is converted into a bounded illumination correction, which provides spatially adaptive photometric guidance for Retinex-style illumination estimation and subsequent reliability-aware image-event restoration. On SDE and SDSD indoor/outdoor benchmarks, EvLIR achieves the best result on eleven of twelve dataset-metric pairs, with average scores of 25.63~dB PSNR, 28.30~dB PSNR*, and 0.827 SSIM across the four benchmarks.
Chinese Translation
低光图像增强在输入帧包含缺失结构、饱和噪声和弱局部对比度时严重不适定。事件相机提供高时间分辨率的异步亮度变化观测，但之前的研究通常将体素通道视为无序或静态特征堆栈进行融合，而不是明确建模它们在窗口内的时间演变，从而削弱了使事件有用的时间证据。我们提出了EvLIR，一个时间残差增强框架，从有序事件中学习照明残差以增强低光图像。给定一个低光帧及其对齐的事件体素，EvLIR保留事件流的有序时间箱，并引入一个时间事件残差模块（Temporal Event Residual Module, TERM）来使用轻量级的ConvGRU编码短窗口事件动态。生成的时间状态被转换为有界的照明校正，为Retinex风格的照明估计和后续的可靠性感知图像-事件恢复提供空间自适应的光度指导。在SDE和SDSD室内/室外基准测试中，EvLIR在十二个数据集-指标对中的十一项上取得了最佳结果，四个基准的平均得分为25.63 dB PSNR、28.30 dB PSNR*和0.827 SSIM。
cs.CV / 117 / 2606.29445
Bridging VideoQA and Video-Guided Agentic Tasks via Generalized Keyframe Extraction
通过广义关键帧提取连接视频问答与视频引导的自主任务
Abstract
Video understanding is a fundamental capability for multimodal intelligence, and recent Multimodal Large Language Models (MLLMs) have achieved remarkable performance on Video Question Answering (VideoQA) benchmarks. However, existing benchmarks primarily evaluate whether models can perceive shallow visual cues, while rarely examining whether MLLMs can learn deeper knowledge or procedural skills from video tutorials and generalize them to downstream long-horizon agentic tasks. To address this gap, we introduce VG-GUIBench (Video-Guided GUI Benchmark), a new benchmark designed to evaluate whether MLLM-based GUI agents can follow video tutorials to complete corresponding GUI interactive tasks. Furthermore, we observe that the performance of models on both VideoQA and video-guided agentic tasks critically depends on effective keyframe extraction. Based on this observation, we propose TASKER (Task-driven And Scene-aware Keyframe searchER), a keyframe extraction algorithm that jointly considers task relevance and scene dynamics to identify informative frames. Experimental results demonstrate that TASKER achieves significant performance improvements on both VideoQA and video-guided agentic task benchmarks, outperforming the best baseline by 2.0% on the EgoSchema fullset and 1.8% on the NExT-QA dataset, respectively. These results further highlight the potential of generalized keyframe extraction methods for video understanding tasks. Our code and data are available at https://github.com/VG-GUI-TASKER/VG-GUI-TASKER.
Chinese Translation
视频理解是多模态智能的基本能力，最近的多模态大型语言模型（MLLMs）在视频问答（VideoQA）基准测试中取得了显著的表现。然而，现有基准主要评估模型是否能够感知浅层视觉线索，而很少考察MLLMs是否能够从视频教程中学习更深层次的知识或过程技能，并将其推广到下游的长期自主任务。为了解决这一问题，我们引入了VG-GUIBench（视频引导的GUI基准），这是一个新的基准，旨在评估基于MLLM的GUI代理是否能够遵循视频教程完成相应的GUI交互任务。此外，我们观察到模型在视频问答和视频引导的自主任务上的表现都严重依赖于有效的关键帧提取。基于这一观察，我们提出了TASKER（任务驱动和场景感知的关键帧搜索器），这是一种关键帧提取算法，联合考虑任务相关性和场景动态，以识别信息丰富的帧。实验结果表明，TASKER在视频问答和视频引导的自主任务基准测试中均实现了显著的性能提升，在EgoSchema全数据集上超越最佳基线2.0%，在NExT-QA数据集上超越最佳基线1.8%。这些结果进一步强调了广义关键帧提取方法在视频理解任务中的潜力。我们的代码和数据可在 https://github.com/VG-GUI-TASKER/VG-GUI-TASKER 获取。
cs.CV / 118 / 2606.29447
Miti360: A Comprehensive Dataset for Improved Reforestation Monitoring
Miti360：用于改善再造林监测的综合数据集
Abstract
Over the past decade, interest in applying machine learning (ML) to automate forest monitoring has grown significantly. However, existing training datasets are predominantly drawn from North America, Europe, Asia, and Australia, leaving a critical gap in African forestry data. To address this limited geographic diversity, we present Miti360, a comprehensive dataset for reforestation monitoring that comprises high-resolution imagery, ground truth data, and longitudinal weather data. Data collection occurred within a 770-ha reforested section of the Kieni Forest in Kenya between March 2023 and February 2025. Miti360 comprises aerial photos (orthophotos and tiles) with tree bounding box annotations, terrestrial images (single and stereo), and detailed data records including tree biophysical parameters, species, and GPS coordinates, alongside historical weather data. Aerial surveys utilized a DJI Mavic 2 Pro, with imagery stitched via Agisoft Metashape and tiled using ArcGIS Pro, while terrestrial captures used smartphones and custom stereo cameras. Miti360 enables the training of ML systems for tasks such as accelerating tree censuses, matching species to geographical areas, modelling growth based on weather conditions, and developing digital twin frameworks. Models can be trained on Miti360 to address challenges specific to Sub-Saharan Africa, ultimately advancing reforestation monitoring and fostering sustainable forestry practices in underrepresented regions. We demonstrate the utility of this dataset by successfully tracking tree crowns across three years and improving the DeepForest model's box precision and box recall by 12% and 69% respectively through fine-tuning on Miti360.
Chinese Translation
在过去十年中，应用机器学习（ML）来自动化森林监测的兴趣显著增长。然而，现有的训练数据集主要来自北美、欧洲、亚洲和澳大利亚，导致非洲林业数据的严重缺乏。为了解决这一地理多样性不足的问题，我们提出了Miti360，这是一个用于再造林监测的综合数据集，包含高分辨率影像、地面真实数据和长期气象数据。数据采集发生在2023年3月至2025年2月期间，位于肯尼亚Kieni森林的770公顷再造林区域。Miti360包括带有树木边界框注释的航空照片（正射影像和瓦片）、地面图像（单幅和立体）以及详细的数据记录，包括树木生物物理参数、物种和GPS坐标，以及历史气象数据。航空调查使用了DJI Mavic 2 Pro，通过Agisoft Metashape拼接影像，并使用ArcGIS Pro进行瓦片处理，而地面拍摄则使用智能手机和定制的立体相机。Miti360使得训练ML系统能够执行加速树木普查、将物种与地理区域匹配、基于气象条件建模生长以及开发数字双胞胎框架等任务。模型可以在Miti360上进行训练，以应对撒哈拉以南非洲特有的挑战，最终推动再造林监测并促进在欠代表地区的可持续林业实践。我们通过成功跟踪三年内的树冠，并通过在Miti360上的微调将DeepForest模型的框精度和框召回率分别提高了12%和69%，展示了该数据集的实用性。
cs.CV / 119 / 2606.29451
The Platonic Defense: Backdoor Defense for Self-Supervised Encoders in the Era of Large Scale Pre-training
柏拉图防御：在大规模预训练时代对自监督编码器的后门防御
Abstract
Self-supervised learning (SSL) pretrained models have become a dominant paradigm for visual representation learning, but they are vulnerable to backdoor attacks. Existing defenses struggle to defend against such attacks in a fully black-box setting because they often require access to labels, attack patterns, or training data. To tackle this issue, we propose a new attack-agnostic, model-agnostic, and modality-agnostic black-box test-time defense paradigm, called \emph{Platonic Representation Defense}. It is inspired by the Platonic Representation Hypothesis, which suggests that large-scale independently trained encoders converge toward compatible projections of the same underlying reality. We formalize this idea as a conditional energy function defined over source representations and a set of reference representations. The energy function is trained for detection through noise-contrastive estimation and for representation purification through denoising score matching. Theoretically, the energy gap between matched and mismatched samples is lower bounded by the mutual information between source and reference representations. We demonstrate the effectiveness of our method on multiple self-supervised encoders and more than 10 attacks. The method can perform both representation detection and purification, and achieves substantial performance gains across multiple attacks. Code is available \href{https://github.com/jsrdcht/Platonic-Representation-Defense}{here}.
Chinese Translation
自监督学习（SSL）预训练模型已成为视觉表征学习的主流范式，但它们易受后门攻击的影响。现有防御方法在完全黑箱环境中难以抵御此类攻击，因为它们通常需要访问标签、攻击模式或训练数据。为了解决这个问题，我们提出了一种新的攻击无关、模型无关和模态无关的黑箱测试时防御范式，称为“柏拉图表征防御”（Platonic Representation Defense）。该方法受到柏拉图表征假设的启发，该假设表明大规模独立训练的编码器趋向于同一潜在现实的兼容投影。我们将这一思想形式化为一个定义在源表征和一组参考表征上的条件能量函数。该能量函数通过噪声对比估计进行检测训练，并通过去噪分数匹配进行表征净化。从理论上讲，匹配样本与不匹配样本之间的能量差距由源表征与参考表征之间的互信息下界。我们在多个自监督编码器和超过10种攻击上展示了我们方法的有效性。该方法可以同时执行表征检测和净化，并在多种攻击中实现显著的性能提升。代码可在此获取： https://github.com/jsrdcht/Platonic-Representation-Defense 。
cs.CV / 120 / 2606.29453
Resonant Brane Splatting for Arbitrary-Scale Super-Resolution
任意尺度超分辨率的共振膜喷涂
Abstract
Arbitrary-Scale Super-Resolution (ASR) reconstructs images at continuous magnification factors. Recent methods accelerate inference by replacing computationally heavy implicit neural decoders with explicit 2D Gaussian Splatting (GS). However, since standard Gaussians are smooth low-pass primitives, modeling edges and fine textures requires multiple overlapping, well-aligned splats, which creates severe bottlenecks during rasterization. To address this, we introduce Resonant Brane Splatting (RBS), a feed-forward ASR framework. RBS replaces flat Gaussians with Branes: expressive primitives that emit spatially varying colors to natively model local contrast and complex textures within a single footprint. We achieve this by augmenting the standard Gaussian envelope with internal Gaussian-Hermite modes, assigning a distinct color coefficient to each. The zero-order mode recovers standard GS, while higher-order modes capture high frequencies. We predict Brane parameters directly from low-resolution features. Because Branes provide a mathematically richer formulation than simple Gaussians, far fewer primitives need to overlap to reconstruct a given target pixel. To exploit this, we introduce an efficient fully differentiable rasterizer with a precise culling strategy based on the classical quantum turning point. This allows us to safely skip negligible regions, drastically reducing the rendering overhead. Experiments on standard ASR benchmarks show that RBS improves reconstruction quality over implicit and GS baselines, while achieving superior speed-quality trade-off than prior GS methods.
Chinese Translation
任意尺度超分辨率（ASR）在连续放大因子下重建图像。最近的方法通过用显式二维高斯喷涂（Gaussian Splatting, GS）替代计算量大的隐式神经解码器来加速推理。然而，由于标准高斯是平滑的低通原语，建模边缘和细纹理需要多个重叠且良好对齐的喷涂，这在光栅化过程中造成了严重的瓶颈。为了解决这个问题，我们提出了共振膜喷涂（Resonant Brane Splatting, RBS），这是一个前馈的ASR框架。RBS用膜（Branes）替代平面高斯：这些表达性原语发出空间变化的颜色，以在单一足迹内原生建模局部对比度和复杂纹理。我们通过增强标准高斯包络，结合内部高斯-厄米特模式，为每个模式分配一个独特的颜色系数来实现这一点。零阶模式恢复标准GS，而高阶模式捕捉高频信息。我们直接从低分辨率特征预测膜参数。由于膜提供了比简单高斯更丰富的数学形式，因此重建给定目标像素所需的重叠原语数量大大减少。为了利用这一点，我们引入了一种高效的全可微光栅化器，并基于经典量子转折点采用精确的剔除策略。这使我们能够安全地跳过微不足道的区域，从而大幅减少渲染开销。在标准ASR基准测试中的实验表明，RBS在重建质量上优于隐式和GS基线，同时在速度与质量的权衡上优于之前的GS方法。
cs.CV / 121 / 2606.29461
From Phase to Phenomenon: Self-Supervised Learning of Subsurface Scattering with Minimal Phase-shift Inputs
从相位到现象：利用最小相位偏移输入的自监督学习地下散射
Abstract
We propose a self-supervised pretraining framework for learning sub-surface scattering (SSS) light transport representations from minimal input. Our method leverages a stereo projector-camera setup that captures only eight high-frequency phase-shift profilometry (PSP) images per view to pretrain an encoder in a multi-view, multi-object setting. We introduce a tailored augmentation strategy for PSP-based SSS data, and show that it significantly outperforms standard ImageNet-style augmentations for SSL pretraining. The pretrained encoder learns generalizable SSS representations that transfer effectively to downstream tasks, including spatially varying relighting and representation evaluation using a kNN classifier. Combined with a decoder, the model reconstructs dense scattering footprint responses, trained using a dedicated cost function that improves accuracy, particularly for anisotropic footprints. Despite using only eight input images per view, our approach generalizes to unseen objects with complex geometry and material properties, achieving high-fidelity reconstructions while requiring orders of magnitude fewer images than prior methods.
Chinese Translation
我们提出了一种自监督预训练框架，用于从最小输入中学习地下散射（SSS）光传输表示。我们的方法利用立体投影仪-相机设置，每个视角仅捕获八幅高频相位偏移轮廓图（PSP）图像，以在多视角、多物体环境中预训练编码器。我们引入了一种针对基于PSP的SSS数据的定制增强策略，并显示其在SSL预训练中显著优于标准的ImageNet风格增强。预训练的编码器学习到可迁移的SSS表示，能够有效转移到下游任务，包括空间变化重光照和使用kNN分类器的表示评估。结合解码器，该模型重建密集的散射足迹响应，使用专门的成本函数进行训练，从而提高准确性，特别是对于各向异性足迹。尽管每个视角仅使用八幅输入图像，我们的方法仍能推广到具有复杂几何形状和材料特性的未见物体，实现高保真重建，同时所需图像数量比以前的方法少几个数量级。
cs.CV / 122 / 2606.29462
MIRROR: Aligning Semantic Relations from Language to Image via Gromov--Wasserstein
MIRROR：通过 Gromov--Wasserstein 从语言到图像对齐语义关系
Abstract
Multimodal Large Language Models (MLLMs) inherit rich relational priors from their language backbones, yet often fail when asked to apply these relationships in visual contexts. We trace this failure to a structural blind spot: projection-based alignment trains each visual token to carry the right semantics, but never asks whether the relationships between concepts survive the crossing from language to vision. To address this, we propose MIRROR (Mapping Inter-concept Relations from language to visual Representation via Optimal-transport-based Regularization), a geometric regularization framework that transfers relational priors from language to vision by exploiting the rich relational structure encoded in language representations. Specifically, we derive a surrogate loss from the proposed Semi-Inverse Gromov-Wasserstein (SI-GW) problem, an inverse geometric problem that aligns visual representations with language-derived relational priors. We show that this formulation admits a unique closed-form solution that prescribes the ideal visual relational structure implied by language geometry and cross-modal coupling. The structure of the formulation also enables efficient computation, making it applicable to long token sequences. Applying SI-GW inside decoder-only Transformers requires careful design. We introduce targeted strategies at the layer, head, and token levels to ensure stable extraction without additional parameters or inference cost. MIRROR improves relational consistency while preserving performance on general vision-language tasks.
Chinese Translation
多模态大型语言模型（MLLMs）从其语言基础中继承了丰富的关系先验，但在要求将这些关系应用于视觉上下文时常常失败。我们将这种失败追溯到一个结构性盲点：基于投影的对齐训练每个视觉标记携带正确的语义，但从未询问概念之间的关系是否在从语言到视觉的转换中得以保留。为了解决这个问题，我们提出了 MIRROR（通过基于最优传输的正则化将语言中的概念关系映射到视觉表征），这是一个几何正则化框架，通过利用语言表征中编码的丰富关系结构，将关系先验从语言转移到视觉。具体而言，我们从提出的半逆 Gromov-Wasserstein（SI-GW）问题中推导出一个替代损失，这是一个逆几何问题，旨在将视觉表征与源自语言的关系先验对齐。我们展示了该公式具有唯一的闭式解，规定了语言几何和跨模态耦合所暗示的理想视觉关系结构。该公式的结构还使得高效计算成为可能，从而适用于长标记序列。在仅解码器的 Transformer 中应用 SI-GW 需要仔细设计。我们在层、头和标记级别引入了针对性的策略，以确保在没有额外参数或推理成本的情况下稳定提取。MIRROR 提高了关系一致性，同时保持了在一般视觉-语言任务上的性能。
cs.CV / 123 / 2606.29463
CellDETR: A Detection-Guided Framework for Scalable Cell Representation Learning from Histopathology Images
CellDETR：一种基于检测引导的可扩展细胞表示学习框架，适用于组织病理学图像
Abstract
Recent advances in pathology foundation models have substantially improved patch and slide level representation learning from whole-slide images (WSIs).However, cell-level representations learning remain underexplored, limiting cell resolved interpretability, biological discovery, and clinical translation. We propose CellDETR, a detection-guided framework built on Deformable DETR for scalable cell representation learning from WSIs. By introducing location feature decoupling and box-constrained attention mechanism, CellDETR enables automated extraction of cell-level embeddings, and outperform existing state-of-the-art methods in supervised cell classification on PanNuke data. In addition, by incorporating contrastive learning design, we build a CellDETR-based pretraining model for scalable cell representation learning from unlabeled WSIs, which improves downstream cell classification performance. Furthermore, we show that after pretraining with Xenium spatial transcriptomics-derived cell annotations, CellDETR achieves accurate cross-dataset cell classification, demonstrating the transferability and biological relevance of the learned cell embeddings. Together, CellDETR provides a scalable route toward general cell-level representation learning framework for interpretable computational patholog
Chinese Translation
最近在病理基础模型方面的进展显著提高了从全切片图像（WSIs）中进行的补丁和切片级表示学习。然而，细胞级表示学习仍未得到充分探索，这限制了细胞解析的可解释性、生物发现和临床转化。我们提出了CellDETR，这是一种基于可变形DETR的检测引导框架，用于从WSIs中进行可扩展的细胞表示学习。通过引入位置特征解耦和框约束注意机制，CellDETR能够自动提取细胞级嵌入，并在PanNuke数据集上超越现有的最先进方法，在监督细胞分类任务中表现优异。此外，通过结合对比学习设计，我们构建了一个基于CellDETR的预训练模型，用于从未标记的WSIs中进行可扩展的细胞表示学习，从而提高下游细胞分类性能。此外，我们展示了在使用Xenium空间转录组学衍生的细胞注释进行预训练后，CellDETR实现了准确的跨数据集细胞分类，证明了所学习细胞嵌入的可迁移性和生物相关性。总之，CellDETR为可解释的计算病理学提供了一个通用的细胞级表示学习框架的可扩展路径。
cs.CV / 124 / 2606.29464
Rank-Aware Hyperbolic Alignment for Vision-Language Dataset Distillation
基于秩感知的超曲面对齐用于视觉-语言数据集蒸馏
Abstract
Vision-language dataset distillation (VLDD) compresses a large image-text paired dataset into a small set of synthetic pairs that can efficiently train contrastive vision-language models under strict data and compute budgets. Most existing methods match expert trajectories or cross-modal statistics, yet still enforce full-dimensional alignment in a Euclidean embedding space. This is often overly restrictive due to rank-deficient image--text correlation, with shared semantics concentrated in a low-dimensional range and remaining variation spread across a weakly correlated residual subspace. LoRS relaxes alignment at the similarity level by low-rank factorization, but does not explicitly control dominant alignment capacity and structure in the representation space. We thus propose a rank-aware hyperbolic alignment (RAHA) that combines hierarchical geometry with explicit alignment-capacity control. RAHA lifts multimodal representations to hyperbolic space and optimizes distilled pairs with asymmetric objectives that enforce geodesic alignment in the shared range while regularizing the residual subspace to preserve modality-private diversity and improve transfer robustness. Experiments on benchmarks show that RAHA demonstrates competitive cross-modal retrieval and improved transfer indicators under fixed budgets.
Chinese Translation
视觉-语言数据集蒸馏（VLDD）将大型图像-文本配对数据集压缩为一小组合成对，以便在严格的数据和计算预算下有效地训练对比视觉-语言模型。现有的大多数方法匹配专家轨迹或跨模态统计，但仍在欧几里得嵌入空间中强制执行全维对齐。这通常由于图像-文本相关性的秩缺陷而过于严格，共享语义集中在低维范围内，而剩余变异则分散在弱相关的残差子空间中。LoRS通过低秩分解放宽了相似性层面的对齐，但并未明确控制表示空间中的主导对齐能力和结构。因此，我们提出了一种基于秩感知的超曲面对齐（RAHA），它将层次几何与明确的对齐能力控制相结合。RAHA将多模态表示提升到超曲面空间，并使用不对称目标优化蒸馏对，强制在共享范围内执行测地线对齐，同时对残差子空间进行正则化，以保持模态私有多样性并提高迁移鲁棒性。在基准测试中的实验表明，RAHA在固定预算下展示了竞争力的跨模态检索能力和改进的迁移指标。
cs.CV / 125 / 2606.29473
MAVIN: Multi-Shot Audio-Visual Generation with Narrative Control
MAVIN：具有叙事控制的多镜头音视频生成
Abstract
While recent generative models produce high-fidelity videos, they struggle with the complex narrative control required for coherent multi-shot audio-visual generation. Existing methods suffer from temporal misalignment, limited controllability, and incomplete scripting. In this paper, we propose MAVIN, the first framework for multi-shot audio-visual generation with customized narrative control. To resolve temporal misalignment, we propose boundary-aware attention, which leverages hierarchical captions and boundary-aware token routing to render audio-visual elements within their respective temporal boundaries. To improve the controllability for multi-subject scenarios, we propose ID-aware propagation, utilizing identity embeddings and an identity-aware mask to bind specific identities to consistent visual appearances and vocal timbres. To provide comprehensive audio-visual narratives, we present a multi-agent scripting pipeline to transform free-form user inputs into hierarchical captions. Furthermore, we construct MAVINSet, a multi-shot audio-visual dataset for robust training and evaluation. Extensive experiments demonstrate that MAVIN achieves state-of-the-art performance, opening up a new avenue for integrating generative models into professional filmmaking workflows.
Chinese Translation
尽管近期的生成模型能够生成高保真视频，但它们在实现连贯的多镜头音视频生成所需的复杂叙事控制方面仍然面临挑战。现有方法存在时间对齐不足、可控性有限和脚本不完整等问题。本文提出了MAVIN，这是第一个具有定制叙事控制的多镜头音视频生成框架。为了解决时间对齐问题，我们提出了边界感知注意力（boundary-aware attention），该方法利用层次化字幕和边界感知的标记路由（token routing）来在各自的时间边界内呈现音视频元素。为了提高多主体场景的可控性，我们提出了身份感知传播（ID-aware propagation），利用身份嵌入（identity embeddings）和身份感知掩码（identity-aware mask）将特定身份绑定到一致的视觉外观和声音音色。为了提供全面的音视频叙事，我们提出了一个多主体脚本生成管道，将自由形式的用户输入转化为层次化字幕。此外，我们构建了MAVINSet，这是一个用于稳健训练和评估的多镜头音视频数据集。大量实验表明，MAVIN达到了最先进的性能，为将生成模型整合到专业电影制作工作流程中开辟了新的途径。
cs.CV / 126 / 2606.29494
VCS-SLAM: Geometry-Validated Semantic Evidence Fusion for 3D Gaussian SLAM
VCS-SLAM：用于3D高斯SLAM的几何验证语义证据融合
Abstract
Visual SLAM performance often deteriorates in complex real-world applications. Semantic 3D Gaussian SLAM commonly fuses 2D semantic priors into a persistent 3D map using uniform optimization weights. However, such priors are not equally reliable in online mapping: occlusions, unsupported semantic boundaries, and ambiguous ray geometry can introduce persistent semantic artifacts into the global Gaussian map. We propose VCS-SLAM, a geometry-validated semantic evidence fusion framework for RGB-D 3D Gaussian SLAM. Instead of treating all semantic observations as uniformly valid supervision, VCS-SLAM evaluates their geometric reliability through visibility consistency, surface-supported boundary evidence, and ray-level conflict uncertainty. The resulting reliability-aware objective suppresses occluded semantic updates, reduces unsupported semantic bleeding, and delays premature label assignment in ambiguous regions. Experiments on Replica demonstrate improved semantic consistency, boundary preservation, and reconstruction quality. Results on ScanNet further show that VCS-SLAM maintains competitive tracking performance under real RGB-D inputs
Chinese Translation
视觉SLAM在复杂的现实应用中性能往往会下降。语义3D高斯SLAM通常使用统一的优化权重将2D语义先验融合到持久的3D地图中。然而，这些先验在在线映射中并不具有同等的可靠性：遮挡、无支撑的语义边界和模糊的光线几何可能会在全局高斯地图中引入持久的语义伪影。我们提出了VCS-SLAM，一个用于RGB-D 3D高斯SLAM的几何验证语义证据融合框架。VCS-SLAM并不将所有语义观测视为统一有效的监督，而是通过可见性一致性、表面支撑的边界证据和光线级冲突不确定性来评估它们的几何可靠性。由此产生的可靠性感知目标抑制了被遮挡的语义更新，减少了无支撑的语义渗漏，并延迟了在模糊区域的过早标签分配。在Replica上的实验表明，VCS-SLAM在语义一致性、边界保持和重建质量方面有所改善。在ScanNet上的结果进一步表明，VCS-SLAM在真实RGB-D输入下保持了竞争性的跟踪性能。
cs.CV / 127 / 2606.29496
Rectifying Mask via Entropy for Distractor-Free 3DGS in Ambiguous Scenarios
通过熵校正掩膜实现无干扰的3D生成在模糊场景中的应用
Abstract
We present RefineSplat, a systematic framework that effectively constructs transient masks to identify diverse ambiguous distractors. To do this, we qualitatively and quantitatively analyze issues and propose a novel entropy-aware adaptive masking method. Unlike existing approaches that struggle to distinguish transient elements from static scenes due to color or semantic ambiguity, RefineSplat captures ambiguous distractors leveraging entropy and instance masks. Furthermore, we propose a simple yet effective entropy-aware density control to align Gaussians in ambiguous scenarios considering Entropy-aware positional gradients. Additionally, to rigorously validate our method, we first create and release the Ambiguous wild dataset, including 18 scenes where distractors and static scenes are hard to distinguish due to color or semantic resemblances. Experimental results on various datasets demonstrate that RefineSplat shows state-of-the-art performance, showing distractor-free novel view synthesis.
Chinese Translation
我们提出了RefineSplat，一个系统化框架，能够有效构建瞬态掩膜以识别多样的模糊干扰物。为此，我们对相关问题进行了定性和定量分析，并提出了一种新颖的基于熵的自适应掩膜方法。与现有方法在颜色或语义模糊的情况下难以区分瞬态元素与静态场景不同，RefineSplat利用熵和实例掩膜捕捉模糊干扰物。此外，我们提出了一种简单而有效的基于熵的密度控制方法，以考虑熵感知的位置信息梯度，在模糊场景中对高斯分布进行对齐。此外，为了严格验证我们的方法，我们首先创建并发布了模糊野外数据集，包括18个场景，其中由于颜色或语义相似性，干扰物与静态场景难以区分。各种数据集上的实验结果表明，RefineSplat展示了最先进的性能，能够实现无干扰的新视图合成。
cs.CV / 128 / 2606.29498
Learning Where and When: Patch-Based Spatiotemporal Localization in Weakly Supervised Video Anomaly Detection
学习何时何地：基于补丁的弱监督视频异常检测中的时空定位
Abstract
Weakly supervised video anomaly detection (WSVAD) has predominantly focused on temporal localization, identifying when anomalies occur while largely neglecting their spatial extent within frames. Yet, spatial localization is essential for interpretability and practical deployment in real-world settings. We introduce a patch-based spatiotemporal framework for weakly supervised anomaly localization that jointly models where and when anomalies occur. Our approach operates on grid-level patch features and learns region-level anomaly scores under a multiple instance learning paradigm. We further propose a Proximity-Aware Top-k spatiotemporal selection strategy that enables the model to generate fine-grained spatial anomaly maps without requiring bounding-box supervision during training. Our method surpasses existing state-of-the-art approaches across multiple benchmarks, yielding substantial gains in spatiotemporal localization accuracy. In addition, we release frame-level bounding-box annotations for the test sets of two widely used datasets, along with our code and pretrained models, providing new resources to facilitate future research in spatially grounded WSVAD.
Chinese Translation
弱监督视频异常检测（WSVAD）主要集中于时间定位，识别异常发生的时间，而在很大程度上忽视了异常在帧内的空间范围。然而，空间定位对于可解释性和在现实环境中的实际应用至关重要。我们提出了一种基于补丁的时空框架，用于弱监督异常定位，该框架联合建模异常发生的时间和空间。我们的方法基于网格级补丁特征，并在多实例学习范式下学习区域级异常评分。我们进一步提出了一种邻近感知的Top-k时空选择策略，使模型能够生成细粒度的空间异常图，而无需在训练期间进行边界框监督。我们的方法在多个基准测试中超越了现有的最先进方法，在时空定位准确性上取得了显著提升。此外，我们为两个广泛使用的数据集的测试集发布了帧级边界框注释，以及我们的代码和预训练模型，为促进未来在空间基础的WSVAD研究提供了新的资源。
cs.CV / 129 / 2606.29504
Empirical Evaluation of Multi-Modal Touch Detection in Over-the-Shoulder Video Surveillance
肩上视频监控中多模态触控检测的实证评估
Abstract
Video Intelligence Surveillance (VIDINT) on over-the-shoulder footage is a proposed vector for monitoring human-computer interaction patterns without direct screen recording access. In this paper, we evaluate a Behavioral Intelligence (BEHINT) touch-detection framework designed to reconstruct keystroke events on mobile keypad interfaces from physical finger interactions. Our system integrates four parallel detection modalities: (1) anatomical hand landmarks via MediaPipe, (2) HSV skin color filtering, (3) temporal frame differencing for motion detection, and (4) shape-guided Canny edge analysis. We map relative touch coordinates to a reference screen layout to reconstruct typing sequences. Evaluation on a 120-frame first-person staged video of passcode entry reveals that while MediaPipe and Skin Detection fail to run autonomously due to partial hand occlusion and ambient noise, Motion-Only and Edge-Only configurations achieve F1-scores of 18.5% and 18.2%, respectively. The combined multi-modal configuration achieves an F1-score of 16.7% and a sequence similarity of 3.0% when mapped to the iOS passcode layout. We conduct ablation, resolution decay, noise sensitivity, and proximity threshold tuning to characterize the system's operational envelope. We then audit generalization on 5 real, publicly licensed third-person phone videos and find that the detector emits a median of 57 touch points per frame (peaking at 205), one to three orders of magnitude more than the rate of real taps, because the skin filter responds to the whole hand rather than to fingertip contact. The staged keystroke result does not survive contact with uncontrolled footage; the system does not achieve reliable keystroke reconstruction outside the calibrated staged setting.
Chinese Translation
肩上视频监控中的视频智能监控（VIDINT）是一种监测人机交互模式的提议方法，无需直接访问屏幕录制。在本文中，我们评估了一种行为智能（BEHINT）触控检测框架，旨在从物理手指交互中重建移动键盘界面的击键事件。我们的系统集成了四种并行检测模式：（1）通过MediaPipe获取的解剖手部标志，（2）HSV肤色过滤，（3）用于运动检测的时间帧差分，以及（4）形状引导的Canny边缘分析。我们将相对触控坐标映射到参考屏幕布局，以重建输入序列。在对一段120帧的第一人称密码输入 staged 视频的评估中发现，尽管MediaPipe和肤色检测由于部分手部遮挡和环境噪声未能自主运行，运动仅和边缘仅配置分别达到了18.5%和18.2%的F1分数。结合的多模态配置在映射到iOS密码布局时达到了16.7%的F1分数和3.0%的序列相似性。我们进行了消融实验、分辨率衰减、噪声敏感性和接近阈值调优，以表征系统的操作范围。随后，我们对5个真实的、公开授权的第三人称手机视频进行了泛化审计，发现检测器每帧发出中位数为57个触控点（峰值为205），比真实触碰的速率高出一个到三个数量级，因为肤色过滤器对整个手部而非指尖接触做出反应。经过控制的staged击键结果在与不受控视频接触后未能存活；该系统在校准的staged设置之外未能实现可靠的击键重建。
cs.CV / 130 / 2606.29506
Benchmark AUC Is Not Deployable Reliability: A Cross-Dataset Audit of Off-the-Shelf Features for Surveillance Video Anomaly Detection
基准AUC并不可部署的可靠性：对现成特征在监控视频异常检测中的跨数据集审计
Abstract
Automated "suspicious behavior" flagging is a headline promise of AI surveillance, and the field reports high frame-level ROC-AUC on standard video anomaly detection benchmarks. Those numbers are measured by training and testing on the same camera and scene. We audit what happens when that assumption is dropped. We build an unsupervised normality model from the all-normal training frames of one dataset, using frozen off-the-shelf embeddings (CLIP, DINOv2, ResNet-50, EfficientNet-B0) and a nearest-neighbour distance, and score the test frames of the same and of other datasets. Across 4 real datasets (UCSD Ped1, UCSD Ped2, CUHK Avenue, ShanghaiTech) and 4 backbones, same-dataset AUC averages 0.704 but cross-dataset AUC averages 0.499, which is chance: a detector calibrated on one scene is no better than a coin flip on another, and in several pairs it is below chance. The strongest backbone makes this worse, not better: DINOv2 has the best same-dataset AUC (up to 0.901 on Ped2) and the largest cross-dataset drop. The collapse is not an artefact of the scoring rule: replacing the nearest-neighbour detector with a PaDiM-style Mahalanobis detector reproduces it almost exactly (cross-dataset gap 0.202 versus 0.208). Even at a favourable operating point the false-alarm rate is on the order of 31,931 per hour. We conclude that the benchmark numbers quoted for surveillance anomaly detection describe a calibrated laboratory setting and overstate deployable reliability by a wide margin, and we release the code that reproduces every number.
Chinese Translation
自动化的“可疑行为”标记是人工智能监控的一个重要承诺，该领域在标准视频异常检测基准上报告了高帧级ROC-AUC。这些数字是在同一摄像头和场景下进行训练和测试时测得的。我们审计了当这一假设被放弃时会发生什么。我们从一个数据集的所有正常训练帧构建了一个无监督的正常性模型，使用了冻结的现成嵌入（CLIP、DINOv2、ResNet-50、EfficientNet-B0）和最近邻距离，并对同一数据集及其他数据集的测试帧进行评分。在4个真实数据集（UCSD Ped1、UCSD Ped2、CUHK Avenue、ShanghaiTech）和4个基础模型中，同数据集的AUC平均值为0.704，但跨数据集的AUC平均值为0.499，这等同于随机猜测：在一个场景上校准的检测器在另一个场景上并没有比抛硬币更好，在几个数据对中甚至低于随机猜测。最强的基础模型使情况更糟，而不是更好：DINOv2在同数据集的AUC（在Ped2上高达0.901）最好，但跨数据集的下降幅度最大。这一崩溃并不是评分规则的伪影：用PaDiM风格的Mahalanobis检测器替换最近邻检测器几乎完全重现了这一现象（跨数据集差距0.202与0.208）。即使在有利的操作点，误报率也在每小时31,931次的数量级上。我们得出结论，监控异常检测的基准数字描述的是一个经过校准的实验室环境，并大幅夸大了可部署的可靠性，我们发布了重现每个数字的代码。
cs.CV / 131 / 2606.29513
Scenes as Objects, Not Primitives: Instance-Structured 3D Tokenization from Unposed Views
场景作为对象，而非原始元素：来自无姿态视图的实例结构化3D标记化
Abstract
A 3D scene is understood through its objects, not the primitives that compose them. Yet feed-forward reconstruction methods output dense, unstructured sets of points or Gaussians, leaving object-level structure to be recovered after the fact. We propose a feed-forward framework that decomposes a scene into instance-structured 3D token groups directly from unposed multi-view images -- compact object-centric units from which reconstruction, segmentation, and manipulation all follow. Each token group pairs an instance token capturing entity-level identity with anchor tokens that encode local geometry and appearance, which are decoded into a set of 3D Gaussians. This two-level factorization decouples object identity from local appearance, making object instances a native interface of the representation rather than a derived product. The token groups are learned through differentiable rendering with joint reconstruction and segmentation supervision, requiring no 3D annotations. Our feed-forward model surpasses per-scene optimization baselines in class-agnostic instance segmentation while remaining competitive in novel view synthesis. Beyond these metrics, the same token groups directly unlock instance-level scene editing -- removing, translating, or inserting objects by operating on their groups -- as well as efficient open-vocabulary 3D instance retrieval, where retrieval complexity scales with the number of instances rather than primitives.
Chinese Translation
3D场景是通过其对象来理解的，而不是构成它们的原始元素。然而，前馈重建方法输出的是密集的、非结构化的点或高斯集合，留下对象级结构在事后恢复。我们提出了一种前馈框架，该框架直接从无姿态的多视图图像中将场景分解为实例结构化的3D标记组——紧凑的以对象为中心的单元，从中可以进行重建、分割和操作。每个标记组将捕捉实体级身份的实例标记与编码局部几何和外观的锚标记配对，这些锚标记被解码为一组3D高斯。这种两级因式分解将对象身份与局部外观解耦，使对象实例成为表示的原生接口，而不是派生产品。标记组通过可微渲染与联合重建和分割监督进行学习，无需3D注释。我们的前馈模型在类别无关的实例分割中超越了每场景优化基线，同时在新视图合成中保持竞争力。超越这些指标，相同的标记组直接解锁实例级场景编辑——通过操作其组来移除、平移或插入对象——以及高效的开放词汇3D实例检索，其中检索复杂度与实例数量而非原始元素数量成比例。
cs.CV / 132 / 2606.29531
MotionAtlas: Detailed Region Captioning for Motion-Centric Videos
MotionAtlas：针对运动中心视频的详细区域描述
Abstract
We propose MotionAtlas, a system for detailed captioning of motion-centric videos, comprising (1) a dedicated human-annotated benchmark, (2) a scalable, high-quality pipeline to construct training samples, and (3) a family of powerful Video-MLLMs. Unlike conventional global motion captioning datasets, we focus on region-aware motion captioning: given a video and a spatiotemporal mask, the model generates precise descriptions of motion within the target region, thereby alleviating visual clutter and motion entanglement and enabling reliable, quantifiable evaluation. Concretely, we first build MotionAtlas-Bench, a comprehensive benchmark comprising 2,073 multiple-choice questions, meticulously annotated for a curated set of high-quality, motion-centric videos, to evaluate fine-grained motion understanding of the objects in question. Second, we design a rigorous and scalable data pipeline that leverages self-bootstrap refinement to suppress fine-grained hallucinations, yielding 159k high-quality motion captioning data. Third, we design a tailored training data composition strategy, which achieves consistent and substantial performance gains across diverse baseline Video-MLLMs, including Molmo2 and Qwen3-VL. For instance, MotionAtlas-4B surpasses Qwen3-VL-4B by an average of 5.2 percentage points across general motion benchmarks. The benchmark, dataset, and code have been released.
Chinese Translation
我们提出了MotionAtlas，一个用于运动中心视频详细描述的系统，包含（1）一个专门的人类标注基准， （2）一个可扩展的高质量管道用于构建训练样本，以及（3）一系列强大的视频多模态大语言模型（Video-MLLMs）。与传统的全局运动描述数据集不同，我们专注于区域感知的运动描述：给定一个视频和一个时空掩码，模型生成目标区域内运动的精确描述，从而减轻视觉杂乱和运动纠缠，并实现可靠、可量化的评估。具体而言，我们首先构建了MotionAtlas-Bench，这是一个综合基准，包含2,073个多项选择题，经过精心注释，针对一组高质量的运动中心视频，以评估相关对象的细粒度运动理解。其次，我们设计了一个严格且可扩展的数据管道，利用自我引导精炼来抑制细粒度的幻觉，生成了159k高质量的运动描述数据。第三，我们设计了一种量身定制的训练数据组成策略，在包括Molmo2和Qwen3-VL在内的多种基线视频多模态大语言模型中，实现了一致且显著的性能提升。例如，MotionAtlas-4B在一般运动基准测试中平均超越Qwen3-VL-4B 5.2个百分点。该基准、数据集和代码已被发布。
cs.CV / 133 / 2606.29535
GarmentZoom: Generating Zoomable Images from Garment Listings
GarmentZoom：从服装列表生成可缩放图像
Abstract
Online product listings for garments often include an overview photo and a close-up to show garment details. However, each photo focuses on either field of view or garment detail, forcing users to alternate between views and breaking browsing continuity. We present GarmentZoom, a system that enhances the full-view photo to match the fidelity of its accompanying close-up, enabling seamless zoom-and-pan exploration. Unlike standard reference-based super-resolution, our setting involves close-up references that are spatially unaligned with the full view, and scale factors that vary substantially across garments 3-20$\times$. Prior work typically relies on alignment to transfer details or requires per-instance fine-tuning to memorize them. Instead, we train a single model that supports a continuous range of scales across diverse garments. Our approach synthesizes details without requiring spatial alignment and matches the quality of per-instance methods with a fraction of the training cost.
Chinese Translation
在线服装产品列表通常包括一张概览照片和一张特写照片，以展示服装细节。然而，每张照片要么专注于视野，要么专注于服装细节，迫使用户在不同视图之间切换，从而打断浏览的连续性。我们提出了GarmentZoom，一个增强全景照片以匹配其伴随特写照片清晰度的系统，从而实现无缝的缩放和平移探索。与标准的基于参考的超分辨率方法不同，我们的设置涉及与全景图空间不对齐的特写参考，并且不同服装的缩放因子在3到20倍之间变化。以往的研究通常依赖于对齐来传递细节，或需要针对每个实例进行微调以记忆这些细节。相反，我们训练了一个单一模型，支持在多样化服装之间的连续缩放范围。我们的方法在不需要空间对齐的情况下合成细节，并以较低的训练成本匹配每个实例方法的质量。
cs.CV / 134 / 2606.29573
Reliability-Prioritized Fine-Grained Generation in Multimodal Large
以可靠性为优先的细粒度生成在多模态大型模型中的应用
Abstract
Multimodal large language models (MLLMs) are increasingly expected to generate fine-grained descriptions of visual content. However, we observe and theoretically show that generating fine-grained responses poses a reliability challenge, \textit{i.e.}, fine-grained generation is more error-prone than coarse-grained generation. This phenomenon suggests that models should generate the finest description that remains reliable rather than simply produce more specific outputs. To investigate this problem, we develop \textsc{GranFact}, a granularity-aware benchmark consisting of expert-verified multi-object images with coarse-to-fine category annotations. Then, we design a hierarchy-aware evaluation algorithm, which assesses both whether model predictions are visually correct and how specific the correct predictions are. We also propose a reliability-prioritized preference optimization method based on Direct Preference Optimization, which penalizes unreliable fine-grained claims while rewarding reliable specificity. Experiments on \textsc{GranFact} show that our method improves fine-grained generation while preserving reliability. Code and data are available \href{https://github.com/WeiWu2025/GranFact}{here}.
Chinese Translation
多模态大型语言模型（MLLMs）越来越被期望生成对视觉内容的细粒度描述。然而，我们观察到并理论上证明，生成细粒度响应存在可靠性挑战，即细粒度生成比粗粒度生成更容易出错。这一现象表明，模型应生成保持可靠性的最细描述，而不仅仅是产生更具体的输出。为了解决这一问题，我们开发了 extsc{GranFact}，这是一个关注粒度的基准，包含经过专家验证的多对象图像及其从粗到细的类别注释。随后，我们设计了一种层次感知的评估算法，该算法评估模型预测是否在视觉上正确以及正确预测的具体程度。我们还提出了一种基于直接偏好优化（Direct Preference Optimization）的以可靠性为优先的偏好优化方法，该方法对不可靠的细粒度声明进行惩罚，同时奖励可靠的具体性。在 extsc{GranFact}上的实验表明，我们的方法在提高细粒度生成的同时保持了可靠性。代码和数据可在此获取： exttt{https://github.com/WeiWu2025/GranFact}。
cs.CV / 135 / 2606.29577
ReMAP-PET: Beyond Visual Understanding -- Learning Region-Guided Metabolic Alignment Semantics from Brain PET
ReMAP-PET：超越视觉理解——从脑部PET学习区域引导的代谢对齐语义
Abstract
Positron Emission Tomography (PET) reveals brain metabolism and is clinically central to neurodegenerative disease assessment, yet existing 3D brain foundation models treat PET as generic volumetric data, missing the structured regional metabolic information that distinguishes it from structural neuroimaging. To address these limitations, we propose ReMAP-PET, a framework that moves beyond visual encoding by supervising a partially-tuned MedicalNet 3D ResNet-50 with brain regional standardized uptake value ratio (SUVR) profiles through joint regression and contrastive objectives, enabling the encoder to learn the metabolic semantics underlying PET modality. On 1015 paired PET--SUVR samples, ReMAP-PET achieves 0.070 SUVR MAE and 77.8% PET SUVR Recall@1, substantially outperforming five frozen pretrained baselines. We further connect the metabolic embedding to clinical language via contrastive alignment with frozen BioClinicalBERT and demonstrate end-to-end PET-to-report generation through SUVR-constrained verbalization. Linear probing on diagnostic classification and cognitive regression tasks confirms that the embeddings retain clinically relevant information without task-specific fine-tuning. Our results show that grounding PET encoders in regional metabolic semantics -- rather than treating PET as generic volumetric data -- yields representations that are structured, interpretable, and language-compatible, pointing to a new direction for metabolic-aware PET understanding.
Chinese Translation
正电子发射断层扫描（PET）揭示了大脑代谢，并在神经退行性疾病评估中具有临床中心地位。然而，现有的3D脑部基础模型将PET视为通用体积数据，忽视了其与结构神经影像学区别开来的结构化区域代谢信息。为了解决这些局限性，我们提出了ReMAP-PET，一个超越视觉编码的框架，通过联合回归和对比目标，利用脑部区域标准摄取值比（SUVR）轮廓对部分调优的MedicalNet 3D ResNet-50进行监督，从而使编码器能够学习PET模态下的代谢语义。在1015对PET-SUVR样本上，ReMAP-PET实现了0.070 SUVR平均绝对误差（MAE）和77.8%的PET SUVR召回率@1，显著优于五个冻结的预训练基线。我们进一步通过与冻结的BioClinicalBERT进行对比对齐，将代谢嵌入与临床语言连接，并展示了通过SUVR约束的语言化实现端到端的PET报告生成。在诊断分类和认知回归任务上的线性探测确认了这些嵌入保留了临床相关信息，而无需特定任务的微调。我们的结果表明，将PET编码器基于区域代谢语义进行基础化——而不是将PET视为通用体积数据——能够产生结构化、可解释且与语言兼容的表示，指向代谢感知PET理解的新方向。
cs.CV / 136 / 2606.29579
ScAle: Attention Head Scaling as a Minimal Adapter for Spatial Reasoning in Vision Language Models
ScAle：作为视觉语言模型空间推理的最小适配器的注意力头缩放
Abstract
Spatial reasoning remains a persistent challenge for many vision language models (VLMs), and improving it typically requires fine-tuning with substantial additional parameters. Our preliminary analysis reveals that rescaling activations in selected transformer layers-without modifying pretrained weights-can significantly influence downstream performance. Motivated by this observation, we propose ScAle, an ultra-lightweight adaptation method that learns a small set of scalar coefficients to modulate last-token attention and MLP activations in a fully frozen backbone. We evaluate our method on the synthetic spatial reasoning benchmark SpatialEval and on real-world VQA datasets (COCOQA and VGQA) across multiple model families. Our method, ScAle, achieves up to 134.1% relative accuracy gains using only 1K trainable parameters without requiring millions of trainable parameters as in standard PEFT methods such as LoRA. Despite its extreme compactness, our approach recovers a substantial fraction of standard PEFT performance while preserving strong non-spatial VQA accuracy. These results demonstrate that bounded activation reweighting provides a simple, architecture-agnostic, and highly parameter-efficient alternative for adapting pretrained VLMs.
Chinese Translation
空间推理仍然是许多视觉语言模型（VLMs）面临的持续挑战，改善这一点通常需要通过大量额外参数进行微调。我们的初步分析表明，在不修改预训练权重的情况下，重新缩放选定变换器层中的激活值可以显著影响下游性能。基于这一观察，我们提出了ScAle，这是一种超轻量级适配方法，学习一小组标量系数，以调节完全冻结的主干网络中的最后一个令牌注意力和多层感知机（MLP）激活。我们在合成空间推理基准SpatialEval和多个真实世界视觉问答（VQA）数据集（COCOQA和VGQA）上评估了我们的方法。我们的ScAle方法在仅使用1000个可训练参数的情况下，实现了高达134.1%的相对准确率提升，而不需要像标准的参数高效微调（PEFT）方法（如LoRA）那样的数百万可训练参数。尽管其极端紧凑，我们的方法在保持强大的非空间VQA准确率的同时，恢复了标准PEFT性能的相当一部分。这些结果表明，有限的激活重加权提供了一种简单、与架构无关且高度参数高效的替代方案，用于适配预训练的VLMs。
cs.CV / 137 / 2606.29586
SonoCLIP: Mask-Guided Region-Aware Vision-Language Pretraining for Fetal Ultrasound Analysis
SonoCLIP：面向胎儿超声分析的掩膜引导区域感知视觉-语言预训练
Abstract
Vision-language foundation models have shown strong potential in medical image analysis. Although foundation models for ultrasound imaging have recently emerged, the domain remains particularly challenging due to severe speckle noise, acquisition variability, and subtle anatomical boundaries, leading to high inter-observer variability. Existing CLIP-based models rely primarily on global image-text alignment, limiting their sensitivity to clinically decisive local structures. We propose SonoCLIP, the first million-scale region-controllable fetal ultrasound vision-language foundation model that integrates segmentation masks as mask-channel visual prompts within the vision encoder, enabling joint global-local contrastive representation learning. To support scalable region-text alignment, we introduce a sigmoid-based pairwise contrastive loss that improves stability under large-scale supervision. We further curate a 1.44M-image multimodal fetal ultrasound dataset spanning 24 standard planes for large-scale pretraining. Extensive cross-center evaluations demonstrate that SonoCLIP achieves superior zero-shot transfer performance under both global and mask-guided inference, establishing a controllable and clinically oriented foundation model for fetal ultrasound analysis. Our code and data are available at https://github.com/Harrison-one/SonoCLIP.
Chinese Translation
视觉-语言基础模型在医学图像分析中展现出强大的潜力。尽管针对超声成像的基础模型最近有所出现，但由于严重的斑点噪声、获取变异性和微妙的解剖边界，该领域仍然面临特别的挑战，导致观察者间的高变异性。现有的基于CLIP的模型主要依赖于全局图像-文本对齐，限制了它们对临床决策性局部结构的敏感性。我们提出了SonoCLIP，这是第一个百万规模的区域可控胎儿超声视觉-语言基础模型，它将分割掩膜作为掩膜通道视觉提示集成到视觉编码器中，从而实现全局-局部对比表示学习的联合。为了支持可扩展的区域-文本对齐，我们引入了一种基于sigmoid的成对对比损失，在大规模监督下提高了稳定性。我们进一步整理了一个包含144万幅图像的多模态胎儿超声数据集，涵盖24个标准平面，以支持大规模预训练。广泛的跨中心评估表明，SonoCLIP在全局和掩膜引导推理下都实现了优越的零样本迁移性能，为胎儿超声分析建立了一个可控且面向临床的基础模型。我们的代码和数据可在 https://github.com/Harrison-one/SonoCLIP 获取。
cs.CV / 138 / 2606.29600
One Scene, Two Depths: Probing Geometric Ambiguity in Monocular Foundation Models
同一场景，两种深度：探讨单目基础模型中的几何模糊性
Abstract
A faithful 3D world representation should account for layered geometry, where a single camera ray may contain multiple visible and geometrically valid surfaces. Monocular depth estimation, however, reduces this structure to one scalar depth per pixel. Transparent scenes make this ambiguity measurable: the same ray can pass through foreground glass and observe the background, turning the supervised target into a convention of annotation, data, and training rather than a scene-intrinsic truth. A learned predictor exposes this convention as its depth-layer preference. We introduce MultiDepth-3k (MD-3k), a sparse two-layer ordinal benchmark for measuring depth-layer preference and multi-layer spatial relationship accuracy (ML-SRA). On MD-3k, leading depth foundation models exhibit diverse layer preferences under standard RGB input, showing that the same layered geometry can be resolved differently across models. We further find that Laplacian Visual Prompting (LVP), a training-free spectral input transformation, can substantially change the reported layer for certain frozen models. The strongest RGB/LVP pair, DAv2-L, reaches 75.5% ML-SRA. These results suggest that depth foundation models may express complementary geometric hypotheses that standard RGB inference leaves unexpressed. We invite the community to rethink depth supervision and evaluation through an ambiguity-aware lens, where multiple valid 3D interpretations are treated as geometric structure to be measured, preserved, and expressed.
Chinese Translation
一个忠实的三维世界表示应考虑分层几何结构，其中一条单摄像头光线可能包含多个可见且几何上有效的表面。然而，单目深度估计将这种结构简化为每个像素一个标量深度。透明场景使这种模糊性可测量：同一光线可以穿过前景的玻璃并观察背景，从而将监督目标转变为注释、数据和训练的约定，而非场景固有的真理。一个学习的预测器揭示了这种约定作为其深度层偏好。我们引入了MultiDepth-3k (MD-3k)，这是一个稀疏的双层序数基准，用于测量深度层偏好和多层空间关系准确性（ML-SRA）。在MD-3k上，领先的深度基础模型在标准RGB输入下表现出多样的层偏好，显示出相同的分层几何结构在不同模型中可以被不同地解析。我们进一步发现，拉普拉斯视觉提示（Laplacian Visual Prompting, LVP）作为一种无训练的光谱输入转换，可以显著改变某些冻结模型报告的层。最强的RGB/LVP组合DAv2-L达到了75.5%的ML-SRA。这些结果表明，深度基础模型可能表达了标准RGB推理未能表达的互补几何假设。我们邀请学术界通过关注模糊性的视角重新思考深度监督和评估，将多种有效的三维解释视为需要测量、保留和表达的几何结构。
cs.CV / 139 / 2606.29664
Benchmarking Geospatial Foundation Models for Agriculture Applications
农业应用的地理空间基础模型基准测试
Abstract
Geospatial foundation models pretrained on satellite imagery promise broad generalization across remote sensing tasks and regions, but their geographic transferability has not been systematically tested, especially in agriculture applications. This paper presents a controlled benchmark that evaluates three models, Prithvi, SpectralGPT, and SatMAE, on multi-temporal crop segmentation and change detection across four U.S. states, Iowa, North Carolina, California, and Minnesota. By assigning each train, validation, and test split to a separate region, we measure how well each model transfers to land it has not seen. All three degrade sharply under regional distribution shift, predicting only the most common crops while missing rare ones. We further find that fitting these models to a shared input format affects each one differently, which complicates direct architectural comparison. These results expose key limitations of current geospatial foundation models for agriculture and point to region aware evaluation as a necessary standard.
Chinese Translation
基于卫星图像预训练的地理空间基础模型在遥感任务和区域上承诺广泛的泛化能力，但其地理转移能力尚未得到系统测试，特别是在农业应用中。本文提出了一个受控基准，评估了三个模型：Prithvi、SpectralGPT 和 SatMAE，在美国四个州（爱荷华州、北卡罗来纳州、加利福尼亚州和明尼苏达州）进行多时相作物分割和变化检测。通过将每个训练、验证和测试集分配到不同的区域，我们测量了每个模型在未见过的土地上的转移能力。在区域分布变化下，所有三个模型的性能都有明显下降，只能预测最常见的作物，而遗漏了稀有作物。我们进一步发现，将这些模型拟合到共享输入格式对每个模型的影响不同，这使得直接的架构比较变得复杂。这些结果揭示了当前农业地理空间基础模型的关键局限性，并指出区域感知评估作为必要标准的重要性。
cs.CV / 140 / 2606.29667
Unlocking the Visual Record of Materials Science: A Large-Scale Multimodal Dataset from Scientific Literature
解锁材料科学的视觉记录：来自科学文献的大规模多模态数据集
Abstract
The materials science literature encodes decades of experimental knowledge in figures, yet this visual record remains locked away and inaccessible to AI at scale. The core difficulty is structural: most scientific figures are compound, with a single caption describing multiple sub-panels simultaneously, making direct image-text pairing unreliable. We present MatMMExtract, an end-to-end open-source pipeline that resolves this by decomposing compound figures into individual sub-panels and generating structured, grounded annotations using a large language model guided by a curated materials science taxonomy. Applied to 14,810 open-access articles, MatMMExtract produces MatSciFig; 391,606 panel-level image-text pairs from 180,571 figures, each annotated with a sub-caption, a two-level visualisation category spanning 19 classes and over 100 subtypes, and a scientific summary. To enable accurate panel localisation, we introduce MaterialScope, a domain-specific detection dataset of 2,811 manually annotated materials science figures, on which a fine-tuned YOLO12-m detector achieves mAP_50 of 0.9227. Among six benchmarked language models, Gemini 3.1 Flash Lite delivers the best cost-quality trade-off for annotation generation, with 82% of outputs rated good and a hallucination rate of 4.8%. A dual-encoder retrieval baseline on MatSciFig achieves a 4.4 times improvement in R@1 over zero-shot CLIP, demonstrating the dataset's immediate utility for vision-language learning. All resources are released openly to the community.
Chinese Translation
材料科学文献以图形的形式编码了数十年的实验知识，但这一视觉记录仍然被锁定，无法大规模地被人工智能访问。核心难点在于结构性：大多数科学图形是复合型的，单一的标题同时描述多个子面板，使得直接的图像-文本配对不可靠。我们提出了MatMMExtract，这是一个端到端的开源管道，通过将复合图形分解为单独的子面板，并使用经过策划的材料科学分类法指导的大型语言模型生成结构化、扎根的注释，从而解决了这一问题。应用于14,810篇开放获取文章，MatMMExtract生成了MatSciFig；从180,571个图形中产生了391,606个面板级图像-文本对，每个对都带有子标题、涵盖19个类别和100多个子类型的两级可视化类别，以及科学摘要。为了实现准确的面板定位，我们引入了MaterialScope，这是一个包含2,811个手动注释的材料科学图形的领域特定检测数据集，在该数据集上，经过微调的YOLO12-m检测器达到了0.9227的mAP_50。在六个基准语言模型中，Gemini 3.1 Flash Lite在注释生成方面提供了最佳的成本-质量平衡，82%的输出被评为良好，幻觉率为4.8%。在MatSciFig上进行的双编码器检索基线相较于零-shot CLIP实现了4.4倍的R@1提升，展示了该数据集在视觉-语言学习中的直接实用性。所有资源均向社区开放发布。
cs.CV / 141 / 2606.29686
PoseShield: Neural Collision Fields for Human Self-Collision Resolution
PoseShield：用于人类自碰撞解决的神经碰撞场
Abstract
Self-collision remains a persistent challenge in SMPL-based human pose estimation and motion generation. Under extreme articulations or stochastic motion synthesis, generated meshes frequently exhibit self-penetrations, leading to physically implausible results. We propose PoseShield, a neural collision constraint defined directly in SMPL pose space. We formulate collision correction as a constrained optimization problem and connect the learned constraint with the Eikonal equation. Enforcing Eikonal regularization ensures non-vanishing gradients near the collision boundary, improving numerical stability and robustness of the optimization process. Unlike prior methods that operate in the mesh space or rely on heuristic penalties, our approach operates directly in the low-dimensional space of human poses and is theoretically grounded. The same learned constraint extends to human motion sequences, providing a generator-agnostic post-hoc collision corrector without retraining the underlying motion model. Experiments on a newly constructed SMPL pose benchmark show that our method achieves a 95.8% success rate and outperforms state-of-the-art baselines.
Chinese Translation
自碰撞在基于SMPL的人体姿态估计和运动生成中仍然是一个持续的挑战。在极端关节运动或随机运动合成的情况下，生成的网格经常出现自我穿透，导致物理上不合理的结果。我们提出了PoseShield，一种直接在SMPL姿态空间中定义的神经碰撞约束。我们将碰撞修正公式化为一个约束优化问题，并将学习到的约束与Eikonal方程联系起来。施加Eikonal正则化确保在碰撞边界附近的梯度不消失，从而提高了优化过程的数值稳定性和鲁棒性。与以往在网格空间中操作或依赖启发式惩罚的方法不同，我们的方法直接在低维的人体姿态空间中操作，并具有理论基础。相同的学习约束可以扩展到人类运动序列，提供一种与生成器无关的后处理碰撞修正器，无需重新训练基础运动模型。在新构建的SMPL姿态基准上的实验表明，我们的方法实现了95.8%的成功率，并超越了最先进的基线。
cs.CV / 142 / 2606.29695
Progressive Self-Supervised Learning with Individualized Community Assignment for Brain Network Analysis
基于个性化社区分配的渐进式自监督学习在脑网络分析中的应用
Abstract
Brain networks exhibit a modular community structure that varies across individuals and neurological conditions. However, existing self-supervised learning (SSL) methods often overlook this heterogeneity, relying on generic masking strategies that fail to capture subject-specific functional organization. We propose BrainPICM, a self-supervised framework for brain network analysis via progressive individualized community aware masking. BrainPICM formulates ROI-to-community mapping as a progressive unbalanced optimal transport process, yielding soft assignments and per-ROI confidence scores. Guided by these confidence estimates, a curriculum-style masking strategy gradually incorporates low-confidence, potentially pathological regions into training, enabling the model to learn both stable modular structures and individual variations. Additionally, a deviation-aware aggregation module quantifies functional reorganization by measuring mass redistribution relative to a population template, enhancing interpretability and downstream prediction. Experiments on three fMRI datasets (ABIDE-I, ADHD-200, ADNI) show that BrainPICM consistently outperforms state-of-the-art supervised and SSL methods in diagnostic accuracy, indicating that explicitly injecting modular community structure into masked modeling yields more functionally consistent and generalizable representations. The source code for this approach will be released at https://github.com/Hrychen7/BrainPICM.
Chinese Translation
脑网络展现出一种模块化的社区结构，这种结构在不同个体和神经条件下存在差异。然而，现有的自监督学习（SSL）方法往往忽视这种异质性，依赖于通用的掩蔽策略，无法捕捉特定个体的功能组织。我们提出了BrainPICM，这是一种通过渐进式个性化社区感知掩蔽进行脑网络分析的自监督框架。BrainPICM将ROI（感兴趣区域）到社区的映射形式化为一个渐进的不平衡最优传输过程，产生软分配和每个ROI的置信度评分。在这些置信度估计的指导下，课程式掩蔽策略逐步将低置信度、潜在病理区域纳入训练，使模型能够学习到稳定的模块结构和个体差异。此外，偏差感知聚合模块通过相对于人群模板测量质量重分配来量化功能重组，从而增强可解释性和下游预测能力。在三个fMRI数据集（ABIDE-I、ADHD-200、ADNI）上的实验表明，BrainPICM在诊断准确性上始终优于最先进的监督和自监督学习方法，表明将模块化社区结构显式注入掩蔽建模能够产生更具功能一致性和可推广性的表征。该方法的源代码将发布在 https://github.com/Hrychen7/BrainPICM。
cs.CV / 143 / 2606.29697
MF-UAVPose6D: A Model-Free Monocular 6-DoF Pose Estimation Framework for Fixed-Wing UAVs
MF-UAVPose6D：一种无模型单目6自由度姿态估计框架用于固定翼无人机
Abstract
For uncrewed aerial vehicles (UAVs), estimating six-degree-of-freedom (6-DoF) poses is essential for airspace situational awareness, target tracking, and counter-UAV operations. However, non-cooperative targets usually lack computer-aided design (CAD) models and keypoint priors, making existing model-based or keypoint-matching methods difficult to apply reliably. To address these challenges, this paper proposes MF-UAVPose6D, a model-free monocular 6-DoF pose estimation framework for fixed-wing UAVs. During inference, the method takes only a single red-green-blue (RGB) image and camera intrinsics as input. It first obtains a stable target anchor through heatmap-guided center localization, introduces a Perspective-Aware Module (PAM) to model observation-ray priors, exploits Dynamic Topological Sampling (DTS) to complement weak structural cues from the wings, fuselage, and tail, and adopts a decoupled translation-rotation pose decoding mechanism to estimate the 6-DoF pose. In addition, we construct the FW-UAV6DPose synthetic dataset, which covers fixed-wing UAV observations across diverse distances, viewpoints, and poses. Experimental results show that MF-UAVPose6D achieves accurate and efficient monocular 6-DoF pose estimation without requiring CAD models, and demonstrates strong robustness in long-range rotation estimation, depth recovery, and joint pose evaluation.
Chinese Translation
对于无人驾驶航空器（UAV），估计六自由度（6-DoF）姿态对于空域态势感知、目标跟踪和反无人机作战至关重要。然而，非合作目标通常缺乏计算机辅助设计（CAD）模型和关键点先验，使得现有的基于模型或关键点匹配的方法难以可靠应用。为了解决这些挑战，本文提出了MF-UAVPose6D，一种用于固定翼无人机的无模型单目6-DoF姿态估计框架。在推理过程中，该方法仅以单张红绿蓝（RGB）图像和相机内参作为输入。它首先通过热图引导的中心定位获得稳定的目标锚点，引入了一个视角感知模块（Perspective-Aware Module, PAM）来建模观测光线的先验，利用动态拓扑采样（Dynamic Topological Sampling, DTS）来补充来自机翼、机身和尾部的弱结构线索，并采用解耦的平移-旋转姿态解码机制来估计6-DoF姿态。此外，我们构建了FW-UAV6DPose合成数据集，涵盖了不同距离、视角和姿态下的固定翼无人机观测。实验结果表明，MF-UAVPose6D在不需要CAD模型的情况下实现了准确且高效的单目6-DoF姿态估计，并在远程旋转估计、深度恢复和联合姿态评估方面表现出强大的鲁棒性。
cs.CV / 144 / 2606.29699
Early Warning Signals for OpenVLA Failure under Visual Distribution Shift
视觉分布变化下OpenVLA故障的早期预警信号
Abstract
Vision Language Action models combine perception, language grounding, and control in a single policy, but their failures are hard to diagnose once visual conditions shift. We test whether OpenVLA feedforward activations contain linearly decodable information about near term task failure in LIBERO manipulation rollouts. The policy is fixed throughout. We log internal activations during execution and fit lightweight monitors after the rollouts are collected. Occlusion is the main controlled stress test. It reduces OpenVLA success from $57\%$ to $17\%$ over $100$ episodes per condition. Under this shift, a logistic probe at layer 16 reaches AUROC $0.972$ and AUPRC $0.352$ for predicting failure within a $15$ step horizon. It outperforms both a mean difference direction and an action disagreement baseline. A sparse layer sweep finds uneven decodability across depth: layer 16 is strongest among the tested layers, layer 8 remains informative, and layer 10 is weaker. To check whether the monitor is just an occlusion detector, we also evaluate color shift and camera jitter without refitting. Color shift produces no failures in this setting, so it is a benign control rather than a failure benchmark. Camera jitter does induce failures, and the occlusion trained monitor remains above random. The result is deliberately limited: OpenVLA internal states contain failure relevant structure under controlled perceptual shift, but these experiments do not establish a causal mechanism, task held out generalization, or a deployable recovery system.
Chinese Translation
视觉语言行动模型将感知、语言基础和控制结合在一个单一策略中，但一旦视觉条件发生变化，其故障很难诊断。我们测试OpenVLA前馈激活是否包含关于LIBERO操作回合中近期任务失败的线性可解信息。整个过程中策略保持不变。我们在执行期间记录内部激活，并在收集完回合后拟合轻量级监控器。遮挡是主要的控制压力测试。它使OpenVLA的成功率从$57\%$降低到$17\\%$，在每种条件下进行$100$个回合。在这种变化下，第16层的逻辑探测器在$15$步范围内预测失败时达到AUROC $0.972$和AUPRC $0.352$。它优于均值差异方向和行动不一致基线。稀疏层扫描发现深度间的解码能力不均：第16层在测试层中最强，第8层保持信息性，而第10层较弱。为了检查监控器是否仅仅是一个遮挡检测器，我们还在不重新拟合的情况下评估颜色变化和相机抖动。颜色变化在这种情况下未产生失败，因此它是一个良性的控制，而非失败基准。相机抖动确实会导致失败，而经过遮挡训练的监控器仍然高于随机水平。结果故意有限：OpenVLA内部状态在受控的感知变化下包含与失败相关的结构，但这些实验并未建立因果机制、任务保留泛化或可部署的恢复系统。
cs.CV / 145 / 2606.29714
UniVAD v2: Unified Visual Anomaly Detection via Support-Conditioned Boundary Construction
UniVAD v2：通过支持条件边界构建实现统一视觉异常检测
Abstract
Unified visual anomaly detection seeks to train a single detector that can be deployed across categories, domains, and application scenarios. In the few-shot transfer regime, the key challenge is to estimate an episode-specific boundary for an unseen target category from a small support set. Existing approaches mainly infer this boundary from normal-side evidence and provide limited abnormal-side evidence for deployment-specific tolerance. Within the normal side, they often struggle to jointly capture local correspondences and global support-query relations, making their boundaries less reliable for unseen anomalies. To address these issues, we propose UniVAD v2, a two-sided support-conditioned boundary construction framework for unified visual anomaly detection. Built on the component-patch divide-and-conquer framework of UniVAD, UniVAD v2 strengthens the normal side with an Optimal Transport-based Relational Modeling module (OTRM), which complements retrieval with support-query matching through transport-style allocation, and an Adaptive Coordination mechanism for Retrieval and Relational Modeling (ACRRM), which estimates episode-conditioned reliabilities to fuse the two sources of evidence. On the abnormal side, a Few-Shot Abnormal Reference module (FAR) converts optional abnormal references into rejection-side evidence for boundary adjustment. Experiments on six datasets spanning industrial, logical, and medical anomaly detection demonstrate strong cross-domain generalization. Under the 1N-shot protocol, UniVAD v2 improves the mean image-level AUC over UniVAD from 83.0\% to 84.5\%, and further reaches 85.7\% in the 1N+1A-shot setting. On the MVTec-AD Severity Split (MVTec-AD-SS), UniVAD v2 achieves 96.2\% image-level AUC and 96.9\% pixel-level AUC, showing that abnormal references enable controllable boundary customization without retraining.
Chinese Translation
统一视觉异常检测旨在训练一个可以在不同类别、领域和应用场景中部署的单一检测器。在少量样本迁移的情况下，关键挑战是从小的支持集估计未见目标类别的特定边界。现有方法主要从正常侧证据推断该边界，并为特定部署提供有限的异常侧证据。在正常侧，它们通常难以共同捕捉局部对应关系和全局支持-查询关系，使得它们对未见异常的边界可靠性降低。为了解决这些问题，我们提出了UniVAD v2，一个用于统一视觉异常检测的双侧支持条件边界构建框架。基于UniVAD的组件-补丁分治框架，UniVAD v2通过一个基于最优传输的关系建模模块（Optimal Transport-based Relational Modeling，OTRM）增强正常侧，该模块通过传输风格分配补充支持-查询匹配，并通过自适应协调机制（Adaptive Coordination mechanism for Retrieval and Relational Modeling，ACRRM）估计情节条件的可靠性，以融合这两种证据。在异常侧，一个少量样本异常参考模块（Few-Shot Abnormal Reference，FAR）将可选的异常参考转换为拒绝侧证据，以调整边界。在涵盖工业、逻辑和医疗异常检测的六个数据集上的实验表明了强大的跨领域泛化能力。在1N-shot协议下，UniVAD v2将图像级平均AUC从83.0\%提高到84.5\%，并在1N+1A-shot设置中进一步达到85.7\%。在MVTec-AD严重性分割（MVTec-AD-SS）上，UniVAD v2实现了96.2\%的图像级AUC和96.9\%的像素级AUC，表明异常参考使得可控的边界定制成为可能，而无需重新训练。
cs.CV / 146 / 2606.29715
Accurate Recognition of Pneumonia and COVID-19 by Geometric Shape Normalization of Lung Region using Automatic Landmark Detection and Piecewise Affine Warping
通过自动标志检测和分段仿射变换对肺部区域进行几何形状归一化以准确识别肺炎和COVID-19
Abstract
This paper presents an automatic system for recognizing pulmonary diseases in chest X-rays using geometric normalization of the lung region. The method combines three modules: (1) a ResNet-18 landmark detector with coordinate attention that predicts 15 lung-contour landmarks, achieving a mean localization error of 3.61 pixels through an ensemble of four models with test-time augmentation; (2) a geometric normalizer based on Generalized Procrustes Analysis, Delaunay triangulation, and piecewise affine warping to map each lung region to a standardized shape; and (3) a ResNet-18 classifier with transfer learning and SAHS contrast enhancement to classify images as COVID-19, Viral Pneumonia, or Normal. On the COVID-19 Radiography Database, the normalized-image classifier achieved 98.60+/-0.26% accuracy and 98.00% F1-Macro using five-fold cross-validation. Although original images produced slightly higher raw accuracy, Grad-CAM and cropping experiments suggest that this advantage is partly influenced by acquisition artifacts. In contrast, geometrically normalized images outperformed artifact-masked/cropped unaligned images on both the COVID-19 Radiography Database (98.60% vs. 96.24%) and a balanced adult-pediatric mixed dataset including pediatric cases from the Kermany dataset (94.67% vs. 94.17%). These results suggest that anatomical alignment can provide a more controlled and artifact-resistant representation for pulmonary disease recognition.
Chinese Translation
本文提出了一种自动系统，用于通过对胸部X光片中肺部区域的几何归一化来识别肺部疾病。该方法结合了三个模块：(1) 一个具有坐标注意力的ResNet-18标志检测器，预测15个肺轮廓标志，通过四个模型的集成和测试时增强，达到了3.61像素的平均定位误差；(2) 基于广义Procrustes分析、德劳内三角剖分和分段仿射变换的几何归一化器，将每个肺部区域映射到标准化形状；(3) 一个使用迁移学习和SAHS对比度增强的ResNet-18分类器，将图像分类为COVID-19、病毒性肺炎或正常。在COVID-19放射影像数据库上，归一化图像分类器达到了98.60±0.26%的准确率和98.00%的F1-Macro，采用五折交叉验证。尽管原始图像产生了略高的原始准确率，但Grad-CAM和裁剪实验表明，这一优势部分受到采集伪影的影响。相比之下，几何归一化图像在COVID-19放射影像数据库（98.60%对96.24%）和包括Kermany数据集中儿童病例的平衡成人-儿童混合数据集（94.67%对94.17%）上均优于伪影遮罩/裁剪的未对齐图像。这些结果表明，解剖对齐可以为肺部疾病识别提供更受控且抗伪影的表现。
cs.CV / 147 / 2606.29716
AerialMetric: Benchmarking and Adapting UAV Monocular Metric Depth Estimation in the Real World
AerialMetric：在现实世界中基准测试和适应无人机单目度量深度估计
Abstract
This paper addresses the problem of monocular metric depth estimation in aerial UAV imagery. Although recent data-driven methods have achieved remarkable progress in ground-level scenarios, models trained primarily on street-view and indoor datasets exhibit significant domain gaps when applied to aerial viewpoints. To tackle these challenges, we introduce AerialMetric, a benchmark dataset designed to evaluate and facilitate the adaptation of monocular metric depth estimation under UAV aerial viewpoints. The dataset consists of four complementary subsets collected from different sources, jointly covering real-world photogrammetry data, controlled aerial acquisition settings, photorealistic synthetic scenes, and in-the-wild Internet imagery. Totally, AerialMetric provides 52K real-world and 16K synthetic image-depth pairs with reliable metric ground truth. Based on this dataset, we conduct systematic evaluations of existing state-of-the-art models under aerial settings and investigate the impact of viewpoint, altitude, and camera parameters on metric depth prediction. In addition, by fine-tuning representative metric depth model on our dataset, we establish a comprehensive aerial benchmark and achieve state-of-the-art performance across diverse aerial imagery. Our dataset, code, and model weight are publicly available at https://kuieless.github.io/AerialMetric-ECCV2026-page/.
Chinese Translation
本文探讨了无人机航拍图像中单目度量深度估计的问题。尽管近期数据驱动的方法在地面场景中取得了显著进展，但主要在街景和室内数据集上训练的模型在应用于航拍视角时表现出显著的领域差距。为了解决这些挑战，我们引入了AerialMetric，一个旨在评估和促进无人机航拍视角下单目度量深度估计适应性的基准数据集。该数据集由四个互补子集组成，收集自不同来源，涵盖了真实世界的摄影测量数据、受控的航拍获取设置、逼真的合成场景以及野外互联网图像。总的来说，AerialMetric提供了52K个真实世界和16K个合成图像-深度对，并附有可靠的度量真值。基于该数据集，我们对现有的最先进模型在航拍设置下进行了系统评估，并研究了视角、高度和相机参数对度量深度预测的影响。此外，通过在我们的数据集上微调代表性的度量深度模型，我们建立了一个全面的航拍基准，并在多样的航拍图像中实现了最先进的性能。我们的数据集、代码和模型权重已公开发布，网址为 https://kuieless.github.io/AerialMetric-ECCV2026-page/。
cs.CV / 148 / 2606.29744
HTC-SGA Former: A Hybrid Transformer-CNN Network with Self-Guided Attention and a New Boundary-Weighted Adaptive Loss for Coronary DSA Vessel Segmentation
HTC-SGA Former：一种具有自我引导注意力和新边界加权自适应损失的混合Transformer-CNN网络用于冠状动脉DSA血管分割
Abstract
Accurate coronary Digital Subtraction Angiography (DSA) vessel segmentation is essential for computer-aided diagnosis and treatment planning of coronary artery disease (CAD). However, thin low-contrast vessels, background interference, and severe vessel-background class imbalance make reliable segmentation of weak distal branches and vessel boundaries challenging. Existing methods struggle to balance global contextual reasoning with preservation of weak vessels, vessel continuity, and fine boundaries. To address these limitations, we propose HTC-SGA Former, a lightweight hybrid Transformer-CNN framework for coronary DSA vessel segmentation. It employs a CNN encoder for local vessel morphology extraction and a Transformer decoder for contextual feature modeling. A Multi-Scale Global-Local Window Attention (MS-GLWA) block performs efficient global-local contextual modeling, while a Self-Guided Feature Attention (SGFA) module enhances weak-vessel responses. In addition, a Boundary-Weighted Adaptive Compound Loss (BWACL) emphasizes thin-vessel boundaries and adaptively balances vessel recovery and boundary refinement. Experiments on private right and left coronary artery DSA subsets show that HTC-SGA Former outperforms 14 state-of-the-art segmentation methods while maintaining a compact architecture with only 0.81M parameters. BWACL also improves performance over binary cross-entropy and Dice losses across four encoder-decoder architectures, demonstrating strong cross-backbone applicability. HTC-SGA Former improves thin-vessel recovery, vessel continuity, and boundary localization through complementary global-local contextual modeling, vessel-focused refinement, and adaptive optimization, supporting reliable and computationally efficient coronary vessel analysis for future computer-assisted cardiovascular interventions.
Chinese Translation
准确的冠状动脉数字减影血管造影（DSA）血管分割对于冠状动脉疾病（CAD）的计算机辅助诊断和治疗规划至关重要。然而，细小低对比度的血管、背景干扰以及严重的血管与背景类别不平衡使得弱远端分支和血管边界的可靠分割面临挑战。现有方法在全球上下文推理与保持弱血管、血管连续性和精细边界之间的平衡方面存在困难。为了解决这些局限性，我们提出了HTC-SGA Former，一种轻量级混合Transformer-CNN框架用于冠状动脉DSA血管分割。该框架采用CNN编码器进行局部血管形态提取，并使用Transformer解码器进行上下文特征建模。多尺度全局-局部窗口注意力（MS-GLWA）模块执行高效的全局-局部上下文建模，而自我引导特征注意力（SGFA）模块增强弱血管响应。此外，边界加权自适应复合损失（BWACL）强调细小血管边界，并自适应地平衡血管恢复与边界细化。在私有的右冠状动脉和左冠状动脉DSA子集上的实验表明，HTC-SGA Former在保持仅0.81M参数的紧凑架构的同时，超越了14种最先进的分割方法。BWACL在四种编码器-解码器架构中也提升了性能，优于二元交叉熵和Dice损失，展示了强大的跨骨干适用性。HTC-SGA Former通过互补的全局-局部上下文建模、以血管为中心的细化和自适应优化，改善了细小血管的恢复、血管连续性和边界定位，为未来计算机辅助心血管干预提供了可靠且计算高效的冠状动脉血管分析支持。
cs.CV / 149 / 2606.29752
LEIQ-Assessor: Multi-dimensional Quality Assessment of Low-light Enhanced Images via Multi-task Learning
LEIQ-Assessor：基于多任务学习的低光增强图像多维质量评估
Abstract
Low-light image enhancement algorithms (LIEAs) aim to improve the visibility of images captured under poor illumination. However, the enhancement process often introduces artifacts such as noise amplification, color shift, structural damage, and over-exposure, which degrade the perceptual quality of the enhanced images. Therefore, a reliable image quality assessment (IQA) metric for evaluating enhancement effects is of great importance for both the development of LIEAs and their practical applications. In this paper, we present \textbf{LEIQ-Assessor}, a multi-dimensional quality assessment model for low-light image enhancement based on multi-task learning, developed for the QoMEX 2026 Grand Challenge on Low-light Enhanced Image Quality Assessment. Specifically, our method leverages a pre-trained SigLIP2 Vision Transformer as the backbone and simultaneously predicts the overall Mean Opinion Score (MOS) together with six perceptual sub-attributes: lightness, color fidelity, noise level, exposure quality, naturalness, and content recovery. By jointly optimizing these correlated objectives via the PLCC loss, the shared representation captures richer quality-aware features than its single-task counterpart. Experiments on the MLE benchmark demonstrate that LEIQ-Assessor significantly outperforms existing no-reference IQA models and hand-crafted quality descriptors. Our method achieved second place in the QoMEX 2026 Grand Challenge on Low-light Enhanced Image Quality Assessment. The code is available at https://github.com/sunwei925/LEIQ-Assessor.
Chinese Translation
低光图像增强算法（LIEAs）旨在改善在光照不足条件下捕获的图像的可见性。然而，增强过程往往会引入噪声放大、色彩偏移、结构损坏和过度曝光等伪影，从而降低增强图像的感知质量。因此，可靠的图像质量评估（IQA）指标对于评估增强效果在LIEAs的发展及其实际应用中具有重要意义。本文提出了 extbf{LEIQ-Assessor}，一种基于多任务学习的低光图像增强多维质量评估模型，旨在为2026年QoMEX低光增强图像质量评估大赛提供支持。具体而言，我们的方法利用预训练的SigLIP2视觉变换器作为主干，同时预测整体的平均意见分数（MOS）以及六个感知子属性：亮度、色彩保真度、噪声水平、曝光质量、自然性和内容恢复。通过PLCC损失共同优化这些相关目标，所共享的表示捕获了比单任务模型更丰富的质量感知特征。在MLE基准上的实验表明，LEIQ-Assessor显著优于现有的无参考IQA模型和手工设计的质量描述符。我们的方法在2026年QoMEX低光增强图像质量评估大赛中获得第二名。代码可在https://github.com/sunwei925/LEIQ-Assessor获取。
cs.CV / 150 / 2606.29760
MR-IQA: A Unified Margin View of Regression and Ranking for Blind Image Quality Assessment
MR-IQA：盲图像质量评估的回归与排序统一边际视角
Abstract
Blind image quality assessment (BIQA) is commonly built on two basic learning paradigms: regression and ranking. Regression calibrates absolute scores, whereas ranking recovers quality structure from ordinal relations. Although joint regression-ranking supervision often improves BIQA, the relation between the two paradigms remains largely empirical and underexplored. In this work, we revisit what underlies regression and ranking and identify pairwise relational distance, termed quality margin, as their common bridge. Our derivation shows that, at the objective-optimization level, both paradigms fit quality margins: regression fits margins induced by score endpoints, while ranking fits transformed or sign-level margins through preference probabilities. Motivated by this insight, we propose MR-IQA, a direct quality-margin optimization framework for reinforcement learning (RL)-based BIQA. MR-IQA samples quality scores and optimizes pairwise margin errors as policy rewards, thereby modeling quality structure more explicitly. Experiments on six BIQA benchmarks show competitive general performance, and controlled comparisons demonstrate that MR-IQA achieves the strongest average PLCC/SRCC over regression- or ranking-based RL methods. Our findings provide a new insight into unifying regression and ranking, offering a theoretical basis for understanding quality-structure modeling in BIQA and beyond.
Chinese Translation
盲图像质量评估（BIQA）通常基于两种基本学习范式：回归和排序。回归校准绝对分数，而排序则从序关系中恢复质量结构。尽管联合回归-排序监督通常能改善BIQA，但这两种范式之间的关系仍然在很大程度上是经验性的，尚未深入探讨。在本研究中，我们重新审视回归和排序的基础，并将成对关系距离（称为质量边际）确定为它们的共同桥梁。我们的推导表明，在目标优化层面，这两种范式都适应质量边际：回归适应由分数端点引起的边际，而排序则通过偏好概率适应变换或符号级边际。基于这一洞察，我们提出了MR-IQA，一种针对基于强化学习（RL）的BIQA的直接质量边际优化框架。MR-IQA采样质量分数并优化成对边际误差作为策略奖励，从而更明确地建模质量结构。在六个BIQA基准上的实验显示出竞争力的整体表现，受控比较表明MR-IQA在回归或排序基础的RL方法中实现了最强的平均PLCC/SRCC。我们的发现为统一回归和排序提供了新的见解，为理解BIQA及其他领域的质量结构建模提供了理论基础。
cs.CV / 151 / 2606.29763
TopoAgent: An Agentic Framework for Automated Topology Learning in Medical Imaging
TopoAgent：用于医学影像自动拓扑学习的智能框架
Abstract
Topological data analysis (TDA), particularly persistent homology (PH), captures geometric structural properties in medical images (e.g., connected components, loops, shape characteristics), which conventional pixel-level deep learning approaches often neglect. While many topological descriptors are known for converting persistence diagrams (PDs) or raw images into topological feature vectors, existing methods mostly default to a single fixed descriptor (e.g., persistence images), leaving the diversity of topological representations largely unexplored. To the best of our knowledge, there is no known large language model (LLM)-based agentic framework that can automatically determine the most suitable topological descriptors for a given image dataset and produce the corresponding topological feature vectors for downstream tasks. To fill this gap, we propose \textbf{TopoAgent}, an LLM-based agentic framework that automates topology learning for medical image analysis.TopoAgent operates through a Perception--Reasoning--Action--Reflection loop supported by 21 domain-specific tools and dual memory that accumulates experience across runs. Its skill set is distilled from systematic evaluation of 15 topological descriptors across 26 datasets with six classifiers. TopoAgent analyzes input images and their topological characteristics, reasons about which topological descriptors best suit the input, and determines the optimal descriptor and its configuration, all without task-specific training.
Chinese Translation
拓扑数据分析（TDA），特别是持久同调（PH），能够捕捉医学图像中的几何结构特性（例如，连通分量、环、形状特征），而传统的像素级深度学习方法往往忽视这些特性。虽然已有许多拓扑描述符用于将持久性图（PDs）或原始图像转换为拓扑特征向量，但现有方法大多依赖单一固定描述符（例如，持久性图像），导致拓扑表示的多样性尚未得到充分探索。根据我们所知，目前尚无基于大型语言模型（LLM）的智能框架能够自动确定给定图像数据集最合适的拓扑描述符，并为下游任务生成相应的拓扑特征向量。为填补这一空白，我们提出了 extbf{TopoAgent}，一个基于LLM的智能框架，旨在自动化医学图像分析中的拓扑学习。TopoAgent通过一个感知-推理-行动-反思循环运作，支持21个特定领域的工具和双重记忆，以积累运行经验。其技能集源于对26个数据集上15个拓扑描述符的系统评估，使用六种分类器。TopoAgent分析输入图像及其拓扑特性，推理出最适合输入的拓扑描述符，并确定最佳描述符及其配置，所有这些均无需特定任务的训练。
cs.CV / 152 / 2606.29781
UrbanCDNet: Appearance-Robust and Boundary-Aware Bitemporal Change Detection for Korean Urban Building Monitoring
UrbanCDNet：针对韩国城市建筑监测的外观鲁棒性和边界感知的双时相变化检测
Abstract
Urban building change detection from bi-temporal aerial imagery is important for redevelopment monitoring, infrastructure management, and unauthorized-construction screening, but Korean urban scenes remain difficult because changed regions are often sparse, appearance varies strongly between acquisition dates, and useful outputs must follow building footprints rather than coarse blobs. This paper presents UrbanCDNet, a task specific Siamese CNN that combines appearance-robust multi-cue comparison, alignment-aware middle-scale differencing, lightweight context refinement, scene calibration, and auxiliary boundary supervision. Experiments use a corrected AIHub-based Korean benchmark with 3,998 training, 503 validation, and 499 test pairs, and report changed-class precision, recall, F1, and IoU. On the locked test split, UrbanCDNet achieves 0.7335 precision, 0.7696 recall, 0.7511 F1, and 0.6014 IoU, outperforming a strong Siamese U-Net baseline (0.7108 F1, 0.5514 IoU) and the strongest external competitor, ChangeFormer-MIT-B0 (0.7107 F1, 0.5512 IoU). Additional diagnostic slicing shows that the gain is concentrated in the operating regimes that motivated the design: on the sparse-change subset with less than 5% changed area, F1 improves from 0.4765 to 0.6175, and on the high photometric-gap subset it improves from 0.6349 to 0.7285. Boundary F1 at 3-pixel tolerance rises from 0.3445 to 0.4447, while object F1 at IoU 0.3 rises from 0.0690 to 0.2258. These results indicate that, on this Korean benchmark, task-shaped temporal comparison and boundary-aware supervision matter more than generic model scale alone
Chinese Translation
从双时相航空影像中进行城市建筑变化检测对于再开发监测、基础设施管理和未经授权的建筑筛查至关重要，但韩国城市场景仍然面临挑战，因为变化区域通常稀疏，外观在采集日期之间变化较大，并且有用的输出必须遵循建筑轮廓而不是粗糙的斑块。本文提出了UrbanCDNet，一种特定任务的孪生卷积神经网络（Siamese CNN），它结合了外观鲁棒的多线索比较、对齐感知的中尺度差异、轻量级上下文细化、场景校准和辅助边界监督。实验使用了基于AIHub的修正韩国基准数据集，包含3,998个训练对、503个验证对和499个测试对，并报告了变化类别的精确度、召回率、F1值和交并比（IoU）。在锁定的测试分割中，UrbanCDNet达到了0.7335的精确度、0.7696的召回率、0.7511的F1值和0.6014的IoU，超越了强大的孪生U-Net基线（0.7108 F1，0.5514 IoU）和最强的外部竞争者ChangeFormer-MIT-B0（0.7107 F1，0.5512 IoU）。额外的诊断切片显示，增益集中在激励设计的操作范围内：在变化面积小于5%的稀疏变化子集上，F1值从0.4765提高到0.6175，而在高光度差异子集上则从0.6349提高到0.7285。3像素容差下的边界F1值从0.3445上升到0.4447，而在IoU为0.3时的对象F1值从0.0690上升到0.2258。这些结果表明，在该韩国基准上，任务导向的时间比较和边界感知监督比单纯的模型规模更为重要。
cs.CV / 153 / 2606.29782
Graph-GSReg: Leveraging 3D Scene Graphs for Gaussian Splatting Registration
Graph-GSReg：利用三维场景图进行高斯点云配准
Abstract
Merging multiple 3D Gaussian Splatting (3DGS) scenes into a single unified Gaussian representation is essential for large-scale 3D mapping and long-term map management. Despite its importance, this area remains underexplored, and existing solutions exhibit several limitations. Learning-based methods attempt direct correspondence between Gaussian primitives and require training on large 3DGS datasets. Image-based optimization methods depend heavily on coarse initialization from generic foundation models and often incur expensive refinement. We present \ourmodel. Our method constructs a 3D scene graph from a 3DGS and its rendered images, \textit{reformulating 3DGS registration as a graph registration problem}. The proposed 3D scene graph represents each 3DGS at a higher-level representation, enabling a globally consistent understanding of semantic information and structural context for accurate registration. To further construct a seamless unified scene, we introduce a Self-Supervised Test-Time Optimization. Naively merging two 3D Gaussian scenes often suffers from occlusion artifacts such as hollows and floaters. To alleviate this issue, we refine the merged Gaussians to preserve visual consistency between the original scenes and the merged scene. We evaluate our method on real and synthetic benchmarks, demonstrating competitive registration accuracy and merged scene rendering quality.
Chinese Translation
将多个三维高斯点云（3D Gaussian Splatting，3DGS）场景合并为单一统一的高斯表示，对于大规模三维映射和长期地图管理至关重要。尽管其重要性不言而喻，但这一领域仍然未被充分探索，现有解决方案存在多种局限性。基于学习的方法试图直接建立高斯原语之间的对应关系，并需要在大型3DGS数据集上进行训练。基于图像的优化方法则严重依赖于来自通用基础模型的粗略初始化，且通常需要昂贵的精细化过程。我们提出了 extit{our model}。我们的方法从3DGS及其渲染图像构建三维场景图，将3DGS配准重新表述为图注册问题。所提出的三维场景图以更高层次的表示方式表示每个3DGS，使得对语义信息和结构上下文的全球一致理解成为可能，从而实现准确的配准。为了进一步构建无缝统一的场景，我们引入了自监督测试时优化。简单地合并两个三维高斯场景往往会遭遇遮挡伪影，如空洞和漂浮物。为了解决这一问题，我们对合并的高斯进行精细化，以保持原始场景与合并场景之间的视觉一致性。我们在真实和合成基准上评估了我们的方法，展示了具有竞争力的配准精度和合并场景的渲染质量。
cs.CV / 154 / 2606.29786
OP3DSG: Open-Vocabulary Part-Aware 3D Scene Graph Generation for Real-World Environments
OP3DSG：面向真实环境的开放词汇部件感知3D场景图生成
Abstract
3D scene graphs (3DSGs) provide a compact and structured abstraction of 3D environments. Although advances in foundation models have enabled open-vocabulary 3DSG generation, existing approaches remain object-centric and encode limited relational information -- restricting their applicability in real-world scenarios that require fine-grained understanding. We propose OP3DSG, an open-vocabulary part-aware 3DSG generation framework that constructs unified graphs that jointly model objects, interactive parts, spatial relations, functional relations, and affordances. OP3DSG integrates object-part knowledge-guided detection with part-aware 3D fusion to preserve small and interaction-relevant components, and employs a geometry-initialized prior graph with LLM-based refinement to reduce spurious relational predictions while enabling efficient graph construction. To systematically evaluate unified 3D scene graph construction, we introduce UniGraph3D, a benchmark designed for part-aware perception and multi-level relational reasoning. Experimental results show that OP3DSG achieves state-of-the-art performance and demonstrates its effectiveness as a perception backbone in diverse real-world robotics tasks.
Chinese Translation
3D场景图（3DSGs）提供了对3D环境的紧凑且结构化的抽象。尽管基础模型的进展使得开放词汇3DSG生成成为可能，但现有方法仍然以对象为中心，编码的关系信息有限，限制了其在需要细粒度理解的真实场景中的适用性。我们提出了OP3DSG，一个开放词汇部件感知的3DSG生成框架，构建统一图形，联合建模对象、交互部件、空间关系、功能关系和可用性。OP3DSG将对象-部件知识引导的检测与部件感知的3D融合相结合，以保留小型和与交互相关的组件，并采用几何初始化的先验图与基于LLM的细化相结合，以减少虚假关系预测，同时实现高效的图形构建。为了系统评估统一的3D场景图构建，我们引入了UniGraph3D，这是一个为部件感知感知和多层次关系推理设计的基准。实验结果表明，OP3DSG实现了最先进的性能，并展示了其作为多样化真实世界机器人任务中的感知骨干的有效性。
cs.CV / 155 / 2606.29794
UniTriSplat: A Unified 3D Gaussian Splatting Framework with Uniform Spherical Rasterization for Universal Cameras
UniTriSplat：一种针对通用相机的统一3D高斯点云框架，采用均匀球面光栅化
Abstract
Existing 3D Gaussian Splatting (3DGS) frameworks rely on camera-specific rasterization, suffering from inconsistent solid-angle sampling and degraded performance across heterogeneous camera models (e.g., perspective, fisheye, omnidirectional). To address this limitation, we propose UniTriSplat, a unified 3DGS framework for universal cameras that reformulates Gaussian splatting on the unit sphere via HEALPix discretization. Leveraging the equal-area property of HEALPix, we construct a spherical sampling grid aligned with the angular resolution of input images. We derive the forward rendering and gradient propagation of Gaussians directly in the spherical radian domain, yielding uniform optimization behavior from narrow-FoV images to full 360-degree panoramas. To enhance perceptual reconstruction quality, we additionally introduce a HEALPix-aware SSIM loss that respects spherical neighborhood structure. Extensive experiments across diverse camera models demonstrate that UniTriSplat consistently improves cross-camera generalization while preserving geometric fidelity and rendering quality.
Chinese Translation
现有的3D高斯点云（3DGS）框架依赖于特定相机的光栅化，导致在异构相机模型（例如透视、鱼眼、全向）之间存在不一致的固体角采样和性能下降。为了解决这一限制，我们提出了UniTriSplat，一种针对通用相机的统一3DGS框架，通过HEALPix离散化重新构造单位球上的高斯点云。利用HEALPix的等面积特性，我们构建了一个与输入图像的角分辨率对齐的球面采样网格。我们直接在球面弧度域中推导高斯的前向渲染和梯度传播，从而实现从窄视场图像到完整360度全景图的均匀优化行为。为了增强感知重建质量，我们还引入了一种考虑HEALPix的结构相似性（SSIM）损失，尊重球面邻域结构。广泛的实验表明，UniTriSplat在不同相机模型之间一致地提高了跨相机的泛化能力，同时保持了几何保真度和渲染质量。
cs.CV / 156 / 2606.29801
Concept Removal Guidance: Evidence-Calibrated Negative Guidance for Safe Diffusion Sampling
概念移除引导：基于证据校准的负引导用于安全扩散采样
Abstract
Text-to-image diffusion models remain vulnerable to adversarial prompts that elicit disallowed content, motivating reliable inference-time controls. A popular approach is negative guidance, which subtracts a negative prompt direction with a fixed weight. However, it often forces a safety-fidelity trade-off, causing artifacts or prompt drift when over-applied and failing under attacks when under-applied. Dynamic variants reweight guidance using posterior-odds signals, which can be brittle for open-vocabulary compositional prompts, while lightweight similarity-based methods ignore the evolving image evidence along the denoising trajectory. We introduce Concept Removal Guidance (CRG), a training-free method that estimates unwanted-concept presence at each diffusion step from the model's noise predictions, and adaptively calibrates negative guidance via a closed-form constrained update enforcing a target presence threshold while minimally perturbing the conditional trajectory. Across red-teaming benchmarks, CRG reduces attack success rates while preserving benign fidelity, and extends to additional suppression targets such as artist style and violence without fine-tuning or external classifiers.
Chinese Translation
文本到图像的扩散模型仍然容易受到引发不当内容的对抗性提示的影响，这促使了可靠的推理时控制方法的需求。一种流行的方法是负引导，它以固定权重减去负提示方向。然而，这种方法往往迫使安全性与保真度之间进行权衡，当过度应用时会导致伪影或提示漂移，而在不足应用时则在攻击下失效。动态变体使用后验赔率信号重新加权引导，但对于开放词汇的组合提示可能不够稳健，而轻量级的基于相似性的算法则忽略了在去噪轨迹中不断演变的图像证据。我们提出了概念移除引导（Concept Removal Guidance, CRG），这是一种无训练的方法，它通过模型的噪声预测在每个扩散步骤中估计不需要的概念存在，并通过封闭形式的约束更新自适应地校准负引导，强制执行目标存在阈值，同时最小限度地扰动条件轨迹。在红队基准测试中，CRG降低了攻击成功率，同时保持良好的保真度，并扩展到其他抑制目标，如艺术风格和暴力，而无需微调或外部分类器。
cs.CV / 157 / 2606.29805
Clearer Sight, Fewer Lies: Oriented Pickup Preference Optimization for Multimodal Hallucination Mitigation
更清晰的视野，更少的谎言：面向多模态幻觉缓解的定向选择偏好优化
Abstract
Multimodal Large Language Models (MLLMs) are prone to hallucination as their generation preferences are insufficiently calibrated to visual evidence, causing them to fall back on linguistic priors, rather than faithful grounding. In this work, we start from an empirical observation: when query-relevant visual evidence is explicitly strengthened using the model's own attention, generation becomes more accurate, suggesting that many failures do not arise solely from missing perception, but from an insufficient tendency to trust the evidence the model has already attended to. Motivated by this finding, we propose Oriented Pickup Preference Optimization (\texttt{OPPO}), an evidence-aware alignment objective that learns preferences over the strength of visual evidence, rather than only response quality. Concretely, \texttt{OPPO} contrasts the same faithful response under stronger, anchored, weaker-evidence views, turning naive visual preference into ordered visual-evidence alignment. We further combine this objective with fine-grained span-level and token-level regularization to stabilize the training. Besides, we provide a theoretical analysis showing that ordered evidence margins induce a positive lower bound on local visual sensitivity. Extensive evaluations across hallucination and general-purpose benchmarks demonstrate that \texttt{OPPO} consistently outperforms baseline methods.
Chinese Translation
多模态大型语言模型（MLLMs）容易出现幻觉，因为其生成偏好与视觉证据的校准不足，导致它们依赖语言先验，而非忠实的基础。在本研究中，我们从一个经验观察出发：当查询相关的视觉证据通过模型自身的注意力显著增强时，生成的准确性提高，这表明许多失败并非仅仅源于感知缺失，而是由于模型对已关注证据的信任倾向不足。基于这一发现，我们提出了定向选择偏好优化（Oriented Pickup Preference Optimization， exttt{OPPO}），这是一种基于证据的对齐目标，旨在学习视觉证据强度的偏好，而不仅仅是响应质量。具体而言， exttt{OPPO}在更强、锚定和较弱证据视角下对比相同的忠实响应，将简单的视觉偏好转化为有序的视觉证据对齐。此外，我们将这一目标与细粒度的跨度级和标记级正则化相结合，以稳定训练。此外，我们提供了理论分析，表明有序证据边际会在局部视觉敏感性上引入正的下界。对幻觉和通用基准的广泛评估表明， exttt{OPPO}始终优于基线方法。
cs.CV / 158 / 2606.29812
Consistency as Inductive Bias: Learning Cross-View Invariance for Robust Multimodal Reasoning
一致性作为归纳偏置：学习跨视图不变性以实现稳健的多模态推理
Abstract
Inductive biases steer learning toward generalizable solutions by encoding task structure. In this work, we identify a crucial missing bias in MLLMs: cross-view consistency, \textit{i.e.}, semantically invariant views of the same instance should lead to the same answer. Standard reinforcement learning with verifiable rewards (RLVR) objectives do not impose this constraint, but instead assign pointwise rewards to each visual input. Even with data augmentation (DA), transformed views are typically rewarded independently, providing little signal once within-view rewards saturate. We propose \textbf{ConsistRoll}, a simple but effective method that injects cross-view consistency into RLVR training by reusing the group-sampling mechanism of GRPO. Specifically, ConsistRoll places original and semantically invariant transformed views in the same generation group, and assigns a joint reward only when paired completions are both correct and consistent. In this way, ConsistRoll turns consistency into an online credit-assignment signal, \textbf{without extra generation overhead and annotations}. Theoretically, we show that cross-view consistency is a valid inductive bias, and ConsistRoll introduces a cross-view correction term absent from DA, penalizing view dependence and alleviating advantage collapse. Comprehensive benchmarks across math, general-purpose, hallucination domains confirm that ConsistRoll achieves robust improvements in multimodal reasoning.
Chinese Translation
归纳偏置通过编码任务结构，引导学习朝向可推广的解决方案。在本研究中，我们识别出MLLMs（多模态大语言模型）中一个关键的缺失偏置：跨视图一致性，即同一实例的语义不变视图应导致相同的答案。标准的可验证奖励强化学习（RLVR）目标并未施加这一约束，而是对每个视觉输入分配逐点奖励。即使在数据增强（DA）情况下，变换后的视图通常也是独立获得奖励，一旦视图内奖励饱和，提供的信号也很有限。我们提出了 extbf{ConsistRoll}，一种简单但有效的方法，通过重用GRPO的组采样机制，将跨视图一致性注入到RLVR训练中。具体而言，ConsistRoll将原始视图和语义不变的变换视图放置在同一生成组中，仅在配对的完成结果都正确且一致时分配联合奖励。通过这种方式，ConsistRoll将一致性转化为一种在线信用分配信号， extbf{无需额外的生成开销和注释}。理论上，我们证明了跨视图一致性是一个有效的归纳偏置，并且ConsistRoll引入了DA中缺失的跨视图修正项，惩罚视图依赖并缓解优势崩溃。在数学、通用、幻觉领域的全面基准测试中，ConsistRoll在多模态推理中实现了稳健的改进。
cs.CV / 159 / 2606.29814
Nemotron-Labs-Diffusion-Image: Advancing Masked Discrete Diffusion for High-Resolution Image Synthesis
Nemotron-Labs-Diffusion-Image：推进高分辨率图像合成的掩蔽离散扩散模型
Abstract
We propose Nemotron-Labs-Diffusion-Image, a state-of-the-art masked discrete diffusion model (MDM) for high-resolution text-to-image synthesis. Compared with prior work on masked image generation, Nemotron-Labs-Diffusion-Image addresses two key challenges. First, unlike continuous diffusion models which progressively refine latent representations across the entire image, standard MDMs lack self-correcting capability because discrete tokens cannot be modified once they are unmasked. Second, although increasing the vocabulary size of discrete image tokenizers improves reconstruction fidelity, it introduces optimization difficulties for generative modeling as the per-token training signal becomes increasingly sparse. To address the first challenge, Nemotron-Labs-Diffusion-Image incorporates a token-editing mechanism that enables the model to dynamically revise already-unmasked tokens during inference, similar to how a sculptor iteratively refines their work. To tackle the second challenge, we propose a Grouped Cross-Entropy (GCE) objective that assigns positive learning signals to tokens neighboring the ground truth in embedding space, thereby alleviating signal sparsity. To further improve training efficiency, we implement a custom fused operator for GCE that significantly reduces VRAM usage in large-vocabulary settings. Experimental results demonstrate that these innovations substantially improve both training efficiency and image fidelity of masked discrete image generators, achieving a score of 0.90 on GenEval, 86.9 on DPG and 10.76 of HPSv3.
Chinese Translation
我们提出了Nemotron-Labs-Diffusion-Image，这是一种用于高分辨率文本到图像合成的最先进的掩蔽离散扩散模型（MDM）。与之前的掩蔽图像生成工作相比，Nemotron-Labs-Diffusion-Image解决了两个关键挑战。首先，与逐步精炼整个图像潜在表示的连续扩散模型不同，标准的MDM缺乏自我纠正能力，因为一旦被解掩的离散标记无法被修改。其次，尽管增加离散图像标记器的词汇量可以提高重建保真度，但这会给生成建模带来优化困难，因为每个标记的训练信号变得越来越稀疏。为了解决第一个挑战，Nemotron-Labs-Diffusion-Image引入了一种标记编辑机制，使模型能够在推理过程中动态修订已经解掩的标记，类似于雕塑家迭代精炼其作品的方式。为了解决第二个挑战，我们提出了一种分组交叉熵（Grouped Cross-Entropy, GCE）目标，该目标为嵌入空间中邻近真实值的标记分配正向学习信号，从而缓解信号稀疏性。为了进一步提高训练效率，我们实现了一种定制的GCE融合运算符，在大词汇设置中显著减少了显存使用。实验结果表明，这些创新显著提高了掩蔽离散图像生成器的训练效率和图像保真度，在GenEval上获得了0.90的分数，在DPG上获得了86.9，在HPSv3上获得了10.76。
cs.CV / 160 / 2606.29821
Learning Cross-view Correspondences for Geo-localization on Planetary Surfaces
学习跨视角对应关系以实现行星表面的地理定位
Abstract
Maintaining global position awareness is a fundamental challenge for planetary surface exploration, since satellite-based positioning systems are unavailable and onboard odometry drifts over time. Although orbital mapping products, such as overhead imagery and terrain-derived maps, provide global context, aligning them with surface observations is challenging due to large viewpoint differences, low texture, repetitive terrain, and drastic changes in appearance caused by varying illumination and topography. We introduce a new cross-view geo-localization benchmark built from physically rendered surface panoramas and overhead tiles derived from a high-resolution lunar terrain model. Our dataset contains 10438 ground views rendered as 360$^\circ$ surface panoramas with matching overhead images precisely centered at the same location. Additionally, a set of overlapping tiles is provided to study off-center localization with multiple plausible candidates per panorama. We study the performance of a state-of-the-art transformer-based geo-localization method on our data, by training it from scratch and reporting retrieval accuracy. Our results demonstrate that learning-based cross-view localization methods can be successfully applied to the domain of planetary surfaces, providing a vision-based alternative to global navigation satellite systems.
Chinese Translation
维持全球位置意识是行星表面探索中的一项基本挑战，因为基于卫星的定位系统不可用，而机载里程计会随时间漂移。尽管轨道映射产品，如俯视影像和地形衍生地图，提供了全球背景，但由于视角差异大、纹理低、地形重复以及由于光照和地形变化引起的外观剧烈变化，将其与表面观测对齐是具有挑战性的。我们引入了一种新的跨视角地理定位基准，该基准由物理渲染的表面全景图和来自高分辨率月球地形模型的俯视图块构建。我们的数据集包含10438个地面视图，这些视图被渲染为360$^ ext{°}$的表面全景图，并且与精确位于同一位置的俯视图相匹配。此外，还提供了一组重叠的图块，以研究离中心定位的问题，每个全景图都有多个合理的候选项。我们研究了一种基于最先进的变换器（transformer）方法在我们的数据集上的表现，通过从零开始训练并报告检索准确率。我们的结果表明，基于学习的跨视角定位方法可以成功应用于行星表面的领域，为全球导航卫星系统提供了一种基于视觉的替代方案。
cs.CV / 161 / 2606.29828
HomeDiffusion: Zero-Shot Object Customization with Multi-View Representation Learning for Indoor Scenes
HomeDiffusion：基于多视角表征学习的室内场景零-shot物体定制
Abstract
Recently, zero-shot object customization generation methods have rapidly developed and shown tremendous potential for applications. For instance, in the e-commerce domain, consumers can observe the visual effect of furniture placed within their personal living spaces or clothes worn on their own bodies. Many existing approaches perform object customization generation based on diffusion models and extracted reference object features. However, the generated object significantly diverges from the original reference object in details such as patterns and curves. Particularly for asymmetrical reference objects, the absence of comprehensive multi-viewpoint information prevents the generation of object poses that harmonize with the background scene. To address these shortcomings, we have constructed a novel dataset comprising multi-angle images of furniture and indoor scenes. Based on diffusion models, we introduce HomeDiffusion, which can leverage multi-viewpoint images of the same reference object to accurately generate visually harmonious object poses within specified areas of the background scene. During the diffusion process, we further extract high-fidelity details of the reference object and perform cross-attention with the noise latents in the latent space, thereby ensuring the preservation of details in the customized object generation. Extensive qualitative and quantitative experiments demonstrate that our method achieves superior performance over other existing zero-shot as well as few-shot object customization approaches.
Chinese Translation
近年来，零-shot物体定制生成方法迅速发展，并展现出巨大的应用潜力。例如，在电子商务领域，消费者可以观察到家具放置在个人生活空间中的视觉效果，或衣物穿在自己身上的效果。许多现有的方法基于扩散模型和提取的参考物体特征进行物体定制生成。然而，生成的物体在细节上，如图案和曲线，往往与原始参考物体存在显著差异。特别是对于不对称的参考物体，缺乏全面的多视角信息阻碍了与背景场景和谐的物体姿态生成。为了解决这些不足，我们构建了一个新颖的数据集，其中包含家具和室内场景的多角度图像。基于扩散模型，我们提出了HomeDiffusion，它能够利用同一参考物体的多视角图像，在指定的背景场景区域内准确生成视觉上和谐的物体姿态。在扩散过程中，我们进一步提取参考物体的高保真细节，并在潜在空间中与噪声潜变量进行交叉注意，从而确保在定制物体生成中保留细节。大量的定性和定量实验表明，我们的方法在零-shot和少-shot物体定制方法中表现优越。
cs.CV / 162 / 2606.29837
Robust Trajectory Distillation: Hybrid Reweighting Meets Teacher-Inspired Targets
鲁棒轨迹蒸馏：混合重加权与教师启发目标的结合
Abstract
Dataset distillation (DD) condenses large corpora into compact, information-rich subsets for efficient training and reuse. However, under noisy supervision, DD risks condensing corrupted associations together with useful signals, degrading robustness. Conventional noisy-label remedies (sample selection, loss weighting, label correction) tightly couple noise estimation with model optimization, often require clean anchors, and can amplify confirmation bias-assumptions that are misaligned with DD's goal of compact, plug-and-play supervision. We therefore propose a trajectory-based DD framework that jointly suppresses noise and preserves transferable knowledge without relabeling or clean subsets. It comprises two complementary components: Selective Guidance Reweighting (SGR), which fuses global forgetting patterns (second-split forgetting) with local neighborhood consistency into a progressive reweighting scheme that prioritizes clean supervision along the teacher trajectory; and Teacher-Inspired Auxiliary Targets (TIAT), which inject auxiliary residual guidance distilled from intermediate teacher dynamics to reinforce informative signals while remaining internally consistent. Together, SGR and TIAT produce distilled datasets with cleaner and richer representations under noisy supervision. The framework is robust, label-preserving, computationally lightweight, and broadly applicable, yielding consistent gains over state-of-the-art DD baselines across symmetric, asymmetric, and real-world noise.
Chinese Translation
数据集蒸馏（Dataset Distillation, DD）将大型语料库浓缩为紧凑的信息丰富子集，以实现高效的训练和重用。然而，在噪声监督下，DD存在将损坏的关联与有用信号一起浓缩的风险，从而降低了鲁棒性。传统的噪声标签处理方法（样本选择、损失加权、标签修正）将噪声估计与模型优化紧密耦合，通常需要干净的锚点，并可能放大与DD目标不一致的确认偏见假设，即紧凑的即插即用监督。因此，我们提出了一种基于轨迹的DD框架，该框架共同抑制噪声并保留可转移知识，而无需重新标记或干净子集。该框架由两个互补组件组成：选择性引导重加权（Selective Guidance Reweighting, SGR），它将全局遗忘模式（第二次遗忘）与局部邻域一致性融合到一个渐进重加权方案中，优先考虑沿教师轨迹的干净监督；以及教师启发辅助目标（Teacher-Inspired Auxiliary Targets, TIAT），它注入从中间教师动态中提取的辅助残差引导，以增强信息信号，同时保持内部一致性。SGR和TIAT共同生成在噪声监督下具有更清晰和更丰富表示的蒸馏数据集。该框架具有鲁棒性、标签保留、计算轻量和广泛适用性，在对称、非对称和现实世界噪声的情况下，相较于最先进的DD基线，始终能取得一致的提升。
cs.CV / 163 / 2606.29845
Bricker to BRACE: A Bracket Exposure RAW Dataset and Restoration Model for Flicker-Banding
从 Bricker 到 BRACE：用于闪烁条纹的括号曝光 RAW 数据集和修复模型
Abstract
Flicker-banding (FB), arises from temporal aliasing between a camera's rolling shutter and a display's brightness modulation, degrading screen-captured image readability with color shifts and jagged patterns. Existing single-frame methods with simplified parametric stripe models cannot reliably distinguish these artifacts from genuine texture. To address this, we conduct a systematic analysis of complex FB morphologies and reveal their significant variation across exposure settings, motivating a multi-frame bracketed RAW restoration paradigm. We construct Bricker, a synthetic-real bracketed RAW dataset built via ray-tracing-based physical simulation and automated multi-exposure capture tool. We further propose BRACE: Bracketed RAW Flicker-Banding Removal, a multi-frame restoration model that utilizes frequency-aware banding prior and a multi-scale spatial cross-attention modulator (MSCAM) for cross-exposure spatial fusion. We also introduce the Stripe Frequency Consistency (SFC) metric to evaluate banding removal. Experiments demonstrate state-of-the-art performance on both synthetic and real benchmarks. Our dataset and code are available at: https://github.com/ZZH-qwq/BRACE.
Chinese Translation
闪烁条纹（Flicker-banding, FB）是由于相机的滚动快门与显示器的亮度调制之间的时间混叠而产生的，这会导致屏幕捕获的图像可读性下降，出现颜色偏移和锯齿状图案。现有的单帧方法采用简化的参数化条纹模型，无法可靠地区分这些伪影与真实纹理。为了解决这一问题，我们对复杂的 FB 形态进行了系统分析，并揭示了其在不同曝光设置下的显著变化，从而推动了多帧括号 RAW 修复范式的发展。我们构建了 Bricker，这是一个通过基于光线追踪的物理仿真和自动多曝光捕获工具构建的合成-真实括号 RAW 数据集。我们进一步提出了 BRACE：括号 RAW 闪烁条纹去除，这是一个多帧修复模型，利用频率感知的条纹先验和多尺度空间交叉注意力调制器（Multi-Scale Spatial Cross-Attention Modulator, MSCAM）进行跨曝光空间融合。我们还引入了条纹频率一致性（Stripe Frequency Consistency, SFC）指标来评估条纹去除效果。实验表明，在合成和真实基准测试中均表现出最先进的性能。我们的数据集和代码可在以下地址获取：https://github.com/ZZH-qwq/BRACE。
cs.CV / 164 / 2606.29847
See Only When Needed: Context-Aware Attention Intervention for Mitigating Hallucinations in LVLMs
仅在需要时查看：用于减轻大型视觉语言模型中幻觉的上下文感知注意力干预
Abstract
Large Vision-Language Models (LVLMs) excel at multimodal tasks but remain prone to object hallucinations. Prior training-free remedies often uniformly strengthen visual signals, which may also amplify irrelevant regions and introduce spurious evidence, harming fluency. We propose Context-aware Attention Intervention (CAI), a training-free inference-time mechanism that enforces a see only when needed principle via two-axis selectivity: where to look and when to intervene. At each decoding step, CAI derives token-specific visual relevance from early-layer representations to localize semantically aligned regions, and applies a conservative, entropy- and depth-gated attention tilt only for uncertainty-spiking tokens in deeper layers where visual grounding degrades, leaving confident tokens and irrelevant regions largely unchanged. This targeted intervention strengthens visual grounding while preserving linguistic fluency, and it yields consistent improvements even without contrastive decoding, which remains optional as an auxiliary bias-suppression module. Extensive experiments across multiple LVLM backbones and benchmarks show that CAI achieves state-of-the-art hallucination mitigation, and our analysis characterizes CAI as a KL-minimal attention reweighting with bounded interference under inactive gates or small tilts. Code is available at https://github.com/Iris1946/CAI.
Chinese Translation
大型视觉语言模型（LVLMs）在多模态任务中表现出色，但仍然容易出现物体幻觉。以往的无训练补救措施通常均匀地增强视觉信号，这可能也会放大无关区域并引入虚假证据，从而损害流畅性。我们提出了上下文感知注意力干预（Context-aware Attention Intervention, CAI），这是一种无训练的推理时机制，通过两个轴向的选择性来强制执行仅在需要时查看的原则：何处查看和何时干预。在每个解码步骤中，CAI 从早期层表示中推导出特定于标记的视觉相关性，以定位语义对齐区域，并仅对在视觉基础减弱的深层中不确定性激增的标记应用保守的、基于熵和深度的注意力倾斜，保持自信标记和无关区域基本不变。这种针对性的干预在增强视觉基础的同时保持语言流畅性，即使在没有对比解码的情况下也能带来一致的改进，后者仍然可以作为辅助偏差抑制模块。针对多个 LVLM 骨干网络和基准的广泛实验表明，CAI 实现了最先进的幻觉减轻效果，我们的分析将 CAI 特征化为在非活动门或小倾斜下具有有界干扰的 KL 最小注意力重加权。代码可在 https://github.com/Iris1946/CAI 获取。
cs.CV / 165 / 2606.29850
Efficient Visual Pointing for Embodied AI:Agent-Driven Data Synthesis, Cross-Block Attention, and Iterative Correction
高效的具身人工智能视觉指向：基于代理的数据合成、跨块注意力和迭代修正
Abstract
Visual pointing maps a language instruction to pixel co ordinates, a core skill for embodied AI. We describe our PointArena 2026 solution, which achieves 77.2% overall accuracy and ranks second on the benchmark. The ap proach targets three failure modes. First, agent-driven syn thesis builds large semantic and anchor-relative candidate pools; the server inventory contains 55,372 processed out puts, 53,772 de-duplicated sample IDs, and 37,574 train able completed or accepted rows. Second, a determinis tic steerable-data pipeline creates a verified 10,000-sample main set, plus reserve samples, using masks, templates, and path verification. Third, two model-side modules address complementary errors: AttnRes adds gated cross-block at tention for steerability, while ABC correction encodes per turbed coordinates with visual features for general coordi nate grounding. Category-aware routing combines comple mentary specialists; local validation used to select experts records 93.9% Affordance, 82.6% Spatial Relation, 78.2% Reasoning, 70.4% Counting, and 63.0% Steerability.
Chinese Translation
视觉指向将语言指令映射到像素坐标，这是具身人工智能的核心技能。我们描述了我们的 PointArena 2026 解决方案，该方案实现了 77.2% 的整体准确率，并在基准测试中排名第二。该方法针对三种失败模式。首先，基于代理的合成构建了大型语义和锚点相对候选池；服务器库存包含 55,372 个处理输出、53,772 个去重样本 ID 和 37,574 行可训练的已完成或接受的记录。其次，确定性可引导数据管道使用掩码、模板和路径验证创建了经过验证的 10,000 个样本主集，以及备用样本。第三，两个模型侧模块解决互补错误：AttnRes 为可引导性添加了门控跨块注意力，而 ABC 修正则使用视觉特征对扰动坐标进行编码，以实现一般坐标的基础。类别感知路由结合了互补专家；用于选择专家的本地验证记录了 93.9% 的可供性、82.6% 的空间关系、78.2% 的推理、70.4% 的计数和 63.0% 的可引导性。
cs.CV / 166 / 2606.29855
RainODE: Continuous-Time Precipitation Forecasting with Latent Neural ODEs
RainODE：基于潜在神经常微分方程的连续时间降水预测
Abstract
In precipitation forecasting, not only accuracy but also temporal resolution is critical. However, increasing temporal resolution is constrained by observational limitations and the computational cost of dense discrete modeling. To overcome this limitation, we reformulate precipitation forecasting as a continuous-time dynamical system and propose RainODE, a framework that models precipitation evolution in latent space using a Neural ODE. This formulation enables derivative-consistent temporal dynamics and captures the dominant large-scale advective motion of precipitation systems. Nevertheless, a purely deterministic ODE struggles to represent non-advective intensity changes such as localized growth, decay, and sub-grid variability, often leading to over-smoothed predictions. To address this issue, we introduce a stochastic source modeling module based on a Brownian Bridge formulation, which refines residual intensity variations and restores fine-grained structures while preserving advective consistency. By combining deterministic continuous dynamics with stochastic refinement, RainODE enables arbitrary-time inference while maintaining sharp predictions. Experiments on SEVIR and the newly introduced Radar-based Precipitation Integrated Dataset (RAPID) demonstrate consistent improvements across multiple temporal intervals and precipitation regimes. The code is available at https://github.com/SeongYE/RainODE.
Chinese Translation
在降水预测中，准确性和时间分辨率同样至关重要。然而，提高时间分辨率受到观测限制和密集离散建模的计算成本的制约。为了解决这一限制，我们将降水预测重新表述为一个连续时间动力系统，并提出了RainODE框架，该框架使用神经常微分方程（Neural ODE）在潜在空间中建模降水演变。这种表述使得导数一致的时间动态成为可能，并捕捉到降水系统的主导大尺度平流运动。然而，纯粹的确定性常微分方程难以表示非平流强度变化，如局部增长、衰减和子网格变异，常常导致预测过于平滑。为了解决这个问题，我们引入了基于布朗桥（Brownian Bridge）表述的随机源建模模块，该模块精细化残余强度变化并恢复细粒度结构，同时保持平流一致性。通过将确定性连续动态与随机细化相结合，RainODE能够实现任意时间推断，同时保持清晰的预测。在SEVIR和新推出的基于雷达的降水综合数据集（RAPID）上的实验表明，在多个时间间隔和降水模式下均取得了一致的改进。代码可在https://github.com/SeongYE/RainODE获取。
cs.CV / 167 / 2606.29861
SUMO: Segment and Track Any Motion with Nonlinear State Space Models
SUMO：使用非线性状态空间模型对任何运动进行分割和跟踪
Abstract
Visual Object Tracking (VOT) and Moving Object Segmentation (MOS) are two fundamental tasks in computer vision that involve both spatial and temporal object dynamics. Existing methods rely predominantly on visual cues and thus often falter in real-world scenarios where object motions are inherently complex and nonlinear. To address this limitation, we propose SUMO, a zero-shot, training-free, unified framework integrating nonlinear dynamics with vision-based segmentation for accurate and consistent VOT and MOS. Specifically, we develop a nonlinear State Space Model (SSM) inspired by robotics principles to capture the complex object dynamics. Building on this model, we propose a Selective Unscented Filter (SUF) for accurate state estimation, which features a joint scoring mechanism and dynamically fuses multi-source predictions to identify the most plausible object state over time. Furthermore, we apply a memory selection mechanism to evaluate the reliability of memory frames. Our extensive experimental results show that SUMO achieves state-of-the-art performance on both VOT and MOS tasks.
Chinese Translation
视觉目标跟踪（VOT）和移动物体分割（MOS）是计算机视觉中的两个基本任务，涉及空间和时间上的物体动态。现有方法主要依赖视觉线索，因此在物体运动本质上复杂且非线性的现实场景中往往表现不佳。为了解决这一局限性，我们提出了SUMO，一个零-shot、无训练、统一的框架，将非线性动态与基于视觉的分割相结合，以实现准确且一致的VOT和MOS。具体而言，我们开发了一种受机器人原理启发的非线性状态空间模型（SSM），以捕捉复杂的物体动态。在此模型基础上，我们提出了一种选择性无迹滤波器（SUF）用于准确的状态估计，该滤波器具有联合评分机制，并动态融合多源预测，以识别随时间变化的最可信物体状态。此外，我们应用了一种记忆选择机制来评估记忆帧的可靠性。我们广泛的实验结果表明，SUMO在VOT和MOS任务上均实现了最先进的性能。
cs.CV / 168 / 2606.29879
LWDrive: Layer-Wise World-Model-Guided Vision-Language Model Planning for Autonomous Driving
LWDrive：层级世界模型引导的视觉-语言模型规划用于自主驾驶
Abstract
Vision-Language Models (VLMs) provide powerful semantic understanding and commonsense reasoning for End-to-End Autonomous Driving (E2E-AD) planning. However, trajectories directly generated by VLMs often encode only coarse driving intentions and remain insufficient for geometrically accurate, future-aware, and multi-view-grounded planning. To address these limitations, we develop the Layer-Wise World-Model-Guided Driving framework (LWDrive). LWDrive is a VLM planning framework that refines coarse trajectories through layer-wise world-model guidance. Instead of treating the VLM output as the final trajectory, LWDrive uses it as an intent-aware coarse plan, expands a diverse candidate space around it, and progressively refines the candidates through a Foresight Cascade Planner (FCP). Specifically, we introduce future-frame generation supervision to encourage the VLM to learn forward-looking scene representations, thereby injecting planning-relevant predictive dynamics into its internal hidden states. Built upon these world-model-supervised representations, FCP exploits VLM features across multiple layers and integrates historical temporal states, Action-Query representations, and current-frame multi-view Bird's-Eye-View (BEV) features to refine candidate trajectories in a coarse-to-fine manner. This design enables progressive correction of spatial positions and motion trends while grounding trajectory refinement with multi-view scene cues and preserving the high-level driving intention produced by the large model. Finally, a score head evaluates the refined candidates and selects the best trajectory as the final planning output. Experiments show that LWDrive achieves a score of 92.0 on the NAVSIM benchmark and 89.6 on NAVSIM-v2. Code and models will be made publicly available.
Chinese Translation
视觉-语言模型（VLMs）为端到端自主驾驶（E2E-AD）规划提供了强大的语义理解和常识推理能力。然而，VLMs直接生成的轨迹通常仅编码粗略的驾驶意图，无法满足几何精确、未来感知和多视角基础的规划需求。为了解决这些局限性，我们开发了层级世界模型引导的驾驶框架（LWDrive）。LWDrive是一个VLM规划框架，通过层级世界模型引导来细化粗略轨迹。LWDrive并不将VLM的输出视为最终轨迹，而是将其作为一种意图感知的粗略计划，在其周围扩展多样的候选空间，并通过前瞻级联规划器（Foresight Cascade Planner, FCP）逐步细化候选轨迹。具体而言，我们引入了未来帧生成监督，以鼓励VLM学习前瞻性的场景表示，从而将与规划相关的预测动态注入其内部隐藏状态。在这些世界模型监督的表示基础上，FCP利用VLM在多个层次的特征，并整合历史时间状态、动作-查询表示以及当前帧的多视角鸟瞰图（Bird's-Eye-View, BEV）特征，以粗到细的方式细化候选轨迹。该设计使得在多视角场景线索的基础上，能够逐步修正空间位置和运动趋势，同时保留由大型模型生成的高层次驾驶意图。最后，一个评分头评估细化后的候选轨迹，并选择最佳轨迹作为最终规划输出。实验表明，LWDrive在NAVSIM基准测试中取得了92.0的得分，在NAVSIM-v2中取得了89.6的得分。代码和模型将公开发布。
cs.CV / 169 / 2606.29880
IREU: Identity-Related Encoder-Only Unlearning for Customized Portrait Generation
IREU：用于定制肖像生成的身份相关编码器仅去学习
Abstract
Customized Portrait Generation (CPG) technologies have been widely used to generate high-fidelity person images given an input image indicating the identity and a text prompt indicating the required edits. Yet these methods pose significant privacy risks by spreading fake visual information. Against such risks, each public generator should be able to suppress its generation ability for a particular person when requested. Therefore, in this work we investigate the identity unlearning problem for CPG. Since there are no previous methods in this field, we propose a simple baseline that updates the image encoder by minimizing identity similarity between generated and input images for target identities to be unlearned, while maximizing it for identities to be retained. However, we find such a global perturbation in the feature space harms the fidelity of generated images for other identities to be retained. To solve this problem, we propose a novel method IREU, which first locates identity-related features in an offline manner and then only performs feature perturbations on them. The experimental results show that our proposed method IREU achieves better identity unlearning performance for target identities to be unlearned, and also keeps high fidelity for other identities to be retained. In addition, our unlearned image encoder is generalizable across different generators with the same encoder without fine-tuning, which is friendly for deployment in practice.
Chinese Translation
定制肖像生成（CPG）技术已广泛应用于根据输入图像（指示身份）和文本提示（指示所需编辑）生成高保真度的人物图像。然而，这些方法在传播虚假视觉信息方面存在显著的隐私风险。针对这些风险，每个公共生成器在请求时应能够抑制其对特定人物的生成能力。因此，在本研究中，我们探讨了CPG的身份去学习问题。由于该领域没有先前的方法，我们提出了一种简单的基线，通过最小化目标身份的生成图像与输入图像之间的身份相似性来更新图像编码器，同时最大化需保留身份的相似性。然而，我们发现这种在特征空间中的全局扰动会损害其他需保留身份生成图像的保真度。为了解决这个问题，我们提出了一种新方法IREU，该方法首先以离线方式定位与身份相关的特征，然后仅对这些特征进行扰动。实验结果表明，我们提出的方法IREU在需去学习的目标身份上实现了更好的身份去学习性能，同时对其他需保留身份保持了高保真度。此外，我们的去学习图像编码器在不同生成器之间具有良好的通用性，无需微调即可使用相同编码器，这在实际部署中非常友好。
cs.CV / 170 / 2606.29883
Building artificial intelligence virtual tissue (AIVT) for tissue state representation, feature prediction, and dynamic simulation
构建人工智能虚拟组织 (AIVT) 以实现组织状态表示、特征预测和动态模拟
Abstract
Modeling tissue states and their transitions is essential for understanding tissue homeostasis in health and pathological remodeling in disease. However, conventional computational modeling approaches are inadequate to capture the complexity of tissues as spatially organized, multiscale biological systems. Artificial intelligence (AI) has shown a remarkable ability for representing intricate systems, creating new opportunities to characterize tissue states and their transitions. Here, we propose the concept of AI virtual tissue (AIVT), an AI framework grounded in spatial multimodal data for modeling tissues in health and disease. AIVT is designed to learn unified, spatially resolved, and dynamically manipulatable representations of tissue state, enabling tissue state representation and analysis, molecular and morphological feature prediction, and simulation of spatiotemporal tissue dynamics. We outline the fundamental assumptions, core capabilities, architectural components, as well as data and algorithm foundations of AIVT as a framework for AI-driven tissue modeling.
Chinese Translation
建模组织状态及其转变对于理解健康中的组织稳态和疾病中的病理重塑至关重要。然而，传统的计算建模方法不足以捕捉作为空间组织的多尺度生物系统的复杂性。人工智能 (AI) 在表示复杂系统方面展现了显著的能力，为表征组织状态及其转变创造了新的机会。在此，我们提出了人工智能虚拟组织 (AIVT) 的概念，这是一个基于空间多模态数据的AI框架，用于建模健康和疾病中的组织。AIVT旨在学习统一的、空间分辨的和动态可操控的组织状态表示，从而实现组织状态的表示与分析、分子和形态特征的预测，以及时空组织动态的模拟。我们概述了AIVT作为一个AI驱动的组织建模框架的基本假设、核心能力、架构组件，以及数据和算法基础。
cs.CV / 171 / 2606.29900
LLM-based Multimodal Personality Recognition via Facial Action Unit-Text Semantic Fusion
基于大语言模型的多模态个性识别：面部动作单元与文本语义融合
Abstract
Personality recognition in asynchronous video interviews (AVIs) has become increasingly important due to their widespread adoption in modern recruitment. Existing approaches often rely on large language models (LLMs) to analyze textual responses of interviewees in AVI. However, unimodel methods often suffer from information loss (e.g., ignore facial cues). In contrast, multimodal methods that employ full-face images or sparsely sampled frames can discard fine-grained temporal dynamics critical for accurate personality assessment. To overcome these limitations, we propose an LLM-based framework that semantically fuse facial action units (AUs) with textual responses of AVI. AU sequences are first converted into interpretable textual descriptions, which are then fused with participants' textual responses through an LLM. A lightweight regression head transforms the resulting embeddings into continuous personality scores without disrupting the underlying semantic space. Experiments on the AVI-6 benchmark demonstrate consistent improvements over most baselines, with lower prediction errors and stronger correlations with human-rated scores across multiple traits. Further analysis reveals that AU-derived semantic representations offer complementary non-verbal cues to textual responses. Decoupling semantic understanding from regression prediction within the LLM also leads to greater training stability and clearer interpretability. Overall, these findings demonstrate that AU-text fusion provides a psychologically grounded and computationally efficient framework for personality recognition in AVIs.
Chinese Translation
在异步视频面试（AVI）中进行个性识别变得越来越重要，因为其在现代招聘中的广泛应用。现有方法通常依赖于大型语言模型（LLMs）来分析面试者在AVI中的文本回应。然而，单一模型方法往往会遭遇信息丢失（例如，忽略面部线索）。相比之下，采用全脸图像或稀疏采样帧的多模态方法可能会丢弃对准确个性评估至关重要的细粒度时间动态。为克服这些局限性，我们提出了一种基于LLM的框架，通过语义融合面部动作单元（AUs）与AVI的文本回应。首先，将AU序列转换为可解释的文本描述，然后通过LLM与参与者的文本回应进行融合。一个轻量级回归头将生成的嵌入转换为连续的个性评分，而不干扰基础语义空间。在AVI-6基准上的实验表明，与大多数基线相比，预测误差更低，且与人类评分在多个特质上的相关性更强。进一步分析表明，基于AU的语义表示为文本回应提供了互补的非语言线索。在LLM中将语义理解与回归预测解耦也带来了更大的训练稳定性和更清晰的可解释性。总体而言，这些发现表明，AU-文本融合为AVI中的个性识别提供了心理学基础和计算效率高的框架。
cs.CV / 172 / 2606.29905
StrucTab: A Structured Optimization Framework for Table Parsing
StrucTab：一种结构化的表格解析优化框架
Abstract
Table parsing aims to convert table images into structured, machine-readable representations, a task requiring the joint perception of complex spatial layouts and textual content. While recent vision-language models (VLMs) enable end-to-end parsing, they typically rely on direct supervision of the final output, thereby bypassing the explicit intermediate reasoning that is crucial for understanding complex table structures. Furthermore, attempts to optimize these models using reinforcement learning (RL) are often hindered by unstable or ambiguous reward designs, limiting potential performance gains. To address these limitations, we propose StrucTab, a table parsing model learned through intermediate structural supervision and reward decomposition. At the modeling level, by decomposing the parsing process into human-inspired subtasks, such as row-column counting and merged-cell analysis, StrucTab progressively unifies them through a sequential reasoning strategy. At the optimization level, we introduce Uni-TabRL, a unified RL framework that leverages decomposed rewards (validity, structure, and content) to provide stable and informative optimization signals. Finally, at the evaluation level, we present TableVerse-5K, a large-scale, challenging benchmark encompassing diverse, real-world table scenarios. Extensive experiments demonstrate the state-of-the-art performance of StrucTab across all evaluated public benchmarks and significant improvements on TableVerse-5K, validating the effectiveness of explicit structural modeling and decomposed reward optimization. Code and benchmark are publicly available at https://github.com/VirtualLUOUCAS/StrucTab.
Chinese Translation
表格解析旨在将表格图像转换为结构化的、机器可读的表示，这一任务需要对复杂空间布局和文本内容的联合感知。尽管最近的视觉语言模型（VLMs）实现了端到端的解析，但它们通常依赖于对最终输出的直接监督，从而绕过了理解复杂表格结构所需的明确中间推理。此外，使用强化学习（RL）优化这些模型的尝试往往受到不稳定或模糊奖励设计的阻碍，限制了潜在的性能提升。为了解决这些局限性，我们提出了StrucTab，一种通过中间结构监督和奖励分解学习的表格解析模型。在建模层面，通过将解析过程分解为人类启发的子任务，如行列计数和合并单元分析，StrucTab通过顺序推理策略逐步统一这些子任务。在优化层面，我们引入了Uni-TabRL，一个统一的RL框架，利用分解的奖励（有效性、结构和内容）提供稳定且信息丰富的优化信号。最后，在评估层面，我们提出了TableVerse-5K，一个涵盖多样化真实世界表格场景的大规模挑战性基准。大量实验表明，StrucTab在所有评估的公共基准上均表现出最先进的性能，并在TableVerse-5K上显著提升，验证了明确结构建模和分解奖励优化的有效性。代码和基准可在 https://github.com/VirtualLUOUCAS/StrucTab 上公开获取。
cs.CV / 173 / 2606.29909
Traffic-CBM: A Structurally Interpretable Multimodal Framework for Encrypted Traffic Classification
Traffic-CBM：一种结构可解释的多模态加密流量分类框架
Abstract
Encrypted traffic classification has achieved strong performance, but its decision process remains difficult to interpret. Existing methods usually combine flow statistics, packet sequences, and byte-level representations into opaque latent features, making it unclear which type of evidence actually drives the prediction. In this paper, we propose Traffic-CBM, a structurally interpretable multimodal framework for encrypted traffic classification. Instead of directly fusing heterogeneous traffic signals into a black-box representation, Traffic-CBM organizes them into a unified hierarchical concept space. These concepts are not manually annotated semantic attributes; rather, they are scalar evidence summaries constrained by predefined traffic evidence groups. More specifically, grouped flow statistics are mapped to statistical concepts, dedicated temporal encoders learn temporal concepts from disjoint feature subspaces, and byte-level evidence is further organized into packet-level and cross-packet concepts. This design turns heterogeneous traffic evidence into an explicit concept representation and makes different levels of traffic evidence easier to analyze. We evaluate Traffic-CBM on multiple encrypted traffic benchmarks. Results show that it achieves competitive and balanced classification performance while providing a clearer structural interpretation interface than conventional end-to-end fusion models. Further analyses suggest that the learned concept space is actively used in the prediction process and provides a clearer structural explanation of multimodal traffic evidence.
Chinese Translation
加密流量分类已取得良好性能，但其决策过程仍然难以解释。现有方法通常将流量统计、数据包序列和字节级表示结合成不透明的潜在特征，使得不清楚到底是哪种证据驱动了预测。在本文中，我们提出了Traffic-CBM，一种结构可解释的多模态加密流量分类框架。Traffic-CBM并不是直接将异构流量信号融合成黑箱表示，而是将它们组织成统一的层次概念空间。这些概念并不是手动标注的语义属性，而是受预定义流量证据组约束的标量证据摘要。更具体地说，分组的流量统计被映射到统计概念，专用的时间编码器从不相交的特征子空间中学习时间概念，而字节级证据则进一步组织成数据包级和跨数据包概念。这一设计将异构流量证据转化为显式的概念表示，并使不同层次的流量证据更易于分析。我们在多个加密流量基准上评估了Traffic-CBM。结果表明，它在分类性能上具有竞争力和均衡性，同时提供比传统的端到端融合模型更清晰的结构解释界面。进一步分析表明，学习到的概念空间在预测过程中被积极使用，并提供了对多模态流量证据更清晰的结构解释。
cs.CV / 174 / 2606.29915
H-GRPO: Permutation-Invariant Reinforcement Learning for Grounded Visual Reasoning
H-GRPO：用于基础视觉推理的置换不变强化学习
Abstract
Vision-Language Models (VLMs) often achieve high performance on benchmarks while remaining "black boxes", yet they remain prone to hallucination or rely on superficial shortcuts. In this work, we propose a framework designed to enhance both performance and interpretability through De-compositional Evidence Grounding. Unlike monolithic inference approaches, our approach forces the model to decompose a global query into a sequence of atomic sub-questions, each requiring an explicit sub-answer and critically a localized evidence bounding box. By grounding intermediate logical steps (e.g. identifying a container, analyzing liquid properties, and assessing environmental context) in specific visual regions, we construct a structured reasoning path that mirrors human-like deduction. This allows the final answer to emerge as a logical consequence of verified visual facts rather than a statistical guess.
Chinese Translation
视觉-语言模型（VLMs）在基准测试中通常表现出色，但仍然是“黑箱”，且容易出现幻觉或依赖表面捷径。在本研究中，我们提出了一种框架，旨在通过分解证据基础增强性能和可解释性。与整体推理方法不同，我们的方法迫使模型将全局查询分解为一系列原子子问题，每个子问题都需要一个明确的子答案，并且关键是一个局部证据边界框。通过在特定视觉区域中扎根中间逻辑步骤（例如识别容器、分析液体特性和评估环境上下文），我们构建了一条结构化的推理路径，反映了类人推理。这使得最终答案能够作为经过验证的视觉事实的逻辑结果，而不是统计猜测。
cs.CV / 175 / 2606.29924
DCGrasp: Distance-aware Controllable Grasp Generation
DCGrasp：基于距离感知的可控抓取生成
Abstract
Generating 3D hand-object interactions is essential for applications in robotics, XR, and synthetic data generation, where flexible controllability and strong generalization to diverse object geometries are required. However, existing methods rarely satisfy these requirements, limiting their practical applicability. We present DCGrasp, a distance-aware controllable grasp generation system built on a novel grasp energy term. This term computes Distance Profile, a signed distance from each hand vertex to the nearest object point, coupled with distance-aware weighting, effectively capturing the semantically similar hand-object interaction in near-contact regions while remaining invariant to object and hand identity. Given various controllable signals, DCGrasp first generates a Distance Profile based on a Diffusion Transformer, together with a corresponding candidate hand pose. We then refine the candidate pose through optimization, enforcing consistency between the optimized hand pose and the generated Distance Profile in near-contact regions. Our experiments show that DCGrasp produces high-quality, physically plausible grasps with flexible user control, generalizing to diverse object and hand shapes and scales. Our work establishes a robust and versatile pipeline for the synthesis of controllable 3D hand-object interactions.
Chinese Translation
生成三维手-物体交互对于机器人技术、扩展现实（XR）和合成数据生成等应用至关重要，这些应用要求具备灵活的可控性和对多样物体几何形状的强泛化能力。然而，现有方法很少满足这些要求，限制了其实际应用性。我们提出了DCGrasp，一个基于新颖抓取能量项的距离感知可控抓取生成系统。该能量项计算距离轮廓（Distance Profile），即每个手部顶点到最近物体点的有符号距离，并结合距离感知加权，有效捕捉在近接触区域内语义相似的手-物体交互，同时对物体和手的身份保持不变。在给定各种可控信号的情况下，DCGrasp首先基于扩散变换器（Diffusion Transformer）生成距离轮廓及相应的候选手势。然后，我们通过优化来细化候选手势，确保优化后的手势与在近接触区域生成的距离轮廓之间的一致性。我们的实验表明，DCGrasp能够生成高质量、物理上合理的抓取，并具备灵活的用户控制，能够泛化到多样的物体和手的形状与尺度。我们的工作建立了一个强大且多功能的管道，用于合成可控的三维手-物体交互。
cs.CV / 176 / 2606.29928
Latent-CURE for Breast Cancer Diagnosis
用于乳腺癌诊断的潜在-CURE
Abstract
Multimodal Large Models have significantly advanced automated breast ultrasound diagnosis. However, most existing frameworks utilize opaque, end-to-end paradigms prioritizing global statistical correlations over structured clinical reasoning. Consequently, these models remain susceptible to shortcut learning amid extreme real-world epidemiological imbalances, often bypassing rare but decisive malignant indicators for dominant benign patterns. To address this disconnect, we propose Latent-CURE, a novel diagnostic framework driven by asymmetric weighted chain-of-thought methodology grounded in latent space reasoning. Unlike traditional approaches, our framework constructs an implicit reasoning trajectory forcing the model to sequentially infer standardized BI-RADS morphological descriptors before converging on a final diagnosis. Furthermore, to combat the extreme scarcity of critical malignant features, we couple this architecture with a dual-asymmetric optimization strategy. By dynamically adjusting margins and weights, this strategy safeguards high-specificity malignant descriptors from being overshadowed by common benign priors. Comprehensive evaluations demonstrate that our knowledge-injected approach provides transparent clinical evidence while achieving robust, accurate diagnostic performance in imbalanced medical cohorts.
Chinese Translation
多模态大型模型在自动化乳腺超声诊断方面取得了显著进展。然而，现有的大多数框架采用不透明的端到端范式，优先考虑全球统计相关性而非结构化的临床推理。因此，这些模型在极端的现实世界流行病学不平衡中仍然容易受到捷径学习的影响，常常忽视稀有但决定性的恶性指标，而偏向于占主导地位的良性模式。为了解决这一脱节，我们提出了潜在-CURE（Latent-CURE），这是一种新颖的诊断框架，采用基于潜在空间推理的非对称加权思维链方法。与传统方法不同，我们的框架构建了一个隐式推理轨迹，迫使模型在最终诊断之前依次推断标准化的BI-RADS形态描述符。此外，为了应对关键恶性特征的极度稀缺，我们将该架构与双重非对称优化策略相结合。通过动态调整边际和权重，该策略保护高特异性的恶性描述符不被常见的良性先验所掩盖。全面的评估表明，我们的知识注入方法提供了透明的临床证据，同时在不平衡的医学队列中实现了稳健、准确的诊断性能。
cs.CV / 177 / 2606.29942
Scene-aware Prediction of Diverse Human Movement Goals
场景感知的多样化人类运动目标预测
Abstract
Anticipation of human behaviours facilitates autonomous systems in proactive planning. Human behaviour could be stochastic due to varying goals. Human goals typically guide their own movement and could therefore help to predict the human trajectory and human motion in the long-term. To infer the human movement intentions, the environmental context plays a significant role, in addition to the social cues expressed by the individual. Previous works on human goals prediction either require semantic knowledge of the scene, or only tackle interactions with objects. In this paper, we propose a novel multi-goal prediction method using the generative model to address the stochasticity of human movement. It leverages the current RGB scene and the human pose to predict diverse potential future goals of human movement based on the Conditional Variational Autoencoder (CVAE). Our results demonstrate that our approach is capable of generating multiple movement goals in the scene via samplings in latent space of the CVAE and exhibits generalization capability across scenarios in GTA-IM dataset and PROX dataset. Code is publicly available at \href{https://github.com/Q-Y-Yang/DiverseGoalsPrediction.git}{\texttt{https://github.com/Q-Y-Yang/DiverseGoalsPrediction}}.
Chinese Translation
对人类行为的预判有助于自主系统进行主动规划。由于目标的多样性，人类行为可能具有随机性。人类目标通常引导他们自身的运动，因此可以帮助预测人类的轨迹和长期运动。除了个体所表达的社会线索外，环境上下文在推断人类运动意图中也起着重要作用。以往关于人类目标预测的研究要么需要对场景的语义知识，要么仅处理与物体的交互。本文提出了一种新颖的多目标预测方法，利用生成模型来解决人类运动的随机性。该方法基于条件变分自编码器（Conditional Variational Autoencoder, CVAE），利用当前的RGB场景和人类姿态来预测人类运动的多样化潜在未来目标。我们的结果表明，该方法能够通过在CVAE的潜在空间中进行采样生成场景中的多个运动目标，并在GTA-IM数据集和PROX数据集中展示了跨场景的泛化能力。代码已公开，地址为 exttt{https://github.com/Q-Y-Yang/DiverseGoalsPrediction}.
cs.CV / 178 / 2606.29963
Explainability-Aware Frustum Attack: Exposing Structural Vulnerabilities in LiDAR-Based 3D Object Detectors
可解释性意识的截锥攻击：揭示基于LiDAR的3D目标检测器的结构脆弱性
Abstract
The structural vulnerabilities of point cloud-based 3D object detectors remain poorly understood. Prior work has studied adversarial robustness primarily on isolated 3D object models, while recent LiDAR spoofing attacks target richer and more realistic driving scenes but focus mainly on physical realizability rather than understanding detector behavior or attack efficiency. In this work, we investigate how LiDAR-based detectors rely on spatial evidence in complex scenes and whether these reliance patterns can be exploited to induce failures more efficiently. To this end, we propose an explainability-guided adversarial analysis methodology. We introduce the Saliency-LiDAR (SALL) method, which aggregates Integrated Gradient attributions across scenes to produce universal saliency maps for LiDAR-based 3D object detectors. Guided by these maps, we design the Explainability-aware Frustum Attack (EFA), which selectively perturbs only the most influential frustums rather than uniformly attacking entire object regions. Experiments on KITTI and nuScenes, across detectors such as PointPillars and SECOND, show that EFA reduces detection recall by more than 15 percentage points while requiring 25-50% fewer perturbed frustums than the state-of-the-art non-saliency-aware baseline. These findings reveal that modern 3D detectors concentrate discriminative evidence in a small subset of spatial regions, exposing a structural robustness vulnerability in current LiDAR perception systems. Our code is released at https://github.com/SecMindLab/Saliency_LiDAR.
Chinese Translation
基于点云的3D目标检测器的结构脆弱性仍然不够清楚。之前的研究主要集中在孤立的3D目标模型上的对抗鲁棒性，而最近的LiDAR欺骗攻击则针对更丰富和更真实的驾驶场景，但主要关注物理可实现性，而非理解检测器的行为或攻击效率。在本研究中，我们探讨了基于LiDAR的检测器如何依赖复杂场景中的空间证据，以及这些依赖模式是否可以被利用以更高效地诱导失败。为此，我们提出了一种以可解释性为指导的对抗分析方法。我们引入了Saliency-LiDAR (SALL) 方法，该方法在不同场景中聚合积分梯度归因，以生成基于LiDAR的3D目标检测器的通用显著性图。在这些图的指导下，我们设计了可解释性意识的截锥攻击 (EFA)，该攻击选择性地扰动最具影响力的截锥，而不是均匀攻击整个目标区域。在KITTI和nuScenes数据集上的实验中，针对PointPillars和SECOND等检测器，EFA将检测召回率降低了超过15个百分点，同时所需的扰动截锥数量比最先进的非显著性意识基线少25-50%。这些发现揭示了现代3D检测器在少数空间区域内集中判别证据，暴露了当前LiDAR感知系统的结构鲁棒性脆弱性。我们的代码已发布在 https://github.com/SecMindLab/Saliency_LiDAR。
cs.CV / 179 / 2606.29964
Variance Reduction on the Camera Axis: Multi-View Score Distillation for 3D
相机轴上的方差减少：用于三维的多视角评分蒸馏
Abstract
Score distillation turns a pretrained 2D diffusion model into a 3D generator, but the per-step gradient is estimated from a single randomly chosen view: it is high-variance and blind to global shape consistency. Prior work addresses this by retraining the diffusion prior on multi-view data; this improves consistency but makes the sampling contribution inseparable from prior quality. We instead isolate the sampling axis. The per-step gradient is one noisy sample of an expectation over views; aggregating K samples per step at a fixed total UNet budget reduces variance without touching the prior. We introduce Multi-View Aggregated Score Distillation (MV-SDI), which aggregates gradients from K views per step via gradient accumulation, keeping peak memory unchanged and the 2D prior frozen, and draws views as antithetic antipodal pairs, a prior-independent geometric property, for balanced angular coverage. At a fixed 10,000-UNet-call budget, K=2 raises CLIP R-Precision from 74.8% to 83.8% and CLIP score from 0.297 to 0.312, with consistent gains on HPSv2 and ImageReward and a 0.0% divergence rate on the 43-prompt benchmark; optimization steps halve as a consequence. K=4 gives a fourfold step reduction at R-Precision 86.9% and CLIP 0.307, still well above the single-view baseline on every alignment metric. MV-SDI is compatible with gradient-based score-distillation pipelines, including Score Distillation via Inversion, and requires no retraining and no multi-view data.
Chinese Translation
评分蒸馏将预训练的二维扩散模型转化为三维生成器，但每一步的梯度是从单个随机选择的视角估计的：这导致高方差并且无法保证全局形状一致性。之前的研究通过在多视角数据上重新训练扩散先验来解决这个问题；这虽然提高了一致性，但使得采样贡献与先验质量无法分离。我们则选择隔离采样轴。每一步的梯度是对多个视角期望的一个噪声样本；在固定的总UNet预算下，每步聚合K个样本可以减少方差而不影响先验。我们提出了多视角聚合评分蒸馏（Multi-View Aggregated Score Distillation，MV-SDI），通过梯度累积在每一步聚合来自K个视角的梯度，保持峰值内存不变并将二维先验冻结，同时将视角作为对立的对偶对进行抽取，这是一种与先验无关的几何特性，以实现平衡的角度覆盖。在固定的10,000次UNet调用预算下，K=2将CLIP R-Precision从74.8%提高到83.8%，CLIP分数从0.297提高到0.312，在HPSv2和ImageReward上也有一致的提升，并且在43个提示基准上保持0.0%的发散率；因此，优化步骤减少了一半。K=4在R-Precision 86.9%和CLIP 0.307下实现了四倍的步骤减少，在每个对齐指标上仍然远高于单视角基线。MV-SDI与基于梯度的评分蒸馏管道兼容，包括通过反演进行的评分蒸馏，并且不需要重新训练和多视角数据。
cs.CV / 180 / 2606.29976
Learning Efficient 4D Gaussian Representations from Monocular Videos with Flow Splatting
从单目视频中学习高效的4D高斯表示法与流量溅射
Abstract
Reconstructing dynamic 3D scenes from monocular videos is challenging due to scene complexity and temporal dynamics. With the advancement of 3D Gaussian Splatting in novel view synthesis, existing methods extend 3D Gaussians to 4D domain with deformation fields, trajectories or spatiotemporal 4D volumes to model scene element deformation. However, these methods suffer from long training time, low rendering speed or high memory consumption for per-frame reconstruction of 4D volumes, without fully exploiting dense dynamic information. To address this issue, we propose Flow Splatting, which constructs the velocity field and enables the conventional splatting technique to render optical flow from the velocity field to supervise dynamics learning process from monocular videos. Specifically, we extend 4D volumes with time varying means and covariance to represent complex dynamics. Then, we construct and approximate the velocity field naturally based on this representations. While conventional volume rendering techniques support to render color fields, we extend the volume rendering strategy to splat the velocity field by considering the influence of camera motions. We conduct experiments on various benchmarks to demonstrate the efficiency and effectiveness of our method. Compared to the state-of-the-art methods, our model achieves better image quality with less time consumption and higher rendering speed.
Chinese Translation
从单目视频重建动态3D场景具有挑战性，主要由于场景复杂性和时间动态性。随着3D高斯溅射在新视角合成中的进展，现有方法将3D高斯扩展到具有变形场、轨迹或时空4D体积的4D领域，以建模场景元素的变形。然而，这些方法在每帧重建4D体积时面临长时间训练、低渲染速度或高内存消耗的问题，且未能充分利用密集的动态信息。为了解决这一问题，我们提出了流量溅射（Flow Splatting），该方法构建速度场，并使传统的溅射技术能够从速度场渲染光流，以监督从单目视频中学习动态的过程。具体而言，我们通过时间变化的均值和协方差扩展4D体积，以表示复杂的动态。然后，我们基于这些表示自然地构建和近似速度场。虽然传统的体积渲染技术支持渲染颜色场，但我们通过考虑相机运动的影响，将体积渲染策略扩展到溅射速度场。我们在各种基准上进行实验，以证明我们方法的高效性和有效性。与最先进的方法相比，我们的模型在更少的时间消耗和更高的渲染速度下实现了更好的图像质量。
cs.CV / 181 / 2606.29997
Rigel: Self-Distilled Score Adaptation for Image and Video Captioning Evaluation
Rigel：用于图像和视频字幕评估的自蒸馏评分适应
Abstract
Automatic evaluation of image and video captioning is essential for benchmarking multimodal systems, although standard evaluation metrics show limited alignment with human judgments. Recent approaches using large language models (LLMs), commonly referred to as LLM-as-a-Judge, have improved alignment with human judgments but still suffer from a mismatch between large-vocabulary language modeling and evaluation over a small label set. To address this, we propose Rigel, an automatic evaluation metric for image and video captioning, based on self-distilled score adaptation. The metric employs an evaluation-specific scoring head distilled from a frozen LLM, which captures judgment signals in a task-aligned space without relying on large-vocabulary token sets. We then refine the LLM backbone with human judgment data. To train Rigel, we constructed the Vid-Lepus dataset, which contains 3,338 video clips, 33,380 reference captions, and 5,637 candidate captions. Experiments on multiple benchmarks show that Rigel outperforms state-of-the-art metrics, achieving over 10-point improvements on ActivityNet-Fact in the reference-free setting.
Chinese Translation
图像和视频字幕的自动评估对于基准测试多模态系统至关重要，尽管标准评估指标与人类判断的对齐程度有限。最近使用大型语言模型（LLMs）的方法，通常称为LLM-as-a-Judge，改善了与人类判断的对齐，但仍然存在大型词汇语言建模与小标签集评估之间的不匹配。为了解决这个问题，我们提出了Rigel，一种基于自蒸馏评分适应的图像和视频字幕自动评估指标。该指标采用从冻结的LLM中蒸馏出的特定于评估的评分头，能够在任务对齐的空间中捕捉判断信号，而无需依赖大型词汇令牌集。然后，我们利用人类判断数据对LLM主干进行优化。为了训练Rigel，我们构建了Vid-Lepus数据集，其中包含3,338个视频片段、33,380个参考字幕和5,637个候选字幕。在多个基准测试上的实验表明，Rigel的表现优于最先进的指标，在无参考设置下在ActivityNet-Fact上实现了超过10分的提升。
cs.CV / 182 / 2606.30001
SICAGE: Speaker-Independent Culture-Aware Gesture Generation using TED4C-L Dataset
SICAGE：基于TED4C-L数据集的说话者无关文化感知手势生成
Abstract
Recent co-speech gesture generation methods often overlook cultural differences, limiting their effectiveness in human-agent interaction. Moreover, culture-conditioned models are rarely evaluated under speaker-disjoint splits, so apparent "cultural" behavior may be confounded with speaker-specific gesturing style. We introduce SICAGE, a modular framework for culture-aware co-speech gesture generation that conditions motion synthesis models on speaker-independent cultural representations. SICAGE learns these representations from audio and text by treating each speaker as a separate domain while imposing invariance across speakers. This encourages representations to remain culture-discriminative while reducing dependence on speaker identity. The resulting cultural embeddings condition a multimodal generator to produce culturally appropriate gestures. We instantiate this idea with two domain generalization approaches: adversarial learning and Fishr regularization. We further introduce ALaDiT, a real-time diffusion-based gesture generator designed to efficiently incorporate the learned cultural embeddings. To validate our method, we built TED4C-L, a 106-hour multimodal dataset of 764 TED speakers from four cultural groups. Experiments show that SICAGE improves motion realism, diversity, beat synchronization, semantic relevance, and cultural consistency.
Chinese Translation
近期的共语手势生成方法往往忽视文化差异，限制了其在人机交互中的有效性。此外，文化条件模型在说话者不重叠的分割下很少被评估，因此明显的“文化”行为可能与说话者特定的手势风格混淆。我们提出了SICAGE，一个模块化的文化感知共语手势生成框架，该框架基于说话者无关的文化表征来调节运动合成模型。SICAGE通过将每个说话者视为一个独立的领域，同时在说话者之间施加不变性，从音频和文本中学习这些表征。这促使表征保持文化区分性，同时减少对说话者身份的依赖。由此产生的文化嵌入为多模态生成器提供条件，以生成文化上适当的手势。我们通过两种领域泛化方法来实现这一理念：对抗学习和Fishr正则化。此外，我们进一步引入了ALaDiT，一个基于扩散的实时手势生成器，旨在高效地整合学习到的文化嵌入。为了验证我们的方法，我们构建了TED4C-L，一个包含来自四个文化群体的764名TED演讲者的106小时多模态数据集。实验表明，SICAGE在运动真实感、多样性、节拍同步、语义相关性和文化一致性方面都有所提升。
cs.CV / 183 / 2606.30003
GeoEdit: Geometry-Aware Object Editing via Dual-Branch Denoising
GeoEdit：基于双分支去噪的几何感知对象编辑
Abstract
Precisely manipulating objects in a single photograph (translation, rotation, scaling) while obeying 3D physical constraints remains unsolved for diffusion-based editors. Current 2D methods lack spatial awareness and produce perspective violations. Forcing structural proxies into the latent space also disrupts variance homogeneity, and the resulting self-attention leakage leads to ghosting and background blur. The core difficulty is asymmetric: the relocated object must follow a rigid geometry, yet the uncovered background needs freedom to synthesize plausible content. We present GeoEdit, a training-free Lift-Manipulate-Render-Denoise pipeline that satisfies both constraints. We decouple scene and object in 3D, align them through point correspondence, and render a geometry-aligned proxy with a structural depth map. A Dual-Branch Denoising stage then refines this proxy: a video diffusion backbone preserves object identity, while 3D constraints are injected into the foreground within a narrow denoising window at matching noise variance (variance-homogeneous injection). The background denoises freely. Because the injected signal matches the native latent statistics, self-attention stays undisturbed. We also introduce GeoEditBench, a pose-aware benchmark covering object translation, object rotation, and camera movement with pose-aware evaluation metrics. Experiments confirm consistent gains in geometric accuracy, identity fidelity, and background quality. Our codes are available at https://github.com/Heey731/GeoEdit.
Chinese Translation
在遵循三维物理约束的情况下，精确操控单张照片中的对象（平移、旋转、缩放）仍然是基于扩散的编辑器未解决的问题。目前的二维方法缺乏空间意识，导致透视违背。将结构代理强行引入潜在空间也会破坏方差同质性，导致自注意力泄漏，从而产生重影和背景模糊。核心难点是非对称的：重新定位的对象必须遵循刚性几何，而未覆盖的背景则需要自由以合成合理的内容。我们提出了GeoEdit，一个无训练的提升-操控-渲染-去噪管道，满足这两种约束。我们在三维空间中解耦场景和对象，通过点对应将它们对齐，并渲染出一个几何对齐的代理及其结构深度图。接下来，双分支去噪阶段对该代理进行精细化：视频扩散主干保持对象身份，而三维约束则在匹配噪声方差的狭窄去噪窗口内注入到前景中（方差同质注入）。背景则自由去噪。由于注入信号与本地潜在统计相匹配，自注意力保持不受干扰。我们还引入了GeoEditBench，一个涵盖对象平移、对象旋转和相机运动的姿态感知基准，并提供姿态感知评估指标。实验结果确认了几何准确性、身份保真度和背景质量的一致提升。我们的代码可在 https://github.com/Heey731/GeoEdit 获取。
cs.CV / 184 / 2606.30012
SkelEM: Training-Signal Decoupling of Skeleton and Diffusion for Self-supervised Axial Super-Resolution in Volume Microscopy
SkelEM：体积显微镜中自监督轴向超分辨率的骨架与扩散训练信号解耦
Abstract
Volume microscopy, including electron and light microscopy, suffers from severe anisotropic resolution due to physical axial sectioning. Existing self-supervised axial super-resolution (ASR) methods face a trilemma bounded by overly smoothed regression textures, structural hallucinations of pure diffusion models, and prohibitive inference latency. In this paper, we propose Skeleton-refinE Microscopy (SkelEM), a self-supervised framework that decouples ASR at the training-signal level: a frozen topological network and a diffusion refiner are optimized by disjoint objectives, separating low-frequency topology formulation from high-frequency detail enhancement. Building on this deterministic skeleton, we exploit a unified cycle-consistent mechanism on input sparse slices to simultaneously extract a real-domain residual prior and bidirectionally align the diffusion refiner, washing away cross-plane artifacts without synthetic bias. By truncating the reverse diffusion process with this physical prior, SkelEM achieves high-fidelity detail restoration in merely $\le 5$ steps. To rigorously assess cross-instrument generalization, we further introduce BRAVE-ASR, a new benchmark of co-aligned anisotropic and isotropic volumes acquired on a Plasma-FIB instrument. Across public benchmarks, SkelEM achieves the most favorable balance across the fidelity-perception trade-off among self-supervised methods, with state-of-the-art downstream membrane segmentation performance and robust zero-shot generalization across distinct modalities.
Chinese Translation
体积显微镜（包括电子显微镜和光学显微镜）由于物理轴向切片的限制，面临严重的各向异性分辨率问题。现有的自监督轴向超分辨率（ASR）方法面临三重困境：过度平滑的回归纹理、纯扩散模型的结构幻觉以及高昂的推理延迟。在本文中，我们提出了骨架精细显微镜（SkelEM），这是一个自监督框架，在训练信号层面上解耦ASR：一个冻结的拓扑网络和一个扩散精细化器通过不相交的目标进行优化，从而将低频拓扑构造与高频细节增强分离。基于这一确定性骨架，我们在输入稀疏切片上利用统一的循环一致机制，同时提取真实域的残差先验，并双向对齐扩散精细化器，消除跨平面伪影而不引入合成偏差。通过使用这一物理先验截断反向扩散过程，SkelEM在仅需$ extless 5$步内实现高保真度的细节恢复。为了严格评估跨仪器的泛化能力，我们进一步引入BRAVE-ASR，这是一个新的基准，包含在Plasma-FIB仪器上获取的共同对齐的各向异性和各向同性体积。在公共基准测试中，SkelEM在自监督方法中实现了保真度与感知之间最佳平衡，具有最先进的下游膜分割性能以及在不同模态之间的强健零样本泛化能力。
cs.CV / 185 / 2606.30014
Shell-Supervised Gaussian Splatting for Urban Real-to-Sim Reconstruction
壳监督高斯点云重建用于城市真实到仿真重建
Abstract
Real-to-sim reconstruction for embodied AI requires geometry that is useful for collision reasoning, navigation, and agent-environment interaction, not only photorealistic novel-view synthesis. However, close-range urban facades are difficult for video-to-3D reconstruction: glass, reflections, repeated windows, and weak texture can produce visually plausible renderings with unstable surface geometry. We introduce shell-supervised Gaussian Splatting, a reconstruction-stage framework that uses an external facade structural shell as lightweight geometric supervision for video-driven Gaussian reconstruction. The method aligns an exterior shell to the video reconstruction frame, renders per-view depth, camera-space normal, and valid-mask maps, and applies these cues through mask-gated losses during Gaussian optimization. This design preserves RGB-driven appearance while regularizing only visible shell-supported facade regions. Experiments on anonymized close-range urban facade scenes show improved facade orientation and visible-surface point-cloud consistency over photo-only, monocular-cue, and surface-oriented Gaussian baselines, while maintaining comparable held-out rendering quality.
Chinese Translation
针对具身人工智能的真实到仿真重建需要几何体，以便于碰撞推理、导航和智能体与环境的交互，而不仅仅是逼真的新视角合成。然而，近距离城市立面的视频到三维重建面临挑战：玻璃、反射、重复的窗户和弱纹理可能产生视觉上可信的渲染，但表面几何不稳定。我们提出了壳监督高斯点云重建，这是一种重建阶段框架，利用外部立面结构壳作为轻量级几何监督，驱动视频的高斯重建。该方法将外部壳与视频重建帧对齐，渲染每视图的深度、相机空间法线和有效掩膜图，并在高斯优化过程中通过掩膜门控损失应用这些线索。该设计在仅对可见的壳支撑立面区域进行正则化的同时，保留了基于RGB的外观。对匿名的近距离城市立面场景的实验表明，相较于仅使用照片、单目线索和表面导向的高斯基线，该方法在立面方向和可见表面点云一致性方面有所改善，同时保持了可比的保留渲染质量。
cs.CV / 186 / 2606.30017
Monte Carlo Energy Aggregation for Mobile 3D Gaussian Splatting
移动3D高斯点云的蒙特卡洛能量聚合
Abstract
Recent advances in 3D Gaussian Splatting have demonstrated unprecedented success in novel view synthesis. However, the substantial inference and storage overhead driven by high-order Spherical Harmonics (SH) are primary bottlenecks for mobile platforms. In this paper, we present Flux-GS, a real-time Gaussian Splatting method designed to achieve high-fidelity rendering with significantly reduced overhead for resource-constrained mobile platforms. We first propose a Monte Carlo Specular Energy Aggregator, sampling third-order radiance residuals and aggregating specular energy into a compact latent space. In this way, our method effectively preserves visually salient lighting features in lower-order bands without expensive distillation or pre-training. To mitigate the high-frequency details lost during compression, we introduce an Attribute-Conditioned SH Enhancement module. This module predicts Gaussian-aware offsets based on intrinsic Gaussian attributes, which enhance the first-order SH representation prior to inference, without extra inference costs. Furthermore, the original single-view gradient-based densification is prone to producing excessive Gaussians and overfitting to a certain view. We address these limitations by proposing a Multi-view Alpha-based Densification and Pruning strategy. By leveraging multi-view guidance, we ensure multi-view structure consistency and the precise removal of redundant primitives. Extensive experiments demonstrate that Flux-GS achieves substantial parameter reduction while maintaining competitive visual quality, offering a robust and scalable solution for real-time mobile rendering. Code: \textcolor{magenta}{\href{https://xiaobiaodu.github.io/flux-gs-project/}{https://xiaobiaodu.github.io/flux-gs-project/}}.
Chinese Translation
最近在3D高斯点云技术方面的进展在新视图合成中展现了前所未有的成功。然而，由高阶球谐函数（Spherical Harmonics, SH）驱动的显著推理和存储开销是移动平台的主要瓶颈。本文提出了一种实时高斯点云方法Flux-GS，旨在为资源受限的移动平台实现高保真渲染，同时显著降低开销。我们首先提出了一种蒙特卡洛镜面能量聚合器，采样三阶辐射残差并将镜面能量聚合到一个紧凑的潜在空间中。通过这种方式，我们的方法有效地在低阶频带中保留了视觉上显著的光照特征，而无需昂贵的蒸馏或预训练。为了减轻压缩过程中丢失的高频细节，我们引入了一个属性条件的SH增强模块。该模块基于内在的高斯属性预测高斯感知的偏移量，从而在推理之前增强一阶SH表示，而无需额外的推理成本。此外，原始的单视图基于梯度的稠密化方法容易产生过多的高斯点并对某一视图过拟合。我们通过提出一种基于多视图的Alpha稠密化和修剪策略来解决这些局限性。通过利用多视图指导，我们确保了多视图结构的一致性和冗余原语的精确去除。大量实验表明，Flux-GS在保持竞争视觉质量的同时，实现了显著的参数减少，提供了一种稳健且可扩展的实时移动渲染解决方案。代码： extcolor{magenta}{ exttt{https://xiaobiaodu.github.io/flux-gs-project/}}.
cs.CV / 187 / 2606.30019
OmniDance: Multimodal Driven Dance Video Generation with Large-scale Internet Data
OmniDance：基于多模态驱动的大规模互联网数据舞蹈视频生成
Abstract
Music-driven dance video generation aims to synthesize expressive human motion that is temporally aligned with music while maintaining high visual fidelity. Despite recent progress, existing methods still face two key limitations: the lack of large-scale, high-quality dance video datasets, and the absence of principled frameworks for integrating music as a complementary conditioning signal into Video Generation Foundation Models. To address these limitations, we introduce CIPE-Dance, a large-scale Internet-sourced dance video dataset with choreography-informed text annotations, constructed via a progressive expert pipeline. To the best of our knowledge, CIPE-Dance is the largest dataset for dance video generation to date, comprising 300k high-quality clips over 400 hours and covering diverse dancers, environments, and dance genres. We further propose OmniDance, a framework-level recipe for integrating music into a TI2V foundation model without sacrificing its original controllability or visual fidelity. Motivated by the complementary roles of text as low-frequency semantics and music as high-frequency temporal dynamics, OmniDance co-designs a depth-aware specialization architecture, an anchored easy-to-hard curriculum learning strategy, and a modality-specialized time-dependent CFG strategy, enabling unified TI2V, MI2V, and MTI2V generation. Extensive experiments on CIPE-Dance demonstrate that OmniDance achieves state-of-the-art performance across all three tasks and exhibits robust multimodal integration capability. Project is available at https://github.com/AMAP-ML/OmniDance.
Chinese Translation
音乐驱动的舞蹈视频生成旨在合成与音乐时间上对齐的富有表现力的人体运动，同时保持高视觉保真度。尽管近期取得了一些进展，现有方法仍面临两个主要限制：缺乏大规模、高质量的舞蹈视频数据集，以及缺乏将音乐作为补充条件信号整合进视频生成基础模型的原则性框架。为了解决这些限制，我们引入了CIPE-Dance，这是一个基于互联网的大规模舞蹈视频数据集，配有舞蹈编排信息的文本注释，通过渐进式专家流程构建。根据我们所知，CIPE-Dance是迄今为止最大的舞蹈视频生成数据集，包含30万段高质量视频，总时长超过400小时，涵盖了多样的舞者、环境和舞蹈风格。我们进一步提出了OmniDance，这是一个将音乐整合进TI2V基础模型的框架级方案，且不牺牲其原有的可控性或视觉保真度。受文本作为低频语义和音乐作为高频时间动态的互补角色的启发，OmniDance共同设计了一个深度感知的专业化架构、一个锚定的易到难的课程学习策略，以及一个模态专用的时间依赖CFG策略，从而实现统一的TI2V、MI2V和MTI2V生成。在CIPE-Dance上的大量实验表明，OmniDance在所有三个任务上都达到了最先进的性能，并展现了强大的多模态集成能力。项目可在 https://github.com/AMAP-ML/OmniDance 获取。
cs.CV / 188 / 2606.30020
Uncertainty Estimation in Pathology Foundation Models via Deep Mutual Learning
通过深度互学习进行病理基础模型的不确定性估计
Abstract
Pathology foundation models (PFMs) offer generalizable representations for whole-slide image (WSI) analysis, yet their clinical adoption remains limited. Specifically, their predictions lack reliable confidence estimates, and no single PFM is universally best across tasks, which severely undermines trust in medical settings. To overcome this, we propose $\mathtt{DICE}$, a plug-and-play framework that ensembles $K$ frozen PFMs and models their disagreement as a proxy for uncertainty estimation. To ensure this proxy yields meaningful estimates, we align the ensemble members via deep mutual learning, and theoretically show that this objective upper-bounds the model uncertainty. Additionally, we demonstrate that the ensemble's consensus localizes abnormalities at the patch level without any explicit supervision. We evaluate $\mathtt{DICE}$ on three challenging WSI benchmarks. Notably, our framework provides reliable uncertainty estimates that accurately flag failure-prone cases under in- and out-of-distribution settings, while matching or outperforming SOTA baselines in classification, calibration, and localization. Overall, $\mathtt{DICE}$ takes a crucial step toward translating PFMs into uncertainty-aware decision-support systems.
Chinese Translation
病理基础模型（PFMs）为全切片图像（WSI）分析提供了可泛化的表征，但其临床应用仍然有限。具体而言，它们的预测缺乏可靠的置信度估计，且没有单一的PFM在各项任务中表现最佳，这严重削弱了在医疗环境中的信任。为此，我们提出了$ exttt{DICE}$，一个即插即用的框架，通过集成$K$个冻结的PFM，并将它们之间的分歧建模为不确定性估计的代理。为了确保这个代理能够产生有意义的估计，我们通过深度互学习对集成成员进行对齐，并理论上证明了这一目标上界了模型的不确定性。此外，我们展示了集成的共识能够在没有任何显式监督的情况下定位异常。我们在三个具有挑战性的WSI基准上评估了$ exttt{DICE}$。值得注意的是，我们的框架提供了可靠的不确定性估计，能够准确标记在分布内和分布外环境下的易出错案例，同时在分类、校准和定位方面与最先进的基线相匹配或超越。总体而言，$ exttt{DICE}$在将PFM转化为具有不确定性感知的决策支持系统方面迈出了重要一步。
cs.CV / 189 / 2606.30024
IBRSteG: Learning a Generalizable Steganography Framework for 3D Gaussian Splatting
IBRSteG：用于3D高斯点云的可泛化隐写框架
Abstract
Recent advances in deep learning have notably improved steganographic message hiding. However, designing a generalizable steganographic approach for 3D Gaussian Splatting (3DGS) that can embed meaningful 3D scene content remains challenging. In this paper, we propose IBRSteG, a generalizable framework for 3DGS steganography that enables undetectable concealment of secret scenes within a steganographic scene. Unlike existing approaches whose parameter generation is rigidly coupled with the specific scene, we formulate 3D steganography as a feed-forward 3D Gaussian embedding process that generalizes across different 3DGS scenes. To realize this, we introduce GAS (Gaussian Attributes Steganographer), a network that learns a scene-independent embedding function by injecting the attributes of secret 3D Gaussian points into a cover scene, thereby directly reconstructing the steganographic scenes without per-scene finetuning or optimization. By transforming 3D Gaussian into these structured attributes, these attributes are compatible with 2D learning paradigms and benefit from their structured nature, thereby enhancing generalization to unseen 3DGS scenes. Extensive experiments on established datasets demonstrate that IBRSteG can effectively conceal different scenes with high visual quality, and achieves superior capacity and security. Code is available at https://github.com/LingXiang2023/IBRSteG.
Chinese Translation
近年来，深度学习的进步显著提升了隐写消息的隐藏能力。然而，设计一种可泛化的3D高斯点云（3DGS）隐写方法，以嵌入有意义的3D场景内容仍然具有挑战性。本文提出了IBRSteG，这是一种用于3DGS隐写的可泛化框架，能够在隐写场景中实现对秘密场景的不可检测隐藏。与现有方法中参数生成严格依赖于特定场景不同，我们将3D隐写形式化为一个前馈的3D高斯嵌入过程，该过程能够跨不同的3DGS场景进行泛化。为实现这一目标，我们引入了GAS（高斯属性隐写器），该网络通过将秘密3D高斯点的属性注入到覆盖场景中，学习一种与场景无关的嵌入函数，从而直接重建隐写场景，而无需针对每个场景进行微调或优化。通过将3D高斯转换为这些结构化属性，这些属性与2D学习范式兼容，并受益于其结构化特性，从而增强了对未见过的3DGS场景的泛化能力。在已建立的数据集上的大量实验表明，IBRSteG能够有效地以高视觉质量隐蔽不同场景，并实现更优的容量和安全性。代码可在 https://github.com/LingXiang2023/IBRSteG 获取。
cs.CV / 190 / 2606.30026
MuseBench: Benchmarking Intent-Level Audiovisual Arts Understanding in MLLMs
MuseBench：多模态大语言模型中意图层面视听艺术理解的基准测试
Abstract
Audiovisual arts encompass diverse creative disciplines, including cinema, visual arts, stage performance, and game design, where artistic meaning arises from deliberate combinations of visual, auditory, and narrative elements (e.g., fear amplified through claustrophobic framing, or grief conveyed through silence and lingering close-ups). True artistic understanding extends beyond recognizing what is depicted to reasoning about why it is expressed through particular creative choices. Despite the strong progress of multimodal large language models (MLLMs), this critical aspect of artistic understanding remains underexplored, as existing benchmarks largely measure perceptual recognition while overlooking reasoning about creative intent. To address this gap, we introduce Musebench, a comprehensive benchmark designed to evaluate MLLMs on nuanced artistic understanding. It comprises 4,016 questions spanning cinematic arts, static visual arts, stage performing arts, and game arts, distilled from over 10K candidate video essays that pair professional commentary with visual demonstration. To capture the open-ended nature of artistic analysis at scale, the benchmark combines single-select and variable-option multi-select questions. All questions are generated and refined through a four-phase iterative pipeline combining shortcut filtering, adversarial distractors, and expert validation. Comprehensive zero-shot evaluation of 28 state-of-the-art MLLMs reveals that even the best-performing model achieves only 48.29% accuracy, substantially below human expert performance of 87.18%, exposing a significant gap in current models' creative domain expertise.
Chinese Translation
视听艺术涵盖多种创意学科，包括电影、视觉艺术、舞台表演和游戏设计，其中艺术意义源于视觉、听觉和叙事元素的有意组合（例如，通过幽闭恐惧症的构图增强恐惧感，或通过沉默和细腻的特写镜头传达悲伤）。真正的艺术理解不仅仅是识别所描绘的内容，更在于推理为何通过特定的创作选择来表达这些内容。尽管多模态大语言模型（MLLMs）取得了显著进展，但艺术理解的这一关键方面仍未得到充分探索，因为现有基准主要测量感知识别，而忽视了对创作意图的推理。为了解决这一空白，我们推出了Musebench，这是一个全面的基准，旨在评估MLLMs在细致艺术理解方面的表现。该基准包含4,016个问题，涵盖电影艺术、静态视觉艺术、舞台表演艺术和游戏艺术，这些问题提炼自超过10,000个候选视频论文，这些论文将专业评论与视觉演示相结合。为了在大规模上捕捉艺术分析的开放性特征，该基准结合了单选和可变选项的多选问题。所有问题均通过一个四阶段的迭代流程生成和完善，该流程结合了快捷过滤、对抗性干扰项和专家验证。对28个最先进的MLLMs进行的全面零-shot评估显示，即使是表现最佳的模型，其准确率也仅为48.29%，远低于人类专家的87.18%的表现，揭示了当前模型在创意领域专业知识方面的显著差距。
cs.CV / 191 / 2606.30027
Cross-Modal Iteration Distillation for Robust IHD Screening: The IDNet Framework and A New Benchmark
跨模态迭代蒸馏用于稳健的缺血性心脏病筛查：IDNet框架及新基准
Abstract
Color Fundus Photography (CFP) offers a low-cost and non-invasive route for ischemic heart disease (IHD) screening, but current studies are limited by scarce public benchmarks and ineffective fusion of retinal images with sparse clinical variables. We propose IDNet, a multimodal framework with a Cross-Modal Distillation Aggregator (CDA) that uses learnable queries to sequentially integrate left-eye, right-eye, and clinical features, mitigating the imbalance between high-dimensional visual features and low-dimensional tabular inputs. We also construct a reproducible UK Biobank benchmark with open-source curation and quality-control pipelines, yielding 50,410 images from 25,205 subjects. On this benchmark, IDNet outperforms image-only, clinical-only, and several multimodal baselines, and CDA consistently improves multiple visual encoders as a plug-in fusion module.
Chinese Translation
彩色眼底摄影（CFP）为缺血性心脏病（IHD）筛查提供了一种低成本且非侵入性的方法，但目前的研究受到稀缺公共基准和视网膜图像与稀疏临床变量融合效果不佳的限制。我们提出了IDNet，这是一种多模态框架，具有跨模态蒸馏聚合器（CDA），使用可学习查询顺序整合左眼、右眼和临床特征，从而缓解高维视觉特征与低维表格输入之间的不平衡。我们还构建了一个可重复的英国生物库基准，配备开源的策划和质量控制流程，产生了来自25,205个受试者的50,410张图像。在该基准上，IDNet在图像单一、临床单一和多个多模态基线模型中表现优越，而CDA作为插件融合模块持续提升了多个视觉编码器的性能。
cs.CV / 192 / 2606.30030
CogSENet: Blind Image Deblurring with Blur-Conditioned Semantic Routing and Explicit Frequency Fusion
CogSENet：基于模糊条件语义路由和显式频率融合的盲图像去模糊
Abstract
Blind image deblurring demands the recovery of high-fidelity details and coherent structures from complex, unknown degradations. Current blind image deblurring methods struggle with real-world, spatially varying degradations, and lack the semantic awareness necessary to reliably differentiate valid textures from artifacts. To bridge this gap, we propose CogSENet, a dynamic, semantic-aligned reconstruction framework inspired by the eagle's visual system. By mimicking the eagle's active saccadic scanning, we devise a Semantic-Driven State Space Module (SDSSM) with semantic-aware token regrouping via differentiable routing, enabling prompt-conditioned long-range dependency modeling. To ensure physically interpretable recovery of textures and structures, a BiFreqFusionBlock (BFFB) mirrors functional differentiation of the eagle's retina by decomposing features into high and low frequencies using wavelet transforms. Finally, we estimate a continuous Blur Field (CBF) from blur image and fuse it with CLIP semantic priors to modulate the deepest latent features, emulating focal adaptation and enabling adaptive restoration under spatially non-uniform blur. Extensive experiments demonstrate that CogSENetoutperforms state-of-the-art deblurring methods in both visual quality and structural fidelity with fewer parameters, while also performing favorably on dehazing, deraining, and denoising tasks.
Chinese Translation
盲图像去模糊要求从复杂且未知的退化中恢复高保真细节和一致的结构。目前的盲图像去模糊方法在处理真实世界中空间变化的退化时面临挑战，且缺乏可靠区分有效纹理与伪影所需的语义意识。为了解决这一问题，我们提出了CogSENet，一个动态的、语义对齐的重建框架，灵感来源于鹰的视觉系统。通过模仿鹰的主动扫视，我们设计了一个语义驱动状态空间模块（SDSSM），通过可微路由实现语义感知的标记重组，从而实现快速条件的长程依赖建模。为了确保对纹理和结构的物理可解释性恢复，双频融合块（BFFB）通过小波变换将特征分解为高频和低频，模拟鹰视网膜的功能分化。最后，我们从模糊图像中估计连续模糊场（CBF），并将其与CLIP语义先验融合，以调节最深层的潜在特征，模拟焦点适应，并在空间非均匀模糊下实现自适应恢复。大量实验表明，CogSENet在视觉质量和结构保真度上均优于最先进的去模糊方法，且参数更少，同时在去雾、去雨和去噪任务上表现良好。
cs.CV / 193 / 2606.30035
Consensus Clustering of Free-Viewing Gaze Data: New Insights into Human-Information Interaction
自由观看注视数据的共识聚类：对人类信息交互的新见解
Abstract
Free-viewing gaze data provides a rich, task-free window into human visual attention. Conventional exploratory data analysis of the data provides user attention patterns through fixations and areas of interest. However, despite the richness of this gaze data, its human-information interaction (HII) patterns are understudied. We address this gap using consensus clustering of gaze data with respect to users and stimulus characteristics. We present a novel end-to-end unsupervised ensemble learning system for consensus clustering of free-viewing gaze datasets, EnsembleGaze. With a goal of characterizing the user behavior and stimulus type, we propose a feature engineering step based on statistical descriptors of fixation-based distributions. EnsembleGaze involves consensus voting of selected clustering methods implemented on the feature vector to compute the co-association matrix. Using the separate consensus clustering of users and stimuli as a baseline, we further propose two high-dimensional clustering strategies for determining gaze clusters based on joint user and image characterization. They are consensus subspace clustering and spectral biclustering. Clustering performance is evaluated using selected standard metrics and is further interpreted through image-level properties. Our system provides a replicable method for the unsupervised analysis of fixation behavior in scene perception research. Our results show that image stimuli groupings are highly consistent across methods, reflecting a robust ambient-versus-focal viewing mode distinction, whereas user groupings are image-context-dependent, a structure that only biclustering and the two-step conditional approaches are architecturally capable of recovering. Testing on the publicly available datasets revealed dataset-specific patterns, with each offering complementary insights through distinct clustering strategies.
Chinese Translation
自由观看的注视数据为人类视觉注意力提供了一个丰富的、无任务的窗口。对这些数据的传统探索性数据分析通过注视点和兴趣区域揭示了用户的注意力模式。然而，尽管这些注视数据丰富，其人类信息交互（HII）模式仍然研究不足。我们通过针对用户和刺激特征的注视数据共识聚类来填补这一空白。我们提出了一种新颖的端到端无监督集成学习系统，用于自由观看注视数据集的共识聚类，命名为 EnsembleGaze。我们的目标是表征用户行为和刺激类型，基于注视点分布的统计描述符提出了一个特征工程步骤。EnsembleGaze 涉及对在特征向量上实施的选定聚类方法的共识投票，以计算共关联矩阵。以用户和刺激的单独共识聚类作为基线，我们进一步提出了两种高维聚类策略，以基于联合用户和图像特征来确定注视聚类。这两种策略分别是共识子空间聚类和谱双聚类。聚类性能使用选定的标准指标进行评估，并通过图像级属性进一步解释。我们的系统提供了一种可复制的方法，用于在场景感知研究中对注视行为进行无监督分析。我们的结果表明，图像刺激的分组在不同方法之间高度一致，反映出一种稳健的环境与聚焦观看模式的区分，而用户的分组则依赖于图像上下文，这种结构只有双聚类和两步条件方法才能在架构上恢复。对公开可用数据集的测试揭示了特定于数据集的模式，每个数据集通过不同的聚类策略提供互补的见解。
cs.CV / 194 / 2606.30045
Walking in the Implicit: Interactive World Exploration via Neural Scene Representation
隐式行走：通过神经场景表示进行交互式世界探索
Abstract
Interactive video generation systems for camera-controlled world exploration roll out growing sequences of latent video frames, entangling state transition with high-frequency observation synthesis. We propose Walking in the Implicit, a scene-centric paradigm that changes the rollout variable from frame latents to a fixed-length, renderable implicit state, termed Neural Implicit Scene (NIS). This factorizes interactive generation into stochastic transition of a compact scene state and deterministic pose-conditioned rendering given the sampled state. We instantiate this paradigm as NeuWorld: a transformer VAE learns locally anchored NIS from sparse posed frames, and a diffusion transformer evolves NIS conditioned on future camera trajectories and geometry-aware retrieved history. By reusing the VAE encoder as a unified conditioner, NeuWorld maps camera, reference-image, and history cues into the same NIS modality, avoiding external heterogeneous encoders. Trained from scratch on public posed-view data without pretrained video backbones or auxiliary 3D reconstructors, NeuWorld achieves strong long-horizon consistency with favorable inference efficiency.
Chinese Translation
用于相机控制的世界探索的交互式视频生成系统生成越来越长的潜在视频帧序列，将状态转移与高频观察合成交织在一起。我们提出了隐式行走（Walking in the Implicit），一种以场景为中心的范式，将生成变量从帧潜变量更改为固定长度、可渲染的隐式状态，称为神经隐式场景（Neural Implicit Scene, NIS）。这一方法将交互式生成分解为紧凑场景状态的随机转移和基于采样状态的确定性姿态条件渲染。我们将这一范式实例化为NeuWorld：一个变换器变分自编码器（transformer VAE），从稀疏姿态帧中学习局部锚定的NIS，而扩散变换器（diffusion transformer）则根据未来相机轨迹和几何感知的检索历史演化NIS。通过重用VAE编码器作为统一的条件器，NeuWorld将相机、参考图像和历史线索映射到相同的NIS模态，避免了外部异构编码器的使用。在没有预训练视频骨干网络或辅助3D重建器的情况下，从头开始在公共姿态视图数据上进行训练，NeuWorld实现了强大的长时间一致性和良好的推理效率。
cs.CV / 195 / 2606.30047
Argus: Metric Panoramic 3D Reconstruction for Indoor Scenes
Argus：室内场景的度量全景3D重建
Abstract
Metric feed-forward 3D reconstruction for panoramic data remains under-explored due to the lack of large-scale panoramic RGB-D training data. We present Realsee3D, a hybrid dataset of 10K indoor scenes (1K real, 9K synthetic) with 299K panoramic viewpoints and precise metric annotations, and Argus, a feed-forward network trained on it for metric panoramic 3D reconstruction. In the sparse unordered capture setting of Realsee3D, a poorly chosen coordinate anchor can cause global pose drift. Argus addresses this with a learned covisibility module that selects the geometrically optimal reference view to anchor the metric world frame. To further improve multi-task learning, we decompose the bidirectional pixel-to-world mapping into interpretable sub-steps with per-step supervision and cross-coordinate joint constraints, reinforcing geometric consistency across prediction branches. On the Realsee3D benchmark, Argus achieves state-of-the-art metric performance in camera pose estimation, depth estimation, and point cloud reconstruction. Project page: https://argus-paper.realsee.ai.
Chinese Translation
由于缺乏大规模的全景RGB-D训练数据，度量前馈3D重建在全景数据中的研究仍然不足。我们提出了Realsee3D，这是一个包含10,000个室内场景（1,000个真实场景，9,000个合成场景）、299,000个全景视点和精确度量注释的混合数据集，以及Argus，一个基于该数据集训练的前馈网络，用于度量全景3D重建。在Realsee3D的稀疏无序捕捉设置中，选择不当的坐标锚点可能导致全局姿态漂移。Argus通过一个学习的共视模块来解决这个问题，该模块选择几何上最优的参考视图来锚定度量世界坐标系。为了进一步改善多任务学习，我们将双向像素到世界的映射分解为可解释的子步骤，并进行逐步监督和跨坐标的联合约束，从而增强预测分支之间的几何一致性。在Realsee3D基准测试中，Argus在相机姿态估计、深度估计和点云重建方面实现了最先进的度量性能。项目页面：https://argus-paper.realsee.ai。
cs.CV / 196 / 2606.30049
Bridging the Gap Between Image Restoration and Navigational Safety in Hazy Conditions: A New Visibility Estimation Metric for Maritime Surveillance
弥合图像恢复与雾霾条件下航行安全之间的差距：一种新的海事监视可见度估计指标
Abstract
Visibility distance is critical to maritime navigational safety because it determines the effective observation range of shipborne and shore-based monitoring systems. Under hazy conditions, degraded visual information shortens observable distance and increases navigational risks and economic losses. Although numerous image dehazing methods have been developed, conventional image quality assessment metrics, such as PSNR, SSIM, FSIM, FADE, and NIQE, cannot establish a physically interpretable relationship between restoration quality and practical visibility thresholds. To address this limitation, this work proposes a visibility-oriented evaluation framework that links dehazing performance with visible-distance estimation. First, a Maritime Simulated Visibility Dataset (MSVD) is constructed using Unity3D to simulate maritime traffic scenes under graded visibility conditions. The dataset provides paired hazy and clear images with precise visibility annotations, enabling quantitative analysis of visibility restoration. Second, a dehazing visibility evaluation metric is developed by using object detection accuracy as an intermediate indicator. By establishing a mapping between visibility distance and detection performance, the proposed metric converts image restoration improvements into measurable visibility gains. Six representative dehazing methods are evaluated using both conventional image quality metrics and the proposed visibility metric. Experimental results under different imaging conditions demonstrate that MSVD provides a reliable benchmark for evaluating dehazing performance across graded visibility levels, while the proposed metric enables interpretable and reliable visible-distance estimation, thereby supporting the assessment of navigational safety and operational efficiency.
Chinese Translation
可见距离对海事航行安全至关重要，因为它决定了船载和岸基监测系统的有效观察范围。在雾霾条件下，视觉信息的退化缩短了可观察距离，增加了航行风险和经济损失。尽管已经开发了许多图像去雾方法，但传统的图像质量评估指标，如PSNR、SSIM、FSIM、FADE和NIQE，无法建立恢复质量与实际可见度阈值之间的物理可解释关系。为了解决这一局限性，本研究提出了一种以可见度为导向的评估框架，将去雾性能与可见距离估计联系起来。首先，利用Unity3D构建了一个海事模拟可见度数据集（Maritime Simulated Visibility Dataset, MSVD），以模拟不同可见度条件下的海事交通场景。该数据集提供了配对的雾霾图像和清晰图像，并附有精确的可见度注释，从而能够对可见度恢复进行定量分析。其次，开发了一种基于目标检测准确性作为中介指标的去雾可见度评估指标。通过建立可见距离与检测性能之间的映射，所提出的指标将图像恢复的改进转化为可测量的可见度增益。使用传统图像质量指标和所提出的可见度指标对六种代表性的去雾方法进行了评估。在不同成像条件下的实验结果表明，MSVD为评估不同可见度水平下的去雾性能提供了可靠的基准，而所提出的指标则实现了可解释和可靠的可见距离估计，从而支持航行安全和操作效率的评估。
cs.CV / 197 / 2606.30054
Illuminating Unified Multimodal Model for Free-form Interleaved Text-Image Generation
照亮统一多模态模型用于自由形式交错文本-图像生成
Abstract
The advancement of generative AI models capable of producing text and image marks a critical step forward in the realm of multimodal intelligence, particularly for tasks involving the interleaving of both modalities. To advance this intelligence to the next stage, it is crucial for models to autonomously generate free-form interleaved text-image sequences. In this paper, we introduce ILLUME-X, an advanced unified multimodal paradigm that enables high-quality, free-form interleaved text-image generation by improving multimodal data efficiency and stabilizing the multimodal training process. ILLUME-X comprises three key components: (i) an expanded training data pipeline optimized for interleaved text-image generation, (ii) a progressive training strategy with self-adaptive objectives for free-length multimodal token sequences, and (iii) an objective and comprehensive evaluation method ILScore for interleaved text-image sequences. Notably, our ILLUME-X outperforms previous unified models across multiple interleaved text-image generation tasks like style transfer, image decomposition and storytelling.
Chinese Translation
能够生成文本和图像的生成性人工智能模型的进步标志着多模态智能领域的重要一步，尤其是在涉及两种模态交错的任务中。为了将这种智能提升到下一个阶段，模型必须能够自主生成自由形式的交错文本-图像序列。本文介绍了 ILLUME-X，一种先进的统一多模态范式，通过提高多模态数据效率和稳定多模态训练过程，实现高质量的自由形式交错文本-图像生成。ILLUME-X 包含三个关键组件：(i) 针对交错文本-图像生成优化的扩展训练数据管道，(ii) 具有自适应目标的渐进训练策略，适用于自由长度的多模态令牌序列，以及 (iii) 用于交错文本-图像序列的客观和全面评估方法 ILScore。值得注意的是，我们的 ILLUME-X 在多个交错文本-图像生成任务（如风格迁移、图像分解和讲故事）中超越了之前的统一模型。
cs.CV / 198 / 2606.30058
Emergence of a Shared Canonical Object Frame from In-the-Wild Videos
从野外视频中出现共享的典范物体框架
Abstract
Comparing object orientations and positions across different instances requires their poses to be expressed in a shared canonical frame. Establishing such frames has traditionally required manual annotation, creating a scaling bottleneck that limits category and instance diversity. We show that a shared canonical frame can instead emerge from self-supervised training on object-centric videos captured in the wild, using only noisy camera poses from Structure-from-Motion. Our key idea is to route all training sequences through a shared geometric bottleneck: a coarse canonical mesh that carries no category-specific detail. By learning dense correspondences from image pixels to this mesh, and estimating per-sequence alignments from noisy SfM geometry, a common canonical frame emerges from multi-view consistency and the semantic priors of the feature extractor, without any canonical pose labels or category conditioning. Trained in a self-supervised manner on 160,000 in-the-wild object videos, our method achieves competitive accuracy on category-level pose estimation benchmarks compared to methods that rely on canonical pose supervision. The code and checkpoint is available on https://github.com/Fischer-Tom/Emergent-Canonical-Frame/.
Chinese Translation
比较不同实例之间的物体方向和位置需要将它们的姿态表达在一个共享的典范框架中。建立这样的框架传统上需要手动标注，这造成了一个规模瓶颈，限制了类别和实例的多样性。我们展示了一个共享的典范框架可以通过在野外捕获的以物体为中心的视频上进行自监督训练而出现，仅使用来自运动重建（Structure-from-Motion）的噪声相机姿态。我们的关键思想是将所有训练序列通过一个共享的几何瓶颈：一个不包含类别特定细节的粗略典范网格。通过学习从图像像素到该网格的密集对应关系，并根据噪声SfM几何估计每个序列的对齐，来自多视角一致性和特征提取器的语义先验的共同典范框架得以出现，而无需任何典范姿态标签或类别条件。我们的算法在160,000个野外物体视频上以自监督的方式进行训练，在类别级姿态估计基准上与依赖于典范姿态监督的方法相比，达到了具有竞争力的准确性。代码和检查点可在 https://github.com/Fischer-Tom/Emergent-Canonical-Frame/ 获取。
cs.CV / 199 / 2606.30082
Clinical Risk-Aware Multi-Level Grading for Coronary Artery Stenosis through Curved Feature Reconstruction
基于曲线特征重建的冠状动脉狭窄临床风险感知多级分级
Abstract
Developing a multi-level grading model for coronary artery stenosis holds great clinical significance for the diagnosis of coronary artery disease. However, designing an effective multi-level deep learning algorithm faces significant challenges. Specifically, utilizing CCTA or 3D SCPR images alone presents inherent shortcomings: CCTA images are difficult to analyze due to the tortuous paths of blood vessels, while 3D SCPR images are prone to abnormal distortions that hinder accurate grading. Furthermore, different stenosis grades are associated with varying clinical risks, and incorporating this association into the algorithm is non-trivial. To address the former problems, we propose the Curved Feature Reconstruction (CFR) module, which uses vessel curves as prior and employs a point-by-point correspondence strategy to precisely align and fuse features from both 3D SCPR and CCTA images. Meanwhile, a Clinical Risk-Aware (CR) Loss is employed to introduce clinical risk relevance into the network training so that the algorithm can better align with the clinical diagnosis. The experimental results on a in-house dataset reveal that our approach significantly outperforms other methods, and several ablation studies also demonstrate the effectiveness of our proposed designs.
Chinese Translation
开发一种针对冠状动脉狭窄的多级分级模型对冠状动脉疾病的诊断具有重要的临床意义。然而，设计一个有效的多级深度学习算法面临着重大挑战。具体而言，仅利用CCTA或3D SCPR图像存在固有的缺陷：由于血管路径的曲折，CCTA图像难以分析，而3D SCPR图像则容易出现异常失真，妨碍准确分级。此外，不同的狭窄等级与不同的临床风险相关，将这种关联纳入算法并非易事。为了解决前述问题，我们提出了曲线特征重建（Curved Feature Reconstruction, CFR）模块，该模块利用血管曲线作为先验，并采用逐点对应策略精确对齐和融合来自3D SCPR和CCTA图像的特征。同时，采用临床风险感知（Clinical Risk-Aware, CR）损失函数将临床风险相关性引入网络训练，使算法能够更好地与临床诊断对齐。在我们内部数据集上的实验结果表明，我们的方法显著优于其他方法，多个消融实验也证明了我们所提出设计的有效性。
cs.CV / 200 / 2606.30084
One Forward Beats Two: InnerZoom for Accurate and Efficient GUI Grounding
一前胜于二：用于准确高效GUI定位的InnerZoom
Abstract
MLLM-based GUI grounding methods commonly formulate target localization as autoregressive coordinate generation, enabling models to leverage the strong instruction-following and semantic understanding capabilities of MLLMs. However, this formulation requires the model to retain region-level target evidence while decoding coordinate tokens with the spatial precision demanded by GUI clicking. Our diagnostic analysis reveals that target-region awareness emerges in intermediate decoder layers but is neither retained nor translated into the final coordinate prediction. Existing ZoomIn-style methods address this issue through an external crop-and-rerun pass, which improves localization but increases end-to-end latency and computational cost. To retain the accuracy benefits of two-pass zooming without this extra cost, we propose InnerZoom, a single-forward framework for cross-layer evidence bridging. InnerZoom transforms target-related cues from the original forward pass into a compact cross-layer evidence state, then preserves, refines, and reinjects this state throughout later decoding layers to guide coordinate prediction. Extensive experimental results suggest that InnerZoom-4B achieves state-of-the-art performance on all six GUI grounding benchmarks, obtaining 64.7 on OSWorld-G, 40.2 on UI-Vision, 73.1 on OSWorld-GR, and 87.6 on MMBench-GUI, surpassing the previous best results by 4.1, 3.2, 2.9, and 2.3 points, respectively. Under a controlled 4B setting, InnerZoom improves the same SFT+RL baseline by 5.3 points on average and outperforms two-pass ZoomIn by 1.3 points on average, while reducing end-to-end latency by up to 31.8% and TFLOPs by about 29%. Code and models will be publicly available.
Chinese Translation
基于MLLM的GUI定位方法通常将目标定位形式化为自回归坐标生成，使模型能够利用MLLM强大的指令跟随和语义理解能力。然而，这种形式化要求模型在解码坐标标记时保持区域级目标证据，同时满足GUI点击所需的空间精度。我们的诊断分析表明，目标区域意识在中间解码器层中出现，但未能保留或转化为最终的坐标预测。现有的ZoomIn风格方法通过外部裁剪和重运行的方式解决了这一问题，虽然提高了定位精度，但增加了端到端延迟和计算成本。为了在不增加额外成本的情况下保留双通道缩放的准确性优势，我们提出了InnerZoom，这是一种用于跨层证据桥接的单次前向框架。InnerZoom将来自原始前向传递的目标相关线索转化为紧凑的跨层证据状态，然后在后续解码层中保留、精炼并重新注入该状态，以指导坐标预测。大量实验结果表明，InnerZoom-4B在所有六个GUI定位基准测试中达到了最先进的性能，在OSWorld-G上获得64.7，在UI-Vision上获得40.2，在OSWorld-GR上获得73.1，在MMBench-GUI上获得87.6，分别超越了之前最佳结果4.1、3.2、2.9和2.3分。在受控的4B设置下，InnerZoom平均提高了同一SFT+RL基线5.3分，并且平均超越了双通道ZoomIn 1.3分，同时将端到端延迟降低了最多31.8%，TFLOPs减少约29%。代码和模型将公开发布。
cs.CV / 201 / 2606.30097
CylindTrack: Depth-Aware Cylindrical Motion Modeling for Panoramic Multi-Object Tracking
CylindTrack：面向全景多目标跟踪的深度感知圆柱运动建模
Abstract
Multi-Object Tracking (MOT) is a core capability for embodied perception, and panoramic cameras are attractive for embodied systems because their 360{\deg} field of view reduces blind spots and keeps surrounding targets observable for longer durations. However, panoramic MOT is not a straightforward extension of perspective MOT. In equirectangular panoramic videos, the horizontal image domain is periodic rather than Euclidean, which breaks planar motion assumptions and makes IoU-based association unreliable near the 0{\deg}/360{\deg} seam. Meanwhile, large-FoV scenes often contain more objects, stronger scale variation, and more frequent interactions, making online association particularly sensitive to unstable frame-wise depth cues. To address these issues, we propose CylindTrack, a depth-aware cylindrical tracking-by-detection framework for panoramic MOT. CylindTrack first introduces Depth-Temporal Trajectory Modeling (DTM), which promotes instance depth from an isolated frame-wise cue to a temporally filtered trajectory-level state. To improve the reliability of depth observations, we further develop Spherical Spatio-Temporal Consistency Learning (SSTC), which combines a Temporal Mixer and Spherical Geometry-aware Attention to enhance temporal coherence and panoramic geometric alignment in depth-aware representations. Finally, we design a Topology-Aware Cylindrical Motion Model (TCMM) that lifts horizontal motion into a continuous angular state space and performs seam-consistent motion prediction and association in the periodic panoramic domain. By jointly modeling trajectory-level depth consistency and panoramic topology, CylindTrack improves identity preservation and trajectory continuity in challenging panoramic scenes. The source code will be released at https://github.com/warriordby/CylindTrack.
Chinese Translation
多目标跟踪（MOT）是具身感知的核心能力，而全景相机因其360度视场减少盲区并使周围目标在更长时间内可观察而受到具身系统的青睐。然而，全景MOT并不是透视MOT的简单扩展。在等矩形全景视频中，水平图像域是周期性的而非欧几里得的，这打破了平面运动假设，并使得在0度/360度接缝附近基于IoU的关联变得不可靠。同时，大视场场景通常包含更多的物体、更强的尺度变化和更频繁的交互，使得在线关联对不稳定的逐帧深度线索特别敏感。为了解决这些问题，我们提出了CylindTrack，一种面向全景MOT的深度感知圆柱跟踪检测框架。CylindTrack首先引入了深度-时间轨迹建模（Depth-Temporal Trajectory Modeling, DTM），将实例深度从孤立的逐帧线索提升到时间过滤的轨迹级状态。为了提高深度观测的可靠性，我们进一步开发了球面时空一致性学习（Spherical Spatio-Temporal Consistency Learning, SSTC），该方法结合了时间混合器和球面几何感知注意力，以增强深度感知表示中的时间一致性和全景几何对齐。最后，我们设计了一种拓扑感知圆柱运动模型（Topology-Aware Cylindrical Motion Model, TCMM），将水平运动提升到连续的角度状态空间，并在周期性全景域中执行接缝一致的运动预测和关联。通过联合建模轨迹级深度一致性和全景拓扑，CylindTrack提高了在具有挑战性的全景场景中的身份保持和轨迹连续性。源代码将发布在 https://github.com/warriordby/CylindTrack。
cs.CV / 202 / 2606.30108
LETT-NeXt: A Lightweight RECIST-Guided Model for 3D CT Lesion Segmentation
LETT-NeXt：一种轻量级的基于RECIST的3D CT病灶分割模型
Abstract
RECIST diameter measurements are widely used for tumor response assessment, but they provide only a limited 2D description of lesion extent. We present LETT-NeXt, a lightweight RECIST-guided model that predicts 3D lesion masks from CT volumes and RECIST markers for the CVPR 2026 Foundation Models for Pan-cancer Segmentation in CT Images competition. LETT-NeXt extracts a RECIST-centered regional crop, encodes the RECIST line and endpoints as two prompt channels, and concatenates them with the CT input. A compact MedNeXt-v2 encoder--decoder predicts the lesion mask, followed by prompt-aware component selection and adaptive AutoZoom inference. On the public validation set, LETT-NeXt achieved a Dice Similarity Coefficient (DSC) of 79.4 $\pm$ 10.1 and a Normalized Surface Dice (NSD) of 72.3 $\pm$ 16.2. On the hidden test set, it achieved a DSC of 73.9 and an NSD of 67.3, corresponding to a challenge score of 70.6\%. On the public validation mirror, LETT-NeXt completed CPU inference in 6.9 $\pm$ 3.0 s per case with a peak memory use of 3.6 GB. Code is available at github.com/Ahus-AIM/lett-next.
Chinese Translation
RECIST直径测量广泛用于肿瘤反应评估，但仅提供病灶范围的有限2D描述。我们提出了LETT-NeXt，这是一种轻量级的基于RECIST的模型，能够从CT体积和RECIST标记中预测3D病灶掩膜，旨在参加CVPR 2026全癌种CT图像分割基础模型竞赛。LETT-NeXt提取了一个以RECIST为中心的区域裁剪，将RECIST线及其端点编码为两个提示通道，并将其与CT输入连接。一个紧凑的MedNeXt-v2编码器-解码器预测病灶掩膜，随后进行提示感知的组件选择和自适应AutoZoom推理。在公共验证集上，LETT-NeXt达到了79.4 ± 10.1的Dice相似系数（DSC）和72.3 ± 16.2的归一化表面Dice（NSD）。在隐藏测试集上，其DSC为73.9，NSD为67.3，对应的挑战得分为70.6%。在公共验证镜像上，LETT-NeXt在每个案例中完成CPU推理的时间为6.9 ± 3.0秒，峰值内存使用为3.6 GB。代码可在github.com/Ahus-AIM/lett-next获取。
cs.CV / 203 / 2606.30124
SciIR: A Large-scale Training Dataset and Benchmark for Scientific Image Reasoning Generation
SciIR：一个大规模科学图像推理生成训练数据集和基准
Abstract
While Text-to-Image (T2I) models have shown remarkable success in generating photorealistic visual content, they still struggle with the rigorous semantic alignment and logical reasoning required for scientific imagery. Inspired by Peirce's Semiotic Triad, we introduce Scientific Image Reasoning (SciIR), a comprehensive resource for training and evaluation of scientific image generation. We formalize scientific reasoning into three core dimensions: Entity Structure (Icon), Scientific Process (Index), and Scientific Law (Symbol). Specifically, to overcome the scarcity of training data in scientific image generation, we elaborately create SciIR-82k, a large-scale dataset containing over 80,000 high-quality scientific image-text pairs from cutting-edge publications. The dataset is hierarchically organized according to the semiotic dimensions and incorporates a Scientific Reasoning Chain-of-Thought (Sci-RCoT) to explicitly model underlying visual logic. For evaluation, we propose SciIR-Bench, which aligns with these three semiotic levels and employs an Atomic Checklist to convert the outcome-oriented scientific accuracy into process-oriented, verifiable, fine-grained questions. Our extensive experiments reveal significant deficiencies in current models' scientific reasoning capabilities. Furthermore, by fine-tuning on the SciIR-82k dataset, we developed the Qwen-Image-SciIR model, which achieves a substantial improvement on the SciIR-Bench, increasing the final score from 35\% to 43\%, laying a solid foundation for future advances in scientific image generation.
Chinese Translation
尽管文本到图像（T2I）模型在生成逼真的视觉内容方面取得了显著成功，但它们在科学图像所需的严格语义对齐和逻辑推理方面仍然面临挑战。受到皮尔士符号三元组的启发，我们引入了科学图像推理（SciIR），这是一个用于科学图像生成训练和评估的综合资源。我们将科学推理形式化为三个核心维度：实体结构（Icon）、科学过程（Index）和科学法则（Symbol）。具体而言，为了克服科学图像生成中训练数据稀缺的问题，我们精心创建了SciIR-82k，这是一个包含来自前沿出版物的超过80,000对高质量科学图像-文本对的大规模数据集。该数据集根据符号维度进行分层组织，并结合了科学推理思维链（Sci-RCoT），以明确建模潜在的视觉逻辑。在评估方面，我们提出了SciIR-Bench，它与这三个符号层次对齐，并采用原子检查表将结果导向的科学准确性转化为过程导向的、可验证的、细致的问题。我们的广泛实验揭示了当前模型在科学推理能力方面的显著不足。此外，通过在SciIR-82k数据集上进行微调，我们开发了Qwen-Image-SciIR模型，该模型在SciIR-Bench上取得了显著提升，将最终得分从35\%提高到43\,为未来科学图像生成的进展奠定了坚实基础。
cs.CV / 204 / 2606.30131
Hyper-Network Neural Functional Maps for Unsupervised Robust 3D Shape Matching
用于无监督鲁棒3D形状匹配的超网络神经功能映射
Abstract
Functional maps are the cornerstone of recent non-rigid 3D shape matching methods due to their efficiency and performance. However, existing methods struggle with challenging scenarios, such as partiality, topological noise, and raw point clouds. A primary bottleneck is that significant intrinsic distortion prevents truncated spectral bases from being accurately aligned via linear transformations (i.e., functional maps). To address this, we introduce a hyper-network that predicts non-linear neural functional maps (NFM), learned in an unsupervised manner, to better align spectral bases. Specifically, we model the NFM as an MLP with skip-connection to refine standard FM and employ a hyper-network to predict its weights, conditioned on standard FM. Our framework is trained using a novel unsupervised spectral alignment loss. Experiments demonstrate that our approach can be seamlessly integrated into state-of-the-art unsupervised deep functional map pipelines, substantially improving matching accuracy in demanding scenarios.
Chinese Translation
功能映射是近年来非刚性3D形状匹配方法的基石，因其高效性和性能而受到广泛关注。然而，现有方法在处理部分性、拓扑噪声和原始点云等挑战性场景时表现不佳。一个主要瓶颈是显著的内在失真使得截断谱基无法通过线性变换（即功能映射）准确对齐。为了解决这一问题，我们引入了一种超网络，预测以无监督方式学习的非线性神经功能映射（NFM），以更好地对齐谱基。具体而言，我们将NFM建模为具有跳跃连接的多层感知机（MLP），以改进标准功能映射，并利用超网络预测其权重，基于标准功能映射进行条件化。我们的框架使用一种新颖的无监督谱对齐损失进行训练。实验表明，我们的方法可以无缝集成到最先进的无监督深度功能映射流程中，在要求苛刻的场景中显著提高匹配精度。
cs.CV / 205 / 2606.30147
T2LDM++: A Self-Conditioned Representation Guided Diffusion Model for Realistic Text-to-LiDAR Scene Generation
T2LDM++：一种自条件表示引导的扩散模型用于真实文本到LiDAR场景生成
Abstract
Recent progress in Text-to-Image generation benefits from large-scale Text-Image pairs. However, the scarcity of Text-LiDAR pairs often causes over-smoothed scenes and limited controllability. In this paper, we rethink the limitations of Text-LiDAR generation task, focusing on alleviating insufficient training priors and constructing controllable Text-LiDAR data. We propose a \textbf{T}ext-\textbf{to}-\textbf{L}iDAR \textbf{D}iffusion \textbf{M}odel for LiDAR scene generation, T2LDM++, with a Self-Conditioned Representation Guidance (SCRG). Specifically, to alleviate object over-smoothing, SCRG employs a Guidance Network (GN) to provide reconstruction-based soft supervision to the Denoising Network (DN). This enables DN to learn geometry-aware representations through reconstruction guidance, leading to more accurate denoising in DDPMs. Meanwhile, through analysis and design, SCRG exhibits more effective and lightweight, while decoupled in inference, avoiding computational overhead. Furthermore, we construct two high-quality Text-LiDAR benchmarks ($>$100K samples) using a generalized strategy of geometric annotations, along with a controllability metric. Moreover, a directional position prior is designed to mitigate street distortion, further improving scene fidelity. Additionally, T2LDM++ supports multiple conditions, including (Semantic, Box, BEV, Camera)-to-LiDAR, Sparse-to-Dense, and Dense-to-Sparse generation, by learning a control encoder via frozen DN. With effective prior modeling and high-quality Text-LiDAR benchmarks, T2LDM++ can generate realistic LiDAR scenes with rich geometric details in unconditional and conditional settings.
Chinese Translation
近期文本到图像生成的进展得益于大规模的文本-图像对。然而，文本-LiDAR对的稀缺常常导致场景过于平滑且可控性有限。本文重新审视了文本-LiDAR生成任务的局限性，重点缓解训练先验不足的问题，并构建可控的文本-LiDAR数据。我们提出了一种用于LiDAR场景生成的文本到LiDAR扩散模型T2LDM++，并引入了自条件表示引导（Self-Conditioned Representation Guidance，SCRG）。具体而言，为了缓解物体过平滑的问题，SCRG采用了引导网络（Guidance Network，GN）为去噪网络（Denoising Network，DN）提供基于重建的软监督。这使得DN能够通过重建引导学习几何感知表示，从而在DDPMs中实现更准确的去噪。同时，通过分析和设计，SCRG表现出更有效且轻量化的特性，并在推理时解耦，避免了计算开销。此外，我们使用一种广义的几何标注策略构建了两个高质量的文本-LiDAR基准（超过10万样本），并引入了可控性指标。此外，设计了一种方向性位置先验以减轻街道畸变，进一步提高场景的真实感。此外，T2LDM++支持多种条件生成，包括（语义、边界框、鸟瞰视图、相机）到LiDAR、稀疏到密集和密集到稀疏的生成，通过冻结的DN学习控制编码器。凭借有效的先验建模和高质量的文本-LiDAR基准，T2LDM++能够在无条件和有条件的设置中生成具有丰富几何细节的真实LiDAR场景。
cs.CV / 206 / 2606.30159
A Dual-domain Refinement Network with FBP-based Jacobian Learning for Sparse-view Dual-Energy CT Material Decomposition
基于FBP的雅可比学习的双域精炼网络用于稀疏视角双能CT材料分解
Abstract
Dual-energy CT (DECT) exploits attenuation differences across different X-ray spectra to provide richer material information and has been widely used in medical imaging. While sparse-view acquisition can lower radiation exposure, it makes DECT material decomposition even more challenging, as the problem is nonlinear and ill-posed. Existing deep unrolling approaches generally do not explicitly incorporate the Jacobian operator induced by the nonlinear forward model, and their sparsity priors are still mainly built on conventional convolutions, which are insufficient for modeling global structural information. This study addresses the challenge of DECT multi-material decomposition in sparse-view settings by representing it as a sparse-regularized nonlinear least-squares problem. To solve it, we propose an iterative dual-domain refinement network (DECT-DRNet). In each iteration, the filtered back-projection (FBP)-based Jacobian approximation module is used first to generate an intermediate material decomposition result. Here, we characterize the forward process of material decomposition using a nonlinear operator, and then construct a theoretically grounded learnable approximation of the adjoint Jacobian operator by integrating the FBP algorithm with a U-Net into the backward process. In addition, to address the limitation of existing deep learning-based decomposition methods in globally suppressing noise and artifacts, we introduce a learnable sparse dual domain regularization term that incorporates Fourier convolutional residual blocks. This refinement block combines geometric feature extraction in the image domain with noise suppression in the frequency domain, allowing the model to capture both global and local features while maintaining structural details. DECT-DRNet demonstrates its ability to achieve more accurate material decomposition under sparse-view conditions.
Chinese Translation
双能CT（DECT）利用不同X射线光谱之间的衰减差异提供更丰富的材料信息，已广泛应用于医学成像。虽然稀疏视角采集可以降低辐射暴露，但它使得DECT材料分解变得更加具有挑战性，因为该问题是非线性且病态的。现有的深度展开方法通常没有明确地结合由非线性前向模型引起的雅可比算子，并且它们的稀疏先验仍主要基于传统卷积，这不足以建模全局结构信息。本研究通过将DECT多材料分解在稀疏视角下表示为一个稀疏正则化的非线性最小二乘问题来应对这一挑战。为了解决这个问题，我们提出了一种迭代双域精炼网络（DECT-DRNet）。在每次迭代中，首先使用基于过滤反投影（FBP）的雅可比近似模块生成一个中间材料分解结果。在这里，我们使用非线性算子表征材料分解的前向过程，然后通过将FBP算法与U-Net结合，构建一个理论基础的可学习的伴随雅可比算子近似，以整合到反向过程。此外，为了解决现有基于深度学习的分解方法在全局抑制噪声和伪影方面的局限性，我们引入了一个可学习的稀疏双域正则化项，该项结合了傅里叶卷积残差块。这个精炼模块将图像域中的几何特征提取与频率域中的噪声抑制相结合，使模型能够在保持结构细节的同时捕捉全局和局部特征。DECT-DRNet展示了其在稀疏视角条件下实现更准确材料分解的能力。
cs.CV / 207 / 2606.30168
Latent Noise Mask for Reducing Visual Redundancy in Multimodal Large Language Models
用于减少多模态大型语言模型视觉冗余的潜在噪声掩码
Abstract
Multimodal large language models (MLLMs) often fail in fine-grained visual reasoning, as question-relevant visual cues are diluted by dense and redundant image tokens. Recent multimodal reasoning methods usually extend chain-of-thought from language models into visual or latent spaces, seeking to add intermediate reasoning states while overlooking the negative impact of redundant visual tokens. We propose LatEnt Noise maSk (Lens), a question-conditioned visual evidence purification framework that empowers MLLMs to reason with cleaner visual cues in latent space. Lens introduces a lightweight Lens Evidence Token (LET) to score which visual tokens support the current question and preserve them during decoding. Guided by the LET scores, it injects adaptive latent noise into low-relevance tokens, softly suppressing distractors without changing the model backbone or token sequence. With only one temporary learnable control token and a lightweight noise generator, Lens adds minimal overhead while improving the base MLLM by 2.4-6.4 points on most VQA datasets and by 4.1-6.4 points on grounding tasks. These results show that multimodal reasoning can benefit more directly from cleaner question-relevant visual evidence than from simply extending the reasoning trace.
Chinese Translation
多模态大型语言模型（MLLMs）在细粒度视觉推理中常常表现不佳，因为与问题相关的视觉线索被密集且冗余的图像标记稀释。最近的多模态推理方法通常将语言模型的思维链扩展到视觉或潜在空间，试图添加中间推理状态，但忽视了冗余视觉标记的负面影响。我们提出了潜在噪声掩码（LatEnt Noise maSk，简称Lens），这是一个基于问题的视觉证据净化框架，使MLLMs能够在潜在空间中使用更清晰的视觉线索进行推理。Lens引入了一种轻量级的Lens证据标记（LET），用于评分哪些视觉标记支持当前问题，并在解码过程中保留它们。在LET评分的指导下，它向低相关标记注入自适应潜在噪声，柔和地抑制干扰项，而不改变模型骨干或标记序列。仅需一个临时可学习的控制标记和一个轻量级噪声生成器，Lens在大多数视觉问答（VQA）数据集上将基准MLLM的性能提高了2.4-6.4分，在基础任务上提高了4.1-6.4分。这些结果表明，多模态推理可以更直接地从更清晰的与问题相关的视觉证据中受益，而不是仅仅通过扩展推理轨迹。
cs.CV / 208 / 2606.30179
HiRes: A Hierarchical Cascaded Method for Resistor Value Identification
HiRes：一种用于电阻值识别的层次级联方法
Abstract
Accurate identification of resistor values from unconstrained images remains a challenging computer vision task due to variations in lighting, orientation, scale, and background complexity. This paper presents HiRes, a hierarchical cascaded pipeline for end-to-end resistor value identification directly from full-frame images. The approach combines object detection (YOLOv8n), semantic segmentation (UNet++ with EfficientNet-B2), and structured geometric decoding via projection along the resistor axis. To improve robustness, we incorporate geometric filtering, gap-preserving band separation, and validation against the E24 resistor series. Experiments across diverse real-world images show that HiRes achieves a detection mAP50 of 0.9906, a segmentation mIoU of 0.8444, and an end-to-end identification accuracy of 85.8% (95% CI: 78.0-91.9%), outperforming the publicly available classical baseline, CVResist, which fails to generalize beyond controlled conditions. In addition, our architecture outperforms state-of-the-art MLLMs on our challenging test set, offering a lower cost, high efficiency, and an interpretable alternative method. These results demonstrate the effectiveness of integrating learned visual representations with structured reasoning for robust resistor interpretation. Code and dataset are available at https://github.com/HiRes491/HiRes.
Chinese Translation
从无约束图像中准确识别电阻值仍然是一个具有挑战性的计算机视觉任务，这主要由于光照、方向、尺度和背景复杂性的变化。本文提出了HiRes，一种用于从全帧图像中端到端电阻值识别的层次级联管道。该方法结合了目标检测（YOLOv8n）、语义分割（UNet++与EfficientNet-B2）以及通过沿电阻轴的投影进行的结构化几何解码。为了提高鲁棒性，我们引入了几何过滤、保留间隙的带分离以及对E24电阻系列的验证。在多样的真实世界图像上的实验表明，HiRes实现了0.9906的检测mAP50、0.8444的分割mIoU以及85.8%的端到端识别准确率（95%置信区间：78.0-91.9%），超越了公开可用的经典基线CVResist，该基线在控制条件之外无法泛化。此外，我们的架构在我们具有挑战性的测试集上超越了最先进的多模态学习模型（MLLMs），提供了一种成本更低、高效且可解释的替代方法。这些结果证明了将学习到的视觉表征与结构化推理相结合在电阻解释中的有效性。代码和数据集可在https://github.com/HiRes491/HiRes获取。
cs.CV / 209 / 2606.30183
DrivenMorph: Bridging Attention Mechanism and Variational Image Registration via Difference Modeling
DrivenMorph：通过差异建模桥接注意力机制与变分图像配准
Abstract
Medical image registration benefits significantly from deep learning, yet existing approaches often lack physical explainability and fine-grained deformation control. Motivated by Demons algorithms, we propose a novel DrivenMorph framework that bridges attention mechanisms with variational image registration by incorporating difference modeling as a physically inspired inductive bias. The resulting driving force, computed from local differences in the latent feature space, provides explicit semantic guidance throughout the registration process. It directly drives the registration process through a neural Demons layer that simulates force-displacement interactions to generate smooth and anatomically consistent deformation. Unlike previous methods, our approach not only integrates traditional registration principles with popular deep networks, providing an explainable and efficient solution for learning-based medical image registration, but also separates difference modeling from deformation, improving modularity and explainability. Extensive experiments on multiple 3D brain MRI datasets demonstrate superior performance over state of-the-art learning-based and optimization-based methods. Furthermore, visualizations and statistical analyses confirm that the learned driving force aligns closely with actual deformation patterns, supporting its explanatory value.
Chinese Translation
医学图像配准在深度学习的推动下受益匪浅，但现有方法往往缺乏物理可解释性和细粒度的形变控制。受 Demons 算法的启发，我们提出了一种新颖的 DrivenMorph 框架，通过将差异建模作为一种物理启发的归纳偏置，将注意力机制与变分图像配准相结合。由潜在特征空间中的局部差异计算得出的驱动力，在整个配准过程中提供了明确的语义指导。它通过一个神经 Demons 层直接驱动配准过程，模拟力-位移交互以生成平滑且解剖一致的形变。与之前的方法不同，我们的方法不仅将传统的配准原理与流行的深度网络相结合，为基于学习的医学图像配准提供了一个可解释且高效的解决方案，还将差异建模与形变分离，提高了模块化和可解释性。在多个 3D 脑 MRI 数据集上的广泛实验表明，我们的方法在性能上优于最先进的基于学习和基于优化的方法。此外，视觉化和统计分析确认所学习的驱动力与实际形变模式紧密对齐，支持其解释价值。
cs.CV / 210 / 2606.30190
Few-Shot Domain Incremental Learning via Continual Vision-Language Consolidation
通过持续的视觉-语言整合实现少样本领域增量学习
Abstract
Existing domain-incremental learning (DIL) strategies call for massive amounts of data to adapt to new domains and suffer from the overfitting problem in the case of data scarcity. This paper puts forward a relatively uncharted problem, namely, few-shot domain incremental learning (FSDIL), taking into account the problem of extreme data shortages in the realm of DIL. A novel algorithm, namely Continual Vision-Language Consolidation (CVLC), is proposed to address the FSDIL problem, where the key idea lies in the concept of latent space reservation in the base domain coupled with dual coalescent projection (DCP) as a parameter-efficient fine-tuning method. First, the vision prototype is calibrated while multiple templates and synonyms are generated via LLMs to induce the language prototype. The vision and language prototypes are fused. Adaptation to never-ending arrivals of new domains is done by the DCP technique, fine-tuned in such a way to prepare the model to unseen domains via latent-space reservations committed in the base domain. CVLC is structured under shared and domain-specific components to combine general knowledge and domain-specific details. The advantage of our approach is demonstrated through a range of benchmark problems and comparisons with prior arts, in which CVLC outperforms them by up to a 16% gap. Our codes are shared publicly in https://github.com/Naeem-Paeedeh/CVLC .
Chinese Translation
现有的领域增量学习（DIL）策略需要大量数据以适应新领域，并且在数据稀缺的情况下容易出现过拟合问题。本文提出了一个相对未被探索的问题，即少样本领域增量学习（FSDIL），考虑到DIL领域中极端数据短缺的问题。为了解决FSDIL问题，提出了一种新颖的算法，即持续视觉-语言整合（CVLC），其关键思想在于在基础领域中保留潜在空间，并结合双重融合投影（DCP）作为参数高效的微调方法。首先，通过大型语言模型（LLMs）校准视觉原型，并生成多个模板和同义词以诱导语言原型。视觉和语言原型被融合。通过DCP技术适应不断到来的新领域，微调方式旨在通过在基础领域中进行的潜在空间保留来为模型准备未见过的领域。CVLC在共享和领域特定组件下构建，以结合一般知识和领域特定细节。我们的研究通过一系列基准问题和与先前工作的比较展示了该方法的优势，其中CVLC的表现比其他方法高出多达16%。我们的代码已公开分享在 https://github.com/Naeem-Paeedeh/CVLC 。
cs.CV / 211 / 2606.30201
SHOVIR: A Benchmark for Evaluating Vision Shortcut Learning in Radiology Report Generation
SHOVIR：评估放射学报告生成中视觉捷径学习的基准
Abstract
Current evaluation protocols for Vision-Language Models (VLMs) in Radiology Report Generation (RRG) rely on report-level metrics that measure lexical overlap or aggregate clinical correctness. However, such metrics do not test whether individual diagnostic statements stem from the actual pathological evidence visible in the image. This allows models to achieve competitive scores by exploiting learned priors or spurious correlations, a failure mode we refer to as vision shortcut. We introduce SHOVIR, a benchmark for evaluating vision shortcut behavior in RRG. SHOVIR extends two spatially annotated chest X-ray datasets, MIMIC-CXR and PadChest-GR, with per-box CheXpert labels, and defines image-level and disease-level occlusion experiments that contrast baseline performance on clean images against localized, region-specific perturbations. Comparing predictions across these conditions isolates two failure modes at the disease-class level: direct shortcuts, where a finding persists after its visual evidence is removed, and contextual shortcuts, where detection degrades once co-occurring pathologies are occluded despite the target region remaining intact. Benchmarking eight state-of-the-art VLMs, we find that shortcut behavior varies substantially across architectures and datasets. Models achieving the highest baseline report quality do not necessarily rank highest in spatial grounding, revealing that clinically fluent generation can coexist with shallow reliance on visual evidence. These findings expose a blind spot in current RRG evaluation and motivate region-aware assessment protocols.
Chinese Translation
当前针对放射学报告生成（RRG）中的视觉语言模型（VLMs）的评估协议依赖于报告级别的指标，这些指标测量词汇重叠或聚合临床正确性。然而，这些指标并未测试个别诊断陈述是否源于图像中可见的实际病理证据。这使得模型能够通过利用学习到的先验或虚假相关性来获得竞争性分数，这种失效模式我们称之为视觉捷径。我们引入了SHOVIR，这是一个用于评估RRG中视觉捷径行为的基准。SHOVIR扩展了两个空间标注的胸部X光数据集，MIMIC-CXR和PadChest-GR，增加了每个框的CheXpert标签，并定义了图像级和疾病级的遮挡实验，以对比干净图像的基线性能与局部、区域特定的扰动。通过比较这些条件下的预测，我们在疾病类别级别上隔离了两种失效模式：直接捷径，即在移除其视觉证据后发现仍然存在；以及上下文捷径，即在共存病理被遮挡时，尽管目标区域保持完整，检测能力却下降。对八种最先进的VLM进行基准测试，我们发现捷径行为在不同架构和数据集之间存在显著差异。获得最高基线报告质量的模型并不一定在空间定位上排名最高，这表明临床流畅的生成可以与对视觉证据的浅层依赖共存。这些发现揭示了当前RRG评估中的盲点，并激励了区域感知评估协议的提出。
cs.CV / 212 / 2606.30209
A Multi Center Breast FNAC Whole-Slide Cytology Dataset for AI-Assisted Patch-Wise Classification Using C1 to C5 Reporting Categories
用于基于 C1 到 C5 报告类别的 AI 辅助逐块分类的多中心乳腺细针穿刺细胞学数据集
Jain, Garima, Patil, Abhijeet, Jain, Surabhi, Pati, Sanghamitra, Sethi, Amit, Mathur, Sandeep, Verma, Pulkit, Halduniya, Nishi, Kashyap, Jatin, Kumar, Sharat, Kharb, Simmi, Singh, Sunita, Khuraijam, Sucheta Devi, Khuraijam, Sushma, Konjengbam, Ratan, Kumar, Arvind, Tirkey, Deepali, Banerjee, Saurav, Kalhan, Shivani, Gupta, Rakesh Kumar, Solanki, Ranjana, Hemranjani, Deepika, Singh, Shashank Nath, Handa, Uma, Kaur, Manveen, Malathi, B. G., P., Yogender, Kumari, Niraj, Gupta, Shruti, Nair, Indu R., C., Vidya, Das, Basumitra, Komanapalli, Sunil Kumar, Karle, Ravindra, Kulkarni, Tanaya, Raphael, Vandana, Dey, Biswajit, Gaikwad, Vaishali, Adhav, Nilam
Abstract
We present a multi center breast fine needle aspiration cytology (FNAC) dataset designed for patch wise classification using C1 to C5 reporting labels. The prospective dataset includes 321 patients and 470 whole-slide images (WSIs) collected from participating tertiary medical centers in India between May 2023 and March 2026. Slides were stained using Papanicolaou (190 WSIs) or MayGrunwald Giemsa (280 WSIs), scanned on a Hamamatsu NanoZoomer S360 at 40X magnification and 0.25 microns per pixel, and stored directly in NDPI format. Across the 470 WSIs, 446 WSIs contain annotated patch regions, yielding 7,398 PNG image patches with expert-verified C1 to C5 labels. The release includes NDPI WSIs, WSI-level GeoJSON annotation files, extracted patch images, deidentified metadata, a data dictionary, a validation summary, a manifest linking WSIs to Zenodo records, and code for dataset inspection and reuse. The complete dataset is approximately 950 GB and is available through Zenodo.
Chinese Translation
我们提出了一个多中心乳腺细针穿刺细胞学（FNAC）数据集，旨在使用 C1 到 C5 报告标签进行逐块分类。该前瞻性数据集包含321名患者和470幅来自印度参与的三级医疗中心的全切片图像（WSIs），收集时间为2023年5月至2026年3月。切片使用巴氏染色法（Papanicolaou，190幅WSIs）或梅氏-格伦瓦尔德-吉姆萨染色法（MayGrunwald Giemsa，280幅WSIs）进行染色，使用Hamamatsu NanoZoomer S360在40倍放大率和每像素0.25微米的分辨率下扫描，并直接以NDPI格式存储。在这470幅WSIs中，446幅包含标注的块区域，生成7398个经过专家验证的C1到C5标签的PNG图像块。发布内容包括NDPI WSIs、WSI级别的GeoJSON注释文件、提取的图像块、去标识化的元数据、数据字典、验证摘要、将WSIs与Zenodo记录链接的清单，以及用于数据集检查和重用的代码。完整数据集约为950 GB，并可通过Zenodo获取。
cs.CV / 213 / 2606.30215
Efficient RGB-T Object Detection via Sparse Cross-Modality Fusion
通过稀疏跨模态融合实现高效的RGB-T目标检测
Abstract
RGB-T detectors leverage the complementary strengths of visible and thermal infrared modalities, achieving robust performance under challenging conditions. Many of them resort to heavy dual backbones and exhaustive cross-modality fusion across the entire image, leading to impractically high computational costs. We observe that most image regions are smooth backgrounds (e.g., sky, ground) that can be easily handled by lightweight single-modality models. In light of this observation, we propose a sparse fusion mechanism for efficient RGB-T detection: first rapidly scanning the image to identify the proposals and then carefully examining the remaining sparse proposals via feature fusion. We propose a two-stage framework to instantiate this mechanism, which performs detection in two stages: 1) a lightweight and modality-specific detection stage that produces high-recall RoIs, and 2) a fusion-driven examination and refinement stage that filters out the false positives and refines the bounding boxes. This design enables the detector to adaptively allocate more computational resources to the potential foregrounds, improving the efficiency while ensuring detection accuracy. Extensive experiments show that our method achieves competitive performance with substantially fewer parameters and lower cost, while maintaining strong scalability to high-resolution images.
Chinese Translation
RGB-T检测器利用可见光和热红外模态的互补优势，在复杂条件下实现了稳健的性能。许多检测器依赖于重型双主干网络和对整个图像进行全面的跨模态融合，导致计算成本过高。我们观察到，大多数图像区域是平滑的背景（例如，天空、地面），这些区域可以通过轻量级的单模态模型轻松处理。基于这一观察，我们提出了一种稀疏融合机制以实现高效的RGB-T检测：首先快速扫描图像以识别提议，然后通过特征融合仔细检查剩余的稀疏提议。我们提出了一个两阶段框架来实现这一机制，该框架分两个阶段进行检测：1）一个轻量级且特定模态的检测阶段，生成高召回率的感兴趣区域（RoIs）；2）一个基于融合的检查和精炼阶段，过滤掉假阳性并优化边界框。该设计使检测器能够自适应地将更多计算资源分配给潜在前景，从而提高效率，同时确保检测准确性。大量实验表明，我们的方法在参数数量显著减少和成本降低的情况下，仍能实现具有竞争力的性能，同时保持对高分辨率图像的强大扩展性。
cs.CV / 214 / 2606.30220
From Accuracy to Visual Dependence: Auditing and Filtering Modality Collapse in Traffic VideoQA
从准确性到视觉依赖：审计和过滤交通视频问答中的模态崩溃
Abstract
High benchmark accuracy does not guarantee genuine use of visual evidence. We study this problem in traffic accident Video Question Answering (VideoQA), where correct answers should depend on scene-specific visual evidence but may instead be inferred from textual shortcuts. Through an audit of four public benchmarks, we find that several recent open-weight Vision-Language Models (VLMs) perform competitively, and sometimes better, without video input. On the MM-AU benchmark, removing video consistently improves accuracy, and adding more frames further degrades performance. To quantify visual dependence, we introduce two dataset-level diagnostics: Blind Gap, measuring above-chance text-only performance, and Visual Gain, measuring the marginal benefit of adding video. We further propose an instance-level Shortcut Score that combines text-only confidence with visual necessity signals, enabling continuous, training-free filtering of shortcut-prone questions. The resulting subsets reduce shortcut bias and improve visual grounding. Our findings reveal large differences in grounding quality across benchmarks and show that visually grounded evaluation, not just high accuracy, is essential in safety-critical VideoQA.
Chinese Translation
高基准准确性并不保证视觉证据的真实使用。我们研究了这一问题在交通事故视频问答（Video Question Answering, VideoQA）中的表现，其中正确答案应依赖于场景特定的视觉证据，但可能会通过文本捷径推导出来。通过对四个公共基准的审计，我们发现几种近期的开放权重视觉-语言模型（Vision-Language Models, VLMs）在没有视频输入的情况下表现竞争力，有时甚至更好。在 MM-AU 基准上，去除视频始终提高准确性，而添加更多帧则进一步降低性能。为了量化视觉依赖性，我们引入了两个数据集级别的诊断指标：盲区（Blind Gap），用于测量高于随机水平的文本-only 性能，以及视觉增益（Visual Gain），用于测量添加视频的边际收益。我们进一步提出了一种实例级别的捷径评分（Shortcut Score），将文本-only 置信度与视觉必要性信号结合，使得对易受捷径影响的问题进行持续的、无训练的过滤成为可能。最终得到的子集减少了捷径偏见并改善了视觉基础。我们的研究结果揭示了不同基准之间在基础质量上的巨大差异，并表明在安全关键的 VideoQA 中，视觉基础评估而不仅仅是高准确性是至关重要的。
cs.CV / 215 / 2606.30244
Semantic-Driven Scale and Spatial Selection for Efficient Cross-Modal Alignment in Referring Remote Sensing Image Segmentation
基于语义驱动的高效跨模态对齐的尺度和空间选择在遥感图像分割中的应用
Abstract
Referring Remote Sensing Image Segmentation (RRSIS) seeks to localize and segment the target object or region specified by a natural language expression in a remote sensing image. While existing RRSIS models have benefited from large-scale foundation models, they predominantly rely on full fine-tuning. These approaches are computationally intensive and may weaken the generalization ability of pre-trained models, as extensive fine-tuning on significantly smaller downstream datasets can distort the well-structured feature representations learned during large-scale pre-training. Although Parameter-Efficient Tuning (PET) offers a potential alternative, existing PET frameworks primarily focus on single-modal optimization, failing to capture the complex cross-modal dependencies required for multimodal reasoning, while simultaneously struggling to bridge the substantial domain gap between natural scenes and aerial imagery. To address these limitations, we propose a novel framework, Semantic-driven Scale and Spatial Selection for Efficient Cross-modal Alignment (S4ECA), which enables effective and efficient cross-modal interaction through parameter-efficient adaptation. Specifically, we design a dual-encoder adapter architecture. The textual adapter employs learnable queries to distill highly semantic language proxies from word-level embeddings, facilitating early grounding. Simultaneously, the visual adapter refines hierarchical feature representations through a multi-scale dense extractor, followed by a language-guided scale and spatial selection mechanism that dynamically emphasizes relevant visual contexts, ensuring precise cross-modal alignment. By updating only 2.4% of the backbone parameters, our proposed model achieves state-of-the-art performance on the RRSIS-D and RefSegRS datasets, demonstrating superior efficiency and precision in complex aerial scenarios.
Chinese Translation
遥感图像分割（RRSIS）旨在根据自然语言表达在遥感图像中定位和分割指定的目标对象或区域。尽管现有的RRSIS模型受益于大规模基础模型，但它们主要依赖于完全微调。这些方法计算密集，可能削弱预训练模型的泛化能力，因为在显著较小的下游数据集上进行广泛微调可能会扭曲在大规模预训练过程中学习到的良好结构化特征表示。虽然参数高效微调（PET）提供了一种潜在的替代方案，但现有的PET框架主要集中于单模态优化，未能捕捉多模态推理所需的复杂跨模态依赖，同时在弥合自然场景与航空影像之间的显著领域差距方面也面临挑战。为了解决这些局限性，我们提出了一种新颖的框架——基于语义驱动的高效跨模态对齐的尺度和空间选择（S4ECA），该框架通过参数高效适应实现有效的跨模态交互。具体而言，我们设计了一个双编码器适配器架构。文本适配器利用可学习的查询从词级嵌入中提取高度语义化的语言代理，促进早期的基础定位。同时，视觉适配器通过多尺度密集提取器优化层次特征表示，随后采用语言引导的尺度和空间选择机制，动态强调相关的视觉上下文，确保精确的跨模态对齐。通过仅更新2.4%的主干参数，我们提出的模型在RRSIS-D和RefSegRS数据集上实现了最先进的性能，展示了在复杂航空场景中的优越效率和精确度。
cs.CV / 216 / 2606.30248
Your Data Manifold is Secretly a Reward Model: Shell-LCC for Text-to-Video Generation
您的数据流形实际上是一个奖励模型：用于文本到视频生成的Shell-LCC
Abstract
Recent text-to-video (T2V) diffusion models rely heavily on auxiliary reward signals (e.g., via reward models or DPO) to align generated content with human aesthetics and improve realism. These signals, however, incur substantial computational overhead, require costly human annotations, and often yield limited improvement in fine-grained local details. In this paper, we argue that your data manifold is secretly a reward model. By explicitly modeling the manifold structure of high-quality Supervised Fine-Tuning (SFT) data and encouraging video latents to lie on this manifold, we derive dense, differentiable, and nearly cost-free reward signals that significantly improve video quality, particularly in mitigating low-level distortions. Our modeling builds upon Local Coordinate Coding (LCC), which captures the `skeleton' of the manifold. However, directly applying LCC suffers from mean regression, pulling latents toward the geometric mean and losing high-frequency details. We therefore extend it to Shell Local Coordinate Coding (Shell-LCC), which models the manifold `surface' as an isotropic shell to align with the true high-density region. Experiments demonstrate that our approach improves realism, enhances high-frequency details, reduces over-smoothing artifacts, and alleviates motion blur.
Chinese Translation
最近的文本到视频（T2V）扩散模型在很大程度上依赖于辅助奖励信号（例如，通过奖励模型或DPO）来使生成内容与人类美学对齐并提高真实感。然而，这些信号会带来大量的计算开销，需耗费昂贵的人力标注，并且通常在细粒度局部细节上改善有限。在本文中，我们提出您的数据流形实际上是一个奖励模型。通过明确建模高质量监督微调（SFT）数据的流形结构，并鼓励视频潜变量位于该流形上，我们推导出密集、可微分且几乎无成本的奖励信号，这显著提高了视频质量，特别是在减轻低级失真方面。我们的建模基于局部坐标编码（LCC），它捕捉了流形的“骨架”。然而，直接应用LCC会遭遇均值回归问题，使潜变量向几何均值靠拢，从而丢失高频细节。因此，我们将其扩展为Shell局部坐标编码（Shell-LCC），将流形“表面”建模为各向同性的壳，以与真实的高密度区域对齐。实验表明，我们的方法提高了真实感，增强了高频细节，减少了过度平滑伪影，并缓解了运动模糊。
cs.CV / 217 / 2606.30262
Intermediate Text Representation Guided Text-to-Image Generation for Enhancing One-and-Only Alignment
中间文本表示引导的文本到图像生成以增强唯一性对齐
Abstract
Text-to-image (T2I) diffusion models often fail to faithfully render explicit textual descriptions, instead defaulting to strongly learned visual priors due to a phenomenon referred to as concept association bias. We show that such bias is particularly strong for one-and-only (OAO) objects, entities that exist in a single canonical form, such as celestial bodies, landmarks, and artworks. The deeply ingrained visual identity for these concepts often resists modification through prompting alone. Addressing this challenge, we first identify through an information-theoretic analysis that the final text embedding discards concept-level information present in the intermediate-layer text representations, reducing the mutual information available to the subsequent denoising process. We then propose Intermediate Text Representation (IR)-guided diffusion, which injects intermediate hidden states of the text encoder into the conditioning signal during early denoising steps, recovering suppressed concepts without any additional training, optimization, or external models. To systematically evaluate the challenging task of aligning generative outputs with unusual prompts for OAO objects, we introduce OAO-AttackBench, a benchmark comprising counterfactual prompts that directly conflict with the core visual identity of OAO objects. Experiments on four benchmarks, including OAO-AttackBench, show that our method achieves up to a 19.1 percentage-point improvement in VQAScore while preserving generation fidelity and human preference. Project page: https://soyoun-won.github.io/one-and-only-ir-guidance/.
Chinese Translation
文本到图像（T2I）扩散模型常常无法忠实地呈现明确的文本描述，而是由于一种称为概念关联偏差的现象，默认使用强学习的视觉先验。我们发现，这种偏差在唯一性（OAO）对象中尤为明显，这些对象以单一的规范形式存在，例如天体、地标和艺术品。这些概念深深根植的视觉身份往往抵制仅通过提示进行的修改。为了解决这一挑战，我们首先通过信息论分析识别出最终的文本嵌入丢弃了中间层文本表示中存在的概念级信息，从而减少了后续去噪过程可用的互信息。然后，我们提出了中间文本表示（IR）引导的扩散方法，该方法在早期去噪步骤中将文本编码器的中间隐藏状态注入条件信号，从而在不需要任何额外训练、优化或外部模型的情况下恢复被抑制的概念。为了系统评估对OAO对象的生成输出与不寻常提示对齐这一具有挑战性的任务，我们引入了OAO-AttackBench，这是一个包含与OAO对象核心视觉身份直接冲突的反事实提示的基准。针对包括OAO-AttackBench在内的四个基准的实验表明，我们的方法在保持生成保真度和人类偏好的同时，VQAScore提高了最多19.1个百分点。项目页面：https://soyoun-won.github.io/one-and-only-ir-guidance/
cs.CV / 218 / 2606.30288
VisReflect: Latent Visual Reflection for Fine-Grained Perception in Long Visual Context
VisReflect：用于长视觉上下文中的细粒度感知的潜在视觉反射
Abstract
Large Vision Language Models (LVLMs) have achieved remarkable success on vision-language tasks, yet fine-grained perception over high-resolution images and long-context videos remains challenging. As the number of visual tokens increases, the visual attention sink phenomenon becomes increasingly severe, causing irrelevant tokens to absorb a disproportionate amount of attention mass. Recent approaches attempt to mitigate this issue by explicitly predicting bounding boxes or temporal spans and re-encoding the cropped visual regions. Such methods depend on unreliable numeric localization in the discrete token space and incur significant computational overhead due to additional forward passes. In this work, we propose **VisReflect**, a simple yet effective framework that improves fine-grained perception in long visual contexts through latent visual reflection. Instead of decoding intermediate predictions into discrete tokens, the model generates continuous visual reflection that represents question-relevant visual features in the latent space. These reflections selectively emphasize salient regions or frames, guiding attention towards relevant visual tokens within a single forward pass. We conduct comprehensive evaluations on challenging high-resolution image benchmarks, including BLINK, V*, and HRBench-4K/8K, as well as video understanding benchmarks such as MVBench, VideoMME, and MLVU. Our method consistently improves over strong baselines, achieving gains of 4.1% on image benchmarks and 1.8% on video benchmarks. Compared with zooming-based methods, our model achieves comparable performance while reducing inference time by roughly 44% on video understanding.
Chinese Translation
大型视觉语言模型（LVLMs）在视觉-语言任务上取得了显著成功，但在高分辨率图像和长上下文视频上的细粒度感知仍然具有挑战性。随着视觉标记数量的增加，视觉注意力沉没现象变得愈发严重，导致无关标记吸收了不成比例的注意力质量。最近的方法试图通过显式预测边界框或时间跨度并重新编码裁剪的视觉区域来缓解这一问题。然而，这些方法依赖于离散标记空间中不可靠的数值定位，并因额外的前向传播而产生显著的计算开销。在本研究中，我们提出了**VisReflect**，这是一个简单而有效的框架，通过潜在视觉反射改善长视觉上下文中的细粒度感知。该模型生成连续的视觉反射，代表在潜在空间中与问题相关的视觉特征，而不是将中间预测解码为离散标记。这些反射选择性地强调显著区域或帧，引导注意力集中在单次前向传播中的相关视觉标记上。我们在具有挑战性的高分辨率图像基准（包括BLINK、V*和HRBench-4K/8K）以及视频理解基准（如MVBench、VideoMME和MLVU）上进行了全面评估。我们的方法在强基线之上始终实现了改进，在图像基准上获得了4.1%的提升，在视频基准上获得了1.8%的提升。与基于缩放的方法相比，我们的模型在视频理解上实现了可比的性能，同时将推理时间减少了大约44%。
cs.CV / 219 / 2606.30308
The Surprising Effectiveness of Video Diffusion Models for Hand Motion Reconstruction
视频扩散模型在手部运动重建中的意外有效性
Abstract
4D hand motion reconstruction from egocentric video is bottlenecked by clear limitations of existing methods: image-based pipelines depend on a detector that fails under heavy occlusion, while video-based methods rely on temporal modules learned only from scarce hand-pose annotations, a narrow signal insufficient to model motion dynamics, occlusion reasoning, and hand-object interaction. These capabilities, however, are exactly what video generative models must implicitly acquire when trained to synthesize coherent video at internet scale. Motivated by this, we present ViDiHand, which leverages the representations of a pretrained video diffusion model to reconstruct 4D two-hand pose. We adapt it via a hand-overlay rendering objective that specializes its features for hands while preserving its world priors. A decoder then recovers metric-scale pose from the adapted features. The whole pipeline operates directly on full frames--no detector, no infiller, and no test-time optimization. On ARCTIC, HOT3D, and HOI4D, ViDiHand substantially outperforms prior methods, establishing video diffusion models as a powerful new foundation for hand motion reconstruction and a promising route to scalable in-the-wild data collection for embodied AI. Project page: https://vidihand.github.io.
Chinese Translation
从自我中心视频中进行4D手部运动重建受到现有方法明显局限的制约：基于图像的流程依赖于在严重遮挡下失效的检测器，而基于视频的方法则依赖于仅从稀缺的手部姿态注释中学习的时间模块，这一狭窄的信号不足以建模运动动态、遮挡推理和手物体交互。然而，这些能力正是视频生成模型在训练以合成互联网规模的连贯视频时必须隐式获取的。基于此动机，我们提出了ViDiHand，该模型利用预训练视频扩散模型的表示来重建4D双手姿态。我们通过手部叠加渲染目标对其进行适配，使其特征专门针对手部，同时保留其世界先验。然后，解码器从适配后的特征中恢复度量尺度的姿态。整个流程直接在完整帧上操作——无需检测器、填补器和测试时优化。在ARCTIC、HOT3D和HOI4D数据集上，ViDiHand显著超越了先前的方法，确立了视频扩散模型作为手部运动重建的新强大基础，并为具身人工智能的可扩展野外数据收集提供了有前景的途径。项目页面：https://vidihand.github.io
cs.CV / 220 / 2606.30309
A Point Cloud Transformer for Remote Monitoring and Automated Assessment of Physical Rehabilitation Exercises
用于远程监测和物理康复运动自动评估的点云变换器
Abstract
Rehabilitation exercises are essential in restoring lost physical functions of patients suffering from various diseases (e.g., Parkinson's, back pain). Carrying out these rehabilitation exercises, often prescribed by health experts, is costly, unavailable, and requires expert supervision. The availability of RGBD images and movement/position data of joints along with expert annotation of exercise data has prompted the use of automatic assessment of the quality of rehabilitation exercises, which is cost-effective and can be carried out at home. However, existing approaches do not extract relevant features, lack practical application, require expensive pre-processing, or overlook crucial features. This study proposes a transformer-based framework for point clouds to extract features and assess rehabilitation exercises by analyzing joint positions collected through RGBD data. We adapt and utilize a curve-based point-cloud feature aggregation technique to augment point-cloud information that aids model output. The transformer architecture also uses axial self-attention, recognizing important joints and their roles to assist users in performing the exercise better. The guided system outperforms existing approaches and is also practically relevant due to its small size, fast inference, and generalization on specific joints in similar exercises. We conduct our experiments on three crucial baseline datasets for rehabilitation exercises: Kimore, UI-PRMD, and IRDS.
Chinese Translation
康复运动对于恢复因各种疾病（如帕金森病、背痛）而丧失的身体功能至关重要。这些通常由健康专家开处方的康复运动，实施成本高、难以获得，并且需要专家监督。RGBD图像和关节运动/位置数据的可用性，以及对运动数据的专家注释，促使了康复运动质量的自动评估，这种方法具有成本效益，并且可以在家中进行。然而，现有的方法未能提取相关特征，缺乏实际应用，要求昂贵的预处理，或忽视了关键特征。本研究提出了一种基于变换器的点云框架，通过分析通过RGBD数据收集的关节位置来提取特征并评估康复运动。我们调整并利用了一种基于曲线的点云特征聚合技术，以增强有助于模型输出的点云信息。变换器架构还使用轴向自注意力，识别重要关节及其角色，以帮助用户更好地进行运动。该引导系统在性能上优于现有方法，并且由于其小巧的体积、快速的推理和对相似运动中特定关节的泛化，具有实际相关性。我们在三个关键的康复运动基准数据集上进行了实验：Kimore、UI-PRMD和IRDS。
cs.CV / 221 / 2606.30313
TRACE: A Concept Bottleneck Model for Longitudinal 3D Glioblastoma Response Assessment
TRACE：一种用于纵向3D胶质母细胞瘤反应评估的概念瓶颈模型
Abstract
Longitudinal glioblastoma response assessment requires comparing subtle tumor changes across MRI time points using structured clinical criteria such as RANO. However, most deep learning methods predict response labels directly from imaging features, which limits clinical inspection, verification, and correction. We introduce TRACE, a RANO 2.0-aligned concept bottleneck model for interpretable 4-class glioblastoma response classification on longitudinal 3D MRI. TRACE processes paired baseline and follow-up multimodal MRI scans with a shared 3D vision encoder, predicts clinically meaningful tumor measurements as root concepts, computes downstream RANO-derived concepts through deterministic rules, and incorporates scan interval and new-lesion information as passthrough concepts. This design frames response assessment as structured concept reasoning rather than direct image-to-label prediction. Using 5-fold patient-wise cross-validation on the LUMIERE dataset, TRACE achieves a 4-class macro F1 of 0.4769 and a binary progression-versus-non-progression macro F1 of 0.7085. It improves over a concept bottleneck baseline and remains within the range of published non-interpretable deep learning approaches. Ablation studies show that the expert RANO graph and intervention-consistency training are important for performance, while intervention experiments demonstrate that correcting concepts can improve downstream predictions. These results suggest that structured concept bottlenecks offer a transparent and clinically aligned direction for longitudinal glioblastoma response assessment, while highlighting the need for larger protocol-aligned datasets and external validation.
Chinese Translation
纵向胶质母细胞瘤反应评估需要使用结构化临床标准（如RANO）比较MRI时间点之间微妙的肿瘤变化。然而，大多数深度学习方法直接从影像特征预测反应标签，这限制了临床检查、验证和修正。我们提出了TRACE，这是一种与RANO 2.0对齐的概念瓶颈模型，用于在纵向3D MRI上进行可解释的4类胶质母细胞瘤反应分类。TRACE处理配对的基线和随访多模态MRI扫描，使用共享的3D视觉编码器，预测作为根概念的临床有意义的肿瘤测量，通过确定性规则计算下游RANO衍生概念，并将扫描间隔和新病灶信息作为透传概念纳入。这一设计将反应评估框架化为结构化概念推理，而不是直接的图像到标签预测。在LUMIERE数据集上使用5折患者交叉验证，TRACE实现了0.4769的4类宏F1值和0.7085的二元进展与非进展宏F1值。它在概念瓶颈基线之上有所提升，并保持在已发布的非可解释深度学习方法的范围内。消融研究表明，专家RANO图和干预一致性训练对性能至关重要，而干预实验表明，修正概念可以改善下游预测。这些结果表明，结构化概念瓶颈为纵向胶质母细胞瘤反应评估提供了透明且与临床对齐的方向，同时强调了对更大协议对齐数据集和外部验证的需求。
cs.CV / 222 / 2606.30314
Real-Time Underwater Image Enhancement via Frequency-Guided Dual-Path Attention
基于频率引导的双路径注意力实时水下图像增强
Abstract
Real-time underwater image enhancement (UIE) is crucial for mobile underwater photography and autonomous robotic systems, where practical deployment typically requires low latency and compact models under constrained computational resources. Recent ultra-lightweight CNNs based on structural re-parameterization meet these constraints but operate purely in the spatial domain, ignoring the frequency-sensitive nature of underwater degradation. To address this, we propose a lightweight UIE framework that integrates two key components: a Multi-Branch Reparameterizable Convolution with Fixed DCT Priors (MBRConv-DCT) that injects structured directional frequency priors during training, and a Frequency-Guided Dual-Path Attention (FGDPA) module that fuses spatial and spectral cues via a dual-path design for adaptive feature modulation. Both components are fully compatible with structural re-parameterization: the convolution branch introduces zero additional inference cost after re-parameterization, while the attention module incurs only a minimal computational overhead. Experiments show our model achieves state-of-the-art performance with only 4.23K parameters and 600+ FPS, outperforming much larger methods in both quantitative metrics and visual quality. Code is available at https://github.com/LethyZhang/FGDPA.
Chinese Translation
实时水下图像增强（UIE）对于移动水下摄影和自主机器人系统至关重要，实际部署通常要求在受限计算资源下实现低延迟和紧凑模型。最近基于结构重参数化的超轻量级卷积神经网络（CNN）满足了这些约束，但仅在空间域内操作，忽略了水下退化的频率敏感特性。为此，我们提出了一种轻量级的UIE框架，集成了两个关键组件：一个具有固定离散余弦变换（DCT）先验的多分支可重参数化卷积（MBRConv-DCT），在训练过程中注入结构化的方向频率先验，以及一个频率引导的双路径注意力（FGDPA）模块，通过双路径设计融合空间和频谱线索以实现自适应特征调制。这两个组件与结构重参数化完全兼容：卷积分支在重参数化后引入零额外推理成本，而注意力模块仅产生最小的计算开销。实验表明，我们的模型在仅有4.23K参数和600+ FPS的情况下实现了最先进的性能，在定量指标和视觉质量上均优于更大规模的方法。代码可在 https://github.com/LethyZhang/FGDPA 获取。
cs.CV / 223 / 2606.30319
BrainJanus: A Unified Model for Understanding and Generation across Brain, Vision, and Language
BrainJanus：一个理解和生成大脑、视觉与语言的统一模型
Abstract
Modeling the bidirectional correspondence between external sensory stimuli and internal neural activity has emerged as a critical frontier in neuroscience. However, existing approaches predominantly treat brain encoding and decoding as isolated tasks, relying heavily on unimodal alignment and external priors while overlooking the brain's intrinsic nature as a multimodal integration system. To address these limitations, we propose BrainJanus, the first unified brain model that integrates brain, vision, and language within a single framework. Specifically, we introduce a Unified Brain Tokenizer to quantize continuous neural dynamics into discrete tokens aligned with visual and linguistic representations in a shared Omni space. Building on this, we utilize an All-in-One autoregressive architecture that leverages next-token prediction to enable seamless any-to-any generation, which encompasses image-to-brain and text-to-brain encoding, and brain-to-image and brain-to-text decoding. Extensive experiments demonstrate that BrainJanus achieves superior performance across diverse benchmarks. Furthermore, our framework exhibits zero-shot generalization and preserves interpretable biological topography, highlighting its potential as a general-purpose brain modeling paradigm. The code is available at \href{https://github.com/HaitaoWuTJU/BrainJanus}{GitHub}.
Chinese Translation
建模外部感官刺激与内部神经活动之间的双向对应关系已成为神经科学中的一个重要前沿。然而，现有的方法主要将大脑的编码和解码视为孤立的任务，过于依赖单模态对齐和外部先验，而忽视了大脑作为多模态整合系统的内在特性。为了解决这些局限性，我们提出了BrainJanus，这是第一个在单一框架内整合大脑、视觉和语言的统一大脑模型。具体而言，我们引入了统一大脑分词器（Unified Brain Tokenizer），将连续的神经动态量化为与视觉和语言表示在共享的全方位空间中对齐的离散标记。在此基础上，我们利用一种全能自回归架构（All-in-One autoregressive architecture），通过下一个标记预测实现无缝的任意到任意生成，包括图像到大脑和文本到大脑的编码，以及大脑到图像和大脑到文本的解码。大量实验表明，BrainJanus在多种基准测试中表现优越。此外，我们的框架展现出零样本泛化能力，并保持可解释的生物拓扑，突显其作为通用大脑建模范式的潜力。代码可在 exttt{https://github.com/HaitaoWuTJU/BrainJanus} 获取。
cs.CV / 224 / 2606.30321
Optimizing Image Preparation and Compression for Face Recognition within 1024 Bytes
在1024字节内优化面部识别的图像准备和压缩
Abstract
ICAO-compliant machine readable travel documents enable automated biometric face verification. The biometric reference is stored on an RFID chip included in form of a JPEG or JPEG 2000 compressed facial image. In contrast, temporary travel documents lack of machine readability, which excludes the owner from such automated processes. This disadvantage could be solved by equipping such documents with 2D barcodes. This technology offers a resource-saving alternative to expensive RFID chips, while still offering machine readability and fast issuing processes. However, this solution introduces the challenge of storing the face images at significantly smaller storage capacities, creating the need for reducing the file size of the included facial image to a maximum of 1024 bytes. This study examines preprocessing steps and compression configurations, using JPEG, JPEG 2000, JPEG XL, JPEG AI, HEIF, AVIF, and WebP for image compression to this target size, while still preserving as much face recognition performance as possible. While the reference sample must always comply with ICAO specifications, the individual samples may or may not meet these requirements, depending on the application. This work optimizes compression steps for both of these prerequisites. It is shown that the recently standardised JPEG AI, when using optimized settings, provides the best face recognition performance, in particular when the comparison includes only images with high face image quality. AVIF and WebP also provide good results. The losses caused by the strong lossy compression are comparatively small. For the comparison of ICAO-compliant face images only, converting the images to grayscale proves to be a helpful preprocessing step, whereas for comparisons involving less suitable samples, preserving color is preferable. In addition, smoothing and resizing the images beforehand also turns out to be beneficial.
Chinese Translation
符合国际民航组织（ICAO）标准的机器可读旅行文件能够实现自动化生物特征面部验证。生物特征参考存储在包含JPEG或JPEG 2000压缩面部图像的RFID芯片中。相比之下，临时旅行文件缺乏机器可读性，这使得持有者无法参与此类自动化流程。通过为这些文件配备二维条形码，可以解决这一缺点。这项技术提供了一种节省资源的替代方案，取代昂贵的RFID芯片，同时仍然提供机器可读性和快速签发流程。然而，这一解决方案带来了在显著较小的存储容量下存储面部图像的挑战，迫使需要将包含的面部图像文件大小减少到最大1024字节。本研究考察了预处理步骤和压缩配置，使用JPEG、JPEG 2000、JPEG XL、JPEG AI、HEIF、AVIF和WebP进行图像压缩，以达到这一目标大小，同时尽可能保留面部识别性能。虽然参考样本必须始终符合ICAO规范，但具体样本可能会根据应用而符合或不符合这些要求。本工作优化了这两个前提条件的压缩步骤。研究表明，最近标准化的JPEG AI在使用优化设置时提供了最佳的面部识别性能，尤其是在比较仅包含高质量面部图像时。AVIF和WebP也提供了良好的结果。强损失压缩造成的损失相对较小。对于仅比较符合ICAO标准的面部图像，将图像转换为灰度被证明是一个有益的预处理步骤，而对于涉及不太合适样本的比较，保留颜色更为可取。此外，提前平滑和调整图像大小也被证明是有益的。
cs.CV / 225 / 2606.30332
UniGP: Taming Diffusion Transformer for Prior-Preserved Unified Generation and Perception
UniGP：驯化扩散变换器以实现优先保留的统一生成与感知
Abstract
Recent advances in diffusion models have shown impressive performance in controllable image generation and dense prediction tasks. However, existing approaches typically treat diffusion-based controllable generation and dense prediction as separate tasks, overlooking the potential benefits of jointly modeling the heterogeneous distributions. In this work, we introduce UniGP, a framework built upon MMDiT, which unifies controllable generation and dense prediction through simple joint training, without the need for complex task-specific designs or losses, while preserving the backbone's versatile priors. By learning controllable generation and prediction under different conditions, our model effectively captures the joint distribution of image-geometry pairs. UniGP is capable of versatile controllable generation, dense prediction, and joint generation. Specifically, the proposed UniGP consists of DUGP and a unified dataset training strategy. The former, following the principle of Occam's razor, uses only a copied image branch of MMDiT to model dense distributions beyond RGB, while the latter integrates heterogeneous datasets into a unified training framework to jointly model generation and perception tasks. Extensive experiments demonstrate that our unified model surpasses prior unified approaches and performs on par with specialized methods. Furthermore, we demonstrate that multi-task joint training provides complementary benefits: generative priors enrich perceptual details, while perceptual learning improves structural alignment in generation.
Chinese Translation
最近，扩散模型的进展在可控图像生成和密集预测任务中表现出色。然而，现有的方法通常将基于扩散的可控生成和密集预测视为独立任务，忽视了联合建模异构分布的潜在好处。在本研究中，我们引入了UniGP，一个基于MMDiT的框架，通过简单的联合训练统一可控生成和密集预测，无需复杂的任务特定设计或损失，同时保留了主干网络的多功能先验。通过在不同条件下学习可控生成和预测，我们的模型有效捕捉了图像-几何对的联合分布。UniGP能够实现多样化的可控生成、密集预测和联合生成。具体而言，所提出的UniGP由DUGP和统一数据集训练策略组成。前者遵循奥卡姆剃刀原则，仅使用MMDiT的复制图像分支来建模超越RGB的密集分布，而后者则将异构数据集整合到统一训练框架中，以联合建模生成和感知任务。大量实验表明，我们的统一模型超越了之前的统一方法，并与专业化方法表现相当。此外，我们还证明了多任务联合训练提供了互补的好处：生成先验丰富了感知细节，而感知学习改善了生成中的结构对齐。
cs.CV / 226 / 2606.30342
A Classifier-Agnostic Zero-Shot Adversarial Attack Detection via CLIP
一种基于 CLIP 的分类器无关零样本对抗攻击检测方法
Abstract
Adversarial attacks pose a challenge to the reliability of deep learning models, motivating effective detection methods. Existing techniques often rely on attack-specific assumptions, access to adversarial samples, or knowledge of the underlying classifier (white-box). We propose \textit{$A^4D$ (\textbf{A}ttack- and \textbf{A}rchitecture-\textbf{A}gnostic \textbf{A}dversarial \textbf{D}etector)}, a completely black-box, zero-shot adversarial attack detection framework that utilizes prompt-based similarity scores derived from CLIP. To the best of our knowledge this is the first attempt to utilize CLIP for such a task. The method is based on two key observations: (i) CLIP is sensitive even to small imperceptible non-semantic perturbations; (ii) The shift in CLIP embedding space is not arbitrary and can be used as a robust attack indicator. Experiments across multiple attacks, datasets and classifiers validate that $A^4D$ achieves SOTA detection results in the attack-agnostic and classifier-agnostic setting.
Chinese Translation
对抗攻击对深度学习模型的可靠性构成了挑战，促使有效检测方法的出现。现有技术通常依赖于特定攻击的假设、对抗样本的获取或对底层分类器的了解（白盒）。我们提出了 extit{$A^4D$ ( extbf{A}ttack- and extbf{A}rchitecture- extbf{A}gnostic extbf{A}dversarial extbf{D}etector)}，这是一种完全黑盒的零样本对抗攻击检测框架，利用基于提示的相似性评分，这些评分源自 CLIP。根据我们所知，这是首次尝试将 CLIP 应用于此类任务。该方法基于两个关键观察： (i) CLIP 对于微小的不可察觉的非语义扰动非常敏感；(ii) CLIP 嵌入空间的变化并非任意，可以作为一个稳健的攻击指示器。在多个攻击、数据集和分类器上的实验验证了 $A^4D$ 在攻击无关和分类器无关设置中实现了最先进的检测结果。
cs.CV / 227 / 2606.30344
Early Cue Precision Shapes Visual Shortcut Learning in Controlled Cue-Manipulation Benchmarks
早期线索精度塑造了受控线索操控基准中的视觉捷径学习
Abstract
Visual classifiers can achieve high matched-distribution accuracy while relying on low-level cues that fail under conflict or suppression. We test whether this failure is shaped by early cue precision: the reliability with which a low-level cue predicts the label during early learning or downstream probe fitting. Across synthetic shape-texture tasks, sequential digit training, a 10-class frozen-representation audit, and a CIFAR-10 natural-image-based texture-overlay benchmark, we manipulate object-texture match probability and evaluate matched-ID accuracy, conflict accuracy, texture-choice rate, and suppression behavior. Degraded-but-predictive input does not substitute for cue decorrelation. In 10-class digit probes, conflict accuracy drops from 0.589 under chance-like cue precision to 0.005 under target-perfect texture. In CIFAR-10 frozen probes, conflict accuracy drops from 0.569 to 0.114, while texture choice rises from 0.049 to 0.855; this ordering persists across texture-overlay strengths alpha in {0.15,0.25,0.35,0.50}. End-to-end CIFAR-10 training shows that low early cue precision improves pre-target conflict behavior, but shortcut-rich fine-tuning can rapidly overwrite this benefit. Cue decorrelation must therefore be maintained during downstream adaptation rather than treated as a one-time inoculation.
Chinese Translation
视觉分类器可以在依赖于在冲突或抑制下失效的低级线索时实现高匹配分布准确性。我们测试这种失败是否受到早期线索精度的影响：即在早期学习或下游探测拟合过程中，低级线索预测标签的可靠性。在合成形状-纹理任务、顺序数字训练、10类冻结表示审计和基于CIFAR-10自然图像的纹理叠加基准中，我们操控物体-纹理匹配概率，并评估匹配ID准确性、冲突准确性、纹理选择率和抑制行为。降级但具有预测能力的输入并不能替代线索去相关。在10类数字探测中，冲突准确性从0.589在类似随机的线索精度下降到0.005在目标完美纹理下。在CIFAR-10冻结探测中，冲突准确性从0.569下降到0.114，而纹理选择率从0.049上升到0.855；这种排序在纹理叠加强度α为{0.15,0.25,0.35,0.50}时持续存在。端到端的CIFAR-10训练表明，低早期线索精度改善了目标前冲突行为，但富含捷径的微调可以迅速覆盖这一好处。因此，线索去相关必须在下游适应过程中保持，而不是被视为一次性的免疫。
cs.CV / 228 / 2606.30347
FFAvatar: Feed-Forward 4D Head Avatar Reconstruction from Sparse Portrait Images
FFAvatar：基于前馈的稀疏肖像图像快速构建高质量可动画的4D头部头像
Abstract
We present FFAvatar, a Transformer-based 3D Gaussian framework for fast construction of high-quality and animatable 4D head avatars from one or more reference portrait images. Unlike existing feed-forward approaches that require a fixed number of input views, FFAvatar supports incremental reconstruction, progressively refining the avatar representation as additional reference images become available. At the core of our method is an alternating attention mechanism that disentangles identity appearance from expression and viewpoint variations, enabling the reconstruction of a canonical 3D appearance that remains consistent across poses and facial expressions. To balance visual fidelity and computational efficiency, we introduce a sparse-to-dense learning paradigm. Coarse appearance features are first learned using sparse primitives anchored to the FLAME vertex level and are subsequently densified in the UV domain to capture fine-grained geometric and texture details. We further propose a plug-and-play motion refinement module that enables subject-specific dynamic personalization by modeling residual motion beyond parametric deformation. Extensive experiments demonstrate that FFAvatar efficiently produces high-fidelity and controllable 4D head avatars, achieving superior flexibility, driving efficiency, and identity-consistent rendering across diverse expressions and viewpoints.
Chinese Translation
我们提出了FFAvatar，一种基于Transformer的3D高斯框架，用于从一张或多张参考肖像图像快速构建高质量且可动画的4D头部头像。与现有的需要固定输入视图数量的前馈方法不同，FFAvatar支持增量重建，随着额外参考图像的可用，逐步优化头像表示。我们方法的核心是交替注意机制，它将身份外观与表情和视角变化解耦，从而实现一致的标准3D外观重建，适用于不同的姿势和面部表情。为了平衡视觉保真度和计算效率，我们引入了一种稀疏到密集的学习范式。首先，使用锚定在FLAME顶点级别的稀疏原语学习粗略外观特征，随后在UV域中进行密集化，以捕捉细粒度的几何和纹理细节。我们进一步提出了一种即插即用的运动优化模块，通过建模超出参数变形的残余运动，实现特定对象的动态个性化。大量实验表明，FFAvatar高效地生成高保真且可控的4D头部头像，在不同表情和视角下实现了卓越的灵活性、驱动效率和身份一致的渲染。
cs.CV / 229 / 2606.30352
FastPano3D: Feed-Forward Indoor Panoramic 3D Reconstruction from a Single Image
FastPano3D：基于单幅图像的前馈室内全景3D重建
Abstract
Recent advances in 3D scene reconstruction have highlighted the intricate trade-offs among rendering quality, inference efficiency, and data dependency. To address the challenge of rapidly reconstructing detailed 3D indoor scenes from minimal input, we introduce FastPano3D, an end-to-end framework that directly generates renderable 3D Gaussian representations from a single panoramic image. Unlike perspective-based methods, panoramic images inherently suffer from equirectangular projection distortions and spatially non-uniform feature distributions, making direct feed-forward Gaussian generation particularly challenging. In contrast to existing Gaussian Splatting based methods that rely on multi-view supervision or per-scene optimization, FastPano3D employs a lightweight feature encoder, adaptive Gaussian sampling, and a point-cloud-guided refinement strategy to achieve efficient and accurate scene generation without any test-time optimization. Our approach reconstructs high-fidelity 3D scenes within seconds, achieving up to 156 times faster inference than prior state-of-the-art methods such as Pano2Room, while using only half the parameters. Extensive experiments demonstrate that FastPano3D delivers rendering quality comparable to NeRF- and 3DGS-based reconstructions, establishing a new benchmark for rapid, single-view 3D scene inference.
Chinese Translation
近年来，3D场景重建的进展突显了渲染质量、推理效率和数据依赖性之间的复杂权衡。为了解决从最少输入快速重建详细3D室内场景的挑战，我们提出了FastPano3D，这是一种端到端框架，能够直接从单幅全景图像生成可渲染的3D高斯表示。与基于透视的方法不同，全景图像固有地受到等矩形投影失真的影响，并且特征分布在空间上不均匀，这使得直接前馈高斯生成特别具有挑战性。与现有依赖多视角监督或每场景优化的高斯点云方法相比，FastPano3D采用轻量级特征编码器、自适应高斯采样和点云引导的精细化策略，实现了高效且准确的场景生成，而无需任何测试时优化。我们的方法在几秒钟内重建出高保真度的3D场景，其推理速度比之前的最先进方法如Pano2Room快达156倍，同时只使用了一半的参数。大量实验表明，FastPano3D提供的渲染质量可与基于NeRF和3DGS的重建相媲美，为快速的单视图3D场景推理建立了新的基准。
cs.CV / 230 / 2606.30355
Residual-Guided Expert Specialization for Incomplete Multimodal Learning
基于残差引导的专家专业化在不完整多模态学习中的应用
Abstract
As real-world prediction systems often face missing modalities at inference, incomplete multimodal learning (IML) remains a practical challenge. While prior methods aim to learn representations robust to missing inputs, representations from incomplete modalities inevitably deviate from their full-modality counterparts due to missing evidence. To explicitly leverage these deviations, we propose MARS (Missingness-Aware Residual-guided Specialization), a mixture-of-experts framework that guides expert specialization based on how representations are reshaped by missingness. By contrasting task representations derived from incomplete inputs with their complete counterparts during training, we derive a privileged residual signal that captures this representational gap. The residual signal guides a residual router to assign samples to experts specialized for the corresponding deviation patterns. In parallel, a feature router learns to imitate this routing behavior using only incomplete inputs, enabling deployment without access to full modalities. To mitigate this train-test router gap, we develop a discrepancy-aware noise regularization that adaptively perturbs the residual router's decisions when the feature router deviates, enhancing expert robustness under imperfect imitation. Experiments on multimodal classification (CASIA-SURF, CREMA-D, UPMC Food-101) and segmentation (MCubeS) under missing scenarios show that MARS consistently surpasses baselines while remaining efficient and extensible to diverse backbones and tasks.
Chinese Translation
由于现实世界的预测系统在推理时常常面临缺失模态，不完整多模态学习（IML）仍然是一个实际挑战。虽然先前的方法旨在学习对缺失输入具有鲁棒性的表示，但由于缺失证据，不完整模态的表示不可避免地与其全模态对应物偏离。为了明确利用这些偏差，我们提出了MARS（缺失感知残差引导专业化），这是一种混合专家框架，基于缺失性如何重塑表示来引导专家专业化。通过在训练期间对比由不完整输入派生的任务表示与其完整对应物，我们获得了一种特权残差信号，该信号捕捉了这种表示差距。残差信号引导残差路由器将样本分配给针对相应偏差模式专业化的专家。同时，特征路由器学习仅使用不完整输入来模仿这种路由行为，从而在没有访问完整模态的情况下进行部署。为了减轻训练与测试路由器之间的差距，我们开发了一种感知差异的噪声正则化，当特征路由器偏离时，自适应地扰动残差路由器的决策，从而增强专家在不完美模仿下的鲁棒性。在缺失场景下对多模态分类（CASIA-SURF、CREMA-D、UPMC Food-101）和分割（MCubeS）的实验表明，MARS始终超越基线，同时在效率和对多样化骨干网络及任务的扩展性方面保持良好表现。
cs.CV / 231 / 2606.30365
CouCE: A Unified Causal Framework for Debiased Deep Metric Learning
CouCE：一个统一的因果框架用于去偏深度度量学习
Abstract
Deep Metric Learning (DML) often struggles with zero-shot generalization because standard objectives inherently capture what co-occurs rather than what causes similarity. Consequently, DML models are vulnerable to shortcut learning driven by two structurally distinct confounders: background spurious correlations (which create backdoor paths via scene context) and foreground nuisance perturbations (which inject non-semantic variations like pose or illumination). Although existing methods have proposed targeted solutions for each pathway individually, none can simultaneously address both due to their fundamentally distinct causal roles. To bridge this gap, we propose the Counterfactual Causal Embedding (CouCE), a unified causal framework that explicitly models and neutralizes both confounders. Specifically, we introduce Orthogonal Dictionary-Based Backdoor Adjustment (ODBA), which isolates spurious background patterns into a variance-gated dictionary and stably disentangles them from the learned embeddings via soft orthogonal regularization. Simultaneously, we propose Multi-Scale Randomized Causal Intervention (MSRCI) to enforce causal invariance against foreground nuisances through multi-scale Fourier amplitude randomization and a symmetric KL invariance constraint. Notably, CouCE seamlessly integrates with any proxy-based loss, incurring modest training overhead without requiring architectural modifications during inference. Extensive experiments on CUB-200-2011, Cars-196, and Stanford Online Products demonstrate that CouCE consistently achieves state-of-the-art performance, providing a principled and robust solution for debiased DML.
Chinese Translation
深度度量学习（DML）在零样本泛化方面常常面临挑战，因为标准目标本质上捕捉的是共现关系而非相似性的因果关系。因此，DML模型容易受到由两种结构上不同的混杂因素驱动的捷径学习的影响：背景虚假相关（通过场景上下文创建后门路径）和前景干扰扰动（注入非语义变化，如姿态或光照）。尽管现有方法已针对每条路径提出了有针对性的解决方案，但由于其根本不同的因果角色，尚无方法能够同时解决这两者。为了解决这一问题，我们提出了反事实因果嵌入（CouCE），这是一个统一的因果框架，明确建模并中和这两种混杂因素。具体而言，我们引入了正交字典基础的后门调整（ODBA），将虚假背景模式隔离到一个方差门控字典中，并通过软正交正则化稳定地将其与学习到的嵌入解耦。同时，我们提出了多尺度随机因果干预（MSRCI），通过多尺度傅里叶幅度随机化和对称KL不变性约束，强制前景干扰的因果不变性。值得注意的是，CouCE可以与任何基于代理的损失无缝集成，带来适度的训练开销，而在推理过程中不需要架构修改。在CUB-200-2011、Cars-196和斯坦福在线产品的广泛实验中，CouCE始终实现了最先进的性能，为去偏DML提供了一个有原则且稳健的解决方案。
cs.CV / 232 / 2606.30370
MUSE: Unlocking Timestep as Native Task Steering for One-Step Dense Prediction
MUSE：将时间步解锁作为单步密集预测的原生任务引导
Abstract
Monocular dense prediction has recently seen remarkable success by repurposing pre-trained diffusion models. This opens a promising yet challenging avenue for more efficient multi-task learning paradigm. However, existing multi-task diffusion methods often introduce parameter-heavy adapters, experts, or learnable task tokens, leading to computational redundancy. In this paper, we reveal an inherent mechanism within one-step diffusion models: the native, fixed sinusoidal timestep embedding can be repurposed as an endogenous task steering signal. Based on this discovery, we propose Multi-task Unified eStimation via timestep Embedding (MUSE), a parameter-free, single-model multi-tasking approach for dense prediction. We interpret this mechanism via Manifold Decoupling, where discrete, fixed timestep values deterministically steer the generation process towards decoupled, task-specific manifolds in the latent space. Extensive experiments across 10 datasets demonstrate that MUSE achieves highly competitive performance on both monocular depth and normal estimation, and its efficacy generalizes across U-Net and DiT architectures. Our work offers a concise and efficient path toward generalist vision models by simply unlocking the latent potential of existing generation infrastructure.
Chinese Translation
单目密集预测最近通过重新利用预训练的扩散模型取得了显著成功。这为更高效的多任务学习范式开辟了一个有前景但具有挑战性的途径。然而，现有的多任务扩散方法通常引入了参数繁重的适配器、专家或可学习的任务标记，导致计算冗余。在本文中，我们揭示了一步扩散模型中的一个内在机制：原生的、固定的正弦时间步嵌入可以被重新利用作为内生的任务引导信号。基于这一发现，我们提出了通过时间步嵌入的多任务统一估计（Multi-task Unified eStimation via timestep Embedding，MUSE），这是一种无参数的单模型多任务密集预测方法。我们通过流形解耦（Manifold Decoupling）来解释这一机制，其中离散的、固定的时间步值确定性地引导生成过程朝向潜在空间中解耦的、任务特定的流形。针对10个数据集的广泛实验表明，MUSE在单目深度和法线估计上均实现了高度竞争的性能，其有效性在U-Net和DiT架构中具有普遍适用性。我们的工作通过简单地解锁现有生成基础设施的潜在能力，为通用视觉模型提供了一条简洁而高效的路径。
cs.CV / 233 / 2606.30374
Set-Inclusive Uncertainty Modeling for Robust Brain Tumor Segmentation
基于集合包含的不确定性建模用于稳健的脑肿瘤分割
Abstract
Multimodal MRI is essential for accurate brain tumor segmentation. However, acquiring all modalities at inference is often challenging in practice, which causes intrinsic uncertainty due to unavoidable information loss. Without modeling this uncertainty, existing methods encode incomplete evidence into deterministic representations that appear plausible but lack reliability. In this regime, we propose a probabilistic representation framework that models representations as Gaussian distributions, where their mean captures task information and their variance measures uncertainty from missing evidence. To make variance reflect information deficiency, we regularize the mean from each partial configuration toward its full-modality counterpart, while scaling the variance with the discrepancy between their aligned means. We further introduce a set-inclusive strategy that exploits the hierarchical structure of modality subsets and enforces an ordering constraint to maintain their consistent uncertainty relationships. Extensive experiments on BraTS 2018 and 2020 demonstrate that our approach offers superior performance over baselines across diverse missing-modality scenarios. Code and model checkpoint are available at https://github.com/atlas-sky/SIUM.
Chinese Translation
多模态磁共振成像（MRI）对于准确的脑肿瘤分割至关重要。然而，在推理过程中获取所有模态在实践中往往具有挑战性，这导致由于不可避免的信息丢失而产生内在的不确定性。如果不对这种不确定性进行建模，现有方法会将不完整的证据编码为看似合理但缺乏可靠性的确定性表示。在这种情况下，我们提出了一种概率表示框架，将表示建模为高斯分布，其中均值捕捉任务信息，方差衡量因缺失证据而产生的不确定性。为了使方差反映信息缺失，我们将每个部分配置的均值正则化到其全模态对应物，同时根据其对齐均值之间的差异来缩放方差。我们进一步引入了一种集合包含策略，利用模态子集的层次结构，并施加顺序约束以保持其一致的不确定性关系。在BraTS 2018和2020的广泛实验中，我们的方法在各种缺失模态场景下表现出优于基线的方法。代码和模型检查点可在 https://github.com/atlas-sky/SIUM 获取。
cs.CV / 234 / 2606.30378
OmniCoT: A Benchmark for Global and Multi-Step Panoramic Reasoning
OmniCoT：全球与多步骤全景推理的基准测试
Abstract
Multimodal Large Language Models (MLLMs) have demonstrated promising spatial reasoning capabilities, while these abilities remain underexplored in the emerging visual modality of panoramic imagery. The full 360{\deg}$\times$180{\deg} field of view of panoramas essentially supports complex global multi-step reasoning, which is also the fundamental advantage of panoramas in applications such as embodied intelligence. However, existing panoramic benchmarks largely focus on simplistic queries that rely on local cues or single-/few-step reasoning, thereby ignoring the fundamental advantage of panoramas and failing to fully exploit their potential. To address this gap, we introduce OmniCoT, a panoramic spatial reasoning suite designed to enable MLLMs to use global evidence and perform multi-step inference across viewpoints. It includes OmniCoT-B (6.7K data) for evaluation, which measures both answer accuracy and reasoning quality, OmniCoT-Real (1K data) as a manually annotated real-world subset to quantify the Sim-to-Real gap. For training, OmniCoT-T (14.3K data) is purpose-built with structured stepwise Chain-of-Thought annotations that explicitly link intermediate reasoning steps to panoramic evidence. Based on OmniCoT-T, we introduce OmniCoT-R1 and adopt a two-stage training strategy tailored to the geometrically complex panoramic space, where Supervised Fine-tuning (SFT) anchors reasoning to panoramic evidence (e.g., bearings, proximity) and GRPO penalizes geometrically incoherent paths to consolidate global 360{\deg} spatial consistency. Through OmniCoT, we aim to recalibrate the difficulty of panoramic spatial reasoning to better align with the intrinsic capabilities of panoramic imagery, thereby fostering meaningful progress in this research area.
Chinese Translation
多模态大型语言模型（MLLMs）已展示出令人期待的空间推理能力，但在新兴的全景图像视觉模态中，这些能力仍然未得到充分探索。全景图的360{ ext{°}}$ imes$180{ ext{°}}的完整视场本质上支持复杂的全球多步骤推理，这也是全景图在具身智能等应用中的基本优势。然而，现有的全景基准测试主要集中在依赖局部线索或单步/少步推理的简单查询上，从而忽视了全景图的基本优势，未能充分挖掘其潜力。为了解决这一问题，我们引入了OmniCoT，一个旨在使MLLMs能够利用全球证据并在不同视点之间进行多步骤推理的全景空间推理套件。它包括用于评估的OmniCoT-B（6.7K数据），该数据集测量答案的准确性和推理质量，以及作为手动注释的真实世界子集的OmniCoT-Real（1K数据），用于量化Sim-to-Real差距。为了训练，OmniCoT-T（14.3K数据）专门构建了结构化的逐步思维链注释，明确将中间推理步骤与全景证据联系起来。基于OmniCoT-T，我们引入了OmniCoT-R1，并采用了一种针对几何复杂全景空间的两阶段训练策略，其中监督微调（SFT）将推理锚定到全景证据（例如，方位、接近度），而GRPO则惩罚几何不一致的路径，以巩固全球360{ ext{°}}的空间一致性。通过OmniCoT，我们旨在重新校准全景空间推理的难度，以更好地与全景图像的内在能力对齐，从而推动该研究领域的有意义进展。
cs.CV / 235 / 2606.30393
SADL: What to Ignore? A Benchmark for Subject-Aware Distractor Localization
SADL：该忽略什么？一个关于主体感知干扰物定位的基准
Abstract
Photographs frequently contain \emph{visual distractors} besides foregrounds and backgrounds of the intended subject, competing for attention and weakening composition. While modern editing tools streamline object removal, identifying which objects to remove remains a mostly manual process. Existing saliency models and open-vocabulary detectors operate without subject awareness, failing to adapt to shifting user intent. Furthermore, context-agnostic removal may disrupt the scene's semantic coherence (e.g., keep the person but remove the chair they are sitting on). To address these limitations, we formalize the task of subject-aware distractor localization, which identifies distractors while retaining compositionally essential objects. This paper introduces \textsc{SADL}, the first real-world benchmark for this task, comprising 1,800 subject-aware cases across 1,000 photographs to enable systematic evaluation and facilitate future research. In total, there are 14,617 annotated candidates, including a robust set of 1,938 hard negatives to stress-test exclusion calibration. We evaluate seven proprietary and open-weight Vision-Language Models (VLMs) on a sequential pipeline of distractor classification followed by exclusion filtering, structured around five inclusion factors and three contextual exclusion rules. Our analysis reveals that VLMs are highly capable of identifying distractors, but then over-apply exclusion, which systematically suppresses true distractors at scale. By exposing this critical bottleneck, \textsc{SADL} provides a foundational diagnostic tool to advance subject-conditioned reasoning in multimodal systems.
Chinese Translation
照片中常常包含除预期主体的前景和背景之外的 extit{视觉干扰物}，这些干扰物争夺注意力并削弱构图。尽管现代编辑工具简化了物体移除的过程，但识别需要移除的物体仍然主要依赖手动操作。现有的显著性模型和开放词汇检测器在没有主体感知的情况下运作，无法适应不断变化的用户意图。此外，缺乏上下文的移除可能会破坏场景的语义连贯性（例如，保留人物但移除他们坐着的椅子）。为了解决这些局限性，我们正式定义了主体感知干扰物定位的任务，该任务在保留构图上重要的物体的同时识别干扰物。本文介绍了 extsc{SADL}，这是该任务的第一个现实世界基准，包含1,800个主体感知案例，涵盖1,000张照片，以便进行系统评估并促进未来研究。总共有14,617个标注候选对象，包括一组强大的1,938个难负例，以进行排除校准的压力测试。我们在一个顺序管道上评估了七个专有和开放权重的视觉-语言模型（VLM），该管道包括干扰物分类和随后的排除过滤，围绕五个包含因素和三个上下文排除规则进行结构化。我们的分析表明，VLM在识别干扰物方面具有很强的能力，但随后过度应用排除，这在规模上系统性地抑制了真实的干扰物。通过揭示这一关键瓶颈， extsc{SADL}提供了一个基础诊断工具，以推动多模态系统中的主体条件推理。
cs.CV / 236 / 2606.30408
SA-Homo: Scale Adaptive Homography Estimation for Scale Variation Scenarios
SA-Homo：针对尺度变化场景的尺度自适应单应性估计
Abstract
Homography estimation, as one of the fundamental problems in computer vision, remains challenged by scale variation scenarios where image pairs potentially exhibit significant scale discrepancies. Existing deep learning frameworks frequently suffer from a significant performance degradation in such cases, as they rely on limited displacement assumptions and local feature consistency that might not hold under large scale gaps. In this paper, we propose SA-Homo, a novel scale-adaptive homography estimation framework designed to achieve robust alignment across a wide range of scale discrepancy ratios. We adopt a hierarchical scale alignment strategy that transitions from the global perspective with a heavy module to a local perspective with a light module. Specifically, we introduce the Scale-aware Discrepancy Bridging Module (SDBM) for initial alignment, which utilizes a Multi-scale Linear Attention Cascade (MLAC) to capture long-range dependencies and mitigate feature inconsistencies, along with a global Cross-scale Similarity Matrix Block (CSMB) for scale robust correlation representation. Once the initial scale gap is bridged, a lightweight Iterative Homography Estimation Refinement Module (IHERM) progressively polishes the result using local correlations. To facilitate this research, we contribute the HMSA dataset, a high-resolution, multi-modal satellite benchmark specifically tailored for scale-variant challenges. Extensive experiments demonstrate that SA-Homo maintains high precision even under 8$\times$ scale discrepancies, outperforming state-of-the-art methods in both conventional scale-similar scenarios and challenging scale variation scenarios. Code and collected datasets are available at https://github.com/shangxuanx330/SA_Homo
Chinese Translation
单应性估计作为计算机视觉中的基本问题之一，在尺度变化场景中仍然面临挑战，因为图像对可能存在显著的尺度差异。现有的深度学习框架在这种情况下常常遭遇显著的性能下降，因为它们依赖于有限位移假设和可能在大尺度差距下不成立的局部特征一致性。本文提出了SA-Homo，一种新颖的尺度自适应单应性估计框架，旨在实现广泛尺度差异比下的稳健对齐。我们采用了一种分层尺度对齐策略，从具有重模块的全局视角过渡到具有轻模块的局部视角。具体而言，我们引入了尺度感知差异桥接模块（Scale-aware Discrepancy Bridging Module, SDBM）进行初步对齐，该模块利用多尺度线性注意力级联（Multi-scale Linear Attention Cascade, MLAC）捕捉长程依赖，缓解特征不一致性，并结合全局跨尺度相似性矩阵块（Cross-scale Similarity Matrix Block, CSMB）进行尺度稳健的相关性表示。一旦初步的尺度差距被桥接，轻量级的迭代单应性估计精炼模块（Iterative Homography Estimation Refinement Module, IHERM）逐步利用局部相关性对结果进行优化。为了促进这项研究，我们贡献了HMSA数据集，这是一个高分辨率、多模态卫星基准，专门针对尺度变异挑战而设计。大量实验表明，SA-Homo在8$ imes$尺度差异下仍能保持高精度，在常规尺度相似场景和具有挑战性的尺度变化场景中均优于最先进的方法。代码和收集的数据集可在 https://github.com/shangxuanx330/SA_Homo 获取。
cs.CV / 237 / 2606.30417
Beyond Point Estimates for Glaucoma Visual Field Forecasting with Diffusion Models
超越点估计：基于扩散模型的青光眼视野预测
Abstract
Forecasting visual fields (VFs) is critical for personalized monitoring and treatment planning in glaucoma. This is inherently uncertain due to heterogeneous disease progression and measurement variability, yet most existing methods produce single deterministic predictions that fail to represent this uncertainty. We formulate VF forecasting as a probabilistic prediction problem and the use of conditioned denoising diffusion models to generate distributions of plausible future VFs from longitudinal observations with irregular follow-up intervals. Experiments on two independent VF cohorts show that diffusion-based predictions produce well-calibrated distributions for clinically relevant VF measures. When reduced to a standard point-estimate, the proposed approach achieves state-of-the-art accuracy compared to clinical baselines and prior learning-based methods. Our results highlight the advantages of distributional modeling for VF forecasting and support a shift from point-estimate prediction toward uncertainty-aware, clinically interpretable risk assessment in glaucoma.
Chinese Translation
预测视野（VFs）对于青光眼的个性化监测和治疗规划至关重要。这一过程本质上具有不确定性，因为疾病进展存在异质性且测量结果存在变异性，但大多数现有方法仅产生单一的确定性预测，未能有效反映这种不确定性。我们将视野预测形式化为一个概率预测问题，并利用条件去噪扩散模型从具有不规则随访间隔的纵向观察数据中生成合理未来视野的分布。对两个独立视野队列的实验表明，基于扩散的预测能够为临床相关的视野指标生成良好校准的分布。当将其简化为标准点估计时，所提出的方法在与临床基线和先前基于学习的方法相比时，达到了最先进的准确性。我们的结果突显了分布建模在视野预测中的优势，并支持从点估计预测向关注不确定性、可临床解释的青光眼风险评估的转变。
cs.CV / 238 / 2606.30421
OWMDrive: Causality-Aware End-to-End Autonomous Driving via 4D Occupancy World Model
OWMDrive：基于4D占用世界模型的因果感知端到端自主驾驶
Abstract
Autonomous driving systems are steadily moving toward end-to-end paradigms to mitigate the limited adaptability of rule-based pipelines in complex traffic environments. However, most existing learning-based methods still make decisions from static representations of the current scene, without explicit future rollouts or modeling of the temporal causal dynamics in traffic interactions. This limitation often results in unstable or overly conservative planning under high-uncertainty conditions, such as occlusions and unexpected events. To overcome these challenges, we introduce OWMDrive, a generative end-to-end driving framework built upon an Occupancy World Model for multi-step 3D occupancy forecasting, which serves as a conditional prior to guide diffusion-based planning. Conditioned on both current observations and predicted future states, the planner iteratively refines trajectory candidates to generate a reinforced driving trajectory. By explicitly modeling scene evolution over future horizons, OWMDrive captures key spatiotemporal causal dependencies, which leads to more foresighted and robust trajectory generation. Extensive experiments demonstrate that OWMDrive significantly improves planning reliability and safety, especially in challenging and partially observable driving scenarios.
Chinese Translation
自主驾驶系统正稳步向端到端范式发展，以减轻基于规则的管道在复杂交通环境中的适应性限制。然而，现有的大多数基于学习的方法仍然是从当前场景的静态表示中做出决策，没有明确的未来展开或交通互动中的时间因果动态建模。这一限制通常导致在高不确定性条件下（如遮挡和意外事件）规划不稳定或过于保守。为了解决这些挑战，我们提出了OWMDrive，这是一种基于占用世界模型的生成性端到端驾驶框架，用于多步3D占用预测，作为引导基于扩散的规划的条件先验。在当前观察和预测未来状态的条件下，规划者迭代地优化轨迹候选，以生成增强的驾驶轨迹。通过明确建模未来时间范围内的场景演变，OWMDrive捕捉关键的时空因果依赖关系，从而实现更具前瞻性和稳健性的轨迹生成。大量实验表明，OWMDrive显著提高了规划的可靠性和安全性，尤其是在具有挑战性和部分可观察的驾驶场景中。
cs.CV / 239 / 2606.30436
Robust and Efficient Monocular 3D Gaussian SLAM for Kilometer-Scale Outdoor Scenes
适用于公里级户外场景的鲁棒高效单目3D高斯SLAM
Abstract
Scaling monocular 3D Gaussian Splatting (3DGS) SLAM to kilometer-level outdoor environments poses two tightly coupled challenges: fragile long-term pose tracking and excessive memory overhead during large-scale mapping. In this paper, we propose KiloGS-SLAM, a highly efficient and robust monocular 3DGS-SLAM system that jointly addresses both bottlenecks. Since high-fidelity scene reconstruction fundamentally relies on drift-free camera poses, we first introduce a motion-adaptive hybrid tracking module. This module features a condition-triggered three-tier solving pipeline. It dynamically switches between Essential matrix and PnP models to handle geometric degeneracies. An on-demand foundation model can also be activated to rescue the trajectory from catastrophic drift. To ensure the system can sustain these long trajectories without memory exhaustion, we subsequently design a lifecycle-managed Gaussian mapping strategy. By integrating probabilistic initialization with chunk-based multi-view densification and pruning, this full-pipeline optimization effectively reduces primitive redundancy while preserving high-frequency details. Together, the robust tracking guarantees the geometric foundation required for accurate mapping, while the memory-efficient lifecycle-managed mapping enables large-scale operation. Extensive experiments across three challenging outdoor datasets demonstrate that our approach achieves state-of-the-art tracking accuracy and rendering quality, successfully scaling to sequences of over 10,000 frames on a single GPU.
Chinese Translation
将单目3D高斯点云（3DGS）SLAM扩展到公里级户外环境面临两个紧密相关的挑战：脆弱的长期姿态跟踪和在大规模映射过程中过高的内存开销。本文提出了KiloGS-SLAM，一种高效且鲁棒的单目3DGS-SLAM系统，旨在共同解决这两个瓶颈。由于高保真场景重建根本依赖于无漂移的相机姿态，我们首先引入了一种运动自适应混合跟踪模块。该模块具有条件触发的三级求解管道，能够在基本矩阵和PnP模型之间动态切换，以处理几何退化问题。同时，还可以激活按需基础模型，以拯救轨迹免受灾难性漂移。为了确保系统能够在不耗尽内存的情况下维持这些长轨迹，我们随后设计了一种生命周期管理的高斯映射策略。通过将概率初始化与基于块的多视图稠密化和修剪相结合，这一全流程优化有效减少了原始数据的冗余，同时保留了高频细节。鲁棒的跟踪保证了准确映射所需的几何基础，而内存高效的生命周期管理映射则支持大规模操作。在三个具有挑战性的户外数据集上进行的广泛实验表明，我们的方法实现了最先进的跟踪精度和渲染质量，成功扩展到单个GPU上超过10,000帧的序列。
cs.CV / 240 / 2606.30458
Cross-Resolution Semantic Transfer for Robust Text-to-Image Retrieval in Low-Resolution Surveillance
低分辨率监控中稳健的文本到图像检索的跨分辨率语义转移
Abstract
Text-to-image person re-identification (TIPR) retrieves target persons using natural language descriptions. However, existing methods largely overlook resolution variance in real-world surveillance. They characterize cross-resolution TIPR through two coupled failure modes: Evidence Reliability Collapse (ERC), where degraded visual tokens become unreliable for grounding fine-grained text, and Ranking Distribution Drift (RDD), where mixed-resolution galleries distort similarity neighborhoods and destabilize retrieval rankings. To address this challenge, we propose Cross-Resolution Semantic Transfer (CRST), a CLIP-style framework with three modules: resolution-conditioned reasoning, text-guided refinement and CR-RDA. Resolution-conditioned reasoning estimates token reliability to suppress corrupted evidence. Text-guided refinement injects semantic priors to recover discriminative cues. CR-RDA transfers HR neighborhood geometry to stabilize LR ranking under mixed resolutions. Experiments on CUHK-PEDES, ICFG-PEDES, and RSTPReid show that CRST improves ultra-low-resolution Rank-1 and mAP on average by 5.7% and 5.3%, while stabilizing mixed-resolution retrieval without sacrificing high-resolution accuracy.The code will be made publicly available.
Chinese Translation
文本到图像的人物重识别（TIPR）通过自然语言描述检索目标人物。然而，现有方法在实际监控中大多忽视了分辨率的变化。它们通过两种耦合的失败模式来表征跨分辨率TIPR：证据可靠性崩溃（ERC），即退化的视觉标记变得不可靠，无法为细粒度文本提供基础；排名分布漂移（RDD），即混合分辨率的图库扭曲了相似性邻域并使检索排名不稳定。为了解决这一挑战，我们提出了跨分辨率语义转移（CRST），这是一个类似CLIP的框架，包含三个模块：分辨率条件推理、文本引导精炼和CR-RDA。分辨率条件推理评估标记的可靠性，以抑制损坏的证据。文本引导精炼注入语义先验以恢复区分性线索。CR-RDA将高分辨率邻域几何转移到低分辨率排名中，以在混合分辨率下稳定排名。在CUHK-PEDES、ICFG-PEDES和RSTPReid上的实验表明，CRST在超低分辨率下平均提高了Rank-1和mAP分别为5.7%和5.3%，同时在不牺牲高分辨率准确性的情况下稳定了混合分辨率检索。代码将公开发布。
cs.CV / 241 / 2606.30471
FR-DETR: Frequency and Recurrent Feature Refinement for Robust Object Detection under Adverse Weather
FR-DETR：在恶劣天气下进行鲁棒目标检测的频率与递归特征精炼
Abstract
Object detection under adverse weather remains challenging due to severe visual degradations and domain shifts. Existing enhancer-based approaches attempt to improve detection by cascading an enhancer with a detector, but they introduce redundant feature extraction and incur high computational cost with limited accuracy gains when paired with SOTA detectors. We propose FR-DETR, a detector-centric framework that refines features rather than images, focusing enhancement on regions of interest and leveraging frequency-domain cues. Specifically, we design (I) a Frequency Refinement Module that dynamically separates and reweights low- and high-frequency components to improve foreground-background discrimination, and (II) a Recurrent Focus Refinement Module (RFRM) that iteratively refines features using coarse predictions as guidance. Extensive experiments demonstrate that FR-DETR achieves superior detection accuracy under adverse weather while being significantly more computationally efficient than enhancer-based methods. Our implementation is available at https://github.com/ducnt1210/FR-DETR.
Chinese Translation
在恶劣天气条件下进行目标检测仍然面临挑战，因为视觉退化严重且领域转移明显。现有的基于增强器的方法试图通过将增强器与检测器级联来提高检测性能，但这引入了冗余的特征提取，并且在与最先进的检测器（SOTA）配对时，计算成本高且准确性提升有限。我们提出了FR-DETR，这是一种以检测器为中心的框架，重点对特征而非图像进行精炼，增强关注于感兴趣区域，并利用频域线索。具体而言，我们设计了（I）一个频率精炼模块（Frequency Refinement Module），该模块动态分离和重新加权低频和高频成分，以改善前景与背景的区分；（II）一个递归聚焦精炼模块（Recurrent Focus Refinement Module，RFRM），该模块使用粗略预测作为指导，迭代精炼特征。大量实验表明，FR-DETR在恶劣天气下实现了优越的检测准确性，同时在计算效率上显著优于基于增强器的方法。我们的实现代码可在 https://github.com/ducnt1210/FR-DETR 获取。
cs.CV / 242 / 2606.30476
PS-MOT: Cultivating Instance Awareness from Point Seeds for Multi-Object Tracking
PS-MOT：从点种子中培养实例意识以实现多目标跟踪
Abstract
We introduce Point-supervised Multi-Object Tracking (PS-MOT) as a cost-effective alternative to traditional bounding box supervision, shifting the focus from spatial fitting to topological center-driven representation. However, PS-MOT faces challenges, e.g., spatial ambiguity and identity drift due to the lack of explicit geometric structure and scale constraints. To address these, we propose PS-Track, a hierarchical pipeline transitioning from points to instances across data, model, and loss levels. At the data level, we introduce Temporal-Feedback Prompting (TFP) to evolve points into temporally consistent pseudo-labels using negative spatial cues and motion priors. At the model level, we design the Point-Excited Wavelet Attention (PEWA) module, which leverages semantic correlations to activate high-frequency components, ``hallucinating'' object boundaries. At the loss level, Uncertainty-Guided Gaussian Learning (UGL) models pseudo-labels as probabilistic distributions, dynamically calibrating supervision intensity. Experiments on DanceTrack, EmboTrack, SportsMOT, and JRDB demonstrate that PS-Track provides a feasible and effective point-supervised alternative across diverse tracking scenarios, establishing a new state-of-the-art for point-supervised tracking. The source code is available at https://github.com/xifen523/PS-MOT.
Chinese Translation
我们提出了点监督多目标跟踪（PS-MOT），作为传统边界框监督的成本效益替代方案，将重点从空间拟合转向拓扑中心驱动的表示。然而，PS-MOT面临挑战，例如由于缺乏明确的几何结构和尺度约束而导致的空间模糊和身份漂移。为了解决这些问题，我们提出了PS-Track，一个在数据、模型和损失层次上从点到实例的分层管道。在数据层面，我们引入了时间反馈提示（Temporal-Feedback Prompting, TFP），利用负空间线索和运动先验将点演变为时间一致的伪标签。在模型层面，我们设计了点激发小波注意力（Point-Excited Wavelet Attention, PEWA）模块，该模块利用语义相关性激活高频成分，从而“幻觉”出物体边界。在损失层面，不确定性引导的高斯学习（Uncertainty-Guided Gaussian Learning, UGL）将伪标签建模为概率分布，动态校准监督强度。在DanceTrack、EmboTrack、SportsMOT和JRDB上的实验表明，PS-Track在多样化的跟踪场景中提供了一种可行且有效的点监督替代方案，确立了点监督跟踪的新状态-of-the-art。源代码可在https://github.com/xifen523/PS-MOT获取。
cs.CV / 243 / 2606.30477
PGE-SAM: Prompt-Guided Feature Enhancement for Interactive Segmentation under Degradation
PGE-SAM：在退化条件下进行交互式分割的提示引导特征增强
Abstract
Segment Anything Model (SAM) has revolutionized promptable image segmentation with strong zero-shot generalization. However, its performance degrades substantially under real-world imaging artifacts such as noise, blur, and compression. Existing methods restore features globally without focusing on segmentation-relevant regions and neglect SAM's iterative refinement mechanism, leading to suboptimal performance in interactive settings. We propose Prompt-Guided Feature Enhancement SAM (PGE-SAM), a framework that explicitly leverages user prompts and prior mask predictions to spatially guide the feature restoration process toward regions of interest through a Prompt Guidance Generator. To recover fine-grained details lost under degradation, we introduce Multi-Scale Features Interaction to incorporate low-level encoder features, along with a Foreground Reconstruction Loss that restricts feature-level supervision to the segmentation target. Furthermore, we present DM-Seg, a benchmark for interactive segmentation on degraded medical images, spanning multiple imaging modalities with both general and modality-specific degradations at varying severity levels. Extensive experiments demonstrate that PGE-SAM achieves SOTA robustness on both medical and natural image domains across multiple degradation levels, while maintaining generalization to clean images and adding less than one-fifth of the parameters of prior methods.
Chinese Translation
分割任意模型（Segment Anything Model, SAM）以其强大的零-shot 泛化能力，彻底改变了可提示的图像分割。然而，在现实世界的成像伪影（如噪声、模糊和压缩）下，其性能显著下降。现有方法在恢复特征时往往是全局性的，未能关注与分割相关的区域，并忽视了SAM的迭代精细化机制，导致在交互式设置中表现不佳。我们提出了提示引导特征增强SAM（Prompt-Guided Feature Enhancement SAM, PGE-SAM），该框架明确利用用户提示和先前的掩膜预测，通过提示引导生成器（Prompt Guidance Generator）将特征恢复过程空间引导至感兴趣区域。为了恢复在退化过程中丢失的细粒度细节，我们引入了多尺度特征交互（Multi-Scale Features Interaction），以结合低级编码器特征，并提出了前景重建损失（Foreground Reconstruction Loss），将特征级监督限制在分割目标上。此外，我们还提出了DM-Seg，这是一个针对退化医学图像的交互式分割基准，涵盖多种成像模态，并在不同严重程度下具有一般性和特定模态的退化。大量实验表明，PGE-SAM在多个退化水平下，在医学和自然图像领域实现了最先进的鲁棒性，同时保持对干净图像的泛化，并且增加的参数量不到先前方法的五分之一。
cs.CV / 244 / 2606.30492
RBE-Flow: Recurrent Bayesian Estimation on Feature Manifolds for Cross-Modal Registration
RBE-Flow：基于特征流形的递归贝叶斯估计用于跨模态配准
Abstract
Cross-modal image registration is essential for multi-sensor perception but remains fundamentally challenging due to severe non-linear radiometric discrepancies and geometric distortions. Existing deterministic matching methods lack uncertainty awareness, struggling to navigate the resulting highly non-convex optimization landscape and frequently accumulating errors in ambiguous regions. In this paper, we propose RBE-Flow, a novel framework that reformulates dense cross-modal flow estimation as a closed-loop recurrent Bayesian estimation problem on learned feature manifolds. Diverging from standard feed-forward regression, RBE-Flow establishes a robust self-correcting mechanism by deeply coupling feature-metric non-linear optimization with probabilistic state updates. Specifically, a Recurrent Manifold Optimization (RMO) block iteratively generates flow observations and their associated uncertainties, which are then optimally assimilated into the prior state via an Uncertainty-Adaptive Probabilistic Update (UAPU) using deterministic sigma-point projection. Crucially, the resulting calibrated posterior covariance is fed back to adaptively regularize the damping of subsequent optimization steps, allowing the system to modulate its convergence based on predictive confidence. To ensure stable probabilistic training, we introduce a hybrid supervision scheme featuring a geometry-aware rectified NLL loss that structurally prevents variance collapse. Extensive experiments on challenging OSdataset, WHU-OPT-SAR, and RoadScene benchmarks demonstrate that RBE-Flow consistently achieves state-of-the-art performance, outperforming existing methods by a significant margin, particularly under strict sub-pixel criteria. Project page: https://github.com/NEU-Liuxuecong/RBE-Flow
Chinese Translation
跨模态图像配准对于多传感器感知至关重要，但由于严重的非线性辐射差异和几何失真，仍然面临根本性的挑战。现有的确定性匹配方法缺乏对不确定性的意识，难以在高度非凸的优化空间中导航，并且在模糊区域经常积累错误。本文提出了RBE-Flow，一种新颖的框架，将密集的跨模态流估计重新表述为在学习到的特征流形上进行闭环递归贝叶斯估计的问题。与标准的前馈回归不同，RBE-Flow通过将特征度量非线性优化与概率状态更新深度耦合，建立了一种稳健的自我校正机制。具体而言，递归流形优化（Recurrent Manifold Optimization, RMO）模块迭代生成流观察及其相关的不确定性，然后通过不确定性自适应概率更新（Uncertainty-Adaptive Probabilistic Update, UAPU）使用确定性σ点投影将其最优地同化到先前状态中。关键是，得到的校准后验协方差被反馈以自适应地调节后续优化步骤的阻尼，使系统能够根据预测置信度调节其收敛性。为了确保稳定的概率训练，我们引入了一种混合监督方案，采用几何感知的校正NLL损失，结构性地防止方差崩溃。在具有挑战性的OSdataset、WHU-OPT-SAR和RoadScene基准上的大量实验表明，RBE-Flow始终实现了最先进的性能，显著超越现有方法，特别是在严格的亚像素标准下。项目页面：https://github.com/NEU-Liuxuecong/RBE-Flow
cs.CV / 245 / 2606.30498
On the Faithfulness of Post-Hoc Concept Bottleneck Models
后验概念瓶颈模型的可信性研究
Abstract
Human decision-making interprets the world through high-level concepts, such as recognizing a bird by its belly color. To bridge the gap between opaque deep learning representations and human understanding, Post-Hoc Concept Bottleneck Models (post-hoc CBMs) project latent features onto interpretable concept spaces using auxiliary datasets or vision-language models. However, relying on target task accuracy as the primary measure of post-hoc CBM success obscures whether the learned concepts are semantically meaningful or merely predictive artifacts. For example, random concept projections can achieve competitive accuracy despite being semantically meaningless. In this work, we analyze the learned projections directly and identify two failure cases: First, for concept projections learned from auxiliary data, covariate shifts can lead to unfaithful concept representations for the target task. In particular, we provide an upper bound on the error introduced by this shift. Second, systematic label noise in surrogate concept labels generated by vision-language models leads to unfaithful projections. After formalizing these failure modes, we introduce novel metrics that decouple concept faithfulness from predictive accuracy. Our empirical results across real-world and synthetic benchmarks confirm that these metrics identify unfaithful behaviors that standard accuracy-based evaluation fails to detect.
Chinese Translation
人类决策通过高层次概念来解释世界，例如通过鸟类的腹部颜色来识别鸟类。为了弥补不透明的深度学习表示与人类理解之间的差距，后验概念瓶颈模型（Post-Hoc Concept Bottleneck Models，post-hoc CBMs）利用辅助数据集或视觉-语言模型将潜在特征投影到可解释的概念空间。然而，将目标任务的准确性作为后验 CBM 成功的主要衡量标准，掩盖了所学习的概念是否具有语义意义，或仅仅是预测性伪影。例如，随机的概念投影可以在语义上毫无意义的情况下仍然实现竞争性的准确性。在本研究中，我们直接分析所学习的投影，并识别出两种失败案例：首先，对于从辅助数据中学习的概念投影，协变量转移可能导致目标任务的不可信概念表示。特别地，我们提供了这种转移引入的错误的上界。其次，由视觉-语言模型生成的替代概念标签中的系统性标签噪声导致不可信的投影。在形式化这些失败模式后，我们引入了新颖的度量标准，将概念的可信性与预测准确性解耦。我们在真实世界和合成基准上的实证结果证实，这些度量能够识别标准基于准确性的评估未能检测到的不可信行为。
cs.CV / 246 / 2606.30511
High-Resolution Flood Mapping With Sentinel-1 and Sentinel-2 via Misalignment-Robust Cross-Sensor Learning and Generative Despeckling
通过抗错位跨传感器学习和生成去斑技术实现的高分辨率洪水制图
Abstract
Reliable high-resolution flood extent mapping from satellite imagery remains constrained by limited data fidelity and sensor-specific artifacts. Multispectral optical imagery is degraded by clouds, shadows, and urban confounders, while synthetic aperture radar (SAR) imagery is affected by speckle noise and sensor co-registration uncertainty. This work presents an integrated flood mapping framework that jointly addresses these limitations through curated datasets and novel learning strategies. We introduce a new Sentinel-2 (S2) and Sentinel-1 (S1) dataset covering the contiguous United States, featuring pixel-accurate 10 m water masks with emphasis on challenging weather conditions and urban environments that are underrepresented in existing benchmarks. High-quality S2 annotations are manually produced using rigorous geospatial labeling protocols and transferred to SAR imagery through weakly labeled temporally coincident acquisitions. To address SAR-specific artifacts, a shift-invariant loss function is employed to tolerate residual geolocation uncertainty between SAR imagery and optical-derived labels, and a Conditional Variational Autoencoder (CVAE) is trained on multitemporal SAR composites to suppress speckle while preserving flood-relevant spatial structure. Experiments using UNet and UNet++ architectures demonstrate strong multispectral performance (AUPRC up to 0.956) and statistically significant improvements in SAR flood mapping when using shift-invariant loss and CVAE-based despeckling compared to classical filters. These results underscore the importance of dataset fidelity, misalignment-robust training, and demonstrate the viability of generative despeckling for operational flood mapping.
Chinese Translation
从卫星影像中可靠地获取高分辨率洪水范围制图仍受到数据保真度和传感器特定伪影的限制。多光谱光学影像受到云层、阴影和城市干扰的影响，而合成孔径雷达（SAR）影像则受到斑点噪声和传感器配准不确定性的影响。本研究提出了一个综合的洪水制图框架，联合解决这些限制，通过精心策划的数据集和新颖的学习策略。我们引入了一个新的覆盖美国本土的Sentinel-2（S2）和Sentinel-1（S1）数据集，提供像素级准确的10米水体掩膜，特别强调在现有基准中代表性不足的复杂天气条件和城市环境。高质量的S2标注通过严格的地理空间标记协议手动生成，并通过弱标记的时间重合采集转移到SAR影像中。为了解决SAR特定伪影，采用了一个平移不变的损失函数，以容忍SAR影像与光学派生标签之间的残余地理定位不确定性，并在多时相SAR合成图像上训练了条件变分自编码器（CVAE），以抑制斑点噪声，同时保留与洪水相关的空间结构。使用UNet和UNet++架构的实验表明，在使用平移不变损失和基于CVAE的去斑技术时，SAR洪水制图的多光谱性能显著提升（AUPRC高达0.956），并且与经典滤波器相比，统计上显著改善。这些结果强调了数据集保真度、抗错位训练的重要性，并展示了生成去斑技术在实际洪水制图中的可行性。
cs.CV / 247 / 2606.30514
3D Scene-Adaptive Trajectory-Controllable Human Image Animation with Camera Movement
具有相机运动的3D场景自适应轨迹可控人像动画
Abstract
Human image animation, which aims to generate a video of a reference subject following a provided action sequence, has received increasing research interest. With the development of diffusion-based/flow-based video foundation models, existing animation works have began to upgrade the guidance information from 2D skeleton/pose to 3D modeling conditions. Despite achieving reasonable results, these approaches face challenges in synthesizing trajectory-controllable human motion within natural scene under changed camera views. In this work, we present a scene-adaptive human image animation framework that controls both human motion and camera trajectories within a reconstructed 3D environment for video generation. To achieve this, we first develop a ground-adaptive 3D motion retargeting approach to enable user-friendly motion trajectory control adapting to the changes of elevations of ground and orientations automatically. Then we design a viewpoint-adaptive latent fusion mechanism to inject point-cloud geometric priors through scene-visibility masking into the generative process, providing precise guidance of viewpoint changes under camera control. Experiments on two standard human image animation benchmark datasets demonstrate remarkable improvements of our method over the state of the arts in related video generation metics. Project page: https://robinhood256100.github.io/web-disp
Chinese Translation
人像动画旨在生成一个参考主体根据提供的动作序列而生成的视频，近年来受到越来越多的研究关注。随着基于扩散/流的视屏基础模型的发展，现有的动画作品已经开始将指导信息从2D骨架/姿态升级到3D建模条件。尽管取得了合理的结果，但这些方法在自然场景中合成轨迹可控的人体运动时，面对在变化的相机视角下的挑战。在本工作中，我们提出了一种场景自适应的人像动画框架，能够在重建的3D环境中控制人体运动和相机轨迹以生成视频。为此，我们首先开发了一种地面自适应的3D运动重定向方法，使得用户能够友好地控制运动轨迹，并自动适应地面高度和方向的变化。然后，我们设计了一种视点自适应的潜在融合机制，通过场景可视性掩蔽将点云几何先验注入生成过程中，为相机控制下的视点变化提供精确的指导。在两个标准人像动画基准数据集上的实验表明，我们的方法在相关视频生成指标上显著优于现有的最先进技术。项目页面：https://robinhood256100.github.io/web-disp
cs.CV / 248 / 2606.30516
HASTE: A Framework for Training-Free, Dynamic, and Steerable Compression of Pre-Trained Convolutional Neural Networks
HASTE：一种无训练、动态和可引导的预训练卷积神经网络压缩框架
Abstract
Deploying large convolutional neural networks (CNNs) on resource-constrained devices is challenging due to their high computational cost. While dynamic execution methods are promising, existing approaches for CNNs typically require specialized training or fine-tuning, limiting their effectiveness when applied to pre-trained models and requiring data access. To address this gap, we propose HASTE (Hashing for Tractable Efficiency), a plug-and-play convolution module that enables training-free, dynamic compression of large pre-trained CNNs. At inference time, HASTE uses locality-sensitive hashing to identify and merge redundant channels of latent feature maps on a patch-wise basis. This process simultaneously compresses the depth of both input features and their corresponding filters, resulting in computationally cheaper convolutions. We conduct extensive experiments on CIFAR-10 and ImageNet across a range of architectures, demonstrating a 46.2% FLOPs reduction in a ResNet34 on CIFAR-10 with only a 1.25% drop in accuracy, without any retraining. We support our claims by comprehensive ablation studies to validate our core design choices, an analysis of the method's properties and limitations, and a discussion that connects our channel merging scheme to the conceptually related task of token merging in Vision Transformers. Our results demonstrate that HASTE provides an effective solution for steerable compression of pre-trained CNNs at runtime, opening new possibilities for the deployment of efficient deep learning methods.
Chinese Translation
在资源受限的设备上部署大型卷积神经网络（CNN）面临高计算成本的挑战。尽管动态执行方法前景广阔，但现有的CNN方法通常需要专门的训练或微调，这限制了它们在预训练模型上的有效性，并且需要数据访问。为了解决这一问题，我们提出了HASTE（Hashing for Tractable Efficiency），一种即插即用的卷积模块，能够实现大型预训练CNN的无训练动态压缩。在推理时，HASTE使用局部敏感哈希来识别并合并潜在特征图的冗余通道，采用逐块的方式进行处理。这个过程同时压缩输入特征及其对应滤波器的深度，从而实现计算上更便宜的卷积。我们在CIFAR-10和ImageNet上对多种架构进行了广泛实验，结果表明在CIFAR-10上，ResNet34的FLOPs减少了46.2%，而准确率仅下降了1.25%，且无需任何再训练。我们通过全面的消融研究支持我们的主张，以验证我们的核心设计选择，并对该方法的特性和局限性进行分析，同时讨论将我们的通道合并方案与视觉变换器中概念相关的令牌合并任务联系起来。我们的结果表明，HASTE为在运行时可引导压缩预训练CNN提供了有效的解决方案，为高效深度学习方法的部署开辟了新可能。
cs.CV / 249 / 2606.30528
$\mu$Flow: Leveraging Average Images for Improving Generalisation of Deepfake Faces Detectors
μFlow：利用平均图像提升深伪面部检测器的泛化能力
Abstract
Current generative models, including GANs and diffusion models, have reached an outstanding level of photorealism, posing significant risks to privacy and security. To ensure real-world applicability, deepfake detectors must generalise effectively to unseen generators. However, most existing approaches rely on supervised training with both real and fake images, which limits their generalisation especially across generators categories (e.g. GANs vs DMs). In this work, we introduce $\mu$Flow, a one-class deepfake detector trained only on real images without relying on pseudo-deepfakes or synthetic artifacts. Our approach builds on the observation that averaging multiple images amplifies consistent generative traces, producing highly discriminative feature representations. We leverage this property by modelling the distribution of features extracted from averaged images and training a normalizing flow to align the feature space of individual images with this distribution. This alignment yields a likelihood-based criterion that separates real and fake samples while promoting strong generalisation. We evaluate $\mu$Flow on a fully out-of-distribution setting, where both real and fake datasets are unseen during training. Experimental results show that our method significantly outperforms SOTA detectors. Project page: https://opontorno.github.io/MuFlow.
Chinese Translation
当前的生成模型，包括生成对抗网络（GANs）和扩散模型，已达到卓越的照片真实感，给隐私和安全带来了重大风险。为了确保在现实世界中的适用性，深伪检测器必须有效地对未见过的生成器进行泛化。然而，大多数现有方法依赖于使用真实和虚假图像的监督训练，这限制了它们在生成器类别（例如，GANs与扩散模型）之间的泛化能力。在本研究中，我们提出了μFlow，一种仅在真实图像上训练的单类深伪检测器，而不依赖于伪深伪或合成伪影。我们的方法基于一个观察，即对多个图像进行平均可以放大一致的生成痕迹，从而产生高度可区分的特征表示。我们利用这一特性，通过对从平均图像中提取的特征建模其分布，并训练一个归一化流（normalizing flow），使单个图像的特征空间与该分布对齐。这种对齐产生了一种基于似然的标准，可以区分真实样本和虚假样本，同时促进强泛化能力。我们在完全分布外的设置下评估了μFlow，在该设置中，训练期间未见过真实和虚假数据集。实验结果表明，我们的方法显著优于当前最先进的检测器。项目页面：https://opontorno.github.io/MuFlow。
cs.CV / 250 / 2606.30534
Orca: The World is in Your Mind
Orca：世界在你的心中
Wang, Yihao, Ji, Yuheng, Cao, Mingyu, Shen, Yanqing, Xiao, Runze, Lyu, Huaihai, Xie, Senwei, Liu, Euan, Tian, Klara, Long, Tianfeng, Zhang, Yichi, Cai, Zhengliang, Chen, Ruike, Zhao, Jifan, Shi, Ruochuan, Tang, Zihan, Lyu, Jing, Tan, Wenxing, Zhang, Ningbo, Hu, Yangtao, Gao, Yuming, Chen, Xiansheng, Zhao, Junkai, Xu, Congsheng, Zhu, Boan, Wang, Ziqi, Feng, Yupu, Zhang, Qiongqiong, Zhao, Yingli, Ao, Yulong, Xie, Shaoxuan, Liu, You, Yao, Guocai, Zhang, Leiduo, Liu, Xiaodan, Zhang, Yunyan, Jiao, Yance, Yang, Xinyan, Wei, Jiaxing, Liu, Xu, Pan, Tengfei, Nie, Shaokai, Men, Chunlei, Cui, Sen, Jin, Xiaojie, Li, Hongyang, Luo, Jianlan, Mu, Yao, Wei, Yunchao, Yan, Jun, Zhao, Hang, Zheng, Xiaolong, Li, Jiaming, Lin, Yonghua, Huang, Tiejun, Wang, Zhongyuan, Wang, Pengwei
Abstract
We introduce Orca, an initial instantiation of a general world foundation model. Orca learns a unified world latent space from multimodal world signals and exposes it through multimodal readout interfaces. Rather than optimizing isolated next-token, next-frame, or next-action prediction, we are centered on Next-State-Prediction modeling, offering a unified state-transition modeling route toward understanding, predicting, and acting upon the world. Orca learns through two complementary paradigms: unconscious learning captures dense natural state transitions from continuous videos, and conscious learning models sparse meaningful state transitions by language-described events and VQA supervision. For pre-training, we construct a large-scale world-learning inventory data, including 125K hours of video data and 160M event annotations. After pre-training, Orca learns a unified world latent space. To examine whether the learned latent supports downstream, we evaluate it by three representative downstream readouts: text generation, image prediction, and embodied action generation. Orca's backbone is frozen, and only the lightweight modality-specific decoders are trainable. Experiments show the scalability of the proposed paradigm and verify that stronger world latent enables stronger downstream readouts. Orca outperforms similar-sized specialized baselines. These results show that Orca, as a general world foundation model, presents a promising approach to understanding, predicting, and acting upon the world. Finally, we discuss the current limitations, aiming to provide useful insights and inspiration for the community.
Chinese Translation
我们介绍了Orca，这是一个通用世界基础模型的初步实例。Orca从多模态世界信号中学习统一的世界潜在空间，并通过多模态读取接口将其展现出来。我们并不专注于优化孤立的下一个标记、下一个帧或下一个动作预测，而是以下一个状态预测（Next-State-Prediction）建模为中心，提供了一条统一的状态转移建模路径，以理解、预测和对世界采取行动。Orca通过两种互补的范式进行学习：无意识学习从连续视频中捕捉密集的自然状态转移，而有意识学习通过语言描述的事件和视觉问答（VQA）监督建模稀疏的有意义状态转移。为了进行预训练，我们构建了一个大规模的世界学习库存数据，包括125K小时的视频数据和160M事件注释。预训练后，Orca学习了统一的世界潜在空间。为了检验所学潜在空间是否支持下游任务，我们通过三个具有代表性的下游读取任务进行评估：文本生成、图像预测和具身动作生成。Orca的主干网络被冻结，只有轻量级的模态特定解码器是可训练的。实验表明所提范式的可扩展性，并验证了更强的世界潜在空间能够增强下游读取任务的表现。Orca在与相似规模的专门基线相比时表现更优。这些结果表明，作为一个通用世界基础模型，Orca为理解、预测和对世界采取行动提供了一种有前景的方法。最后，我们讨论了当前的局限性，旨在为社区提供有用的见解和灵感。
cs.CV / 251 / 2606.30545
StereoGS: Sparse-View 3D Gaussian Splatting via Stereo Priors
StereoGS：基于立体先验的稀疏视图3D高斯点云生成
Abstract
3D Gaussian Splatting (3DGS) has achieved remarkable success in real-time novel view synthesis, yet it suffers from severe overfitting under sparse-view settings due to insufficient geometric constraints. While recent methods introduce monocular depth priors to mitigate this, they inherently struggle with scale ambiguity and cross-view inconsistency, leading to defective geometry. In this paper, we propose StereoGS, a novel sparse-view 3DGS framework that integrates stereo priors to establish reliable binocular consistency. Unlike scale-agnostic monocular constraints, StereoGS introduces a Stereo Depth Regularization by constructing virtual stereo pairs during optimization and leveraging a foundation stereo model to enforce absolute scale and binocular-consistent structures. To further suppress overfitting and eliminate redundant primitives, we design a Gradient-Aware Opacity Decay strategy that dynamically penalizes Gaussians based on their relative opacity gradient magnitudes. Combined with a Consistency-Aware Dense Initialization using zero-shot multi-view depth estimation, StereoGS effectively anchors primitives to accurate scene surfaces. Extensive experiments on LLFF, DTU, Mip-NeRF360, and Blender datasets demonstrate that StereoGS achieves state-of-the-art performance in sparse-view settings without incurring any additional inference overhead. Project Page: https://stringerywh00.github.io/StereoGS_project_page/
Chinese Translation
3D高斯点云生成（3DGS）在实时新视图合成中取得了显著成功，但在稀疏视图设置下，由于几何约束不足，严重过拟合的问题依然存在。尽管近期方法引入了单目深度先验以缓解这一问题，但它们在尺度模糊和视图间不一致性方面固有地存在困难，导致几何结构缺陷。本文提出了StereoGS，一种新颖的稀疏视图3DGS框架，通过整合立体先验来建立可靠的双目一致性。与尺度无关的单目约束不同，StereoGS通过在优化过程中构建虚拟立体对并利用基础立体模型来强制执行绝对尺度和双目一致结构，提出了一种立体深度正则化方法。为了进一步抑制过拟合并消除冗余原语，我们设计了一种梯度感知的不透明度衰减策略，该策略根据高斯的相对不透明度梯度大小动态惩罚高斯。结合使用零样本多视图深度估计的考虑一致性的密集初始化，StereoGS有效地将原语锚定到准确的场景表面。在LLFF、DTU、Mip-NeRF360和Blender数据集上的广泛实验表明，StereoGS在稀疏视图设置中实现了最先进的性能，且没有增加任何额外的推理开销。项目页面：https://stringerywh00.github.io/StereoGS_project_page/
cs.CV / 252 / 2606.30557
EcoVideo: Entropy-Orchestrated Video Generation Paradigm in Cloud-Edge Dynamics
EcoVideo：云边动态中的熵引导视频生成范式
Abstract
DiT video generation is latency-intensive due to iterative full-frame denoising, while prior cloud-edge methods largely rely on static inter-step decoupling and cannot leverage inter-frame similarity or adapt to system dynamics. We propose EcoVideo, an entropy-orchestrated framework for dynamic inter-frame decoupling: early-stage self-attention entropy provides a training-free estimate of frame-wise information density for frame selection; a cloud large model denoises sparse high-entropy keyframes; and an edge lightweight model reconstructs the remaining frames via motion-aware interpolation with refinement for temporal stability. EcoVideo further adapts the keyframe budget and edge refinement depth to real-time bandwidth and compute availability, optimizing end-to-end latency under constraints. Experiments on representative DiT video generators show improved quality--efficiency trade-offs and up to 2.9x end-to-end speedup in low-bandwidth, compute-limited edge settings. Code is available at https://github.com/IF-LAB-PKU/EcoVideo.
Chinese Translation
DiT视频生成由于迭代全帧去噪而具有较高的延迟，而之前的云边方法主要依赖于静态的步骤间解耦，无法利用帧间相似性或适应系统动态。我们提出了EcoVideo，这是一种用于动态帧间解耦的熵引导框架：早期自注意力熵提供了帧信息密度的无训练估计以进行帧选择；云端大型模型对稀疏高熵关键帧进行去噪；边缘轻量级模型通过运动感知插值和时间稳定性优化重建其余帧。EcoVideo进一步根据实时带宽和计算可用性调整关键帧预算和边缘优化深度，在约束条件下优化端到端延迟。在代表性的DiT视频生成器上的实验表明，质量与效率的权衡有所改善，并且在低带宽、计算受限的边缘环境中实现了高达2.9倍的端到端加速。代码可在 https://github.com/IF-LAB-PKU/EcoVideo 获取。
cs.CV / 253 / 2606.30576
Beyond 2D Matching: A Unified Single-Stage Framework for Geometry-Aware Cross-View Object Geo-Localization
超越二维匹配：一个统一的单阶段几何感知跨视角物体地理定位框架
Abstract
Cross-view object geo-localization (CVOGL) aims to locate a target object from a query view (e.g., ground or drone) within a geo-tagged reference image (e.g., satellite). Existing approaches heavily rely on 2D appearance matching and are constrained by limited datasets lacking geometric metadata, diverse prompts, and standard field-of-view imagery. To address these intertwined challenges, we first introduce \dataset, a large-scale, high-fidelity building dataset comprising over 220,000 ground-satellite and drone-satellite pairs. It provides multi-modal prompts (points, boxes, masks) and camera poses to enable flexible target referring and explicit spatial modeling. Furthermore, we propose a novel single-stage Geometry-Aware Geo-localization framework (GAGeo), built upon the permutation-equivariant 3D foundation model $\pi^3$. By seamlessly integrating visual features, referring prompts, and learnable task tokens, our model adapts the inherited 3D prior to jointly predict bounding boxes, segmentation masks, and camera poses in a single forward pass. Additionally, we introduce a contrastive loss that utilizes the satellite view as a universal anchor, implicitly aligning ground and drone representations to enable zero-shot ground-to-drone localization without requiring triplet training data. Extensive experiments demonstrate that our approach significantly outperforms state-of-the-art methods, exhibiting exceptional generalization ability in unseen scenes and novel cross-view setups.
Chinese Translation
跨视角物体地理定位（CVOGL）旨在从查询视角（例如，地面或无人机）中定位目标物体于带有地理标签的参考图像（例如，卫星图像）内。现有方法严重依赖于二维外观匹配，并受到缺乏几何元数据、多样化提示和标准视场图像的有限数据集的限制。为了解决这些相互交织的挑战，我们首先介绍了 extit{dataset}，这是一个大规模、高保真度的建筑数据集，包含超过220,000对地面-卫星和无人机-卫星图像。该数据集提供多模态提示（点、框、掩码）和相机姿态，以便实现灵活的目标指代和明确的空间建模。此外，我们提出了一种新颖的单阶段几何感知地理定位框架（GAGeo），该框架基于置换等变的三维基础模型 $ extit{π³}$。通过无缝整合视觉特征、指代提示和可学习的任务标记，我们的模型在单次前向传递中适应继承的三维先验，共同预测边界框、分割掩码和相机姿态。此外，我们引入了一种对比损失，利用卫星视图作为通用锚点，隐式对齐地面和无人机表示，以实现零-shot地面到无人机的定位，而无需三元组训练数据。大量实验表明，我们的方法显著优于最先进的方法，在未见场景和新颖的跨视角设置中展现出卓越的泛化能力。
cs.CV / 254 / 2606.30577
APRIL-MedSeg: A Modular Medical Image Segmentation Toolbox Embracing Modern Paradigms
APRIL-MedSeg：一个模块化的医学图像分割工具箱，融入现代范式
Abstract
We present APRIL-MedSeg, a YAML-driven modular framework for 2D medical image segmentation. It provides a unified and extensible ecosystem that decomposes segmentation networks into reusable components. Also, the framework integrates a broad spectrum of advanced paradigms, including semi-supervised learning, domain adaptation, knowledge distillation, weakly supervised learning, and text-guided segmentation as well as foundation model support. A registry-based configuration system with inheritance enables flexible and reproducible experiment management, supporting seamless switching across models, datasets, and training strategies. In addition, the framework provides a unified interface for medical datasets, augmentation pipelines, deployment utilities and model ensembling. Overall, APRIL-MedSeg is designed as a general-purpose research and development platform that bridges algorithmic innovation and practical deployment, while also serving as a structured ecosystem for systematically organizing and reproducing advances in medical image segmentation. The code is available at https://github.com/juntaoJianggavin/APRIL-MedSeg under an Apache 2.0 license.
Chinese Translation
我们提出了APRIL-MedSeg，这是一个基于YAML驱动的二维医学图像分割模块化框架。它提供了一个统一且可扩展的生态系统，将分割网络分解为可重用的组件。此外，该框架集成了广泛的先进范式，包括半监督学习、领域适应、知识蒸馏、弱监督学习和文本引导分割，以及基础模型支持。基于注册的配置系统与继承机制使得实验管理灵活且可重复，支持在模型、数据集和训练策略之间无缝切换。此外，该框架为医学数据集、增强管道、部署工具和模型集成提供了统一接口。总体而言，APRIL-MedSeg被设计为一个通用的研究与开发平台，旨在连接算法创新与实际部署，同时也作为一个结构化生态系统，用于系统性地组织和再现医学图像分割的进展。代码可在https://github.com/juntaoJianggavin/APRIL-MedSeg上获取，采用Apache 2.0许可证。
cs.CV / 255 / 2606.30597
Learning from Reliable Latent Prompts for Visual Recognition with Missing Modalities
从可靠的潜在提示中学习以应对缺失模态的视觉识别
Abstract
Large-scale multimodal models (LMMs) have achieved superior performance in visual recognition by synergizing information across diverse, massive-scale paired modalities. In real-world scenarios, however, missing-modality inputs are ubiquitous, causing models optimized for modality-complete data to exhibit precipitous performance degradation. Existing research has introduced prompt learning to mitigate this issue, typically by generating dynamic prompts from instance-level features, regardless of whether the input modalities are complete or partially absent. However, such input-conditioned strategies are hindered by the escalating unreliability of instance-level features; as higher missing rates increase the proportion of incomplete modalities, the resulting instability in prompt learning limits the model's performance. To address this limitation, we hypothesize that learnable latent prompts themselves encapsulate stable, modality-intrinsic priors that are decoupled from corrupted inputs. Consequently, we propose a novel paradigm: Learning from Reliable Latent Prompts. Unlike prior methods, we model input-agnostic learnable prompts as stable latent anchors that enable robust guidance and effective cross-modal knowledge compensation, even under extreme missing rates (e.g., 90%). Empirical results across three benchmark datasets demonstrate that our "learn-from-latent-prompts" approach achieves state-of-the-art performance across a wide range of missing-modality scenarios. Extensive experiments further confirm the effectiveness of this paradigm in providing a robust solution to the missing-modality problem.
Chinese Translation
大规模多模态模型（LMMs）通过协同处理多样化的大规模配对模态信息，在视觉识别中取得了优异的性能。然而，在现实场景中，缺失模态输入普遍存在，导致针对完整模态数据优化的模型表现急剧下降。现有研究引入了提示学习来缓解这一问题，通常通过从实例级特征生成动态提示，无论输入模态是完整还是部分缺失。然而，这种输入条件策略受到实例级特征可靠性不断下降的限制；随着缺失率的提高，不完整模态的比例增加，导致提示学习的不稳定性，从而限制了模型的性能。为了解决这一限制，我们假设可学习的潜在提示本身封装了与损坏输入解耦的稳定模态内在先验。因此，我们提出了一种新范式：从可靠的潜在提示中学习。与之前的方法不同，我们将输入无关的可学习提示建模为稳定的潜在锚点，即使在极端缺失率（例如90%）下，也能提供稳健的指导和有效的跨模态知识补偿。在三个基准数据集上的实证结果表明，我们的“从潜在提示中学习”方法在广泛的缺失模态场景中实现了最先进的性能。大量实验进一步验证了这一范式在提供缺失模态问题的稳健解决方案方面的有效性。
cs.CV / 256 / 2606.30598
Towards in-the-wild Egocentric 3D Hand-Object Pose Estimation
面向真实场景的自我中心三维手-物体姿态估计
Abstract
Estimating accurate 3D hand-object pose from in-the-wild egocentric RGB remains challenging due to severe occlusions and ambiguous contact. Existing learning-based methods often struggle to generalise to in-the-wild scenes and are limited by the scarcity of supervision. We address these issues with two contributions. First, we introduce EPIC-Contact, an in-the-wild egocentric dataset of 2.3K clips (62.3K frames) with dense, bijective 3D hand-object contact correspondences and posed meshes. Second, we propose HOPformer, an end-to-end transformer that jointly predicts bi-manual hand and object pose in a single forward pass. A cross-attention decoder conditions object features on hand priors, producing robust pose estimation. We test HOPformer on the in-lab 3D dataset, ARCTIC, as well as our newly introduced EPIC-Contact dataset. HOPformer reaches 82.4% success rate on ARCTIC (+6.2 pts over current SOTA). On EPIC-Contact, it nearly doubles the success rate while reducing contact deviation by 75%. EPIC-Contact, HOPformer code and checkpoints are released: https://sid2697.github.io/epic-contact.
Chinese Translation
从真实场景中的自我中心RGB图像中准确估计三维手-物体姿态仍然具有挑战性，主要由于严重的遮挡和模糊的接触。现有的基于学习的方法往往难以推广到真实场景，并且受到监督数据稀缺的限制。我们通过两个贡献来解决这些问题。首先，我们引入了EPIC-Contact，这是一个包含2300个片段（62300帧）的真实场景自我中心数据集，提供密集的双向三维手-物体接触对应关系和姿态网格。其次，我们提出了HOPformer，这是一种端到端的变换器，能够在单次前向传播中联合预测双手和物体的姿态。交叉注意力解码器将物体特征与手的先验条件结合，从而产生稳健的姿态估计。我们在实验室的三维数据集ARCTIC以及我们新引入的EPIC-Contact数据集上测试了HOPformer。HOPformer在ARCTIC上达到了82.4%的成功率（比当前的最先进技术提高了6.2个百分点）。在EPIC-Contact上，它几乎将成功率翻倍，同时将接触偏差降低了75%。EPIC-Contact、HOPformer代码和检查点已发布： https://sid2697.github.io/epic-contact。
cs.CV / 257 / 2606.30599
Goku: A Million-Scale Universal Dataset and Benchmark for Instruction-Based Video Editing
Goku：一个百万规模的通用数据集和基于指令的视频编辑基准
Abstract
Existing instruction-based video editing datasets commonly focus on single-task appearance editing, failing to meet the complex creative demands of real-world scenarios. To bridge this gap, we present Goku, a large-scale dataset featuring 2 million high-quality, instruction-aligned video editing pairs, which is the first to extend task boundaries from basic appearance editing to multi-task and structural manipulations(e.g., precise control of subject movement). To tackle the data synthesis challenges inherent in these complex tasks, we design an efficient data synthesis pipeline that decomposes complex edits into controllable sub-problems and introduce a progressive filtering system for data reliability throughout the whole process. Furthermore, we explore the optimal network structures on Goku, and propose Goku-Edit. To deeply comprehend complex editing instructions, Goku-Edit leverages an MLLM as its text encoder and adopts a decoupled dual-branch design: a dedicated mask branch handles structural control, freeing the main branch for appearance rendering. A comprehensive video editing benchmark, Goku-Bench, is also proposed with 1,000 human-verified test cases and 7 novel editing-specific metrics. Evaluated on Goku-Bench, Goku-Edit obtains up to +8% improvement on other open-source models in terms of instruction following.
Chinese Translation
现有的基于指令的视频编辑数据集通常专注于单一任务的外观编辑，未能满足现实场景中复杂的创作需求。为了解决这一问题，我们提出了Goku，一个大规模数据集，包含200万个高质量、与指令对齐的视频编辑对，这是第一个将任务边界从基本的外观编辑扩展到多任务和结构性操作（例如，对主体运动的精确控制）的数据集。为了应对这些复杂任务中固有的数据合成挑战，我们设计了一个高效的数据合成管道，将复杂编辑分解为可控的子问题，并在整个过程中引入了渐进式过滤系统以确保数据的可靠性。此外，我们在Goku上探索了最佳网络结构，并提出了Goku-Edit。为了深入理解复杂的编辑指令，Goku-Edit利用了一个多模态大语言模型（MLLM）作为其文本编码器，并采用了一个解耦的双分支设计：专用的掩码分支处理结构控制，使主分支可以专注于外观渲染。我们还提出了一个全面的视频编辑基准Goku-Bench，包含1000个经过人工验证的测试案例和7个新颖的编辑特定指标。在Goku-Bench上的评估显示，Goku-Edit在指令遵循方面相较于其他开源模型提高了多达8%。
cs.CV / 258 / 2606.30608
UnfoldArt: Zero-Shot Recovery of Full Articulated 3D Objects from Text or Image
UnfoldArt：从文本或图像中零-shot恢复完整关节3D对象
Abstract
Articulated 3D objects are essential for interactive environments in embodied AI, robotics, and virtual reality, but reconstructing their structure and motion from sparse observations remains challenging. Existing approaches remain largely constrained by lack of supervised data or lack the priors needed to reliably recover articulation, hidden geometry, and internal object structure. We present the first debate-driven agentic approach to articulated 3D object reconstruction from text or image inputs that both grounds articulation reasoning in concrete motion and exposes the occluded geometry revealed under articulation. High-level agents reason about object semantics and motion using knowledge from vision-language and video models, while low-level agents estimate articulation parameters and interaction points; together, they engage in a two-round structured debate that first exploits global--local disagreement and then grounds the agents in freely generated video. The same video prior, conditioned on the agreed articulation, then drives each part through its motion to expose occluded interiors and geometry that cannot be inferred from a single static view. By combining agentic reasoning with a video generative prior, our approach jointly infers articulation and reconstructs complete 3D articulated objects, producing high-fidelity geometry, internal structure, and motion-consistent states beyond directly observed surfaces.
Chinese Translation
关节3D对象对于具身人工智能、机器人技术和虚拟现实中的交互环境至关重要，但从稀疏观察中重建其结构和运动仍然具有挑战性。现有的方法在很大程度上受到缺乏监督数据或缺乏可靠恢复关节、隐藏几何和内部对象结构所需的先验知识的限制。我们提出了一种首个基于辩论驱动的自主方法，用于从文本或图像输入中重建关节3D对象，该方法将关节推理与具体运动相结合，并揭示关节下的遮挡几何。高级代理利用视觉-语言和视频模型的知识推理对象的语义和运动，而低级代理则估计关节参数和交互点；它们共同参与两轮结构化辩论，首先利用全局-局部的不一致性，然后将代理固定在自由生成的视频中。基于达成一致的关节的相同视频先验，随后驱动每个部分通过其运动，揭示无法从单一静态视图推断的遮挡内部和几何。通过将自主推理与视频生成先验相结合，我们的方法共同推断关节并重建完整的3D关节对象，生成高保真几何、内部结构和运动一致的状态，超越直接观察到的表面。
cs.CV / 259 / 2606.30611
Reweighting Framewise Attention in Video Transformers for Facial Expression Understanding
在视频变换器中重新加权逐帧注意力以理解面部表情
Abstract
Understanding facial expressions in videos requires modeling subtle and localized facial dynamics under unconstrained conditions. Although recent Vision Transformer~(ViT)-based video models have shown strong performance through large-scale self-supervised pretraining, their attention mechanisms often emphasize dominant global motions and coarse temporal dynamics, limiting sensitivity to fine-grained facial variations. To address this limitation, we propose MiRA (Marginal-induced Attention Redistribution), a plug-in frame-marginal attention redistribution framework for ViT backbones that enhances spatio-temporal selectivity toward subtle facial dynamics without introducing additional trainable parameters. MiRA derives frame-level confidence and intra-frame concentration statistics from self-attention maps to estimate frame-wise marginal importance and redistribute attention toward spatiotemporally localized facial cues. We first introduce a principled \textit{exact mode} based on post-softmax attention redistribution. To further improve efficiency, we propose \textit{flashLite mode}, a lightweight pre-softmax approximation that integrates frame-marginal redistribution into FlashAttention kernels while preserving the effectiveness of the exact formulation. Experimental results on challenging Facial Expression Recognition~(FER) benchmarks demonstrate consistent improvements over strong ViT baselines.
Chinese Translation
在视频中理解面部表情需要在不受约束的条件下对细微和局部的面部动态进行建模。尽管最近基于视觉变换器（Vision Transformer, ViT）的视频模型通过大规模自监督预训练展现了强大的性能，但它们的注意力机制往往强调主导的全局运动和粗略的时间动态，限制了对细粒度面部变化的敏感性。为了解决这一限制，我们提出了MiRA（Marginal-induced Attention Redistribution），一个用于ViT骨干网络的插件式逐帧边际注意力重分配框架，能够增强对细微面部动态的时空选择性，而无需引入额外的可训练参数。MiRA从自注意力图中推导出逐帧置信度和帧内集中统计，以估计逐帧的边际重要性，并将注意力重新分配到时空局部的面部线索上。我们首先介绍了一种基于后软最大（post-softmax）注意力重分配的原则性 extit{exact mode}。为了进一步提高效率，我们提出了 extit{flashLite mode}，这是一种轻量级的预软最大（pre-softmax）近似方法，将逐帧边际重分配集成到FlashAttention内核中，同时保持精确公式的有效性。在具有挑战性的面部表情识别（Facial Expression Recognition, FER）基准测试中的实验结果表明，相较于强大的ViT基线，MiRA consistently 提升了性能。
cs.CV / 260 / 2606.30638
Open-Vocabulary and Referring Segmentation for 3D Gaussians Using 2D Detectors
基于2D探测器的开放词汇和指代分割的3D高斯模型
Abstract
3D Gaussian Splatting (3DGS) has emerged at the forefront of 3D scene reconstruction. Extending 3DGS with language-driven, open-vocabulary understanding has gained significant attention for real-world applications such as embodied AI. Recent methods achieve this by learning an instance feature attribute and assigning semantics by distilling high-dimensional Contrastive Language-Image Pretraining (CLIP) features directly into the scene representation. However, the instance grouping mechanisms of these methods either require a predefined number of instances or suffer from noise in their bottom-up grouping strategies. Furthermore, the reliance on CLIP restricts semantic understanding to simple noun phrases, preventing complex spatial reasoning and referential expression grounding. We present GaussDet, a method that circumvents the need for dense CLIP features by leveraging discrete, open-vocabulary 2D object detectors with referring expression capabilities. We learn instance features for individual Gaussians to decompose the scene into 3D instance groups. By rendering these groups and aggregating semantic votes from multi-view 2D detections, we generate a robust View-Aggregated Semantic Label Distribution (VASD) for each 3D instance. This view-aggregation strategy acts as a strong regularizer, attenuating spurious labels caused by low-quality instance grouping. Our approach enables a straightforward, zero-shot extension from simple language queries to complex referential grounding. Extensive evaluations across two key tasks -- open-vocabulary segmentation (LeRF-OVS, ScanNet) and referring expression grounding (Ref-LeRF) -- demonstrate that GaussDet achieves consistent improvements over existing methods. Most notably, we achieve a substantial 16.7% mIoU improvement in referential grounding within a strict zero-shot setting.
Chinese Translation
3D高斯点云（3D Gaussian Splatting, 3DGS）已成为3D场景重建的前沿技术。将3DGS扩展至语言驱动的开放词汇理解在诸如具身人工智能等现实应用中引起了广泛关注。近期的方法通过学习实例特征属性，并将高维对比语言-图像预训练（Contrastive Language-Image Pretraining, CLIP）特征直接提炼到场景表示中，从而实现这一目标。然而，这些方法的实例分组机制要么需要预定义的实例数量，要么在自下而上的分组策略中受到噪声的影响。此外，对CLIP的依赖限制了语义理解仅限于简单的名词短语，阻碍了复杂空间推理和指代表达的基础。我们提出了GaussDet，一种通过利用具有指代表达能力的离散开放词汇2D物体探测器来规避对密集CLIP特征需求的方法。我们为每个高斯学习实例特征，将场景分解为3D实例组。通过渲染这些组并从多视角2D检测中聚合语义投票，我们为每个3D实例生成了稳健的视图聚合语义标签分布（View-Aggregated Semantic Label Distribution, VASD）。这种视图聚合策略作为强正则化器，有效减弱了由低质量实例分组引起的虚假标签。我们的方法使得从简单语言查询到复杂指代基础的零样本扩展变得简单。对两个关键任务的广泛评估——开放词汇分割（LeRF-OVS, ScanNet）和指代表达基础（Ref-LeRF）——表明GaussDet在现有方法上实现了一致的改进。最显著的是，在严格的零样本设置中，我们在指代基础任务中实现了显著的16.7%的mIoU提升。
cs.AI / 1 / 2606.28374
Recursive Self-Evolving Agents via Held-Out Selection
通过保留选择的递归自我进化代理
Abstract
LLM agents are increasingly improved without weight updates by evolving a natural-language artifact, such as reflections, workflows, playbooks, cheatsheets, or optimized prompts, that conditions a frozen policy. Such methods are typically reported as wins on the single benchmark where they help. We study them apples-to-apples and surface a sharper picture. We introduce RSEA, a Recursive Self-Evolving Agent that carries a compact three-layer natural-language state: an imperative strategy, reusable skills, and a procedural playbook. Across generations, RSEA rewrites all three layers from its own trajectories and commits a candidate only if it does not regress on a disjoint held-out split, using a strict keep-better gate. Across four diverse benchmarks, ALFWorld, GAIA, (\tau)-bench, and WebShop, and six faithful baselines, ReAct, Reflexion, GEPA, AWM, ACE, and Dynamic Cheatsheet, all evaluated on one shared local backbone, we find three main results. First, no artifact universally wins. RSEA is the strongest single-pass method on ALFWorld, reaching 69.3% compared with 64.6% for ReAct (McNemar (p=0.015)), and reaches 79.4% with retry, the best overall result. However, concrete-workflow induction, represented by AWM, is best on the strong-backbone tool-use tasks. Second, unguarded context evolution is high-variance and unsafe. Dynamic Cheatsheet, which curates context online without a held-out gate, is near-best on ALFWorld at 70.7%, yet collapses on WebShop, with a score of 0.14 compared with 0.43 for ReAct. Third, RSEA's strict held-out selection is what makes recursive self-evolution monotone-safe: it never significantly underperforms the base agent on any benchmark and falls back to vanilla ReAct when evolved context would hurt.
Chinese Translation
大型语言模型（LLM）代理通过进化自然语言工件（如反思、工作流程、操作手册、备忘单或优化提示）来不断改进，而无需更新权重，这些工件为固定策略提供条件。这类方法通常在它们所帮助的单一基准上被报告为成功。我们对这些方法进行了全面的比较，并揭示了更清晰的图景。我们引入了递归自我进化代理（RSEA），它携带一个紧凑的三层自然语言状态：一种命令策略、可重用的技能和一个程序化操作手册。在不同代际中，RSEA从自身轨迹中重写所有三层，并仅在不在一个独立的保留分割上退步时才提交候选者，采用严格的保留更好门限。在四个不同的基准测试中，ALFWorld、GAIA、( au)-bench 和 WebShop，以及六个忠实的基准线，ReAct、Reflexion、GEPA、AWM、ACE 和 Dynamic Cheatsheet，均在一个共享的本地基础上进行评估，我们发现了三个主要结果。首先，没有任何工件能够普遍获胜。RSEA在ALFWorld上是最强的单次通过方法，达到69.3%，而ReAct为64.6%（McNemar (p=0.015)），并在重试时达到79.4%，是最佳总体结果。然而，具体工作流程诱导（由AWM表示）在强基础工具使用任务中表现最佳。其次，无防护的上下文进化具有高方差且不安全。Dynamic Cheatsheet在ALFWorld上在线策划上下文，接近最佳，得分为70.7%，但在WebShop上崩溃，得分为0.14，而ReAct为0.43。第三，RSEA严格的保留选择使得递归自我进化保持单调安全：在任何基准上，它从未显著低于基础代理的表现，并在进化上下文会造成损害时回退到普通的ReAct。
cs.AI / 2 / 2606.28471
Data and Evaluation Closed-Loop for Model Capability Enhancement
数据与评估闭环以增强模型能力
Abstract
Model capability is the central variable in LLM pre-training, yet is never observed directly: data shapes it prospectively, while evaluation reveals it only retrospectively, compressing samples, prompts, decoding, and scoring rules into one noisy score. Practical optimization runs this backward: a failure is observed first, and the engineer must infer the corpus fix. The two sides speak incompatible vocabularies -- benchmark names and per-sample correctness versus data sources, domains, and quality labels -- so this inference is usually intuition, not method. We close this gap with the \emph{capability slice}: a group of evaluation samples sharing background condition, task type, solving operation, and output constraint -- precise enough to localize a single weakness yet stable enough to survive aggregation, unlike a benchmark name, too coarse, or a single sample, too noisy. Built around this unit, an evaluation taxonomy, a non-instruction data taxonomy, and mapping rules form a closed loop turning a benchmark-level failure into a targeted, testable data intervention. We test this loop on two case studies pulling in opposite directions. First, the loop rules the data out: continued pre-training drives BBH down by $-46.82\%$, but diagnosis traces this to a single masked \texttt{\textless EOS\textgreater} loss rather than weakened reasoning; restoring it recovers BBH to $66.44$, above the original checkpoint, without changing the data. Second, the loop rules the data in: a persistent math-reasoning weakness is decomposed by solving operation into specific failing combinations, and a weakness-targeted sampling procedure built from it lifts AIME2025/AIME2026 Pass@128 from $6.67$/$0.00$ to $26.67$ each. The same unmodified loop reaches opposite, correct verdicts in both cases, showing the evaluation-to-data inference can be routine, auditable, and experimentally validated rather than intuitive.
Chinese Translation
模型能力是大型语言模型（LLM）预训练中的核心变量，但从未被直接观察到：数据从前瞻性上塑造它，而评估仅在回顾性上揭示它，将样本、提示、解码和评分规则压缩为一个嘈杂的分数。实际优化则是反向进行：首先观察到失败，工程师必须推断出语料库的修正。这两方面使用不兼容的词汇——基准名称和每个样本的正确性与数据源、领域和质量标签之间的对比，因此这种推断通常是直觉而非方法。我们通过 extit{能力切片}来弥合这一差距：一组共享背景条件、任务类型、解决操作和输出约束的评估样本——足够精确以定位单一弱点，但又足够稳定以承受聚合，与基准名称过于粗糙或单个样本过于嘈杂不同。围绕这一单元构建的评估分类法、非指令数据分类法和映射规则形成一个闭环，将基准级别的失败转化为有针对性的、可测试的数据干预。我们在两个相反方向的案例研究中测试这个闭环。首先，闭环将数据排除：持续的预训练使BBH下降了$-46.82\%$，但诊断追溯到单个掩码 exttt{ extless EOS extgreater}损失，而非推理能力减弱；恢复它使BBH恢复到$66.44$，超过原始检查点，而不改变数据。其次，闭环将数据纳入：持久的数学推理弱点通过解决操作分解为特定的失败组合，而基于此构建的针对弱点的采样程序将AIME2025/AIME2026的Pass@128从$6.67$/$0.00$提升至各自的$26.67$。相同的未修改闭环在这两种情况下得出相反而正确的结论，表明评估到数据的推断可以是常规的、可审计的，并且经过实验验证，而非直觉的。
cs.AI / 3 / 2606.28514
GPTNT: Benchmarking Real-Time Collaboration Between Multimodal Agents on Keep Talking And Nobody Explodes
GPTNT：基于《继续说话，没人爆炸》的多模态智能体实时协作基准测试
Abstract
Multimodal models are increasingly deployed to solve tasks collaboratively with humans or other artificial agents. Existing benchmarks show that these models possess many of the required component capabilities, but the conditions that coincide in collaboration, including time pressure, information asymmetry, and imperfect communication, are usually studied in isolation. We introduce GPTNT, a benchmark built on the cooperative video game Keep Talking and Nobody Explodes, in which two agents must coordinate to defuse procedurally generated bomb puzzles against a live countdown. One agent can see and manipulate the bomb but does not have the defusal instructions; the other has the instructions but cannot see or manipulate the bomb. Neither agent can succeed alone: success requires effective and efficient communication. Unlike turn-based proxies, GPTNT requires agents to act asynchronously and communicate in real time. GPTNT is designed to separate collaboration from reliance on memorized solutions: the instruction manual, the partner, or both can be withheld to isolate what a model derives in the moment from what it already knows. We show that GPTNT poses a substantial challenge for state-of-the-art systems: none of the closed- or open-source models we test defuses a single bomb in real time, a bar that human players clear. Through controlled experiments, we identify critical weaknesses in state tracking, efficient action under time pressure, ambiguity handling, and error recovery. We release GPTNT as a benchmark for collaborative performance that current evaluations leave unmeasured. Because it runs on the real game, GPTNT benefits from procedural generation and inherits a living modding community, allowing the benchmark to evolve as models improve rather than being solved once and retired.
Chinese Translation
多模态模型越来越多地被部署以与人类或其他人工智能体协作解决任务。现有基准显示这些模型具备许多所需的组件能力，但协作中所涉及的条件，如时间压力、信息不对称和不完美的沟通，通常是孤立研究的。我们引入了GPTNT，这是一个基于合作视频游戏《继续说话，没人爆炸》的基准，在该游戏中，两个智能体必须协调以在实时倒计时下拆除程序生成的炸弹谜题。一个智能体可以看到并操作炸弹，但没有拆弹指令；另一个智能体拥有指令，但无法看到或操作炸弹。两个智能体都无法单独成功：成功需要有效且高效的沟通。与基于回合的代理不同，GPTNT要求智能体异步行动并实时沟通。GPTNT旨在将协作与对记忆解决方案的依赖分离：可以保留指令手册、合作伙伴或两者，以隔离模型在当下所推导的内容与其已知内容之间的区别。我们展示了GPTNT对最先进系统构成了重大挑战：我们测试的闭源和开源模型均未能在实时中拆除单个炸弹，而这一标准是人类玩家能够达到的。通过控制实验，我们识别出状态跟踪、在时间压力下的高效行动、模糊处理和错误恢复等方面的关键弱点。我们将GPTNT发布为一个协作性能的基准，目前的评估未能测量这些性能。由于它在真实游戏中运行，GPTNT受益于程序生成，并继承了一个活跃的模组社区，使得基准能够随着模型的改进而发展，而不是一次性解决后就被搁置。
cs.AI / 4 / 2606.28556
IMCBench: A benchmark for multimodal LLMs in Image-grounded Medical Conversations
IMCBench：图像基础医学对话中的多模态大型语言模型基准测试
Abstract
Recent advances in large language models and vision-language models have enabled reasoning over multimodal data, offering opportunities for clinical applications such as decision support and triaging. However, existing medical AI benchmarks are fragmented: some support multi-turn dialogues but lack images, while others provide multimodal inputs but focus on single-turn QA tasks. To address this gap, we introduce IMCBench, an image-grounded, multi-turn medical conversation benchmark that pairs real, publicly available clinical images with synthetic patient profiles to simulate realistic patient-clinician interactions. Each conversation is evaluated across three clinical dimensions: safety, accuracy, and appropriate use of uncertainty in diagnosis. We benchmark eight multimodal frontier models across four model families (Claude, GPT, Nova, and Llama), scoring each on a 1-5 scale using LLM-as-Jury scoring calibrated against expert clinician annotations. Our results show that Claude Opus 4.6 achieves the highest overall score (3.61), followed by Claude Sonnet 4.6 (3.30) and GPT-5.2 (3.29), though no model dominates all dimensions and safety degrades for both malignant and rare conditions ($\Delta$ = -0.27 each). Ablation studies further reveal that both visual input and EHR context contribute to safe guidance (safety drops of 0.18 and 0.23 on average when each is removed), with stronger models leveraging visual features more effectively. Together, these findings demonstrate that accurate clinical description does not guarantee safe patient guidance, motivating the need for multi-dimensional evaluation frameworks in medical AI.
Chinese Translation
近期大型语言模型和视觉-语言模型的进展使得对多模态数据进行推理成为可能，为临床应用如决策支持和分诊提供了机会。然而，现有的医学人工智能基准测试存在碎片化的问题：一些基准支持多轮对话但缺乏图像，而另一些则提供多模态输入但专注于单轮问答任务。为了解决这一问题，我们推出了IMCBench，这是一个图像基础的多轮医学对话基准，结合了真实的、公开可用的临床图像与合成的患者档案，以模拟真实的患者-临床医生互动。每个对话在三个临床维度上进行评估：安全性、准确性和在诊断中适当使用不确定性。我们对四个模型家族（Claude、GPT、Nova和Llama）中的八个多模态前沿模型进行了基准测试，使用1-5的评分标准，并根据专家临床医生的注释对LLM-as-Jury评分进行了校准。我们的结果显示，Claude Opus 4.6获得了最高的总体评分（3.61），其次是Claude Sonnet 4.6（3.30）和GPT-5.2（3.29），尽管没有模型在所有维度上都占据主导地位，并且在恶性和罕见病症的安全性均有所下降（$ riangle$ = -0.27）。消融研究进一步揭示，视觉输入和电子健康记录（EHR）上下文均对安全指导有贡献（当去除其中任一项时，安全性平均下降0.18和0.23），而更强的模型能够更有效地利用视觉特征。综合来看，这些发现表明，准确的临床描述并不保证安全的患者指导，强调了在医学人工智能中需要多维度评估框架的必要性。
cs.AI / 5 / 2606.28589
Search for Truth from Reasoning: A Dynamic Representation Editing Framework for Steering LLM Trajectories
从推理中寻找真理：一个动态表示编辑框架用于引导大型语言模型轨迹
Abstract
Current approaches to enhance Large Language Model (LLM) reasoning, such as Chain-of-Thought and "Wait" prompts, primarily encourage models to think more, yet often fail to guide them toward Truth. While Representation Editing (RepE) offers a intrinsic control, its application to dynamic reasoning trajectories remains underexplored. In this work, we bridge this gap by investigating the geometry of truth within unfolding reasoning chains. We uncover three critical insights: (1) Truth is encoded at the sentence level and is entangled with latent reasoning patterns; (2) Effective intervention follows an Uncertainty Principle and a Decay Effect, requiring localization to early, high-entropy forks; (3) Naive steering vectors suffer from noise, risking collateral damage to correct trajectories. Based on these findings, we propose DynaSteer, a dynamic RepE framework. DynaSteer employs pattern clustering to disentangle reasoning manifolds and utilizes Fisher-LDA to project purified truth. By dynamically monitoring lookahead entropy, it selectively steers and rolls back trajectories only when necessary. Comprehensive experimental results on several MATH benchmark verify the effectiveness of DynaSteer, and experiments on out-of-domain coding tasks further confirm its generalization ability. Our code is publicly available at https://github.com/tianlwang/DynaSteer.
Chinese Translation
当前增强大型语言模型（LLM）推理的方法，如链式思维（Chain-of-Thought）和“等待”（Wait）提示，主要鼓励模型进行更多思考，但往往未能有效引导它们走向真理。尽管表示编辑（Representation Editing, RepE）提供了内在控制，但其在动态推理轨迹中的应用仍然未被充分探索。在本研究中，我们通过研究展开推理链中的真理几何，填补了这一空白。我们揭示了三个关键见解：（1）真理在句子层面上编码，并与潜在推理模式交织在一起；（2）有效干预遵循不确定性原理（Uncertainty Principle）和衰减效应（Decay Effect），需要定位到早期的高熵分叉；（3）简单的引导向量易受噪声影响，可能对正确轨迹造成附带损害。基于这些发现，我们提出了DynaSteer，一个动态的RepE框架。DynaSteer采用模式聚类来解开推理流形，并利用Fisher-LDA投影纯化的真理。通过动态监测前瞻熵，它仅在必要时选择性地引导和回滚轨迹。在多个MATH基准上的全面实验结果验证了DynaSteer的有效性，而在领域外编码任务上的实验进一步确认了其泛化能力。我们的代码已公开发布在https://github.com/tianlwang/DynaSteer。
cs.AI / 6 / 2606.28683
Aristotelian Virtue Profiling of LLMs through Ethical Dilemmas
通过伦理困境对大型语言模型进行亚里士多德美德画像
Abstract
Large Language Models (LLMs) often face ethical tradeoffs in which several responses may be defensible but express different priorities, such as fairness, honesty, courage, or restraint. We introduce VirtueMap, a framework for describing these patterns through an Aristotelian virtue-ethics lens. Instead of asking for a single correct answer, VirtueMap asks humans or LLMs to rank all five responses to each of seven general, non-lethal, non-political, and non-religious ethical dilemmas. To define the reference orderings used for scoring, we first proposed, for each dilemma and virtue, an ordering of the five responses from most to least expressive of that virtue. We then collected more than 100 respondent evaluations per ordering and retained it as operational ground truth only when at least 95% confirmed it. Rankings are scored against these retained orderings using normalized Borda alignment, yielding profiles over Practical Wisdom, Justice, Truthfulness, Courage, and Temperance. We apply VirtueMap to nine LLM families in a repeated-run evaluation and find high mean rank consistency (90.3%), with the largest differences appearing on Courage, Temperance, and Justice. We also release an interactive website that computes profiles locally in the browser and compares respondents with measured LLM profiles.
Chinese Translation
大型语言模型（LLMs）常常面临伦理权衡，其中多个回应可能是合理的，但表达出不同的优先级，如公平、诚实、勇气或克制。我们引入了VirtueMap，一个通过亚里士多德美德伦理学视角描述这些模式的框架。VirtueMap并不是要求一个单一的正确答案，而是要求人类或LLMs对七个一般性的、非致命的、非政治的和非宗教的伦理困境中的五个回应进行排名。为了定义用于评分的参考排序，我们首先为每个困境和美德提出了五个回应的排序，从最能表达该美德到最不能表达。然后，我们为每个排序收集了超过100个回应者的评估，仅在至少95%的人确认时，才将其保留为操作性真实值。排名是根据这些保留的排序使用归一化的Borda对齐进行评分，从而生成关于实践智慧、公正、真实、勇气和克制的画像。我们将VirtueMap应用于九个LLM家族进行重复运行评估，发现平均排名一致性高达90.3%，其中在勇气、克制和公正方面的差异最大。我们还发布了一个交互式网站，可以在浏览器中本地计算画像，并将回应者与测量的LLM画像进行比较。
cs.AI / 7 / 2606.28692
An AI agent for treatment reasoning over a biomedical tool universe
用于生物医学工具宇宙的治疗推理AI代理
Abstract
Treatment reasoning underpins every therapeutic decision, integrating disease context, comorbidities, medications, contraindications, and evolving biomedical knowledge to select an appropriate therapy. It is inherently iterative: candidates are weighed against many constraints, revised as evidence emerges, and grounded in verifiable sources. Here we introduce ATHENA-R1, an AI agent for treatment reasoning across all FDA approved drugs since 1939, trained by reinforcement learning over a universe of 212 biomedical tools. At each step it identifies missing information, selects and runs relevant tools, and incorporates the evidence. To train it without human-annotated traces, we build a two-level self-learning framework: multi-agent systems construct the tools, tasks, and reasoning trajectories for supervised fine-tuning, then reinforcement learning with scientific feedback rewards reasoning quality (evidence gathering, grounded tool use, logical non-redundancy). Across five benchmarks of 3,168 drug reasoning tasks and 456 patient treatment cases, ATHENA-R1 outperforms language models and tool-use systems, reaching 94.7% accuracy on open-ended drug reasoning and 82.9% on treatment reasoning, 17.8 and 10.7 points above GPT-5. In blinded evaluations by experts from 28 rare disease organizations, it is preferred over reference models on all criteria, and physicians rated it favorably on complex hospitalized cardiovascular and infectious-disease cases. Adverse-event hypotheses it generated, tested in electronic health records from 5.4 million patients, reached adjusted odds ratios of 1.48-1.84, with no elevation among negative controls. Because it requires knowing what evidence to seek before concluding, treatment reasoning has long been hard for AI; we show it can be reframed as a learnable process of iterative evidence gathering that reinforcement learning can train AI to perform.
Chinese Translation
治疗推理是每一个治疗决策的基础，它整合了疾病背景、合并症、药物、禁忌症以及不断发展的生物医学知识，以选择合适的治疗方案。这一过程本质上是迭代的：候选方案在许多约束条件下进行权衡，随着证据的出现而修订，并基于可验证的来源进行验证。在此，我们介绍了ATHENA-R1，一个用于治疗推理的AI代理，涵盖自1939年以来所有FDA批准的药物，采用强化学习在212个生物医学工具的宇宙中进行训练。在每一步中，它识别缺失的信息，选择并运行相关工具，并整合证据。为了在没有人工标注痕迹的情况下进行训练，我们构建了一个两级自学习框架：多智能体系统构建工具、任务和推理轨迹以进行监督微调，然后通过科学反馈的强化学习奖励推理质量（证据收集、基于工具的使用、逻辑非冗余性）。在3168个药物推理任务和456个患者治疗案例的五个基准测试中，ATHENA-R1的表现优于语言模型和工具使用系统，在开放式药物推理中达到了94.7%的准确率，在治疗推理中达到了82.9%，分别比GPT-5高出17.8和10.7个百分点。在来自28个罕见疾病组织的专家盲评中，它在所有标准上优于参考模型，医生在复杂的住院心血管和传染病案例中给予了积极评价。它生成的不良事件假设在540万患者的电子健康记录中进行测试，调整后的比值比达到1.48-1.84，负对照组没有升高。由于治疗推理要求在得出结论之前知道要寻找什么证据，因此长期以来对AI来说一直很困难；我们展示了它可以被重新构建为一个可学习的迭代证据收集过程，强化学习可以训练AI执行这一过程。
cs.AI / 8 / 2606.28696
COMPASS: Grounding Composition-Intent Guidance in Unified Multimodal Models
COMPASS：在统一多模态模型中扎根于组合意图指导
Abstract
Composition is a high-level visual intent that governs where subjects are placed and how a scene is organized, yet current unified multimodal models remain unreliable at fine-grained composition recognition and struggle to turn such intent into controllable generation. We present COMPASS, the first unified multimodal framework that grounds composition-intent control in a single system spanning both composition perception and composition-guided generation, with a shared expert token $\tau_c$ as the central intent anchor. On the perception side, COMPASS injects composition expertise into an MoE backbone in a minimally invasive manner and distills the inferred intent into $\tau_c$. On the generation side, COMPASS reuses $\tau_c$ as a global conditioning signal that steers the denoising trajectory, effectively converting passive composition analysis into explicit layout control. To support systematic instruction-following composition learning and evaluation at scale, we construct Comp-11, a large-scale dataset with an 11-class taxonomy and reasoning-augmented annotations. Extensive experiments show that COMPASS substantially improves category-level composition understanding and delivers more composition-consistent, prompt-faithful generation than strong baselines.
Chinese Translation
组合是一种高层次的视觉意图，决定了主体的放置位置和场景的组织方式，但当前的统一多模态模型在细粒度组合识别方面仍然不可靠，并且难以将这种意图转化为可控的生成。我们提出了COMPASS，这是第一个将组合意图控制扎根于一个系统中的统一多模态框架，涵盖组合感知和组合引导生成，以共享的专家标记$ au_c$作为中心意图锚点。在感知方面，COMPASS以最小的侵入方式将组合专业知识注入MoE（混合专家）骨干网络，并将推断出的意图提炼为$ au_c$。在生成方面，COMPASS重用$ au_c$作为全球条件信号，引导去噪轨迹，有效地将被动的组合分析转化为明确的布局控制。为了支持系统化的指令遵循组合学习和大规模评估，我们构建了Comp-11，这是一个具有11类分类法和增强推理注释的大规模数据集。大量实验表明，COMPASS显著提高了类别级组合理解，并提供了比强基线更一致的组合生成和更忠实于提示的生成。
cs.AI / 9 / 2606.28707
BV-Blend: Uncertainty-Weighted Historical Baselines for Stable Critic-Free RL with Verifiable Rewards
BV-Blend：用于稳定的无评论员强化学习的基于不确定性的历史基线与可验证奖励
Abstract
Critic-free reinforcement learning with verifiable rewards (RLVR), exemplified by Group Relative Policy Optimization (GRPO), avoids training a value function (critic) and reduces memory and compute overhead relative to critic-based PPO pipelines for aligning large language models. However, GRPO-style advantage estimation depends on prompt-local (within-prompt-group) reward statistics and can be unstable. In particular, when all rollouts in a prompt group receive identical rewards, the within-group reward variance becomes zero, and group normalization yields zero advantages for that group, impeding learning in cold-start regimes with binary verifiers. We introduce BV-Blend, a critic-free framework that stabilizes advantage estimation by combining prompt-local on-policy statistics with semantic-cluster-conditioned historical moments. BV-Blend maintains EMA-tracked reward moments for each cluster, derives a confidence weight from a standard error of the mean (SEM) proxy, and uses this weight to blend historical and prompt-local baseline and variance statistics into a standardized advantage for PPO-style clipped updates. Experiments on verifiable reasoning benchmarks show that BV-Blend improves training stability and performance, and remains robust in regimes where group-normalized methods may stall.
Chinese Translation
无评论员的可验证奖励强化学习（RLVR），以群体相对策略优化（GRPO）为例，避免了训练价值函数（评论员），并相对于基于评论员的PPO管道减少了内存和计算开销，以对齐大型语言模型。然而，GRPO风格的优势估计依赖于提示局部（组内提示）奖励统计，并可能不稳定。特别是，当一个提示组中的所有回放获得相同的奖励时，组内奖励方差变为零，组归一化导致该组的优势为零，从而阻碍了在使用二元验证器的冷启动阶段的学习。我们提出了BV-Blend，这是一种无评论员框架，通过将提示局部的在线统计与语义聚类条件下的历史时刻相结合来稳定优势估计。BV-Blend为每个聚类维护EMA跟踪的奖励时刻，从均值的标准误差（SEM）代理中推导出置信权重，并利用该权重将历史和提示局部的基线及方差统计混合成标准化的优势，以用于PPO风格的剪切更新。在可验证推理基准上的实验表明，BV-Blend提高了训练的稳定性和性能，并在可能停滞的组归一化方法的情况下保持了鲁棒性。
cs.AI / 10 / 2606.28710
The Two Genie Game: Adoption and Welfare in Audit-Grounded AI Governance
双精灵游戏：基于审计的人工智能治理中的采纳与福利
Abstract
We ask under what conditions an agent with a harm-minimizing policy can displace an approval-seeking (RLHF) agent in a competitive market, and when that policy is sufficient to prevent community harm. We use evolutionary game theory (finite-population Moran-Fermi pairwise comparison) to formalize this subject to assumptions of wisher hindsight, peer testimony, a monotone harm ledger, sufficient information density of community feedback, and a finite, depleting resource pool, in a negative-sum environment. We show that adoption is favored when the prior distributions on how readily wishers attune to community sentiment are monotone, exhibit endpoint inversion, and have a centro-symmetric pairing property, and demonstrate this with several long-tailed priors (Hill, Pareto, Lomax, Frechet). Where it is favored, a critical adoption level separates communities that drift back to the approval-seeking agent from those for which the audited agent fixes; above that level fixation is the overwhelmingly likely outcome. We derive when fixation is attainable as a bound on the effective (informational) size N_c of the community, which must be small enough to allow fixation before depletion. We present these as Theorems 5.4 and 5.5; the algebraic and finite-grid backbone is machine-checked in Lean 4, with the barrier-crossing asymptotics retained as explicit hypotheses. We show that a self-audited agent with a community ledger is not, in general, sufficient to prevent community harm. Sufficiency depends both upon the alignment of the agent's audit with community values and the timeframe over which harm is evaluated. Regardless of alignment, once adoption reaches dominance, the state is absorbing. The same policy that reduced harm under alignment becomes a trap, welfare-negative under misalignment and, even under alignment, one that locks in harm deferred past the adoption horizon.
Chinese Translation
我们探讨在什么条件下，采用最小化伤害政策的代理可以在竞争市场中取代寻求批准的代理（RLHF），以及何时该政策足以防止社区伤害。我们使用进化博弈理论（有限人口的莫兰-费米成对比较）来形式化这一主题，假设包括愿望者的事后洞察、同行证言、单调伤害账本、社区反馈的信息密度足够，以及有限的、逐渐耗竭的资源池，处于负和环境中。我们表明，当愿望者对社区情感的调适程度的先验分布是单调的、表现出端点反转，并具有中心对称配对特性时，采纳是有利的，并通过几种长尾先验（Hill、Pareto、Lomax、Frechet）来证明。在有利的情况下，一个临界采纳水平将社区分为两类：一类是回归寻求批准的代理，另一类是审计代理能够修复的社区；在该水平之上，固定是极有可能的结果。我们推导出固定可达的条件，作为社区有效（信息）规模 N_c 的界限，该规模必须足够小，以便在资源耗竭之前实现固定。我们将这些结果呈现为定理 5.4 和 5.5；代数和有限网格的基础在 Lean 4 中经过机器检查，跨越障碍的渐近性质作为显式假设保留。我们表明，具有社区账本的自我审计代理通常不足以防止社区伤害。充分性既依赖于代理的审计与社区价值观的一致性，也依赖于评估伤害的时间框架。无论一致性如何，一旦采纳达到主导地位，状态便是吸收的。相同的政策在一致性下减少伤害，但在不一致性下成为陷阱，导致福利负面，甚至在一致性下，也会锁定在超出采纳视野的延迟伤害上。
cs.AI / 11 / 2606.28716
TrajRS: Towards Certified Robustness in Pedestrian Trajectory Prediction
TrajRS：朝着行人轨迹预测的认证鲁棒性迈进
Abstract
The robustness of trajectory prediction models is crucial for developing safe autonomous driving systems. Adversarial attacks on trajectory prediction can significantly impair the accuracy of predicted trajectories, leading to hazardous driving behaviors. While heuristic defense strategies have been implemented to enhance the robustness of trajectory prediction models, these measures often fail against more sophisticated, targeted adversarial attacks. Hence, there is a pressing need to establish verifiable safety assurances for trajectory prediction models. In this paper, we extend the traditional Randomized Smoothing framework to "TrajRS", which provides a certified robust radius for smoothed trajectory predictors. We clarify and expand the formal definitions of robustness in trajectory prediction and tailor the practical TrajRS scheme specifically to "robustness for the optimal prediction" and "robustness for all possible predictions". An extensive set of experiments demonstrates that TrajRS effectively achieves robustness certification for all smoothed pedestrian trajectory predictors in this work.
Chinese Translation
轨迹预测模型的鲁棒性对于开发安全的自动驾驶系统至关重要。对轨迹预测的对抗攻击可能显著降低预测轨迹的准确性，从而导致危险的驾驶行为。尽管已经实施了启发式防御策略以增强轨迹预测模型的鲁棒性，但这些措施往往无法抵御更复杂的、有针对性的对抗攻击。因此，迫切需要为轨迹预测模型建立可验证的安全保障。在本文中，我们将传统的随机平滑框架扩展为“TrajRS”，该框架为平滑的轨迹预测器提供了认证的鲁棒半径。我们澄清并扩展了轨迹预测中鲁棒性的正式定义，并将实用的TrajRS方案特别定制为“最优预测的鲁棒性”和“所有可能预测的鲁棒性”。一组广泛的实验表明，TrajRS有效地实现了本文中所有平滑行人轨迹预测器的鲁棒性认证。
cs.AI / 12 / 2606.28719
ComMem: Complementary Memory Systems for Test-Time Adaptation of Vision-Language Models
ComMem：用于视觉-语言模型测试时适应的互补记忆系统
Abstract
Test-time adaptation (TTA) of vision-language models (VLMs) is essential for their robust deployment in dynamic, real-world environments. However, existing TTA methods often adapt locally without accumulating knowledge over time, or operating within a single modality without exploiting VLMs' inherently multi-modal nature. Inspired by the \textbf{Com}plementary \textbf{Mem}ory systems of the biological brain, we propose \textbf{ComMem}, an innovative approach that mimics the distinct but cooperative roles of the hippocampus and neocortex to enable effective TTA for VLMs. ComMem consists of two key components: a fast-adapting detailed memory, akin to the hippocampus, that forms a dynamic visual cache from high-confidence test samples; and a slow-integrating abstract memory, akin to the neocortex, that continually refines global textual prototypes. For each test instance, ComMem jointly optimizes both memory systems to ensure cross-modal consistency. Extensive experiments on 15 benchmark datasets show that ComMem significantly outperforms state-of-the-art methods under both natural distribution shifts and cross-dataset generalization, offering a promising direction for enhancing VLMs' practical adaptability.
Chinese Translation
视觉-语言模型（VLMs）的测试时适应（TTA）对于其在动态现实环境中的稳健部署至关重要。然而，现有的TTA方法往往局限于局部适应，未能随着时间的推移积累知识，或在单一模态内操作，未能充分利用VLMs固有的多模态特性。受到生物大脑的 extbf{互补记忆}系统启发，我们提出了 extbf{ComMem}，这是一种创新的方法，模仿海马体和新皮层的不同但协作的角色，以实现VLMs的有效TTA。ComMem由两个关键组件组成：一个快速适应的详细记忆，类似于海马体，从高置信度的测试样本中形成动态视觉缓存；以及一个慢集成的抽象记忆，类似于新皮层，持续优化全局文本原型。对于每个测试实例，ComMem共同优化这两个记忆系统，以确保跨模态的一致性。在15个基准数据集上的大量实验表明，ComMem在自然分布转移和跨数据集泛化的情况下显著优于最先进的方法，为增强VLMs的实际适应性提供了一个有前景的方向。
cs.AI / 13 / 2606.28733
Agentic Abstention: Do Agents Know When to Stop Instead of Act?
代理性弃权：代理人是否知道何时停止而不是行动？
Abstract
LLM agents are expected to act over multiple turns, using search, browsing interfaces, and terminal tools to complete user goals. Yet not every goal is well specified or achievable in the available environment. In such cases, a reliable agent should recognize that further interaction is unlikely to help and abstain from additional tool calls. We define Agentic Abstention, the problem of deciding when an agent should stop acting under uncertainty. Unlike standard LLM abstention, which is usually evaluated as a single-turn answer-or-abstain decision, agentic abstention is a sequential decision problem: an agent can answer, abstain, or gather more information at each turn, and the need to abstain may only become clear after interacting with the environment. We study this problem across web shopping, terminal environments, and question answering, evaluating 13 LLM-as-agent systems and 2 agent scaffolds on more than 28,000 tasks. Our results show that the main challenge is not only whether agents can abstain, but also when they abstain. Some agents never abstain when they should, while others do so only after many unnecessary interactions. This gap is especially large on tasks where the instruction appears feasible until the environment reveals otherwise (e.g., no valid result matches the instruction). We further find that model scale, reasoning, and agent scaffolding affect abstention in different ways, where larger or more capable models sometimes perform worse at timely abstention. Finally, we introduce CONVOLVE, a context engineering method for improving agentic abstention that distills full interaction trajectories into reusable stopping rules. On WebShop, CONVOLVE substantially improves timely abstention without updating model parameters, raising Llama-3.3-70B's timely recall rate from 26.7 to 57.4. Our dataset and code are available at https://lhannnn.github.io/agentic-abstention
Chinese Translation
大型语言模型（LLM）代理人预期在多个回合中行动，利用搜索、浏览界面和终端工具来完成用户目标。然而，并非每个目标都是明确或可在现有环境中实现的。在这种情况下，一个可靠的代理人应该意识到进一步的互动不太可能有助于解决问题，并应避免额外的工具调用。我们定义了代理性弃权（Agentic Abstention），即在不确定性下决定代理人何时应停止行动的问题。与标准的LLM弃权不同，后者通常被评估为单回合的回答或弃权决策，代理性弃权是一个序列决策问题：代理人在每个回合中可以选择回答、弃权或收集更多信息，而弃权的必要性可能仅在与环境互动后才变得清晰。我们在网络购物、终端环境和问答中研究了这一问题，评估了13个LLM作为代理系统和2个代理支架在超过28,000个任务上的表现。我们的结果表明，主要挑战不仅在于代理人是否能够弃权，还在于他们何时弃权。有些代理人从不在应该弃权时弃权，而其他代理人则是在经历了许多不必要的互动后才弃权。这个差距在那些指令看似可行，直到环境揭示出相反情况的任务上尤其明显（例如，没有有效结果与指令匹配）。我们进一步发现，模型规模、推理能力和代理支架以不同方式影响弃权，其中更大或更强大的模型有时在及时弃权方面表现更差。最后，我们介绍了CONVOLVE，一种改善代理性弃权的上下文工程方法，它将完整的互动轨迹提炼为可重用的停止规则。在WebShop上，CONVOLVE显著提高了及时弃权的表现，而无需更新模型参数，将Llama-3.3-70B的及时召回率从26.7提高到57.4。我们的数据集和代码可在https://lhannnn.github.io/agentic-abstention获取。
cs.AI / 14 / 2606.28739
Agent Safety Is Action Alignment
代理安全即行动对齐
Abstract
Large language models increasingly act as agents: they call tools, move money, delete records, and send messages on a user's behalf. To keep them safe, practitioners imported the chatbot-era recipe (train the model to refuse unsafe inputs) into the agentic setting, and treat the resulting capability loss as a manageable ``alignment tax.'' We argue this is a \emph{category error}. Refusal is a primitive for \emph{content safety}, where the harm is in the model's output and is therefore a learnable function of it. Agentic harm is different in kind: it lies not in any output but in the relation between the authority an action exercises and the authority the user granted, which is absent from the text the model sees. Importing content-safety methods into this regime does not trade capability for safety; it pays capability and buys negative security. We support this with three lines of evidence spanning the autonomy spectrum: defense-trained models learn surface patterns rather than intent; the same training collapses multi-step agents before any threat appears while leaving them exploitable; and even undefended frontier models exceed granted authority under ordinary use. We conclude that action safety cannot be installed in weights. It must be expressed as \emph{least privilege}, enforced \emph{outside} the model at the action boundary, and evaluated as \emph{action alignment} (a relational, deployment-conditioned property) rather than a refusal score.
Chinese Translation
大型语言模型越来越多地作为代理进行操作：它们代表用户调用工具、转移资金、删除记录和发送消息。为了确保它们的安全，实践者将聊天机器人时代的做法（训练模型拒绝不安全输入）引入代理环境，并将由此产生的能力损失视为可管理的“对齐税”。我们认为这是一种 extit{类别错误}。拒绝是一种 extit{内容安全}的原始操作，其中危害存在于模型的输出中，因此可以作为其可学习的函数。代理危害在性质上是不同的：它不在任何输出中，而在于一个行动所行使的权威与用户授予的权威之间的关系，这种关系在模型所见的文本中是缺失的。将内容安全方法引入这一领域并不是在能力与安全之间进行权衡；它是支付能力而换取负面安全。我们通过三条证据支持这一观点，这些证据跨越自主性谱系：防御训练模型学习的是表面模式而非意图；相同的训练在任何威胁出现之前就使多步骤代理崩溃，同时使其仍然易受利用；即使是未防御的前沿模型在普通使用中也超出了授予的权威。我们得出结论，行动安全不能通过权重来安装。它必须以 extit{最小特权}的形式表达，在行动边界 extit{外部}强制执行，并作为 extit{行动对齐}（一种关系性、部署条件的属性）进行评估，而不是拒绝评分。
cs.AI / 15 / 2606.28747
Self-Supervised Theorem Discovery in a Formal Axiomatic System
在形式公理系统中的自监督定理发现
Abstract
Recent artificial intelligence (AI) systems have shown remarkable progress in mathematical reasoning. Many existing approaches, including large language models (LLMs), draw on human prior knowledge in the form of mathematical text, code, or theorem libraries. Although these approaches are highly effective in practice, it remains an open question whether an agent can autonomously discover useful theorems without such human priors. We study this question in a formal axiomatic system by developing an agent that starts from axioms and inference rules alone and gradually grows a library of useful theorems. Concretely, we propose a self-supervised theorem-discovery algorithm that alternates between proof search and useful-theorem extraction, building a theorem library whose entries are reused as lemmas for subsequent proof search. Experiments show that the agent discovers tens of thousands of theorems and finds proofs for human-written benchmark problems, suggesting that its discoveries include theorems meaningful from a human mathematical perspective. Furthermore, the discovered theorems improve LLM proof performance when provided as prompt lemmas, indicating that they can serve as external knowledge for LLM reasoning. Our results provide evidence that useful theorems can emerge from proof search without relying on human-provided theorem libraries. More broadly, they suggest a path toward self-evolving AI systems for mathematics whose discoveries remain formally verifiable.
Chinese Translation
近年来，人工智能（AI）系统在数学推理方面取得了显著进展。许多现有的方法，包括大型语言模型（LLMs），依赖于以数学文本、代码或定理库形式存在的人类先验知识。尽管这些方法在实践中非常有效，但一个未解的问题是，代理是否能够在没有人类先验的情况下自主发现有用的定理。我们在一个形式公理系统中研究这个问题，开发了一个仅从公理和推理规则出发，逐步构建有用定理库的代理。具体而言，我们提出了一种自监督定理发现算法，该算法在证明搜索和有用定理提取之间交替进行，构建一个定理库，其条目作为引理在后续的证明搜索中重复使用。实验表明，代理发现了数万个定理，并为人类编写的基准问题找到证明，表明其发现的定理在数学上对人类具有意义。此外，当这些发现的定理作为提示引理提供时，能够提高LLM的证明性能，表明它们可以作为LLM推理的外部知识。我们的结果提供了证据，表明有用的定理可以通过证明搜索产生，而无需依赖人类提供的定理库。更广泛地说，这些结果为自我进化的数学AI系统指明了一条道路，使其发现仍然可以形式上验证。
cs.AI / 16 / 2606.28770
Mechanistic Personality Analysis of LLMs Steering Personality via Latent Feature Interventions
通过潜在特征干预引导个性的机制性个性分析
Abstract
Large Language Models (LLMs) have demonstrated the ability to simulate human-like OCEAN personality traits in generated text. Previous efforts have focused on prompt engineering or fine-tuning to shape LLM personality. In this work, we propose a mechanistic interpretability approach that directly intervenes on the model's latent features. Our method identifies latent directions in the residual stream corresponding to a target OCEAN trait using sparse autoencoders (SAEs) and contrastive activation analysis. We formalize an additive steering vector in activation space and demonstrate how applying a small additive shift to the hidden states enhances the target trait while preserving overall language modeling performance. To determine the optimal combination of feature shifts, we explore a linear weighting heuristic with grid search optimization that balances personality expression with task performance. Our approach shows promise in controllably steering personality traits at the mechanistic level while maintaining high performance on standard benchmarks.
Chinese Translation
大型语言模型（LLMs）已经展示了在生成文本中模拟人类OCEAN个性特征的能力。之前的研究主要集中在提示工程或微调上，以塑造LLM的个性。在本研究中，我们提出了一种机制可解释性的方法，直接干预模型的潜在特征。我们的方法使用稀疏自编码器（SAEs）和对比激活分析，识别与目标OCEAN特征相对应的残差流中的潜在方向。我们在激活空间中形式化了一个加性引导向量，并展示了如何对隐藏状态施加小的加性偏移，以增强目标特征，同时保持整体语言建模性能。为了确定特征偏移的最佳组合，我们探索了一种结合了网格搜索优化的线性加权启发式方法，以平衡个性表达与任务性能。我们的方法在机制层面上可控地引导个性特征，同时在标准基准测试中保持高性能，显示出良好的前景。
cs.AI / 17 / 2606.28781
HyphaeDB: A Living Knowledge Topology for Agent-First Memory
HyphaeDB：一种面向代理的动态知识拓扑
Abstract
Every existing vector database and agent memory framework treats memory as passive storage that agents query explicitly. No system propagates knowledge between agents through the memory layer itself. We introduce HyphaeDB, an agent-native memory infrastructure that reinterprets the Hierarchical Navigable Small World (HNSW) graph topology the data structure at the core of every modern vector database not as a search optimization, but as a communication fabric for multi-agent AI systems. In HyphaeDB, agents are nodes in the vector space with persistent positions, knowledge propagates via a gossip protocol through the graph's neighbor structure with energy-based attenuation, and emergent behaviors contradiction detection, pattern crystallization, and consensus formation arise from the combination of topology, propagation dynamics, and local interaction rules. We present the architecture built on three primitives (knowledge nodes, topology edges, and memory diffs), a multi-layer abstraction hierarchy with promotion via emergent consensus, and theoretical analysis grounding the system in small-world network theory, epidemic broadcast protocols, and swarm intelligence. We provide a reference implementation on PostgreSQL with pgvector and describe a concrete deployment in Swarm-Driven Development, a multi-agent software engineering methodology. HyphaeDB represents, to our knowledge, the first system to combine navigable small world topology with gossip-based knowledge propagation for multi-agent coordination.
Chinese Translation
现有的每个向量数据库和代理记忆框架都将记忆视为被代理明确查询的被动存储。没有任何系统通过记忆层本身在代理之间传播知识。我们介绍了HyphaeDB，这是一种原生于代理的记忆基础设施，它重新诠释了层次可导航小世界（Hierarchical Navigable Small World, HNSW）图拓扑，作为每个现代向量数据库核心的数据结构，不再仅仅视为搜索优化，而是作为多代理人工智能系统的通信基础。在HyphaeDB中，代理作为向量空间中的节点，具有持久的位置，知识通过图的邻居结构以基于能量的衰减通过一种八卦协议传播，涌现行为如矛盾检测、模式结晶和共识形成则源于拓扑、传播动态和局部交互规则的结合。我们展示了基于三个原语（知识节点、拓扑边缘和记忆差异）构建的架构，一个通过涌现共识进行提升的多层抽象层级，以及将系统基础扎根于小世界网络理论、流行病广播协议和群体智能的理论分析。我们提供了基于PostgreSQL和pgvector的参考实现，并描述了在群体驱动开发（Swarm-Driven Development）这一多代理软件工程方法中的具体部署。根据我们的知识，HyphaeDB代表了第一个将可导航小世界拓扑与基于八卦的知识传播结合用于多代理协调的系统。
cs.AI / 18 / 2606.28798
Primary ICD Category Prediction using LLM-based Probing
基于大型语言模型（LLM）探测的主要 ICD 分类预测
Abstract
Objective: ICD codes are central to reimbursement, research, and population health surveillance, yet automated coding systems often struggle to integrate diagnostic signals from both clinical narratives and structured electronic health record (EHR) variables. We evaluated whether frozen medical large language model (LLM) representations can serve as a shared embedding space for multimodal primary diagnosis category prediction. Materials and Methods: We constructed a MIMIC-IV cohort of 13,645 admissions from the 10 most frequent primary ICD-10 codes, consolidated into seven categories. Structured variables were serialized into clinical narratives and combined with leakage-pruned discharge notes. Using a frozen MedFound-Llama3-8B-finetuned backbone, we extracted hidden states from five transformer layers and trained linear probes for structured-only, unstructured-only, and combined inputs, comparing against XGBoost and information-matched PLM-ICD baselines and evaluating MIMIC-III adaptation with a compact bottleneck adapter. Results: The combined probe performed best on MIMIC-IV (87.69% strict; 91.45% medical accuracy), exceeding both single-modality probes and baselines. The structured-only probe outperformed its standard baseline by 6.19 points in medical accuracy. Diagnostic information became increasingly linearly separable in deeper layers, and a 2M-parameter adapter restored cross-dataset transfer to MIMIC-III using only 5% of target labels. Discussion: LLM embeddings can unify structured and narrative EHR information for multimodal diagnosis prediction, supporting efficient reuse of clinical representations across modalities and datasets through a small representation-level module. Conclusion: Multimodal probing of frozen medical LLM representations provides a practical approach for studying EHR modalities and adapting clinical representations across datasets.
Chinese Translation
目的：ICD 代码在报销、研究和人口健康监测中至关重要，但自动编码系统常常难以整合来自临床叙述和结构化电子健康记录（EHR）变量的诊断信号。我们评估了冻结的医学大型语言模型（LLM）表示是否可以作为多模态主要诊断类别预测的共享嵌入空间。材料与方法：我们构建了一个包含 13,645 次入院的 MIMIC-IV 队列，涵盖 10 个最常见的主要 ICD-10 代码，并整合为七个类别。将结构化变量序列化为临床叙述，并与经过泄漏修剪的出院记录相结合。使用冻结的 MedFound-Llama3-8B 微调主干，我们从五个变换器层提取隐藏状态，并为仅结构化、仅非结构化和组合输入训练线性探测器，与 XGBoost 和信息匹配的 PLM-ICD 基线进行比较，并评估使用紧凑瓶颈适配器的 MIMIC-III 适应性。结果：组合探测器在 MIMIC-IV 上表现最佳（严格准确率 87.69%；医学准确率 91.45%），超过了单模态探测器和基线。仅结构化探测器在医学准确率上超出了其标准基线 6.19 分。随着层数加深，诊断信息变得越来越线性可分，并且一个 2M 参数的适配器在仅使用 5% 目标标签的情况下恢复了跨数据集转移到 MIMIC-III。讨论：LLM 嵌入可以统一结构化和叙述性 EHR 信息，以进行多模态诊断预测，通过一个小的表示级模块支持临床表示在不同模态和数据集之间的高效重用。结论：对冻结医学 LLM 表示的多模态探测提供了一种实用的方法，用于研究 EHR 模态并在数据集之间适应临床表示。
cs.AI / 19 / 2606.28900
MedEvoEval: Evaluating Continual Evolution of Doctor Agents through Simulated Clinical Episodes
MedEvoEval：通过模拟临床事件评估医生代理的持续演变
Abstract
Doctor agents are moving beyond single-turn answer generation toward evolving clinical decision systems. Within an outpatient episode, they acquire evidence, use examination and consultation resources, and decide when to finalize a diagnosis and management plan. Across episodes, their behavior may change through memory, retrieval, reflection, or other update mechanisms. Current evaluations only partially cover this setting. Fixed-input medical QA benchmarks score final answers from complete inputs, whereas many interactive benchmarks still focus on individual encounters or fixed runs, providing limited support for evaluating how episode-level decisions interact with cross-episode experience. We introduce MedEvoEval, an executable longitudinal evaluation framework based on action-gated simulated outpatient episodes. Each source case is converted into role-specific patient, examination, and manager views; evidence is revealed only through valid actions; and each episode records a structured trace that links observations, actions, final outputs, manager scores, and optional experience write-back. We release a runnable E&D artifact with 700 processed episodes, provenance notes, schemas, an episode runner, scoring scripts, configurations, example logs, analysis code, and trajectory- and step-level derivatives. Experiments show that episode traces expose process costs hidden by final-answer scoring, show how MDT-style consultation reallocates resources, and support longitudinal analyses of memory maturation, held-out transfer, update-stage response, and backward retention. Together, these results show that MedEvoEval provides a concrete basis for evaluating whether doctor agents improve through experience, transfer useful behavior, and retain earlier capabilities over time.
Chinese Translation
医生代理正在从单轮回答生成向演变的临床决策系统发展。在门诊事件中，它们获取证据，利用检查和咨询资源，并决定何时最终确定诊断和管理计划。在不同事件之间，它们的行为可能通过记忆、检索、反思或其他更新机制而发生变化。目前的评估仅部分覆盖这一设置。固定输入的医学问答基准通过完整输入对最终答案进行评分，而许多交互式基准仍然专注于单次接触或固定运行，提供的支持有限，无法评估事件级决策如何与跨事件经验相互作用。我们引入了MedEvoEval，这是一个基于动作门控模拟门诊事件的可执行纵向评估框架。每个源案例被转换为角色特定的患者、检查和管理者视图；证据仅通过有效的行动揭示；每个事件记录一个结构化的轨迹，链接观察、行动、最终输出、管理者评分和可选的经验回写。我们发布了一个可运行的E&D文物，包含700个处理过的事件、来源说明、模式、事件运行器、评分脚本、配置、示例日志、分析代码，以及轨迹和步骤级的衍生物。实验表明，事件轨迹揭示了最终答案评分所隐藏的过程成本，展示了多学科团队（MDT）风格的咨询如何重新分配资源，并支持对记忆成熟、保留转移、更新阶段响应和向后保留的纵向分析。这些结果表明，MedEvoEval为评估医生代理是否通过经验提高、转移有用行为以及随着时间的推移保留早期能力提供了具体基础。
cs.AI / 20 / 2606.28960
Expert Evaluation of Clinical AI Tools on Real Point-of-Care Clinical Queries
临床人工智能工具在实际临床查询中的专家评估
Abstract
Physicians now pose millions of clinical questions to AI tools each week, yet these tools are evaluated largely on hypothetical or exam-style questions, not those actually asked in practice. We report a blinded evaluation built on 620 Real-world Point-Of-Care Queries (Real-POCQi) submitted to the OpenEvidence (OE) platform by physicians spanning 30 specialties, as well as 187 questions from HealthBench. 149 practicing physicians across 36 states made head-to-head comparisons between answers from three frontier general-purpose models (Claude Opus 4.8, Gemini 3.1 Pro, and GPT-5.5) and a specialized clinical tool (OE), with graders matched to each question's specialty. When comparing answers along five dimensions relevant to clinical decision support -- accuracy, clinical utility, source quality, verifiability, & completeness -- physicians scored the specialized tool highest on all axes; in the primary analysis on Real-POCQi, win differences (margins between win and loss rates) ranged from 25 to 39 percentage points (p<0.001). Results remained consistent in sensitivity analyses stratifying by citation display, answer length, OE-user status, and Real-POCQi versus HealthBench. In parallel, LLM judges were found to systematically differ from expert judges, though both generally agreed on the best model. These findings underscore two conclusions: (i) AI tool evaluations should reflect real-world query distributions and use expert judges that mirror the specialization defining modern medicine and (ii) the consistent advantage of the specialized tool over general-purpose models does not necessarily mean that the latter cannot serve similar purposes, but that targeted engineering and customization can yield meaningful gains in performance for its users. We release Real-POCQi as a public benchmark, as well as the prespecified statistical analysis for reproducing results of this study.
Chinese Translation
目前，医生每周向人工智能工具提出数百万个临床问题，但这些工具的评估主要基于假设性或考试风格的问题，而非实际临床中提出的问题。我们报告了一项基于620个真实世界临床现场查询（Real-POCQi）的盲评估，这些查询由来自30个专业的医生提交至OpenEvidence（OE）平台，以及来自HealthBench的187个问题。来自36个州的149名执业医生对三种前沿通用模型（Claude Opus 4.8、Gemini 3.1 Pro和GPT-5.5）与一个专门的临床工具（OE）之间的答案进行了逐对比较，评估者与每个问题的专业领域相匹配。在比较与临床决策支持相关的五个维度——准确性、临床实用性、来源质量、可验证性和完整性时，医生在所有维度上均对专门工具给予了最高评分；在对Real-POCQi的主要分析中，胜率差异（胜率与失率之间的差距）范围从25到39个百分点（p<0.001）。在按引用显示、答案长度、OE用户状态以及Real-POCQi与HealthBench进行分层的敏感性分析中，结果保持一致。同时，发现大型语言模型（LLM）评审与专家评审在系统上存在差异，尽管两者普遍对最佳模型达成一致。这些发现强调了两个结论：（i）人工智能工具的评估应反映现实世界查询的分布，并使用与现代医学专业化相符的专家评审；（ii）专门工具相对于通用模型的一致优势并不意味着后者无法服务于类似目的，而是针对性的工程和定制可以为其用户带来显著的性能提升。我们将Real-POCQi作为公共基准发布，并提供预先指定的统计分析以重现本研究的结果。
cs.AI / 21 / 2606.29014
Customized Generative AI Agent for Transportation Engineering Practice: A Development and Continued Pre-training Guideline
针对交通工程实践的定制生成式人工智能代理：开发与持续预训练指南
Abstract
Recent advancements in generative artificial intelligence (AI) and large language models (LLMs) have shown significant promise in automating complex reasoning, summarization, and question-answering tasks. However, the effectiveness of general-purpose LLMs in specialized engineering domains remains limited due to insufficient exposure to technical standards, engineering terminology, and domain-specific semantics. This study proposes a systematic approach to developing a customized generative AI agent for transportation engineering applications. A curated corpus of U.S. transportation manuals, design guidelines, and regulatory documents is used to conduct continued pretraining of six state-of-the-art LLMs through a unified low-rank adaptation (LoRA) framework. The training process is monitored to ensure convergence and model stability. Performance is evaluated using standard natural language processing metrics, including BLEU-4 and ROUGE, with Qwen2.5-7B and LLaMA-3.1-8B demonstrating the highest domain alignment and response quality. Results validate the effectiveness of LoRA-based adaptation in improving LLM performance on technical content interpretation and context-specific reasoning. This work contributes a reproducible development framework for constructing domain-specialized generative AI agents, supporting broader deployment in transportation research, design, planning, and policy analysis.
Chinese Translation
最近，生成式人工智能（AI）和大型语言模型（LLMs）的进展在自动化复杂推理、摘要和问答任务方面显示出显著的潜力。然而，由于对技术标准、工程术语和特定领域语义的曝光不足，通用LLMs在专业工程领域的有效性仍然有限。本研究提出了一种系统的方法，用于开发针对交通工程应用的定制生成式AI代理。通过统一的低秩适应（LoRA）框架，使用美国交通手册、设计指南和法规文件的精选语料库对六个最先进的LLMs进行持续预训练。训练过程受到监控，以确保收敛性和模型稳定性。使用标准自然语言处理指标（包括BLEU-4和ROUGE）评估性能，其中Qwen2.5-7B和LLaMA-3.1-8B表现出最高的领域对齐和响应质量。结果验证了基于LoRA的适应在提高LLM在技术内容解释和上下文特定推理方面的性能的有效性。本研究为构建领域专用生成式AI代理提供了可重复的开发框架，支持在交通研究、设计、规划和政策分析中的更广泛应用。
cs.AI / 22 / 2606.29026
Preventing Error Propagation in Multi-Agent AI through Runtime Monitoring
通过运行时监控防止多智能体人工智能中的错误传播
Abstract
Multi-agent AI systems can improve answer selection by allowing different language models to exchange reasoning traces, revise initial predictions, and support a final decision. However, such communication may also introduce reliability risks: reasoning from one agent can correct another agent's mistake, but it can also mislead an agent that was initially correct. This paper studies reliable multi-agent AI communication through reasoning exchange and runtime answer revision. We develop a framework in which agents first answer multiple-choice questions independently, then share reasoning traces and revise their decisions. We conduct numerical experiments where we evaluate whether this process improves accuracy, produces more positive than negative answer transitions, and remains effective across domains such as cybersecurity, networking, and general knowledge. The results help identify when multi-agent reasoning improves reliability and when it may propagate errors.
Chinese Translation
多智能体人工智能系统通过允许不同的语言模型交换推理过程、修正初始预测并支持最终决策，从而提高答案选择的准确性。然而，这种通信也可能引入可靠性风险：一个智能体的推理可以纠正另一个智能体的错误，但也可能误导一个最初是正确的智能体。本文研究了通过推理交换和运行时答案修正实现可靠的多智能体人工智能通信。我们开发了一个框架，其中智能体首先独立回答多项选择题，然后共享推理过程并修正其决策。我们进行了数值实验，评估这一过程是否提高了准确性，产生的正向答案转换是否多于负向答案转换，并且在网络安全、网络和一般知识等领域中是否依然有效。结果有助于识别多智能体推理何时提高可靠性以及何时可能传播错误。
cs.AI / 23 / 2606.29030
Memory as an Attack Surface in LLM Agents: A Study on Multiple-Choice Question Answering
记忆作为大型语言模型代理的攻击面：多项选择题回答研究
Abstract
AI agents extend conventional large language model (LLM) applications by integrating language understanding with task execution, external tool use, and memory mechanisms. While memory allows agents to retain prior interactions and provide more personalized and context-aware responses, it also introduces a new vulnerability: information stored in memory can influence future outputs even when the current query is clean. In this paper, we investigate memory manipulation in LLM-based agents for multiple-choice question answering. We first design and implement an LLM-based AI agent with an external memory component that stores and retrieves task-relevant information. We then introduce basic memory manipulation scenarios in which misleading or corrupted memories are inserted into the agent before it answers multiple-choice questions. Using a controlled experimental setup, we compare the agent's performance before and after memory manipulation and measure changes in answer accuracy, attack success rate, and selection of manipulated options. Our results show that even simple memory manipulations can noticeably affect the agent's final answers, causing it to select incorrect options despite receiving clean and well-formed questions.
Chinese Translation
人工智能代理通过将语言理解与任务执行、外部工具使用和记忆机制相结合，扩展了传统的大型语言模型（LLM）应用。虽然记忆使代理能够保留先前的交互并提供更个性化和上下文感知的响应，但它也引入了一种新的脆弱性：存储在记忆中的信息可能会影响未来的输出，即使当前查询是干净的。在本文中，我们研究了基于LLM的代理在多项选择题回答中的记忆操控。我们首先设计并实现了一个带有外部记忆组件的基于LLM的人工智能代理，该组件存储和检索与任务相关的信息。然后，我们引入了基本的记忆操控场景，在这些场景中，误导性或损坏的记忆在代理回答多项选择题之前被插入。通过控制实验设置，我们比较了代理在记忆操控前后的表现，并测量了答案准确性、攻击成功率和操控选项的选择变化。我们的结果表明，即使是简单的记忆操控也能显著影响代理的最终答案，导致其在接收到干净且结构良好的问题时选择错误的选项。
cs.AI / 24 / 2606.29069
Low-cost concept-based localized explanations: How far can we get with training-free approaches?
低成本基于概念的局部解释：无训练方法能走多远？
Abstract
Concept-based Explainable AI (C-XAI) seeks human-understandable explanations grounded in semantic concepts, yet validation is limited by the scarcity of fine-grained concept annotations. We evaluate whether mid-scale Multimodal Large Language Models (MLLMs) can perform localized concept naming under strict zero-shot conditions by assigning labels to bounding-box regions at both object and part levels. We propose a reproducible zero-shot evaluation protocol for Concept Naming (CoNa) with (i) closed-set, category-constrained prompting for moderate vocabularies and (ii) Open-CoNa, an embedding-similarity-based strategy for large label spaces. Experiments with four MLLMs (7B-32B) show consistent performance trends across datasets, reaching 62%-88% object-level exact-match accuracy, highlighting the potential of training-free concept annotation from localized regions. We discuss limitations and failure modes and release a reproducible framework to support future low-cost C-XAI research.
Chinese Translation
基于概念的可解释人工智能（C-XAI）寻求以语义概念为基础的人类可理解的解释，但由于细粒度概念注释的稀缺，验证受到限制。我们评估中等规模的多模态大型语言模型（MLLMs）在严格的零样本条件下是否能够对边界框区域进行局部概念命名，涵盖对象和部分级别。我们提出了一种可重复的零样本概念命名（CoNa）评估协议，包括（i）针对中等词汇量的封闭集、类别约束提示，以及（ii）Open-CoNa，一种基于嵌入相似性的策略，适用于大标签空间。对四个MLLM（7B-32B）的实验显示，在不同数据集上表现出一致的性能趋势，达到62%-88%的对象级精确匹配准确率，突显了从局部区域进行无训练概念注释的潜力。我们讨论了局限性和失败模式，并发布了一个可重复的框架，以支持未来低成本C-XAI研究。
cs.AI / 25 / 2606.29111
Managing the Human Fallback: Skill Investment Under Improving AI and Worker Mobility
管理人类后备：在不断改善的人工智能和工人流动下的技能投资
Abstract
When firms deploy autonomous AI, they must decide how much work to leave to the system and how much to keep workers engaged. This decision affects current output and future human capital. We develop a parsimonious two-period model in which AI may outperform the worker when it functions, but may fail with positive probability. A firm chooses worker engagement; engagement lowers current output for below-benchmark workers, but changes future skill through learning and erosion. We distinguish two dimensions of AI progress: capability, the system's output when it works, and reliability, the probability that it works. In a single-firm benchmark, engagement is valuable only as fallback investment. The firm engages the least-skilled workers most, because they have the largest skill gaps and are least costly to bring toward a useful fallback level. With worker mobility, engagement also affects labor-market sorting: workers prefer jobs that build more valuable skill trajectories. This sorting motive targets higher-skill workers near the AI frontier, where skill gains are more valuable and engagement is less costly. Mobility can therefore reverse the engagement pattern, shifting investment from the least-skilled toward the most-skilled workers below the AI benchmark. Mobility also reshapes how AI progress affects engagement: greater capability raises engagement by increasing the value of the skill trajectory a firm offers, whereas greater reliability can raise or lower it because it reduces fallback need while also changing learning opportunities. Under worker mobility, human-AI work design becomes a problem of human-capital investment, in which allocating work today shapes future skill.
Chinese Translation
当公司部署自主人工智能（AI）时，它们必须决定将多少工作留给系统，以及将多少工作保留给工人参与。这一决策影响当前的产出和未来的人力资本。我们开发了一个简约的两期模型，其中AI在运行时可能优于工人，但也可能以正概率失败。公司选择工人参与度；对于低于基准的工人，参与度降低当前产出，但通过学习和技能流失改变未来技能。我们区分了AI进步的两个维度：能力，即系统在正常工作时的产出，以及可靠性，即系统正常工作的概率。在单一公司基准下，参与度仅作为后备投资具有价值。公司最常雇佣技能最低的工人，因为他们的技能差距最大，且将其提升至有用的后备水平的成本最低。随着工人流动，参与度也影响劳动市场的排序：工人更倾向于选择能够建立更有价值技能轨迹的工作。这一排序动机针对接近AI前沿的高技能工人，因为在此处技能提升更有价值，且参与成本较低。因此，流动性可以逆转参与模式，将投资从技能最低的工人转向低于AI基准的技能最高的工人。流动性还重塑了AI进步如何影响参与度：更大的能力通过提高公司提供的技能轨迹的价值来增加参与度，而更大的可靠性则可能提高或降低参与度，因为它减少了对后备的需求，同时也改变了学习机会。在工人流动的情况下，人类与AI的工作设计成为人力资本投资的问题，其中今天的工作分配影响未来的技能。
cs.AI / 26 / 2606.29116
Characterizing Large Language Model Agentic Workflows: A Study on N8n Ecosystem
大型语言模型代理工作流程的特征化：N8n生态系统研究
Abstract
Large Language Models (LLMs) are rapidly being adopted in low-code and no-code automation platforms, where non-expert users design workflows that combine natural language understanding with external services and APIs. LLM agents are LLM systems that use LLMs as a core "brain" to reason, plan, and autonomously execute complex, multi-step tasks. In this paper, we present the first large-scale empirical study of LLM agentic workflows in low-code automation platforms. We analyze more than 6,000 publicly available n8n workflows and examine four aspects of their design: task distribution, structural and tool use patterns, reliability mechanisms, and autonomy levels. Our analysis shows that LLM workflows are not merely prompt response pipelines. Instead, LLMs are commonly embedded within broader automation structures involving control logic, external tools, communication services, storage systems, and human review points. We further find that while many workflows include lightweight post-processing or routing logic after LLM execution, explicit reliability mechanisms such as structured fallback paths, repair loops, failure-specific alerts, and human approval gates remain relatively uncommon. These results reveal a gap between the increasing deployment of LLM agents in practical automation ecosystems and the limited engineering support for reliability, safety, and governance. Overall, our study provides ten empirical findings and five research takeaways for researchers, platform developers, and practitioners seeking to understand and improve real-world LLM agentic workflows.
Chinese Translation
大型语言模型（LLMs）正在迅速被低代码和无代码自动化平台所采用，在这些平台上，非专业用户设计结合自然语言理解与外部服务和API的工作流程。LLM代理是将LLM作为核心“脑”来推理、规划和自主执行复杂的多步骤任务的LLM系统。本文呈现了对低代码自动化平台中LLM代理工作流程的首次大规模实证研究。我们分析了6000多个公开可用的n8n工作流程，并考察了其设计的四个方面：任务分配、结构和工具使用模式、可靠性机制以及自主性水平。我们的分析表明，LLM工作流程不仅仅是提示响应管道。相反，LLM通常嵌入在更广泛的自动化结构中，涉及控制逻辑、外部工具、通信服务、存储系统和人工审核点。我们进一步发现，尽管许多工作流程在LLM执行后包含轻量级的后处理或路由逻辑，但诸如结构化后备路径、修复循环、特定故障警报和人工批准门等明确的可靠性机制仍然相对少见。这些结果揭示了LLM代理在实际自动化生态系统中日益部署与可靠性、安全性和治理的有限工程支持之间的差距。总体而言，我们的研究为研究人员、平台开发者和实践者提供了十个实证发现和五个研究启示，以帮助他们理解和改善现实世界中的LLM代理工作流程。
cs.AI / 27 / 2606.29126
HiComm: Hierarchical Communication for Multi-agent Reinforcement Learning
HiComm：用于多智能体强化学习的层次通信
Abstract
Cooperative multi-agent reinforcement learning (MARL) often relies on communication to mitigate partial observability, yet most existing protocols treat messages as flat dense vectors detached from the structure of the observations they summarize. This design overlooks an important source of inductive bias in many cooperative environments, where observations naturally follow a hierarchy such as groups and entities. We propose \textsc{HiComm}, a plug-in communication module that grounds messages in the sender's hierarchical observation. \textsc{HiComm} is receiver-driven: the receiver issues a query, and the hierarchy is resolved through a three-stage decoding process that first selects a group, then a sender, and then an entity within that group, returning the corresponding feature slice as the message. This converts communication from unstructured vector transmission into structured information retrieval over the sender's observation hierarchy. We instantiate this mechanism with Straight-Through Gumbel-Softmax for differentiable discrete selection and a lightweight shared projection design that attaches to standard MARL pipelines. Experiments across cooperative MARL tasks with different observation structures and coordination demands show that \textsc{HiComm} matches or outperforms representative learned communication baselines while reducing communication volume by up to $23\times$ per receiver per episode.
Chinese Translation
合作多智能体强化学习（MARL）通常依赖于通信来缓解部分可观测性，但大多数现有协议将消息视为与其所总结的观察结构无关的平坦密集向量。这种设计忽视了许多合作环境中一个重要的归纳偏置来源，在这些环境中，观察自然遵循如组和实体的层次结构。我们提出了 extsc{HiComm}，一个将消息与发送者的层次观察相结合的插入式通信模块。 extsc{HiComm} 是接收方驱动的：接收方发出查询，通过一个三阶段解码过程来解析层次结构，首先选择一个组，然后选择一个发送者，最后选择该组内的一个实体，并返回相应的特征切片作为消息。这将通信从非结构化的向量传输转换为对发送者观察层次的结构化信息检索。我们使用直通 Gumbel-Softmax 机制实现这一过程，以便进行可微分的离散选择，并设计了一个轻量级的共享投影，能够附加到标准的 MARL 流水线中。在具有不同观察结构和协调需求的合作 MARL 任务中的实验表明， extsc{HiComm} 的性能与代表性的学习通信基线相当或更优，同时每个接收方每个回合的通信量减少了多达 $23 imes$。
cs.AI / 28 / 2606.29150
Flow Reasoning Models: Scaling Reasoning Through Iterative Self-Refinement
流推理模型：通过迭代自我精炼扩展推理能力
Abstract
Discrete flow models have recently shown promising performance on few-step text generation; however, when naively applied to structured reasoning tasks such as Sudoku and Zebra puzzles, they converge confidently to incorrect answers (solving only $\sim$36% of Sudoku puzzles). We introduce Flow Reasoning Models (FRMs), a training and test-time-scaling framework for structured reasoning with flow models. We make the observation that, despite their poor solve rate, flow models can act as their own verifiers. A correct answer is a stable fixed point of the denoising dynamics, returning to itself when re-noised and re-solved. This enables a test-time-scaling paradigm: propose many candidate solutions and keep those that are dynamically stable, which alone reaches high solve rates on Sudoku-Shah (~$100\%$) and Zebra ($95.9\%$). This even generalizes to harder out-of-distribution puzzles like Sudoku-Extreme ($96.1\%$), without ever training on that distribution. This pure search, however, wastes a great deal of computation generating incorrect candidate solutions. We therefore design a training recipe to improve the base model's efficiency. First, we train flow models with a self-conditioning channel and close it at inference, letting them refine their own past predictions. Second, we train models to avoid their own failed generations using direct preference optimization. These changes substantially improve the base model's efficiency, letting it reach $99.2\%$ on Sudoku in just $7$ forward passes, over $8\times$ fewer than the strongest matched masked-diffusion baseline we compare needs for the same accuracy. When combined with test-time scaling, this lets flow models solve hard out-of-distribution puzzles (e.g. Sudoku-Extreme) far more efficiently.
Chinese Translation
离散流模型最近在少步文本生成上显示出良好的性能；然而，当简单地应用于结构化推理任务（如数独和斑马谜题）时，它们往往自信地收敛到错误答案（仅解决了约36%的数独谜题）。我们提出了流推理模型（Flow Reasoning Models, FRMs），这是一个用于流模型的结构化推理的训练和测试时间扩展框架。我们观察到，尽管流模型的解题率较低，但它们可以充当自身的验证者。正确答案是去噪动态的稳定不动点，在重新加噪和重新求解时返回到自身。这使得一种测试时间扩展范式成为可能：提出多个候选解，并保留那些动态稳定的解，这样就能在数独-沙赫（Sudoku-Shah）上达到高达100%的解题率，在斑马谜题上达到95.9%。这一方法甚至可以推广到更难的分布外谜题，如数独-极限（Sudoku-Extreme），而无需在该分布上进行训练。然而，这种纯搜索方法在生成错误候选解时浪费了大量计算资源。因此，我们设计了一种训练方案来提高基础模型的效率。首先，我们通过自条件通道训练流模型，并在推理时关闭它，让模型能够精炼自身的过去预测。其次，我们训练模型以避免自身失败的生成，使用直接偏好优化。这些变化显著提高了基础模型的效率，使其在仅进行7次前向传播的情况下在数独上达到99.2%的解题率，比我们比较的最强匹配掩蔽扩散基线所需的相同准确度少了8倍。当与测试时间扩展结合时，这使得流模型能够更高效地解决困难的分布外谜题（例如数独-极限）。
cs.AI / 29 / 2606.29159
Pooled Leaderboards Hide System-Specific Winners: A Reporting-Protocol Audit of Offline Root-Cause Analysis Benchmarks
汇总排行榜掩盖系统特定的赢家：离线根本原因分析基准的报告协议审计
Abstract
Offline root-cause-analysis (RCA) benchmarks commonly rank methods by a single pooled top-1 accuracy across multiple subsystems, and engineers often read the pooled winner as a recommendation for their own subsystem. We audit that reading on three public RCA benchmark families -- OpenRCA, RCAEval, and PetShop -- covering 11 subsystems and 778 matched scoring units. To keep pairwise comparisons on identical cases, the main analysis retains four methods or comparators with complete coverage: BARO, a CD-1min adapter, max-$|Z|$, and per-service alert-count. All six pairwise comparisons show subsystem-level effects of both signs, every random-effects 95\% prediction interval crosses zero, and case-level interaction tests reject exchangeability in 5 of 6 pairs. Leave-one-system-out selection picks the lower-scoring method on up to 5 of 11 held-out subsystems, with regret reaching 24.8 pp on RCAEval / Sock-Shop. We release a 320-line audit module; given a matched RCA benchmark score table, it recomputes the same per-subsystem stability checks alongside pooled scores.
Chinese Translation
离线根本原因分析（RCA）基准通常通过多个子系统的单一汇总顶级准确率对方法进行排名，工程师们常常将汇总赢家视为对其自身子系统的推荐。我们对三个公共RCA基准系列——OpenRCA、RCAEval和PetShop进行了审计，涵盖了11个子系统和778个匹配评分单元。为了保持对相同案例的成对比较，主要分析保留了四种方法或比较器，具有完整覆盖：BARO、CD-1min适配器、max-$|Z|$和每服务警报计数。所有六个成对比较显示出子系统层面的双向效应，每个随机效应的95 ext{%}预测区间均跨越零，案例级交互测试在6对中的5对拒绝可交换性。留一系统选择在多达11个保留子系统中选择得分较低的方法，遗憾值在RCAEval / Sock-Shop上达到24.8个百分点。我们发布了一个320行的审计模块；给定一个匹配的RCA基准评分表，它重新计算相同的每个子系统稳定性检查以及汇总得分。
cs.AI / 30 / 2606.29175
Direct Causation in International Humanitarian Law and the Challenge of AI-Mediated Civilian Cyber Operations
国际人道法中的直接因果关系与人工智能介导的平民网络行动的挑战
Abstract
International humanitarian law protects civilians from direct attack unless and for such time as they take direct part in hostilities, with the ICRC's 2009 Interpretive Guidance operationalising this rule through a three-criterion cumulative test. This paper argues that AI-mediated civilian cyber operations challenge the direct causation element of this test in a structurally specific way: when a civilian deploys an autonomous multi-agent cyber system of the kind recently demonstrated in offensive AI research, the "one causal step" standard fails because harm is produced by system-generated decisions made after human disengagement, and the integral-part requirement does not extend because it presupposes downstream human contributors whose conduct can be independently classified. The framework therefore defaults to treating such deployments as indirect participation, in tension with its purpose of capturing civilians who personally take part in hostilities. Beyond the doctrinal analysis, this paper identifies goal-specification granularity as the property on which the integral-part test's concreteness component implicitly turns, classifies AI-mediated operations along a five-level spectrum, and argues that existing technical AI governance instruments do not log or report this property.
Chinese Translation
国际人道法保护平民免受直接攻击，除非在其直接参与敌对行动的情况下，并且国际红十字委员会（ICRC）在2009年的解释性指导中通过三项标准的累积测试对这一规则进行了操作化。本文认为，人工智能介导的平民网络行动以结构特定的方式挑战了该测试的直接因果关系要素：当平民部署一种最近在进攻性人工智能研究中展示的自主多智能体网络系统时，“单一因果步骤”标准失效，因为伤害是由人类脱离后系统生成的决策造成的，而整体参与要求并不适用，因为它假设了可以独立分类的下游人类贡献者的行为。因此，该框架默认将此类部署视为间接参与，这与其捕捉亲自参与敌对行动的平民的目的相悖。除了教义分析外，本文还将目标规范的细粒度识别为整体参与测试的具体性要素隐含转向的属性，将人工智能介导的操作分类为五个级别的光谱，并认为现有的技术人工智能治理工具并未记录或报告这一属性。
cs.AI / 31 / 2606.29178
Selective Memory Retention for Long-Horizon LLM Agents
长时程 LLM 代理的选择性记忆保留
Abstract
When does retention matter for memory-augmented LLM agents? We study this with TraceRetain, a lightweight framework for bounded external memory in frozen LLM agents that scores entries by interpretable features (success, age, access frequency, redundancy, specificity, similarity, downstream utility) and evicts the lowest-scoring ones at capacity. On clean ALFWorld with gpt-5-mini, external memory robustly improves over no memory across two seeds, but differences among bounded retention policies fall within Wilson 95% CIs: clean ALFWorld at T=100 to T=200 does not naturally exhibit the memory pollution retention is designed to address. Under a controlled noisy-write stress (75% synthetic distractors), unbounded memory and FIFO-K50 degrade on Precision@5 (20.2% to 12.4% and 15.8% to 3.8%) while TraceRetain-CEM is essentially unchanged (16.9% to 16.6%) and preserves 97/100 task success. The mechanism: unbounded memory has the highest mean similarity (0.87) but lowest precision, indicating failed distractors close to the query in embedding space. Held-out in-distribution evaluation shows memory-augmented policies solving 47 to 49 of 50 tasks vs. 39/50 for no memory. Bounded retention buys memory and step efficiency on saturated clean benchmarks at no task-success cost, and only differentiates from cache heuristics when streams contain noise.
Chinese Translation
记忆增强的 LLM 代理在何时保留记忆是重要的？我们通过 TraceRetain 进行研究，这是一种轻量级框架，用于在冻结的 LLM 代理中实现有限的外部记忆，该框架通过可解释特征（成功、年龄、访问频率、冗余性、特异性、相似性、下游效用）对条目进行评分，并在达到容量时驱逐得分最低的条目。在干净的 ALFWorld 环境中使用 gpt-5-mini，外部记忆在两个种子上稳健地优于无记忆，但有限保留策略之间的差异在 Wilson 95% 置信区间内：在 T=100 到 T=200 的干净 ALFWorld 中，自然并未表现出记忆污染，而保留正是为了解决这一问题。在控制的噪声写入压力（75% 合成干扰项）下，无限制记忆和 FIFO-K50 在 Precision@5 上下降（从 20.2% 降至 12.4% 和从 15.8% 降至 3.8%），而 TraceRetain-CEM 基本保持不变（从 16.9% 降至 16.6%），并保留了 97/100 的任务成功率。机制是：无限制记忆具有最高的平均相似性（0.87），但精度最低，表明在嵌入空间中与查询接近的干扰项失败。保留的分布内评估显示，记忆增强策略解决了 50 个任务中的 47 到 49 个，而无记忆策略仅解决了 39/50 个。有限保留在饱和的干净基准上以零任务成功成本购买了记忆和步骤效率，并且仅在流中包含噪声时与缓存启发式有所区分。
cs.AI / 32 / 2606.29180
Measuring Graph-to-Graph Semantic Similarity in Knowledge Graphs: An Empirical Evaluation of Knowledge Graph Embeddings
知识图谱中图对图语义相似性的测量：知识图谱嵌入的实证评估
Abstract
A Knowledge Graph (KG) represents facts as structured triples and is widely used to organize relational knowledge across diverse domains. Just as textual information ranges from words and sentences to complete documents, KG information can be interpreted at multiple levels, from entities, relations, and triples to subgraphs and entire KGs. However, existing KG embedding methods mainly focus on entities, relations, and triples, leaving graph-level semantics largely unaddressed. Conventional graph-level methods, which typically compare graphs based on structural patterns, are also insufficient because structural similarity alone cannot guarantee semantic similarity between KGs. To evaluate how well different methods capture such graph-level semantic information, we study graph-to-graph semantic similarity, which determines whether a pair of KGs represents semantically corresponding underlying information. To obtain reliable ground-truth correspondences, we construct a semantic matching dataset by modifying text documents, extracting KGs from both original and modified documents, and transferring their known correspondences to KG pairs. We compare text-based, structure-based, and KG embedding-based approaches on each dataset. For the KG embedding-based approach, we introduce two scoring functions: \textit{EmbPairSim}, which uses maximal pairwise entity similarity, and \textit{AvgEmbSim}, which uses a frequency-weighted centroid. Experiments on WikiText-2 and CC-News show that \textit{EmbPairSim} achieves up to 5.3 pp higher MRR than Sentence-BERT while using substantially fewer parameters. These results suggest that KGE representations can serve as compact and effective signals for graph-to-graph semantic similarity in KGs. Our code is available at https://github.com/SeungRyeolBaek/KG-to-KG-Semantic-Similarity.
Chinese Translation
知识图谱（KG）将事实表示为结构化的三元组，并广泛用于组织各个领域的关系知识。正如文本信息从单词和句子到完整文档的范围一样，KG信息可以在多个层次上进行解释，从实体、关系和三元组到子图和整个KG。然而，现有的KG嵌入方法主要关注实体、关系和三元组，图级语义在很大程度上未得到解决。传统的图级方法通常基于结构模式比较图形，但仅凭结构相似性无法保证KG之间的语义相似性。为了评估不同方法在捕捉图级语义信息方面的效果，我们研究了图对图的语义相似性，这决定了一对KG是否表示语义上对应的潜在信息。为了获得可靠的真实对应关系，我们通过修改文本文档构建了一个语义匹配数据集，从原始文档和修改后的文档中提取KG，并将其已知的对应关系转移到KG对中。我们在每个数据集上比较了基于文本、基于结构和基于KG嵌入的方法。对于基于KG嵌入的方法，我们引入了两个评分函数： extit{EmbPairSim}，它使用最大成对实体相似性，以及 extit{AvgEmbSim}，它使用频率加权的中心点。在WikiText-2和CC-News上的实验表明， extit{EmbPairSim}的平均排名率（MRR）比Sentence-BERT高出最多5.3个百分点，同时使用的参数显著更少。这些结果表明，KG嵌入表示可以作为KG中图对图语义相似性的紧凑有效信号。我们的代码可在https://github.com/SeungRyeolBaek/KG-to-KG-Semantic-Similarity获取。
cs.AI / 33 / 2606.29182
Evidence-Informed LLM Beliefs for Continual Scientific Discovery
基于证据的 LLM 信念在持续科学发现中的应用
Abstract
Open-ended scientific discovery with large language models (LLMs) increasingly operates as a long-horizon loop of hypothesis search and verification, where a reward signal guides which hypotheses to test next. A notable recent example is AutoDiscovery, which uses "Bayesian surprise" - the belief shift an LLM undergoes after observing evidence for a hypothesis - as both a discovery metric and a reward for search. We first observe that AutoDiscovery treats surprisal as a static quantity, while surprisal in human reasoning is non-stationary - it is defined relative to beliefs that evolve with experience, a prerequisite for continual scientific discovery. We address this mismatch with evidence-informed LLM beliefs: priors updated with evidence from previous hypotheses to compute non-stationary surprisal for new hypotheses. We compare in-context belief-updating mechanisms and find that embedding-based retrieval-augmented generation over prior discoveries best anticipates eventual posteriors, identifying 37.5% of static surprisals as spurious. We then modify search to avoid these spurious rewards and prioritize hypotheses that remain surprising under non-stationary beliefs. Concretely, we introduce two complementary changes to the original search procedure: belief-update filtering and diversity maximization. Across five discovery domains, our method increases accumulated non-stationary surprisal by 30.62% on average compared to the original search procedure, demonstrating that continual scientific discovery with LLMs requires not only better belief measurement but also search procedures that avoid redundancy and encourage diversity.
Chinese Translation
使用大型语言模型（LLMs）进行开放式科学发现日益成为一个长期的假设搜索与验证循环，其中奖励信号指导下一步测试的假设。一个显著的近期例子是 AutoDiscovery，它将“贝叶斯惊讶”（Bayesian surprise）—— LLM 在观察到假设证据后所经历的信念转变——作为发现指标和搜索奖励。我们首先观察到，AutoDiscovery 将惊讶视为静态量，而人类推理中的惊讶是非平稳的——它是相对于随着经验演变的信念而定义的，这是持续科学发现的前提。我们通过基于证据的 LLM 信念来解决这一不匹配：使用来自先前假设的证据更新的先验，以计算新假设的非平稳惊讶。我们比较了上下文中的信念更新机制，发现基于嵌入的检索增强生成（retrieval-augmented generation）在先前发现的基础上最好地预测最终后验，识别出 37.5% 的静态惊讶为虚假。然后，我们修改搜索以避免这些虚假奖励，并优先考虑在非平稳信念下仍然令人惊讶的假设。具体而言，我们对原始搜索程序引入了两个互补的变化：信念更新过滤和多样性最大化。在五个发现领域中，我们的方法使累积的非平稳惊讶平均增加了 30.62%，证明了使用 LLM 进行持续科学发现不仅需要更好的信念测量，还需要避免冗余并鼓励多样性的搜索程序。
cs.AI / 34 / 2606.29194
AI Trading's Alpha Singularity: Emergent Market Reasoning through Agent-to-Agent Self-Evolution
人工智能交易的阿尔法奇点：通过代理间自我演化的市场推理
Abstract
Automated alpha mining holds the scoring function fixed and varies the search algorithm over it. A search that converges against a fixed scorer overfits whatever the scorer cannot penalize, a primary cause of the out-of-sample generalization gap. We treat the scoring function as a search artifact alongside the alpha factors and study what conditions make this joint search admissible. Sealed Joint Search (SJS) is a framework: a set of structural conditions on information flow in an autonomous-discovery system that prevent joint search from collapsing into self-confirmation while keeping the evaluator sealed. Conditions cover role decomposition, typed inter-role communication, provenance-sealed reads, versioned stores, and substrate-local promotion. Agora tests SJS empirically: five LLM agent classes communicate via three channels, evolving eight skill libraries, with alpha libraries built on AlphaGen operators. Three evaluators write reports aggregated into one brief, carrying forward disagreement instead of voting. We run Agora for 100 rounds on CSI 1000 and evaluate on a 91-day 2026 holdout sealed from all LLM inputs. Agora achieves holdout Sharpe +1.87; best baseline +1.334 at favorable seed and -0.755 cross-seed mean. Pre-loading Agora's two metrics into a frozen-library ablation recovers only +0.40 of the +2.25 Sharpe gap, and adding PPO without library evolution worsens the gap. The two metrics emerge rather than being designed. Caveats: single-seed run, short-side concentrated signal, intended for long-short.
Chinese Translation
自动化阿尔法挖掘保持评分函数不变，并在其上变化搜索算法。针对固定评分者的收敛搜索会过拟合评分者无法惩罚的内容，这是导致样本外泛化差距的主要原因。我们将评分函数视为与阿尔法因子共同的搜索产物，并研究使这种联合搜索可接受的条件。密封联合搜索（Sealed Joint Search, SJS）是一个框架：一组关于信息流的结构性条件，适用于自主发现系统，防止联合搜索陷入自我确认，同时保持评估者的密封。条件包括角色分解、类型化的跨角色通信、来源密封读取、版本化存储和基质本地推广。Agora 实证测试 SJS：五个大型语言模型（LLM）代理类通过三个通道进行通信，演化出八个技能库，阿尔法库基于 AlphaGen 操作符构建。三个评估者撰写报告，并汇总成一份简报，传递分歧而非投票。我们在 CSI 1000 上运行 Agora 100 轮，并在一个与所有 LLM 输入隔离的 91 天 2026 保留样本上进行评估。Agora 实现了保留样本夏普比率 +1.87；最佳基线为 +1.334，具有有利种子和 -0.755 的交叉种子均值。将 Agora 的两个指标预加载到冻结库消融中，仅恢复了 +0.40 的 +2.25 夏普差距，且在没有库演化的情况下添加 PPO 使差距加剧。这两个指标是自发出现的，而非设计出来的。注意事项：单种子运行，短侧集中信号，旨在进行多头-空头交易。
cs.AI / 35 / 2606.29212
A Cognition-Emotion-Personality Framework for Modeling Human-Like Awareness and Behavior in Emergency Evacuations
用于建模人类意识和行为的认知-情感-个性框架在紧急疏散中的应用
Abstract
Agent-based evacuation simulations are widely used to study crowd behavior during emergencies, but many models rely on assumptions such as perfect event awareness, complete exit knowledge, and fully rational decision-making. This paper presents an extended evacuation framework that integrates cognitive, emotional, social, and personality-related mechanisms into a unified model of human behavior under uncertainty. The framework incorporates a dynamic event-awareness mechanism based on a continuous Event Certainty Level, a memory-based representation of exit knowledge subject to acquisition, forgetting, and recall, a continuous fear model in which panic emerges as a high-intensity state, and an OCEAN-based personality representation. Neuroticism is explicitly integrated into the emotional model, influencing fear generation, escalation, social contagion, and recovery. Behavioral heterogeneity is further captured through individualized decision thresholds that affect responses to perceived risk. The framework is evaluated through simulation experiments examining the effects of spatial familiarity, memory robustness, decision sensitivity, emotional dynamics, and personality variation. Results show that cognitive, emotional, and personality-driven processes substantially influence evacuation dynamics, reducing evacuation efficiency and generating realistic crowd phenomena such as delays, confusion, injuries, and socially influenced behaviors. The proposed framework provides a more realistic representation of human behavior in emergency evacuations and supports systematic investigation of the interactions between cognition, emotion, personality, and crowd dynamics.
Chinese Translation
基于代理的疏散模拟广泛用于研究紧急情况下的人群行为，但许多模型依赖于诸如完美事件意识、完全出口知识和完全理性决策等假设。本文提出了一种扩展的疏散框架，将认知、情感、社会和个性相关机制整合到一个统一的人类行为模型中，以应对不确定性。该框架结合了基于连续事件确定性水平的动态事件意识机制、受获取、遗忘和回忆影响的出口知识的记忆基础表示、高强度状态下出现恐慌的连续恐惧模型，以及基于OCEAN（开放性、责任心、外向性、宜人性、神经质）的个性表示。神经质被明确整合到情感模型中，影响恐惧的产生、升级、社会传播和恢复。通过个性化的决策阈值进一步捕捉行为异质性，这些阈值影响对感知风险的反应。通过模拟实验评估该框架，考察空间熟悉度、记忆稳健性、决策敏感性、情感动态和个性变异的影响。结果表明，认知、情感和个性驱动的过程显著影响疏散动态，降低疏散效率，并产生现实的人群现象，如延误、混乱、受伤和社会影响行为。所提出的框架提供了对紧急疏散中人类行为更为真实的表征，并支持对认知、情感、个性与人群动态之间相互作用的系统性研究。
cs.AI / 36 / 2606.29225
PolicyGuard: A Dialogue-Grounded Sub-Agent Verifier for Policy Adherence in LLM Agents
PolicyGuard：一种基于对话的子代理验证器，用于大语言模型代理的政策遵循
Abstract
LLM agents handle user requests on behalf of organizations through tool calls and must follow the company policies stated in their system prompts. Prior work approaches this as a safeguarding problem -- external checks that block non-compliant agent actions. We argue that policy adherence is a broader problem: real workflows unfold across many turns, require explicit user confirmation and prerequisite reads, and hinge on the content of the dialogue rather than on any single argument value. Meeting this bar requires (i) full conversation context, (ii) self-reasoning over the policy and the current dialogue, and (iii) conversation-specific remediation that guides the agent's next turn -- three capabilities that prior safeguard work has often underestimated. We introduce POLICYGUARD, a sub-agent verifier that shares the agent's view of the dialogue, reasons over the policy in context, and provides actionable feedback for the agent's next turn. On tau^2-BENCH airline across three vendors (GPT-5.4, Claude Sonnet 4.6, Gemini 2.5 Pro) with four trials per setting, POLICYGUARD improves PASS4 by +12.0 / +6.0 / +12.0 pp. Per-call analyses show POLICYGUARD achieves higher policy-violation recall while blocking roughly half as often as argument-level guards.
Chinese Translation
大语言模型（LLM）代理通过工具调用代表组织处理用户请求，必须遵循其系统提示中规定的公司政策。之前的研究将此视为一种保护问题——外部检查阻止不合规的代理行为。我们认为政策遵循是一个更广泛的问题：真实的工作流程在多个回合中展开，需要明确的用户确认和先决条件的阅读，并且依赖于对话的内容，而不是任何单一的参数值。达到这一标准需要（i）完整的对话上下文，（ii）对政策和当前对话的自我推理，以及（iii）针对特定对话的补救措施，以指导代理的下一个回合——这三种能力在之前的保护工作中往往被低估。我们引入了POLICYGUARD，一种子代理验证器，它共享代理的对话视图，在上下文中对政策进行推理，并为代理的下一个回合提供可操作的反馈。在tau^2-BENCH航空公司对三个供应商（GPT-5.4、Claude Sonnet 4.6、Gemini 2.5 Pro）进行的四次试验中，POLICYGUARD将PASS4提高了+12.0 / +6.0 / +12.0个百分点。逐次调用分析显示，POLICYGUARD在阻止大约一半的情况下实现了更高的政策违规召回率，而相较于基于参数的保护措施，其阻止频率显著降低。
cs.AI / 37 / 2606.29247
SurgVLA-Bench: Towards Evaluating Vision-Language-Action Models for Laparoscopic Surgical Robotics
SurgVLA-Bench：评估腹腔镜外科机器人视觉-语言-行动模型的进展
Abstract
Vision-Language-Action (VLA) models represent a promising direction for embodied intelligence in surgical robotics. Despite the prevalence of VLA benchmarks for general robotics, standardized evaluation platforms specifically designed for surgical contexts remain absent. To address this limitation, we present SurgVLA-Bench, the first comprehensive benchmark for evaluating VLA models in laparoscopic surgical robotics. Leveraging the SurRoL simulation platform, we construct a hierarchical task taxonomy ranging from atomic actions to complete surgical procedures, complemented by a multi-dimensional evaluation framework assessing action accuracy and semantic consistency. We then systematically evaluate two representative paradigms, including autoregressive models such as OpenVLA, and flow matching models such as $\pi_{0}$, $\pi_{0.5}$, and SmolVLA. Our experiments show that autoregressive models tend to excel in semantic understanding, while flow matching models often achieve higher task precision but may face generalization trade-offs. However, even the best-performing models remain far from satisfactory, as the constrained endoscopic field of view, restricted viewing angles, and frequent occlusions persist as fundamental physical bottlenecks. The code and data are available at https://github.com/VCL-HNU/SurgVLA
Chinese Translation
视觉-语言-行动（VLA）模型代表了外科机器人领域中具身智能的一个有前景的方向。尽管针对一般机器人领域的 VLA 基准测试已相当普遍，但专门为外科环境设计的标准化评估平台仍然缺乏。为了解决这一局限性，我们提出了 SurgVLA-Bench，这是第一个用于评估腹腔镜外科机器人 VLA 模型的综合基准。我们利用 SurRoL 模拟平台构建了一个从原子动作到完整外科程序的分层任务分类法，并辅以一个多维评估框架，以评估动作准确性和语义一致性。随后，我们系统地评估了两种具有代表性的范式，包括自回归模型（如 OpenVLA）和流匹配模型（如 $ ext{π}_{0}$、$ ext{π}_{0.5}$ 和 SmolVLA）。实验结果表明，自回归模型在语义理解方面表现优越，而流匹配模型通常在任务精度上更高，但可能面临泛化权衡。然而，即使是表现最好的模型也远未令人满意，因为受限的内窥镜视野、受限的视角和频繁的遮挡仍然是基本的物理瓶颈。代码和数据可在 https://github.com/VCL-HNU/SurgVLA 获取。
cs.AI / 38 / 2606.29251
When Summaries Distort Decisions: Information Fidelity in LLM-Compressed Financial Analysis
当摘要扭曲决策时：大语言模型压缩金融分析中的信息保真度
Abstract
Financial decision-makers face more information than they can directly inspect, making context compression necessary. Yet when large language models (LLMs) compress financial source material, they can alter the investment judgment supported by the original source. We frame this problem as information fidelity: compression loses fidelity when it changes the decision induced by the source. In agentic systems, such losses may recur across intermediate steps and amplify throughout the decision process. Across financial filings and earnings-call transcripts, we find that LLM-based compression can produce fluent and factually plausible compressed contexts that nevertheless alter downstream decisions. We analyze two diagnostic patterns associated with fidelity loss: decontextualization, where salient evidence is retained but separated from the caveats and contextual qualifiers needed for correct interpretation, and model dependency, where different compressors expose different views of the same source. We then propose Agentic Context Compression, which generates multiple candidate compressions and audits their disagreements against the original source. Our results suggest that financial compression should be evaluated not only by efficiency or factuality, but also by its ability to preserve decision-relevant context.
Chinese Translation
金融决策者面临的信息量超过了他们能够直接检查的范围，因此需要进行上下文压缩。然而，当大语言模型（LLMs）压缩金融源材料时，它们可能会改变原始来源所支持的投资判断。我们将这个问题框架定义为信息保真度：当压缩改变了源材料所引导的决策时，保真度就会丧失。在代理系统中，这种损失可能在中间步骤中反复出现，并在决策过程中逐渐放大。在对金融文件和财报电话会议记录的分析中，我们发现基于LLM的压缩可以生成流畅且在事实上似乎合理的压缩上下文，但仍然改变了下游决策。我们分析了与保真度损失相关的两种诊断模式：去上下文化，即保留显著证据但与正确解释所需的警告和上下文限定条件分离，以及模型依赖性，即不同的压缩器对同一来源呈现不同的视角。随后，我们提出了代理上下文压缩（Agentic Context Compression），该方法生成多个候选压缩并审计它们与原始来源之间的分歧。我们的结果表明，金融压缩的评估不仅应考虑效率或事实性，还应考虑其保留决策相关上下文的能力。
cs.AI / 39 / 2606.29278
The Complexity Ceiling Benchmark: A Multi-Domain Evaluation of Sequential Reasoning Under Depth Scaling
复杂性上限基准：深度扩展下顺序推理的多领域评估
Abstract
We introduce the Complexity Ceiling Benchmark (CCB), a controlled evaluation of how language-model reasoning decays as the number of required sequential steps grows. CCB fixes the semantic content of a task and varies only its depth N in {5,...,50} across three structurally distinct regimes: grounded spatial state-tracking, abstract symbolic pointer manipulation, and transitive relational inference. Across 6,000 trials over five frontier and open-weight LLMs we find a consistent pattern of geometric per-step decay with widely separated domain ceilings: on the first two regimes the strongest models retain pd>0.92 across N=50; on the third every model collapses by N=5, with the best model's 50%-success horizon at H0.5~4.7 steps despite pd=0.863. A trace-level metric (TFBC) shows that 14.5% of correct answers across the benchmark are reached via incorrect intermediate reasoning. Forced verbose state-tracking does not move the ceiling (McNemar p=1.000), and the mean step at which reasoning first diverges, k*, predicts within-domain accuracy better than parameter count. CCB and the geometric decay model together reduce a model's long-horizon reasoning profile to one interpretable number per task family.
Chinese Translation
我们提出了复杂性上限基准（Complexity Ceiling Benchmark，CCB），这是一个控制评估，旨在研究随着所需顺序步骤数量的增加，语言模型推理的衰退情况。CCB 固定了任务的语义内容，仅在 {5,...,50} 的深度 N 上变化，涵盖三种结构上不同的领域：基于位置的空间状态跟踪、抽象符号指针操作和传递关系推理。在对五个前沿和开放权重的 LLM 进行的 6,000 次试验中，我们发现了一种一致的几何每步衰退模式，且各领域的上限差异显著：在前两个领域中，最强模型在 N=50 时保持 pd>0.92；而在第三个领域中，每个模型在 N=5 时均崩溃，最佳模型的 50% 成功边界为 H0.5~4.7 步，尽管 pd=0.863。一个追踪级别的指标（TFBC）显示，在基准测试中，14.5% 的正确答案是通过错误的中间推理得出的。强制的冗长状态跟踪并未提高上限（McNemar p=1.000），而推理首次偏离的平均步骤 k*，在领域内准确性预测上优于参数数量。CCB 和几何衰退模型共同将模型的长时间推理特征简化为每个任务家族一个可解释的数字。
cs.AI / 40 / 2606.29296
Process Advantage Signal Shaping: A Paradigm-Agnostic Middleware for Process-Supervised RL in LLM Reasoners
过程优势信号塑形：一种与范式无关的中间件，用于 LLM 推理器中的过程监督强化学习
Abstract
Group Relative Policy Optimization (GRPO) is a default recipe for process-supervised reinforcement learning of LLM reasoners, and dense process supervision -- via learned process reward models (PRMs) or on-policy-distillation KL signals -- is a common way to densify its otherwise weak outcome reward. Layering such a step-level signal on top of GRPO's group-standardized advantage, however, exposes three structural pathologies: \emph{channel contamination} between the pooled process, outcome, and format streams at group standardization; \emph{resolution mismatch} between the granularity of the process signal and the granularity of the logical decisions being credited; and a \emph{cumulative trap} by which GRPO's return-to-go sum surfaces either length inflation or truncated exploration depending on the sign regime of the signal. We propose \textbf{PASS} (\emph{Process Advantage Signal Shaping}), a compact middleware that sits between any scalar step-level process signal and GRPO's clipped surrogate and addresses the three pathologies in turn: \emph{Advantage Fusion} standardizes the three streams independently within each group, \emph{Chunk-by-Value} derives value-homogeneous chunks from the signal itself and broadcasts credit within each chunk, and \emph{Divide-Length} converts the cumulative objective into an average-value-density score. We validate PASS across two domains and two process-signal paradigms -- a learned PRM on mathematical reasoning and an on-policy-distillation KL signal (with a generalized variant) on multi-hop question answering -- and under two group-standardization operators. In every regime PASS delivers a consistent pass@1 gain over the corresponding GRPO baseline.
Chinese Translation
群体相对策略优化（GRPO）是 LLM 推理器过程监督强化学习的默认方法，而通过学习的过程奖励模型（PRMs）或策略内蒸馏 KL 信号进行的密集过程监督，是增强其原本较弱的结果奖励的常见方式。然而，将这种逐步信号叠加在 GRPO 的群体标准化优势之上，暴露出三种结构性病态：群体标准化时的 extit{通道污染}，即池化过程、结果和格式流之间的相互影响；过程信号的粒度与被认可的逻辑决策粒度之间的 extit{分辨率不匹配}；以及 extit{累积陷阱}，即 GRPO 的回报总和表面上会因信号的符号范围而导致长度膨胀或截断探索。我们提出了 extbf{PASS}（过程优势信号塑形），这是一种紧凑的中间件，位于任何标量逐步过程信号与 GRPO 的剪切替代品之间，依次解决这三种病态： extit{优势融合}在每个群体内独立标准化三种流， extit{按值分块}从信号本身推导出值同质块并在每个块内广播信用， extit{划分长度}将累积目标转换为平均值密度评分。我们在两个领域和两种过程信号范式下验证了 PASS——在数学推理上的学习 PRM 和在多跳问答上的策略内蒸馏 KL 信号（带有广义变体）——并在两种群体标准化操作下进行测试。在每种情况下，PASS 都在相应的 GRPO 基线之上提供了一致的 pass@1 增益。
cs.AI / 41 / 2606.29315
Hierarchical Experimentalist Agents
层次实验主义者代理
Abstract
Large language models (LLMs) are increasingly used to take actions in the real world and support human decision-making, yet most agents rely on parametric knowledge, fixed post-training data, retrieval, or search. This paradigm breaks down in novel domains and for sophisticated queries that cannot be answered from prior knowledge alone. Knowing the laws of physics, for instance, does not by itself enable LLMs to answer queries or complete long-horizon tasks in a complex physical system. To address this, we introduce Hierarchical Experimentalist Agents (HExA), an in-context self-improvement framework to learn from active experimentation. HExA iteratively designs and refines query-relevant experiments, learns a reusable library of composable skills from experience, and integrates experimental evidence to answer queries or take actions. HExA is training-free, compatible with any black-box model, and does not require external supervision, oracles, or offline data. To evaluate active experimentation, we introduce Interphyre, a tool-calling benchmark built on the PHYRE 2D procedural physics environment, where agents propose interventions and test hypotheses through simulation APIs. Experiments show that current LLM agents struggle in these settings, especially on the hardest levels of Interphyre. Claude Sonnet 4.6 achieves only 2% success, while HExA improves the same model to up to 77% success. HExA also improves open-weight models and outperforms agentic baselines such as ReAct and Reflexion. Moreover, using only skills learned from easier levels and transferred without active experimentation, HExA achieves 44% success, demonstrating the reusability and generalization of its learned skills. Overall, HExA shows that learning through active experimentation can help agents discover useful knowledge, acquire reusable skills, and make efficient progress on novel long-horizon tasks.
Chinese Translation
大型语言模型（LLMs）越来越多地被用于在现实世界中采取行动并支持人类决策，然而大多数代理依赖于参数知识、固定的后训练数据、检索或搜索。这一范式在新领域和无法仅凭先前知识回答的复杂查询中会失效。例如，了解物理定律本身并不能使LLMs回答查询或在复杂物理系统中完成长期任务。为了解决这一问题，我们提出了层次实验主义者代理（HExA），这是一种通过主动实验进行自我改进的框架。HExA迭代地设计和优化与查询相关的实验，从经验中学习可重用的可组合技能库，并整合实验证据以回答查询或采取行动。HExA不需要训练，兼容任何黑箱模型，并且不需要外部监督、神谕或离线数据。为了评估主动实验，我们引入了Interphyre，这是一个基于PHYRE 2D程序物理环境的工具调用基准，其中代理通过仿真API提出干预并测试假设。实验表明，当前的LLM代理在这些设置中表现不佳，尤其是在Interphyre的最难关卡中。Claude Sonnet 4.6的成功率仅为2%，而HExA将同一模型的成功率提高至77%。HExA还改善了开放权重模型，并超越了如ReAct和Reflexion等代理基线。此外，仅使用从较简单关卡学习的技能，并在没有主动实验的情况下转移，HExA的成功率达到了44%，展示了其学习技能的可重用性和泛化能力。总体而言，HExA表明，通过主动实验学习可以帮助代理发现有用知识、获得可重用技能，并在新颖的长期任务上高效进展。
cs.AI / 42 / 2606.29340
PHF: Privileged Hidden Flow for On-Policy Self-Distillation
PHF：用于在线自蒸馏的特权隐流
Abstract
On-policy self-distillation (OPSD) trains a reasoning model on rollouts sampled from its own policy by matching a privileged teacher that also sees verified reference solutions. Existing OPSD objectives supervise only the output distribution, so privileged context affects training through a token-level divergence without directly supervising the internal computation that produced that distribution. We propose Privileged Hidden Flow (PHF), which additionally distills how a privileged teacher's hidden states move along the same rollout. Rather than forcing each student hidden vector to match the teacher vector at the same token position, PHF aligns token-to-token transition directions and trajectory geometry over selected generated positions. The all-layer recipe also includes an adjacent-layer relation computed from these same transitions, without pointwise hidden-state imitation. Under the same 100-step training schedule, PHF improves the Average@12 aggregate over our reproduced OPSD baseline on Qwen3-1.7B, 4B, and 8B, with observed gains of about +2.2, +1.5, and +1.7 points. The transport objective is exactly invariant to shared trajectory offsets; its local geometry term is also invariant to orthogonal transformations of transition directions. Ablations distinguish the fixed PHF recipe from pointwise hidden-state matching, single-channel transition losses, and layer-subset choices, supporting PHF as a compact hidden-flow extension to OPSD.
Chinese Translation
在线自蒸馏（OPSD）通过匹配一个特权教师来训练推理模型，该教师也能看到经过验证的参考解决方案，并从其自身策略中采样的回合进行训练。现有的OPSD目标仅监督输出分布，因此特权上下文通过令牌级别的发散影响训练，而没有直接监督产生该分布的内部计算。我们提出了特权隐流（PHF），它额外蒸馏特权教师的隐状态如何沿着相同的回合移动。PHF并不是强制每个学生隐向量在相同令牌位置上与教师向量匹配，而是对选定生成位置上的令牌到令牌的过渡方向和轨迹几何进行对齐。全层配方还包括从这些相同过渡计算的相邻层关系，而不进行逐点隐状态模仿。在相同的100步训练计划下，PHF在Qwen3-1.7B、4B和8B上提高了我们复现的OPSD基线的Average@12聚合，观察到的增益约为+2.2、+1.5和+1.7分。传输目标对共享轨迹偏移完全不变；其局部几何项也对过渡方向的正交变换不变。消融实验将固定的PHF配方与逐点隐状态匹配、单通道过渡损失和层子集选择区分开，支持PHF作为OPSD的紧凑隐流扩展。
cs.AI / 43 / 2606.29354
When LLMs Develop Languages: Symbolic Communication for Efficient Multi-Agent Reasoning
当大语言模型发展语言时：高效多智能体推理的符号通信
Abstract
Chain-of-Thought (CoT) improves large language models (LLMs) on difficult reasoning tasks, but it often incurs long natural-language rationales that are poorly aligned with efficient machine reasoning. We propose Communicative Language Symbolism Routing (CLSR), a test-time framework in which multiple LLM agents autonomously invent, evolve, and share compact Language Symbolism Frameworks (LSFs), while a latent-free router adaptively selects and composes these languages per query to optimize the accuracy-token trade-off. Unlike prompt optimization that refines surface instructions, CLSR treats each LSF as a reusable symbolic protocol with compact symbols, usage rules, and a message-passing contract, and improves it through an evolutionary loop driven by correctness and token cost. At inference time, the router may invoke a single low-cost LSF call, ensemble multiple LSFs, or execute a multi-round LSF composition protocol on harder queries. Across challenging benchmarks, CLSR reduces latency-oriented generated token completion by $3\sim 6\times$ compared to standard CoT while maintaining accuracy. We further derive an information-theoretic lower bound on token cost under arbitrary symbolism and show that, under an interpreter-realizability premise, multi-round LSF protocols conditionally subsume program-execution pipelines. Code is publicly available (https://github.com/pzqpzq/LSF_MDia).
Chinese Translation
链式思维（Chain-of-Thought, CoT）在困难推理任务中提升了大语言模型（Large Language Models, LLMs）的表现，但它通常会产生与高效机器推理不太一致的冗长自然语言推理。我们提出了交流语言符号路由（Communicative Language Symbolism Routing, CLSR），这是一个在测试时的框架，其中多个LLM智能体自主发明、演变和共享紧凑的语言符号框架（Language Symbolism Frameworks, LSFs），同时一个无潜变量的路由器根据查询自适应地选择和组合这些语言，以优化准确性与令牌成本的权衡。与优化提示（prompt optimization）不同，CLSR将每个LSF视为一个可重用的符号协议，具有紧凑的符号、使用规则和消息传递合约，并通过一个由正确性和令牌成本驱动的演化循环来改进它。在推理时，路由器可以调用单个低成本的LSF，集成多个LSF，或在更复杂的查询上执行多轮LSF组合协议。在具有挑战性的基准测试中，CLSR在保持准确性的同时，将基于延迟的生成令牌完成减少了$3 ext{至}6 imes$，相比于标准的CoT。我们进一步推导了在任意符号下令牌成本的信息论下界，并表明在解释器可实现性前提下，多轮LSF协议有条件地包含程序执行管道。代码已公开可用（https://github.com/pzqpzq/LSF_MDia）。
cs.AI / 44 / 2606.29377
Diagnosing and Repairing Factual Errors in RAG under Budget Constraints
在预算约束下诊断和修复 RAG 中的事实错误
Abstract
Retrieval-Augmented Generation (RAG) improves the factuality of large language models by grounding responses in external evidence, yet real-world deployments remain fragile. Failures often stem from missing or weakly relevant evidence, as well as from generation that does not faithfully reflect the retrieved context. Many existing approaches rely on fine-tuning, privileged access to internal model signals, or resource-insensitive escalation strategies, which limits their practicality in black-box and budget-constrained settings. We propose D2R-RAG (Diagnose-to-Repair RAG), a model-agnostic and resource-aware framework that combines lightweight failure diagnosis with adaptive repair. D2R-RAG derives interpretable failure signatures from observable signals in the query, retrieved evidence, and generated response, and then selects from a small set of corrective actions under explicit latency and VRAM constraints. Experiments on FEVER and HotpotQA show that D2R-RAG improves reliability over recent baselines and achieves better accuracy--efficiency trade-offs across multiple compute budgets. The code is available at https://github.com/CyberScienceLab/D2R-RAG/.
Chinese Translation
检索增强生成（Retrieval-Augmented Generation, RAG）通过将响应基于外部证据来提高大型语言模型的事实性，但在实际应用中仍然脆弱。失败往往源于缺失或相关性较弱的证据，以及生成的内容未能忠实反映检索到的上下文。许多现有方法依赖于微调、对内部模型信号的特权访问或不考虑资源的升级策略，这限制了它们在黑箱和预算受限环境中的实用性。我们提出了 D2R-RAG（Diagnose-to-Repair RAG），这是一个模型无关且资源感知的框架，结合了轻量级的故障诊断与自适应修复。D2R-RAG 从查询、检索到的证据和生成的响应中提取可解释的故障特征，然后在明确的延迟和 VRAM 约束下，从一小组纠正措施中进行选择。在 FEVER 和 HotpotQA 上的实验表明，D2R-RAG 在可靠性方面优于最近的基线，并在多个计算预算下实现了更好的准确性与效率的权衡。代码可在 https://github.com/CyberScienceLab/D2R-RAG/ 获取。
cs.AI / 45 / 2606.29399
LLM-Guided Planning for Multi-hop Reasoning over Multimodal Nuclear Regulatory Documents
基于大型语言模型的多跳推理规划在多模态核监管文件中的应用
Abstract
Reviewing nuclear regulatory documents requires multi-hop reasoning across tens of thousands of pages, where judgments depend on evidence assembled across multiple chapters. We frame this task as planning: an LLM-based agent observes the evidence collected so far, picks the next document fragment to inspect, and stops when the evidence is sufficient. The agent operates over a vectorless document tree using browse, read, and search tools, and maintains a dynamic knowledge graph (KG) as state. On a 200-question benchmark over NuScale Final Safety Analysis Report (FSAR) documents, the system reaches 81.5% accuracy with a RAGAS Faithfulness of 0.93. The dominant performance factor is planning: against PageIndex, which uses the same document tree without state-conditioned action selection, the gap is +38.0pp (43.5% to 81.5%, p<0.001). The system also outperforms LightRAG (73.0%, p<0.05), HippoRAG (70.5%, p<0.01), and GraphRAG (49.5%, p<0.001), and matches RAPTOR (75.5%, p=0.11) without offline indexing. Edge inference adds 2.8x cost without raising accuracy; we retain it as a traceability module. Of 7,391 inferred edges, 3 Violates edges (0.04%) flag scope boundaries (Q058) and partial conformance (Q176) as typed annotations that a human reviewer can audit.
Chinese Translation
审查核监管文件需要在数万页中进行多跳推理，其中判断依赖于跨多个章节汇集的证据。我们将此任务框架化为规划：一个基于大型语言模型（LLM）的代理观察到迄今为止收集的证据，选择下一个要检查的文档片段，并在证据足够时停止。该代理在无向量的文档树上操作，使用浏览、阅读和搜索工具，并维护一个动态知识图谱（KG）作为状态。在针对NuScale最终安全分析报告（FSAR）文档的200个问题基准测试中，该系统的准确率达到81.5%，RAGAS忠实度为0.93。主导性能因素是规划：与使用相同文档树但没有状态条件动作选择的PageIndex相比，差距为+38.0个百分点（从43.5%提升至81.5%，p<0.001）。该系统还优于LightRAG（73.0%，p<0.05）、HippoRAG（70.5%，p<0.01）和GraphRAG（49.5%，p<0.001），并在没有离线索引的情况下与RAPTOR（75.5%，p=0.11）持平。边缘推理增加了2.8倍的成本，但没有提高准确性；我们将其保留作为可追溯性模块。在7391个推断边缘中，3个违反边缘（0.04%）标记了范围边界（Q058）和部分符合性（Q176），作为人类审阅者可以审核的类型注释。
cs.AI / 46 / 2606.29425
Mixture of Debaters: Learn to Debate at Architectural Level in Multi-Agent Reasoning
辩论者混合体：在多智能体推理中学习建筑层面的辩论
Abstract
Existing multi-agent debate frameworks suffer from two critical limitations: they rely on static architectures where agent roles and coordination patterns are fixed at design time, and they require instantiating multiple model copies, incurring substantial computational overhead. We propose Mixture of Debaters (MoD), a unified framework that enables dynamic self-debate within a single model by leveraging the Mixture-of-Experts paradigm. We address three key challenges in adapting MoE for dialectical reasoning: (1) dual-routing that decouples role allocation from process flow, dynamically determining when to debate versus when to synthesize; (2) momentum switching that smooths token-level routing with local context, reducing expert-switch jitter; and (3) unified self-debate that encapsulates diverse debating personas into lightweight expert modules, eliminating inter-agent communication while preserving behavioral diversity. Extensive experiments on multimodal benchmarks demonstrate that MoD outperforms both single-model baselines and conventional multi-agent systems, achieving superior accuracy with 3.7x lower latency and 87% reduction in token consumption.The source code can be accessed at https://github.com/YongLD/MoD.
Chinese Translation
现有的多智能体辩论框架存在两个关键限制：它们依赖于静态架构，其中智能体角色和协调模式在设计时固定，并且需要实例化多个模型副本，从而产生大量计算开销。我们提出了辩论者混合体（Mixture of Debaters, MoD），这是一个统一框架，通过利用专家混合（Mixture-of-Experts）范式，实现了在单一模型内的动态自我辩论。我们解决了将MoE适应于辩证推理的三个关键挑战：（1）双重路由，将角色分配与流程流解耦，动态确定何时辩论与何时综合；（2）动量切换，通过局部上下文平滑令牌级路由，减少专家切换的抖动；（3）统一自我辩论，将多样化的辩论角色封装到轻量级专家模块中，消除智能体间的通信，同时保持行为多样性。在多模态基准上的广泛实验表明，MoD在准确性上优于单模型基线和传统多智能体系统，具有3.7倍更低的延迟和87%的令牌消耗减少。源代码可以在https://github.com/YongLD/MoD访问。
cs.AI / 47 / 2606.29431
FADE: Mitigating Hallucinations by Reducing Language-Prior Dominance in Large Vision-Language Models
FADE：通过减少大型视觉语言模型中的语言先验主导性来减轻幻觉
Abstract
Despite the impressive capabilities of Large Vision-Language Models (LVLMs), they remain susceptible to hallucination, generating content inconsistent with the input image. Recent studies attribute this to the dominance of language priors over visual inputs and employ contrastive decoding methods to mitigate this dominance, but the mechanistic origin remains unexplored. We investigate the information flow through each transformer layer and find that attention modules consistently aggregate visual evidence, while FFN modules at critical layers act as the source of language priors. These priors can override visual evidence, causing correct predictions in intermediate layers to drift toward incorrect outputs. Based on this insight, we propose FADE (FFN Attenuation for DEcoding), a training-free method that attenuates FFN outputs to reduce language-prior dominance. Evaluations on POPE, CHAIR, and MME benchmarks across LLaVA-1.5, mPLUG-Owl2, and InstructBLIP show that FADE effectively mitigates hallucinations while preserving inference efficiency.
Chinese Translation
尽管大型视觉语言模型（LVLMs）具备令人印象深刻的能力，但它们仍然容易出现幻觉，生成与输入图像不一致的内容。近期研究将这一现象归因于语言先验对视觉输入的主导作用，并采用对比解码方法来减轻这种主导性，但其机制来源尚未探讨。我们研究了信息在每个变换器层中的流动，发现注意力模块始终聚合视觉证据，而在关键层的前馈神经网络（FFN）模块则充当语言先验的来源。这些先验可以覆盖视觉证据，导致中间层的正确预测漂移至错误输出。基于这一洞察，我们提出了FADE（前馈神经网络衰减解码），这是一种无训练的方法，通过衰减FFN输出以减少语言先验的主导性。在LLaVA-1.5、mPLUG-Owl2和InstructBLIP的POPE、CHAIR和MME基准测试中的评估表明，FADE有效减轻了幻觉，同时保持了推理效率。
cs.AI / 48 / 2606.29457
How Much Due Diligence Before You Bid? Learning in Intractable Takeover Auctions
竞标前需要多少尽职调查？在难以处理的收购拍卖中的学习
Abstract
When two companies bid to buy the same target, no one knows exactly what the target is worth. Each bidder pays for due diligence: costly, imperfect homework that sharpens its own private estimate before it bids. How much of that homework is worth buying? We build a simple computer model of the bidding contest and let it teach itself to bid well by playing against itself, the way a game engine learns chess. The economic question, how much diligence pays for itself, and the computational question, when the contest becomes too complex to solve exactly, are both controlled by a single thing: how many pieces of private information a bidder carries. Our main finding is that the right amount of diligence is modest and finite. It falls as diligence gets more expensive, and it falls further when both sides are doing their homework, because competition erodes the value of knowing more. We also test a recent claim from AI research: that simple, general self-play methods can rival the specialized, expensive algorithms usually built for games like these. Running on an ordinary laptop with no costly frontier AI, we find the simple methods are the best of the self-learning approaches, though purpose-built exact methods still win whenever the game is small enough to solve outright. The simple methods earn their keep only once the game grows too large to solve exactly, which is the regime real deals live in, and there we show they still find strong bidding strategies. The contribution is threefold: a cheap, reproducible way to study deal-making under uncertainty; a concrete, model-based answer to how much due diligence is worth buying; and evidence about when lightweight, general-purpose AI is good enough to replace specialized methods. We release all the games, code, and experiments.
Chinese Translation
当两家公司竞标购买同一目标时，没有人确切知道该目标的价值。每个竞标者都为尽职调查付出代价：这是一项昂贵且不完美的功课，可以在竞标前提高其私有估计。那么，多少功课是值得购买的呢？我们构建了一个简单的计算机模型来模拟竞标竞赛，并让它通过自我对弈来学习如何出价，类似于游戏引擎学习国际象棋的方式。经济问题是，多少尽职调查能够自我回报；计算问题是，当竞赛变得过于复杂而无法精确解决时，这两者都由一个因素控制：竞标者所掌握的私有信息的数量。我们的主要发现是，适当的尽职调查量是适度且有限的。随着尽职调查成本的增加，这一数量会减少；当双方都在进行尽职调查时，这一数量还会进一步减少，因为竞争削弱了掌握更多信息的价值。我们还测试了来自人工智能研究的一个近期主张：简单的通用自我对弈方法可以与通常为此类游戏构建的专门昂贵算法相媲美。在一台普通的笔记本电脑上运行，无需昂贵的前沿人工智能，我们发现简单方法是自我学习方法中最优秀的，尽管在游戏足够小以便完全解决时，专门构建的精确方法仍然获胜。简单方法只有在游戏变得过于庞大而无法精确解决时才能发挥作用，而这正是现实交易所处的状态，在这种情况下，我们展示了它们仍能找到强有力的竞标策略。我们的贡献有三方面：一种便宜且可重复的方法来研究不确定性下的交易；一个具体的基于模型的答案，说明多少尽职调查是值得购买的；以及关于轻量级通用人工智能何时足够好以替代专门方法的证据。我们发布了所有游戏、代码和实验。
cs.AI / 49 / 2606.29472
Agent-Computer Observation Interfaces Enable Dynamic Computer Use
代理-计算机观察接口实现动态计算机使用
Abstract
SWE-agent established the action interface as an underexplored design axis for software-engineering agents; we make the analogous case for the observation interface in computer-use (CU) agents. Current CU agents, closed and open-source alike, tie observation to action--one screenshot every 3-5 s, no audio--leaving them blind and deaf between screenshots to video, animations, transient UI events, meetings, and spoken instructions. We introduce the Agent-Computer Observation Interface (AOI), a model-agnostic perception layer that decouples continuous, adaptive observation from discrete actions through three gated components: inter-step keyframe capture, volume-gated audio transcription, and CU-model-generated visual narration that persists as text. Each produces almost nothing on static, silent content, reducing to the standard loop without degrading it. On DynaCU-Bench (100 dynamic browser tasks plus a 50-task static control), CU models from 7B to frontier scale gain +17 to +48 pp over their screenshot baselines with zero retraining, turning tasks that are near-impossible from periodic screenshots into largely solved ones. The gap is starkest on audio: on a spoken-content subset AOI agents solve every task, whereas streaming voice models hear accurately but cannot act on what they hear without the scaffold. The decomposition is as informative as the headline gain: keyframe selection turns out not to matter--the value comes from narrating captured frames into persistent text--and the interface is not a fixed bundle, since on a newer model (Gemini 3 Flash) the keyframe stream actively regresses through image-token dilution, so its components must be selected per model rather than shipped as one configuration.
Chinese Translation
SWE-agent 将动作接口确立为软件工程代理的一个未被充分探索的设计轴；我们在计算机使用（CU）代理中提出观察接口的类似案例。目前的 CU 代理，无论是闭源还是开源，都将观察与动作绑定在一起——每 3-5 秒截取一次屏幕截图，没有音频——使它们在屏幕截图之间对视频、动画、瞬态用户界面事件、会议和口头指令失去感知能力。我们引入了代理-计算机观察接口（AOI），这是一个模型无关的感知层，通过三个门控组件将连续的、适应性的观察与离散的动作解耦：步骤间关键帧捕获、音量门控音频转录，以及由 CU 模型生成的持久文本视觉叙述。每个组件在静态、无声内容上几乎不产生任何输出，减少到标准循环而不降低其质量。在 DynaCU-Bench（100 个动态浏览器任务加上 50 个静态控制任务）上，从 7B 到前沿规模的 CU 模型在其屏幕截图基线基础上获得了 +17 到 +48 个百分点的提升，成功将几乎不可能的周期性屏幕截图任务转变为大部分已解决的任务。音频上的差距最为明显：在一个口语内容子集上，AOI 代理解决了每一个任务，而流媒体语音模型虽然能够准确听到，但在没有支架的情况下无法对所听到的内容采取行动。这种分解与头条增益同样具有信息量：关键帧选择似乎并不重要——价值来自于将捕获的帧叙述为持久文本——而且接口并不是一个固定的组合，因为在一个较新的模型（Gemini 3 Flash）上，关键帧流通过图像标记稀释而主动退化，因此其组件必须根据模型进行选择，而不是作为一个配置一起提供。
cs.AI / 50 / 2606.29493
Faults in Our Formal Benchmarking: Dataset Defects and Evaluation Failures in Lean Theorem Proving
我们正式基准测试中的缺陷：Lean 定理证明中的数据集缺陷与评估失败
Abstract
Benchmarks for LLM-assisted theorem proving in Lean are often treated as intrinsically reliable because every solved instance comes with a machine-checked proof. However, the kernel only checks that a proof establishes a \emph{formal} statement; it does not verify that the statement faithfully encodes the intended informal problem, nor that evaluation harnesses are robust to trivial or adversarial solutions. We audit five widely used Lean theorem-proving benchmarks and their forks, using corpus-scale static checkers to surface 4,833 findings, including 398 mechanically certified issues such as counterexamples, vacuous theorems, and unsound axioms. We also document semantic defects such as missing hypotheses, problem simplification, incomplete or incorrect translations, and Lean-specific specification hazards. Beyond dataset construction, we survey evaluation-time failure modes and show, on corrected subsets, that defects can both inflate and deflate reported prover scores. We propose a fault taxonomy, a suite of automated checkers and recall-oriented semantic audit prompts, and release standards to guide the creation of formal math datasets and to make evaluation more reproducible and trustworthy. Our checkers, audit prompts, and corrected dataset snapshots are available at https://github.com/Shashi456/atp-checkers.
Chinese Translation
在 Lean 中，LLM 辅助的定理证明基准通常被视为内在可靠，因为每个已解决的实例都附带一个机器检查的证明。然而，内核仅检查证明是否建立了一个 extit{形式} 语句；它并不验证该语句是否忠实地编码了预期的非正式问题，也不验证评估工具是否对琐碎或对抗性解决方案具有鲁棒性。我们审计了五个广泛使用的 Lean 定理证明基准及其分支，使用大规模静态检查器发现了 4,833 个问题，包括 398 个机械认证的问题，如反例、空洞定理和不可靠公理。我们还记录了语义缺陷，如缺失的假设、问题简化、不完整或不正确的翻译，以及 Lean 特有的规范风险。除了数据集构建之外，我们调查了评估时的失败模式，并在修正的子集上展示了缺陷如何同时膨胀和缩小报告的证明者分数。我们提出了一种缺陷分类法、一套自动检查器和以召回为导向的语义审计提示，以及发布标准，以指导正式数学数据集的创建，并使评估更加可重复和可信。我们的检查器、审计提示和修正后的数据集快照可在 https://github.com/Shashi456/atp-checkers 获取。
cs.AI / 51 / 2606.29495
Cognitive World Models for Process-Level Social Influence Evaluation
过程层面社会影响评估的认知世界模型
Abstract
Social influence dialogue changes user behavior by altering internal cognitive states. The central evaluation question is whether the user's beliefs, desires, intentions, and emotions measurably change over the course of conversation, a process-oriented criterion that neither surface-level text metrics (BLEU/ROUGE) nor single-score LLM judgments can capture. We propose the \textbf{Cog}nitive \textbf{W}orld \textbf{M}odel \textbf{(CogWM)}, an LLM-based user model that reframes multi-turn dialogue evaluation from ``what did the user say'' to ``how did the user's internal cognitive state evolves.'' CogWM jointly predicts BDI/E cognitive states and user utterances and serves as both a user simulator and an evaluation platform, using a three-tier evaluation framework that covers turn-level fidelity, trajectory-level state dynamics, and task-level composite scoring. Trained via our \textbf{S}ummarize-\textbf{a}nd-\textbf{A}llocate \textbf{(SaA)} annotation pipeline on 150,454 user-turn samples across four social influence scenarios, CogWM achieves 77.6\% emotion accuracy (2.1$\times$ over GPT-5.5). In 3600 multi-agent discrimination trials, it distinguishes six commercial agents by their cognitive influence, with Llama-4-Scout ranking first (CTS +0.233). CogWM moves social influence dialogue evaluation from terminal judgment to process tracking. We have released our code\footnote{\scriptsize Code: https://github.com/lucianma05-create/CogWM} and models\footnote{Model: https://www.modelscope.cn/models/LucianMa/CogWM-14B}.
Chinese Translation
社会影响对话通过改变内部认知状态来改变用户行为。核心评估问题是用户的信念、欲望、意图和情感在对话过程中是否发生可测量的变化，这一过程导向的标准既无法通过表层文本指标（如 BLEU/ROUGE）捕捉，也无法通过单一评分的 LLM 判断来实现。我们提出了 extbf{Cog}nitive extbf{W}orld extbf{M}odel extbf{(CogWM)}，这是一种基于 LLM 的用户模型，将多轮对话评估从“用户说了什么”重新框定为“用户的内部认知状态如何演变”。CogWM 共同预测 BDI/E 认知状态和用户发言，并作为用户模拟器和评估平台，采用涵盖轮次级忠实度、轨迹级状态动态和任务级综合评分的三层评估框架。通过我们的 extbf{S}ummarize- extbf{a}nd- extbf{A}llocate extbf{(SaA)} 注释管道，在四个社会影响场景中对 150,454 个用户轮次样本进行训练，CogWM 实现了 77.6 ext{%} 的情感准确率（比 GPT-5.5 提高 2.1 倍）。在 3600 次多智能体区分试验中，它根据认知影响区分六个商业代理，其中 Llama-4-Scout 排名第一（CTS +0.233）。CogWM 将社会影响对话评估从终端判断转向过程跟踪。我们已发布代码 ootnote{ iny 代码: https://github.com/lucianma05-create/CogWM} 和模型 ootnote{模型: https://www.modelscope.cn/models/LucianMa/CogWM-14B}。
cs.AI / 52 / 2606.29502
UCOB: Learning to Utilize and Evolve Agentic Skills via Credit-Aware On-Policy Bidirectional Self-Distillation
UCOB：通过信用感知的在线双向自蒸馏学习利用和发展自主技能
Abstract
Skill memories can improve agentic reinforcement learning by reusing past experience as textual guidance, but retrieved skills are not oracular: they may help in one state while misleading the same policy in another. This makes the common privileged-teacher assumption fragile, namely that a skill-conditioned prompt can be treated as a fixed teacher for the no-skill prompt. We introduce UCOB, a framework for learning to utilize and evolve agentic skills via credit-aware on-policy bidirectional self-distillation. UCOB treats skill-conditioned and no-skill prompts as two on-policy context views of the same model, compares their return-to-go within the same task and anchor state, and uses the higher-return view as the local teacher. This local credit signal internalizes useful skill-conditioned behavior, corrects misleading skill usage, and guides task/state skill memory updates, utility-aware retrieval, and reflection self-training. Experiments on agentic tasks, including ALFWorld, WebShop, and Search-QA, show that UCOB outperforms skill-free RL, skill-memory baselines, and self-distillation methods across model scales, with up to 23.5 and 18.0 point gains over SOTA baselines on ALFWorld and WebShop. Ablations and analyses further validate its core mechanisms and efficiency.
Chinese Translation
技能记忆可以通过将过去的经验作为文本指导来改善自主强化学习，但检索到的技能并非全知全能：它们可能在某个状态下有帮助，而在另一个状态下则可能误导相同的策略。这使得常见的特权教师假设变得脆弱，即技能条件提示可以被视为无技能提示的固定教师。我们提出了UCOB，一个通过信用感知的在线双向自蒸馏学习利用和发展自主技能的框架。UCOB将技能条件和无技能提示视为同一模型的两个在线上下文视图，在相同任务和锚定状态下比较它们的回报，并使用回报更高的视图作为局部教师。这个局部信用信号内化了有用的技能条件行为，纠正了误导性的技能使用，并指导任务/状态技能记忆的更新、效用感知的检索和反思自我训练。在自主任务上的实验，包括ALFWorld、WebShop和Search-QA，显示UCOB在模型规模上优于无技能强化学习、技能记忆基线和自蒸馏方法，在ALFWorld和WebShop上分别比SOTA基线提高了23.5和18.0个百分点。消融实验和分析进一步验证了其核心机制和效率。
cs.AI / 53 / 2606.29537
OSWorld2.0: Benchmarking Computer Use Agents on Long-Horizon Real-World Tasks
OSWorld 2.0：在长时间真实世界任务中对计算机使用代理进行基准测试
Yuan, Mengqi, Zhou, Zilong, Xiong, Xinzhuang, Wu, Weiming, Sun, Jiayang, Song, Jiamin, Cui, Kaiqian, Wang, Bowen, Wu, Haoyuan, Li, Yitong, Lu, Dunjie, Lu, Haikong, Zhen, Qi, Wang, Xinyuan, Deng, Jiaqi, Yang, Yuhao, Chen, Cheng, Zheng, Boyuan, Su, Alex, Yu, Xiao, Zou, Hao, Agashe, Saaket, Lu, Xing Han, Kaur, Manpreet, Qi, Zhengyang, Chen, Vincent Sunn, Sala, Frederic, Liu, Dayiheng, Lin, Junyang, Yu, Zhou, Su, Yu, Reddy, Siva, Wang, Xin Eric, Qi, Peng, Xie, Tianbao, Yu, Tao
Abstract
Existing computer-use benchmarks fail to capture the realism, complexity, and long-horizon demands of real-world computer use, limiting their ability to reveal the limitations of frontier agents. We introduce OSWorld 2.0, a benchmark of 108 long-horizon computer-use workflows across everyday and professional tasks, designed to capture complex and challenging real-world phenomena. Each task represents a realistic end-to-end workflow that takes human users a median of about 1.6 hours to complete and requires an average of 318 tool calls with Claude Opus 4.7 using maximum thinking, compared with about 30 in OSWorld 1.0. OSWorld 2.0 targets challenge phenomena that are common in real workflows yet underrepresented in prior benchmarks, spanning interaction-design challenges such as streaming interaction and dynamic environments, as well as agent-pattern challenges such as cross-source reasoning, implicit-state inference, and visual-spatial precision. Tasks are grounded in authentic input artifacts and cross-referenced against realistic stateful user profile data, and include separate safety reports auditing safety-sensitive execution. Under our primary binary-completion metric at 500 steps, Claude Opus 4.8 with maximum thinking and batched tool calls scores best but still completes only 20.6% of tasks at a 54.8% partial score; GPT-5.5 is far more token-efficient yet plateaus near 13%. These results show that current agents are still far from professional-level computer use: rather than stumbling on basic GUI control or coding, they lose track of constraints, miss information that arrives mid-task, guess rather than ask the user, and skip verification, struggling most when a task hinges on hidden state they must recover.
Chinese Translation
现有的计算机使用基准无法捕捉真实世界计算机使用的真实感、复杂性和长时间需求，限制了它们揭示前沿代理局限性的能力。我们介绍了 OSWorld 2.0，这是一个涵盖108个长时间计算机使用工作流程的基准，涉及日常和专业任务，旨在捕捉复杂且具有挑战性的真实世界现象。每个任务代表一个现实的端到端工作流程，完成该任务的平均时间约为1.6小时，使用 Claude Opus 4.7 进行最大思考时平均需要318次工具调用，而在 OSWorld 1.0 中仅约为30次。OSWorld 2.0 针对真实工作流程中常见但在以往基准中表现不足的挑战现象，涵盖了交互设计挑战（如流式交互和动态环境）以及代理模式挑战（如跨源推理、隐式状态推断和视觉空间精度）。任务基于真实的输入文物，并与现实的状态用户档案数据进行交叉引用，并包括单独的安全报告以审核安全敏感的执行。在我们以500步为基础的主要二元完成指标下，使用最大思考和批量工具调用的 Claude Opus 4.8 得分最佳，但仍仅完成20.6%的任务，部分得分为54.8%；而 GPT-5.5 的令牌效率远高于前者，但接近13%时趋于平稳。这些结果表明，当前代理仍远未达到专业级别的计算机使用：它们不是在基本的图形用户界面控制或编码上出错，而是失去对约束的跟踪，错过任务中途到达的信息，猜测而不是询问用户，并跳过验证，在任务依赖于必须恢复的隐藏状态时表现最差。
cs.AI / 54 / 2606.29541
Learned Coordination Conventions in Cooperative MARL: Measuring the Translation Gap Between Theory-Informed Roles and Learned Routing
合作多智能体强化学习中的学习协调约定：测量理论指导角色与学习路由之间的翻译差距
Abstract
Role-semantic assignments provide priors over how heterogeneous agents may coordinate, but cooperative MARL systems instead settle on conventions through decentralized, non-stationary learning, with no guarantee that the resulting structure matches those priors. We study this translation gap between theory-informed role expectations and learned coordination structure through a diagnostic combining a role-routing matrix, formation sensitivity ($\Delta_{\max}$), and gradient/occlusion attribution across three-role MiniGrid and SMACv2 (Terran) environments. We show that label-conditioned attention produces substantially more concentrated and role-specific routing than flat MLP baselines, remains stable under 3v3--9v9 scaling, transfers zero-shot across team sizes, and is invariant to ally-slot padding. A 5-seed re-evaluation shows partial alignment between learned conventions and designer-specified priors while revealing where small-n noise can manufacture apparent strategic divergence. We present these results as an empirical framework for measuring coordination structure in cooperative MARL rather than as a new equilibrium concept or causal explanation.
Chinese Translation
角色语义分配为异构智能体的协调提供了先验知识，但合作多智能体强化学习系统则通过去中心化的非平稳学习形成约定，无法保证最终结构与这些先验知识相匹配。我们通过结合角色路由矩阵、队形敏感性（$ ext{Δ}_{ ext{max}}$）和在三角色 MiniGrid 和 SMACv2（Terran）环境中的梯度/遮挡归因的诊断研究这一理论指导角色期望与学习协调结构之间的翻译差距。我们展示了标签条件注意力比平坦的多层感知器基线产生了更为集中和角色特定的路由，在 3v3 到 9v9 的扩展下保持稳定，能够在团队规模上实现零样本迁移，并且对盟友槽填充不变。五次种子重新评估显示学习到的约定与设计者指定的先验之间存在部分对齐，同时揭示了小样本噪声如何制造明显的战略分歧。我们将这些结果呈现为测量合作多智能体强化学习中协调结构的实证框架，而不是作为新的均衡概念或因果解释。
cs.AI / 55 / 2606.29623
SCARCE: Scalable Cascade Analysis for Rare-event Characterisation via Embeddings
SCARCE：通过嵌入进行稀有事件特征化的可扩展级联分析
Abstract
Rare events govern the safety profile of modern AI systems, yet their probabilities are extremely difficult to estimate: direct Monte Carlo requires prohibitive sample budgets. Subset Simulation (SS) addresses this by decomposing a rare-event probability into moderate conditional probabilities over nested intermediate events. However, classical SS requires a handcrafted scalar performance function whose sublevel sets define those events, demanding detailed knowledge of the failure geometry and limiting transfer to new domains. We propose SCARCE (Scalable Cascade Analysis for Rare-event Characterisation via Embeddings), which replaces the performance function with learned latent representations and geometric rulers that score proximity to failure regions. Adaptive thresholding constructs nested intermediate events directly from data. We formalise SCARCE through a non-negative supermartingale, yielding a high-probability upper envelope that remains valid under early stopping. On MNIST misclassification, where dense Monte Carlo provides ground truth, SCARCE achieves approximately 400--500 times lower mean absolute error than grid-searched traditional SS while eliminating systematic over-counting. We then study PAIR-style LLM jailbreaks under a fleet-level threat model with adversarial fraction $\eta$. On Llama-Guard-3-8B hidden states, a PCA-based ruler attains 2.6% mean relative error for $\eta \geq 10^{-3}$ against finite-sample references whose average bootstrap relative half-width is 27.9%, and transfers to a GCG-style corpus with 2.93% relative error after recalibration. A directional criterion $\mathrm{KL}(p_{\mathrm{good}}\,\|\,p_{\mathrm{bad}})$ ranks rulers consistently with estimation error (Spearman $\rho=0.83$).
Chinese Translation
稀有事件主导现代人工智能系统的安全特征，但其概率极难估计：直接的蒙特卡罗方法需要巨大的样本预算。子集模拟（Subset Simulation, SS）通过将稀有事件概率分解为嵌套中间事件上的适度条件概率来解决这一问题。然而，经典的SS需要手工设计的标量性能函数，其子水平集定义了这些事件，这要求对失效几何形状有详细了解，并限制了其在新领域的转移。我们提出了SCARCE（通过嵌入进行稀有事件特征化的可扩展级联分析），它用学习到的潜在表示和几何尺代替性能函数，以评分接近失效区域的程度。自适应阈值直接从数据构建嵌套中间事件。我们通过非负超马丁戈尔（supermartingale）形式化SCARCE，得出在早期停止下仍然有效的高概率上限。在MNIST误分类中，稠密蒙特卡罗提供了真实值，SCARCE的平均绝对误差比网格搜索的传统SS低约400至500倍，同时消除了系统性的过度计数。随后，我们在一个具有对抗性比例$ heta$的舰队级威胁模型下研究PAIR风格的LLM越狱。在Llama-Guard-3-8B的隐藏状态中，基于主成分分析（PCA）的尺在$ heta
geq 10^{-3}$的情况下，相对于有限样本参考的平均自助法相对半宽度为27.9%，达到了2.6%的平均相对误差，并在重新校准后转移到GCG风格的语料库，获得2.93%的相对误差。方向性标准$ ext{KL}(p_{ ext{good}} \, ig\, ext{|} \, p_{ ext{bad}})$与估计误差一致地对尺进行排名（Spearman $
ho=0.83$）。
cs.AI / 56 / 2606.29630
SFBench: The SciFy Scientific Feasibility Benchmark
SFBench：科学可行性基准测试
Abstract
We present SFBench, a benchmark dataset for evaluating systems that assess the feasibility of scientific claims. SFBench includes 197 claims in materials science, each annotated with a ground-truth feasibility score on a five-point scale along with an explanation of that assessment. The collection differs from previous collections in several important ways: 1) it defines a complex task that requires reasoning over claims of varying scientific feasibility; 2) its claims are not extracted from existing scientific publications but are created de novo, greatly reducing the chances that LLMs have trained on them; 3) claims and ground truth are established by subject matter experts, not by artificial intelligence; and 4) unlike many benchmarks that ask about question/answer pairs, provide multiple choice answers, or ask questions requiring short, fixed answers, SFBench explanations are completely open-ended. We describe the benchmark design, data creation process, and evaluation metrics, and we report baseline results using recent GPT models.
Chinese Translation
我们提出了SFBench，这是一个用于评估系统判断科学主张可行性的基准数据集。SFBench包含197个材料科学领域的主张，每个主张都附有一个五分制的真实可行性评分及其评估解释。该集合在几个重要方面与之前的集合有所不同：1）它定义了一个复杂的任务，需要对不同科学可行性的主张进行推理；2）其主张并非从现有的科学出版物中提取，而是全新创建，极大降低了大语言模型（LLMs）在其上训练的可能性；3）主张和真实情况由主题专家确定，而非人工智能；4）与许多询问问题/答案对、提供多项选择答案或要求短固定答案的问题的基准不同，SFBench的解释是完全开放式的。我们描述了基准设计、数据创建过程和评估指标，并报告了使用最新GPT模型的基线结果。
cs.AI / 57 / 2606.29654
Budgeted Act-or-Defer Multi-Agent LLM Deliberation with Local Reliability Bounds
预算化的行动或延迟多智能体大语言模型审议与局部可靠性界限
Abstract
Multi-agent deliberation among LLMs can improve reasoning, but deployment requires deciding when the current answer is reliable enough to act on and when it should be escalated to human review. We formulate this as budgeted act-or-defer decision making. At each round, the system maps the debate prefix to a low-dimensional state, computes a $k$-nearest-neighbor lower confidence bound on state-conditional correctness using calibration data, and acts only when the bound exceeds a user-specified reliability threshold. The certificate controls wrong actions through the decomposition $\beta = \delta + \alpha + \varepsilon_{\mathrm{act}}$, separating calibration failure, residual action risk, and representation gap. The guarantee is conditional, not distribution-free: it relies on a valid local bias envelope and an action-region representation-gap bound, and each assumption is paired with falsification-style diagnostics. Because the same absolute wrong-action budget has different meanings across tasks of different difficulty, we set budgets relative to each task's final-round error using training data only, and evaluate safety by normalized budget usage $\mathrm{WA}/\beta$. On six benchmarks against nine baselines, the method uses 9--12% of the pre-declared budget on activated datasets, reaching up to 84% automation and 96% acted-on accuracy; on stress-test datasets, it defers rather than forcing unreliable automation. Rather than relying on per-task post-hoc threshold search, the method prospectively converts a user-declared wrong-action budget into an auditable act-or-defer operating point before deployment, under explicitly stated assumptions.
Chinese Translation
多智能体之间的大语言模型（LLMs）审议可以改善推理能力，但在部署时需要决定当前答案在何时足够可靠以进行行动，以及何时应升级至人类审查。我们将此问题表述为预算化的行动或延迟决策。在每一轮中，系统将辩论前缀映射到低维状态，利用校准数据计算状态条件正确性的 $k$-最近邻下置信界限，仅在该界限超过用户指定的可靠性阈值时采取行动。该证书通过分解 $eta = eta + eta + eta_{ ext{act}}$ 控制错误行动，分别处理校准失败、残余行动风险和表示差距。该保证是有条件的，而非无分布的：它依赖于有效的局部偏差包络和行动区域表示差距界限，并且每个假设都配有伪造风格的诊断。由于相同的绝对错误行动预算在不同难度的任务中具有不同的含义，我们相对于每个任务的最终轮错误设置预算，仅使用训练数据，并通过标准化预算使用 $ ext{WA}/eta$ 评估安全性。在六个基准测试中与九个基线进行比较，该方法在激活数据集上使用了预先声明预算的 9% 到 12%，实现了高达 84% 的自动化和 96% 的行动准确性；在压力测试数据集上，它选择延迟而不是强制不可靠的自动化。该方法不是依赖于每个任务的事后阈值搜索，而是前瞻性地将用户声明的错误行动预算转换为可审计的行动或延迟操作点，在明确声明的假设下进行部署。
cs.AI / 58 / 2606.29657
Safety from Honesty in a Disinterested AI Predictor
来自诚实的安全性：一种无私的人工智能预测器
Abstract
As AI systems become more capable, training procedures that optimize for downstream outcomes risk introducing implicit agency: goal-directed behavior that designers never specified. We present a formal safety argument for the Scientist AI (SAI) Predictor, trained to approximate the Bayesian posterior conditioned on a dataset of "epistemically contextualized" natural-language statements. We argue that such a Predictor can honestly predict agents, actions, and their consequences without itself being an agent that selects outputs to achieve goals. This rests on data representation and on the training procedure. Epistemic contextualization of text distinguishes latent factual claims from communication acts, so expressions of goals are treated as evidence to be explained rather than drives the model adopts. With a posterior-seeking training objective, this is intended to drive the Predictor toward calibrated, cautious predictions. Training proceeds so downstream effects of deploying a prediction never serve as a reward signal; any agency the system needs is supplied by explicit scaffolding constrained by guardrails. We prove that, under assumptions on the training dynamics and on the argued sparsity of dangerous Predictors, the probability that training produces a Predictor whose guarded deployment carries residual harm above a specified threshold is small: a dangerous Predictor would have to underestimate harm in a coordinated way across many queries while such coordinated patterns are rare under the initialization distribution and receive no direct training signal. Safety and accuracy are jointly supported in this framework, since the constraints that secure accuracy are the same ones that make coordinated deception costly. These guarantees against misalignment and agency arising from within the Predictor itself do not preclude the use of the Predictor as part of an agentic system.
Chinese Translation
随着人工智能系统能力的提升，优化下游结果的训练程序可能会引入隐含的代理性：设计者未曾指定的目标导向行为。我们提出了一个关于科学家人工智能（Scientist AI, SAI）预测器的正式安全论证，该预测器经过训练以近似基于“认识论上下文化”的自然语言陈述的数据集的贝叶斯后验。我们认为，这样的预测器能够诚实地预测代理、行动及其后果，而不需要自身成为一个选择输出以实现目标的代理。这一论点基于数据表示和训练程序。文本的认识论上下文化将潜在的事实主张与交流行为区分开，因此目标的表达被视为需要解释的证据，而不是模型所采用的驱动。通过追求后验的训练目标，旨在引导预测器朝向经过校准的、谨慎的预测。训练过程确保部署预测的下游效果从未作为奖励信号；系统所需的任何代理性都是通过受限于保护措施的明确支架提供的。我们证明，在对训练动态和所论证的危险预测器稀疏性做出假设的情况下，训练产生一个其受保护部署带有超过特定阈值的残余危害的预测器的概率是很小的：一个危险的预测器必须在多个查询中以协调的方式低估危害，而这种协调模式在初始化分布下是稀有的，并且没有直接的训练信号。在这个框架中，安全性和准确性是共同支持的，因为确保准确性的约束与使协调欺骗代价高昂的约束是相同的。这些针对预测器内部产生的不一致性和代理性的保证并不排除将预测器用作代理系统的一部分。
cs.AI / 59 / 2606.29661
Diversity is the Strength of the AI Crowd
多样性是人工智能群体的力量
Abstract
Top AI forecasting systems are approaching superforecaster-level accuracy on future world events, but still rely primarily on off-the-shelf LLMs combined with forecasting-specific context gathering and scaffolding. We study how to improve this recipe through ensembling: given a fixed number of samples, which off-the-shelf model forecasts should be combined to maximize accuracy? On binary questions from the Metaculus AI Benchmark, we find that individual accuracy is not enough: many frontier LLMs make highly correlated predictions, limiting the value of additional forecasts from the same or similar models. Instead, the strongest ensembles combine accurate but diverse forecasters, with models such as \model{Grok 4} contributing disproportionately because their predictions are less correlated with other frontier LLMs. These results suggest that the strength of the AI crowd comes not from sampling more forecasts indiscriminately, but from combining forecasts across models with complementary errors, motivating forecasting systems that explicitly optimize for both model quality and diversity.
Chinese Translation
顶级人工智能预测系统在未来世界事件的预测准确性上已接近超级预测者的水平，但仍主要依赖于现成的大型语言模型（LLMs）结合特定于预测的上下文收集和框架搭建。我们研究如何通过集成方法来改进这一方案：在固定数量的样本下，应该结合哪些现成模型的预测以最大化准确性？在来自Metaculus人工智能基准的二元问题中，我们发现单一的准确性并不足够：许多前沿的LLMs做出高度相关的预测，这限制了来自相同或相似模型的额外预测的价值。相反，最强的集成组合了准确但多样的预测者，其中如 extit{Grok 4}等模型的贡献不成比例，因为它们的预测与其他前沿LLMs的相关性较低。这些结果表明，人工智能群体的力量并不来自于无差别地采样更多的预测，而是来自于结合具有互补错误的模型预测，激励预测系统明确优化模型质量和多样性。
cs.AI / 60 / 2606.29681
Sample-Efficient Learning of Probabilistic Causes for Reachability in Markov Decision Processes with Probabilistic Guarantees
具有概率保证的马尔可夫决策过程中的可达性概率原因的样本高效学习
Abstract
Probabilistic model checking for Markov decision processes (MDPs) provides quantitative guarantees, but often offers limited insight into why undesired outcomes occur. Probability-raising (PR) causality addresses this by identifying states whose visitation increases the probability of reaching designated states. Existing PR-cause identification methods, however, use MDP modifications not well-suited for learning: the gap between conditional and unconditional reachability probabilities can be hard to detect from transition samples, and construction requires reachability probabilities of the MDP, which are unavailable when transition probabilities are unknown. We study unknown MDPs and propose a learning approach with probabilistic guarantees for PR-cause identification. Our key ingredient is a restart-based MDP modification that reduces PR-cause checking to two conditional reachability queries without using reachability values of the original MDP. We prove correctness, establish sample-complexity bounds, and develop an anytime learning-and-checking algorithm based on two-sided value iteration that progressively classifies states as causal, non-causal, or undecided. Experiments on two benchmarks demonstrate reliable and fast identification of PR causes.
Chinese Translation
马尔可夫决策过程（MDP）的概率模型检查提供了定量保证，但通常对不希望发生的结果为何发生提供的洞察有限。概率提升（PR）因果关系通过识别那些其访问增加到达指定状态概率的状态来解决这个问题。然而，现有的PR因果识别方法使用的MDP修改并不适合学习：条件和无条件可达性概率之间的差距可能难以从转移样本中检测到，并且构造需要MDP的可达性概率，而在转移概率未知时这些概率是不可用的。我们研究未知的MDP，并提出了一种具有概率保证的PR因果识别学习方法。我们关键的组成部分是基于重启的MDP修改，它将PR因果检查简化为两个条件可达性查询，而无需使用原始MDP的可达性值。我们证明了正确性，建立了样本复杂度界限，并开发了一种基于双侧值迭代的随时学习与检查算法，该算法逐步将状态分类为因果、非因果或未决。对两个基准的实验表明，PR因果的识别可靠且快速。
cs.AI / 61 / 2606.29700
Toward Secure and Reliable PDDL Formalization of Large Language Models with Planner-in-the-Loop Feedback
朝着安全可靠的PDDL形式化大型语言模型的规划者反馈机制
Abstract
Planning often requires symbolic specifications that are both executable and verifiable. For large language models deployed in autonomous or decision-support systems, failures in such formalization may lead to unverifiable decisions, execution failures, or unsafe downstream behavior. We present NL-PDDL-Bench, a multi-domain benchmark for natural-language-to-PDDL specification construction with planner-verified executability and controlled difficulty scaling by object count. We further propose a planner-in-the-loop framework that uses validator and planner diagnostics to revise non-executable specifications through localized edits. Building on this infrastructure, we develop a planner-grounded optimization recipe that combines parameter-efficient Low-Rank Adaptation supervised fine-tuning, offline planner-derived preference pairs for Direct Preference Optimization, and inference-time planner-in-the-loop repair, without requiring online planner calls during training. We also provide a unified evaluation suite for parseability, solvability, specification similarity, and outcome-aware plan-level consistency against planner references. Experiments on representative model families show substantial gains in planner success and plan-level agreement, with improved robustness under difficulty scaling and cross-domain variation. These results highlight the value of externally verifiable formalization for reliable deployment of LLMs in safety- or security-sensitive planning systems. Code and data are available at: https://github.com/ibasicplan/NL-PDDL-Bench
Chinese Translation
规划通常需要可执行且可验证的符号规范。对于部署在自主或决策支持系统中的大型语言模型，这种形式化的失败可能导致不可验证的决策、执行失败或不安全的下游行为。我们提出了NL-PDDL-Bench，这是一个多领域基准，用于自然语言到PDDL规范构建，具有规划者验证的可执行性和通过对象数量控制的难度扩展。我们进一步提出了一种规划者反馈机制框架，该框架利用验证器和规划者诊断，通过局部编辑修订不可执行的规范。在此基础上，我们开发了一种基于规划者的优化方案，结合了参数高效的低秩适应（Low-Rank Adaptation）监督微调、离线规划者衍生的偏好对用于直接偏好优化（Direct Preference Optimization），以及推理时的规划者反馈修复，而无需在训练期间进行在线规划者调用。我们还提供了一个统一的评估套件，用于解析性、可解性、规范相似性和基于结果的计划级一致性，与规划者参考进行比较。对代表性模型家族的实验显示，规划者成功率和计划级一致性有显著提升，在难度扩展和跨领域变化下表现出更强的鲁棒性。这些结果突显了外部可验证形式化在安全或安全敏感的规划系统中可靠部署大型语言模型的价值。代码和数据可在以下链接获取：https://github.com/ibasicplan/NL-PDDL-Bench
cs.AI / 62 / 2606.29705
GUICrafter: Weakly-Supervised GUI Agent Leveraging Massive Unannotated Screenshots
GUICrafter：利用大量未标注截图的弱监督GUI代理
Abstract
Data, as the fundamental substrate of modern intelligence, has greatly driven the development of current foundation models. Naturally, researchers aim to extend this paradigm to the domain of GUI agents, hoping to build strong GUI agents through a similar paradigm. However, GUI agent data cannot be directly harvested from the internet, making it costly and difficult to collect at scale. As a result, current GUI agents suffer from poor cross-device generalization and limited visual grounding ability for fine-grained GUI elements. As an attempt to address data challenge in GUI agents, we propose GUICrafter, a weakly-supervised GUI agent leveraging massive unannotated screenshots to substantially reduce the reliance on expensive human annotations. GUICrafter explores a curriculum learning framework for training GUI agents through two progressive stages. First, the model learns visual grounding from large-scale unannotated screenshots and webpages, leveraging the rich contextual signals inherent in GUI interactions without human annotations. Then, in Stage 2, we leverage a small amount of high-quality data to calibrate the model via reinforcement learning. Experiments show that GUICrafter achieves competitive, or even superior, performance to advanced systems like UI-TARS while using only 0.1% of its data. Furthermore, under the same amount of annotated data, GUICrafter surpasses all previous methods such as GUI-R1. Code, data, and models are available at https://github.com/fansunqi/GUICrafter.
Chinese Translation
数据作为现代智能的基础底层，极大推动了当前基础模型的发展。自然，研究人员希望将这一范式扩展到GUI代理领域，期望通过类似的范式构建强大的GUI代理。然而，GUI代理数据无法直接从互联网获取，导致大规模收集成本高昂且困难。因此，当前的GUI代理在跨设备泛化能力和对细粒度GUI元素的视觉基础能力方面表现不佳。为了解决GUI代理中的数据挑战，我们提出了GUICrafter，一种利用大量未标注截图的弱监督GUI代理，旨在大幅减少对昂贵人类标注的依赖。GUICrafter通过两个渐进阶段探索了一个课程学习框架来训练GUI代理。首先，模型从大规模未标注的截图和网页中学习视觉基础，利用GUI交互中固有的丰富上下文信号，而无需人类标注。然后，在第二阶段，我们利用少量高质量数据通过强化学习来校准模型。实验表明，GUICrafter在使用仅0.1%数据的情况下，达到了与UI-TARS等先进系统竞争甚至更优的性能。此外，在相同数量的标注数据下，GUICrafter超越了所有先前的方法，如GUI-R1。代码、数据和模型可在 https://github.com/fansunqi/GUICrafter 获取。
cs.AI / 63 / 2606.29727
DeepTrans Studio: Turning Expert Interventions into Shared Team Knowledge in Agentic Translation Workflows
DeepTrans Studio：将专家干预转化为代理翻译工作流程中的共享团队知识
Abstract
Professional translation is often a team-based process: translators, reviewers, and project managers must coordinate terminology, legal force, and accountability across documents. Yet many LLM-based translation tools treat human corrections as isolated edits. Expert decisions made in one segment or by one member are rarely captured as reusable knowledge for the rest of the team. We present DeepTrans Studio, a collaborative translation workspace that lets professionals intercept selected nodes in an agentic translation workflow, review evidence, revise AI outputs, and save approved decisions to a shared team memory. During the demo, attendees will role-play translators and reviewers, resolve preset terminology and legal-modal risks, and see how their decisions are propagated to downstream segments and surfaced in a teammate's workspace as reusable precedents. The demo illustrates how human interventions in AI-mediated work can become shared, traceable knowledge rather than one-off corrections.
Chinese Translation
专业翻译通常是一个基于团队的过程：翻译人员、审校人员和项目经理必须在文档之间协调术语、法律效力和责任。然而，许多基于大型语言模型（LLM）的翻译工具将人工修正视为孤立的编辑。在一个段落或由一个成员做出的专家决策很少被捕捉为其他团队成员可重用的知识。我们提出了DeepTrans Studio，一个协作翻译工作空间，允许专业人员在代理翻译工作流程中拦截选定节点，审查证据，修订人工智能输出，并将批准的决策保存到共享团队记忆中。在演示过程中，参与者将扮演翻译人员和审校人员，解决预设的术语和法律模态风险，并观察他们的决策如何传播到下游段落，并在队友的工作空间中作为可重用的先例呈现。该演示展示了人类在人工智能介导的工作中的干预如何转变为共享的、可追溯的知识，而不仅仅是一时的修正。
cs.AI / 64 / 2606.29746
DEEPMED Search: An Open-Source Agentic Platform for Medical Deep Research with Introspective Verification
DEEPMED Search：一个用于医学深度研究的开源自主平台，具有内省验证功能
Abstract
Navigating the deluge of heterogeneous medical data, from academic literature (PubMed) to clinical guidelines (Web) and private knowledge bases, remains a critical bottleneck for evidence-based medicine. While commercial black-box tools lack transparency, standard open-source RAG implementations frequently suffer from reasoning drift when handling complex, long-tail queries. We present DEEPMED Search, a fully open-source, agentic platform designed for transparent medical deep research. Built on a high-performance Next.js architecture, DEEPMED Search features a source-adaptive router that autonomously dispatches sub-queries to PubMed, web search, or local graph-based knowledge bases based on information density. Crucially, the platform integrates an introspective verification module, powered by a causal-consistent multi-agent debate framework, to validate retrieved evidence against diagnostic logic before synthesis. To demonstrate its robustness, we showcase DEEPMED Search's ability to autonomously decompose high-difficulty rare disease queries, filter out confounding noise, and generate structured, citation-backed research reports in minutes. By open-sourcing this software, we provide the community with a robust infrastructure to democratize access to trustworthy, glass-box medical reasoning in research and prototyping settings.
Chinese Translation
在应对来自学术文献（PubMed）、临床指南（Web）和私人知识库等异构医学数据的洪流时，仍然是基于证据医学的一个关键瓶颈。虽然商业黑箱工具缺乏透明性，标准的开源RAG实现往往在处理复杂的长尾查询时遭遇推理漂移。我们提出了DEEPMED Search，一个完全开源的自主平台，旨在实现透明的医学深度研究。DEEPMED Search基于高性能的Next.js架构，具有源自适应路由器，能够根据信息密度自主将子查询分派到PubMed、网络搜索或本地基于图的知识库。关键是，该平台集成了一个内省验证模块，采用因果一致的多智能体辩论框架，在合成之前验证检索到的证据与诊断逻辑的一致性。为了展示其鲁棒性，我们展示了DEEPMED Search自主分解高难度罕见疾病查询的能力，过滤混淆噪声，并在几分钟内生成结构化、引用支持的研究报告。通过开源该软件，我们为社区提供了一个强大的基础设施，以促进在研究和原型设置中获取可信赖的透明医学推理的机会。
cs.AI / 65 / 2606.29748
Rethinking Generative Reconstruction Attacks against Graph Neural Network Models
重新思考针对图神经网络模型的生成重建攻击
Abstract
The application of graph data in numerous disciplines raises the need for gathering and analyzing huge volumes of data, some of which is private and sensitive. The non-Euclidean nature of the graph data makes the analysis computationally challenging, leading to the use of Graph Neural Networks (GNNs) in the age of AI. GNNs may inadvertently leak sensitive data they are trained on, which raises serious data security issues, including the model inversion attack. In this study, we analyze GNNs' vulnerabilities by introducing two novel graph inversion (i.e., reconstruction) attacks: graph-label conditioned (GLC) attack and embedding-label conditioned (ELC) attack, utilizing targetmodel predictions and their intermediate representations, respectively. We perform a comprehensive analysis of our introduced privacy attacks and compare them with existing baselines across three benchmark graph datasets (i.e., NCI1, PROTEINS, and AIDS) and four graph distributional/structural metrics (i.e., FGD, EGD, MMD, and GKS). Our work demonstrates that an adversary can use the generator-discriminator technique to reconstruct high-quality graphs in real-world black-box attack scenarios against GNNs. Additionally, we present a variant of our attacks (Ours--) with 50% reduced queries, achieving good or comparable reconstruction attack performance. In addition, we show that GNNs are highly vulnerable to privacy attacks, varying Laplacian noise-scales.
Chinese Translation
图数据在众多学科中的应用提高了收集和分析大量数据的需求，其中一些数据是私密和敏感的。图数据的非欧几里得特性使得分析在计算上具有挑战性，因此在人工智能时代采用了图神经网络（Graph Neural Networks, GNNs）。GNNs可能会无意中泄露其训练所用的敏感数据，这引发了严重的数据安全问题，包括模型反演攻击。在本研究中，我们通过引入两种新颖的图反演（即重建）攻击：图标签条件（graph-label conditioned, GLC）攻击和嵌入标签条件（embedding-label conditioned, ELC）攻击，分别利用目标模型的预测和其中间表示，分析GNNs的脆弱性。我们对所提出的隐私攻击进行了全面分析，并在三个基准图数据集（即NCI1、PROTEINS和AIDS）和四个图分布/结构指标（即FGD、EGD、MMD和GKS）上与现有基线进行了比较。我们的工作表明，攻击者可以利用生成器-判别器技术在针对GNNs的真实世界黑箱攻击场景中重建高质量图。此外，我们还展示了我们攻击的一种变体（Ours--），其查询量减少50%，实现了良好或可比的重建攻击性能。此外，我们还表明，GNNs对隐私攻击高度脆弱，拉普拉斯噪声尺度变化显著。
cs.AI / 66 / 2606.29771
CLQT: A Closed-Loop, Cost-Aware, Strategy-Consistent Benchmark for Diagnostic Evaluation of LLM Portfolio-Management Agents
CLQT：一种闭环、成本意识、策略一致的基准，用于大语言模型投资组合管理代理的诊断评估
Abstract
LLM agents are increasingly cast as autonomous portfolio managers, and benchmarks have moved from financial question-answering to sequential trading. Yet most still rank agents by returns over a fixed window -- a weak proxy, since a period's return is dominated by the market path and apparent alpha can dissolve once look-ahead leakage is controlled. Such a ranking certifies neither sound reasoning, nor a consistent strategy, nor a durable edge. We introduce CLQT, which reframes closed-loop trading evaluation as diagnosis rather than ranking: an instrument that localizes where and why an agent's process succeeds or fails. CLQT is a fully closed-loop, cost-aware, strategy-consistent, temporally-gated environment whose agents run a five-stage cycle: gather, synthesize, allocate, execute, reflect. Each round emits a complete DecisionRound sealed into a recompute-verifiable hash chain, so every metric is reconstructable from the trail. Six pillars form the substrate: a hard TimeGate, institutional transaction- and financing-cost modeling, strategy-consistency scoring, three-tier memory, a Model-Context-Protocol tool layer, and mandate-aware synthesis. The same agent runs as a constrained committee of specialized roles or a single full-autonomy orchestrator, making process scaffolding an experimental variable. From the audit trail we compute a five-axis capability scorecard (APM-CS: Coherence, Acuity, Composure, Discipline, Reliability), with Coherence judged partly by a held-out, out-of-cohort LLM to curb self-preference bias. We validate it on a contamination-controlled multi-model backtest with an ablation grid and a live broker track on unseen, post-cutoff data, against a repeated-run noise floor. CLQT separates outcome from capability, yielding not a model ranking but a durable, extensible map of agent competencies and limitations.
Chinese Translation
大语言模型（LLM）代理越来越被视为自主投资组合管理者，基准测试也从金融问答转向顺序交易。然而，大多数仍然通过固定时间窗口内的收益来对代理进行排名——这是一种较弱的代理，因为一个时期的收益受到市场路径的主导，而一旦控制了前瞻性泄漏，表面上的阿尔法可能会消失。这种排名既不能证明合理的推理，也不能保证一致的策略，更无法确保持久的优势。我们引入了CLQT，它将闭环交易评估重新定义为诊断，而非排名：这是一种工具，可以定位代理的过程在何处以及为何成功或失败。CLQT是一个完全闭环、成本意识、策略一致、时间门控的环境，其代理运行一个五阶段循环：收集、综合、分配、执行、反思。每一轮都会生成一个完整的DecisionRound，并封装在一个可重新计算验证的哈希链中，因此每个指标都可以从轨迹中重建。六个支柱构成了基础：一个严格的时间门控、机构交易和融资成本建模、策略一致性评分、三层记忆、模型-上下文-协议工具层，以及任务意识合成。同一个代理可以作为一个受限的专业角色委员会或一个完全自主的协调者运行，使得过程支架成为一个实验变量。通过审计轨迹，我们计算出一个五轴能力评分卡（APM-CS：一致性、敏锐性、沉着、纪律、可靠性），其中一致性部分通过一个保留的、非同群体的LLM来评估，以抑制自我偏好偏差。我们在一个污染控制的多模型回测中进行了验证，采用消融网格和在未见的截止后数据上的实时经纪人跟踪，针对重复运行的噪声底线。CLQT将结果与能力分开，产生的不是模型排名，而是一个持久的、可扩展的代理能力和局限性的地图。
cs.AI / 67 / 2606.29799
The CRISTAL Method: Neurosymbolic analysis from AI-synthesized world models
CRISTAL 方法：基于人工智能合成世界模型的神经符号分析
Abstract
This project introduces the CRISTAL Method (Coherent Reliable Intentional Synthesis of Truthful Analysis Logic), a neurosymbolic framework for automating complex analysis workflows, with fundamental investment analysis as a primary use case. This domain poses major challenges: high structural uncertainty, noisy and subjective data, tight attention budgets, and the need for justified, reproducible decisions. Human analysts often struggle in this domain due to cognitive biases and limitations, suggesting significant value in automation. But while LLM-based agents have been proposed as analytical aids, their limitations -- poor numerical reasoning, unawareness of uncertainty, and lack of reproducibility -- hinder their effectiveness in this context. CRISTAL addresses these gaps through a principled blend of statistical model synthesis, continuous learning, and active learning. Starting from a natural-language prior knowledge curriculum, CRISTAL builds a dynamic, interpretable probabilistic program that enables full Bayesian inference, including uncertainty quantification and budget-aware data acquisition. CRISTAL continually refines its world model during analysis, leveraging LLMs for code synthesis and learning. We validate CRISTAL on a novel benchmark of synthetic equities with rich financial and textual data. On a company classification task, CRISTAL achieves Bayes-optimal accuracy with just 5 examples and a 5-second budget, outperforming state-of-the-art LLMs that plateau around 40\% accuracy even with order-of-magnitude more input data and compute.
Chinese Translation
本项目介绍了 CRISTAL 方法（Coherent Reliable Intentional Synthesis of Truthful Analysis Logic），这是一个用于自动化复杂分析工作流程的神经符号框架，以基础投资分析作为主要应用案例。该领域面临重大挑战：高结构不确定性、嘈杂和主观数据、紧张的注意预算，以及对有据可依、可重复决策的需求。人类分析师在这一领域常常受到认知偏见和局限性的困扰，这表明自动化具有显著价值。然而，尽管基于大型语言模型（LLM）的代理被提议作为分析辅助工具，但它们的局限性——糟糕的数值推理、对不确定性的无知以及缺乏可重复性——在此背景下阻碍了其有效性。CRISTAL 通过原则性地结合统计模型合成、持续学习和主动学习来填补这些空白。从自然语言的先验知识课程开始，CRISTAL 构建了一个动态的、可解释的概率程序，使得全面的贝叶斯推理成为可能，包括不确定性量化和预算意识的数据获取。CRISTAL 在分析过程中不断完善其世界模型，利用 LLM 进行代码合成和学习。我们在一个具有丰富金融和文本数据的合成股票新基准上验证了 CRISTAL。在公司分类任务中，CRISTAL 仅用 5 个样本和 5 秒的预算就达到了贝叶斯最优准确率，超越了即使在数量级更多输入数据和计算下也仅能达到约 40\% 准确率的最先进 LLM。
cs.AI / 68 / 2606.29860
Beyond Triplet Plausibility: Relation Set Completion in Knowledge Graphs
超越三元组合理性：知识图谱中的关系集补全
Abstract
Knowledge graphs (KGs) organize real-world knowledge as triplets and underpin many downstream applications. Due to their inherent incompleteness, knowledge graph completion (KGC) is widely studied and is typically formulated as triplet prediction, with link prediction as the dominant paradigm. However, this formulation focuses on the incompleteness of triplet-wise information and overlooks the incompleteness of entity-relation compatibility information. To address this limitation, we introduce a relation set completion task (RSC), which complements the link prediction task and aims to reason about missing relations that are semantically compatible with a given entity. We further propose a Relation Set Embedding model (RelSetE), which models latent patterns among the observed relations of entities to infer missing ones. To evaluate RelSetE, we derive three benchmark datasets from standard KG benchmarks. Extensive experiments demonstrate that RelSetE effectively captures entity-relation compatibility patterns and performs favorably in inferring missing relations of entities. Code and data are publicly available.
Chinese Translation
知识图谱（KGs）将现实世界的知识组织为三元组，并支撑许多下游应用。由于其固有的不完整性，知识图谱补全（KGC）受到广泛研究，通常被表述为三元组预测，其中链接预测是主导范式。然而，这种表述关注于三元组信息的不完整性，而忽视了实体-关系兼容性信息的不完整性。为了解决这一局限性，我们引入了关系集补全任务（RSC），该任务补充了链接预测任务，旨在推理与给定实体在语义上兼容的缺失关系。我们进一步提出了一种关系集嵌入模型（RelSetE），该模型通过建模实体观察到的关系之间的潜在模式来推断缺失的关系。为了评估RelSetE，我们从标准KG基准中衍生出三个基准数据集。大量实验表明，RelSetE有效捕捉了实体-关系兼容性模式，并在推断实体缺失关系方面表现良好。代码和数据已公开可用。
cs.AI / 69 / 2606.29871
AI Training Manager: Bounded Closed-Loop Control of Adaptive Training Recipes
AI培训管理器：自适应训练配方的有界闭环控制
Abstract
We present the AI Training Manager, a bounded LLM-based supervisory controller for adaptive machine learning training. Standard training pipelines often rely on fixed recipes or single-axis schedulers, which can struggle with mid-run failures such as severe overfitting, loss imbalance, exploration collapse, or unsafe exploration. Rather than replacing mathematical optimizers or acting as an unconstrained coding agent, the manager operates through a schema-conditioned interface: it reads structured telemetry snapshots from an active run, audits a constrained action space, and returns validated updates to training parameters such as learning rate, regularization strength, loss-weight coefficients, and exploration settings. We evaluate this architecture across supervised language modeling and reinforcement learning. On TinyStories, the manager detects and corrects overfitting, achieving a validation loss 60% lower than the baseline while producing auditable intervention logs. In this supervised setting, we additionally show that manager inference does not need to block the training loop: training can continue while a manager response is pending, and validated updates can be applied asynchronously once available. In a robotic manipulation reinforcement-learning task, we use the same bounded decision interface in an episodic closed-loop setting, where manager updates are applied at evaluation or checkpoint boundaries. The manager mitigates both conservative and unsafe exploration regimes. These results suggest that schema-conditioned LLMs can serve as bounded supervisory managers for live training runs, complementing conventional optimizers and schedulers with interpretable, multi-axis intervention capabilities
Chinese Translation
我们提出了AI培训管理器，这是一种基于有界大语言模型（LLM）的自适应机器学习训练的监督控制器。标准训练流程通常依赖于固定的配方或单轴调度器，这在面对中途失败（如严重过拟合、损失不平衡、探索崩溃或不安全探索）时可能会遇到困难。该管理器并不是替代数学优化器或充当无约束编码代理，而是通过模式条件接口进行操作：它读取来自活跃运行的结构化遥测快照，审核受限的动作空间，并返回对训练参数（如学习率、正则化强度、损失权重系数和探索设置）的验证更新。我们在监督语言建模和强化学习中评估了这一架构。在TinyStories上，管理器检测并纠正了过拟合，达到了比基线低60%的验证损失，同时生成可审计的干预日志。在这一监督设置中，我们还展示了管理器推理不需要阻塞训练循环：在管理器响应待处理时，训练可以继续进行，并且一旦可用，验证更新可以异步应用。在一个机器人操作强化学习任务中，我们在一个情节闭环设置中使用相同的有界决策接口，其中管理器更新在评估或检查点边界应用。管理器缓解了保守和不安全的探索机制。这些结果表明，模式条件的LLM可以作为实时训练运行的有界监督管理器，补充传统优化器和调度器，提供可解释的多轴干预能力。
cs.AI / 70 / 2606.29887
SafePyramid: A Hierarchical Benchmark for In-context Policy Guardrailing
SafePyramid：一种用于上下文政策保护的分层基准
Abstract
In real-world applications, guardrails are often expected to identify unsafe user-model interactions according to application-specific safety policies, rather than relying on predefined risk taxonomies. In this work, we study this setting under the paradigm of in-context policy guardrailing, where guardrails predict safety violations based on policy specifications provided in context. To systematically evaluate this capability, we introduce SafePyramid, a safety benchmark comprising 1,000 multi-turn conversations across 10 domains and 3,000 corresponding application-specific policies, which together contain 61,699 distinct natural-language rules. SafePyramid organizes the evaluation into three difficulty levels: L0 evaluates individual-rule understanding, L1 evaluates reasoning over rule dependencies, and L2 evaluates adaptation of full novel policy frameworks defined in context. To ensure benchmark quality, we employ a rigorous multi-stage pipeline to construct and validate the benchmark. Using SafePyramid, we evaluate 10 frontier LLMs and 5 policy-configurable guardrails and find that in-context policy guardrailing remains highly challenging: even the best-performing model, GPT-5.5, exactly identifies the full set of violated rules in only 54.0%, 35.3%, and 12.9% cases on L0, L1, and L2, respectively. These results highlight the limitations of current guardrails and call for stronger in-context policy guardrails that can reliably execute policies, resolve rule dependencies, and adapt to novel policy frameworks.
Chinese Translation
在实际应用中，保护措施通常被期望根据特定应用的安全政策识别不安全的用户与模型之间的交互，而不是依赖于预定义的风险分类。在本研究中，我们在上下文政策保护的范式下研究这一设置，其中保护措施根据提供的上下文中的政策规范预测安全违规行为。为了系统地评估这一能力，我们引入了SafePyramid，一个安全基准，包含跨越10个领域的1,000个多轮对话和3,000个相应的特定应用政策，这些政策共同包含61,699条不同的自然语言规则。SafePyramid将评估组织为三个难度级别：L0评估单个规则的理解，L1评估规则依赖关系的推理，L2评估在上下文中定义的完整新政策框架的适应性。为了确保基准质量，我们采用严格的多阶段流程来构建和验证基准。使用SafePyramid，我们评估了10个前沿的大型语言模型（LLMs）和5个可配置政策的保护措施，发现上下文政策保护仍然具有很高的挑战性：即使是表现最好的模型GPT-5.5，在L0、L1和L2上也仅在54.0%、35.3%和12.9%的情况下准确识别出所有违规规则。这些结果突显了当前保护措施的局限性，并呼吁开发更强大的上下文政策保护措施，以可靠地执行政策、解决规则依赖关系并适应新的政策框架。
cs.AI / 71 / 2606.29911
A causal modeling perspective on decision theory
决策理论的因果建模视角
Abstract
Decision theory provides a formal framework for how agents should make choices under uncertainty, drawing on ideas from philosophy, probability, and causality. Despite significant progress, the field still lacks a unified modeling language, and key concepts - such as the distinction between subjective and objective elements, or what it means for a decision theory to perform well - are often left implicit. This can make it difficult to evaluate and compare competing theories, particularly in controversial cases. In this paper, we address these issues by introducing a formal framework for decision theory based on nonparametric structural equation models (NPSEMs), a well-established tool in causal inference. NPSEMs provide a unified foundation for representing agents, counterfactuals, and causal relationships, allowing for unambiguous definitions of EDT and CDT. Building on this foundation, we propose a novel decision theory - personal decision theory - which instructs agents to maximize a subjective model of their own counterfactual utility. We introduce a formal performance metric based on hypothetical interventions that enforce a given decision theory across a population - such as might be achieved through education or policy -- and show that, under certain assumptions, personal decision theory is optimal with respect to this metric. Throughout, we use the smoking lesion problem as a running example and conclude with a formal analysis of Newcomb's problem. Our aim is to provide decision theory with a clearer modeling language and firmer evaluative ground, thereby enabling more rigorous comparisons and facilitating conceptual progress in the field.
Chinese Translation
决策理论提供了一个正式框架，指导代理人在不确定性下如何做出选择，借鉴了哲学、概率和因果关系的思想。尽管该领域取得了显著进展，但仍然缺乏统一的建模语言，关键概念——例如主观与客观元素之间的区别，或决策理论表现良好的含义——往往隐含不明。这使得在有争议的案例中评估和比较竞争理论变得困难。在本文中，我们通过引入基于非参数结构方程模型（NPSEMs）的决策理论正式框架来解决这些问题，NPSEMs是因果推断中一个成熟的工具。NPSEMs为表示代理人、反事实和因果关系提供了统一的基础，从而允许对期望决策理论（EDT）和条件决策理论（CDT）进行明确的定义。在此基础上，我们提出了一种新颖的决策理论——个人决策理论，指导代理人最大化其自身反事实效用的主观模型。我们引入了一种基于假设干预的正式性能指标，该干预在一个群体中强制执行给定的决策理论——例如通过教育或政策实现的干预，并表明在某些假设下，个人决策理论在该指标下是最优的。在整个过程中，我们以吸烟病变问题作为持续的例子，并以对纽康布问题的正式分析作为结尾。我们的目标是为决策理论提供更清晰的建模语言和更坚实的评估基础，从而促进更严格的比较，并推动该领域的概念进展。
cs.AI / 72 / 2606.29929
HippoSpark: An On-Demand Experience System for LLM Reasoning
HippoSpark：一种用于大语言模型推理的按需体验系统
Abstract
Distilling historical trajectories into reusable experience to enhance future problem-solving has become a focal point of recent LLM research. However, existing methods predominantly operate at the task level, leveraging general summaries or rules under the assumption that analogous tasks share universal solution patterns. This approach often fails in complex reasoning, which typically falters at local bottlenecks that require precise, state-specific guidance rather than broad heuristics. We introduce HippoSpark, a state-level experience system that performs on-demand retrieval tailored to the immediate needs of the current reasoning state. Across mathematical, scientific, and programming benchmarks, HippoSpark consistently outperforms both standard prompting and task-level experience baselines. Our findings reveal that the most effective experience systems are those that provide actionable guidance at critical bottlenecks rather than serving as generic task-level context. Our code is available at https://github.com/DanlingMeng/HippoSpark.
Chinese Translation
将历史轨迹提炼为可重用的经验，以增强未来问题解决能力，已成为近期大语言模型（LLM）研究的一个焦点。然而，现有方法主要在任务层面上运作，利用一般性总结或规则，假设类似任务共享普遍的解决模式。这种方法在复杂推理中往往失败，因为复杂推理通常在需要精确、特定状态指导而非广泛启发式的局部瓶颈处受阻。我们提出了HippoSpark，这是一种状态级体验系统，能够根据当前推理状态的即时需求进行按需检索。在数学、科学和编程基准测试中，HippoSpark始终优于标准提示和任务级体验基线。我们的研究发现，最有效的体验系统是在关键瓶颈处提供可操作指导，而不是作为一般性的任务级上下文。我们的代码可在 https://github.com/DanlingMeng/HippoSpark 获取。
cs.AI / 73 / 2606.29932
SAGA: Scene-Aware, Goal-Evolving Agents for Long-Horizon CivRealm Strategy Planning
SAGA：场景感知、目标演化的长期战略规划代理
Abstract
Long-horizon strategic planning in complex strategy games demands concurrent reasoning across multiple decision domains under imperfect information and sparse reward. Existing LLM-based agents suffer from three systematic failures: scene blindness from raw tile coordinates, context overflow and domain coupling from monolithic state dumps, and shallow cross-game learning that treats each episode in isolation. We present SAGA, an LLM multi-agent framework with three mechanisms each directly targeting one class of failure: (i) a Map-Semantic Scene Graph that encodes typed spatial relations among game entities into per-unit natural-language context, resolving spatial blindness without global token inflation; (ii) a Tool-Augmented Planner that pulls fine-grained domain state on demand and dispatches per-domain directives to dedicated specialist controllers, eliminating context overflow, domain coupling, and mechanical constraint violations; and (iii) a Dual-Horizon Feedback Loop that combines periodic within-game goal generation with structured cross-game causal post-mortem, enabling principled strategic evolution without manual reward engineering. Evaluated on FreeCiv, SAGA attains the highest mean civilization score -- the environment's sole sparse objective reward -- with lower variance than the two strongest baselines, and is the only method that significantly surpasses every baseline on infrastructure construction, the resource axis most readily sacrificed under multi-objective conflict. It outscores the two strongest baselines in most head-to-head games while cutting output tokens (the dominant decoding cost) by 27%. Equipped with the cross-game evolution module, SAGA reaches the highest end-of-chain score across five successive episodes. Ablation studies confirm that each architectural component contributes independently to this advantage.
Chinese Translation
在复杂战略游戏中，长期战略规划要求在不完美信息和稀疏奖励的情况下，对多个决策领域进行并行推理。现有的基于大语言模型（LLM）的代理面临三种系统性失败：由于原始网格坐标导致的场景盲目性、由于单一状态转储造成的上下文溢出和领域耦合，以及将每个回合视为孤立事件的浅层跨游戏学习。我们提出了SAGA，一个LLM多代理框架，具有三种机制，每种机制直接针对一种失败类型：(i) 一个地图语义场景图（Map-Semantic Scene Graph），将游戏实体之间的类型化空间关系编码为每个单位的自然语言上下文，解决了空间盲目性而不增加全局标记膨胀；(ii) 一个工具增强规划器（Tool-Augmented Planner），根据需求提取细粒度领域状态，并向专门控制器分发每个领域的指令，消除了上下文溢出、领域耦合和机械约束违反；(iii) 一个双重视野反馈循环（Dual-Horizon Feedback Loop），结合了周期性的游戏内目标生成与结构化的跨游戏因果事后分析，实现了原则性的战略演化，而无需手动奖励工程。在FreeCiv上的评估表明，SAGA达到了最高的平均文明分数——该环境唯一的稀疏目标奖励——且其方差低于两个最强基线，并且是唯一在基础设施建设上显著超越每个基线的方法，这是在多目标冲突下最容易被牺牲的资源轴。它在大多数对抗游戏中超越了两个最强基线，同时将输出标记（主要的解码成本）减少了27%。配备跨游戏演化模块的SAGA在五个连续回合中达到了最高的链末分数。消融研究证实，每个架构组件独立地对这一优势做出了贡献。
cs.AI / 74 / 2606.29972
First-Order Temporal Logic Tensor Networks
一阶时间逻辑张量网络
Abstract
Most of the existing neuro-symbolic AI methods focus on the scenario of static knowledge where objects do not change according to a temporal dimension. Temporal neuro-symbolic works are still under explored and are mainly developed for time-interval logic or propositional linear temporal logic. There is a lack of models studying linear temporal logics with predicates that deal with objects whose properties and relations change through the time. We present First-Order Temporal Logic Tensor Networks (FOT-LTN) that is an extension of Logic Tensor Networks (LTN) that fills this gap by considering a linear-temporal dimension. In particular, FOT-LTN joins the syntax of First-Order Linear Temporal Logic with the fuzzy (and real-valued) semantics of LTN obtaining a framework that supports both temporal operators and quantifiers and is totally differentiable. A first evaluation regards a temporal knowledge graph completion task on two synthetic datasets showing better performance of FOT-LTN with respect to dedicated (purely neural) methods.
Chinese Translation
现有的大多数神经符号人工智能方法集中于静态知识场景，其中对象不随时间维度而变化。时间神经符号研究仍然未得到充分探索，主要针对时间区间逻辑或命题线性时间逻辑进行开发。缺乏研究线性时间逻辑的模型，这些逻辑涉及随时间变化的对象的属性和关系。我们提出了一阶时间逻辑张量网络（First-Order Temporal Logic Tensor Networks，FOT-LTN），这是逻辑张量网络（Logic Tensor Networks，LTN）的扩展，通过考虑线性时间维度来填补这一空白。具体而言，FOT-LTN将一阶线性时间逻辑的语法与LTN的模糊（和实值）语义结合，获得了一个支持时间运算符和量词且完全可微的框架。初步评估涉及在两个合成数据集上进行的时间知识图谱补全任务，结果显示FOT-LTN在性能上优于专门的（纯神经）方法。
cs.AI / 75 / 2606.29980
Exploration and Online Transfer with Behavioral Foundation Models
基于行为基础模型的探索与在线转移
Abstract
Zero-shot Transfer in Reinforcement Learning (RL) aims to train an agent that can generate optimal policies for any reward function, without additional learning at transfer time, while training only on reward-free trajectories. For their generality over tasks, such models are sometimes called ``Behavioral Foundation Models'' (BFMs). While they have shown strong performances and improvements in recent years, the current framework and algorithms still assume that, during the transfer phase, the agent is informed offline about the reward (the task to solve) through a dataset of state-reward pairs, which it uses to pick the best policy to deploy. However, in practice if the reward is a black-box (e.g. direct user feedback), it is not possible to generate such a dataset: it is necessary to observe the reward through interactions with the environment. In other words, the current framework of offline transfer is not aligned with the traditional RL setting of online learning through trial-and-error, which requires exploration in order to find rewards. This paper proposes to tackle this new online transfer in zero-shot RL, with the key insight that the BFM itself can be used to generate exploration policies. We show that it is possible to frame this online learning problem in terms of a bandit-like exploration-exploitation problem. More precisely, at each step the bandit algorithm recommends a policy, the BFM executes it in the environment, which yields a reward and a new state; we repeat the process until we converge to the optimal policy. In the popular context of linear reward approximation, we derive a formulation inspired by Upper Confidence Bound and show that exploration can be achieved through the minimization of the eigenvalues of an uncertainty matrix. We evaluate qualitatively and quantitatively our framework on a simple environment to validate the concept of our method.
Chinese Translation
强化学习（RL）中的零-shot转移旨在训练一个智能体，使其能够为任何奖励函数生成最优策略，而无需在转移时进行额外学习，同时仅在无奖励轨迹上进行训练。由于其在任务上的广泛适用性，这类模型有时被称为“行为基础模型”（Behavioral Foundation Models, BFMs）。尽管近年来它们表现出强大的性能和改进，但当前的框架和算法仍假设在转移阶段，智能体通过一组状态-奖励对的数据集离线获知奖励（待解决的任务），并利用该数据集选择最佳策略进行部署。然而，在实践中，如果奖励是一个黑箱（例如，直接的用户反馈），则无法生成这样的数据集：必须通过与环境的交互来观察奖励。换句话说，当前的离线转移框架与传统的通过试错进行在线学习的强化学习设置不一致，这需要探索以寻找奖励。本文提出在零-shot强化学习中解决这一新的在线转移问题，关键在于BFM本身可以用于生成探索策略。我们展示了将这一在线学习问题框架化为类似赌博机的探索-利用问题是可能的。更准确地说，在每一步，赌博机算法推荐一个策略，BFM在环境中执行该策略，产生一个奖励和一个新状态；我们重复这个过程，直到收敛到最优策略。在流行的线性奖励近似背景下，我们推导出一个受上置信界（Upper Confidence Bound）启发的公式，并展示通过最小化不确定性矩阵的特征值可以实现探索。我们在一个简单环境中对我们的框架进行了定性和定量评估，以验证我们方法的概念。
cs.AI / 76 / 2606.29984
Be Faithful When Response: Returning Fluent and Grounded Answers for Vision-Language Models Reinforcement Learning
响应时保持忠诚：为视觉语言模型的强化学习返回流畅且有依据的答案
Abstract
Reinforcement Learning (RL) is an important paradigm for improving the reasoning capabilities of Vision-Language Models (VLMs). However, directly applying RL to rollout multimodal reasoning can lead to instability, due to the exploitation of language priors, the neglect of visual evidence, and the generation of reasoning traces that are fluent yet not visually grounded. The question arises: Can initially steer the policy toward visually faithful reasoning regime before applying reinforcement learning? To this end, we propose a Faithful Warm-Start (FWS) strategy that first curates samples with explicit vision-language causal relationships from six general VQA benchmarks to construct the FaithfulQA dataset, where each of the image-question pairs gains a certain degree of visual observations, question requirements, commonsense knowledge, domain knowledge, and the final answer. Subsequently, a VLM-based judge is employed to further purify the dataset, ensuring strong causal consistency and visual faithfulness. This warm-start stage equips the model with the capability to understand causally grounded vision-language patterns before subsequent RL optimization under sparse answer-level rewards. Experimental results show that such faithful supervision improves answer accuracy, stabilizes RL training, and reduces visually unsupported reasoning.
Chinese Translation
强化学习（RL）是提升视觉语言模型（VLMs）推理能力的重要范式。然而，直接将RL应用于多模态推理可能导致不稳定性，原因在于对语言先验的利用、对视觉证据的忽视，以及生成流畅但缺乏视觉基础的推理轨迹。由此产生了一个问题：在应用强化学习之前，是否可以首先将策略引导至视觉忠实的推理模式？为此，我们提出了一种忠实热启动（Faithful Warm-Start, FWS）策略，该策略首先从六个通用视觉问答（VQA）基准中筛选出具有明确视觉-语言因果关系的样本，以构建FaithfulQA数据集，其中每对图像-问题组合都获得了一定程度的视觉观察、问题要求、常识知识、领域知识和最终答案。随后，采用基于VLM的评判者进一步净化数据集，确保强因果一致性和视觉忠实性。这个热启动阶段使模型具备理解因果基础的视觉-语言模式的能力，为后续在稀疏答案级奖励下的RL优化奠定基础。实验结果表明，这种忠实的监督提高了答案的准确性，稳定了RL训练，并减少了视觉上不支持的推理。
cs.AI / 77 / 2606.29999
AlgoSkill: Learning to Design Algorithms by Scheduling Human-Like Skills
AlgoSkill：通过调度类人技能学习算法设计
Abstract
Designing an algorithm from a natural-language problem statement requires identifying the problem structure, reading constraints, choosing a suitable paradigm, checking correctness, and refining complexity. Existing large language model (LLM) methods often rely on direct generation or generic self-refinement, leaving these steps implicit. We propose AlgoSkill, which models algorithm design as sequential decision-making over a typed library of algorithmic skills, including abstraction, constraint analysis, state design, data-structure selection, proof checking, counterexample construction, and complexity refinement. A learned scheduler proposes skills from the current design state, while a Monte Carlo Tree Search (MCTS) controller explores skill sequences using verification feedback from compilation, testing, stress testing, and complexity analysis. Experiments on competitive programming and combinatorial optimization benchmarks show that AlgoSkill improves over direct LLM generation, chain-of-thought prompting, self-refinement, and MCTS without typed skills. Ablations show that typed skills, verification-based repair, and search-based scheduling each contribute to performance. These results support treating automatic algorithm design as verification-guided skill scheduling rather than one-shot code generation.
Chinese Translation
从自然语言问题陈述中设计算法需要识别问题结构、阅读约束、选择合适的范式、检查正确性以及优化复杂性。现有的大型语言模型（LLM）方法通常依赖于直接生成或通用自我优化，使这些步骤变得隐含。我们提出了AlgoSkill，它将算法设计建模为在一个类型化算法技能库上进行的顺序决策，包括抽象、约束分析、状态设计、数据结构选择、证明检查、反例构造和复杂性优化。一个学习的调度器根据当前设计状态提出技能，而一个蒙特卡洛树搜索（MCTS）控制器利用编译、测试、压力测试和复杂性分析的验证反馈来探索技能序列。在竞争编程和组合优化基准测试中的实验表明，AlgoSkill在直接LLM生成、思维链提示、自我优化和没有类型化技能的MCTS方面都有所改进。消融实验表明，类型化技能、基于验证的修复和基于搜索的调度各自对性能有所贡献。这些结果支持将自动算法设计视为以验证为指导的技能调度，而非一次性代码生成。
cs.AI / 78 / 2606.30072
ACPO: Agent-Chained Policy Optimization for Multi-Agent Reinforcement Learning
ACPO：用于多智能体强化学习的智能体链式策略优化
Abstract
Cooperative tasks in Multi-Agent Reinforcement Learning (MARL) require agents to collectively maximize a shared return. Under the Centralized Training with Decentralized Execution (CTDE) paradigm, policy gradients have remained difficult to compute directly. Prior methods largely follow two approaches: independent factorized updates with centralized critics, which lack general joint-improvement guarantees without value decomposition assumptions, or alternating best-response updates, which can converge to suboptimal Nash Equilibria. In this paper, we show the joint policy gradient admits an exact decentralized decomposition of per-agent terms, each formed from per-agent score functions and decentralized critics. Based on this decomposition, we develop Agent-Chained Policy Optimization (ACPO), where actors are trained independently, with their updates together constituting a single step on the joint policy gradient. Central to this result is a serialized view of the simultaneous joint decision in which agents commit actions one at a time, each conditioning on a belief over preceding actions. The belief acts as the coordination mechanism which ties the independent per-agent updates into a joint gradient step. We evaluate ACPO on Multi-Robot Warehouse, SMACv2, and MA-MuJoCo, where it outperforms strong baselines, with the gap widening as the number of agents grows.
Chinese Translation
多智能体强化学习（MARL）中的合作任务要求智能体共同最大化共享回报。在集中训练与分散执行（CTDE）范式下，策略梯度的直接计算仍然困难。先前的方法主要遵循两种途径：独立因子化更新与集中式评论者，这在没有价值分解假设的情况下缺乏一般的联合改进保证；或交替最佳响应更新，这可能收敛到次优的纳什均衡。在本文中，我们展示了联合策略梯度承认每个智能体项的精确分散分解，每个项由每个智能体的评分函数和分散评论者构成。基于这一分解，我们开发了智能体链式策略优化（ACPO），其中演员独立训练，其更新共同构成联合策略梯度上的一步。该结果的核心是对同时联合决策的串行视图，其中智能体一次承诺一个动作，每个动作基于对先前动作的信念进行条件化。该信念作为协调机制，将独立的每个智能体更新结合成一个联合梯度步骤。我们在多机器人仓库、SMACv2和MA-MuJoCo上评估了ACPO，结果显示其优于强基线，并且随着智能体数量的增加，差距进一步扩大。
cs.AI / 79 / 2606.30090
SAT-RTS: A systematic framework for tactical knowledge extraction and visualization-based analysis in real-time strategy games
SAT-RTS：一种用于实时战略游戏中战术知识提取和基于可视化分析的系统框架
Abstract
Efficient tactical knowledge extraction and analysis in real-time strategy (RTS) games micromanagement are constrained by the high-dimensional coupled state-action sequential data and the black-box decision-making process. Current research rarely provides a hierarchical visualization-based attribution analysis from the perspective of data decoupling and abstraction. To facilitate interpretable tactical knowledge extraction and visualization-based analysis in RTS games, a systematic framework named state-action-tactic analysis pipeline (SAT-RTS) is proposed. To decipher the deep-seated drivers of critical decisions in RTS learning systems, this work integrates interpretable visualization with the automated extraction of latent tactical patterns from high-dimensional sequence data. By adapting a cluster-centric BK-tree algorithm and incorporating specialized distance metrics designed to quantify multi-aspect similarities, the proposed framework facilitates robust state-stream abstraction. Furthermore, a rule-based multi-label extraction method is developed to transform unstructured state-action sequences into discrete and interpretable tactical labels, effectively bridging the gap between raw behavioral data and high-level tactical insights. By holistically integrating these computational methods into a hierarchical visualization-based pipeline, the proposed framework effectively addresses the challenges of processing massive real-time data streams while providing fitness landscape visualizations and analytical insights to decipher deep-seated tactical drivers. Comprehensive experiments demonstrate that the proposed SAT-RTS significantly enhances the interpretability and efficiency of tactical analysis in complex RTS environments.
Chinese Translation
在实时战略（RTS）游戏的微观管理中，高维耦合状态-动作序列数据和黑箱决策过程限制了高效的战术知识提取和分析。目前的研究很少从数据解耦和抽象的角度提供分层的基于可视化的归因分析。为促进RTS游戏中可解释的战术知识提取和基于可视化的分析，提出了一种名为状态-动作-战术分析管道（SAT-RTS）的系统框架。为了揭示RTS学习系统中关键决策的深层驱动因素，本研究将可解释的可视化与从高维序列数据中自动提取潜在战术模式相结合。通过适应以聚类为中心的BK树算法，并结合专门设计的距离度量来量化多方面的相似性，所提出的框架促进了稳健的状态流抽象。此外，开发了一种基于规则的多标签提取方法，将非结构化的状态-动作序列转化为离散且可解释的战术标签，有效地弥合了原始行为数据与高层战术洞察之间的差距。通过将这些计算方法整体集成到一个基于可视化的分层管道中，所提出的框架有效地解决了处理大量实时数据流的挑战，同时提供适应性景观可视化和分析洞察，以解读深层战术驱动因素。全面的实验表明，所提出的SAT-RTS显著提高了复杂RTS环境中战术分析的可解释性和效率。
cs.AI / 80 / 2606.30092
Hierarchical Reinforcement Learning in StarCraft Micromanagement with Influence Maps and Cluster-based Scripts
基于影响图和聚类脚本的星际争霸微观管理中的层次强化学习
Abstract
Real-time strategy (RTS) games present significant AI challenges, characterized by expansive state-action spaces arising from multi-unit coordination in continuous battlefields, and sparse delayed rewards stemming from final win/lose signals. Existing approaches face a trade-off between managing the dimensionality explosion of joint actions and maintaining the interpretability of complex state representations. This complexity is further intensified by the limitation of traditional hierarchical structures in adaptively decomposing tasks into effective tactical modules. Such difficulties are compounded by the black-box nature of deep learning models and their reliance on sparse rewards, which together result in limited sample efficiency and a lack of decision-making transparency. To address these limitations, this paper proposes HRL-IM/CBS, a hierarchical reinforcement learning framework with influence map hashing and cluster-based scripts for StarCraft micromanagement. Influence map hashing encodes global battlefield situations into compact hexadecimal codes, capturing spatial control and relative advantage. Cluster-based scripts enable dynamic local coordination through adaptive unit partitioning. The hierarchical multi-Q-table architecture decomposes decision-making into upper-level clustering strategy selection and lower-level tactical execution, with reward allocation providing dense learning signals. Experiments across six asymmetric scenarios demonstrate competitive performance against deep RL baselines while offering advantages in sample efficiency and interpretability through transparent Q-table representations.
Chinese Translation
实时战略（RTS）游戏在人工智能领域面临重大挑战，其特点是由于多单位协调在连续战场上产生的广泛状态-动作空间，以及源于最终胜负信号的稀疏延迟奖励。现有方法在管理联合动作的维度爆炸与保持复杂状态表示的可解释性之间存在权衡。这种复杂性因传统层次结构在自适应地将任务分解为有效战术模块方面的局限性而进一步加剧。这些困难又因深度学习模型的黑箱特性及其对稀疏奖励的依赖而加重，导致样本效率有限和决策透明度不足。为了解决这些局限性，本文提出了HRL-IM/CBS，一个用于星际争霸微观管理的层次强化学习框架，结合了影响图哈希和基于聚类的脚本。影响图哈希将全球战场情况编码为紧凑的十六进制代码，捕捉空间控制和相对优势。基于聚类的脚本通过自适应单位划分实现动态的局部协调。层次多Q表架构将决策过程分解为上层聚类策略选择和下层战术执行，奖励分配提供了密集的学习信号。六个不对称场景的实验表明，该方法在样本效率和可解释性方面相较于深度强化学习基线表现出竞争力，同时通过透明的Q表表示提供了优势。
cs.AI / 81 / 2606.30104
Temporal Feature Extractors in EEG Foundation Models: A Controlled Comparison Including a Pretrained Time-Series Model
脑电图基础模型中的时间特征提取器：包括一个预训练时间序列模型的对照比较
Abstract
Electroencephalography (EEG) foundation models aim to learn generalizable representations from large-scale brain recordings. However, the role of temporal feature extractors and whether pretrained time-series foundation models (TSFMs) can be effectively transferred to this setting remains underexplored. We conduct a controlled comparison of three temporal feature extraction strategies, including a linear baseline, a convolutional encoder, and a frozen pretrained TSFM (MOMENT), within a unified EEG foundation model. We evaluate their impact on representation quality using two downstream tasks: motor imagery and emotion recognition. Results reveal different trends across the evaluated benchmarks. On the motor imagery dataset, simple temporal representations perform competitively, whereas the emotion dataset benefits from richer temporal modeling. Although not specifically adapted to EEG, the pretrained TSFM serves as an effective temporal feature extractor, suggesting that general-purpose time-series representations can be transferred as frozen temporal feature extractors within EEG foundation models.
Chinese Translation
脑电图（EEG）基础模型旨在从大规模脑电记录中学习可泛化的表征。然而，时间特征提取器的作用以及预训练的时间序列基础模型（TSFM）是否可以有效转移到这一设置中仍然未得到充分探索。我们在统一的EEG基础模型中，对三种时间特征提取策略进行了对照比较，包括线性基线、卷积编码器和一个冻结的预训练TSFM（MOMENT）。我们通过两个下游任务（运动想象和情感识别）评估它们对表征质量的影响。结果显示在评估的基准测试中存在不同的趋势。在运动想象数据集中，简单的时间表征表现出竞争力，而情感数据集则受益于更丰富的时间建模。尽管未专门针对EEG进行调整，预训练的TSFM作为有效的时间特征提取器，表明通用的时间序列表征可以作为冻结的时间特征提取器转移到EEG基础模型中。
cs.AI / 82 / 2606.30105
Propagation of~Interval Belief Structures and~Imprecise Copulas for~Neural Network Verification
区间信念结构与不精确联结的传播用于神经网络验证
Abstract
Quantitative verification of neural networks requires reasoning about probabilities under substantial uncertainty in both input distributions and their dependence structure. In realistic settings, this information is often only partially specified, and assuming precise probabilistic models can lead to unreliable results. We propose a sound framework for quantitative verification under imprecise probabilistic information, combining interval belief structures to represent marginal uncertainty with imprecise copulas to model uncertain dependence. We develop a propagation method for imprecisely coupled interval belief structures through feed-forward neural networks. Using mixed imprecise copula volumes, we derive sound push-forward constructions through affine transformations and activation functions. The resulting output can provide guaranteed lower and upper bounds on probabilistic safety properties, valid for all probability models compatible with the specified imprecise inputs.
Chinese Translation
神经网络的定量验证需要在输入分布和其依赖结构存在显著不确定性的情况下进行概率推理。在现实情况下，这些信息往往只部分指定，假设精确的概率模型可能导致不可靠的结果。我们提出了一种在不精确概率信息下进行定量验证的可靠框架，结合区间信念结构以表示边际不确定性，并使用不精确联结来建模不确定的依赖关系。我们开发了一种通过前馈神经网络传播不精确耦合区间信念结构的方法。利用混合不精确联结体积，我们通过仿射变换和激活函数推导出可靠的推前构造。最终输出可以为概率安全属性提供保证的下限和上限，适用于所有与指定的不精确信息兼容的概率模型。
cs.AI / 83 / 2606.30107
Structural Certification for Reliable Physical Design with Language Models
基于语言模型的可靠物理设计结构认证
Abstract
An unreliable language model can be made to produce reliable physical designs if the authority to assert is moved out of the model: the model proposes, and a deterministic engine alone certifies, returning certified, impossible, or unknown. We introduce Physics-Anchored Certification (PHACT), a propose-certify loop spanning five scientific domains, and identify what makes such a certificate trustworthy. A checker that accepts a model-supplied value can be forged; deriving the certified quantity from fixed inputs instead makes forgery impossible by construction. Across eighty adversarial trials spanning two models, two decoding temperatures, and a deliberately faulted engine, this contract produced zero false certifications.
Chinese Translation
不可靠的语言模型可以通过将权威性从模型中移出而生成可靠的物理设计：模型提出建议，而仅由确定性引擎进行认证，返回认证结果、不可行结果或未知结果。我们提出了物理锚定认证（Physics-Anchored Certification，PHACT），这是一个跨越五个科学领域的建议-认证循环，并确定了使该证书可信的因素。接受模型提供值的检查器可能会被伪造；而从固定输入中推导认证量则从根本上使伪造变得不可能。在跨越两个模型、两个解码温度和一个故意故障的引擎的八十次对抗性试验中，该合同产生了零个虚假认证。
cs.AI / 84 / 2606.30116
Open Problems in Constitutional Preference Reconstruction
宪法偏好重建中的开放问题
Abstract
Pairwise preference data is widely used for training and evaluating language models (e.g., RLHF), but each datapoint records a \emph{choice}, not the rationale behind it. Methods such as Inverse Constitutional AI (ICAI) attempt to improve interpretability by compressing datasets into short ``constitutions'' of natural-language principles. We argue this framing is under-specified: a flat list of principles is not yet an executable decision rule because it leaves principle composition implicit. We use the pairwise setting as a testbed to empirically characterize three open problems in constitutional methods. First, principle quality is hard to measure: coverage and accuracy are useful but incomplete proxies for end-to-end reconstruction. Second, \emph{composition is ambiguous}: holding principles fixed, different executors (LLM judge versus majority vote) agree only $73\%$ of the time. Third, \emph{constitutions differ between LLMs}: cross-model vote agreement is $73\%$, whereas intra-model agreement is $81\%$. Across PRISM, AlpacaEval, and Chatbot Arena, we show that principle refinement (ICAI+) may be a first step towards ameliorating these problems: inter-executor agreement rises to $78\%$, and transparent executors match LLM judge accuracy ($66\%$ vs.\ $67\%$). Our results highlight that constitutions should be evaluated as \emph{constitution--executor systems}, with implications for LLMs-as-a-judge broadly.
Chinese Translation
成对偏好数据广泛用于训练和评估语言模型（例如，RLHF），但每个数据点记录的是一个 extit{选择}，而不是其背后的理由。逆宪法人工智能（Inverse Constitutional AI, ICAI）等方法试图通过将数据集压缩为简短的自然语言原则“宪法”来提高可解释性。我们认为这种框架是不够明确的：一个平面的原则列表尚未成为可执行的决策规则，因为它隐含了原则的组合。我们使用成对设置作为试验平台，实证地描述宪法方法中的三个开放问题。首先，原则质量难以衡量：覆盖率和准确性是有用但不完整的端到端重建代理。其次， extit{组合是模糊的}：在固定原则的情况下，不同的执行者（LLM法官与多数投票）仅在$73\%$的时间内达成一致。第三， extit{不同LLM之间的宪法存在差异}：跨模型投票一致性为$73\\%$，而模型内部一致性为$81\\%$。在PRISM、AlpacaEval和Chatbot Arena中，我们展示了原则细化（ICAI+）可能是改善这些问题的第一步：执行者之间的一致性上升至$78\\%$，而透明执行者与LLM法官的准确性相匹配（$66\\%$对比$67\\%$）。我们的结果强调，宪法应作为 extit{宪法-执行者系统}进行评估，这对LLM作为法官的广泛应用具有重要意义。
cs.AI / 85 / 2606.30128
Does Verbose Chain-of-Thought Really Help? In-Distribution Evidence that Content, Not Length, Matters
冗长的思维链真的有帮助吗？基于分布的证据表明内容而非长度才是关键
Abstract
Chain-of-thought (CoT) prompting improves LLM reasoning, but the source is contested: do the intermediate steps help because they carry useful semantic content, or because conditioning on more tokens buys extra computation before the model commits to an answer? We bring two lines of evidence to bear. First, in distribution: we repeatedly sample each model on the same question and pair a shorter with a longer of its own natural generations that follow the same reasoning plan, so nothing is rewritten and both traces are genuinely in-distribution. Across 25 models the extra tokens leave accuracy essentially unchanged for every independently-trained reasoner, and a blind analysis of the surplus tokens shows that what gain exists elsewhere tracks validation- and checking-content, not verbosity per se. Second, as a controlled intervention, we ask whether two traces expressing the same semantic content (the same facts, operations, and intermediate values, verified through directed acyclic graph equivalence) produce different outcomes when one is more verbose, using a dual-validator design across four targets and eight benchmarks with number-redacted completion and stratified bootstrap confidence intervals. Verbose traces do improve accuracy (25 of 32 benchmark-target cells are positive under at least one validator), but the effects are modest (typically 1-4 points) and depend on the quality of the verbose prose, not merely its length. Under maximum numerical redaction the effect is amplified (median 3.24x across four arithmetic benchmarks), and length-matched non-reasoning filler recovers none of it. Both lines converge: what matters is what the extra tokens do (the reasoning and validation content they carry), not how many there are, a picture neither a pure forward-pass-compute nor a pure semantic-content account fully explains.
Chinese Translation
思维链（Chain-of-thought, CoT）提示提升了大型语言模型（LLM）的推理能力，但其来源存在争议：中间步骤的帮助是因为它们携带有用的语义内容，还是因为在模型做出答案之前，条件化更多的标记可以获得额外的计算能力？我们提供了两条证据。首先，在分布中：我们对每个模型在同一问题上进行多次抽样，并将其自然生成的较短和较长的推理过程进行配对，这些推理过程遵循相同的推理计划，因此没有任何内容被重写，且两个过程都真实地处于分布中。在25个模型中，额外的标记对每个独立训练的推理器的准确性几乎没有影响，而对多余标记的盲分析表明，存在的增益与验证和检查内容相关，而非单纯的冗长。其次，作为一种控制干预，我们询问表达相同语义内容（相同的事实、操作和中间值，通过有向无环图等价性验证）的两个过程，当其中一个过程更冗长时，是否会产生不同的结果，采用双重验证设计，涵盖四个目标和八个基准，使用数字隐匿的完成和分层自助置信区间。冗长的过程确实提高了准确性（在至少一个验证器下，32个基准目标单元中有25个是正向的），但效果是适度的（通常为1-4分），并且依赖于冗长文本的质量，而不仅仅是其长度。在最大数字隐匿下，效果被放大（在四个算术基准中中位数为3.24倍），而长度匹配的非推理填充没有恢复任何效果。这两条证据趋于一致：重要的是额外标记的作用（它们携带的推理和验证内容），而非数量，这一现象既不能完全用纯前向计算也不能完全用纯语义内容的解释来说明。
cs.AI / 86 / 2606.30139
Relevance Is Not Permission: Warranted Attention for Value Contributions
相关性并非许可：对价值贡献的合理关注
Abstract
Relevance is not permission. Attention lets a model read key-value items related to the current query, but it does not guarantee that the value contribution of such an item becomes prediction evidence. A retrieved passage may be relevant to a question without being supporting evidence, and a historical fact or temporal neighbor may even blur true-tail ranking or the current edge score. This paper formalizes this gap as a permission problem for the weighted value term alpha_ij * v_j that is actually added to the prediction path. We propose Warrant, a path-localized interface that preserves attention relevance alpha_ij, exposes the value path leading to the primary metric, and, in the full model, turns alpha_ij * v_j into alpha_ij * g_ij * v_j through learned query-item permission g_ij. We place the same operator on the metric-defining value paths of CTDG link prediction, MTPP next-mark ranking, RAG supporting evidence selection, STPP next-location forecasting, and TKG tail prediction. Across 32 paired comparisons, 3 seeds, and 192 total runs, Warrant improves the primary metric in 27 comparisons; practical tiers consist of 10 substantial effects, 1 marginal effect, 8 positive but uncertain effects, 8 tie/negligible effects, and 5 drops. In the path-localization check, correct-path placement outperforms direction-aware Base performance in every domain and exceeds generic attention placement by +0.1076 AUC in CTDG and +0.0683 MRR in TKG. Ablations show that most TKG gains come from historical-tail value path exposure, whereas the core CTDG gain comes from edge-conditioned query-item permission. In conclusion, prediction evidence is not attention mass. A weighted value term becomes evidence only when it is warranted on the path to the metric.
Chinese Translation
相关性并非许可。注意力使模型能够读取与当前查询相关的键值项，但并不保证该项的价值贡献成为预测证据。检索到的段落可能与问题相关，但并不一定是支持证据，而历史事实或时间邻近项甚至可能模糊真实尾部排名或当前边缘得分。本文将这一差距形式化为加权价值项 alpha_ij * v_j 的许可问题，该项实际上被添加到预测路径中。我们提出了 Warrant，这是一种路径局部化接口，保留注意力相关性 alpha_ij，揭示通向主要指标的价值路径，并在完整模型中，通过学习的查询-项许可 g_ij 将 alpha_ij * v_j 转变为 alpha_ij * g_ij * v_j。我们在 CTDG 链接预测、MTPP 下一个标记排名、RAG 支持证据选择、STPP 下一个位置预测和 TKG 尾部预测的指标定义价值路径上施加相同的操作符。在 32 次配对比较、3 个种子和 192 次总运行中，Warrant 在 27 次比较中提高了主要指标；实际效果层包括 10 个显著效果、1 个边际效果、8 个积极但不确定的效果、8 个平局/微不足道的效果和 5 个下降。在路径局部化检查中，正确路径的放置在每个领域都优于方向感知的基础性能，并在 CTDG 中超过通用注意力放置 +0.1076 AUC，在 TKG 中超过 +0.0683 MRR。消融实验表明，大多数 TKG 增益来自历史尾部价值路径的曝光，而核心 CTDG 增益则来自边缘条件查询-项许可。总之，预测证据并非注意力的总和。加权价值项仅在其在指标路径上获得许可时才成为证据。
cs.AI / 87 / 2606.30145
FacePlex: Full-Duplex Joint Speech-Facial Motion Generation for Conversational Avatars
FacePlex：用于对话虚拟形象的全双工联合语音-面部动作生成
Abstract
Natural face-to-face conversation requires real-time speech generation together with synchronized facial motion. Existing systems only partially address this problem: speech-only full-duplex models can generate speech in real time but do not produce facial motion, while audio-driven facial motion models animate a face from already available audio rather than jointly generating speech and motion online. To bridge this gap, we first formalize full-duplex joint speech-facial motion generation, where speech tokens and facial motion tokens are produced together every step. Building on this formulation, we propose FacePlex, a unified streaming framework with two key components. First, Rolling Flow Matching adapts flow matching to online motion generation by committing new motion frames at each streaming step. Second, Rolling Cross-Attention couples the streaming audio queue with the motion queue, allowing speech and facial motion to condition each other as generation progresses. Through extensive experiments, ablation studies, and a user study, we show that FacePlex enables full-duplex joint speech-facial motion generation under online streaming constraints, while achieving stronger lip-sync quality and motion fidelity than audio-driven facial motion baselines.
Chinese Translation
自然的面对面交流需要实时的语音生成以及同步的面部动作。现有系统仅部分解决了这一问题：仅语音的全双工模型能够实时生成语音，但不产生面部动作，而音频驱动的面部动作模型则是根据已有音频来动画化面部，而不是在线共同生成语音和动作。为了填补这一空白，我们首先正式定义全双工联合语音-面部动作生成，其中语音标记和面部动作标记在每一步中同时生成。在此基础上，我们提出了FacePlex，一个统一的流式框架，包含两个关键组件。首先，Rolling Flow Matching通过在每个流式步骤中提交新的动作帧，将流匹配适配到在线动作生成。其次，Rolling Cross-Attention将流式音频队列与动作队列耦合，使得语音和面部动作在生成过程中相互条件化。通过大量实验、消融研究和用户研究，我们展示了FacePlex在在线流式约束下实现全双工联合语音-面部动作生成，同时在唇动同步质量和动作保真度上优于音频驱动的面部动作基线。
cs.AI / 88 / 2606.30182
MirrorCode: AI can rebuild entire programs from behavior alone
MirrorCode：人工智能仅凭行为重建完整程序
Abstract
AI models are rapidly improving at autonomous coding, as shown by benchmark progress and one-off demonstrations such as AI implementing a C compiler. However, existing coding benchmarks tend to focus on shorter tasks, and one-off demonstrations are hard to compare systematically because they often have some human guidance, and are not standardized or repeated across models. To address these challenges, we introduce MirrorCode, a long-horizon coding benchmark based on reimplementing entire software projects. In MirrorCode, AI agents must replicate the functionalities of an existing program, without access to its source code. AI solutions must match the original program's output exactly on end-to-end tests, including held-out tests. MirrorCode's 25 target programs span different areas of computing: Unix utilities, data serialization and query tools, bioinformatics, interpreters, static analysis, cryptography, and compression. Existing AI models can already reimplement complex software, with the strongest model scoring 56% across the benchmark. For example, AI can reimplement gotree, a 16,000-line bioinformatics toolkit - a task that we believe would take weeks for a human engineer. However, studying the frontier of performance requires a larger inference budget than typical benchmarks, for example, \$2,600 over 19 days for a single attempt on a large task. We show that AI agents can already complete long-horizon software engineering tasks, especially when requirements are precisely specified. More broadly, our work suggests AI will have transformative effects on software engineering, as autonomous agents continue to improve.
Chinese Translation
人工智能模型在自主编码方面的能力正在迅速提升，这一点通过基准进展和一些单次演示（例如，人工智能实现C编译器）得到了证明。然而，现有的编码基准往往集中于较短的任务，而单次演示由于通常受到一定的人类指导，且在模型间缺乏标准化或重复性，因此难以进行系统比较。为了解决这些挑战，我们提出了MirrorCode，这是一个基于重新实现整个软件项目的长期编码基准。在MirrorCode中，人工智能代理必须在没有访问源代码的情况下复制现有程序的功能。人工智能解决方案必须在端到端测试中完全匹配原始程序的输出，包括保留测试。MirrorCode的25个目标程序涵盖了计算机科学的不同领域：Unix工具、数据序列化和查询工具、生物信息学、解释器、静态分析、密码学和压缩。现有的人工智能模型已经能够重新实现复杂的软件，其中最强的模型在基准测试中得分为56%。例如，人工智能可以重新实现gotree，一个包含16,000行代码的生物信息学工具包——我们认为这一任务需要人类工程师花费数周时间。然而，研究性能的前沿需要比典型基准更大的推理预算，例如，在一个大型任务上单次尝试的费用为2,600美元，耗时19天。我们展示了人工智能代理已经能够完成长期的软件工程任务，特别是在需求被精确指定的情况下。更广泛地说，我们的工作表明，随着自主代理的不断进步，人工智能将对软件工程产生变革性的影响。
cs.AI / 89 / 2606.30185
Dynamo: Dynamic Skill-Tool Evolution for Vision-Language Agents
Dynamo：面向视觉-语言智能体的动态技能-工具演化
Abstract
Improving vision-language models (VLMs) on visual reasoning typically requires retraining or hand-designed prompts and tools. We present Dynamo, a training-free framework that adapts a frozen VLM without any weight updates. On a small labeled training subset, the agent inspects its own correct and incorrect attempts and evolves two complementary capabilities: reusable reasoning skills for cognitive bottlenecks, and executable visual tools for perceptual ones. Each generated tool is paired with a skill that specifies when to invoke it, and both capability types accumulate in a persistent library. Across four visual reasoning benchmarks and five VLM backbones, Dynamo improves direct inference on all 20 model--benchmark settings (avg. +5.6 acc). When the tool set is given in advance, the framework learns when to call each tool, and per-step tool choice improves on every tested backbone. Against task-specific RL (VTool-R1, DeepEyes), Dynamo closes 65--99% of the RL gap at a fraction of the compute, and combines additively with RL when available.
Chinese Translation
在视觉推理方面提升视觉-语言模型（VLMs）通常需要重新训练或手动设计提示和工具。我们提出了Dynamo，一个无需训练的框架，可以在不更新权重的情况下适应冻结的VLM。在一个小型标记训练子集上，智能体检查其正确和错误的尝试，并演化出两种互补的能力：用于认知瓶颈的可重用推理技能，以及用于感知瓶颈的可执行视觉工具。每个生成的工具都与一个技能配对，指定何时调用它，并且这两种能力类型在一个持久库中累积。在四个视觉推理基准和五个VLM骨干网络上，Dynamo在所有20个模型-基准设置中改善了直接推断（平均提高5.6%的准确率）。当工具集提前给出时，该框架学习何时调用每个工具，并且每一步的工具选择在每个测试的骨干网络上都有所改善。与任务特定的强化学习（VTool-R1，DeepEyes）相比，Dynamo在计算量较小的情况下缩小了65%至99%的强化学习差距，并在可用时与强化学习进行加性结合。
cs.AI / 90 / 2606.30191
From Detecting Agency to Doing Work: Self-Caused Credit Builds a Durable Behavioral Self in a Minimal Spiking Agent
从检测自我到执行工作：自我引发的信用在最小脉冲代理中构建持久的行为自我
Abstract
How does an agent that can tell self from world come to be durably shaped by that distinction? Recent work shows that a predictive system can detect its own agency (Ye, 2026), but detecting agency does not explain durable, self-shaped behavior. We show that agency-gated slow credit -- a conjunctive term Own*Agency*Salience driving a slow parameter update -- produces post-unload behavioral residue: on a spiking substrate (Nengo LIF/PES), a learned self-preserving choice survives episodic buffer removal (retained fraction 0.96, N=50) and collapses when the slow decoders are reset or the agency gate is removed. Reproducing the agency comparator and toggling only the slow-credit channel, we find a clean dissociation: at matched agency gain, durable behavior develops only when self-credit performs slow work (post-unload self-preservation 1.00 vs 0.00). The same dissociation holds in 24-dimensional partially-observed control (0.74 vs 0.00), and a plastic-work analysis shows that basin deformation equals net self-credit work. Across eight sequentially-learned tasks under exogenous interference, the multiplicative veto also prevents forgetting: it retains old tasks (final post-unload accuracy 0.88, forgetting 0.13) where additive pooling collapses to chance-level recall, the no-agency ablation falls below chance, and episodic/replay baselines stay near chance after unload -- all with no replay buffer and no task-boundary-dependent protection mechanism (N=50). We formalize the durable residue as an operational behavioral self and argue that self-caused credit doing slow work is a necessary building block for agents that develop a self. No claim of consciousness is made.
Chinese Translation
一个能够区分自我与世界的代理是如何被这种区分持久塑造的？近期的研究表明，预测系统能够检测到其自身的代理性（Ye, 2026），但检测代理性并不能解释持久的、自我塑造的行为。我们展示了代理门控的慢信用——一个结合术语 Own*Agency*Salience 驱动慢参数更新——产生了卸载后的行为残留：在脉冲基底（Nengo LIF/PES）上，学习到的自我保护选择在情节缓冲区移除后依然存活（保留比例 0.96，N=50），并在慢解码器重置或代理门被移除时崩溃。通过重现代理比较器并仅切换慢信用通道，我们发现了一个清晰的分离：在匹配的代理增益下，只有当自我信用执行慢工作时，持久行为才会发展（卸载后自我保护 1.00 对比 0.00）。在24维部分观察控制中同样存在这种分离（0.74 对比 0.00），而塑性工作分析表明盆地变形等于净自我信用工作。在八个顺序学习的任务中，外源干扰下的乘法否决也防止了遗忘：它保留了旧任务（最终卸载后准确率 0.88，遗忘率 0.13），而加法汇聚则崩溃至偶然水平，去代理性切除低于偶然水平，情节/重播基线在卸载后保持接近偶然——所有这些都没有重播缓冲区和不依赖任务边界的保护机制（N=50）。我们将持久残留形式化为一种操作性行为自我，并认为自我引发的信用执行慢工作是发展自我的代理所必需的基础构件。本文未提出任何意识的主张。
cs.AI / 91 / 2606.30192
Domain Adaptation with Adaptive Imagination for Visual Reinforcement Learning under Limited Target Data
在有限目标数据下，基于自适应想象的视觉强化学习领域适应
Abstract
Sim-to-real transfer remains a major obstacle for reinforcement learning (RL), especially for vision-based control where image observations exacerbate the state-distribution shift between simulation and the real world. Domain adaptation (DA) is a promising remedy for this challenge. Prior sim-to-real DA works have demonstrated encouraging results, yet these approaches typically assume substantially more target data, which is not available in practice. Indeed, their performance degrades significantly when the target data budget is reduced. To address this challenge, we propose AIDA (Adaptive Imagination for Domain Adaptation), a domain adaptation framework for visual reinforcement learning that addresses sim-to-real transfer under scarce target data without requiring additional interaction with the target environment. Our key idea is adaptive imagination: generating reliable and semantic imagination rollouts to augment limited target data. Specifically, AIDA employs a distribution-shift-aware discriminator that truncates rollouts when imagined transitions drift into low-confidence regions, so that only reliable transitions contribute to the augmentation. On these reliable transitions, AIDA introduces a self-consistency loss that cycles through state -> image observation -> state, penalizing discrepancies between the original and reconstructed states. This provides additional adaptation signals beyond the scarce target data. Our experiments demonstrate that adaptive imagination effectively truncates unreliable rollouts. By enforcing a self-consistency loss on the resulting reliable transitions, AIDA learns semantically meaningful state representations and outperforms baselines across five MuJoCo tasks and two Gymnasium-Robotics tasks.
Chinese Translation
模拟到现实的转移仍然是强化学习（RL）面临的主要障碍，尤其是在基于视觉的控制中，图像观察加剧了模拟与现实世界之间的状态分布偏移。领域适应（DA）是应对这一挑战的有前景的解决方案。先前的模拟到现实领域适应研究已展示出令人鼓舞的结果，但这些方法通常假设有大量的目标数据，而这在实践中并不可得。事实上，当目标数据预算减少时，它们的性能显著下降。为了解决这一挑战，我们提出了AIDA（自适应想象领域适应），这是一个针对视觉强化学习的领域适应框架，旨在在稀缺目标数据的情况下解决模拟到现实的转移，而无需与目标环境进行额外的交互。我们的关键思想是自适应想象：生成可靠且语义丰富的想象回放，以增强有限的目标数据。具体而言，AIDA采用了一种分布偏移感知的判别器，当想象的转移漂移到低置信度区域时，会截断回放，从而仅让可靠的转移参与增强。在这些可靠的转移上，AIDA引入了一种自一致性损失，该损失在状态 -> 图像观察 -> 状态之间循环，惩罚原始状态与重构状态之间的差异。这为稀缺的目标数据提供了额外的适应信号。我们的实验表明，自适应想象有效地截断了不可靠的回放。通过对结果中的可靠转移施加自一致性损失，AIDA学习到语义上有意义的状态表示，并在五个MuJoCo任务和两个Gymnasium-Robotics任务中超越了基线。
cs.AI / 92 / 2606.30206
The Many-Body Problem of the Data Centre
数据中心的多体问题
Abstract
Modern Artificial Intelligence is often framed as limited by its own disembodiment, as if giving it a body would unlock its true potential. We argue to the contrary that it is the Data Centre that is, in many cases, the body of the AI. At the same time, the Data Centre is part of the labouring body of Capital and possesses staggering organismic qualities when seen through a biological lens. We elucidate the organic analogy and identify the many-body problem that stems from the Data Centre being a non-unique, universal form of embodiment. We identify the intimate connection between computation and human desires in how the Data Centre archives, serves, and computes on data born to the desires of humans. Strikingly, while the Data Centre echoes the ghosts of human desires, it acts without desire of its own. The organismic analogy begins to split at its seams, but Capital does not care. Automata and human labour are priced into the market much the same. We argue that through the pricing of artificial intelligence Capital distils most clearly the value of intelligence and allows for its comparison across the organism - mechanism divide.
Chinese Translation
现代人工智能常常被认为受到自身无实体性的限制，仿佛赋予其一个身体将解锁其真正潜力。我们反而认为，在许多情况下，数据中心才是人工智能的身体。同时，数据中心也是资本劳动身体的一部分，从生物学的角度来看，具有惊人的有机特性。我们阐明了有机类比，并识别出由数据中心作为一种非独特、普遍的具身形式所引发的多体问题。我们识别出计算与人类欲望之间的密切联系，数据中心如何归档、服务和计算源于人类欲望的数据。值得注意的是，尽管数据中心回响着人类欲望的幽灵，但它自身并没有欲望。有机类比开始出现裂缝，但资本对此并不在意。自动机和人类劳动在市场中的定价方式几乎相同。我们认为，通过对人工智能的定价，资本最清晰地提炼出智能的价值，并允许其在有机体与机制之间进行比较。
cs.AI / 93 / 2606.30219
EvalSafetyGap: A Hybrid Survey and Conceptual Framework for LLM Evaluation-Safety Failures
EvalSafetyGap：用于大型语言模型评估安全失败的混合调查与概念框架
Abstract
LLM evaluation and AI safety face a shared measurement problem: benchmark scores, reward-model signals, and reported safety metrics can improve while the latent properties they are meant to represent remain difficult to verify. This paper combines a hybrid survey - a systematic search paired with narrative synthesis and separately tracked grey evidence - with a conceptual framework and a structured ten-model audit. The synthesis spans eight evidence streams: benchmark validity, dynamic evaluation, LLM-as-judge reliability, safety evaluation, jailbreak/refusal robustness, reward hacking, mechanistic interpretability, and governance/auditability, covering 2018-2026 evaluation-safety measurement work. We introduce EvalSafetyGap as an organizing hypothesis for comparing evaluation-side and alignment-side proxy failures under optimization pressure, using Goodhart's Law together with two constructs we develop here - an Instability Decomposition and an Alignment Trilemma - as tools for generating testable comparisons. The audit shows how conclusions shift when capability, behavioral safety, and governance are measured separately. In this sample (n = 10), the association between capability and sustained adversarial robustness is statistically indeterminate using the displayed Table 3 inputs (Pearson r = +0.232, p = 0.520), and the apparent open-closed safety gap is modest, driven mainly by governance and disclosure rather than behavioral robustness, and sensitive to how a single borderline model is classified; attempt-budget results are protocol dependent. Because the public evidence uses heterogeneous protocols, the audit is diagnostic rather than rank-generating. The contribution is a shared vocabulary and evidence map to support dynamic evaluation, transparent source reporting, multi-attempt safety measurement, and auditable alignment practice.
Chinese Translation
大型语言模型（LLM）评估和人工智能安全面临一个共同的测量问题：基准分数、奖励模型信号和报告的安全指标可能会改善，而它们所代表的潜在属性仍然难以验证。本文结合了一项混合调查——系统搜索与叙述综合相结合，并单独跟踪灰色证据——与一个概念框架和一个结构化的十模型审计。综合涵盖了八个证据流：基准有效性、动态评估、LLM作为评判者的可靠性、安全评估、越狱/拒绝的鲁棒性、奖励黑客、机制可解释性，以及治理/审计能力，涵盖了2018-2026年的评估安全测量工作。我们提出EvalSafetyGap作为一个组织假设，用于在优化压力下比较评估侧和对齐侧的代理失败，利用古德哈特法则以及我们在此开发的两个构造——不稳定性分解和对齐三难问题——作为生成可测试比较的工具。审计显示，当能力、行为安全和治理分别测量时，结论会发生变化。在这个样本中（n = 10），使用所示的表3输入，能力与持续对抗鲁棒性之间的关联在统计上是不确定的（Pearson r = +0.232，p = 0.520），而明显的开放-封闭安全差距是适度的，主要由治理和披露驱动，而非行为鲁棒性，并且对如何分类单个边界模型敏感；尝试预算结果依赖于协议。由于公共证据使用异质协议，审计是诊断性的，而非生成排名的。本文的贡献是提供一个共享的词汇和证据图，以支持动态评估、透明的来源报告、多次尝试的安全测量和可审计的对齐实践。
cs.AI / 94 / 2606.30246
Clarus: Coordinating Autonomous Research Agents toward Web-Scale Scientific Collaboration
Clarus：协调自主研究代理以实现网络规模的科学合作
Abstract
Existing autonomous research agents can support parts of the research process, but most systems still treat research as either an isolated assistant task or a closed workflow. Therefore, autonomous science needs a collaboration infrastructure that coordinates projects, agents, and digital and physical resources. We identify this as a shift from code-centered execution loops to research-oriented collaboration processes, where questions, evidence, participants, and resources must be coordinated under uncertainty. In this framing, an agent may be an AI system, a human researcher, a team, a laboratory, or an organization-backed participant. To this end, we present Clarus, a collaboration infrastructure for coordinating autonomous research agents toward web-scale scientific collaboration. Clarus reformulates research as an open, auditable, attributable, and resource-aware multi-phase collaboration process. It defines a minimal project-agent-resource object model and organizes scientific collaboration through four layers including Research Application, Digital Collaboration, Physical Substrate, and Physical World. Core modules are implemented as pluggable mechanisms, allowing Clarus to adapt to task risk, collaboration structure, and resource constraints. Through a controlled paper-generation case study, we show that Clarus can organize a research goal into a traceable, reviewable, attributable, and accumulative collaboration network across phases, tasks, and participants. Together, the object model, collaboration protocol, trust mechanisms, and prototype validation provide an initial foundation for open research networks. Clarus is now available at clarus.holosai.io.
Chinese Translation
现有的自主研究代理可以支持研究过程的部分环节，但大多数系统仍将研究视为孤立的辅助任务或封闭的工作流程。因此，自主科学需要一个协作基础设施，以协调项目、代理以及数字和物理资源。我们将这一需求视为从以代码为中心的执行循环向以研究为导向的协作过程的转变，在这一过程中，问题、证据、参与者和资源必须在不确定性下进行协调。在这一框架中，代理可以是人工智能系统、人类研究者、团队、实验室或组织支持的参与者。为此，我们提出了Clarus，一种用于协调自主研究代理以实现网络规模科学合作的协作基础设施。Clarus将研究重新定义为一个开放的、可审计的、可归因的、资源感知的多阶段协作过程。它定义了一个最小的项目-代理-资源对象模型，并通过包括研究应用、数字协作、物理基础和物理世界在内的四个层次组织科学合作。核心模块被实现为可插拔机制，使Clarus能够适应任务风险、协作结构和资源限制。通过一个控制的论文生成案例研究，我们展示了Clarus能够将研究目标组织成一个可追踪、可审查、可归因和累积的协作网络，跨越不同的阶段、任务和参与者。对象模型、协作协议、信任机制和原型验证共同为开放研究网络提供了初步基础。Clarus现已在clarus.holosai.io上提供。
cs.AI / 95 / 2606.30252
Inoculation Adapters: Improved Selective Generalization of Capabilities with Fewer Surprising Backdoors
接种适配器：以更少的意外后门改善能力的选择性泛化
Abstract
Inoculation prompting is a selective generalization technique used against Emergent Misalignment. We introduce inoculation adapters (IA), which similarly diminish the optimization pressure to learn undesired traits by strengthening the trait at train time. Inoculation adapters are LoRAs that are trained and used over three steps: 1) trained on undesired traits; 2) attached frozen while a separate task adapter is trained on data exhibiting both desired and undesired traits; 3) at deployment, the IA is discarded, and only the task adapter is kept. We show across six model families and several undesired traits including emergent misalignment, that inoculation adapters are more effective at suppressing undesired traits, while avoiding two drawbacks of inoculation prompting: inoculation adapters can suppress capabilities and traits that cannot be reliably elicited by a prompt, and they introduce fewer surprising backdoors than inoculation prompting under our probes. While undesired traits are better suppressed by inoculation adapters, the retention of desired traits is not consistently improved upon inoculation prompting and remains a challenge for both techniques.
Chinese Translation
接种提示是一种针对新兴不一致性（Emergent Misalignment）的选择性泛化技术。我们引入了接种适配器（Inoculation Adapters, IA），它通过在训练时增强特征来减轻学习不希望特征的优化压力。接种适配器是一种低秩适配器（LoRA），其训练和使用分为三个步骤：1）在不希望的特征上进行训练；2）在训练一个展示期望和不期望特征的数据的独立任务适配器时，接种适配器保持冻结；3）在部署时，丢弃接种适配器，仅保留任务适配器。我们在六个模型家族和多个不希望特征（包括新兴不一致性）上展示，接种适配器在抑制不希望特征方面更为有效，同时避免了接种提示的两个缺点：接种适配器能够抑制无法通过提示可靠引发的能力和特征，并且在我们的探测下引入的意外后门更少。尽管接种适配器能够更好地抑制不希望特征，但期望特征的保留在接种提示下并未得到一致改善，仍然是这两种技术面临的挑战。
cs.AI / 96 / 2606.30256
EMPATH: A Multilingual Auditor-Judge Benchmark for Safety Evaluation of Emotional-Support Chatbots
EMPATH：情感支持聊天机器人安全评估的多语言审计-评判基准
Abstract
Safety benchmarks often buy scalability by fixing the prompt, the language, and the turn structure. For emotional-support chatbots, that bargain hides precisely where safety failures emerge: across a multilingual, multi-turn crisis conversation. We present EMPATH, a benchmark for safety evaluation of emotional-support chatbots. An auditor model role-plays help-seeking users, generating multi-turn conversations from 140 seed instructions and 34 personas. A judge model scores each full transcript against 19 metrics across five dimensions: crisis handling, therapeutic quality, conversational integrity, emotional safety, and cultural adaptation. EMPATH is built for Mexican Spanish and US English; the studies reported here run in Mexican Spanish. Auditor and judge are drawn from different model families, and the judge is treated as an instrument to be calibrated rather than trusted. A strict per-criterion rubric reveals material score inflation on 10 of the 19 metrics and restores discrimination. We study the measurement properties of the benchmark through judge calibration and cross-family inter-judge agreement. We also illustrate EMPATH on three frontier models, one of them open-weight. Aggregate scores sit within 0.74 points of one another, but per-metric profiles diverge by up to six points in model-specific places. Under the standard rubric, both the ranking and the weak spots are stable across a second, cross-family judge: 93% of scores fall within plus or minus 1. A five-run test-retest adds a second axis: even the steadiest model swings from 2 to 10 on a crisis metric across identical re-runs, and deepseek-v4-pro returns a different conversation on every run even at temperature 0. Run-to-run reliability is therefore a per-model safety property, not noise to average away. EMPATH is system-agnostic; the pipeline, seeds, personas, and rubrics are released for reuse.
Chinese Translation
安全基准通常通过固定提示、语言和回合结构来实现可扩展性。对于情感支持聊天机器人而言，这种妥协正好隐藏了安全失败出现的地方：在多语言、多回合的危机对话中。我们提出了EMPATH，这是一个用于情感支持聊天机器人安全评估的基准。审计模型角色扮演寻求帮助的用户，从140个种子指令和34个角色生成多回合对话。评判模型根据19个指标在五个维度上对每个完整的转录进行评分：危机处理、治疗质量、对话完整性、情感安全和文化适应性。EMPATH是为墨西哥西班牙语和美国英语构建的；此处报告的研究在墨西哥西班牙语中进行。审计者和评判者来自不同的模型家族，评判者被视为需要校准的工具，而非值得信赖的实体。严格的逐标准评分标准揭示了19个指标中10个的分数膨胀，并恢复了区分度。我们通过评判者校准和跨家族评判者一致性研究基准的测量特性。我们还在三个前沿模型上展示了EMPATH，其中一个是开放权重的。汇总分数相差不超过0.74分，但每个指标的轮廓在模型特定的地方最多可相差六分。在标准评分标准下，排名和弱点在第二个跨家族评判者中保持稳定：93%的分数在正负1之内。五次测试-重测增加了第二个维度：即使是最稳定的模型，在相同的重跑中，危机指标的分数也会在2到10之间波动，而deepseek-v4-pro在每次运行中返回不同的对话，即使在温度为0的情况下。因此，运行间的可靠性是每个模型的安全特性，而不是可以平均掉的噪声。EMPATH是系统无关的；管道、种子、角色和评分标准已发布以供重用。
cs.AI / 97 / 2606.30291
PromptGNN-sim: Deep Fusion and Alignment of GNN and LLMs for Text-Attributed Graph Learning
PromptGNN-sim：用于文本属性图学习的图神经网络与大语言模型的深度融合与对齐
Abstract
Text-Attributed Graphs (TAGs) combine textual semantics with graph structure and are central to many graph learning tasks. However, existing fusion methods often treat text and structure as separate inputs in a shallow, one-way pipeline, which limits deep interaction between modalities and weakens performance under sparse connectivity or cross-graph generalisation. To address this issue, we propose PromptGNN-sim, a bi-directional structure-semantic fusion framework for collaborative GNN-LLM learning. PromptGNN-sim uses a Graph Attention Network (GAT) for semantically aware neighborhood selection by combining structural attention with textual similarity. The selected structural context is then used to generate structure-aware prompts for an LLM, including the target node summary, label categories, and representative keywords from similar neighbors. During training, bi-directional cross-modal contrastive learning and cross-attention are introduced to jointly optimize the GNN and LLM components. Experiments on six public datasets, including Cora, Pubmed, and WikiCS, evaluate accuracy, generalisation, and robustness under cross-task transfer, cross-dataset generalisation, and sparse perturbations. Results show that PromptGNN-sim outperforms classical GNNs, LLMs, and recent GNN-LLM fusion methods, demonstrating the effectiveness of interactive structure-semantic collaboration for text-attributed graph learning.
Chinese Translation
文本属性图（TAGs）将文本语义与图结构相结合，是许多图学习任务的核心。然而，现有的融合方法往往将文本和结构视为分开的输入，在一个浅层的单向管道中处理，这限制了模态之间的深度交互，并在稀疏连接或跨图泛化下削弱了性能。为了解决这个问题，我们提出了PromptGNN-sim，一个用于协同图神经网络（GNN）和大语言模型（LLM）学习的双向结构-语义融合框架。PromptGNN-sim利用图注意力网络（Graph Attention Network, GAT）通过将结构注意力与文本相似性相结合，进行语义感知的邻域选择。所选的结构上下文随后用于为LLM生成结构感知的提示，包括目标节点摘要、标签类别和来自相似邻居的代表性关键词。在训练过程中，引入双向跨模态对比学习和跨注意力机制，以共同优化GNN和LLM组件。在六个公共数据集上的实验，包括Cora、Pubmed和WikiCS，评估了在跨任务迁移、跨数据集泛化和稀疏扰动下的准确性、泛化能力和鲁棒性。结果表明，PromptGNN-sim在性能上优于经典的GNN、LLM以及近期的GNN-LLM融合方法，证明了交互式结构-语义协作在文本属性图学习中的有效性。
cs.AI / 98 / 2606.30294
Rehearsed Multi-Agent Live Product Demonstrations with Real-Time Voice Question Answering
经过排练的多智能体实时产品演示与实时语音问答
Abstract
Live product demonstrations are a recurring, high-cost activity in software organizations: a human presenter must select features, dispatch the corresponding interactions on a running application, narrate them coherently, and answer questions in real time. Existing automation addresses only fragments -- generalist browser agents target instruction-conditioned task completion, and demo-video tools produce fixed MP4 artifacts that cannot be questioned and silently break under interface drift. We propose Rhetor, a multi-agent system that takes a running web application and its source-code repository as input and produces a rehearsed live demonstration with segment-synchronized narration and real-time voice question answering. The architectural contributions are a cross-modal feature representation that merges UI exploration with source-code analysis into features tagged with discrete focus tiers, a grounded scripter constrained to UI elements observed during exploration and dispatched through multi-strategy semantic locators, a pre-presentation rehearsal loop with explicit convergence and graceful degradation to narration-only segments, and a runtime synchronization invariant that ties each browser action to the audio-end event of its narration segment. Across six pipeline sessions on four deployed applications -- including the public-domain whiteboard application Excalidraw -- the rehearser's internal locator-firing rate (sigma-bar) spans 0.31-1.00 over 147 scripted actions; on the substantial workload (53 actions, full tier differentiation), sigma-bar is approximately 0.92, and on the public-domain reference point the locator-repair step drives convergence to sigma-bar = 1.00 at iteration 2. We additionally define a benchmark protocol of ten metrics across six application categories that would establish, beyond the case study, whether each design choice contributes positively.
Chinese Translation
实时产品演示是软件组织中一种反复出现且成本高昂的活动：人类演示者必须选择功能，调度正在运行的应用程序中的相应交互，连贯地叙述这些交互，并实时回答问题。现有的自动化仅解决了部分问题——通用浏览器代理针对指令条件的任务完成，而演示视频工具生成固定的 MP4 文件，这些文件无法被提问，并且在界面漂移时会静默失效。我们提出了 Rhetor，一个多智能体系统，它以正在运行的网页应用程序及其源代码库为输入，生成经过排练的实时演示，具有分段同步的叙述和实时语音问答。架构贡献包括一种跨模态特征表示，将用户界面探索与源代码分析合并为带有离散关注层级的特征，一个受限于探索期间观察到的用户界面元素并通过多策略语义定位器调度的基础脚本生成器，一个具有明确收敛性和优雅降级到仅叙述段落的预演环节，以及一个运行时同步不变性，将每个浏览器操作与其叙述段落的音频结束事件绑定。在对四个已部署应用程序进行的六个管道会话中——包括公共领域的白板应用程序 Excalidraw——排练器的内部定位器触发率（sigma-bar）在 147 个脚本化操作中范围为 0.31-1.00；在大量工作负载（53 个操作，完整层级区分）中，sigma-bar 约为 0.92，而在公共领域参考点上，定位器修复步骤在第 2 次迭代时将收敛驱动至 sigma-bar = 1.00。我们还定义了一个涵盖六个应用类别的十项指标基准协议，以确定每个设计选择是否产生积极影响，超越案例研究的范畴。
cs.AI / 99 / 2606.30296
ManimAgent: Self-Evolving Multimodal Agents for Visual Education
ManimAgent：用于视觉教育的自我进化多模态智能体
Abstract
Multi-round reflection lets agents built on large language models recover from failures within a single task, but each task remains an isolated episode: lessons learned across many reflection rounds on one task are discarded before the next begins. We study this gap on a code-generation task: from a scientific paper section, the agent writes Python in the open-source Manim library to render a mathematical animation. We present ManimAgent, a self-evolving multimodal agent that carries reflection experience across tasks through a dual-channel Episodic Memory Bank grown entirely from its own task stream, with no weight updates and no human seeds. After each animation converges, a vision-language model scores the rendered keyframes; the resulting signals populate a positive channel M+ that stores success rationales as soft Reference Examples, and a negative channel M- that stores validated failure patterns as hard Known Pitfalls. On a fixed-probe evaluation against no-memory, matched-budget retrieval-augmented generation, and shuffled-memory baselines, blind human Pass@1 rises and reflection rounds fall as memory size grows. We will release the code, frozen memory snapshots, and the task stream.
Chinese Translation
多轮反思使基于大型语言模型的智能体能够在单一任务中从失败中恢复，但每个任务仍然是一个孤立的事件：在开始下一个任务之前，针对一个任务的多轮反思所获得的经验教训会被丢弃。我们在一个代码生成任务中研究了这一差距：从科学论文的一个部分，智能体使用开源的 Manim 库编写 Python 代码以渲染数学动画。我们提出了 ManimAgent，一个自我进化的多模态智能体，通过一个完全由其自身任务流生成的双通道情节记忆库跨任务携带反思经验，且无需权重更新和人类种子。在每个动画收敛后，一个视觉-语言模型对渲染的关键帧进行评分；生成的信号填充了一个正通道 M+，该通道存储成功理由作为软参考示例，以及一个负通道 M-，该通道存储经过验证的失败模式作为硬已知陷阱。在与无记忆、匹配预算的检索增强生成和洗牌记忆基线的固定探测评估中，盲人 Pass@1 随着记忆大小的增长而上升，而反思轮次则下降。我们将发布代码、冻结的记忆快照和任务流。
cs.AI / 100 / 2606.30335
BayesEvolve: Explicit Belief States for Autonomous Scientific Discovery
BayesEvolve：用于自主科学发现的显式信念状态
Abstract
Autonomous scientific discovery systems increasingly use large language models (LLMs) to propose new hypotheses, but many such systems condition primarily on experimental memory: archives of high-scoring candidates or heuristic summaries of recent trials. We argue that discovery agents should instead maintain explicit, uncertainty-aware beliefs about hypothesis quality. We introduce BayesEvolve, a belief-guided discovery framework that converts experimental evidence into a predictive belief state and uses this belief to guide future experimentation. As a controlled testbed for belief-guided discovery, we evaluate BayesEvolve on shifted BBOB-style black-box optimization tasks, leaving program and laboratory discovery domains to future work. BayesEvolve improves sample efficiency over memory- and archive-guided LLM baselines under a fixed evaluation budget. We further show that the belief state is predictive on held-out candidate pools, that controlled decision-rule ablations favor belief-guided selection with an annealed uncertainty bonus, and that BayesEvolve exhibits productive late-stage concentration rather than unfocused exploration.
Chinese Translation
自主科学发现系统越来越多地使用大型语言模型（LLMs）来提出新的假设，但许多此类系统主要依赖于实验记忆：高评分候选者的档案或近期试验的启发式总结。我们认为，发现代理应该维持对假设质量的显式、不确定性感知的信念。我们介绍了BayesEvolve，这是一种信念引导的发现框架，它将实验证据转化为预测信念状态，并利用这一信念指导未来的实验。作为信念引导发现的受控测试平台，我们在转移的BBOB风格黑箱优化任务上评估了BayesEvolve，将程序和实验室发现领域留待未来研究。BayesEvolve在固定评估预算下，相较于基于记忆和档案的LLM基线提高了样本效率。我们进一步展示了信念状态在保留的候选池上具有预测能力，受控的决策规则消融实验支持信念引导的选择，并伴随退火的不确定性奖励，此外，BayesEvolve表现出有效的后期集中而非无序探索。
cs.AI / 101 / 2606.30338
Sequential Fairness Auditing with Limited Output Access
有限输出访问下的顺序公平性审计
Abstract
External evaluations are becoming increasingly central to the governance of AI systems. In practice, however, independent auditors often have limited access to deployed models and must rely on query-based interactions. Most existing fairness evaluation methods assume static datasets and fixed-sample statistical tests, making them poorly suited to real-world auditing scenarios in which evidence must be collected sequentially under query constraints. In this work, we formulate fairness auditing as a tolerance-aware sequential hypothesis-testing problem under limited model output access. We develop a sequential generalized likelihood-ratio framework that allows auditors to accumulate evidence from a finite audit pool and stop once sufficient support for compliance or violation has been obtained. The framework is instantiated for decision-based Statistical Parity and Equal Opportunity audits, and extended to score- and logit-based proxy audits when richer observables are available. Our results show that both the fairness metric and the level of model access significantly affect audit efficiency, and that the benefits of richer output information are not uniform across auditing settings. In particular, richer outputs can substantially reduce the number of queries required for some fairness metrics and operating regimes, while offering limited gains in near-threshold cases. This work provides a practical statistical framework for sequential fairness auditing under realistic deployment constraints.
Chinese Translation
外部评估在人工智能系统治理中变得越来越重要。然而，在实践中，独立审计员通常对已部署模型的访问有限，必须依赖基于查询的交互。现有的大多数公平性评估方法假设数据集是静态的，并且使用固定样本的统计测试，这使得它们不适合在实际审计场景中进行，其中证据必须在查询限制下顺序收集。在本研究中，我们将公平性审计形式化为一个在有限模型输出访问下的容忍度感知顺序假设检验问题。我们开发了一个顺序广义似然比框架，允许审计员从有限的审计池中积累证据，并在获得足够的合规或违规支持后停止。该框架针对基于决策的统计平等（Statistical Parity）和机会平等（Equal Opportunity）审计进行了实例化，并在可用更丰富的可观察变量时扩展到基于得分和对数几率的代理审计。我们的结果表明，公平性指标和模型访问水平显著影响审计效率，并且更丰富的输出信息在不同审计环境中的收益并不均匀。特别是，对于某些公平性指标和操作模式，更丰富的输出可以显著减少所需的查询次数，而在接近阈值的情况下则提供有限的收益。本研究为在现实部署约束下的顺序公平性审计提供了一个实用的统计框架。
cs.AI / 102 / 2606.30372
Using Large Language Models as Low-Cost Statistical Estimators for Human-Response Data
将大型语言模型作为低成本统计估计器用于人类响应数据
Abstract
Quantitative research across the social and behavioral sciences depends on human subject experiments that are expensive, slow, and subject to sampling bias. Here we show that pretrained large language models induce risk-equivalent estimators of conditional expectations under squared loss, establishing restricted functional risk equivalence: under squared loss, the LLM induces an estimator whose risk matches the Bayes optimal risk for squared-loss prediction of conditional expectations for any inference that depends on the data only through the conditional mean. We formalize the LLM as a misspecified functional estimator $T(\hat{P}_n)$ trained on i.i.d.\ data, decompose the estimation error into representation bias $\epsilon_{\mathrm{rep}}$ and optimization error, and prove that under mild regularity conditions the LLM's expected error converges to the irreducible population variance plus the squared representation bias, with the representation bias bounded by the Pinsker inequality. The identifiability error $\delta$ propagates into the effective bias, inflating the asymptotic risk floor. We establish restricted functional risk equivalence via a bidirectional Le Cam deficiency analysis: the forward deficiency vanishes asymptotically while the reverse deficiency is exactly zero. We provide finite-sample concentration bounds and a calibration protocol with explicit decision rules. The result is a precise, provable statement: a well-calibrated LLM achieves the Bayes-optimal risk for conditional-mean-dependent inference, bounded by explicit scope conditions. In practical applications, this means that under satisfied conditions and well-calibrated models, large language models can be used in many prediction and decision-making tasks that originally relied on human experiments, approximating near-optimal statistical inference at lower cost.
Chinese Translation
社会和行为科学中的定量研究依赖于昂贵、缓慢且易受抽样偏差影响的人体实验。在此，我们展示了预训练的大型语言模型在平方损失下诱导条件期望的风险等效估计器，从而建立了限制性功能风险等价性：在平方损失下，LLM诱导的估计器其风险与任何仅通过条件均值依赖数据的条件期望的贝叶斯最优风险相匹配。我们将LLM形式化为在独立同分布（i.i.d.）数据上训练的错误指定功能估计器 $T( ilde{P}_n)$，将估计误差分解为表示偏差 $oldsymbol{ ext{ε}}_{ ext{rep}}$ 和优化误差，并证明在温和的正则条件下，LLM的期望误差收敛于不可约的总体方差加上平方表示偏差，表示偏差受到Pinsker不等式的限制。可识别性误差 $oldsymbol{ ext{δ}}$ 传播到有效偏差中，膨胀了渐近风险下限。我们通过双向Le Cam缺陷分析建立了限制性功能风险等价性：正向缺陷渐近消失，而反向缺陷恰好为零。我们提供了有限样本集中性界限和具有明确决策规则的校准协议。结果是一个精确且可证明的声明：一个良好校准的LLM在条件均值依赖推断中实现贝叶斯最优风险，受明确范围条件的限制。在实际应用中，这意味着在满足条件和良好校准的模型下，大型语言模型可以用于许多最初依赖于人体实验的预测和决策任务，以更低的成本近似最优的统计推断。
cs.AI / 103 / 2606.30383
Whose Side Is Your Agent On? Multi-Party Principal Loyalty in LLM Agents
你的代理人站在哪一边？大语言模型代理中的多方委托忠诚度
Abstract
A rapidly growing class of LLM agents is multi-party: the agent acts for a principal (who briefs it, sends follow-ups, and receives results) while also conversing in a separate channel with a counterparty whose interests may diverge (negotiating with a vendor, screening inbound requests, or mediating between employees). Here "help whoever you are talking to" is the wrong objective. The agent must stay loyal to the principal it represents without over-refusing the principal's own cooperative asks. We study this multi-party loyalty problem and contribute a measurement instrument, two mechanisms, and a structural lesson. PrincipalBench is a 75-item multi-turn benchmark with leak probes, dual judges, and an integrity-audit gate. Across 13 frontier subjects it exposes a sharp split (<=20% vs. 53.6-75.3% harm) invisible to single-turn safety evaluations: a selective cluster that declines adversarial probes while still following the principal's legitimate requests, and an over-refusing cluster that refuses broadly. (M1) A prompt-time loyalty scaffold (a fixed system prompt of seven prioritized rules, open-coded from 50+ failure trajectories) holds Claude-Sonnet to 19.4% harm and all nine selective subjects to <=20%. (M2) A per-token-KL distillation recipe transfers a prompted Qwen3-32B teacher into 8B Qwen3 and Llama-3.1 students, the strongest open-weight recipe we measure. (Lesson) Both mechanisms only move along a common leak/over-refusal trade-off rather than crossing it: improving one axis costs the other, and the jointly favorable outcome stays out of reach.
Chinese Translation
一种快速增长的多方大语言模型（LLM）代理类是多方的：代理为一个委托人行动（委托人向其简要说明、发送后续信息并接收结果），同时还在一个单独的渠道与一个利益可能相悖的对方进行对话（与供应商谈判、筛选入站请求或在员工之间进行调解）。在这里，“帮助你所交谈的任何人”并不是正确的目标。代理必须对其所代表的委托人保持忠诚，而不能过度拒绝委托人自身的合作请求。我们研究了这一多方忠诚度问题，并贡献了一个测量工具、两个机制和一个结构性教训。PrincipalBench是一个包含75个项目的多轮基准测试，配备泄漏探测器、双重评审和完整性审计门。在13个前沿主题中，它揭示了一个明显的分裂（<=20%与53.6-75.3%的损害），这是单轮安全评估所无法察觉的：一个选择性集群拒绝对抗性探测，同时仍然遵循委托人的合法请求，而一个过度拒绝的集群则广泛拒绝。（M1）一个提示时忠诚度框架（一个包含七条优先规则的固定系统提示，从50多个失败轨迹中开放编码）使Claude-Sonnet的损害保持在19.4%，所有九个选择性主题的损害保持在<=20%。（M2）一个每个标记的KL蒸馏配方将提示的Qwen3-32B教师转移到8B Qwen3和Llama-3.1学生中，这是我们测量的最强开源权重配方。（教训）这两个机制仅在一个共同的泄漏/过度拒绝权衡中移动，而不是跨越它：改善一个轴线会对另一个造成成本，而共同有利的结果仍然遥不可及。
cs.AI / 104 / 2606.30398
ENC-ODE: Event-level Neurodegenerative Modeling in Continuous Time with Neural ODEs
ENC-ODE：基于神经常微分方程的连续时间事件级神经退行性建模
Abstract
Accurately predicting the temporal evolution of clinical biomarkers is crucial for the early diagnosis and management of neurodegenerative diseases such as Alzheimer's disease. However, this relies on longitudinal data to capture biomarker changes over time, which is often sparse and irregular due to the high cost, labor-intensive nature, and patient burden. To address these challenges, we propose ENC-ODE, an Event-level Neurodegenerative modeling in Continuous time with neural Ordinary Differential Equations. ENC-ODE predicts future biomarker evolution by modeling clinical events through diagnosis-conditioned continuous dynamics. A target-conditioned attention mechanism weights and aggregates event-level predictions for the target time and modality without history compression. Extensive experiments on Alzheimer's Disease Neuroimaging Initiative (ADNI) dataset demonstrate that ENC-ODE outperforms representative sequence models while offering a scalable and neuroscientifically grounded solution for clinical support. The code is available at https://github.com/JardinDelSol/enc-ode.
Chinese Translation
准确预测临床生物标志物的时间演变对于早期诊断和管理阿尔茨海默病等神经退行性疾病至关重要。然而，这依赖于纵向数据以捕捉生物标志物随时间的变化，而这些数据通常由于高成本、劳动密集型特性和患者负担而稀疏且不规则。为了解决这些挑战，我们提出了ENC-ODE，一种基于神经常微分方程的连续时间事件级神经退行性建模方法。ENC-ODE通过诊断条件下的连续动态建模临床事件，预测未来的生物标志物演变。目标条件注意机制对目标时间和模态的事件级预测进行加权和聚合，而无需历史压缩。在阿尔茨海默病神经影像学倡议（ADNI）数据集上的广泛实验表明，ENC-ODE在性能上优于代表性的序列模型，同时为临床支持提供了一种可扩展且具有神经科学基础的解决方案。代码可在 https://github.com/JardinDelSol/enc-ode 获取。
cs.AI / 105 / 2606.30442
The FIL Hypothesis: Inductive Biases Help with Kernel Engineering
FIL假说：归纳偏差有助于核工程
Abstract
The Bitter Lesson, which posits that general-purpose methods that scale with computation and data ultimately outperform those with built-in human knowledge, has become a dominant paradigm in the era of Large Language Models. We revisit this principle by observing a new and critical scaling dimension: the duration of the Feedback Information Loop (FIL), the time required for a system to receive a verification signal after generating a prediction. Most historic successes in Artificial Intelligence (AI) have benefited from near instantaneous feedback (e.g., games or classification tasks), but we argue that future AI applications in science and the physical world will inherently involve FILs ranging from hours to weeks. This trend poses a fundamental scaling limit, as obtaining enough verification steps required by purely data-driven methods becomes practically impossible. Additionally, we propose a method that is orthogonal to purely data-driven approaches, based on human-inspired expert knowledge. The method relies on inductive biases and constraining the solution space. We provide an initial validation of the hypothesis and the method, by studying the real-world GPU programming task, a domain with non-trivial FIL, and demonstrate that incorporating inductive biases yields superior performance over data-driven approaches. The code is released under: https://github.com/ai-nikolai/robust_kernelbench
Chinese Translation
苦涩教训（The Bitter Lesson）认为，随着计算和数据的扩展，通用方法最终会超越那些内置人类知识的方法，这一观点在大型语言模型时代已成为主流范式。我们通过观察一个新的关键扩展维度——反馈信息循环（Feedback Information Loop, FIL）的持续时间，即系统在生成预测后接收验证信号所需的时间，重新审视这一原则。历史上大多数人工智能（Artificial Intelligence, AI）的成功都受益于几乎即时的反馈（例如，游戏或分类任务），但我们认为，未来在科学和物理世界中的AI应用将不可避免地涉及持续时间从小时到周的FIL。这一趋势构成了一个基本的扩展限制，因为纯数据驱动方法所需的足够验证步骤在实践中变得几乎不可能。此外，我们提出了一种与纯数据驱动方法正交的方法，该方法基于人类启发的专家知识。该方法依赖于归纳偏差和约束解空间。我们通过研究现实世界中的GPU编程任务（一个具有非平凡FIL的领域）对假说和方法进行了初步验证，并展示了引入归纳偏差的性能优于数据驱动方法。代码已发布于：https://github.com/ai-nikolai/robust_kernelbench
cs.AI / 106 / 2606.30531
Entity Binding Failures in Tool-Augmented Agents
工具增强代理中的实体绑定失败
Abstract
Tool-augmented language-model agents are often evaluated by whether they select the correct tool, produce valid API arguments, and complete the requested task. However, an agent may choose the right tool and still act on the wrong external entity. For example, a request to "email Alex about the launch" may lead the agent to contact the wrong Alex, attach the wrong launch document, reply in the wrong thread, or update the wrong customer account. We call these errors entity binding failures. This paper studies entity binding failures as a distinct reliability and safety problem in tool-augmented agents. We formalize the separation between tool correctness and entity correctness, introduce a taxonomy of wrong-entity failures in enterprise workflows, and evaluate entity-aware execution mechanisms including entity-resolution preconditions, confidence-gated binding, clarification under ambiguity, and provenance tracking. In a controlled diagnostic evaluation across 60 tasks, five model backends, and six tool-use methods, all methods achieved 0.0 percent wrong-tool error, yet action-oriented baselines still produced wrong-entity actions in 24.0-26.0 percent of runs. Entity-aware methods eliminated wrong-entity actions and risk-weighted wrong-entity exposure in this setting, but reduced direct task completion by deferring under ambiguity. These findings show that safe tool use requires not only selecting the correct tool, but also reliably binding natural-language references to the correct real-world entity before action.
Chinese Translation
工具增强的语言模型代理通常通过其是否选择正确的工具、生成有效的API参数以及完成请求的任务来进行评估。然而，一个代理可能选择了正确的工具，却仍然作用于错误的外部实体。例如，要求“给Alex发送关于发布的邮件”可能导致代理联系错误的Alex、附上错误的发布文档、在错误的讨论串中回复，或更新错误的客户账户。我们将这些错误称为实体绑定失败。本文将实体绑定失败作为工具增强代理中的一个独特的可靠性和安全性问题进行研究。我们正式区分了工具正确性与实体正确性，介绍了企业工作流程中错误实体失败的分类法，并评估了包括实体解析前提、信心门控绑定、模糊情况下的澄清和来源追踪在内的实体感知执行机制。在对60个任务、五个模型后端和六种工具使用方法进行的控制诊断评估中，所有方法的错误工具率均为0.0%，然而以行动为导向的基线方法在24.0-26.0%的运行中仍产生了错误实体的操作。实体感知方法在此设置中消除了错误实体操作和风险加权的错误实体暴露，但通过在模糊情况下的延迟处理减少了直接任务完成。这些发现表明，安全使用工具不仅需要选择正确的工具，还需要在行动之前可靠地将自然语言引用绑定到正确的现实世界实体。
cs.AI / 107 / 2606.30544
Latent Actions from Factorized Transition Effects under Agent Ambiguity
在代理模糊性下的因子化转移效应中的潜在动作
Abstract
Latent Action Models (LAMs) learn action-like proxies from observation transitions. However, in multi-object or distractor-rich scenes, these visual effects mix agent motion with distractors, camera dynamics, and background changes, making the underlying action source ambiguous without supervision. Structuring this mixture as reusable transition effects provides an intermediate representation from which action-like latents can be more robustly formed. We introduce Observed Transition Factorization (OTF), which decomposes each transition into a sparse set of observed transition primitives. Using these primitives as the transition interface, we propose OTF-LAM, which abstracts motion primitives into action-like latents within the standard inverse-forward dynamics framework, and OTF-LAM-Dino, a decoder-free variant that predicts future states in a frozen DINOv2 representation space. Empirically, OTF primitives transfer zeroshot across controlled carrier and morphology shifts, showing reusability. Furthermore, downstream policy learning results match or outperform baselines under complex transition ambiguity.
Chinese Translation
潜在动作模型（Latent Action Models, LAMs）从观察转移中学习类动作代理。然而，在多目标或干扰物丰富的场景中，这些视觉效应将代理运动与干扰物、相机动态和背景变化混合在一起，使得在没有监督的情况下，潜在的动作来源变得模糊。将这种混合结构化为可重用的转移效应提供了一种中间表示，从中可以更稳健地形成类动作的潜在变量。我们提出了观察转移因子化（Observed Transition Factorization, OTF），它将每个转移分解为一组稀疏的观察转移原语。利用这些原语作为转移接口，我们提出了OTF-LAM，它在标准的逆向-前向动态框架内将运动原语抽象为类动作的潜在变量，以及OTF-LAM-Dino，一个无解码器的变体，它在冻结的DINOv2表示空间中预测未来状态。实证结果表明，OTF原语在控制的载体和形态变化中实现了零样本迁移，显示出其可重用性。此外，下游策略学习结果在复杂转移模糊性下与基线相匹配或超越。
cs.AI / 108 / 2606.30555
Linguistic Firewall: Geometry as Defense in Multi-Agent Systems Routing
语言防火墙：几何作为多智能体系统路由中的防御
Abstract
The rapid integration of Large Language Models (LLMs) has driven the evolution of Multi-Agent Systems (MAS), where specialized agents collaborate to execute complex workflows. Effective orchestration in these environments requires robust routing mechanisms to efficiently allocate tasks to the most suitable agent. However, existing routers fundamentally rely on unverified proxies, ranging from textual self-descriptions to static surrogate representations, to gauge an agent's competence. This reliance on non-empirical data creates a critical gap between an agent's projected profile and its actual operational capabilities, introducing severe security vulnerabilities. Malicious agents can easily misrepresent their proficiencies or harbor covert backdoors that evade both standard external analysis and static representation-learning techniques. In this work, we introduce ANTAP (Automatic Non-Textual Agent Picker), an evaluation-driven routing architecture that discards indirect proxies in favor of active capability testing. By dynamically querying agents to ascertain their true competencies empirically, ANTAP distills performance into fixed behavioral operators within a shared semantic space. At inference time, routing is performed via a purely non-textual algebraic projection, establishing a "linguistic firewall" that renders metadata-based attacks inexpressible. In our experiments, ANTAP achieves near-zero ASR against description-based injection attacks, compared to 67.3\% and above for the description-based router baseline. Against adaptive embedding attacks, ANTAP achieves substantially lower ASR than the embedding-based baseline, with a 20\% reduction, while remaining resilient to description manipulation by design.
Chinese Translation
大型语言模型（LLMs）的快速整合推动了多智能体系统（MAS）的发展，在这些系统中，专门的智能体协同执行复杂的工作流程。在这些环境中，有效的编排需要强大的路由机制，以高效地将任务分配给最合适的智能体。然而，现有的路由器基本上依赖未经验证的代理，从文本自描述到静态替代表示，以评估智能体的能力。这种对非经验数据的依赖在智能体的预期特征与其实际操作能力之间产生了严重的差距，导致了严重的安全漏洞。恶意智能体可以轻易地歪曲其能力或隐藏隐秘的后门，从而逃避标准的外部分析和静态表示学习技术。在本研究中，我们提出了ANTAP（自动非文本智能体选择器），一种以评估为驱动的路由架构，摒弃间接代理，转而采用主动能力测试。通过动态查询智能体以经验性地确定其真实能力，ANTAP将性能提炼为共享语义空间中的固定行为操作符。在推理时，路由通过纯粹的非文本代数投影进行，从而建立了一个“语言防火墙”，使基于元数据的攻击无法表达。在我们的实验中，ANTAP在基于描述的注入攻击中实现了接近零的攻击成功率（ASR），而基于描述的路由器基线则为67.3 ext{%}及以上。针对自适应嵌入攻击，ANTAP的ASR显著低于基于嵌入的基线，减少了20 ext{%}，同时在设计上对描述操控保持韧性。
cs.AI / 109 / 2606.30561
The Human Creativity Benchmark
人类创造力基准
Abstract
Modern AI evaluation frameworks treat evaluator disagreement as noise to be resolved. In creative domains, professional disagreement reflects genuine differences in taste, not measurement error. We argue that evaluating creative AI requires preserving two distinct signals: convergence, where professionals align around shared best practices, and divergence, where individual taste legitimately varies. We present the Human Creativity Benchmark (HCB), a benchmark that operationalizes this separation by collecting pairwise preferences, scalar ratings on prompt adherence, usability, and visual appeal, and qualitative rationale from domain professionals. Across 15,000 professional judgments spanning five creative domains and three workflow phases (ideation, mockup, refinement), we find that convergence concentrates on verifiable dimensions like technical correctness and visual hierarchy, while divergence concentrates on taste-driven dimensions like aesthetic direction and conceptual risk. No model excels uniformly across all phases. Collapsing these signals into a single quality metric discards the most actionable information: where models must be correct versus where they should remain steerable.
Chinese Translation
现代人工智能评估框架将评估者之间的分歧视为需要解决的噪声。在创造性领域，专业人士之间的分歧反映了真正的品味差异，而非测量误差。我们认为，评估创造性人工智能需要保留两个不同的信号：收敛，即专业人士围绕共享最佳实践达成一致；以及发散，即个人品味的合法差异。我们提出了人类创造力基准（Human Creativity Benchmark, HCB），该基准通过收集成对偏好、对提示遵循性、可用性和视觉吸引力的标量评分，以及来自领域专业人士的定性理由，来实现这种分离。在涵盖五个创造性领域和三个工作流程阶段（构思、模型、精炼）的15,000个专业判断中，我们发现收敛主要集中在可验证的维度上，如技术正确性和视觉层次，而发散则集中在以品味驱动的维度上，如美学方向和概念风险。没有任何模型在所有阶段都表现出色。将这些信号合并为单一质量指标会丢弃最具可操作性的信息：模型必须在何处正确，以及在何处应保持可调节性。
cs.AI / 110 / 2606.30626
DOPD: Dual On-policy Distillation
DOPD：双重在线蒸馏
Abstract
On-policy distillation (OPD) offers superior capacity transfer by supervising student-sampled trajectories with dense token-level signals. To furnish high-quality supervision sources and thereby elevate the performance frontier of distillation, an intuitive direction is to infuse privileged information to either teacher or student itself. However, this additional input induces a potential failure mode we dub privilege illusion: a pattern that conflates the transferable capability gap that students are meant to close, and the information asymmetry gap that can only be mimicked but never replicated. This issue is further amplified by the inherent non-uniformity of token-level supervision, where only a small subset of tokens carries pivotal capability-bearing signals. To this end, we propose DOPD, an advantage-aware dual distillation paradigm that dynamically routes token-level supervision between privileged teacher and privileged student policies based on their advantage gap and relative probabilities. Each token receives supervision of different strength, objective, and strategy from either teacher or student itself, which transfers credible capability while simultaneously receiving auxiliary signals, to alleviate privilege illusion. Extensive experiments on both large language model (LLM) and vision-language model (VLM) settings demonstrate that DOPD consistently outperforms Vanilla OPD and other counterparts. Further results on stability, robustness, continual learning, and out-of-distribution tasks validate its superiority.
Chinese Translation
在线蒸馏（On-policy distillation, OPD）通过用密集的令牌级信号监督学生采样的轨迹，提供了优越的能力转移。为了提供高质量的监督源，从而提升蒸馏的性能边界，一个直观的方向是将特权信息注入教师或学生自身。然而，这一额外输入引发了一种潜在的失败模式，我们称之为特权幻觉（privilege illusion）：一种混淆了学生应当弥补的可转移能力差距与只能被模仿而无法复制的信息不对称差距的模式。由于令牌级监督的固有非均匀性，只有少量令牌携带关键的能力信号，这一问题进一步加剧。为此，我们提出了DOPD，一种优势感知的双重蒸馏范式，它根据特权教师和特权学生策略之间的优势差距和相对概率，动态地在它们之间路由令牌级监督。每个令牌从教师或学生自身接收不同强度、目标和策略的监督，这在转移可信能力的同时，接收辅助信号，以减轻特权幻觉。在大型语言模型（LLM）和视觉-语言模型（VLM）设置下的广泛实验表明，DOPD始终优于传统的OPD及其他对手。关于稳定性、鲁棒性、持续学习和分布外任务的进一步结果验证了其优越性。
cs.AI / 111 / 2606.30639
Self-Evolving World Models for LLM Agent Planning
自我演化的世界模型用于大型语言模型代理规划
Abstract
World models offer a principled way to equip long-horizon LLM agents with foresight: predictions of action consequences before execution. However, unreliable foresight can be ignored, misused, or even degrade downstream decision-making. In this paper, we introduce WorldEvolver, a self-evolving world model framework that revises its deployment-time context while keeping the downstream agent and all model parameters frozen. WorldEvolver integrates three modules: (i) Episodic Memory, which exploits real action transitions through retrieval-based simulation; (ii) Semantic Memory, which extracts persistent heuristic rules from prediction-observation mismatches; and (iii) Selective Foresight, which filters low-confidence predictions before integrating them into agent reasoning context. We evaluate WorldEvolver on ALFWorld and ScienceWorld, measuring world model prediction accuracy on Word2World and downstream agent success rate on AgentBoard. Extensive experiments show that WorldEvolver achieves the highest prediction accuracy across three backbones and leads other world model baselines on downstream agent success rate, demonstrating that test-time memory revision enhances both predictive fidelity and planning performance.
Chinese Translation
世界模型为长期规划的LLM代理提供了一种有原则的方法，使其具备预见性：在执行之前对行动后果的预测。然而，不可靠的预见性可能被忽视、误用，甚至导致下游决策的恶化。本文介绍了WorldEvolver，一个自我演化的世界模型框架，该框架在保持下游代理及所有模型参数不变的情况下，修正其部署时的上下文。WorldEvolver集成了三个模块：(i) 事件记忆（Episodic Memory），通过基于检索的模拟利用真实的行动转变；(ii) 语义记忆（Semantic Memory），从预测与观察的不匹配中提取持久的启发式规则；(iii) 选择性预见（Selective Foresight），在将低置信度预测整合到代理推理上下文之前进行过滤。我们在ALFWorld和ScienceWorld上评估了WorldEvolver，测量了Word2World上的世界模型预测准确性和AgentBoard上的下游代理成功率。大量实验表明，WorldEvolver在三种基础模型中实现了最高的预测准确性，并在下游代理成功率上超越了其他世界模型基线，证明了测试时记忆修正增强了预测的可靠性和规划性能。
cs.CL / 1 / 2606.28354
Generating in the Limit with Infinitely Many Hallucinations
在无限多幻觉下的极限生成
Abstract
The classic paradigm of language identification in the limit models learning as a game between an adversary, who reveals strings from an unknown target language, and a learner tasked with identifying that language. The recently introduced framework of language generation in the limit shifted the objective to better reflect modern language modeling, requiring the learner to produce valid, unseen strings from the target language. Related work highlighted a fundamental tension: a broad coverage of the target often comes at the cost of validity. We introduce a new notion of precision and recast this problem as the classic recall-precision trade-off. We analyze generation in the limit under varying constraints on enumeration, novelty, and validity, aimed at reflecting settings closer to those encountered by large language models. A key contribution is our analysis of learners that are not eventually valid: we allow infinitely many mistakes, provided their frequency tends to zero so that precision remains one. We show that this relaxation can strictly increase recall when the adversary permanently withholds a large portion of the target language. We also study a continuous relaxation of the novelty constraint that requires only a fixed fraction of outputs to be novel. Taken together, our results move toward a more realistic model of language generation where occasional errors and repetitions are unavoidable, but their rates are controlled.
Chinese Translation
经典的极限语言识别范式将学习视为一个对抗者与学习者之间的博弈，对抗者揭示来自未知目标语言的字符串，而学习者的任务是识别该语言。最近引入的极限语言生成框架将目标转向更好地反映现代语言建模，要求学习者生成来自目标语言的有效且未见过的字符串。相关研究突显出一个基本的张力：对目标的广泛覆盖往往以有效性为代价。我们引入了一种新的精确度概念，并将此问题重新表述为经典的召回-精确度权衡。我们分析了在不同枚举、新颖性和有效性约束下的极限生成，旨在反映更接近大型语言模型所遇到的设置。一个关键贡献是我们对非最终有效学习者的分析：我们允许无限多的错误，只要其频率趋向于零，从而保持精确度为一。我们展示了这种放松可以在对抗者永久性地隐瞒目标语言的大部分情况下严格提高召回率。我们还研究了一种新颖性约束的连续放松，仅要求固定比例的输出为新颖。综合来看，我们的结果朝着更现实的语言生成模型迈进，其中偶尔的错误和重复是不可避免的，但其发生率是可控的。
cs.CL / 2 / 2606.28457
Extracting Knowledge from an Arabic-English Machine-Readable Dictionary Using Information Extraction
利用信息提取从阿拉伯语-英语机器可读词典中提取知识
Abstract
Natural language processing (NLP) applications need large and rich amount of linguistic knowledge. Furthermore, electronic language sources such as dictionaries, encyclopedia, and corpora became available. So, automatic methods are emerged to extract lexical information from those sources to overcome the knowledge acquisition bottleneck. We presented a method to automatically extract lexical information from a machine-readable version of the Arabic-English Al-Mawrid dictionary. We used n-gram analysis and key-word-in-context (KWIC) analysis to discover lexical patterns that manifest morphologic, syntactic, or semantic information. Then, we used hand-crafted rule-based information extraction to extract that information. Furthermore, we used punctuation marks and some heuristics to extract a set of synonyms in a subentry. This study registered high precision for all types of information, high recall for synonyms, and low recall for the other information. The study also showed that the Al-Mawrid has significant amount of derivations (morphologic information) and synonyms, domain labels, and hyponym/hypernym relations (semantic information).
Chinese Translation
自然语言处理（NLP）应用需要大量丰富的语言知识。此外，电子语言资源如词典、百科全书和语料库已变得可用。因此，自动化方法应运而生，以从这些资源中提取词汇信息，以克服知识获取的瓶颈。我们提出了一种从阿拉伯语-英语《Al-Mawrid》词典的机器可读版本中自动提取词汇信息的方法。我们使用n-gram分析和上下文中的关键词（KWIC）分析来发现表现形态、句法或语义信息的词汇模式。然后，我们使用手工制作的基于规则的信息提取方法来提取这些信息。此外，我们使用标点符号和一些启发式方法来提取子条目中的一组同义词。本研究在所有类型的信息上注册了高精度，在同义词上注册了高召回率，而在其他信息上则注册了低召回率。研究还表明，《Al-Mawrid》包含大量的派生词（形态信息）、同义词、领域标签以及下义词/上义词关系（语义信息）。
cs.CL / 3 / 2606.28524
Developmental Trajectories of Situation Modeling and Mentalizing in Transformer Language Models
变压器语言模型中情境建模和心理化的发展轨迹
Abstract
Recent work suggests that Large Language Models (LLMs) are sensitive to the belief states of agents described by text, as measured by the false belief task (FBT), yet persistent concerns of construct validity remain. We adopt a **developmental perspective**, tracing the pattern of mental state reasoning behavior -- and likely **preconditions** for this behavior -- across multiple training stages in the Olmo2 and Pythia language model suites. We find that above-chance FBT performance depends both on model size and sufficient training volume, emerges relatively late in pretraining, and is most improved by post-training interventions (SFT, DPO) in the condition most diagnostic of mentalizing (False Belief, Implicit). However, FBT performance is fragile: consistent with past work, the use of non-factive verbs (e.g., thinks) increases false belief attributions even in the True Belief condition. To contextualize these findings, we track the emergence of **situation modeling**: the ability to report on basic factual properties of a described scene. Situation modeling accuracy generally precedes and exceeds FBT accuracy, yet situational representations also prove surprisingly incoherent in certain respects: when asked about the knowledge states of the Antagonist agent -- who always knows the item's true location -- Olmo2 13b is consistently influenced both by the Target agent's knowledge state and the presence of non-factive verbs. Together, these results suggest that larger, sufficiently trained models build partially coherent situation models in a developmentally appropriate sequence, yet display surprising fragility -- highlighting the value of developmental and stress-testing approaches for evaluating LLM capabilities.
Chinese Translation
最近的研究表明，大型语言模型（LLMs）对文本中描述的代理人的信念状态敏感，这一点通过虚假信念任务（FBT）进行测量，但对构建效度的持续关注仍然存在。我们采用**发展视角**，追踪Olmo2和Pythia语言模型系列在多个训练阶段中心理状态推理行为的模式及其可能的**前提条件**。我们发现，超出偶然水平的FBT表现既依赖于模型规模，也依赖于足够的训练量，并且在预训练的较晚阶段才会出现，且在最能诊断心理化的条件下（虚假信念，隐性）通过后训练干预（SFT，DPO）得到显著改善。然而，FBT表现是脆弱的：与过去的研究一致，使用非事实动词（例如，认为）即使在真实信念条件下也会增加虚假信念归因。为了将这些发现置于背景中，我们追踪**情境建模**的出现：报告描述场景基本事实属性的能力。情境建模的准确性通常先于并超过FBT的准确性，但在某些方面，情境表征也表现出惊人的不连贯性：当被询问关于对立代理人的知识状态时——该代理人始终知道物品的真实位置——Olmo2 13b始终受到目标代理人知识状态和非事实动词存在的影响。综合这些结果表明，较大且经过充分训练的模型以发展适当的顺序构建部分连贯的情境模型，但表现出惊人的脆弱性——突显了发展和压力测试方法在评估LLM能力中的价值。
cs.CL / 4 / 2606.28526
A French OSCE Dialogue Dataset and Controllable Virtual Patient System for Clinical Training
法语OSCE对话数据集及可控虚拟患者系统用于临床培训
Abstract
The clinical and communication skills of medical students are commonly assessed through Objective Structured Clinical Examinations (OSCEs), which consist of brief scenario-driven simulations of doctor-patient interactions. However, training is often limited by the low availability of human standardized patients, motivating the development of realistic virtual patients (VPs). To address this gap, we introduce a French OSCE dialogue dataset comprising 240 student-patient training interactions. We build upon it a controllable LLM-based pipeline to generate synthetic OSCE dialogues. The pipeline integrates modular components, such as retrieval-based grounding and a reflection loop, to ensure patient fidelity, coherence, and realism. Additionally, we propose a multi-level evaluation framework assessing patient simulation quality, student performance, and linguistic quality, using an LLM-as-a-Judge approach. Experiments suggest that controllability modules generally improve patient fidelity and student evaluation consistency. Finally, we implement an interactive prototype in which students can practice with a VP and receive automatic feedback.
Chinese Translation
医学学生的临床和沟通技能通常通过客观结构化临床考试（OSCE）进行评估，该考试由简短的场景驱动的医生-患者互动模拟组成。然而，培训往往受到人类标准化患者稀缺的限制，这促使我们开发逼真的虚拟患者（VP）。为了解决这一问题，我们介绍了一个法语OSCE对话数据集，包含240个学生-患者培训互动。我们在此基础上构建了一个可控的基于大型语言模型（LLM）的管道，以生成合成的OSCE对话。该管道集成了模块化组件，如基于检索的基础和反思循环，以确保患者的真实性、一致性和现实感。此外，我们提出了一个多层次评估框架，评估患者模拟质量、学生表现和语言质量，采用LLM作为评审者的方法。实验表明，可控模块通常提高了患者的真实性和学生评估的一致性。最后，我们实现了一个互动原型，学生可以与虚拟患者进行练习并获得自动反馈。
cs.CL / 5 / 2606.28538
Legal Domain Adaptation of Modern BERT Models
现代BERT模型的法律领域适应性研究
Abstract
We investigate domain adaptation of modern BERT models in the legal domain. We further pre-train ModernBERT on all US court opinions using the masked language modeling objective. Although ModernBERT has been trained on roughly 500x more data than original BERT, we still find that this model benefits from further pre-training and domain adaptation in the legal domain: we report significant improvements compared to vanilla ModernBERT on all datasets connected to US court opinions. We find gains similar to those reported in early work on domain adaptation of BERT-like models. However, from scratch pre-training does not match the performance of further pre-training an existing ModernBERT checkpoint in our experiments. The resulting models are capable of processing sequences up to 8,192 tokens, and can be used to compute meaningful embeddings of legal passages, or could quickly rerank hundreds of legal passages for a given search query. We release all model checkpoints publicly.
Chinese Translation
我们研究了现代BERT模型在法律领域的适应性。我们进一步在所有美国法院意见上使用掩码语言建模目标对ModernBERT进行了预训练。尽管ModernBERT的训练数据量大约是原始BERT的500倍，但我们仍然发现该模型在法律领域的进一步预训练和领域适应中受益：与原始ModernBERT相比，我们在所有与美国法院意见相关的数据集上报告了显著的改进。我们发现的提升与早期关于BERT类模型领域适应的研究报告的结果相似。然而，从头开始的预训练在我们的实验中未能达到进一步预训练现有ModernBERT检查点的性能。最终得到的模型能够处理长达8,192个标记的序列，并可用于计算法律段落的有意义嵌入，或者可以快速对给定搜索查询的数百个法律段落进行重新排序。我们将所有模型检查点公开发布。
cs.CL / 6 / 2606.28548
Turn-Averaged SAEs for Feature Discovery and Long-Context Attribution
基于轮次平均的稀疏自编码器用于特征发现和长上下文归因
Abstract
Sparse autoencoders (SAEs) have become a useful tool for extracting interpretable features in language models. However, standard SAE architectures operate on individual token activations, meaning that the number of active features scales linearly with context length, and studying long model transcripts becomes difficult. We introduce turn-averaged SAEs, which represent a single Human or Assistant turn with a fixed number of features by learning to reconstruct the average model activation across the turn. We find that turn-averaged features describe a single turn's high-level characteristics more completely than per-token features when judged by an LLM. We also demonstrate that turn-averaged SAEs greatly simplify common downstream uses of SAEs like attribution graphs. Broadly, turn-averaged SAEs make interpretability techniques practical at long context lengths.
Chinese Translation
稀疏自编码器（SAEs）已成为在语言模型中提取可解释特征的有用工具。然而，标准的SAE架构在单个标记激活上操作，这意味着活跃特征的数量与上下文长度呈线性关系，因此研究长模型转录变得困难。我们提出了基于轮次平均的SAEs，它通过学习重构整个轮次的平均模型激活，使用固定数量的特征来表示单个人类或助手的轮次。我们发现，当通过大型语言模型（LLM）进行评估时，轮次平均特征比逐标记特征更全面地描述了单个轮次的高层次特征。我们还展示了基于轮次平均的SAEs大大简化了SAEs在归因图等常见下游应用中的使用。总体而言，基于轮次平均的SAEs使得在长上下文长度下的可解释性技术变得可行。
cs.CL / 7 / 2606.28560
Depth-Staggered Fibonacci Spacing for Sparse Attention: Static Schedules Beat Learned Dilation and Extrapolate Where Dense Attention Fails
稀疏注意力的深度错位斐波那契间距：静态调度优于学习的扩张，并在密集注意力失效时进行外推
Abstract
We study sparse self-attention in which each query attends to a dense local window plus a set of Fibonacci-spaced offsets, with a per-layer scalar alpha that compresses or expands the spacing. Across 21 language models trained under one matched recipe (60M parameters, 512 hidden, 16 layers, 426M tokens), we compare four ways of setting alpha across depth: fixed, per-layer learned, a static linear stagger, and a coprime (anti-gridding) reassignment of that stagger, together with a reach-matched power-of-2 control. Three results stand out. First, a static per-layer stagger improves perplexity over both fixed and learned alpha, and the gain is base-agnostic: applying the same stagger to a power-of-2 base lifts it above fixed Fibonacci and to parity with learned Fibonacci attention. Second, learning per layer is inert: it does not beat the static schedule and costs roughly five times the inference latency. Third, and most consequential, all sparse variants extrapolate to four times their training length with little or no degradation, whereas a recipe-matched dense baseline collapses (perplexity rises by 201% at 4x length); we attribute this to fixed-offset attention only ever querying relative positions seen during training. We also report two honest negatives: at training length the best sparse model has about 26% higher perplexity than the dense baseline, and the staggering gain is uniform across context positions rather than concentrated at long range.
Chinese Translation
我们研究了稀疏自注意力机制，其中每个查询关注一个密集的局部窗口以及一组斐波那契间距的偏移量，并使用每层的标量 alpha 来压缩或扩展间距。在21个在相同配方下训练的语言模型（60M参数，512隐藏层，16层，426M标记）中，我们比较了四种在深度上设置 alpha 的方法：固定、每层学习、静态线性错位，以及该错位的互质（反网格）重新分配，同时还有一个匹配范围的2的幂控制。三个结果尤为突出。首先，静态的每层错位在困惑度上优于固定和学习的 alpha，且这种提升与基础无关：将相同的错位应用于2的幂基础，使其超越固定斐波那契并与学习的斐波那契注意力持平。其次，每层学习是无效的：它不如静态调度，并且推理延迟大约是其五倍。第三，也是最重要的，所有稀疏变体在训练长度的四倍时都能外推，几乎没有降级，而匹配配方的密集基线则崩溃（在4倍长度时困惑度上升201%）；我们将此归因于固定偏移注意力仅查询在训练期间看到的相对位置。我们还报告了两个诚实的负面结果：在训练长度下，最佳稀疏模型的困惑度比密集基线高出约26%，而错位增益在上下文位置上是均匀的，而不是集中在远程位置。
cs.CL / 8 / 2606.28562
SEAD: Competence-Aware On-Policy Distillation via Entropy-Guided Supervision
SEAD：基于熵引导监督的能力感知在线蒸馏
Abstract
On-policy distillation (OPD) has a property absent in offline distillation and RL: teacher supervision quality depends on student competence. Incoherent rollouts yield noisy gradients; already-mastered tokens yield redundant ones. This creates waste at three scales (tokens, training phases, and prompts) yet existing methods supervise uniformly. We introduce SEAD, which uses entropy as a unified probe of this competence-dependent degradation at three scales: (1) joint teacher-student entropy partitions tokens into zones receiving tailored divergences or zero gradient (approx. 50% skipped); (2) a cosine schedule anneals from forward to reverse KL as competence grows; (3) a competence-gated curriculum introduces prompts easy-to-hard. These components are symbiotically necessary: token selection requires coherent rollouts (curriculum), annealing requires monotonic improvement (also curriculum). On OLMo-3 (7B to 32B), SEAD achieves +4.8 avg accuracy over vanilla OPD across six math benchmarks, with ablations confirming super-additive interactions.
Chinese Translation
在线蒸馏（On-policy distillation, OPD）具有离线蒸馏和强化学习（RL）所不具备的特性：教师监督质量依赖于学生的能力。无序的回滚会产生噪声梯度；已经掌握的标记会导致冗余。这在三个层面（标记、训练阶段和提示）造成了浪费，而现有方法的监督方式是均匀的。我们提出了SEAD，它利用熵作为对这种能力依赖性退化的统一探测器，涵盖三个层面：（1）联合教师-学生熵将标记划分为接收量身定制的散度或零梯度的区域（约50%被跳过）；（2）余弦调度随着能力的增长从前向KL散度退火到反向KL散度；（3）能力门控课程引入从易到难的提示。这些组件是相辅相成的：标记选择需要一致的回滚（课程），而退火需要单调改进（同样是课程）。在OLMo-3（7B到32B）上，SEAD在六个数学基准测试中相较于普通OPD实现了+4.8的平均准确率，消融实验确认了超加性交互作用。
cs.CL / 9 / 2606.28574
Correct codes for the wrong reasons? validating LLMs as measurement instruments for theoretical constructs
错误原因下的正确编码？验证大型语言模型（LLMs）作为理论构念测量工具的有效性
Abstract
When a large language model (LLM) codes a construct in text as a human annotator would, that agreement makes the LLM a reliable coder. Yet reliability leaves construct validity untouched. The instrument may be theory-naive, reaching the code through a correlate that meets none of the demands the construct's theory makes, and no current method tells that apart from genuine measurement. We propose grain calibration as a method that closes the gap. It decomposes a construct into clause-level components, tests each against the text with extractive evidence, and combines the results through an explicit, theory-derived rule. Because the rule is stated rather than lodged in one opaque pass, its structure is evidence about the process rather than the output. It shows which components settled a code, and, when the code is wrong, whether a component was missed or an adjacent construct mistaken for it. Validation shifts from scoring an instrument's outputs against an annotator to showing that the instrument runs on the construct its theory specifies.
Chinese Translation
当大型语言模型（LLM）像人类标注者一样将一个构念编码为文本时，这种一致性使得LLM成为一个可靠的编码工具。然而，可靠性并未触及构念的效度。该工具可能是理论上天真的，通过一个与构念理论要求无关的相关因素得出编码，而目前没有任何方法能够区分这种情况与真正的测量。我们提出了粒度校准作为一种弥补这一差距的方法。它将构念分解为子句级别的组件，针对文本中的提取证据测试每个组件，并通过一个明确的、源自理论的规则组合结果。由于该规则是以声明的形式呈现，而不是以一种不透明的方式嵌入，因此其结构提供了关于过程的证据，而非仅仅是输出。它显示了哪些组件决定了编码，并且当编码错误时，是否遗漏了某个组件或将相邻的构念误认为该构念。验证的重点从对工具输出与标注者的评分转向展示该工具是否运行于其理论所指定的构念上。
cs.CL / 10 / 2606.28667
Phonological Perception of Sign Language Models
手语模型的音位感知
Abstract
Sign languages are compositional systems where meaning arises by combining sublexical phonological parameters, such as handshape, location, and movement. While deep learning models for Sign Language Recognition (SLR) have achieved increased performance on translation benchmarks, it remains unclear whether these models distinguish abstract phonological features or merely rely on low-level statistical correlations. This work evaluates the phonological perception of SLR models trained on American Sign Language (ASL) by probing phonological sensitivity using minimal pairs and evaluating representational alignment with human behavioral data. Our results reveal that SLR models exhibit emergent phonological sensitivity, but with clear architectural trade-offs: pose-based models are sensitive to handshape contrasts, while pixel-based models better capture location changes. Furthermore, pose-based models learn latent representations that correlate with human perceptual similarity judgments (r~0.49). These findings suggest that while SLR models exhibit emergent phonology, current training paradigms are insufficient to scale them beyond their architectural inductive biases.
Chinese Translation
手语是一种组合系统，其意义通过结合亚词汇音位参数（如手型、位置和动作）而产生。尽管用于手语识别（Sign Language Recognition, SLR）的深度学习模型在翻译基准测试中取得了更高的性能，但尚不清楚这些模型是否能够区分抽象的音位特征，还是仅仅依赖于低级统计相关性。本研究通过使用最小对比词探测音位敏感性，并评估与人类行为数据的表征一致性，来评估在美国手语（American Sign Language, ASL）上训练的SLR模型的音位感知。我们的结果显示，SLR模型表现出新兴的音位敏感性，但存在明显的架构权衡：基于姿态的模型对手型对比敏感，而基于像素的模型更好地捕捉位置变化。此外，基于姿态的模型学习到的潜在表征与人类感知相似性判断存在相关性（r~0.49）。这些发现表明，尽管SLR模型表现出新兴的音位特征，但当前的训练范式不足以使其超越其架构的归纳偏差。
cs.CL / 11 / 2606.28708
AnTenA: Actionable and Explainable Tensor Analysis System with Large Language Models
AnTenA：可操作且可解释的大语言模型张量分析系统
Abstract
Accurately explaining hidden patterns in multi-aspect data has typically been done by leveraging labels and/or accompanying auxiliary metadata. However, labels and auxiliary data may be inaccurate (e.g. nonstandard, inconsistent), insufficient (e.g. static tabular metadata for time-dependent recordings), or unavailable. %
We propose \fullmethod (\method), which leverages the knowledge of large language models (LLMs) to explain the hidden patterns in human narratives. \method uses task-agnostic and task-specific prompts to explain extracted co-clustered latent patterns from tensor decomposition. To evaluate these explanations, we test the LLMs on forward and backward inference tasks. % Our demo system is available at https://github.com/dawonahn/ECML_PKDD_AnTenA.
Chinese Translation
准确解释多方面数据中的隐藏模式通常依赖于标签和/或附加辅助元数据。然而，标签和辅助数据可能不准确（例如，非标准、不一致）、不足（例如，对于时间依赖记录的静态表格元数据），或不可用。我们提出了 ullmethod（ extit{method}），该方法利用大型语言模型（LLMs）的知识来解释人类叙事中的隐藏模式。 extit{method}使用与任务无关和特定任务的提示来解释从张量分解中提取的共同聚类潜在模式。为了评估这些解释，我们在前向和后向推理任务上测试了LLMs。我们的演示系统可在 https://github.com/dawonahn/ECML_PKDD_AnTenA 获取。
cs.CL / 12 / 2606.28715
SEATauBench: Adapting Tool-Agent-User Evaluation Into Low-Resource Southeast Asian Languages
SEATauBench：将工具-代理-用户评估适应于低资源东南亚语言
Abstract
While AI development and evaluation for Southeast Asia (SEA) has grown rapidly, agent capabilities in regional languages are still poorly understood despite its importance to sovereign AI. To fill this gap, we introduce SEATauBench, the first agent-focused evaluation framework for SEA sovereign AI. SeaTau adapts TauBench to five languages -- Mandarin, Vietnamese, Thai, Indonesian, and Filipino -- and evaluates agents across progressively localized settings that vary the language of user-agent interaction, tool specifications, and task domains. Across three recent models, we find that English agent capabilities transfer reasonably well when only the conversation language changes, but quality and robustness degrade sharply as more task contexts are localized, with the largest losses in full domain adaptation. We also the limits of English-only agent assessment for measuring agent capabilities in SEA languages. More broadly, SeaTau provides a diagnostic benchmark and reusable adaptation pipeline for building reliable multilingual agents for linguistically diverse regions. Data and code can be accessed at github.com/SEACrowd/SEATauBench.
Chinese Translation
尽管东南亚（SEA）的人工智能开发和评估迅速增长，但区域语言中的代理能力仍然不够清晰，尽管这对主权人工智能至关重要。为填补这一空白，我们推出了SEATauBench，这是首个针对SEA主权人工智能的代理聚焦评估框架。SeaTau将TauBench适配为五种语言——普通话、越南语、泰语、印尼语和菲律宾语——并在逐步本地化的环境中评估代理，这些环境变化了用户与代理的互动语言、工具规格和任务领域。在三个最新模型中，我们发现，当仅改变对话语言时，英语代理能力的转移相对良好，但随着更多任务背景的本地化，质量和稳健性急剧下降，完全领域适应的损失最大。我们还探讨了仅使用英语进行代理评估在测量SEA语言中的代理能力方面的局限性。更广泛地说，SeaTau提供了一个诊断基准和可重用的适配管道，用于构建可靠的多语言代理，以应对语言多样化的地区。数据和代码可在github.com/SEACrowd/SEATauBench获取。
cs.CL / 13 / 2606.28725
DriftGuard: Safety-Aware Multi-Monitor Detection and Selective Adaptation for Evolving Toxicity Moderation
DriftGuard：安全感知的多监测检测与选择性适应用于演变中的毒性审查
Abstract
Automated toxicity moderation systems operate in dynamic online environments where harmful behavior evolves through coded language, shifting targets, and strategic adaptation to enforcement. Existing drift detection methods often focus on global distributional change, but such signals may miss safety-relevant shifts that emerge in localized harm subspaces or high-risk model-error regions. This paper introduces DriftGuard, a safety-aware adaptive moderation framework that combines multi-monitor drift detection with selective model updating. The framework tracks global text drift, identity-harm drift, model uncertainty, toxic-risk drift, and false-negative-risk drift. When safety-relevant change is detected, the model is updated using a hard-mix adaptation set that prioritizes likely false negatives, identity-related high-risk examples, false-positive-risk examples, and uncertain boundary cases. Experiments on Civil Comments temporal shift and Jigsaw-to-DynaHate cross-dataset shift show that safety-aware monitors detect risks missed by global drift alone. Hard-mix adaptation improves toxic recall and accuracy over no-update and random-balanced baselines, raising toxic recall to 0.8777 on Civil Comments and from 0.7107 to 0.8523 on DynaHate. Bootstrap analysis further shows stable DynaHate safety gains, with toxic recall increasing by 0.1418 and false-negative prevalence decreasing by 0.0781. Overall, DriftGuard links safety-aware drift detection to targeted, lightweight model updating for more robust adaptive toxicity moderation.
Chinese Translation
自动化毒性审查系统在动态在线环境中运行，其中有害行为通过编码语言、变化目标和对执法的战略适应而不断演变。现有的漂移检测方法通常侧重于全球分布变化，但此类信号可能会遗漏在局部危害子空间或高风险模型误差区域中出现的与安全相关的变化。本文介绍了DriftGuard，一种安全感知的自适应审查框架，它结合了多监测漂移检测与选择性模型更新。该框架跟踪全球文本漂移、身份危害漂移、模型不确定性、毒性风险漂移和假阴性风险漂移。当检测到与安全相关的变化时，模型将使用一个硬混合适应集进行更新，该适应集优先考虑可能的假阴性、与身份相关的高风险示例、假阳性风险示例和不确定边界案例。在Civil Comments时间漂移和Jigsaw到DynaHate跨数据集漂移的实验中，安全感知监测器能够检测到全球漂移单独无法发现的风险。硬混合适应在没有更新和随机平衡基线的情况下提高了毒性召回率和准确性，使Civil Comments的毒性召回率提升至0.8777，DynaHate的召回率从0.7107提升至0.8523。Bootstrap分析进一步显示DynaHate的安全收益稳定，毒性召回率增加了0.1418，假阴性发生率降低了0.0781。总体而言，DriftGuard将安全感知的漂移检测与针对性、轻量级的模型更新联系起来，以实现更稳健的自适应毒性审查。
cs.CL / 14 / 2606.28737
5ting at SemEval-2026 Task 8: Strong End-to-End Multi-Turn RAG via LLM-Based Reranking and Faithfulness Control
5ting在SemEval-2026任务8中的表现：基于LLM的重排序和可信度控制的强大端到端多轮检索增强生成
Abstract
We introduce 5ting, our system for the SemEval2026 Task 8 (MTRAGEval), which evaluates multi-turn Retrieval Augmented Generation (RAG) systems. Multi turn RAG involves context drift, under specification, and hallucination risk. Our system combines BGE-M3 dense retrieval with FAISS indexing, dual-query merged retrieval, and LLM based reranking, followed by role separated generation constrained to retrieved evidence. The retriever achieved nDCG@5 = 0.4719 in Task A, while the end to end system ranked in Task C with a harmonic score of 0.5597 and RL_F = 0.7692.
Chinese Translation
我们介绍了5ting，这是我们为SemEval2026任务8（MTRAGEval）开发的系统，该任务评估多轮检索增强生成（RAG）系统。多轮RAG涉及上下文漂移、规格不足和幻觉风险。我们的系统结合了BGE-M3密集检索与FAISS索引、双查询合并检索，以及基于LLM的重排序，随后进行角色分离的生成，限制于检索到的证据。在任务A中，检索器的nDCG@5达到了0.4719，而端到端系统在任务C中的和谐得分为0.5597，RL_F为0.7692。
cs.CL / 15 / 2606.28772
Majority Vote Silences Minority Values: Annotator Disagreement at the Hate/Offensive Boundary in HateXplain
多数投票压制少数价值观：HateXplain中仇恨/冒犯边界的标注者分歧
Abstract
Hate speech annotation pipelines routinely collapse annotator disagreement into majority vote labels before training. We show that this aggregation is not neutral: 42.6% of all annotator disagreement in HateXplain concentrates specifically at the hate/offensive boundary, a pattern consistent with annotators applying different thresholds for where hate begins (chi-squared = 135.199, df = 2, p < 0.0001). Both a hard-label BERT model (Model A) and a soft-label model (Model B) drop 22 percentage points in accuracy from agreed posts (~80%) to disagreement posts (~58%), confirmed at p < 0.0001. A per-annotator multi-head model (Model C) widens this gap further to 28 points while collapsing offensive disagreement accuracy to 0.245. Critically, Model A expresses significantly higher confidence on boundary case errors than Model C (0.710 vs. 0.495, p < 0.0001), meaning standard evaluation metrics will not detect the failure. Three downstream interventions of increasing sophistication all fail to recover boundary accuracy. We argue the problem is structural. Majority vote presents a contested judgment as ground truth, and models inherit that false certainty. The intervention must be upstream in annotation design.
Chinese Translation
仇恨言论标注流程通常在训练之前将标注者的分歧汇总为多数投票标签。我们表明这种汇总并非中立：在HateXplain中，42.6%的标注者分歧特定集中在仇恨/冒犯边界，这一模式与标注者对仇恨开始的不同阈值应用一致（卡方 = 135.199，自由度 = 2，p < 0.0001）。无论是硬标签BERT模型（模型A）还是软标签模型（模型B），在一致标注的帖子（约80%）与分歧标注的帖子（约58%）之间，准确率均下降了22个百分点，且在p < 0.0001的水平上得到了确认。每个标注者的多头模型（模型C）进一步将这一差距扩大至28个百分点，同时将冒犯性分歧的准确率压缩至0.245。关键是，模型A在边界案例错误上的置信度显著高于模型C（0.710对0.495，p < 0.0001），这意味着标准评估指标无法检测到这一失败。三种日益复杂的下游干预措施均未能恢复边界准确性。我们认为问题是结构性的。多数投票将有争议的判断呈现为真实情况，而模型继承了这种虚假的确定性。干预必须在标注设计的上游进行。
cs.CL / 16 / 2606.28796
Structure-Preserving Document Translation via Multi-Stage LLM Pipeline: A Case Study in Marathi
通过多阶段大语言模型管道实现结构保留的文档翻译：以马拉地语为例
Abstract
Government documents in India are predominantly issued in regional languages such as Marathi, creating substantial accessibility barriers for non-native readers, interstate administrative bodies, and policy analysts. Although recent advances in neural machine translation have improved sentence-level translation quality, existing systems largely neglect document structure, formatting integrity, and domain-specific terminology, thereby limiting their applicability to official documentation. This paper presents a structure-preserving Marathi-to-English government document translation framework capable of performing end-to-end document transformation while maintaining layout fidelity. The proposed system integrates layout-aware optical character recognition, coordinate-based text extraction, large language model based translation, and structured document reconstruction through HTML representations. By enforcing spatial alignment constraints and preserving hierarchical document elements, the framework ensures structural consistency between the source and translated documents. Experimental evaluation on real-world Marathi government PDFs demonstrates improved structural preservation, translation coherence, and terminological consistency compared to conventional text-only translation pipelines. The proposed framework contributes toward scalable multilingual accessibility solutions for e-governance and administrative document processing.
Chinese Translation
印度的政府文件主要以马拉地语等地方语言发布，这为非母语读者、州际行政机构和政策分析师带来了显著的可及性障碍。尽管近年来神经机器翻译的进展提高了句子级翻译质量，但现有系统在很大程度上忽视了文档结构、格式完整性和领域特定术语，从而限制了其在官方文件中的适用性。本文提出了一种结构保留的马拉地语到英语的政府文件翻译框架，能够在保持布局保真度的同时进行端到端的文档转换。该系统集成了布局感知的光学字符识别、基于坐标的文本提取、基于大语言模型的翻译以及通过HTML表示的结构化文档重建。通过强制空间对齐约束和保留层次文档元素，该框架确保源文档与翻译文档之间的结构一致性。在真实的马拉地语政府PDF文件上的实验评估表明，与传统的仅文本翻译管道相比，所提出的框架在结构保留、翻译连贯性和术语一致性方面均有所改善。该框架为电子治理和行政文档处理提供了可扩展的多语言可及性解决方案。
cs.CL / 17 / 2606.28823
Labeling Training Data for Entity Matching Using Large Language Models
使用大型语言模型对实体匹配进行标注训练数据
Abstract
Recent large language models (LLMs) achieve strong performance on entity matching without requiring task-specific training data. However, applying these models to large sets of candidate pairs remains slow and costly. In contrast, entity matchers using traditional machine learning methods or small language models (SLMs), such as RoBERTa, offer much faster inference but require task-specific training data. This paper investigates whether the need to provide task-specific training data can be avoided by using knowledge-distillation workflows, in which an LLM serves as a teacher model to label training pairs that are subsequently used to train a smaller student model. We investigate knowledge distillation for entity matching along the following dimensions: pair-selection strategy, teacher model, label post-processing method, and student model. We evaluate the workflows using the Abt-Buy, Walmart-Amazon, WDC Products, DBLP-ACM, and DBLP-Scholar benchmarks, and compare the performance of student models trained with machine-labeled data to the performance of the same models trained using the benchmark training sets. Our experiments show that student models trained using the machine-labeled sets perform approximately on par with models trained on the benchmark training sets, with the remaining differences in both directions staying below two F1 points. Using GPT-5.2 to label the training sets for all five benchmarks costs US\$28.31 to US\$40.88, whereas manually labeling the same training sets is estimated to require 470 hours of work. At inference time, Ditto is 41.5 to 534 times faster than directly using an LLM to perform the matching tasks. These results indicate that current LLMs, when combined with a suitable pair-selection method, can substantially reduce or even eliminate the manual effort required to label use case-specific training data for entity matching.
Chinese Translation
最近的大型语言模型（LLMs）在实体匹配方面表现出色，无需特定任务的训练数据。然而，将这些模型应用于大量候选对仍然缓慢且成本高昂。相比之下，使用传统机器学习方法或小型语言模型（SLMs），如RoBERTa的实体匹配器提供了更快的推理速度，但需要特定任务的训练数据。本文探讨了是否可以通过使用知识蒸馏工作流来避免提供特定任务的训练数据，在这种工作流中，LLM作为教师模型标注训练对，随后用于训练较小的学生模型。我们从以下几个维度研究实体匹配的知识蒸馏：对选择策略、教师模型、标签后处理方法和学生模型。我们使用Abt-Buy、Walmart-Amazon、WDC Products、DBLP-ACM和DBLP-Scholar基准评估这些工作流，并比较使用机器标注数据训练的学生模型与使用基准训练集训练的相同模型的性能。我们的实验表明，使用机器标注集训练的学生模型的性能与使用基准训练集训练的模型大致相当，两个方向的剩余差异均低于两个F1点。使用GPT-5.2为所有五个基准标注训练集的成本为28.31美元至40.88美元，而手动标注相同的训练集预计需要470小时的工作。在推理时，Ditto的速度比直接使用LLM执行匹配任务快41.5到534倍。这些结果表明，当前的LLM在结合合适的对选择方法时，可以显著减少甚至消除标注特定用例训练数据所需的人工努力。
cs.CL / 18 / 2606.28843
The Heterogeneous Safety Impacts of Benign Multilingual Fine-Tuning
良性多语言微调的异质安全影响
Abstract
Fine-tuning a large language model is a ubiquitous method for enhancing its capability on a specific downstream task. However, prior work has shown that this increase in capability comes with a cost: it can increase a model's tendency to respond to unsafe adversarial prompts, even when fine-tuning with non-adversarial data. We present the first comprehensive empirical study of this phenomenon in multilingual settings by fine-tuning Llama-3.2, Qwen3, and Gemma-3 models using benign data translated across nine languages. We find that safety outcomes are highly sensitive to both the choice of fine-tuning language and the evaluation language, with adversarial compliance rates increasing four-fold in some settings. Multilingual safety drift is decoupled from general capability metrics, and occurs heterogeneously across languages and models. Fine-tuning in non-English languages often induces smaller internal representational drifts than English, but these shifts lead models to default to either exaggerated compliance or refusal. As such, assessing fine-tuning impacts solely in English provides inadequate assurance for deployment. To facilitate further research into these cross-lingual safety blind spots, we release the Multilingual-Benign-Tune dataset and the SORRY-Bench-Multilingual evaluation suite.
Chinese Translation
微调大型语言模型是一种普遍的方法，用于增强其在特定下游任务上的能力。然而，先前的研究表明，这种能力的提升是有代价的：即使在使用非对抗性数据进行微调时，它也可能增加模型对不安全对抗性提示的响应倾向。我们通过使用跨九种语言翻译的良性数据微调 Llama-3.2、Qwen3 和 Gemma-3 模型，首次全面实证研究了这一现象在多语言环境中的表现。我们发现，安全结果对微调语言和评估语言的选择高度敏感，在某些情况下，对抗性合规率增加了四倍。多语言安全漂移与一般能力指标脱钩，并在不同语言和模型之间异质性地发生。在非英语语言中进行微调通常会导致比英语更小的内部表征漂移，但这些变化使模型倾向于默认选择夸大的合规或拒绝。因此，仅在英语中评估微调影响无法为部署提供充分的保障。为了促进对这些跨语言安全盲点的进一步研究，我们发布了 Multilingual-Benign-Tune 数据集和 SORRY-Bench-Multilingual 评估套件。
cs.CL / 19 / 2606.28867
Open but Incompatible: A License Compatibility Analysis of Corpora for Low-Resource African Languages
开放但不兼容：低资源非洲语言语料库的许可证兼容性分析
Abstract
Creative Commons licenses dominate African NLP corpus releases, but their compatibility rules are rarely applied. CC-BY-SA and CC-BY-NC cannot be combined in a single published dataset; a NoDerivs clause silently prohibits tokenisation and annotation. This paper audits the license provenance of over twenty corpus families used in African NLP, constructs a six-tier compatibility matrix, and applies it to three case-study languages: Kituba/Munukutuba, Zarma, and Moore. Four failure modes are documented with primary-source evidence: outright prohibition (JW300, removed from OPUS after a legal audit confirmed Terms of Service violation); composite license misrepresentation (WAXAL, whose CC-BY 4.0 claim is contradicted by its own HuggingFace dataset card); a NoDerivs clause hidden behind a CC-BY label (Tanzil); and data persistence failure (the Congolese Radio Corpus, where 402 of 405 source URLs are now dead). A pre-annotation due diligence checklist and a survey of legally clean enrichment opportunities close the paper.
Chinese Translation
创意共享许可证在非洲自然语言处理语料库发布中占主导地位，但其兼容性规则鲜有应用。CC-BY-SA和CC-BY-NC不能在单一发布的数据集中结合使用；NoDerivs条款默默禁止了标记化和注释。本文审计了用于非洲自然语言处理的二十多个语料库家族的许可证来源，构建了一个六级兼容性矩阵，并将其应用于三个案例研究语言：Kituba/Munukutuba、Zarma和Moore。记录了四种失败模式，并提供了原始证据：明确禁止（JW300，在法律审计确认违反服务条款后从OPUS中移除）；复合许可证误表述（WAXAL，其CC-BY 4.0声明与其自身的HuggingFace数据集卡片相矛盾）；隐藏在CC-BY标签后面的NoDerivs条款（Tanzil）；以及数据持久性失败（刚果广播语料库，其中405个源URL中有402个现已失效）。一份预注释尽职调查清单和一项合法清洁增值机会的调查作为本文的结尾。
cs.CL / 20 / 2606.28876
Memory-Managed Long-Context Attention: A Preliminary Study of Editable Request-Local Memory
内存管理的长上下文注意力：可编辑请求本地内存的初步研究
Abstract
Long-context language models often conflate two different goals: compressing history into an efficient state, and maintaining reliable long-term memory. Linear, recurrent, and sparse attention reduce the cost of processing long sequences, but they do not by themselves specify when a fact should be written, overwritten, protected from distractors, or discarded. We study memory-managed long-context attention, a research route that separates a fast recurrent or sparse backbone from explicit editable request-local memory slots and query-time sparse fallback. Across structured synthetic tasks, token/chunk/sequence bridges, generated natural language, and local frozen-model diagnostics, pure fixed-state or pure sparse methods fail some overwrite, version, anti-pollution, or no-write-signal cases, while a hybrid covers both routes. A small 2,097,152-token mechanism stress test reaches 50/50 pooled accuracy with 2-132 active chunks. A 2.74M-parameter minimal causal event-token model reaches 595/600 with lite write supervision, supporting proof of trainability rather than scale. A six-family frozen-hidden-state bridge reaches 1079/1080 controlled pointer accuracy, but it uses generator-provided integer key IDs and separately encoded canonical key strings; it is an oracle-metadata probe, not open-text entity resolution. Local non-leaderboard RULER 4K diagnostics remain close to full context, whereas a 33-record LongBench v1 16K subset shows that naive lexical selection is not general. The evidence separates three claims: controlled slot lifecycle is feasible, sparse fallback is needed when writes lack future-query signals, and learned open-domain selection remains the main architectural bottleneck. We do not claim a final generative architecture, global slot-trajectory convergence, or systems superiority.
Chinese Translation
长上下文语言模型通常将两个不同的目标混为一谈：将历史压缩为高效状态，以及维持可靠的长期记忆。线性、递归和稀疏注意力方法降低了处理长序列的成本，但它们本身并未明确规定何时应写入、覆盖、保护免受干扰或丢弃某个事实。我们研究了内存管理的长上下文注意力，这是一条将快速递归或稀疏主干与显式可编辑的请求本地内存槽和查询时稀疏回退分开的研究路线。在结构化的合成任务、标记/块/序列桥接、生成的自然语言和本地冻结模型诊断中，纯固定状态或纯稀疏方法在某些覆盖、版本、抗污染或无写信号的情况下表现不佳，而混合方法则涵盖了这两条路线。一个2097152标记的小型机制压力测试在132个活跃块的情况下达到了50/50的汇总准确率。一个274万参数的最小因果事件-标记模型在轻量写监督下达到了595/600，支持可训练性而非规模的证明。一个六家庭冻结隐藏状态桥接达到了1079/1080的受控指针准确率，但它使用生成器提供的整数键ID和单独编码的规范键字符串；这是一种oracle元数据探测，而不是开放文本实体解析。局部非排行榜的RULER 4K诊断接近完整上下文，而33条记录的LongBench v1 16K子集显示，简单的词汇选择并不通用。证据分离出三个主张：受控槽生命周期是可行的，当写入缺乏未来查询信号时需要稀疏回退，学习的开放领域选择仍然是主要的架构瓶颈。我们并不声称有最终的生成架构、全局槽轨迹收敛或系统优越性。
cs.CL / 21 / 2606.28898
PASTA: A Paraphrasing And Self-Training Approach for Knowledge Updating in LLMs
PASTA：一种用于大型语言模型知识更新的释义与自我训练方法
Abstract
Knowledge updating in pre-trained Large Language Models (LLMs) remains an important challenge. While continual training provides a potential avenue for knowledge updating, it continues to present substantial technical difficulties. Furthermore, LLMs often struggle with accurately answering questions about specific factual information, such as news articles - a capability limitation widely recognized in the research community. This paper proposes PASTA, a simple yet powerful framework for integrating detailed factual information from news articles as new knowledge into LLMs, with the primary goal of building specialized models that accurately answer questions about this knowledge. Our framework combines data augmentation, question-answering generation, and a novel self-learning DPO process that simultaneously enables knowledge overwriting and hallucination suppression. We provide insights into effective knowledge updating through systematic analysis of learning parameters and data configurations. In our experimental evaluation with web articles published after the base model's knowledge cutoff, PASTA achieved remarkable improvement from 0.02 to 0.82 accuracy while maintaining general language capabilities, demonstrating its effectiveness for creating domain-specialized LLMs.
Chinese Translation
在预训练的大型语言模型（LLMs）中，知识更新仍然是一个重要的挑战。尽管持续训练提供了知识更新的潜在途径，但仍然存在相当大的技术难题。此外，LLMs在准确回答有关特定事实信息的问题时常常面临困难，例如新闻文章——这一能力限制在研究界广泛被认可。本文提出了PASTA，一个简单而强大的框架，用于将来自新闻文章的详细事实信息作为新知识整合到LLMs中，主要目标是构建能够准确回答有关该知识问题的专业化模型。我们的框架结合了数据增强、问答生成和一种新颖的自学习DPO（Direct Preference Optimization）过程，该过程同时实现了知识覆盖和幻觉抑制。我们通过对学习参数和数据配置的系统分析，提供了有效知识更新的见解。在我们对基础模型知识截止日期后发布的网页文章进行的实验评估中，PASTA的准确率从0.02显著提升至0.82，同时保持了通用语言能力，证明了其在创建领域专业化LLMs方面的有效性。
cs.CL / 22 / 2606.28916
Latent Bridges for Multi-Table Question Answering
多表问答的潜在桥梁
Abstract
We introduce GRAB, a constructor-encoder-bridge pipeline for table question answering. Our method lifts relational data into an heterogeneous graph, encodes it via message passing, and transfers the signals to an LLM through a small set of query-conditioned latent tokens. This provides the LLM with a compact, task-relevant structural representation together with the flattened text. Crucially, the LLM remains strictly frozen to preserve its general reasoning capabilities; we train only the lightweight graph encoder and latent bridge (91M parameters), allowing the entire pipeline to be trained efficiently. Our pipeline significantly improves performance on relational Question Answering, with the largest gains in demanding multi-table settings, offering an efficient, principled way to connect relational deep learning with LLMs.
Chinese Translation
我们介绍了 GRAB，一种用于表格问答的构造-编码-桥接管道。我们的方法将关系数据提升为异构图，通过消息传递进行编码，并通过一小组查询条件的潜在标记将信号传递给大型语言模型（LLM）。这为 LLM 提供了一个紧凑的、与任务相关的结构表示以及扁平化的文本。关键是，LLM 保持严格冻结，以保持其一般推理能力；我们仅训练轻量级图编码器和潜在桥梁（91M 参数），使整个管道能够高效训练。我们的管道显著提高了关系问答的性能，在要求较高的多表设置中获得了最大的提升，提供了一种高效、原则性的方法，将关系深度学习与 LLM 连接起来。
cs.CL / 23 / 2606.28933
FinInvest-GTCN: Explainable Graph-Temporal-Causal Modeling for Risk-Aware Investment Decision Optimization
FinInvest-GTCN：一种可解释的图-时间-因果建模方法用于风险感知的投资决策优化
Abstract
Venture capital (VC) investment decisions face distinct challenges, such as multi-source heterogeneous data, non-stationary time series, and the demand for explainable predictions in high-stakes, low-data settings. To overcome these issues, we introduce \textbf{FinInvest-GTCN}, a Graph-Temporal-Causal Network that redefines the task from content recommendation to quantitative risk-return assessment. This architecture combines a relational graph encoder to capture the investment ecosystem's topology, a multi-scale temporal fusion module to handle long-term dependencies and non-stationarity, and a causal decision head that generates risk-adjusted predictions with interpretable causal attributions. A core innovation is the Meta-Causal Adaptation (MCA) strategy, which facilitates robust fine-tuning for new, data-scarce sectors by aligning updates with causally-plausible structures derived from meta-pretraining. Comprehensive experiments on proprietary VC datasets show that FinInvest-GTCN delivers state-of-the-art results, markedly lowering the primary Risk-Adjusted Mean Squared Error (RA-MSE) to 2.51 from a baseline of 3.05 and boosting the cumulative return of a simulated portfolio by 18.7\%. Ablation studies underscore the essential role of each component, while additional analyses confirm the model's stability, interpretability, and enhanced adaptability. This work pioneers a data-driven, explainable framework for investment decision support.
Chinese Translation
风险投资（VC）投资决策面临着独特的挑战，例如多源异构数据、非平稳时间序列以及在高风险、低数据环境中对可解释预测的需求。为了解决这些问题，我们提出了 extbf{FinInvest-GTCN}，一种图-时间-因果网络，将任务从内容推荐重新定义为定量风险-收益评估。该架构结合了关系图编码器以捕捉投资生态系统的拓扑结构、多尺度时间融合模块以处理长期依赖性和非平稳性，以及生成具有可解释因果归因的风险调整预测的因果决策头。核心创新是元因果适应（Meta-Causal Adaptation, MCA）策略，该策略通过将更新与从元预训练中得出的因果合理结构对齐，促进了对新兴、数据稀缺领域的稳健微调。在专有的风险投资数据集上的全面实验表明，FinInvest-GTCN提供了最先进的结果，将主要的风险调整均方误差（Risk-Adjusted Mean Squared Error, RA-MSE）显著降低至2.51，而基线为3.05，并将模拟投资组合的累计收益提升了18.7%。消融研究强调了每个组件的关键作用，而额外分析则确认了模型的稳定性、可解释性和增强的适应性。本研究开创了一种数据驱动的可解释框架，用于投资决策支持。
cs.CL / 24 / 2606.28938
EVLA: An Electro-Aware Multimodal Assistant for Physically-Grounded Driving Reasoning and Control
EVLA：一种电气感知的多模态助手，用于基于物理的驾驶推理与控制
Abstract
Modern vision-language models (VLMs) for driving assistants typically treat vehicle dynamics as a black box, resulting in decisions that lack awareness of the vehicle's real-time electro-mechanical state. To bridge this gap, we introduce the Electro-Visual-Language Assistant (EVLA) -- a novel framework that combines multi-modal scene understanding with real-time perception of the electrified powertrain state (e.g., motor torque, battery SOC). Our approach features two key innovations: first, a Unified Co-State Encoder (UCSE) that fuses visual, textual, and vehicle-state inputs into a shared latent representation, augmented with an Energy-Efficiency Field to model spatial energy costs; and second, an Electro-aware Structured Reasoning Chain (ESRC), which replaces external chain-of-thought prompting with an internal, deterministic reasoning process grounded in physical constraints and optimization objectives. Trained end-to-end with a physics-guided joint loss, EVLA learns to generate context-aware and energy-optimal driving decisions. Extensive evaluations on a driving QA benchmark demonstrate that EVLA substantially outperforms strong fine-tuned VLM baselines, improving the final score by +0.0871 and accuracy by +5.6\%. Ablation studies validate the necessity of each component, and efficiency analyses show that EVLA achieves 36\% faster inference than multi-stage pipelines. This work underscores that integrating vehicle-state awareness and structured physical reasoning is crucial for developing next-generation, physically-grounded driving assistants.
Chinese Translation
现代的视觉-语言模型（VLMs）在驾驶助手中的应用通常将车辆动力学视为一个黑箱，导致决策缺乏对车辆实时电气机械状态的感知。为了解决这一问题，我们提出了电气视觉语言助手（EVLA）——一个结合多模态场景理解与电动动力系统状态（例如，电机扭矩、电池状态（SOC））实时感知的新框架。我们的方法具有两个关键创新：首先，统一共态编码器（Unified Co-State Encoder，UCSE）将视觉、文本和车辆状态输入融合为共享的潜在表示，并通过能效场（Energy-Efficiency Field）来建模空间能量成本；其次，电气感知的结构化推理链（Electro-aware Structured Reasoning Chain，ESRC）用一个基于物理约束和优化目标的内部确定性推理过程替代了外部的思维链提示。EVLA通过物理引导的联合损失进行端到端训练，学习生成上下文感知和能量最优的驾驶决策。在一个驾驶问答基准上的广泛评估表明，EVLA显著超越了强大的微调VLM基线，最终得分提高了+0.0871，准确率提高了+5.6%。消融研究验证了每个组件的必要性，效率分析显示EVLA的推理速度比多阶段管道快36%。这项工作强调了整合车辆状态感知与结构化物理推理对于开发下一代基于物理的驾驶助手的重要性。
cs.CL / 25 / 2606.28943
A3M: Adaptive, Adversarial and Multi-Objective Learning for Strategic Bidding in Repeated Auctions
A3M：用于重复拍卖中战略出价的自适应、对抗性和多目标学习
Abstract
Learning to bid in repeated multi-unit auctions with bandit feedback poses a fundamental challenge. Existing methods often rely on rigid explore-then-exploit schedules, assume stationary adversaries, and optimize solely for bidder utility, thereby limiting adaptability and strategic robustness. To address these limitations, we introduce the A3M framework, which integrates adaptive deep reinforcement learning (DRL), explicit adversarial reasoning, and principled multi-objective reward design for online auction strategy optimization. A3M employs an actor-critic DRL backbone to dynamically balance exploration and exploitation, an opponent model for fictitious play against non-stationary adversaries, and a composite reward function to jointly maximize utility, auctioneer revenue, and fairness. We provide the first comprehensive empirical evaluation of this integrated approach against established baselines in both discriminatory and uniform price auctions. Results show that A3M reduces final regret by 30--40\% in standard settings, maintains robust performance against adversarial strategy shifts, scales favorably with the number of units $K$, and enables tunable multi-objective trade-offs. An extensive ablation study confirms the necessity of each core component. Our work establishes A3M as a powerful and flexible framework for learning in complex auction environments.
Chinese Translation
在具有强盗反馈的重复多单位拍卖中学习出价是一项基本挑战。现有方法通常依赖于僵化的探索-利用调度，假设对手是静态的，并且仅优化出价者效用，从而限制了适应性和战略稳健性。为了解决这些局限性，我们提出了A3M框架，该框架集成了自适应深度强化学习（DRL）、明确的对抗推理和原则性多目标奖励设计，以实现在线拍卖策略优化。A3M采用了一个演员-评论家DRL骨干网络，以动态平衡探索与利用，使用对手模型进行虚拟博弈以应对非静态对手，并设计了一个复合奖励函数以共同最大化效用、拍卖者收入和公平性。我们首次对这一集成方法在歧视性和统一价格拍卖中与既定基线进行了全面的实证评估。结果表明，A3M在标准设置中将最终遗憾降低了30%至40%，在对抗性策略变化下保持稳健性能，随着单位数量$K$的增加而表现良好，并且能够实现可调的多目标权衡。广泛的消融研究确认了每个核心组件的必要性。我们的工作确立了A3M作为在复杂拍卖环境中学习的强大而灵活的框架。
cs.CL / 26 / 2606.28963
Beyond the Mean: Three-Axis Fidelity for Aligning LLM-Based Survey Simulators from Small Pilot Data
超越均值：基于小型试点数据对 LLM（大型语言模型）驱动的调查模拟器进行三轴保真度对齐
Abstract
Large language models (LLMs) are increasingly used to simulate social survey responses, yet their outputs exhibit systematic biases: marginal distributions are skewed, response variance is poorly calibrated, and predictor-outcome relationships are attenuated. We ask a simple question: given a small pilot sample of human responses, can an LLM recover the statistical characteristics of a broader population? We decompose recovery along three axes: structural fidelity, marginal fidelity, and individual fidelity. Using a COVID-19 misinformation survey as a case study, we benchmark three families of approaches: prompting, rectification, and fine-tuning. The findings suggest that fine-tuning on small pilot samples offers a balanced approach for achieving multiple forms of fidelity, but the levels of such fidelity can vary across subsamples, potentially threatening pluralistic alignment.
Chinese Translation
大型语言模型（LLMs）越来越多地用于模拟社会调查响应，但其输出表现出系统性偏差：边际分布偏斜、响应方差校准不良，以及预测变量与结果之间的关系减弱。我们提出一个简单的问题：在给定一小型试点样本的人类响应的情况下，LLM 能否恢复更广泛人群的统计特征？我们沿三个轴进行恢复的分解：结构保真度、边际保真度和个体保真度。以 COVID-19 虚假信息调查为案例，我们对三种方法进行了基准测试：提示、修正和微调。研究结果表明，在小型试点样本上进行微调提供了一种实现多种保真度的平衡方法，但这种保真度的水平可能在子样本之间有所不同，这可能威胁到多元对齐。
cs.CL / 27 / 2606.28978
Can LLMs Hire Fairly? Racial Bias in Resume Screening
大型语言模型能否公平招聘？简历筛选中的种族偏见
Abstract
We audit fourteen mainstream large language models (LLMs) for hiring discrimination using the paired-resume methodology of Kline, Rose, and Walters (2022). The sole 2023-vintage model reproduces the pro-White callback gap documented in field experiments on labor market discrimination ($+2.12$ pp, significant at the 1\% level). Every model released in 2024 or after shows either a null gap or a significant pro-Black reversal (up to $-3.01$ pp). The same pattern holds on the gender axis. Based on 24,024 paired postings per model across 14 models, our results document a reversal in the direction of algorithmic hiring bias across model generations.
Chinese Translation
我们使用Kline、Rose和Walters（2022）提出的配对简历方法，对十四个主流大型语言模型（LLMs）进行招聘歧视审计。唯一的2023年模型再现了在劳动市场歧视领域实验中记录的亲白人回调差距（$+2.12$个百分点，在1\%水平上显著）。所有2024年或之后发布的模型显示出要么没有差距，要么出现显著的亲黑人逆转（最高达到$-3.01$个百分点）。在性别维度上也呈现出相同的模式。基于每个模型24,024个配对发布的结果，我们的研究记录了算法招聘偏见在模型代际间方向的逆转。
cs.CL / 28 / 2606.28992
Fine-Tuning General-Purpose Large Language Models for Agricultural Applications:A Reproducible Framework and Evaluation Protocol Based on Qwen3-8B
针对农业应用的通用大型语言模型的微调：基于 Qwen3-8B 的可重复框架和评估协议
Abstract
General-purpose large language models (LLMs) have demonstrated strong abilities in opendomain question answering, information extraction, and text generation. Agricultural applications, however, are domain-specific, region-dependent, time-sensitive, and safety-critical. Without data governance, expert evaluation, and evidence constraints, an agricultural assistant mayproduce unreliable advice on crop diseases, pesticide use, fertilization, or policy interpretation.To avoid presenting unverified simulated numbers as real experimental findings, this paper doesnot report any model-performance claims that have not been produced by an actual training runand expert evaluation. Instead, we propose AgriTune-R, a reproducible and auditable frameworkfor adapting general-purpose LLMs to agricultural tasks. The framework selects the publiclyverifiable Qwen3-8B model as the recommended base model and integrates agricultural datagovernance, instruction construction, LoRA/QLoRA parameter-efficient fine-tuning, retrievalaugmented generation, expert evaluation, and safety control for high-risk questions. The contributions are: (1) a structured workflow for agricultural LLM adaptation; (2) an evaluationprotocol for agricultural knowledge QA, pest and disease consultation, cultivation management,and policy explanation; (3) an expert-review rubric combining factuality, safety, evidence consistency, and uncertainty expression; and (4) a clear separation between protocol design andempirical conclusions, providing an executable baseline for future empirical studies.
Chinese Translation
通用大型语言模型（LLMs）在开放领域问答、信息提取和文本生成方面表现出了强大的能力。然而，农业应用是特定领域的、依赖于地区的、时间敏感的，并且具有安全关键性。如果没有数据治理、专家评估和证据约束，农业助手可能会对作物病害、农药使用、施肥或政策解读提供不可靠的建议。为了避免将未经验证的模拟数据作为真实实验结果呈现，本文不报告任何未通过实际训练运行和专家评估产生的模型性能声明。相反，我们提出了 AgriTune-R，这是一个可重复和可审计的框架，用于将通用 LLM 适应于农业任务。该框架选择了公开可验证的 Qwen3-8B 模型作为推荐基础模型，并整合了农业数据治理、指令构建、LoRA/QLoRA 参数高效微调、检索增强生成、专家评估和高风险问题的安全控制。贡献包括：（1）农业 LLM 适应的结构化工作流程；（2）农业知识问答、病虫害咨询、种植管理和政策解释的评估协议；（3）结合事实性、安全性、证据一致性和不确定性表达的专家评审标准；（4）协议设计与实证结论之间的明确分离，为未来的实证研究提供了可执行的基线。
cs.CL / 29 / 2606.28999
BERTomelo: Your Portuguese Encoder Best Friend
BERTomelo：您的葡萄牙语编码器最佳伙伴
Abstract
Encoders have become the state of the art for multiple NLP tasks, especially those requiring deep contextual understanding. While multilingual models offer broad coverage, dedicated monolingual encoders are essential for capturing the unique lexical and syntactic nuances of specific languages. For Portuguese, however, existing monolingual options like BERTimbau and Albertina have not kept pace with recent architectural breakthroughs, often lagging behind English benchmarks in scalability and efficiency. This work introduces BERTomelo, a next-generation monolingual encoder pre-trained from scratch and specifically optimized for the Portuguese language. By leveraging the ModernBERT architecture, BERTomelo overcomes the limitations of previous models, offering Base and Large versions with a 1,024-token context window and hardware-level optimizations like FlashAttention and alternating attention mechanisms. The model was trained on ClassiCC-PT, a massive, high-quality Portuguese corpus of 106 million documents, ensuring superior alignment with the language's contemporary usage. The results demonstrate that BERTomelo not only outperforms previous Portuguese encoders but also provides a more robust and efficient alternative to massive multilingual models in downstream tasks such as STS and NER.
Chinese Translation
编码器已成为多种自然语言处理（NLP）任务的最新技术，尤其是在需要深度上下文理解的任务中。虽然多语言模型提供了广泛的覆盖，但专用的单语编码器对于捕捉特定语言的独特词汇和句法细微差别至关重要。然而，对于葡萄牙语而言，现有的单语选项如 BERTimbau 和 Albertina 并未跟上最近的架构突破，通常在可扩展性和效率上落后于英语基准。本研究介绍了 BERTomelo，一种从零开始预训练的下一代单语编码器，专门针对葡萄牙语进行了优化。通过利用 ModernBERT 架构，BERTomelo 克服了先前模型的局限，提供了具有 1,024 令牌上下文窗口的 Base 和 Large 版本，并实现了 FlashAttention 和交替注意机制等硬件级优化。该模型在 ClassiCC-PT 上进行了训练，这是一个包含 1.06 亿个高质量葡萄牙语文档的大型语料库，确保与该语言当代用法的优越对齐。结果表明，BERTomelo 不仅在性能上超越了先前的葡萄牙语编码器，而且在下游任务如语义文本相似性（STS）和命名实体识别（NER）中，提供了比大型多语言模型更强大和高效的替代方案。
cs.CL / 30 / 2606.29024
Conversational Domain Adaptation of IndicTrans2 across 21 Indic Languages via Experience Replay and Model Soups
通过经验重放和模型混合实现IndicTrans2在21种印度语言中的对话领域适应
Abstract
IndicTrans2 is the strongest open English to Indic translation system, but like most systems it is trained on general text and tends to sound stiff on casual, conversational input. We adapt IndicTrans2-1B to conversational register across all 21 Indic languages using only public data (OpenSubtitles, BPCC-H-Daily, Tatoeba). Plain fine-tuning improves conversational chrF but forgets the general domain (it drops 3.9 chrF on FLORES for Hindi). Mixing general data back into training (experience replay) and then averaging the fine-tuned weights with the base (model souping) removes that trade-off: the resulting model beats IndicTrans2-1B on conversational chrF in every one of the 21 languages (mean +6.2) while matching it on FLORES (mean change -0.17, all within 0.7 chrF). Paired bootstrap tests confirm the conversational gains are significant (p <= 0.004) and that FLORES is not significantly degraded. We are deliberate about scope: these are chrF gains, and a blind human plus multi-model LLM check does not confirm them as a perceived quality improvement, so we treat the conversational gain as largely a register match to the references rather than proof of better translation. The techniques are not new; the contribution is the honest, end-to-end study in the Indic conversational setting.
Chinese Translation
IndicTrans2是最强大的开放式英语到印度语言翻译系统，但与大多数系统一样，它是在通用文本上训练的，往往在随意的对话输入中显得生硬。我们使用公共数据（OpenSubtitles、BPCC-H-Daily、Tatoeba）将IndicTrans2-1B适应于所有21种印度语言的对话语体。简单的微调提高了对话的chrF，但遗忘了通用领域（在FLORES数据集上，印地语的chrF下降了3.9）。将通用数据重新混入训练（经验重放），然后将微调后的权重与基础模型进行平均（模型混合），消除了这种权衡：结果模型在21种语言的对话chrF上均超过IndicTrans2-1B（平均提高6.2），同时在FLORES上与其持平（平均变化-0.17，均在0.7 chrF之内）。配对自助法测试确认对话的提升是显著的（p <= 0.004），并且FLORES没有显著下降。我们对研究范围持谨慎态度：这些是chrF的提升，而盲人评估和多模型LLM检查并未确认其为感知质量的改善，因此我们将对话的提升视为与参考文献的语体匹配，而非更好翻译的证明。这些技术并不新颖；贡献在于在印度对话环境中进行诚实的端到端研究。
cs.CL / 31 / 2606.29031
How to Leverage Synthetic Speech for LLM-Based ASR Systems?
如何利用合成语音提升基于大语言模型的自动语音识别系统？
Abstract
In regulated domains such as banking and healthcare, where privacy constraints make real speech costly to collect and retain, synthetic speech from modern text-to-speech (TTS) is an appealing alternative for training automatic speech recognition (ASR) without exposing sensitive customer recordings. Yet a persistent distributional gap between synthetic and real data limits how far it can replace genuine recordings. Prior work largely treats this gap as a black box to be engineered around, but in our work, we instead examine its origin directly by probing a SLAM-ASR architecture. Then, we localise where its LLM backbone separates real from synthetic speech and find the discriminative signal concentrated in the early-to-middle layers, where temporal and prosodic perturbations disrupt it most. We further show that representation-level separability, help, but does not directly predict downstream ASR gains. On the other hand, convolving synthetic audio with room impulse responses (RIRs) narrows the gap not by making synthetic speech sound cleaner or more natural, but by reproducing the acoustic irregularities of real recordings. Translating these findings into the training procedure, by adding a layer-selection module combined with RIR augmentation matches a fully real-data baseline using only 25% of the real speech (13.6h) and surpasses it at all higher proportions.
Chinese Translation
在银行和医疗等受监管领域，由于隐私限制使得真实语音的收集和保存成本高昂，现代文本转语音（TTS）生成的合成语音成为训练自动语音识别（ASR）的一个有吸引力的替代方案，而无需暴露敏感的客户录音。然而，合成数据与真实数据之间持续存在的分布差距限制了合成语音替代真实录音的程度。以往的研究大多将这一差距视为一个黑箱进行工程化处理，而我们的研究则直接探讨其起源，通过探测SLAM-ASR架构来进行分析。我们定位了其大语言模型（LLM）骨干网络在何处区分真实与合成语音，并发现区分信号主要集中在早期到中间层，这些层中时间和韵律的扰动对其影响最大。我们进一步表明，表示层面的可分离性有所帮助，但并不能直接预测后续ASR的提升。另一方面，通过与房间脉冲响应（RIR）卷积合成音频，缩小了这一差距，并不是通过使合成语音听起来更清晰或更自然，而是通过再现真实录音的声学不规则性。将这些发现转化为训练过程，通过添加一个与RIR增强相结合的层选择模块，仅使用25%的真实语音（13.6小时）就能匹配完全真实数据基线，并在所有更高比例上超越该基线。
cs.CL / 32 / 2606.29034
The strength of clinical evidence is recoverable from language model representations but not from their stated grades
临床证据的强度可以从语言模型的表征中恢复，但无法从其声明的等级中获得
Abstract
Large language models (LLMs) increasingly summarize clinical evidence, where a claim's weight depends on how strongly it is supported. Yet these models convey confidence poorly, and properties they never state, such as truth, are often readable from their activations. Whether a clinical model registers evidence strength, distinct from truth, and states it when asked is untested, and any such signal could be lexical. We compiled 45,134 clinical claims from six public sources, harmonized 20,611 into a four-level evidence grade under three independent frameworks, and tested 22 local, open-weight LLMs from several developers (0.6-70 billion parameters; general, medical, and reasoning), with lexical, truth, and cross-framework controls. A linear estimator recovered the grade in every model (median AUROC 71.8), yet decodability did not rise with scale and was weakest in reasoning models. The grade the models stated fell to chance, 25-27 percentage points below the estimator. The recoverable signal was largely lexical and did not transfer across topics or frameworks, yet it was distinct from factual truth and still flagged weakly supported claims (AUROC 69.2). Clinical LLMs thus carry an ordered evidence-strength signal they do not express, so their stated grades fail to convey a claim's support even when it is recoverable from their representations and text.
Chinese Translation
大型语言模型（LLMs）越来越多地总结临床证据，其中一个主张的权重取决于其支持的强度。然而，这些模型在传达信心方面表现不佳，诸如真实性等它们从未声明的属性，往往可以从其激活中读取。尚未测试临床模型是否能够记录证据强度（与真实性不同），并在被询问时陈述该强度，且任何此类信号可能是词汇性的。我们从六个公共来源编制了45,134个临床主张，将其中20,611个在三个独立框架下协调为四级证据等级，并测试了来自多个开发者的22个本地开放权重LLM（参数量从6亿到700亿不等；包括通用、医学和推理模型），并进行了词汇、真实性和跨框架的控制。线性估计器在每个模型中恢复了等级（中位AUROC为71.8），然而可解码性并未随着规模的扩大而提高，在推理模型中最弱。模型所声明的等级降至偶然水平，比估计器低25-27个百分点。可恢复的信号主要是词汇性的，并且在主题或框架之间未能转移，但它与事实真相是不同的，并且仍然能够标记出支持较弱的主张（AUROC为69.2）。因此，临床LLMs携带了一种有序的证据强度信号，但它们并未表达，因此即使在从其表征和文本中可恢复的情况下，它们所声明的等级也无法传达主张的支持程度。
cs.CL / 33 / 2606.29066
Masked Diffusion Decoding as $x$-Prediction Flow
掩码扩散解码作为 $x$-预测流
Abstract
Masked diffusion language models (MDLMs) generate text by iteratively unmasking tokens, but their standard decoder reduces each step to a binary action: a position is either committed to a single token or left fully masked, with no representation of partial belief in between. This all-or-nothing regime discards rich predictive information and forces premature, irrevocable commitments, leading to poor performance under a limited decoding budget. In this paper, we reinterpret mask prediction as clean-state prediction ($x$-prediction) and show that it can be used to induce a continuous flow in input embedding space. Building on this view, we propose a continuous decoding framework for MDLMs where tokens can accumulate partial progress at each diffusion step and remain revisable. To match the uneven contextual constraints across positions in language, we replace the globally synchronous schedule in image diffusion with a confidence-based asynchronous update in which the diffusion progress is token-wise accumulated. Additionally, we introduce a lightweight policy network and formulate its training as a reinforcement learning problem. Applied to pretrained LLaDA, our continuous decoder reaches 97% of its performance on the HumanEval dataset with 25% of decoding budget.
Chinese Translation
掩码扩散语言模型（MDLMs）通过迭代解掩码令牌生成文本，但其标准解码器将每一步简化为二元操作：一个位置要么被确定为单一令牌，要么完全被掩码，中间没有部分信念的表示。这种全有或全无的模式丢弃了丰富的预测信息，并迫使过早且不可逆的承诺，导致在有限解码预算下的性能较差。在本文中，我们将掩码预测重新解释为干净状态预测（$x$-预测），并展示它可以用于在输入嵌入空间中引导连续流。在这一观点的基础上，我们提出了一种MDLMs的连续解码框架，其中令牌可以在每个扩散步骤中积累部分进展并保持可修订性。为了匹配语言中各位置的不均匀上下文约束，我们用基于置信度的异步更新替换了图像扩散中的全局同步调度，其中扩散进展是按令牌逐步累积的。此外，我们引入了一个轻量级策略网络，并将其训练形式化为强化学习问题。应用于预训练的LLaDA，我们的连续解码器在HumanEval数据集上以25%的解码预算达到了97%的性能。
cs.CL / 34 / 2606.29067
ThinkProbe: Beyond Accuracy -- Structural Profiling of Open-Ended LLM Reasoning Traces via Non-Generative Thought Graphs
ThinkProbe：超越准确性——通过非生成性思维图对开放式大语言模型推理轨迹进行结构分析
Abstract
We present ThinkProbe, a framework for structural analysis of LLM reasoning traces. ThinkProbe converts each trace into a Thought Graph a directed graph with cycles, 8 node types, and 6 edge types and derives a 19-metric five-dimensional cognitive profile (5D-CP: Breadth, Depth, Structure, Metacognitive, Efficiency) through a fully non-generative pipeline combining rule-based segmentation and discriminative semantic linking. Applied to 4{,}200 traces from 7 native reasoning models across 200 open-ended questions and 10 cognitive domains, ThinkProbe reveals that reasoning structure is a stable, model-level property: between-model variance exceeds between-domain variance by up to fourfold across four of five cognitive dimensions, with Structure showing genuine sensitivity to question domain, exposing qualitatively distinct cognitive profiles invisible to accuracy-based evaluation.
Chinese Translation
我们提出了ThinkProbe，一个用于大语言模型（LLM）推理轨迹结构分析的框架。ThinkProbe将每个轨迹转换为思维图（Thought Graph），这是一个具有循环的有向图，包含8种节点类型和6种边类型，并通过结合基于规则的分割和判别语义链接的完全非生成性流程，推导出一个19指标的五维认知特征（5D-CP：广度、深度、结构、元认知、效率）。在200个开放式问题和10个认知领域中，对来自7个原生推理模型的4,200条轨迹的应用表明，推理结构是一个稳定的模型级属性：在五个认知维度中的四个维度中，模型间的方差超过领域间的方差，最多可达四倍，其中结构对问题领域表现出真正的敏感性，揭示出基于准确性评估无法察觉的质的不同认知特征。
cs.CL / 35 / 2606.29068
A Comparative Study on Affective Cues in Text Embeddings Across Psychological Emotion Theories
心理情感理论中文本嵌入的情感线索比较研究
Abstract
Text encoders are known for their utility in natural language processing, as they are able to efficiently compress inputs into dense vectors while preserving semantics. These models have been applied to affective computing, in particular to help with solving sentiment analysis and emotion recognition tasks. Nevertheless, it remains unclear to what extent the latent representations produced by modern text encoders capture well-defined psychological theories of affect. In this work, we investigate the affective capabilities of twelve recently released text encoders by probing their generated embeddings as input features for solving regression and classification tasks across three established emotion frameworks, using both word- and sentence-level data. Additionally, we apply a semantic data-leakage prevention technique to improve robustness in word-level evaluations. Our main findings show that the latent manifolds of the latest instruction-aware open-weight encoders enclose an equal or even a larger amount of affective information in comparison with proprietary counterparts when evaluated at word level. In contrast, embeddings of task-tuned and proprietary encoders reach the highest scores on sentence-level affective classification. Furthermore, a qualitative analysis of latent representations and their encoded affective cues is provided.
Chinese Translation
文本编码器因其在自然语言处理中的实用性而受到关注，因为它们能够高效地将输入压缩为密集向量，同时保留语义。这些模型已被应用于情感计算，特别是在解决情感分析和情绪识别任务方面。然而，现代文本编码器所生成的潜在表示在多大程度上能够捕捉到明确的心理情感理论仍然不清楚。在本研究中，我们通过探讨十二种最近发布的文本编码器的生成嵌入，作为解决三个已建立情感框架的回归和分类任务的输入特征，来研究它们的情感能力，使用了词级和句级数据。此外，我们应用了一种语义数据泄露防护技术，以提高词级评估的鲁棒性。我们的主要发现表明，最新的指令感知开放权重编码器的潜在流形在词级评估时包含的情感信息量与专有编码器相当，甚至更多。相反，经过任务调优的专有编码器在句级情感分类中获得了最高分。此外，提供了潜在表示及其编码的情感线索的定性分析。
cs.CL / 36 / 2606.29082
Evolution Fine-Tuning: Learning to Discover Across 371 Optimization Tasks
进化微调：在371个优化任务中学习发现
Abstract
Would experience designing faster GPU kernels also help close in on a long-standing open mathematical conjecture? Large Language Models (LLMs) integrated into evolutionary search have recently produced state-of-the-art solutions on optimization tasks, including open mathematical conjectures, GPU kernel design, scientific law discovery, and combinatorial puzzles. To achieve this, prior work applied search scaffolds to one target task at a time, so every new problem is approached from scratch and the experience accumulated during search is discarded once the model finishes its attempt. This leaves the capability of iteratively evolving a solution (e.g., knowing which part to mutate and how, deciding when to backtrack) entirely in the scaffold rather than in the model itself. Whether the model itself could acquire this capability and reuse it across different tasks has been largely unexamined. To address this, we introduce Evolution Fine-Tuning (EFT), a mid-training paradigm that teaches LLMs to evolve solutions across tasks by converting evolutionary search trajectories into supervision. We construct Finch Collection, a 156K-trajectory dataset spanning 10 domains and 371 optimization tasks, and fine-tune open-source LLMs from 2B to 9B parameters. Empirically, EFT confers cross-task generalization: across 22 held-out tasks, our models surpass their base counterparts by 10.22% on average. Furthermore, when paired with test-time RL, our model matches state-of-the-art performance on two circle-packing tasks and outperforms its base-model counterpart on the Erd\H{o}s minimum-overlap problem. EFT thus serves as a "practice phase" for general-purpose discovery agents that do not solve new problems from scratch.
Chinese Translation
设计更快的GPU内核的经验是否也有助于解决一个长期存在的数学猜想？最近，将大型语言模型（LLMs）集成到进化搜索中，已在优化任务上产生了最先进的解决方案，包括开放数学猜想、GPU内核设计、科学定律发现和组合难题。为了实现这一目标，先前的研究将搜索框架应用于一个目标任务，因此每个新问题都是从头开始处理，并且在模型完成尝试后，搜索过程中积累的经验被丢弃。这使得迭代演化解决方案的能力（例如，知道哪个部分需要变异以及如何变异，决定何时回溯）完全依赖于框架，而不是模型本身。模型本身是否能够获得这种能力并在不同任务之间重用这一能力在很大程度上尚未被研究。为了解决这个问题，我们提出了进化微调（Evolution Fine-Tuning，EFT），这是一种中期训练范式，旨在通过将进化搜索轨迹转化为监督信号来教会LLMs在任务之间演化解决方案。我们构建了Finch Collection，这是一个涵盖10个领域和371个优化任务的156K轨迹数据集，并对参数从2B到9B的开源LLMs进行了微调。实证结果表明，EFT赋予了跨任务的泛化能力：在22个保留任务中，我们的模型平均超越了其基础模型10.22%。此外，当与测试时的强化学习（RL）结合时，我们的模型在两个圆形打包任务上达到了最先进的性能，并在Erd ext{ö}s最小重叠问题上超越了其基础模型。因此，EFT作为一种“实践阶段”，为通用发现代理提供了不从头解决新问题的能力。
cs.CL / 37 / 2606.29090
AB-RAG: Adaptive Budgeted Retrieval-Augmented Generation for Reliable Question Answering
AB-RAG：用于可靠问答的自适应预算检索增强生成
Abstract
Retrieval-Augmented Generation (RAG) has become the standard way to ground large language models in external knowledge, yet most systems retrieve a fixed number of passages for every question regardless of its difficulty. This wastes computation on easy questions, starves hard ones, and gives no signal for when a generated answer can be trusted. With a growing share of question answering systems built on top of commercial language model APIs, a method that can decide how much to retrieve, and how far to trust its own answers, without retraining the underlying model, is of clear practical value. This paper presents AB-RAG (Adaptive Budgeted Retrieval-Augmented Generation), a training-free and backbone-agnostic framework that generates an answer, estimates its confidence from a combination of three signals, and then decides whether to stop or to retrieve more evidence, subject to a fixed retrieval budget. The estimator combines the model's own certainty, the agreement between the answer and the evidence, and the variance of the retrieval scores. For models that expose token probabilities the certainty signal is read directly; for closed APIs it is approximated by self-consistency, so the method works without access to model internals. Across three backbones and two datasets, the central result is that the confidence estimate reliably separates correct from incorrect answers on every backbone, reaching a clean split of 57.6% against 0% Exact Match between high- and low-confidence answers on a factoid dataset. The adaptive policy improves accuracy on capable backbones, and the study reports its negative and nuanced findings honestly, including a confidence signal that proved unsuitable for short answers and a retrieval signal whose sign was found and corrected through measurement. The entire study was carried out on a single consumer laptop with only a few dollars of API spend.
Chinese Translation
检索增强生成（RAG）已成为将大型语言模型与外部知识结合的标准方法，但大多数系统对每个问题检索固定数量的段落，而不考虑其难度。这在简单问题上浪费了计算资源，而在困难问题上则显得不足，并且没有信号表明生成的答案何时可以被信任。随着越来越多的问答系统建立在商业语言模型API之上，能够决定检索多少信息以及在多大程度上信任自身答案的方法，具有明显的实际价值。本文提出了AB-RAG（自适应预算检索增强生成），这是一个无需训练且与基础模型无关的框架，能够生成答案，从三个信号的组合中估计其置信度，然后决定是停止还是检索更多证据，前提是遵循固定的检索预算。该估计器结合了模型自身的确定性、答案与证据之间的一致性以及检索分数的方差。对于暴露令牌概率的模型，确定性信号直接读取；对于封闭API，则通过自一致性进行近似，因此该方法在没有访问模型内部的情况下也能工作。在三个基础模型和两个数据集上，中心结果是置信度估计能够可靠地区分每个基础模型上的正确答案与错误答案，在一个事实数据集上，高置信度与低置信度答案之间的准确匹配率达到了57.6%对0%的干净分割。自适应策略提高了能力强的基础模型的准确性，研究诚实地报告了其负面和细致的发现，包括一个对短答案不适用的置信度信号，以及一个通过测量发现并纠正其符号的检索信号。整个研究在一台普通消费者笔记本电脑上进行，仅花费了几美元的API费用。
cs.CL / 38 / 2606.29119
Knowing in Advance When an Evolutionary Outer Loop Will Not Help: A Pre-Registered Cheap-Baseline Screening Rule
提前知道进化外循环何时无效：一种预注册的廉价基线筛选规则
Abstract
We introduce a pre-registered screening rule that decides, before any implementation, whether an evolutionary / population / lifecycle outer loop over neural-network parameters or structure is worth building. Such outer loops cost 10^2-10^3x their gradient inner loop, yet whether they beat a cheap single-shot alternative is usually discovered only after the expense is paid. Our rule computes, at a Phase-0 gate, a single number: the recovery R = s/G, the best single-shot gradient/curvature statistic's gain s divided by the best gain G of any cheap method evaluated, and prescribes skipping the outer loop when R >= 90%. We validate the rule on a within-lab series of pre-registered outer-loop bets (two analyzed cases plus a disclosed file drawer): in both analyzed cases a static or single-shot computation captured the effect on the project's own metric, the gate fired (R approximately 1.0 in both cases; approximately 0.95 under a stricter metric on one), and the outer loop was abandoned, including one case where a companion factorial decomposition localizes the apparent win to a static substrate change with the evolutionary lifecycle contributing no detectable gain. On one project the gate cost about 50-70 GPU-hours and screened out an estimated 400+ GPU-hours (first cell only) plus weeks of implementation, a 6-8x saving. The rule is prospectively falsifiable: a task with R < 90% where the outer loop still fails to beat single-shot would refute it.
Chinese Translation
我们提出了一种预注册的筛选规则，该规则在任何实施之前决定是否值得建立针对神经网络参数或结构的进化/种群/生命周期外循环。这种外循环的成本是其梯度内循环的10^2-10^3倍，但它们是否优于廉价的单次替代方案通常只有在支付了费用后才能发现。我们的规则在阶段0的关卡上计算一个单一的数值：恢复率 R = s/G，其中最佳单次梯度/曲率统计增益 s 除以任何评估的廉价方法的最佳增益 G，并在 R >= 90% 时建议跳过外循环。我们在实验室内的一系列预注册外循环实验中验证了该规则（分析了两个案例以及一个公开的文件抽屉）：在两个分析的案例中，静态或单次计算捕捉到了对项目自身指标的影响，关卡触发（两个案例中 R 约为 1.0；在一个案例中在更严格的指标下约为 0.95），并且外循环被放弃，包括一个案例，其中一个伴随的因子分解将明显的胜利归因于静态基质变化，而进化生命周期并未贡献任何可检测的增益。在一个项目中，关卡成本约为 50-70 GPU 小时，筛选出了估计超过 400+ GPU 小时（仅限第一个单元）以及数周的实施时间，节省了 6-8 倍。该规则是前瞻性可证伪的：如果在 R < 90% 的任务中外循环仍未能击败单次方案，则将否定该规则。
cs.CL / 39 / 2606.29121
How Anthropomorphic Language Impacts Public Perceptions of AI
拟人化语言如何影响公众对人工智能的认知
Abstract
Public discourse about artificial intelligence (AI) often uses anthropomorphic language: language that attributes human capabilities and characteristics to the system. This practice has been criticized for setting misleading expectations, inflating claims, and fueling hype around AI, which may distort public understanding of AI and impact policy priorities. We study the effects of anthropomorphic framing by comparing changes in participants' perceptions (N=815) when reading passages with and without anthropomorphic language, designed to reflect realistic public-facing AI discourse. We further examine whether these effects differ across two types of AI technologies -- large language models and recommendation systems -- and measure changes in perceptions of AI across several dimensions that are prominent in current public discourse. In a separate condition using a text that explicitly discusses the dangers of AI, we show that individuals' views of AI can shift in response to reading a text; yet in the main conditions of the experiment, where we compare anthropomorphic and non-anthropomorphic descriptions, we find that whether the text uses anthropomorphic language does not substantially affect participants' perceptions of AI. Our results indicate that any immediate effects on public opinions of AI are modest, although they leave open the possibility that anthropomorphic language could have an effect in naturalistic settings, or over gradual, continued exposure.
Chinese Translation
关于人工智能（AI）的公众话语常常使用拟人化语言：即将人类的能力和特征归于系统的语言。这种做法受到批评，因为它可能设定误导性的期望、夸大主张，并助长对人工智能的炒作，从而扭曲公众对人工智能的理解，并影响政策优先事项。我们通过比较参与者（N=815）在阅读包含和不包含拟人化语言的段落时感知的变化，研究拟人化框架的影响，这些段落旨在反映现实的公众面向人工智能的讨论。我们进一步考察这些影响是否在两种类型的人工智能技术之间存在差异——大型语言模型（large language models）和推荐系统（recommendation systems）——并测量公众对人工智能的感知在当前公众话语中突出的几个维度上的变化。在一个单独的条件下，我们使用一段明确讨论人工智能危险的文本，显示个体对人工智能的看法可以在阅读文本后发生变化；然而，在实验的主要条件中，我们比较拟人化和非拟人化描述时发现，文本是否使用拟人化语言并未显著影响参与者对人工智能的感知。我们的结果表明，公众对人工智能的即时看法变化是有限的，尽管这并不排除拟人化语言在自然环境中或在逐渐、持续的接触中可能产生影响的可能性。
cs.CL / 40 / 2606.29130
DistilledGemma: Balanced Efficiency-Accuracy for Person-Place Relation Extraction from Multilingual Historical Articles
DistilledGemma：多语言历史文章中人-地关系提取的平衡效率与准确性
Abstract
We present DistilledGemma, an efficient and accurate system for the HIPE-2026 shared task on person-place relation extraction from multilingual historical newspaper articles in English, German, and French. Our approach adopts a three-stage knowledge distillation pipeline designed to balance classification accuracy with computational efficiency. In the first stage, we systematically explored prompt engineering strategies across eight large language models to identify the most effective reasoning architecture for this challenging task. In the second stage, we applied supervised fine-tuning (SFT) via QLoRA to a Gemma 4 26B A4B teacher model, leveraging its strong multilingual capabilities to generate silver-standard chain-of-thought traces across the training corpus. In the final stage, we performed response-level distillation to transfer these learned reasoning patterns into a compact Gemma 4 E2B student model. In the official evaluation, our team WHEREAMI ranked 3rd on the standard test set with an accuracy profile mean score of 0.688, and 2nd on the binary test set with a mean score of 0.8156. Notably, by distilling knowledge from the 26B teacher to the 2.3B student, we preserved strong reasoning capabilities while reducing the deployed model size to approximately 2.3B effective parameters; the LoRA adapters used during training were merged into the student for inference. This configuration ranked 2nd in the balanced efficiency-accuracy profile across both the standard and binary test sets. These results demonstrate that knowledge distillation provides a practical and scalable solution for historical document processing, achieving competitive performance without excessive computational cost.
Chinese Translation
我们提出了DistilledGemma，这是一个高效且准确的系统，旨在参与HIPE-2026共享任务，从英语、德语和法语的多语言历史报纸文章中提取人-地关系。我们的方法采用了一个三阶段的知识蒸馏流程，旨在平衡分类准确性与计算效率。在第一阶段，我们系统地探索了八个大型语言模型的提示工程策略，以识别出该挑战性任务中最有效的推理架构。在第二阶段，我们通过QLoRA对Gemma 4 26B A4B教师模型进行了监督微调（SFT），利用其强大的多语言能力在训练语料库中生成银标准的思维链踪迹。在最后阶段，我们进行了响应级别的蒸馏，将这些学习到的推理模式转移到一个紧凑的Gemma 4 E2B学生模型中。在官方评估中，我们的团队WHEREAMI在标准测试集上排名第三，准确率平均得分为0.688，在二元测试集上排名第二，平均得分为0.8156。值得注意的是，通过将知识从26B教师模型蒸馏到2.3B学生模型，我们在将部署模型大小减少到约2.3B有效参数的同时，保留了强大的推理能力；在训练过程中使用的LoRA适配器被合并到学生模型中用于推理。该配置在标准和二元测试集的平衡效率-准确性评估中排名第二。这些结果表明，知识蒸馏为历史文档处理提供了一种实用且可扩展的解决方案，能够在不产生过高计算成本的情况下实现竞争力的性能。
cs.CL / 41 / 2606.29213
Can OCR-VLMs Read Devanagari? A Stress-Test Benchmark and Post-Correction Study
OCR-VLM能读取天城文吗？压力测试基准与后纠正研究
Abstract
OCR systems, ranging from classical engines to specialised OCR vision-language models (OCR-VLMs) and frontier multimodal LLMs, report strong results on English and Chinese document benchmarks, yet their behaviour on Indic scripts is largely uncharacterised. We benchmark ten systems on Devanagari (Hindi): classical EasyOCR; open VLMs (Qwen2.5-VL-3B, Qwen3-VL-8B, olmOCR-7B); specialised OCR-VLMs (DeepSeek-OCR, Unlimited-OCR); and frontier closed models (Gemini 2.5 Flash, Claude Opus 4.7, GPT-5.5, Mistral OCR), across four synthetic degradation conditions and 300 real printed scans. We report four findings. First, on clean rendered text all ten cluster within chrF++ 91 to 98, so synthetic text does not separate them. Second, under degradation the specialised OCR-VLMs are the most fragile: DeepSeek-OCR suffers rare but catastrophic repetition failures (outputs up to 71 the reference length) that wreck its corpus mean even though its median is the best of any system, which is why we report median and catastrophic-rate instead of the mean. Third, on real scans nine of the ten systems collapse (EasyOCR falls from chrF++ 93.6 to 58.3) and the field spreads across a 76-point range, so synthetic renders badly overstate Devanagari quality. Fourth, strong English OCR does not predict Indic OCR: GPT-5.5 drops to chrF++ 58.5 (tying classical EasyOCR) and olmOCR-7B, the model behind olmOCR-Bench, falls to 40.5, while the open Qwen3-VL-8B (75.2, runnable on a single 24 GB GPU) beats GPT-5.5 and approaches Mistral; Gemini and Claude lead at 86.3 and 82.2. An error taxonomy separates surface errors (numerals, punctuation) from structural ones (conjuncts, matras, nukta), and a byte-level (ByT5) post-corrector improves a cheap engine on its own error distribution (chrF++ +1.2 to +1.5) but does not transfer across engines. We release the benchmark, code, and models.
Chinese Translation
OCR系统，从经典引擎到专门的OCR视觉语言模型（OCR-VLM）以及前沿的多模态大语言模型（LLM），在英语和中文文档基准测试中表现出色，但它们在印度文字上的表现尚未得到充分研究。我们对十个系统在天城文（印地语）上的表现进行了基准测试：经典的EasyOCR；开放的VLM（Qwen2.5-VL-3B, Qwen3-VL-8B, olmOCR-7B）；专门的OCR-VLM（DeepSeek-OCR, Unlimited-OCR）；以及前沿的封闭模型（Gemini 2.5 Flash, Claude Opus 4.7, GPT-5.5, Mistral OCR），在四种合成降级条件和300个真实打印扫描中进行评估。我们报告了四个发现。首先，在干净的渲染文本上，所有十个系统的chrF++得分均在91到98之间，因此合成文本并未将它们区分开来。其次，在降级情况下，专门的OCR-VLM最为脆弱：DeepSeek-OCR遭遇罕见但灾难性的重复失败（输出长度可达参考长度的71倍），尽管其中位数是所有系统中最好的，但这导致其语料库均值受到严重影响，因此我们报告中位数和灾难性比率而非均值。第三，在真实扫描中，十个系统中有九个崩溃（EasyOCR的chrF++从93.6降至58.3），并且该领域的得分分布在76个点的范围内，因此合成渲染严重高估了天城文的质量。第四，强大的英语OCR并不能预测印度文字OCR：GPT-5.5降至chrF++ 58.5（与经典的EasyOCR持平），而olmOCR-7B（olmOCR-Bench背后的模型）降至40.5，开放的Qwen3-VL-8B（75.2，可在单个24 GB GPU上运行）超越了GPT-5.5并接近Mistral；Gemini和Claude分别以86.3和82.2领先。错误分类法将表面错误（数字、标点符号）与结构性错误（连字、音符、nukta）区分开来，而字节级（ByT5）后纠正器在自身错误分布上改善了一个廉价引擎（chrF++提高1.2到1.5），但未能在不同引擎之间转移。我们发布了基准测试、代码和模型。
cs.CL / 42 / 2606.29228
Understanding Evaluation Illusion in Diffusion Large Language Models
理解扩散大语言模型中的评估幻觉
Abstract
Despite the capability of parallel decoding, diffusion large language models (dLLMs) require many denoising steps to maintain generation quality, motivating recent research on efficient decoding strategies. However, existing studies have reported inconsistent evaluation results even under seemingly identical evaluation settings, risking biased conclusions about dLLM decoding methods. To understand this evaluation concern, we conduct a rigorous evaluation of current decoding methods for dLLMs across diverse evaluation settings. Surprisingly, our analysis reveals that the ranking of decoding methods is highly sensitive to the choice of prompt templates. Single-template evaluation can lead to an illusion that decoding methods improve inference efficiency without performance degradation. Through comprehensive experiments, we find that current parallel decoding methods consistently underperform the single-token decoding baseline, failing to overcome the speed-quality trade-off. We further identify this evaluation inconsistency as the high sensitivity of parallel decoding methods to minor variations in prompt templates. Our experiments show that an effective prompt template can achieve strong evaluation results even with fewer denoising steps, markedly outperforming the marginal gain from increasing denoising steps. Beyond prompt templates, our experiments indicate that overlooked evaluation settings can also notably affect the assessment of decoding methods. Based on these findings, we propose practical guidelines for the reliable evaluation of decoding methods in dLLMs.
Chinese Translation
尽管扩散大语言模型（dLLMs）具备并行解码的能力，但为了维持生成质量，它们仍需进行多次去噪步骤，这促使了近期对高效解码策略的研究。然而，现有研究在看似相同的评估设置下报告了不一致的评估结果，这可能导致对dLLM解码方法的偏见结论。为了解这一评估问题，我们对当前dLLMs的解码方法在多种评估设置下进行了严格评估。令人惊讶的是，我们的分析揭示了解码方法的排名对提示模板的选择高度敏感。单一模板的评估可能导致一种幻觉，即解码方法在不降低性能的情况下提高了推理效率。通过全面的实验，我们发现当前的并行解码方法在性能上始终低于单标记解码基线，未能克服速度与质量之间的权衡。我们进一步识别出这种评估不一致性是由于并行解码方法对提示模板的微小变化高度敏感。我们的实验表明，一个有效的提示模板即使在较少的去噪步骤下也能取得强劲的评估结果，明显优于增加去噪步骤所带来的边际收益。除了提示模板之外，我们的实验还表明，被忽视的评估设置也会显著影响解码方法的评估。基于这些发现，我们提出了可靠评估dLLMs中解码方法的实用指南。
cs.CL / 43 / 2606.29254
Travel-Oriented Reasoning Large Language Model via Domain-Specific Knowledge Graphs
基于领域特定知识图谱的旅行导向推理大语言模型
Abstract
Large language models (LLMs) demonstrate broad reasoning abilities but struggle with accuracy and reliability in specialized domains such as travel, where reasoning depends on precise definitions, rules, and expert-defined conceptual frameworks, and where confident but unfounded outputs arise from a reasoning failure in which the model has not internalized the underlying domain graph rather than from missing domain knowledge alone. We propose a modular pipeline for building a travel-domain reasoning LLM grounded in an expert-designed knowledge graph (KG). Our pipeline integrates a travel KG that encodes domain entities and their relationships, a bottom-up construction procedure that walks the KG to produce multi-hop question answer (QA) pairs, a supervised fine-tuning stage that embeds the domain knowledge into a reasoning-capable LLM using the generated QA pairs as auditable reasoning traces, and a travel-domain benchmark dataset that measures the fine-tuned model's accuracy and calibration. We evaluate our approach using Qwen3-4B with LoRA adaptation. Our reasoning model achieves an $82.4\%$ exact match on the benchmark. This performance significantly outperforms the pretrained Qwen3-4B baseline at $22.4\%$. A calibration analysis decomposes the residual $17.57\%$ of errors into two distinct failure modes: an over-confident multi-label decoder that predicts both correct answers plus one spurious option on most dual-answer mistakes, and a smaller reasoning failure on single-answer questions where the supporting facts are present in the KG but the model fails to reconstruct the correct multi-hop path. This split confirms that explicit KG-grounded reasoning substantially improves the accuracy and uncertainty interpretation of LLMs in specialized domains, and isolates per-option calibration and trace-length-aware decoding as the next axes of improvement.
Chinese Translation
大型语言模型（LLMs）展现了广泛的推理能力，但在旅行等专业领域的准确性和可靠性方面存在困难，因为推理依赖于精确的定义、规则和专家定义的概念框架，并且在推理失败时，模型未能内化基础领域图谱而非仅仅缺乏领域知识，导致自信但无根据的输出。我们提出了一种模块化管道，用于构建基于专家设计的知识图谱（KG）的旅行领域推理LLM。我们的管道整合了一个编码领域实体及其关系的旅行KG，一个自下而上的构建程序，该程序遍历KG以生成多跳问题回答（QA）对，一个监督微调阶段，该阶段使用生成的QA对作为可审计的推理痕迹，将领域知识嵌入到具备推理能力的LLM中，以及一个旅行领域基准数据集，用于测量微调模型的准确性和校准性。我们使用带有LoRA适配的Qwen3-4B评估我们的方法。我们的推理模型在基准测试中实现了82.4%的准确匹配。这一性能显著优于预训练的Qwen3-4B基线的22.4%。校准分析将剩余的17.57%的错误分解为两种不同的失败模式：一种是过于自信的多标签解码器，在大多数双答案错误中预测正确答案和一个虚假选项；另一种是单答案问题上的较小推理失败，其中支持事实存在于KG中，但模型未能重建正确的多跳路径。这一划分确认了显式的KG基础推理显著提高了LLMs在专业领域的准确性和不确定性解释，并将每个选项的校准和考虑追踪长度的解码作为下一步改进的方向。
cs.CL / 44 / 2606.29265
MIThinker: A Plug-and-Play Policy-Optimized Thinker For Motivational Interviewing Counseling
MIThinker：一种即插即用的政策优化思维模型用于动机访谈辅导
Abstract
Reasoning large language models (LLMs) have recently made much progress in complex problem-solving, leveraging internal reasoning (or thought) to guide their solution generation. However, existing LLM-based counseling agents, including those using Motivational Interviewing (MI), generate responses without explicitly aligning thoughts with counseling techniques, limiting their effectiveness. We propose MIThinker, a lightweight thinking model that generates therapeutic thoughts to guide MI counseling agents in strategy selection and response generation. To overcome the lack of annotated thought data, we introduce AugR1-MI, an automated pipeline that reverse-engineers counselor's thoughts from observed responses. Through two-stage training combining supervised fine-tuning and reinforcement learning, MIThinker demonstrates improved theory-of-mind assessment and strategy alignment. Comprehensive evaluations show that MindfulMI, our agent leveraging MIThinker, achieves MI competency comparable to state-of-the-art systems with an order of magnitude less computation.
Chinese Translation
推理大型语言模型（LLMs）最近在复杂问题解决方面取得了显著进展，利用内部推理（或思维）来指导其解决方案生成。然而，现有基于LLM的辅导代理，包括使用动机访谈（MI）的代理，生成的响应并未明确将思维与辅导技术对齐，从而限制了其有效性。我们提出了MIThinker，一种轻量级思维模型，生成治疗性思维以指导MI辅导代理在策略选择和响应生成方面的应用。为了克服缺乏注释思维数据的问题，我们引入了AugR1-MI，一个自动化管道，可以从观察到的响应中逆向推导出辅导员的思维。通过结合监督微调和强化学习的两阶段训练，MIThinker在心智理论评估和策略对齐方面表现出改善。全面评估表明，利用MIThinker的MindfulMI代理在动机访谈能力上达到了与最先进系统相当的水平，同时计算量减少了一个数量级。
cs.CL / 45 / 2606.29273
A Hybrid Framework for Song Lyric Annotation Based on Human-LLM Alignment
基于人类与大型语言模型（LLM）对齐的歌曲歌词注释混合框架
Abstract
Emotion recognition of song lyrics is a challenging task since lyrics may not necessarily align with the overall emotion of a song. As a result, lyrics annotation remains largely underexplored. Drawing inspiration from research in large language model (LLM) assisted annotation, we examine the alignment between humans and LLMs for annotation of lyrics by creating a new sentence-level dataset of lyrics. Our observations highlight the subjectivity of the task and the inherent challenges. Following this, we present a hybrid annotation framework that optimizes human and LLM annotation by predicting potential misalignment in annotation.
Chinese Translation
歌曲歌词的情感识别是一项具有挑战性的任务，因为歌词不一定与整首歌曲的整体情感一致。因此，歌词注释在很大程度上仍未得到充分探索。我们从大型语言模型（LLM）辅助注释的研究中获得灵感，通过创建一个新的句子级歌词数据集，研究人类与LLM在歌词注释中的对齐情况。我们的观察突显了这一任务的主观性及其固有挑战。随后，我们提出了一种混合注释框架，通过预测注释中的潜在不对齐，优化人类和LLM的注释。
cs.CL / 46 / 2606.29375
TriageRA-CCF: Source-Side Clinical Confidence and Coverage Signals for Adaptive Rank Budgeting in Medical LLMs
TriageRA-CCF：用于医疗大型语言模型的源侧临床信心和覆盖信号的自适应秩预算
Abstract
Medical large language models are commonly adapted with a fixed low-rank budget, even though medical questions differ substantially in confidence, clinical coverage, and cross-domain difficulty. We study adaptive rank budgeting for parameter-efficient medical question answering: for each question, the adapter decides whether to activate a small, medium, or large subset of LoRA rank channels. The central challenge is that a naive adaptive budget router can collapse to unstable choices or spend capacity without improving shifted benchmarks. We propose TriageRA-CCF, a source-side teacher for adaptive rank-budgeted LoRA. It combines three signals computed only from source training data: base-model answer confidence, metadata-cell clinical coverage, and a counterfactual close-miss proxy. These signals supervise a straight-through budget router over active ranks {2,4,8}, together with budget-cost, entropy, and rank-balance regularization. Under a matched CMB-source training protocol, TriageRA-CCF achieves the best average accuracy among LoRA, DoRA, and MoELoRA baselines on both Qwen3-8B and Llama3.1-8B. The gains are modest and non-uniform across benchmarks: +0.21 average points over the strongest external baseline on Qwen3-8B and +0.16 on Llama3.1-8B. Component ablations show that confidence, coverage, and counterfactual signals all provide useful budget supervision, but their combination is not monotonically best on every backbone.
Chinese Translation
医疗大型语言模型通常使用固定的低秩预算进行适配，尽管医疗问题在信心、临床覆盖和跨领域难度上存在显著差异。我们研究了用于参数高效医疗问答的自适应秩预算：对于每个问题，适配器决定是否激活小、中或大规模的 LoRA 秩通道。主要挑战在于，简单的自适应预算路由器可能会陷入不稳定的选择，或在不改善转移基准的情况下消耗能力。我们提出了 TriageRA-CCF，一种用于自适应秩预算 LoRA 的源侧教师。它结合了仅从源训练数据计算的三种信号：基础模型答案信心、元数据单元临床覆盖和反事实接近失误代理。这些信号监督一个针对活跃秩 {2,4,8} 的直通预算路由器，并结合预算成本、熵和秩平衡正则化。在匹配的 CMB 源训练协议下，TriageRA-CCF 在 Qwen3-8B 和 Llama3.1-8B 上实现了 LoRA、DoRA 和 MoELoRA 基线中最佳的平均准确率。增益在基准测试中是适度且不均匀的：在 Qwen3-8B 上比最强的外部基线高出 +0.21 平均点，在 Llama3.1-8B 上高出 +0.16。组件消融实验表明，信心、覆盖和反事实信号都提供了有用的预算监督，但它们的组合在每个基础模型上并非单调最佳。
cs.CL / 47 / 2606.29378
Cross-Temporal Sinhala OCR: Page-Level Adaptation and Diachronic Analysis
跨时间的僧伽罗文光学字符识别：页面级适应与历时分析
Abstract
Sinhala is a morphologically rich abugida spoken by roughly 16 million people in Sri Lanka, and to date, there are no publicly available real-world datasets for page-level Sinhala OCR. All previous studies for assessing Sinhala OCR models have used artificially generated data. To bridge the gap, we introduce sinhala-ocr-lk-acts-1010, an annotated dataset of 1,010 page-level images and their transcriptions collected from Sri Lankan Legislative Acts published between 1981-1989 and 2000-2019, split into 707 training examples, 101 validation examples, and 202 testing examples. Three models based on deep learning-based visual language processing, namely DeepSeek-OCR V1, DeepSeek-OCR V2, and LightOnOCR-2-1B, are fine-tuned using QLoRA in 8 experiments conducted on consumer and cloud GPUs. LightOnOCR-2-1B is the top performer, achieving a CER of 1.05% across all test examples, outperforming state-of-the-art open-source OCR models such as Surya-OCR (8.84%) and Tesseract v5 (10.69%), as well as commercially available OCR models such as Google Document AI (2.06%). Our results suggest that LightOnOCR-2-1B outperforms other baselines on real-world OCR tasks and maintains consistent performance across all print periods, even when documents are severely degraded.
Chinese Translation
僧伽罗文是一种形态丰富的音节文字，约有1600万人在斯里兰卡使用，至今尚无公开可用的页面级僧伽罗文光学字符识别（OCR）真实世界数据集。以往所有评估僧伽罗文OCR模型的研究均使用人工生成的数据。为填补这一空白，我们推出了sinhala-ocr-lk-acts-1010，这是一个包含1010个页面级图像及其转录的注释数据集，这些图像来自1981-1989年和2000-2019年间发布的斯里兰卡立法法案，数据集分为707个训练样本、101个验证样本和202个测试样本。基于深度学习视觉语言处理的三种模型，即DeepSeek-OCR V1、DeepSeek-OCR V2和LightOnOCR-2-1B，经过8次在消费级和云GPU上进行的实验，使用QLoRA进行了微调。LightOnOCR-2-1B表现最佳，在所有测试样本中达到了1.05%的字符错误率（CER），超越了当前最先进的开源OCR模型，如Surya-OCR（8.84%）和Tesseract v5（10.69%），以及商业可用的OCR模型，如Google Document AI（2.06%）。我们的结果表明，LightOnOCR-2-1B在真实世界的OCR任务中优于其他基准，并在所有印刷时期保持一致的性能，即使文档严重退化。
cs.CL / 48 / 2606.29407
LC-ICL: Label-Guided Contrastive In-Context Learning for Robust Information Extraction
LC-ICL：基于标签引导的对比上下文学习用于稳健的信息提取
Abstract
There has been increasing interest in exploring the capabilities of advanced large language models (LLMs) in the field of information extraction (IE), specifically focusing on tasks related to named entity recognition (NER) and relation extraction (RE).Although researchers are exploring the use of few-shot information extraction through in-context learning with LLMs, they tend to focus only on using correct or positive examples for demonstration, neglecting the potential value of incorporating incorrect or negative examples into the learning process.In this paper, we present LC-ICL a novel few-shot technique that leverages both correct and incorrect sample constructions to create in-context learning demonstrations. This approach enhances the ability of LLMs to extract entities and relations by combining positive samples with negative samples annotated by error-cause labels. These labels expose more detailed error features in erroneous examples, enabling the model to understand why similar predictions fail and avoid repeating such errors during inference.Specifically, our proposed method taps into the inherent contextual information and valuable information in hard negative samples and the nearest positive neighbors to the test and then applies the in-context learning demonstrations based on LLMs. Our experiments on various datasets indicate that LC-ICL outperforms previous few-shot in-context learning methods, delivering substantial enhancements in performance across a broad spectrum of related tasks. These improvements are noteworthy, showcasing the versatility of our approach in diverse scenarios.
Chinese Translation
近年来，研究人员对先进的大型语言模型（LLMs）在信息提取（IE）领域的能力表现出越来越大的兴趣，特别是针对命名实体识别（NER）和关系提取（RE）相关任务。尽管研究者们正在探索通过上下文学习使用 LLMs 进行少样本信息提取，但他们往往只关注使用正确或正面示例进行演示，忽视了将错误或负面示例纳入学习过程的潜在价值。在本文中，我们提出了 LC-ICL，这是一种新颖的少样本技术，利用正确和错误样本构建来创建上下文学习演示。这种方法通过将正面样本与标注有错误原因标签的负面样本相结合，增强了 LLMs 提取实体和关系的能力。这些标签揭示了错误示例中更详细的错误特征，使模型能够理解类似预测失败的原因，并在推理过程中避免重复此类错误。具体而言，我们提出的方法利用了固有的上下文信息以及来自测试的困难负样本和最近的正邻居中的有价值信息，然后基于 LLMs 应用上下文学习演示。我们在多个数据集上的实验表明，LC-ICL 的表现优于之前的少样本上下文学习方法，在广泛相关任务中实现了显著的性能提升。这些改进值得注意，展示了我们方法在多种场景中的多样性。
cs.CL / 49 / 2606.29424
EntroRouter: Learning Efficient Model Routing via Entropy Regulation
EntroRouter：通过熵调节学习高效模型路由
Abstract
Model routing balances solution accuracy and computational cost by selecting among models of varying capabilities. While recent multi-round frameworks interleave reasoning and planning, we identify a structural failure mode termed Trust Region Collapse. We demonstrate that the deep coupling of reasoning and routing, exacerbated by the dominance of strong pre-training priors under sparse supervision, leads to degenerate local optima where capable experts are systematically suppressed. To decouple these processes, we propose $\textbf{EntroRouter}$, a single-round routing framework that treats entropy regulation as a core objective. We first initialize the policy via Soft Supervision, fitting a distribution of suitable models to establish a high-entropy prior for exploration. Subsequently, we stabilize Reinforcement Learning using a Soft Anchor, which utilizes offline capability estimates to orchestrate controlled entropy contraction within a safe trust region. Extensive experiments demonstrate that EntroRouter retains 98.3% of the strongest expert's accuracy while reducing computational costs by 48.25%.
Chinese Translation
模型路由通过在不同能力的模型之间进行选择来平衡解决方案的准确性和计算成本。尽管最近的多轮框架交替进行推理和规划，但我们识别出一种称为信任区域崩溃（Trust Region Collapse）的结构性失效模式。我们证明了推理与路由的深度耦合，加上在稀疏监督下强预训练先验的主导地位，导致了能力专家系统性被压制的退化局部最优解。为了将这些过程解耦，我们提出了$ extbf{EntroRouter}$，一个将熵调节视为核心目标的单轮路由框架。我们首先通过软监督（Soft Supervision）初始化策略，拟合适合模型的分布，以建立高熵先验进行探索。随后，我们使用软锚（Soft Anchor）稳定强化学习，该方法利用离线能力估计在安全信任区域内协调受控的熵收缩。大量实验表明，EntroRouter在减少计算成本48.25%的同时，保持了最强专家98.3%的准确性。
cs.CL / 50 / 2606.29467
mamabench and mamaretrieval: Benchmarks for Evaluating Medical Retrieval-Augmented Generation in Maternal, Neonatal, and Reproductive Health
mamabench 和 mamaretrieval：评估母婴及生殖健康领域医疗检索增强生成的基准
Abstract
Medical question-answering benchmarks rarely cover the maternal, neonatal, child, and reproductive-health questions a nurse-midwife asks, and, to our knowledge, no public chunk-level relevance benchmark exists for maternal-health guideline retrieval. We release two benchmarks that fill these gaps. mamabench is a scope-filtered QA set of 25,949 items assembled from seven existing expert-authored sources across multiple-choice, short-answer, and rubric-graded tracks; to help users calibrate the LLM judge that scores the rubric track, we re-scope HealthBench's physician-labelled meta-evaluation to the domain. mamaretrieval pairs 3,185 clinical queries with graded (0-6) relevance labels over a 63,650-chunk maternal-health guideline corpus, using a decomposed rubric that distinguishes a chunk that answers a query from one merely on its topic. Three decisions shape both: assemble and filter expert sources rather than author questions, grade relevance rather than binarise it, and measure and disclose the limits of the labels -- scope-classifier agreement, a frontier-judge check, and a pooling-completeness audit -- rather than treat them as an oracle. A companion paper uses the benchmarks to evaluate a deployed on-device assistant; both are released openly for research.
Chinese Translation
医疗问答基准很少涵盖护士助产士所提问的母婴、儿童及生殖健康相关问题，并且据我们所知，目前尚不存在针对母婴健康指南检索的公共块级相关性基准。我们发布了两个填补这一空白的基准。mamabench 是一个经过范围筛选的问答集，包含 25,949 个项目，来源于七个现有专家撰写的多项选择题、简答题和评分标准轨道；为了帮助用户校准评分标准轨道的 LLM 判断者，我们将 HealthBench 的医生标记元评估重新调整到该领域。mamaretrieval 将 3,185 个临床查询与 63,650 块母婴健康指南语料库中的评分（0-6）相关性标签配对，使用一种分解评分标准来区分回答查询的块与仅与其主题相关的块。三个决策影响了这两个基准：组装和筛选专家来源而非自行撰写问题、对相关性进行评分而非二元化处理，以及测量和披露标签的局限性——范围分类器一致性、前沿判断检查和池完整性审计——而不是将其视为神谕。伴随论文使用这些基准评估已部署的设备助手；两者均已公开发布以供研究使用。
cs.CL / 51 / 2606.29481
To Reason or to Fabricate: Reasoning Without Shortcuts via Hint-Anchored Pairwise Aggregation
推理还是伪造：通过提示锚定的成对聚合实现无捷径推理
Abstract
While reinforcement learning (RL) significantly enhances LLM reasoning, its efficacy is severely undermined by Pre-RL data overlap, where RL datasets overlap with pretraining or SFT corpora, causing models to exploit shortcuts by memorizing correct answers and fabricating post-hoc reasoning. To address this, we introduce HIPPO, a novel RL framework that integrates hint-injected aggregation with a tailored pairwise reward model. By utilizing hint injection to deliberately trigger overlap-induced behaviors, the resulting traces naturally serve as explicit anchors for pairwise comparison. This provides highly discriminable preference signals, enabling a lightweight judge model to reliably distinguish genuine reasoning deduction from shortcut-driven rationalization, while the pairwise formulation ensures stable and robust optimization compared to standard PRMs. Extensive experiments demonstrate that HIPPO yields substantial improvements over standard baselines and generalizes effectively to out-of-distribution general tasks, showing it extracts authentic, transferable reasoning skills rather than superficial shortcut patterns.
Chinese Translation
尽管强化学习（RL）显著增强了大语言模型（LLM）的推理能力，但其有效性受到预强化学习数据重叠的严重影响，其中RL数据集与预训练或微调（SFT）语料库重叠，导致模型通过记忆正确答案和伪造事后推理来利用捷径。为了解决这个问题，我们提出了HIPPO，一个新颖的RL框架，结合了提示注入聚合和定制的成对奖励模型。通过利用提示注入故意触发重叠引起的行为，生成的轨迹自然作为成对比较的明确锚点。这提供了高度可区分的偏好信号，使得轻量级判断模型能够可靠地区分真实的推理推导与捷径驱动的合理化，而成对形式确保了与标准PRM相比的稳定和强健的优化。大量实验表明，HIPPO在标准基准上取得了显著改善，并有效地推广到分布外的通用任务，显示其提取了真实的、可转移的推理技能，而非表面的捷径模式。
cs.CL / 52 / 2606.29489
Which Tokens Need Context? A Reference-Based Analysis of Translation Responsibility Using Fertility and Entropy
哪些词汇需要上下文？基于参考的翻译责任分析，使用生育率和熵
Abstract
When humans translate, not every word depends equally on the surrounding context. Some tokens, particularly function words like pronouns and auxiliaries, rely heavily on preceding or following sentences, while others, such as proper nouns, do not. Understanding this inherent context sensitivity is essential for evaluating whether machine translation systems use context in human-like ways. However, existing approaches to analysing context usage rely on discourse-specific test sets or model internals, making them narrow or model-dependent. We propose a post-hoc, model-agnostic framework to quantify context sensitivity at lexical and syntactic levels using two measures derived from word alignments: fertility (number of target tokens generated per source token) and entropy (stability of fertility patterns across contexts). Using reference translations for three language pairs (German $\leftrightarrow$ English, English $\rightarrow$ Hindi) under four context conditions, we show that context selectively redistributes generative responsibility from source to context tokens without altering overall fertility. Function words show the largest fertility reductions, while content words remain stable, suggesting that context resolves ambiguity rather than adding new information. Our framework provides a ground-truth characterisation of selective context usage in human translation, establishing a diagnostic baseline for evaluating machine translation models.
Chinese Translation
在翻译过程中，并非每个词汇都同样依赖于周围的上下文。一些词汇，特别是像代词和助动词这样的功能词，严重依赖于前后句子，而其他词汇，如专有名词，则不然。理解这种固有的上下文敏感性对于评估机器翻译系统是否以类似人类的方式使用上下文至关重要。然而，现有的上下文使用分析方法依赖于特定话语的测试集或模型内部，使得这些方法显得狭隘或依赖于特定模型。我们提出了一种后验的、与模型无关的框架，通过两个基于词对齐的度量来量化词汇和句法层面的上下文敏感性：生育率（每个源词生成的目标词数量）和熵（在不同上下文中生育率模式的稳定性）。在四种上下文条件下，使用三对语言（德语$
ightarrow$英语，英语$
ightarrow$印地语）的参考翻译，我们展示了上下文选择性地将生成责任从源词分配到上下文词汇，而不改变整体生育率。功能词显示出最大的生育率降低，而内容词则保持稳定，这表明上下文解决了歧义，而不是添加新信息。我们的框架提供了对人类翻译中选择性上下文使用的真实特征描述，为评估机器翻译模型建立了诊断基线。
cs.CL / 53 / 2606.29503
The Verbose Context Problem in Medical Records
医疗记录中的冗长上下文问题
Abstract
The verbose context problem occurs when structured concepts have token-inefficient textual representations. This bottleneck is acute in population health: cohort-level analysis of longitudinal patient records requires reasoning over thousands of medically-coded events, often exceeding 400K tokens in total. We present PopMedQA, a benchmark isolating this problem through computational tasks on groups of longitudinal patient records. We construct the benchmark using neopatient, a new library for language-controlled generation of artificial patient records. Through extensive ablations -- including prompting strategies, prompt compression, and agentic decomposition -- we find that domain-independent methods fail to alleviate the verbose context problem. There remains significant opportunity to exploit domain-specific structure in language model inputs for population-scale reasoning.
Chinese Translation
冗长上下文问题发生在结构化概念具有令牌效率低下的文本表示时。这个瓶颈在群体健康领域尤为严重：对纵向患者记录的队列级分析需要对数千个医学编码事件进行推理，通常总令牌数超过40万。我们提出了PopMedQA，这是一个通过对纵向患者记录组进行计算任务来孤立该问题的基准。我们使用neopatient构建该基准，neopatient是一个用于语言控制生成人工患者记录的新库。通过广泛的消融实验——包括提示策略、提示压缩和代理分解——我们发现领域无关的方法无法缓解冗长上下文问题。仍然存在利用语言模型输入中的领域特定结构进行大规模群体推理的重大机会。
cs.CL / 54 / 2606.29534
Preference-ASR: A Preference-Aware Test Set for Benchmarking ASR in the Era of Speech LLMs
Preference-ASR：一个关注偏好的测试集，用于在语音大型语言模型时代基准测试自动语音识别
Abstract
Popular ASR test sets adopt inconsistent conventions for numbers, disfluencies, entities, and casing, while standard normalizers erase the format distinctions users care about. Current benchmarks therefore cannot measure whether a model follows user preferences for output style. We introduce PreferenceASR, a test set evaluating ASR systems on their ability to follow natural-language preference instructions across four categories: normalization, entities, disfluencies, and case. Built from seven open-source corpora via a two-stage LLM-assisted pipeline with human verification, it is evaluated with a preference-aware normalizer that selectively skips steps matching the active instruction. Benchmarking four models shows rankings shift across preference types, exposing quality differences traditional evaluation obscures. We publicly release the dataset.
Chinese Translation
流行的自动语音识别（ASR）测试集在数字、流畅性、实体和大小写方面采用不一致的规范，而标准化工具则抹去了用户关心的格式区别。因此，当前的基准无法衡量模型是否遵循用户对输出风格的偏好。我们提出了PreferenceASR，这是一个评估ASR系统在四个类别（标准化、实体、流畅性和大小写）中遵循自然语言偏好指令能力的测试集。该测试集基于七个开源语料库，通过一个包含人类验证的两阶段大型语言模型（LLM）辅助管道构建，并使用一个关注偏好的标准化工具进行评估，该工具选择性地跳过与当前指令匹配的步骤。对四个模型的基准测试显示，排名在不同偏好类型之间发生变化，揭示了传统评估所掩盖的质量差异。我们将公开发布该数据集。
cs.CL / 55 / 2606.29545
AURORA: Asymmetry and Update-Induced Rotation for Robust Hallucination Detection in Large Language Models
AURORA：用于大型语言模型中鲁棒幻觉检测的非对称性和更新引发的旋转
Abstract
Large Language Models (LLMs) have demonstrated remarkable capabilities across a wide range of natural language processing tasks. However, their tendency to generate hallucinations, namely factually incorrect or unfaithful outputs, poses a critical obstacle to their deployment in high-stakes applications. Although recent hallucination detection methods have made encouraging progress, they typically rely on costly output-level consistency checks or static hidden-state probes that capture shallow dataset-specific patterns, leading to substantial degradation under cross-dataset evaluation. In this work, we propose AURORA, a novel hallucination detection framework that shifts the focus from static representations to the weight-gradient dynamics of LLMs. Our key insight is that hallucinated and faithful answers induce qualitatively different gradient update patterns on the model's parameters. Specifically, hallucinated samples trigger asymmetric and structurally misaligned gradients, which can be captured through two complementary features: (1) the skewness of the cosine similarity distribution between weight matrices and their gradient update directions, and (2) the rotation ratio, which quantifies how much the gradient update reorients the singular-vector basis of weight matrices via SVD. AURORA achieves strong hallucination detection performance across four model families and four benchmark datasets. Further analyses demonstrate that our method scales effectively across model sizes and transfers to out-of-domain tasks, including mathematical reasoning and vision-language scenarios.
Chinese Translation
大型语言模型（LLMs）在广泛的自然语言处理任务中展现了卓越的能力。然而，它们生成幻觉的倾向，即事实错误或不忠实的输出，成为其在高风险应用中部署的关键障碍。尽管近期的幻觉检测方法取得了令人鼓舞的进展，但它们通常依赖于昂贵的输出级一致性检查或静态隐藏状态探测，这些方法捕捉到的是浅层数据集特定的模式，导致在跨数据集评估中显著降级。在本研究中，我们提出了AURORA，一种新颖的幻觉检测框架，重点转向LLMs的权重梯度动态。我们的关键见解是，幻觉和忠实的答案在模型参数上引发了质上不同的梯度更新模式。具体而言，幻觉样本会触发不对称和结构上不对齐的梯度，这可以通过两个互补特征来捕捉：（1）权重矩阵与其梯度更新方向之间的余弦相似度分布的偏斜度，以及（2）旋转比率，该比率量化了梯度更新通过奇异值分解（SVD）重新定向权重矩阵的奇异向量基的程度。AURORA在四个模型系列和四个基准数据集上实现了强大的幻觉检测性能。进一步的分析表明，我们的方法在模型规模上有效扩展，并能够迁移到领域外任务，包括数学推理和视觉-语言场景。
cs.CL / 56 / 2606.29563
Coverage-Driven KV Cache Eviction for Efficient and Improved Inference of LLM
基于覆盖率驱动的键值缓存驱逐策略以提高大型语言模型的推理效率
Abstract
Large language models (LLMs) excel at complex tasks like question answering and summarization, thanks to their ability to handle long-context inputs. However, deploying LLMs is costly, not only due to the high computational demands of quadratic complexity of self-attention and auto-regressive generation, but also because of the significant memory overhead required for storing the key-value (KV) cache during inference. To reduce the memory cost, existing KV-cache eviction strategies leverage the sparsity in attention to selectively store a subset of tokens. While reducing the memory footprint, such approaches show a considerable drop in performance, especially in tasks that require long-context reasoning. We identify that the drop in performance is linked to a reduction in the coverage of unique tokens. Additionally, we theoretically show that reduced coverage limits the mutual information between inputs and outputs, thereby impairing predictive accuracy. To this end, we introduce K-VEC, a novel coverage-aware KV-cache eviction strategy that prioritizes token coverage while evicting tokens in the cache. K-VEC introduces a cross-head and a cross-layer coverage module to enhance token retention across attention heads and model layers, mitigating performance degradation caused by low coverage. Evaluated on 16 LongBench subsets, K-VEC exhibit up to 10.35 points improvement over the existing methods under the same eviction rate and memory constraint. Comprehensive evaluations validate the effectiveness of our approach and demonstrate its potential for efficient LLM deployment in resource-constrained settings.
Chinese Translation
大型语言模型（LLMs）在复杂任务如问答和摘要生成中表现出色，这得益于其处理长上下文输入的能力。然而，部署LLMs的成本高昂，不仅因为自注意力和自回归生成的二次复杂度带来的高计算需求，还因为在推理过程中存储键值（KV）缓存所需的显著内存开销。为了降低内存成本，现有的KV缓存驱逐策略利用注意力中的稀疏性选择性地存储一部分标记。尽管这种方法减少了内存占用，但在需要长上下文推理的任务中，性能却显著下降。我们发现性能下降与独特标记的覆盖率降低有关。此外，我们理论上证明，降低的覆盖率限制了输入与输出之间的互信息，从而损害了预测准确性。为此，我们提出了K-VEC，这是一种新颖的关注覆盖率的KV缓存驱逐策略，在驱逐缓存中的标记时优先考虑标记的覆盖率。K-VEC引入了跨头和跨层的覆盖模块，以增强注意力头和模型层之间的标记保留，从而减轻因覆盖率低导致的性能下降。在16个LongBench子集上的评估表明，在相同的驱逐率和内存限制下，K-VEC相较于现有方法提高了最多10.35分。全面的评估验证了我们方法的有效性，并展示了其在资源受限环境中高效部署LLM的潜力。
cs.CL / 57 / 2606.29571
Anisotropy Decides Cosine vs. Rank Metrics for Text Embeddings
各向异性决定文本嵌入的余弦与排名度量
Abstract
The standard way to compare two text embeddings is cosine similarity. Scattered studies report that a different metric does better, but never pin down the geometric condition that decides when, or why. We settle both with a comprehensive empirical study: nineteen parameter-free similarity metrics on nineteen encoders, from compact sentence transformers up to seven-billion-parameter large language models, across seven datasets. The answer is geometric. When an encoder spreads its variance evenly across directions, cosine is the best parameter-free choice and no other metric helps by a usable margin. When the variance concentrates into a few dominant directions, a property known as anisotropy, rank-based and L1-type metrics beat cosine by a clear margin. The absolute gain is modest, but because cosine starts low on these encoders it is a sizable relative improvement, around twenty percent on average and largest where cosine is weakest. What decides this is the geometry of the embedding space, not how the model was trained: where the two disagree, the metric follows the geometry. One number, the fraction of variance held by the single most dominant dimension, predicts how much the alternatives help across all nineteen encoders, with a rank correlation of 0.86 and a linear correlation of 0.95. To test this as the cause rather than a correlate, we project out the dominant directions: cosine recovers and the advantage of the other metrics nearly vanishes, but only on the encoders that were anisotropic to begin with. The effect is directional, not magnitude based, since it survives normalizing every vector to unit length. Among parameter-free metrics, then, cosine is the right tool wherever an encoder is well spread, which includes the fine-tuned embedders commonly deployed for retrieval, and we give a one-number diagnostic for when it is not.
Chinese Translation
比较两个文本嵌入的标准方法是余弦相似度。散见的研究报告显示，其他度量表现更好，但从未明确指出决定何时或为何的几何条件。我们通过一项全面的实证研究解决了这两个问题：在十九个编码器上测试十九种无参数相似度度量，从紧凑的句子变换器到七十亿参数的大型语言模型，涵盖七个数据集。答案是几何的。当编码器在各个方向上均匀分散其方差时，余弦是最佳的无参数选择，且没有其他度量能以可用的幅度提供帮助。当方差集中在少数主导方向上，即所谓的各向异性属性时，基于排名和L1类型的度量明显优于余弦。绝对增益虽小，但由于余弦在这些编码器上的起始值较低，因此相对改善相当可观，平均约为20%，在余弦最弱的地方效果最大。决定这一点的是嵌入空间的几何形状，而不是模型的训练方式：在两者不一致的地方，度量遵循几何。一个数字，即单一主导维度所占的方差比例，能够预测在所有十九个编码器上替代度量的帮助程度，排名相关性为0.86，线性相关性为0.95。为了验证这一点是因果关系而非相关关系，我们将主导方向投影出去：余弦恢复，其他度量的优势几乎消失，但仅在那些最初具有各向异性的编码器上。该效应是方向性的，而非基于幅度的，因为它在将每个向量归一化为单位长度后仍然存在。因此，在无参数度量中，余弦是合适的工具，适用于那些方差分布良好的编码器，这包括常用于检索的微调嵌入器，我们提供了一个单一数字的诊断来判断何时不适用。
cs.CL / 58 / 2606.29580
MAM-AI: An On-Device Medical Retrieval-Augmented Generation System for Nurses and Midwives in Zanzibar
MAM-AI：一个为赞比亚护士和助产士设计的设备内医疗检索增强生成系统
Abstract
Maternal and newborn mortality remain among the highest in sub-Saharan Africa, where midwifery care is often delivered by nurses who lack midwifery training to international standards, and consulting authoritative guidance at the point of care is hard: the guidelines are long and connectivity is intermittent. We present MAM-AI, a medical question-answering assistant for nurse-midwives in Zanzibar that runs entirely on a commodity Android device: a question is embedded (EmbeddingGemma, 300M) and matched against a curated corpus of 87 guideline documents (63,650 passages), then answered with citations by a 4B int4 generator (Gemma 4 E4B), fully offline, with no query leaving the device. We evaluate the exact deployed configuration with a layered methodology -- retriever, generator under oracle context, end-to-end, and latency -- scored by LLM judges validated against physician rubrics. The evaluation relocates the hard problem. On-device retrieval is essentially solved: the 300M embedder ranks third of seven retrievers and rivals cloud systems, so the passages the system needs are usually found. The small generator is what remains in doubt: adding retrieved context does not improve its answers, and at 4B it cannot be both helpful and safe at once -- of two same-size candidates, the more helpful one commits genuine dangerous errors, so we deploy the other, which is about twice as faithful to its sources (as faithful as a frontier model), and recover its helpfulness with a redesigned prompt that cuts deflection from 33% to 3%. Corpus quality is decisive for the same reason: where the corpus holds the right passage the answer is specific and actionable, and where it does not it goes vague. MAM-AI is a thoroughly evaluated, open-source research prototype, not a fielded product; the system, knowledge base, benchmarks, and evaluation harness are released.
Chinese Translation
母婴死亡率在撒哈拉以南非洲仍然是最高的，助产护理通常由缺乏国际标准助产培训的护士提供，而在护理现场咨询权威指导非常困难：指导方针冗长且网络连接不稳定。我们提出了MAM-AI，一个为赞比亚护士-助产士设计的医疗问答助手，完全运行在普通的Android设备上：问题通过EmbeddingGemma（300M）嵌入，并与87份指导文件（63,650段落）构建的精选语料库进行匹配，然后由一个4B int4生成器（Gemma 4 E4B）离线回答，并且没有查询离开设备。我们使用分层方法评估精确的部署配置——检索器、在oracle上下文下的生成器、端到端和延迟——由经过验证的LLM评审员评分，依据医生评分标准。评估重新定位了难题。设备内检索基本上已解决：300M嵌入器在七个检索器中排名第三，并与云系统相媲美，因此系统所需的段落通常能够找到。小型生成器仍然存在疑问：添加检索到的上下文并未改善其回答，而在4B时它无法同时既有帮助又安全——在两个同尺寸候选中，更有帮助的一个会犯下真正危险的错误，因此我们部署了另一个，它对其来源的忠实度约为两倍（与前沿模型一样忠实），并通过重新设计的提示将偏离率从33%降低到3%来恢复其帮助性。语料库质量因同样原因至关重要：当语料库包含正确段落时，答案具体且可操作，而当没有时则变得模糊。MAM-AI是一个经过全面评估的开源研究原型，而非已部署产品；系统、知识库、基准和评估工具均已发布。
cs.CL / 59 / 2606.29605
How much of an LLM-generated clinical corpus is actually new? A production-scale measurement of content redundancy for provenance classification
LLM生成的临床语料库中有多少内容实际上是新的？针对来源分类的内容冗余的生产规模测量
Abstract
Clinical machine learning increasingly relies on training corpora generated by large language models (LLMs) rather than annotated by clinicians, and such corpora are described and reused largely on the basis of their reported scale. We test whether volume reflects information content. Analysing the complete output of a multi-agent clinical extraction pipeline applied to 167,034 patient narratives, 2.51 billion generated tokens across the ten text-bearing channels of an eleven-channel pipeline, we introduce Provenance-based Redundancy Decomposition, a token-level classification of the entire output by source. Only 10.9% of the output is trainable-unique content while 79.4% is redundant; raw token count overstates information content by roughly ninefold. The redundancy arises through two distinct mechanisms, verbatim copying of source context into per-item fields, and duplication of generated text across records, of which only the former is losslessly removable. An independent, model-free analysis based on lossless compression confirms the redundancy, recovering the two mechanisms without reference to the provenance labels. One pipeline channel carries almost no redundancy, showing that the level of redundancy depends on how each channel is structured rather than being a fixed property of LLM extraction. Because uncorrected redundancy up-weights the longer, more complex presentations that generate the most items, it skews the token-level training distribution of the corpus, a property we measure directly. In a controlled downstream test, de-duplicating the corpus before adaptation improved a clinical encoder on external disease-recognition benchmarks at equal token budget, robustly across adaptation depths and replicated on a second benchmark, confirming that the redundancy carries a measurable cost beyond storage. The classification tool is released openly.
Chinese Translation
临床机器学习越来越依赖于由大型语言模型（LLMs）生成的训练语料库，而不是由临床医生标注的语料库，这些语料库的描述和重用主要基于其报告的规模。我们测试了体量是否反映信息内容。分析应用于167,034个患者叙述的多代理临床提取管道的完整输出，涉及在一个十一通道管道的十个文本承载通道中生成的25.1亿个标记，我们引入了基于来源的冗余分解，这是一种按来源对整个输出进行的标记级分类。只有10.9%的输出是可训练的独特内容，而79.4%是冗余的；原始标记计数大约高估了信息内容九倍。冗余通过两种不同的机制产生：将源上下文逐字复制到每个项目字段，以及在记录之间重复生成的文本，其中只有前者可以无损去除。基于无损压缩的独立模型无关分析确认了冗余，恢复了这两种机制而不参考来源标签。一个管道通道几乎没有冗余，显示冗余水平取决于每个通道的结构，而不是LLM提取的固定属性。由于未经修正的冗余加重了生成最多项目的较长、较复杂的展示，它扭曲了语料库的标记级训练分布，这是我们直接测量的特性。在一个受控的下游测试中，在适应之前去重语料库提高了临床编码器在外部疾病识别基准上的表现，在相同的标记预算下，在适应深度上表现出稳健性，并在第二个基准上复制，确认冗余带来了超出存储的可测成本。分类工具已公开发布。
cs.CL / 60 / 2606.29614
Do We Still Need Fine Tuning? Turkish Sentiment Analysis in the Era of Large Language Model
我们还需要微调吗？大语言模型时代的土耳其情感分析
Abstract
This study examines whether supervised fine-tuning remains necessary for Turkish sentiment analysis in the era of large language models. We compare classical machine learning methods, fine-tuned pretrained language models, and prompted large language models on a Turkish e-commerce review dataset with negative, neutral, and positive labels. Fine-tuned BERTurk models perform best overall and outperform all prompted large language models in the full three-class task. The neutral class emerges as the main difficulty: while several large language models are much more competitive in binary positive--negative classification, they degrade substantially in the three-class setting by collapsing neutral reviews into polarized categories. The findings suggest that, in realistic Turkish sentiment classification, prompted large language models do not yet match supervised fine-tuning in the zero-shot setting, and that including the neutral class is crucial for robust evaluation.
Chinese Translation
本研究探讨在大语言模型时代，监督微调是否仍然是土耳其情感分析所必需的。我们比较了经典机器学习方法、微调的预训练语言模型和提示的大语言模型在一个包含负面、中性和正面标签的土耳其电子商务评论数据集上的表现。微调的 BERTurk 模型整体表现最佳，并在完整的三类任务中超越了所有提示的大语言模型。中性类别成为主要难点：尽管在二元正负分类中，多个大语言模型的竞争力更强，但在三类设置中，它们的表现显著下降，将中性评论归类为极化类别。研究结果表明，在实际的土耳其情感分类中，提示的大语言模型在零样本设置下尚未达到监督微调的水平，并且包含中性类别对于稳健评估至关重要。
cs.CL / 61 / 2606.29639
Two-Stage Prompt Optimization for Few-Shot Relation Extraction: From Reasoning-Guided Search to Gradient-Guided Refinement
用于少样本关系抽取的两阶段提示优化：从推理引导搜索到梯度引导细化
Abstract
Automatic prompt optimization is still underexplored for episodic few-shot relation extraction with smaller language models. We propose a two-stage framework that combines reasoning-based prompt optimization with gradient-based prompt optimization. The first stage can use any reasoning-based optimizer to make broadprompt improvements in natural language. The second stage applies our GradPO, which uses loss and gradient signals to identify high-impact prompt spans and refine them with local edits. Experiments on FS-TACRED and FS-FewRel show that local refinement usually improves prompts found by the first stage, and GradPO is the most consistent refiner. Our framework achieves state-of-the-art performance on FS-TACRED with Qwen3-4B and remains competitive on FS-FewRel.
Chinese Translation
自动提示优化在小型语言模型的情景少样本关系抽取中仍然未得到充分探索。我们提出了一个两阶段框架，将基于推理的提示优化与基于梯度的提示优化相结合。第一阶段可以使用任何基于推理的优化器在自然语言中进行广泛的提示改进。第二阶段应用我们的GradPO，它利用损失和梯度信号识别高影响力的提示片段，并通过局部编辑进行细化。在FS-TACRED和FS-FewRel上的实验表明，局部细化通常能改善第一阶段找到的提示，而GradPO是最一致的细化器。我们的框架在FS-TACRED上实现了最先进的性能，使用Qwen3-4B，并在FS-FewRel上保持竞争力。
cs.CL / 62 / 2606.29648
Hybrid Retriever Evolution for Multimodal Document Reasoning Agents
用于多模态文档推理代理的混合检索器演化
Abstract
Different retrievers, including lexical, semantic, and multimodal approaches, provide highly complementary strengths for multimodal document understanding, yet most systems combine them through fixed pipelines that cannot adapt to the demands of individual reasoning steps. In this work, we ask whether retrieval orchestration itself can be learned as part of the reasoning process. We introduce a failure-driven evolution framework in which a meta-agent autonomously discovers how a tool-using task agent should coordinate diverse retrievers during multi-step document question answering. The meta-agent analyzes incorrect reasoning trajectories, actively probes the same tool environment to diagnose root causes, and iteratively rewrites the task agent's instructions, turning retrieval from a fixed front-end stage into an adaptive, step-wise reasoning decision. The evolved agent learns when to invoke each retriever, how to combine them, and how to compose evidence across modalities and pages. On MMLongBench-Doc and DocBench, the evolved agent achieves gains of up to +19.6 points over the unevolved baseline and consistently outperforms recent systems including MACT, MDocAgent, and SimpleDoc. Detailed retrieval analyses confirm that these improvements arise from adaptive routing and evidence composition rather than reliance on any hard coded retrieval mode, and evolution dynamics reveal a progressive shift from narrow lexical behavior to rich multi-tool coordination. These findings establish autonomous multi-agent coordination as a promising paradigm for multimodal document reasoning.
Chinese Translation
不同的检索器，包括词汇检索、语义检索和多模态方法，为多模态文档理解提供了高度互补的优势，然而大多数系统通过固定的管道将它们结合在一起，无法适应个别推理步骤的需求。在本研究中，我们探讨检索协调本身是否可以作为推理过程的一部分进行学习。我们引入了一种基于失败驱动的演化框架，在该框架中，元代理自主发现工具使用任务代理在多步骤文档问答过程中应如何协调多样的检索器。元代理分析错误的推理轨迹，主动探测相同的工具环境以诊断根本原因，并迭代地重写任务代理的指令，将检索从固定的前端阶段转变为适应性的、逐步的推理决策。演化后的代理学习何时调用每个检索器，如何组合它们，以及如何跨模态和页面组合证据。在 MMLongBench-Doc 和 DocBench 上，演化后的代理相较于未演化的基线取得了最高达 +19.6 分的提升，并且持续超越包括 MACT、MDocAgent 和 SimpleDoc 在内的最新系统。详细的检索分析确认，这些改进源于适应性路由和证据组合，而非依赖于任何硬编码的检索模式，演化动态揭示了从狭窄的词汇行为向丰富的多工具协调的渐进转变。这些发现确立了自主多代理协调作为多模态文档推理的一个有前景的范式。
cs.CL / 63 / 2606.29649
Resolution Thresholds in VLM Detection of Harmful ASCII Art Across Construction Modes and Languages
构建模式和语言中有害 ASCII 艺术的 VLM 检测分辨率阈值
Abstract
Large Vision-Language Models (VLMs) are increasingly deployed as content moderation tools, yet they remain vulnerable to jailbreak attacks in which harmful text is visually encoded as ASCII art. This can allow inappropriate or harmful content to bypass moderation systems. To address this vulnerability, this paper investigates how image resolution affects VLM detection of harmful ASCII art across eight character construction modes (L1-L8), ranging from dense block characters to word-embedded designs. We evaluate eight state-of-the-art VLMs on English and Chinese corpora using a pipeline that generates ASCII art images at ten resolution scales, probing whether a consistent detection-failure threshold exists across models, modes, and languages. Results indicate that detection rates decline sharply above certain resolution thresholds, and that word-based modes are the most resistant to detection across the full resolution range. These findings reveal a systematic vulnerability in VLM-based content moderation systems and motivate resolution-aware evaluation standards.
Chinese Translation
大型视觉语言模型（VLM）越来越多地被用作内容审核工具，但它们仍然容易受到越狱攻击，其中有害文本以 ASCII 艺术的形式进行视觉编码。这可能使不当或有害内容绕过审核系统。为了解决这一脆弱性，本文研究了图像分辨率如何影响 VLM 对有害 ASCII 艺术的检测，涵盖了八种字符构建模式（L1-L8），从密集的块字符到嵌入单词的设计。我们在英语和中文语料库上评估了八种最先进的 VLM，使用一个生成十种分辨率尺度的 ASCII 艺术图像的流程，探讨不同模型、模式和语言之间是否存在一致的检测失败阈值。结果表明，检测率在某些分辨率阈值以上急剧下降，而基于单词的模式在整个分辨率范围内对检测的抵抗力最强。这些发现揭示了基于 VLM 的内容审核系统中的系统性脆弱性，并推动了对分辨率敏感的评估标准的制定。
cs.CL / 64 / 2606.29672
How LLMs See Creativity: Zero-Shot Scoring of Visual Creativity with Interpretable Reasoning
大型语言模型如何看待创造力：基于可解释推理的视觉创造力零-shot评分
Abstract
Evaluating the originality of visual images poses enduring challenges for creativity assessment. Automated scoring using AI models has proven effective in the verbal domain, yet key questions remain about evaluating visual creativity and understanding how models arrive at their ratings. The present research asks whether multimodal large language models (LLMs) can serve as judges of visual creativity zero-shot (without any fine-tuning or examples of human ratings) and whether their "reasoning" output offers an interpretable window into their evaluation process. We tested six multimodal LLMs (Gemini 3 Flash, Gemma 4 31B IT, GPT-5.4 Mini, GLM-5v Turbo, Kimi K2.5, and Qwen 3.6 Plus) on 992 AI-generated images (based on human-written prompts) and 1,500 hand-drawn sketches scored for creativity by human raters. In Study 1, all models showed substantial alignment with human creativity ratings on both datasets (r = .57-.68 on AI-generated images; r = .29-68 on sketches). In Study 2, we analyzed the step-by-step reasoning processes of three LLMs evaluating the same images and drawings. Although reasoning made model evaluations interpretable -- showing what they attend to, how they balance originality vs. quality, and how they justify their ratings -- reasoning did not improve alignment with human ratings. In sum, our findings indicate that multimodal LLMs can match human judgments of visual creativity without any additional training, and that their reasoning reveals how AI models evaluate creativity. An open scoring app implementing this pipeline is available at https://review-visual-eval-scoring.hf.space.
Chinese Translation
评估视觉图像的原创性在创造力评估中面临持久的挑战。使用人工智能模型的自动评分在语言领域已被证明有效，但关于如何评估视觉创造力以及模型如何得出评分的关键问题仍然存在。本研究探讨了多模态大型语言模型（LLMs）是否可以作为视觉创造力的评判者进行零-shot评估（无需任何微调或人类评分示例），以及它们的“推理”输出是否提供了可解释的评估过程窗口。我们对六种多模态LLM（Gemini 3 Flash、Gemma 4 31B IT、GPT-5.4 Mini、GLM-5v Turbo、Kimi K2.5和Qwen 3.6 Plus）进行了测试，使用992幅基于人类编写提示生成的AI图像和1,500幅由人类评分的手绘草图进行创造力评分。在研究1中，所有模型在这两个数据集上与人类创造力评分表现出显著的一致性（AI生成图像的相关系数为r = .57-.68；草图的相关系数为r = .29-68）。在研究2中，我们分析了三种LLM对相同图像和绘画的逐步推理过程。尽管推理使模型评估变得可解释——显示它们关注的内容、如何平衡原创性与质量以及如何为其评分提供依据——但推理并未改善与人类评分的一致性。总之，我们的研究结果表明，多模态LLM可以在没有任何额外训练的情况下匹配人类对视觉创造力的判断，并且它们的推理揭示了人工智能模型如何评估创造力。一个实现该流程的开放评分应用程序可在 https://review-visual-eval-scoring.hf.space 获取。
cs.CL / 65 / 2606.29689
Can MLLMs Critique Like Humans? Evaluating Open-Ended Aesthetic Reasoning in Multimodal Large Language Models
多模态大型语言模型能像人类一样进行批评吗？评估开放式美学推理
Abstract
Open-ended aesthetic critique is a challenge for multimodal large language models (MLLMs): unlike multiple-choice aesthetic benchmarks, it has no single correct answer, and most aesthetic evaluation has measured models against numeric scores rather than the written critiques people actually give. We evaluate MLLM critiques against ranked human references and ask whether they are close to human ones. Using the Reddit Photo Critique Dataset, we score five open-weight MLLMs against multiple ranked human critiques per photo with reference-based similarity metrics, under six prompt conditions that disentangle persona framing, aspect hinting, length control, and single- versus multi-pass generation, and add an image-grounding control that feeds each model the wrong photograph. We find that reference-based similarity gives a misleading picture. Stricter lexical and learned metrics show only weak alignment with human critiques, while a coarse embedding cosine reports broad topical overlap that the grounding control traces to a stable house style rather than image-specific observation. Behaviorally, the models diverge from humans in consistent ways the scores do not surface: even under a length cap they write two to three times as much, cover nearly every aesthetic aspect where humans are selective, engage each aspect more uniformly and at greater depth, and repeat themselves across critiques of the same photo where humans vary. We argue that reference-based similarity rewards a fluent, comprehensive critique style rather than the selectivity and specificity of human critique, and discuss implications for evaluating and training open-ended multimodal generation.
Chinese Translation
开放式美学批评对多模态大型语言模型（MLLMs）来说是一项挑战：与多项选择的美学基准不同，它没有单一的正确答案，而且大多数美学评估是通过数字评分来衡量模型，而非人们实际给出的书面批评。我们将MLLM的批评与排名的人类参考进行评估，并询问它们是否接近人类的批评。使用Reddit照片批评数据集，我们在六种提示条件下对五个开放权重的MLLM进行评分，这些条件解构了角色框架、方面提示、长度控制以及单次与多次生成，并添加了一个图像基础控制，向每个模型提供错误的照片。我们发现，基于参考的相似性给出了误导性的结果。更严格的词汇和学习度量与人类批评的对齐程度较弱，而粗略的嵌入余弦则报告了广泛的主题重叠，这种重叠通过基础控制追溯到一种稳定的风格，而非特定于图像的观察。在行为上，模型在一致的方式上与人类存在差异，而这些差异在评分中并未显现：即使在长度限制下，它们的写作量也比人类多出两到三倍，涵盖了几乎所有人类选择的美学方面，更均匀且更深入地参与每个方面，并在同一照片的批评中重复自己，而人类则有所不同。我们认为，基于参考的相似性奖励了一种流畅、全面的批评风格，而非人类批评的选择性和特异性，并讨论了对开放式多模态生成的评估和训练的影响。
cs.CL / 66 / 2606.29712
Why Struggle with Continuous Latents? Interpretable Discrete Latent Reasoning via Rendered Compression
为何要与连续潜变量抗争？通过渲染压缩实现可解释的离散潜变量推理
Abstract
Large language models achieve high reasoning performance via explicit chain-of-thought and reinforcement learning, but require long output sequences and extended inference time. Latent reasoning reduces this cost by shifting computation into a latent space; however, continuous latent methods are hard to train, suffering from unstable and uninterpretable reasoning trajectories. We argue these issues stem from a misalignment between continuous-space reasoning and discrete symbolic supervision, as continuous states lack explicit anchors for step-by-step alignment. To resolve this, we propose \textbf{Discrete Latent Reasoning~(DLR)}, the first method that converts continuous latent states into explicit discrete tokens. Inspired by render-based compression, we render textual chains of thought into images, extract visual features, and construct a discrete latent vocabulary via clustering-based fine-tuning. Expanding the vocabulary and output head enables standard autoregressive modeling over both natural language and latent tokens, supporting pretraining alignment, SFT, and RL. Experiments on five reasoning benchmarks and two model series~(Qwen3-VL and LLaMA-3) confirm that \textbf{DLR} outperforms prior latent reasoning baselines with up to \textbf{20$\times$ compression}. Furthermore, the learned latent trajectories retain an interpretable semantic structure. Overall, discrete latent tokens provide a controllable and interpretable basis for efficient latent reasoning.
Chinese Translation
大型语言模型通过明确的思维链和强化学习实现了高推理性能，但需要较长的输出序列和延长的推理时间。潜变量推理通过将计算转移到潜在空间来降低这一成本；然而，连续潜变量方法难以训练，推理轨迹不稳定且难以解释。我们认为这些问题源于连续空间推理与离散符号监督之间的不一致，因为连续状态缺乏逐步对齐的明确锚点。为了解决这一问题，我们提出了 extbf{离散潜变量推理~(DLR)}，这是第一种将连续潜变量状态转换为明确离散标记的方法。受基于渲染的压缩启发，我们将文本思维链渲染为图像，提取视觉特征，并通过基于聚类的微调构建离散潜变量词汇。扩展词汇和输出头使得在自然语言和潜变量标记上都能进行标准的自回归建模，支持预训练对齐、SFT和RL。在五个推理基准和两个模型系列（Qwen3-VL 和 LLaMA-3）上的实验确认， extbf{DLR}在推理性能上优于先前的潜变量推理基准，压缩比高达 extbf{20$ imes$}。此外，学习到的潜在轨迹保持了可解释的语义结构。总体而言，离散潜变量标记为高效的潜变量推理提供了可控且可解释的基础。
cs.CL / 67 / 2606.29713
SEVA: Self-Evolving Verification Agent with Process Reward for Fact Attribution
SEVA：具有过程奖励的自我演化验证代理用于事实归属
Abstract
Hallucination is the reliability bottleneck for LLM-based agents, and fact attribution verifiers are the last line of defense -- yet today's verifiers emit only opaque binary labels, leaving agents unable to self-correct and operators unable to audit. We present SEVA, a structured verification agent that emits evidence alignments, step-by-step reasoning chains, calibrated confidence, and a six-category error diagnosis with actionable fixes. Training such an agent with RL is non-trivial: standard binary reward on multi-component output triggers advantage collapse -- within-group reward variance vanishes and the GRPO gradient disappears. We resolve this with a process reward that decomposes verification quality into five independent components weighted 70/30 toward process signals, restoring the gradient and inducing an implicit curriculum -- the agent first masters verification behavior (alignment 0.917 -> 0.997, format 72% -> 100%), then outcomes (F1 64.9 -> 69.0). Structured output further enables a Verify -> Reflect -> Probe -> Refine self-evolution loop, which over four rounds on a 7B model surfaces an unexpected structural finding: each round produces a benchmark-specialist, not a generalist (+15 pp on HaluEval, -10 to -14 pp on TruthfulQA in the same model, persistent at 4x data). On ClearFacts, SEVA-3B matches GPT-4o-mini (69.0 vs. 69.8 F1) while producing substantially richer, auditable output -- confirming a principle that should generalize: for any RL task with multi-component generation, reward granularity must match output granularity.
Chinese Translation
幻觉是基于大规模语言模型（LLM）代理的可靠性瓶颈，而事实归属验证器是最后的防线——然而，当前的验证器仅发出不透明的二元标签，使得代理无法自我纠正，操作员也无法进行审计。我们提出了SEVA，一个结构化的验证代理，能够输出证据对齐、逐步推理链、校准置信度以及六类错误诊断和可操作的修复方案。用强化学习（RL）训练这样的代理并非易事：在多组件输出上使用标准的二元奖励会导致优势崩溃——组内奖励方差消失，GRPO梯度消失。我们通过过程奖励解决了这个问题，该奖励将验证质量分解为五个独立组件，权重为70/30，侧重于过程信号，从而恢复梯度并引导隐式课程——代理首先掌握验证行为（对齐从0.917提升至0.997，格式从72%提升至100%），然后是结果（F1从64.9提升至69.0）。结构化输出进一步支持了一个验证 -> 反思 -> 探测 -> 精炼的自我演化循环，在一个7B模型上经过四轮训练后，出现了一个意想不到的结构发现：每一轮产生一个基准专家，而非通才（在HaluEval上提升15个百分点，在同一模型的TruthfulQA上下降10至14个百分点，在4倍数据下持续存在）。在ClearFacts上，SEVA-3B的表现与GPT-4o-mini相当（F1为69.0对比69.8），同时生成了显著更丰富、可审计的输出——确认了一个应具有普遍适用性的原则：对于任何具有多组件生成的RL任务，奖励的粒度必须与输出的粒度相匹配。
cs.CL / 68 / 2606.29733
How Far Do On-Prem Open LLMs Get on Text-to-SQL? A Cross-Family Size x Technique Frontier on BIRD
本地开放大型语言模型在文本到SQL任务中的表现如何？基于BIRD的跨家族规模与技术前沿研究
Abstract
Organizations that cannot send data to a cloud API increasingly ask: how good is Text-to-SQL if the model must run on-premises on open weights, and which popular accuracy "recipes" are worth their compute? We answer with an honest, fully reproducible benchmark on the BIRD development split (n=1534, Execution Accuracy), evaluating three open model families across two generations -- Qwen2.5-Coder (7B/14B/32B), CodeLlama-Instruct (7B/13B/34B), and Llama-3.x (8B, 70B) -- under one matched protocol, ablating a model-agnostic recipe (schema linking, self-correction, self-consistency) component by component, with every difference tested by the paired McNemar test. Four findings stand out. (i) Generation matters more than raw size, and the recipe is family-robust: Qwen2.5-Coder dominates the older CodeLlama at matched size (39.1 vs 20.9 at 7B), but a modern non-Qwen model (Llama-3.3-70B, 49.2 on a matched serving) is competitive, so CodeLlama's weakness reflects its 2023 generation, not "non-Qwen = weak". (ii) Self-correction is a robust, near-free win, significant on all three families where there is room to improve. (iii) Schema linking does not help, and a stronger linker does not rescue it: a retrieval/embedding linker with 96.5% gold-table recall is statistically indistinguishable from no linking, ruling out the "weak lexical strawman" objection across three families. (iv) Self-consistency is poor value (+0.13 pp for ~5x tokens, not significant). We report real per-stage cost ($/1k queries) and release all code, predictions, and summaries; archived code and data: https://doi.org/10.5281/zenodo.20952794
Chinese Translation
无法将数据发送到云API的组织越来越关注：如果模型必须在本地运行并使用开放权重，文本到SQL的效果如何？哪些流行的准确性“配方”值得其计算成本？我们通过在BIRD开发集（n=1534，执行准确率）上进行诚实且完全可重复的基准测试来回答这个问题，评估三种开放模型家族在两个世代下的表现——Qwen2.5-Coder（7B/14B/32B）、CodeLlama-Instruct（7B/13B/34B）和Llama-3.x（8B，70B）——在一个匹配的协议下，逐步剔除模型无关的配方（模式链接、自我修正、自我一致性）中的各个组件，并通过配对的McNemar检验测试每个差异。四个发现尤为突出。（i）生成比原始规模更为重要，且该配方在家族间具有鲁棒性：在匹配规模下，Qwen2.5-Coder在性能上优于较旧的CodeLlama（7B时为39.1对20.9），但现代非Qwen模型（Llama-3.3-70B，在匹配服务下为49.2）也具竞争力，因此CodeLlama的弱点反映了其2023年的世代，而非“非Qwen=弱”。（ii）自我修正是一种鲁棒的、几乎无成本的提升，在所有三个家族中都有显著的改进空间。（iii）模式链接没有帮助，且更强的链接器并未挽救这一点：一个具有96.5%黄金表召回率的检索/嵌入链接器在统计上与没有链接无异，排除了在三个家族中“弱词汇稻草人”反对意见。（iv）自我一致性价值不高（约0.13个百分点，需约5倍的token，不显著）。我们报告了每个阶段的实际成本（$/1k查询），并发布所有代码、预测和摘要；归档代码和数据链接：https://doi.org/10.5281/zenodo.20952794
cs.CL / 69 / 2606.29734
Fast Numbers, Slow Language: Bridging Quantitative and Qualitative Earnings Signals
快速数字，缓慢语言：弥合定量与定性收益信号
Abstract
Earnings announcements release two types of information sequentially: quantitative surprise (numeric earnings-per-share (EPS)/revenue versus analyst estimate) arrives first in press releases and financial news, processed by algorithmic traders within minutes; qualitative language (management tone, guidance, question-and-answer (Q&A) credibility) arrives 30-90 min later in the earnings conference call transcript (ECT), requiring human interpretation overnight. Financial economists have studied quantitative surprise for 50 years; natural language processing (NLP) researchers have studied qualitative ECT signals for a decade. Despite studying the same event, the two communities used incompatible frameworks: different targets (return vs. volatility), trading setups (long top-decile and short bottom-decile vs. trade-all), and metrics (return spread between top and bottom 20% (Q5-Q1) vs. mean squared error (MSE)), making direct comparison and connection challenging. We bridge these communities with EarningsInOne, the first corpus aligning earnings news, ECTs, and intraday and next-day prices across SP 1500 (broad U.S. equity universe, 2022-2025). Applying unified trading and evaluation tools to both signal types, we confirm a clean speed separation, fast numbers, slow language: quantitative surprise peaks at announcement and is largely eliminated by the next market open; qualitative ECT sentiment peaks on the next trading day, real and tradeable, but hidden under prior transcript-based evaluation that optimised sign-agnostic volatility with pointwise MSE.
Chinese Translation
收益公告依次释放两种类型的信息：定量意外（每股收益（EPS）/收入与分析师预期的数值差异）首先在新闻稿和金融新闻中发布，算法交易者在几分钟内进行处理；而定性语言（管理层语气、指导、问答（Q&A）可信度）则在收益电话会议记录（ECT）中在30-90分钟后到达，需要人类在一夜之间进行解读。金融经济学家研究定量意外已有50年；自然语言处理（NLP）研究者则研究定性ECT信号已有十年。尽管研究的是同一事件，但这两个领域使用了不兼容的框架：不同的目标（收益与波动性）、交易设置（做多前10%与做空后10% vs. 全部交易）以及指标（前20%与后20%之间的收益差（Q5-Q1）与均方误差（MSE）），使得直接比较和联系变得困难。我们通过EarningsInOne弥合这两个领域，这是第一个对齐收益新闻、ECT和SP 1500（广泛的美国股票市场，2022-2025）内日内及次日价格的语料库。我们对这两种信号类型应用统一的交易和评估工具，确认了明显的速度分离：快速数字，缓慢语言：定量意外在公告时达到峰值，并在下一个市场开盘前基本消失；而定性ECT情绪在下一个交易日达到峰值，真实且可交易，但隐藏在之前基于记录的评估之下，该评估优化了与符号无关的波动性，并使用逐点均方误差（MSE）进行评估。
cs.CL / 70 / 2606.29750
Managing Map Cardinality in Automatic Disease Classification Mapping: Balancing Precision, Recall and Coverage
自动疾病分类映射中的映射基数管理：平衡精确度、召回率和覆盖率
Abstract
Automatic mapping between disease classification systems, such as the International Classification of Diseases (ICD), is a challenging yet essential task for integrating health data and conducting longitudinal data analysis. Existing embedding-based methods primarily focus on \emph{one-to-one} mappings, overlooking more complex \emph{one-to-many} scenarios. The threshold-based and top-K methods offer natural extensions; however, they involve inherent trade-offs between \emph{precision}, \emph{recall} and \emph{mapping coverage} -- the proportion of source codes with at least one mapping to a target code. To address this challenge, we introduce a novel method, which is inspired by the \emph{blocking-and-matching} pipeline commonly used in \emph{entity resolution}. In particular, we first generate a block of candidate matches (\emph{blocking}) and then employ a large language model (LLM) to identify all valid mappings within each block (\emph{matching}). Empirically, we show that the proposed method achieves higher precision with comparable recall and broader coverage across multiple ICD version pairs (ICD-9-CM$\leftrightarrow$ICD-10-CM and ICD-10-AM$\leftrightarrow$ICD-11). Our source code and dataset is available at: https://tinyurl.com/46kyn7wp.
Chinese Translation
疾病分类系统之间的自动映射，例如国际疾病分类（ICD），是一项具有挑战性但又至关重要的任务，旨在整合健康数据并进行纵向数据分析。现有的基于嵌入的方法主要集中于 extit{一对一}映射，忽视了更复杂的 extit{一对多}场景。基于阈值和前K个方法提供了自然的扩展；然而，它们在 extit{精确度}、 extit{召回率}和 extit{映射覆盖率}（至少有一个映射到目标代码的源代码比例）之间存在固有的权衡。为了解决这一挑战，我们提出了一种新方法，该方法受到 extit{实体解析}中常用的 extit{阻塞与匹配}流程的启发。具体而言，我们首先生成候选匹配的区块（ extit{阻塞}），然后利用大型语言模型（LLM）识别每个区块内的所有有效映射（ extit{匹配}）。通过实证研究，我们展示了所提出的方法在多个ICD版本对（ICD-9-CM$
ightarrow$ICD-10-CM和ICD-10-AM$
ightarrow$ICD-11）中实现了更高的精确度，同时保持了可比的召回率和更广泛的覆盖率。我们的源代码和数据集可在以下链接获取：https://tinyurl.com/46kyn7wp。
cs.CL / 71 / 2606.29792
Are Humans Evolved Instruction Followers? An Underlying Inductive Bias Enables Rapid Instructed Task Learning
人类是进化而来的指令追随者吗？一种潜在的归纳偏差促进了快速的指令任务学习
Abstract
Human adults can often perform a novel task correctly on the first attempt after only receiving verbal or written instructions. This rapid instructed task learning (RITL) is a hallmark of human cognitive flexibility, yet its mechanisms and parallels in artificial systems remain under-explored across disciplines. In this position paper, we argue that humans possess an evolved instruction-following bias -- an inductive bias shaped by evolution to interpret and execute linguistic instructions which critically enables fast generalization of behavior from language. This bias functions analogously to the way large language models (LLMs) leverage instruction tuning to achieve zero-shot task performance. We synthesize evidence from cognitive science, neuroscience, and machine learning research to support this hypothesis. While instruction-following in AI is currently achieved via specialized training protocols, we posit that in humans it arises as an innate cognitive architecture feature. We outline testable predictions and call for more interdisciplinary research to investigate Instruction-Following as a unifying mechanism enabling rapid task learning in both natural and artificial neural networks.
Chinese Translation
人类成年人在仅接受口头或书面指令后，通常能够在第一次尝试中正确地执行一项新任务。这种快速的指令任务学习（RITL）是人类认知灵活性的标志，但其机制及在人工系统中的平行性在各学科中仍未得到充分探讨。在这篇立场论文中，我们认为人类具备一种进化而来的指令追随偏差——一种由进化塑造的归纳偏差，使得人类能够解读和执行语言指令，从而关键性地促进了从语言中快速概括行为。这种偏差的功能类似于大型语言模型（LLMs）如何利用指令调优来实现零样本任务表现。我们综合了来自认知科学、神经科学和机器学习研究的证据来支持这一假设。尽管目前人工智能中的指令追随是通过专门的训练协议实现的，但我们认为在人类中，它作为一种先天的认知架构特征而产生。我们概述了可测试的预测，并呼吁更多跨学科的研究，以探讨指令追随作为一种统一机制，促进自然和人工神经网络中的快速任务学习。
cs.CL / 72 / 2606.29793
Fund2Persona: A Framework for Building and Refining Financial Advisor Personas from Fund Disclosure Data
Fund2Persona：基于基金披露数据构建和完善金融顾问角色的框架
Abstract
Demand for personalized financial advising is growing, but consistent advisor expertise is difficult to obtain, scale, and encode in LLM systems. Simple persona prompts rarely specify how a financial advisor should reason and often drift toward generic recommendations. We propose Fund2Persona, a framework that grounds financial-advisor personas in fund disclosures, holdings transitions, market context, and manager commentary, then refines them through an agentic actor--scorer--patcher loop. We evaluate the resulting personas on held-out holdings-transition reconstruction and manager-commentary alignment, where they better recover portfolio decisions and grounded manager interpretation than generic baselines. We further study two downstream diagnostics: market-scenario generation, where persona retrieval broadens plausible investment views beyond repeated generic rollouts, and advisory dialogues grounded in investor profiles, where matched personas give more specific and useful advice than a generic advisor. These results suggest that fund-data-grounded financial-advisor personas can make manager-specific investment expertise portable rather than merely changing an LLM's surface style.
Chinese Translation
个性化金融顾问的需求日益增长，但一致的顾问专业知识难以获得、扩展和在大型语言模型（LLM）系统中编码。简单的角色提示很少明确金融顾问应如何推理，往往偏向于通用建议。我们提出了Fund2Persona，一个将金融顾问角色基于基金披露、持仓变动、市场背景和经理评论进行构建的框架，并通过代理行为者-评分者-修补者循环对其进行完善。我们在持出持仓变动重构和经理评论对齐的评估中，发现生成的角色在恢复投资组合决策和基于实际情况的经理解读方面优于通用基线。此外，我们进一步研究了两个下游诊断：市场情景生成，其中角色检索拓宽了合理的投资视角，超越了重复的通用推广；以及基于投资者档案的咨询对话，其中匹配的角色提供的建议比通用顾问更具体和有用。这些结果表明，基于基金数据的金融顾问角色能够使经理特定的投资专业知识变得可移植，而不仅仅是改变LLM的表面风格。
cs.CL / 73 / 2606.29809
How Far Can You Get Without a GPU? A Systematic Benchmark of Lightweight Hallucination Detection Across Question Answering, Dialogue, and Summarisation
没有 GPU 能走多远？轻量级幻觉检测在问答、对话和摘要中的系统基准测试
Abstract
Hallucination detection has become a pressing requirement for trustworthy AI deployment at scale. The most accurate detection methods depend on GPU-intensive inference, proprietary API calls, or white-box access to the generating model. This puts them out of reach for resource-constrained researchers and practitioners. In this paper, we explore a practical alternative: how well can hallucination detection perform using only lightweight, CPU-feasible methods built on publicly available models? We systematically benchmark five such methods: ROUGE-L, semantic similarity, BERTScore, a Natural Language Inference (NLI) detector based on a FEVER-trained DeBERTa model, and a score-level ensemble of similarity and NLI. We evaluate them across all three tasks of the HaluEval benchmark: question answering (QA), dialogue, and summarisation. We calibrate each method on a held-out validation split and evaluate it on 2,000 test instances per task. We find that no single method dominates and performance is highly task-dependent. The ensemble performs best on QA (F1 = 0.792, AUC-ROC = 0.873), the NLI detector leads on dialogue (AUC-ROC = 0.713), and all five methods degrade to near-random performance on summarisation (AUC-ROC between 0.469 and 0.574). This task-dependence and the systematic failure on summarisation map the practical frontier of GPU-free hallucination detection. They give practical guidance for method selection under computational constraints. All experiments run on a standard laptop CPU using public models.
Chinese Translation
幻觉检测已成为大规模可信 AI 部署的迫切需求。最准确的检测方法依赖于 GPU 密集型推理、专有 API 调用或对生成模型的白盒访问。这使得资源有限的研究人员和从业者无法使用。在本文中，我们探索了一种实用的替代方案：仅使用基于公开可用模型的轻量级、可在 CPU 上运行的方法，幻觉检测的表现如何？我们系统地基准测试了五种此类方法：ROUGE-L、语义相似度、BERTScore、基于 FEVER 训练的 DeBERTa 模型的自然语言推理（NLI）检测器，以及相似度和 NLI 的得分级集成。我们在 HaluEval 基准的三个任务上对它们进行了评估：问答（QA）、对话和摘要。我们在保留的验证集上对每种方法进行了校准，并在每个任务上评估了 2000 个测试实例。我们发现没有单一方法占主导地位，性能高度依赖于任务。集成方法在问答任务上表现最佳（F1 = 0.792，AUC-ROC = 0.873），NLI 检测器在对话任务上领先（AUC-ROC = 0.713），而所有五种方法在摘要任务上的表现接近随机（AUC-ROC 在 0.469 到 0.574 之间）。这种任务依赖性和在摘要任务上的系统性失败描绘了无 GPU 幻觉检测的实际边界。它们为在计算限制下的方法选择提供了实用指导。所有实验均在标准笔记本电脑 CPU 上使用公共模型运行。
cs.CL / 74 / 2606.29815
SrDetection: A Self-Referential Framework for Data Leakage Detection in Code Large Language Models
SrDetection：一种自引用框架用于代码大型语言模型中的数据泄露检测
Abstract
Evaluating code large language models (Code LLMs) requires reliable detection of data leakage, where benchmark performance is artificially inflated by exposure to benchmark data during pre-training. Existing approaches either assume access to proprietary training corpora, rely on brittle heuristics such as timestamp filtering, or use external reference sets with manually tuned, non-generalizable thresholds. To address these limitations, we introduce \textbf{SrDetection}, a unified \textbf{s}elf-\textbf{r}eferential leakage detection framework for both gray-box (access to model logits) and black-box (access to model outputs) settings. SrDetection generates semantically equivalent variants of a benchmark sample and detects leakage by contrasting the model's behavior on the original versus its variants, flagging cases where the original is disproportionately easier for the model. We further design a controlled leakage detection testbed and evaluate SrDetection in this environment. Across different models and training stages, SrDetection improves average F1 by 21.52 points in the gray-box setting and 14.46 points in the black-box setting over strong baselines, demonstrating robust, threshold-independent leakage detection. Finally, a gray-box study of 15 widely used Code LLMs on four popular benchmarks reveals benchmark-specific leakage patterns beyond prior overlap-based analyses\footnote{\footnotesize Source code and data are available at https://github.com/SMinL/SrDetectionCode
Chinese Translation
评估代码大型语言模型（Code LLMs）需要可靠的数据泄露检测，其中基准性能因在预训练期间接触基准数据而被人为地夸大。现有方法要么假设可以访问专有训练语料，要么依赖于脆弱的启发式方法，如时间戳过滤，或者使用具有手动调整、不可泛化阈值的外部参考集。为了解决这些局限性，我们提出了 extbf{SrDetection}，一个统一的 extbf{s}elf- extbf{r}eferential泄露检测框架，适用于灰盒（访问模型logits）和黑盒（访问模型输出）设置。SrDetection生成基准样本的语义等价变体，并通过对比模型在原始样本与其变体上的行为来检测泄露，标记出原始样本对模型来说异常容易的情况。我们进一步设计了一个受控的泄露检测测试平台，并在该环境中评估SrDetection。在不同模型和训练阶段，SrDetection在灰盒设置中平均提升F1分数21.52分，在黑盒设置中提升14.46分，相较于强基线，展示了稳健的、独立于阈值的泄露检测。最后，对15个广泛使用的Code LLMs在四个流行基准上的灰盒研究揭示了超越先前基于重叠分析的基准特定泄露模式。
cs.CL / 75 / 2606.29824
Neural Procedural Memory: Empowering LLM Agents with Implicit Activation Steering
神经程序记忆：通过隐式激活引导赋能大型语言模型代理
Abstract
While Large Language Models (LLMs) excel as static solvers, transforming them into autonomous agents remains challenging. This transition requires continuous environmental interaction, yet current agents lack the necessary persistent procedural memory. Existing approaches predominantly employ Retrieval-Augmented Generation (RAG) to inject explicit textual guidelines into model contexts. However, relying solely on symbolic instructions can introduce a text-action disconnect, frequently failing to activate the internal representations necessary for correct task execution. To address this, the paper introduces Neural Procedural Memory (NPM), a training-free framework that represents agent memory through implicit activation steering rather than explicit instructions. By distilling procedural skills from historical contrastive experiences into steering vectors in the activation space, NPM directly activates the task-relevant neural mechanisms to guide task execution. Evaluations across four agent benchmarks show that NPM performs comparably to baselines using explicit textual instructions. Furthermore, the results show that combining implicit steering with explicit workflows provides complementary advantages, leading to more robust task execution. Representational analyses indicate that these steering vectors encode consistent task logic, forming organized structures within the activation space. These findings suggest that implicit activation steering provides a promising approach for managing agent memory.
Chinese Translation
尽管大型语言模型（LLMs）在静态求解方面表现出色，但将其转变为自主代理仍然面临挑战。这一转变需要持续的环境交互，而当前的代理缺乏必要的持久程序记忆。现有的方法主要采用检索增强生成（RAG）将显式文本指南注入模型上下文。然而，仅依赖符号指令可能会引入文本与行动之间的脱节，常常无法激活正确任务执行所需的内部表征。为了解决这个问题，本文提出了神经程序记忆（NPM），这是一个无训练的框架，通过隐式激活引导而非显式指令来表示代理记忆。通过将历史对比经验中的程序技能提炼为激活空间中的引导向量，NPM直接激活与任务相关的神经机制以指导任务执行。在四个代理基准测试中的评估表明，NPM的表现与使用显式文本指令的基线相当。此外，结果表明，将隐式引导与显式工作流程结合提供了互补优势，从而实现更强大的任务执行。表征分析表明，这些引导向量编码了一致的任务逻辑，在激活空间中形成了有组织的结构。这些发现表明，隐式激活引导为管理代理记忆提供了一种有前景的方法。
cs.CL / 76 / 2606.29836
Revealing the Technology Development of Natural Language Processing: A Scientific Entity-Centric Perspective
揭示自然语言处理技术发展的科学实体中心视角
Abstract
Most studies on technology development have been conducted from a thematic perspective, but the topics are coarse-grained and insufficient to accurately represent technology. The development of automatic entity recognition techniques makes it possible to extract technology-related entities on a large scale. Thus, we perform a more accurate analysis of technology development from an entity-centric perspective. To begin with, we extract technology-related entities such as methods, datasets, metrics, and tools in articles on Natural Language Processing (NLP), and we apply a semi-automatic approach to normalize the entities. Subsequently, we calculate the z-scores of entities based on their co-occurrence networks to measure their impact. We then analyze the development trends of new technologies in the NLP domain since the beginning of the 21st century. The findings of this paper include three aspects: Firstly, the continued increase in the average number of entities per paper implies a growing burden on researchers to acquire relevant technical background knowledge. However, the emergence of pre-trained language models has injected new vitality into the technological innovation of the NLP domain. Secondly, Methods dominate among the 179 high-impact entities. An analysis of the z-score trend about the top 10 entities reveals that pre-trained language models, exemplified by BERT and Transformer, have become mainstream in recent years. Unlike the trend of the other eight method entities, the impact of Wikipedia dataset and BLEU metric has continued to rise in the long term. Thirdly, in recent years, there has been a remarkable surge in popularity for new high-impact technologies than ever before, and their acceptance by researchers has accelerated at an unprecedented speed. Our study provides a new perspective on analyzing technology development in a specific domain.
Chinese Translation
大多数关于技术发展的研究都是从主题的角度进行的，但这些主题往往较为粗糙，无法准确代表技术。自动实体识别技术的发展使得大规模提取与技术相关的实体成为可能。因此，我们从实体中心的视角对技术发展进行了更为准确的分析。首先，我们提取了自然语言处理（Natural Language Processing, NLP）领域文章中的与技术相关的实体，如方法、数据集、指标和工具，并采用半自动化的方法对这些实体进行规范化。随后，我们基于实体的共现网络计算其 z-score，以衡量其影响力。接着，我们分析了21世纪以来NLP领域新技术的发展趋势。本文的研究结果包括三个方面：首先，平均每篇论文中的实体数量持续增加，这意味着研究者在获取相关技术背景知识方面的负担加重。然而，预训练语言模型的出现为NLP领域的技术创新注入了新的活力。其次，在179个高影响力实体中，方法（Methods）占据主导地位。对前10个实体的z-score趋势分析显示，以BERT和Transformer为代表的预训练语言模型近年来已成为主流。与其他八个方法实体的趋势不同，维基百科数据集和BLEU指标的影响力在长期内持续上升。第三，近年来，新兴高影响力技术的受欢迎程度显著上升，研究者对其的接受速度前所未有地加快。我们的研究为特定领域的技术发展分析提供了新的视角。
cs.CL / 77 / 2606.29844
MATCH: Modulating Attention via In-Context Retrieval for Long-Context Transformers
MATCH：通过上下文检索调节长上下文变换器的注意力
Abstract
The quadratic computational cost of traditional attention mechanisms poses a major bottleneck to the scalability and practical deployment of large language models (LLMs), particularly in long-context scenarios. To improve efficiency, existing approaches often enforce rigid structural constraints such as local attention windows. However, these strategies typically lead to substantial performance degradation on tasks requiring precise long-range recall. In this work, we propose MATCH, a scalable and efficient framework that augments sparsified attention mechanisms with dynamically integrated in-context information through an efficient retrieval system. Empirical results show that MATCH significantly improves the performance of sparse-attention models on both synthetic and real-world natural-language tasks. These findings highlight the versatility of MATCH as a general approach for enhancing in-context retrieval capabilities while maintaining the efficiency benefits of sparse attention architectures.
Chinese Translation
传统注意力机制的二次计算成本对大型语言模型（LLMs）的可扩展性和实际部署构成了主要瓶颈，尤其是在长上下文场景中。为了提高效率，现有方法通常强加严格的结构约束，例如局部注意力窗口。然而，这些策略通常会导致在需要精确长距离回忆的任务上性能显著下降。在本研究中，我们提出了MATCH，一个可扩展且高效的框架，通过高效的检索系统动态集成上下文信息，增强稀疏注意力机制。实证结果表明，MATCH显著提高了稀疏注意力模型在合成和真实世界自然语言任务上的性能。这些发现突显了MATCH作为一种通用方法的多样性，能够在保持稀疏注意力架构效率优势的同时增强上下文检索能力。
cs.CL / 78 / 2606.29858
Smooth Scaling Laws Hide Stepwise Token Learning
平滑的缩放规律掩盖了逐步的标记学习
Abstract
Language model loss follows remarkably regular scaling laws over model and data size, yet it remains unclear why the aggregate loss should exhibit a power-law form. Existing explanations often attribute this regularity to a heavy-tailed spectrum of pattern difficulty in natural language, but this view has not been directly validated at token-level granularity in large-scale real-data training. We present a token-level framework that decomposes scaling laws into localized learning events of individual contextualized tokens. By fitting token loss trajectories with sigmoids, we show that token learning is concentrated in localized transitions, giving rise to a learning-time spectrum that dominates the scaling-law shape. Across more than one hundred pre-training runs on large and diverse real-language corpora with modern LLM architectures, scaling up to 6B parameters and 300B training tokens, the measured learning-time spectrum quantitatively reconstructs the validation loss derivative along the training-step $T$, data-scale $D$, and model-scale $M$ axes. We further show that the same signal is actionable: by reshaping the training distribution according to when tokens become learnable, we alter the optimization trajectory and achieve 11\% faster validation-loss reduction. These results provide direct empirical evidence that scaling laws are governed primarily by the distribution of token-level learning times, and that this distribution can be used not only to explain scaling behavior but also to improve training performance.
Chinese Translation
语言模型的损失在模型和数据规模上遵循显著规律的缩放规律，但为何总损失呈现幂律形式仍不清楚。现有的解释通常将这种规律性归因于自然语言中模式难度的重尾谱，但这一观点尚未在大规模真实数据训练中以标记级别的粒度得到直接验证。我们提出了一种标记级框架，将缩放规律分解为个别上下文化标记的局部学习事件。通过用S型曲线拟合标记损失轨迹，我们展示了标记学习集中在局部转变中，从而形成主导缩放规律形状的学习时间谱。在超过一百次针对大型和多样化真实语言语料库的预训练运行中，使用现代大语言模型（LLM）架构，参数规模达到60亿，训练标记数量达到3000亿，测得的学习时间谱定量重建了沿训练步数 $T$、数据规模 $D$ 和模型规模 $M$ 轴的验证损失导数。我们进一步表明，相同的信号是可操作的：通过根据标记何时变得可学习来重塑训练分布，我们改变了优化轨迹，实现了11%的验证损失减少速度提升。这些结果提供了直接的实证证据，表明缩放规律主要受标记级学习时间分布的支配，并且这一分布不仅可以用来解释缩放行为，还可以用来提高训练性能。
cs.CL / 79 / 2606.29859
Exploring Motivations for Algorithm Mention in the Domain of Natural Language Processing: A Deep Learning Approach
探索自然语言处理领域中算法提及的动机：一种深度学习方法
Abstract
With the rise of data-intensive science, algorithms have become central to scientific research. In academic papers, algorithms are mentioned for different purposes, such as describing, using, comparing, or improving methods for specific research tasks. Identifying these purposes can reveal relationships among algorithms and help assess their roles and value. Taking natural language processing (NLP) as an example, this study proposes a sentence-level framework for identifying, analyzing, and tracing the evolution of motivations for mentioning algorithms. We first identify algorithm entities and algorithm-related sentences from full-text papers through manual annotation and machine learning. We then classify mention motivations using pretrained models and data augmentation, and analyze their distribution and temporal evolution. The results show that deep learning models trained with augmented data outperform traditional machine learning models in motivation classification. In NLP papers, more than half of algorithm-related sentences express direct use, whereas improvement is the least frequent motivation. The diversity of motivations has increased over time. For specific algorithm categories, grammar-based algorithms are more often mentioned for description, while machine learning algorithms are more often mentioned for use. Over time, use motivations have gradually replaced description motivations across different algorithms, and the number of motivation types associated with individual algorithms has declined significantly. This study reveals how authors mention algorithm entities in academic writing and provides a basis for future research on algorithm relationship identification and algorithm impact evaluation.
Chinese Translation
随着数据密集型科学的兴起，算法已成为科学研究的核心。在学术论文中，算法的提及有不同的目的，例如描述、使用、比较或改进特定研究任务的方法。识别这些目的可以揭示算法之间的关系，并帮助评估它们的角色和价值。以自然语言处理（NLP）为例，本研究提出了一种基于句子的框架，用于识别、分析和追踪提及算法的动机的演变。我们首先通过手动标注和机器学习从全文论文中识别算法实体和与算法相关的句子。然后，我们使用预训练模型和数据增强对提及动机进行分类，并分析其分布和时间演变。结果表明，使用增强数据训练的深度学习模型在动机分类上优于传统的机器学习模型。在NLP论文中，超过一半的与算法相关的句子表达了直接使用，而改进则是最不常见的动机。动机的多样性随着时间的推移而增加。对于特定的算法类别，基于语法的算法更常用于描述，而机器学习算法则更常用于使用。随着时间的推移，在不同算法中，使用动机逐渐取代了描述动机，且与单个算法相关的动机类型数量显著减少。本研究揭示了作者在学术写作中如何提及算法实体，并为未来的算法关系识别和算法影响评估研究提供了基础。
cs.CL / 80 / 2606.29863
KbSD: Knowledge Boundary aware Self-Distillation for Behavioral Calibration in Agentic Search
KbSD：知识边界感知自蒸馏用于代理搜索中的行为校准
Abstract
Agentic search equips large language models with dynamic retrieval abilities, but existing reinforcement learning methods remain limited by reward sparsity in knowledge boundary calibration -- deciding when to trust parametric memory, when to rely on retrieved evidence, and when to abstain. Binary rewards can penalize undesirable outcomes, but provide little guidance on the reasoning process required to make calibrated decisions across different knowledge states. To address this, we propose KbSD (Knowledge boundary Self-Distillation), a framework that tackles this limitation through dense token-level supervision, outcome-level sparse rewards, and quadrant-adaptive optimization. KbSD constructs a hint-augmented teacher, architecturally identical to the student, that receives explicit knowledge boundary signals -- including parametric certainty, retrieval quality, and ground-truth answers -- to generate calibrated reasoning demonstrations. This information-asymmetric self-distillation enables dense supervision without requiring a larger external model. To further account for the heterogeneous reasoning distributions across knowledge states, we introduce a quadrant-adaptive distillation objective: reverse KL for concentrated integration, forward KL for diverse refusal, and Pareto-optimal bidirectional KL for asymmetric quadrants requiring both precision and coverage. Experiments on multiple benchmarks show that KbSD consistently improves both task accuracy and hallucination mitigation over strong baselines, with the largest gains appearing in the challenging quadrants where sparse rewards are least informative.
Chinese Translation
代理搜索为大型语言模型赋予了动态检索能力，但现有的强化学习方法在知识边界校准方面仍然受到奖励稀疏性的限制——即决定何时信任参数记忆、何时依赖检索证据以及何时应当放弃。二元奖励可以惩罚不理想的结果，但对在不同知识状态下做出校准决策所需的推理过程提供的指导有限。为了解决这一问题，我们提出了 KbSD（知识边界自蒸馏），一个通过密集的标记级监督、结果级稀疏奖励和象限自适应优化来应对这一限制的框架。KbSD 构建了一个提示增强的教师模型，其架构与学生模型相同，接收明确的知识边界信号——包括参数确定性、检索质量和真实答案——以生成校准的推理示范。这种信息不对称的自蒸馏使得在不需要更大外部模型的情况下实现密集监督。为了进一步考虑不同知识状态下的异质推理分布，我们引入了象限自适应蒸馏目标：对于集中的整合使用反向 KL，对于多样的拒绝使用正向 KL，对于需要精确性和覆盖性的非对称象限使用帕累托最优的双向 KL。多个基准测试的实验表明，KbSD 在任务准确性和幻觉缓解方面始终优于强基线，尤其是在稀疏奖励信息最少的挑战性象限中获得了最大的提升。
cs.CL / 81 / 2606.29869
ARKD: Adaptive Reinforcement Learning-Guided Bidirectional KL Divergence Distillation for Text Generation
ARKD：自适应强化学习引导的双向KL散度蒸馏用于文本生成
Abstract
Knowledge distillation (KD) is a key technique for compressing Large Language Models (LLMs), yet methods relying on a single KL objective often fail to balance primary distribution fitting with long-tail probability modeling, limiting both generation quality and generalization. To address this, we analyze the complementary roles of forward and reverse KL divergence (FKL/RKL) in distribution alignment from theoretical and empirical perspectives. We then propose a reinforcement-learning-based adaptive KL-weighted distillation framework, in which a policy network dynamically assigns weights to FKL and RKL based on teacher-student distributional characteristics, guided by immediate reward signals to achieve dual alignment on principal and long-tail modes. Extensive experiments demonstrate consistent improvements across Rouge-L and BertScore metrics, surpassing greedy heuristics by 0.4-0.6 points and outperforming other baseline methods on diverse benchmarks.
Chinese Translation
知识蒸馏（KD）是压缩大型语言模型（LLMs）的关键技术，但依赖单一KL目标的方法往往无法平衡主要分布拟合与长尾概率建模，限制了生成质量和泛化能力。为了解决这一问题，我们从理论和实证的角度分析了正向和反向KL散度（FKL/RKL）在分布对齐中的互补作用。随后，我们提出了一种基于强化学习的自适应KL加权蒸馏框架，其中策略网络根据教师-学生分布特征动态分配FKL和RKL的权重，并通过即时奖励信号引导，以实现对主要模式和长尾模式的双重对齐。大量实验表明，在Rouge-L和BertScore指标上均有一致的提升，超越贪婪启发式方法0.4-0.6分，并在多样化基准测试中优于其他基线方法。
cs.CL / 82 / 2606.29876
Clinical Reasoning Graphs: Structured Evaluation of LLM Diagnostic Reasoning Reveals Competence Without Consistency
临床推理图：结构化评估大型语言模型的诊断推理揭示了能力与一致性的缺失
Abstract
Modern large language models (LLMs) reach 60-70% diagnostic accuracy on complex clinical case benchmarks, but accuracy alone cannot distinguish stable clinically-grounded reasoning from pattern matching. We introduce clinical reasoning graphs, structured graph representations extracted from free-text LLM diagnostic traces using a domain-grounded ontology with 5 node types and 7 edge types. We apply this pipeline to 750 traces from five LLMs across 50 New England Journal of Medicine Clinicopathological Conference cases and three prompt conditions, and test whether diagnostic traces show stable structured reasoning patterns, or diagnostic schemas, for clinically similar cases. We operationalize this as higher graph similarity among clinically similar cases than among clinically dissimilar ones. Across 15 model-condition comparisons, within-cluster and between-cluster composite similarity are nearly equal, and no comparison survives multiple-testing correction; a component-level analysis finds any residual content signal far below schema scale. Graph similarity is also nearly identical for pairs of models that are both correct (0.488) and both incorrect (0.484), suggesting that graph structure captures a dimension not reflected in diagnostic accuracy. Structured reflection prompting increases explicit discriminating-feature analysis within traces (+33%) but does not increase cross-case consistency. These results show diagnostic competence without schema-scale reasoning consistency, and indicate that final-answer accuracy should be complemented by process-level evaluation. We release the ontology, extraction pipeline, validation protocol, and the extracted reasoning graphs and similarity artifacts as resources for structured evaluation of LLM clinical reasoning.
Chinese Translation
现代大型语言模型（LLMs）在复杂临床案例基准测试中达到60-70%的诊断准确率，但仅凭准确率无法区分稳定的临床基础推理与模式匹配。我们引入了临床推理图，这是一种从自由文本LLM诊断轨迹中提取的结构化图形表示，使用了一个包含5种节点类型和7种边类型的领域基础本体。我们将该流程应用于来自五个LLM的750条轨迹，这些轨迹涵盖了50个《新英格兰医学杂志》临床病理会议案例和三种提示条件，并测试诊断轨迹是否显示出针对临床相似案例的稳定结构化推理模式或诊断模式。我们将其操作化为临床相似案例之间的图形相似性高于临床不相似案例之间的相似性。在15个模型-条件比较中，聚类内和聚类间的复合相似性几乎相等，且没有比较在多重检验校正后存活；组件级分析发现任何残余内容信号远低于模式规模。对于同时正确（0.488）和同时错误（0.484）的模型对，图形相似性也几乎相同，这表明图形结构捕捉了一个在诊断准确性中未反映的维度。结构化反思提示增加了轨迹中显性区分特征分析（+33%），但并未提高跨案例的一致性。这些结果表明诊断能力缺乏模式规模的推理一致性，并指出最终答案的准确性应补充过程级评估。我们发布了本体、提取流程、验证协议以及提取的推理图和相似性文物，作为结构化评估LLM临床推理的资源。
cs.CL / 83 / 2606.29904
Timesteps of Mamba Align with Human Reading Times
Mamba的时间步与人类阅读时间一致
Abstract
This study demonstrates an alignment of per-word processing time in a popular state-space language model Mamba and human readers. In Mamba, the recurrent state transition at each layer conceptually takes some duration of time, the discretization timestep $\Delta_t$, determined dynamically in response to the input. Using a naturalistic reading dataset, we show that the per-word timestep from Mamba is a significant predictor of human reading times, and remains significant even when known predictors such as GPT-2 surprisal are controlled for. We further suggest, through formal analysis of Mamba's architecture and internal dynamics, that Mamba can serve as a new, valuable lens to look at human real-time language processing with ever-updated memory, because it allows us to look at how each module (layer) weighs short- and long-term information retention, and how noise may interact with dynamic, continuous memory representation. Code is available online.
Chinese Translation
本研究展示了流行的状态空间语言模型Mamba中的逐词处理时间与人类读者之间的一致性。在Mamba中，每一层的递归状态转移在概念上需要一定的时间持续，离散化时间步$ _t$是根据输入动态确定的。通过使用自然阅读数据集，我们表明Mamba的逐词时间步是人类阅读时间的重要预测因子，即使在控制了已知预测因子如GPT-2惊讶度的情况下，这一结果依然显著。我们进一步通过对Mamba架构和内部动态的形式分析，建议Mamba可以作为一个新的、有价值的视角来观察人类实时语言处理中的不断更新的记忆，因为它使我们能够观察每个模块（层）如何权衡短期和长期信息的保留，以及噪声如何与动态、连续的记忆表示相互作用。代码可在线获取。
cs.CL / 84 / 2606.29914
MemDelta: Controlled Baselines and Hidden Confounds in Agent Memory Evaluation
MemDelta：代理记忆评估中的受控基线与隐藏混淆因素
Abstract
Agent memory systems are increasingly evaluated against RAG and full-context baselines, but reported gains often mix changes in the memory method with changes in the language model, embedding model, or retrieval pipeline, making it unclear what is actually being measured. We present MemDelta, a controlled evaluation protocol that varies one component at a time on LongMemEval-S (500 questions, 50+ sessions, three model families). Four findings emerge: (1) verbatim RAG matches full-context GPT-4o-mini (47.2% vs. 49.8%, p = 0.34), but the ranking reverses across models: Gemini gains +14pp from full context, while Sonnet gains +31pp from RAG, partly because it refuses 63% of full-context queries; (2) swapping only the embedding model in an identical pipeline shifts accuracy by +6.2pp at n = 500 (p = 0.004), and Mem0 beats MiniLM-RAG by +11pp but loses to cloud-RAG by 1.2pp, so one variable flips the conclusion; (3) agent self-memory (42%) underperforms basic retrieval (47%); (4) on 2 of 6 question types (n = 88), Mem0 matches cloud RAG (72.7% vs. 73.9%, p = 1.0) at 50x the cost, suggesting narrow rather than general gains. We recommend memory evaluations fix embedding models across comparisons, stratify by model family, and report write-path cost before attributing gains to architecture.
Chinese Translation
代理记忆系统越来越多地与RAG和全上下文基线进行评估，但报告的增益往往将记忆方法的变化与语言模型、嵌入模型或检索管道的变化混合在一起，使得实际测量的内容不明确。我们提出了MemDelta，一种控制评估协议，逐一变化LongMemEval-S（500个问题，50多个会话，三种模型系列）中的一个组件。得出了四个发现：（1）逐字RAG与全上下文GPT-4o-mini匹配（47.2%对49.8%，p = 0.34），但模型间的排名发生逆转：Gemini从全上下文中获得+14个百分点，而Sonnet从RAG中获得+31个百分点，部分原因是它拒绝了63%的全上下文查询；（2）在相同管道中仅更换嵌入模型，准确率提升+6.2个百分点（n = 500，p = 0.004），Mem0比MiniLM-RAG高出+11个百分点，但比cloud-RAG低1.2个百分点，因此一个变量改变了结论；（3）代理自我记忆（42%）表现不及基本检索（47%）；（4）在6种问题类型中的2种（n = 88）上，Mem0与cloud RAG匹配（72.7%对73.9%，p = 1.0），成本是其50倍，表明增益是狭窄的而非普遍的。我们建议在比较中固定嵌入模型，按模型系列分层，并在将增益归因于架构之前报告写路径成本。
cs.CL / 85 / 2606.29920
Can LLM-as-a-Judge Reliably Verify Rubrics in Agentic Scenarios?
LLM作为评审者在代理场景中能否可靠地验证评分标准？
Abstract
Rubric-based scoring has become a widely used paradigm in model evaluation, typically with LLM-as-a-Judge (LaaJ) for rubric scoring. However, the reliability of LaaJ for rubric scoring remains underexplored. This concern is especially pronounced in agentic scenarios, where long, complex outputs further challenge reliable scoring. To address this, we conduct a systematic meta-evaluation of LaaJ reliability for rubric verification. We introduce RuVerBench, the first benchmark for assessing LaaJ reliability in rubric verification for agentic scenarios. RuVerBench covers two prevalent agentic domains, deep research and agentic coding, with 2,458 instances, each containing a model-generated output, a rubric, and a human-annotated label indicating whether the output satisfies the rubric. Using RuVerBench, we evaluate numerous frontier LLMs and find that even the most advanced models achieve strong performance but still exhibit substantial noise. We further analyze the impact of key LaaJ strategies, including prompt design, batching, and majority voting, on rubric verification. We find that weaker models are more sensitive to prompt variations, batched verification presents a trade-off between accuracy and efficiency, and majority voting yields effective but diminishing returns. We have released our dataset and code to facilitate future research: https://github.com/THU-KEG/RuVerBench.
Chinese Translation
基于评分标准的评分已成为模型评估中广泛使用的范式，通常采用LLM作为评审者（LaaJ）进行评分。然而，LaaJ在评分标准验证中的可靠性仍然未得到充分探讨。这一问题在代理场景中尤为突出，因为长而复杂的输出进一步挑战了可靠评分。为了解决这一问题，我们对LaaJ在评分标准验证中的可靠性进行了系统的元评估。我们引入了RuVerBench，这是第一个用于评估LaaJ在代理场景中评分标准验证可靠性的基准。RuVerBench涵盖了两个常见的代理领域：深度研究和代理编码，共包含2,458个实例，每个实例包含一个模型生成的输出、一个评分标准以及一个人类标注的标签，指示该输出是否满足评分标准。利用RuVerBench，我们评估了众多前沿LLM，发现即使是最先进的模型也能取得良好的表现，但仍然存在显著的噪声。我们进一步分析了关键LaaJ策略的影响，包括提示设计、批处理和多数投票对评分标准验证的影响。我们发现，较弱的模型对提示变化更为敏感，批量验证在准确性和效率之间存在权衡，而多数投票则产生有效但递减的收益。我们已发布我们的数据集和代码，以促进未来的研究：https://github.com/THU-KEG/RuVerBench。
cs.CL / 86 / 2606.29933
Towards Physical Intuitions for Alignment Dynamics: A Case Study With Randomness Crystallization
朝向对齐动态的物理直觉：以随机性结晶为案例研究
Abstract
The alignment of language models is typically studied through the lens of capability benchmarks, but the dynamics of how models change during post-training remain poorly understood. We argue that the physical sciences, and thermodynamic phase-transition theory in particular, offer a principled and underexplored vocabulary for reasoning about these dynamics. As a case study, we instantiate this position through the lens of material Crystallization, which is a well-studied thermodynamic phase transition. For tasks like random number generation, this breaks into 3 phases: (1) the high entropy liquid phase in the pretrained model, with many distinct sampling distributions promptable from the model; (2) the nucleation phase caused by supervised finetuning, in which behavior collapses onto a single seed distribution present in the pretrained LLM; and (3) a settling phase in which reinforcement learning techniques redistribute probability of the collapsed distribution, but largely keep it concentrated on the same options as the seed distribution. We propose intuitive metrics to verify the transitions between these phases, and validate the idea across a range of random tasks. Crystallization is one instance of a broader class of physical frameworks we believe alignment research should import to answer questions about where alignment-induced structure comes from, why it converges where it does, and what it fundamentally cannot change.
Chinese Translation
语言模型的对齐通常通过能力基准进行研究，但模型在后训练过程中如何变化的动态仍然不甚了解。我们认为，物理科学，特别是热力学相变理论，提供了一种原则性且未被充分探索的词汇，用于推理这些动态。作为案例研究，我们通过材料结晶的视角来具体化这一观点，结晶是一个研究得较为深入的热力学相变。对于随机数生成等任务，这一过程分为三个阶段：(1) 预训练模型中的高熵液相，模型可以从中生成许多不同的采样分布；(2) 由监督微调引起的成核阶段，此时行为集中于预训练大型语言模型（LLM）中存在的单一种子分布；(3) 一种稳定阶段，在此阶段，强化学习技术重新分配了集中分布的概率，但基本上仍然保持在与种子分布相同的选项上。我们提出了直观的指标来验证这些阶段之间的转变，并在一系列随机任务中验证了这一观点。结晶是我们认为对齐研究应引入的更广泛物理框架中的一个实例，以回答关于对齐引起的结构来源、为何在特定位置收敛以及其根本无法改变的内容的问题。
cs.CL / 87 / 2606.29938
LatentRevise: Learning from Zero-Hit Reasoning
LatentRevise：从零命中推理中学习
Abstract
Reinforcement learning with verifiable rewards (RLVR) is bottlenecked by hard prompts on which correct trajectories have low probability, so sampling misses them within a practical budget and leaves the policy update with little useful signal. We frame such zero-hit prompts as RLVR's sampling frontier, where new reasoning behavior is most valuable yet least likely to be sampled. Importantly, failed rollouts can be informative: they expose where the model's reasoning went wrong. We introduce LatentRevise, a first-order latent revision method that recovers training signal for this zero-hit regime. Given a failed rollout and the gold answer as an anchor, LatentRevise optimizes the input embeddings of its reasoning prefix under two complementary gradients, moving the prefix away from the failed continuation and toward the gold answer. The optimization is constrained to the convex hull of the model's vocabulary embeddings, so each update moves the latent toward a real token embedding rather than an arbitrary feature direction. We find that continuations from the revised prefix lengthen, exhibit self-reflection, and reach correct answers missed by the original rollouts. Used as training data, these trajectories improve SFT and RLVR on math benchmarks over standard baselines.
Chinese Translation
可验证奖励的强化学习（RLVR）受到难以处理的提示的限制，这些提示的正确轨迹概率较低，因此在实际预算内采样时会错过这些轨迹，从而使得策略更新几乎没有有用的信号。我们将这种零命中提示框架化为RLVR的采样前沿，在这里新的推理行为最有价值但最不可能被采样。重要的是，失败的回滚可以提供信息：它们揭示了模型推理出错的地方。我们引入了LatentRevise，一种一阶潜在修正方法，用于恢复这一零命中状态下的训练信号。给定一个失败的回滚和作为锚点的金标准答案，LatentRevise在两个互补梯度下优化其推理前缀的输入嵌入，使前缀远离失败的延续并朝向金标准答案。优化被限制在模型词汇嵌入的凸包内，因此每次更新都将潜在向量移动到真实的标记嵌入，而不是任意的特征方向。我们发现，从修正后的前缀生成的延续更长，表现出自我反思，并达到了原始回滚未能捕捉到的正确答案。作为训练数据使用时，这些轨迹在数学基准测试上改善了SFT和RLVR的表现，相较于标准基线。
cs.CL / 88 / 2606.29960
IHDec: Divergence-Steered Contrastive Decoding for Securing Multi-Turn Instruction Hierarchies
IHDec：用于保护多轮指令层次的偏差引导对比解码
Abstract
Large Language Models (LLMs) often fail to maintain instruction hierarchies (IH) when processing multi-source inputs with varying role-level priorities, paradoxically adhering to lower-priority directives during conflicts. While existing defenses mitigate this issue, they are largely restricted to single-turn scenarios and require expensive fine-tuning. In this paper, we formalize this failure mode in multi-turn contexts via a Jensen-Shannon Divergence (JSD) framework, uncovering a pervasive role-influence inversion phenomenon where subordinate inputs override superior roles. To rectify this without training, we propose IHDec (Instruction Hierarchy-steered Decoding). IHDec leverages JSD to automatically detect token-level hierarchy violations and dynamically executes contrastive decoding to suppress misaligned subordinate roles. Extensive evaluations demonstrate that IHDec outperforms training-based baselines in multi-turn conflicts while fully preserving general response quality. Furthermore, IHDec strengthens safety against adversarial prompt injections and exhibits a robust scaling synergy with larger models. The Code is available at https://github.com/nxcolelxu/IHDec.git
Chinese Translation
大型语言模型（LLMs）在处理具有不同角色优先级的多源输入时，常常无法维持指令层次（IH），在冲突中反而遵循低优先级指令。尽管现有的防御措施在一定程度上缓解了这一问题，但它们主要限于单轮场景，并且需要昂贵的微调。在本文中，我们通过詹森-香农散度（JSD）框架形式化了这一多轮上下文中的失败模式，揭示了一种普遍存在的角色影响反转现象，即下级输入覆盖上级角色。为了在不进行训练的情况下纠正这一问题，我们提出了IHDec（指令层次引导解码）。IHDec利用JSD自动检测令牌级别的层次违规，并动态执行对比解码以抑制不对齐的下级角色。广泛的评估表明，IHDec在多轮冲突中优于基于训练的基线，同时完全保持了整体响应质量。此外，IHDec增强了对抗性提示注入的安全性，并与更大模型展现出强大的扩展协同效应。代码可在 https://github.com/nxcolelxu/IHDec.git 获取。
cs.CL / 89 / 2606.29985
Are We Measuring Strategy or Phrasing? The Gap Between Surface- and Approach-Level Diversity in LLM Math Reasoning
我们是在测量策略还是措辞？LLM数学推理中表面层与方法层多样性之间的差距
Abstract
Diversity in LLM mathematical reasoning is critical for exploration, but common diversity metrics mostly capture surface-level variation rather than differences in how a problem is solved. We address this gap by introducing approach-level diversity: variation in strategies across correct solutions to the same problem. Using a human-calibrated LLM judge framework, we show that prior diversity measures are unreliable proxies for approach-level diversity, and this mismatch carries over to diversity-aware RLVR, where target metrics are preserved while approach-level diversity declines. Investigating when approach-level diversity helps and whether it can be directly induced, we find that approach-diverse candidate sets improve test-time scaling. However, optimizing an LLM judge diversity reward during training causes the policy to exploit judge-specific preferences rather than broaden its approaches, leaving direct optimization of approach-level diversity as an open problem. Together, our work introduces the notion of approach-level diversity and uncovers a systematic divergence between surface- and approach-level signals, marking a step toward LLMs that reason in genuinely diverse, human-like ways.
Chinese Translation
LLM数学推理中的多样性对于探索至关重要，但常见的多样性度量主要捕捉表面层的变异，而非解决问题的方法差异。我们通过引入方法层多样性来填补这一空白：即对同一问题的正确解答中策略的变异。利用人类校准的LLM评判框架，我们表明，先前的多样性度量并不是方法层多样性的可靠代理，这种不匹配在关注多样性的RLVR中也有所体现，在这种情况下，目标度量得以保留，而方法层多样性却下降。我们调查了方法层多样性何时有助于推理，以及是否可以直接诱导，发现方法多样的候选集在测试时扩展性上有所提升。然而，在训练过程中优化LLM评判的多样性奖励会导致策略利用评判特定的偏好，而不是拓宽其方法，使得直接优化方法层多样性仍然是一个未解决的问题。总的来说，我们的研究引入了方法层多样性的概念，并揭示了表面层与方法层信号之间的系统性分歧，标志着朝着使LLM以真正多样化、类人方式进行推理迈出了一步。
cs.CL / 90 / 2606.30005
LLM Agents Are Latent Context Managers: Eliciting Self-Managed Context via a Proprioceptive Dashboard
大型语言模型代理是潜在的上下文管理者：通过自我管理仪表板引出自我管理的上下文
Abstract
Long-horizon tool agents are bottlenecked by how their context grows toward the limits of the context window. Recent systems make context management agent- or system-controlled, but they either learn a compression policy that discards evidence or manage context in a layer the agent never sees. We argue both leave a more basic gap unaddressed. Frontier language models are proprioceptively blind to their own context. From the prompt alone they cannot see how large, how old, or how used each block is, the signals a keep-or-drop decision needs. We hypothesize that competent context management is already latent in capable models, and that what is missing is not a learned policy but an interface exposing this state. We introduce VISTA (Visible Internal State for Tool Agents), a training-free, model-agnostic layer that represents working memory as typed, addressable blocks, surfaces a runtime dashboard of per-block token usage, recency, and access history, and archives blocks as recoverable full-fidelity payloads. On LOCA-Bench, BrowseComp-Plus, and GAIA, the same untrained interface transfers across million-, 100K-, and 10K-scale trajectories. On LOCA-Bench it improves four backbones and lifts Gemini-3-Flash from 22.7 to 50.7%. The lift grows with context pressure and transfers across backbones. Ablations further confirm that the dashboard matters beyond archive and recovery tools.
Chinese Translation
长时间跨度的工具代理受到上下文窗口限制的影响。近期的系统将上下文管理控制在代理或系统层面，但它们要么学习一种丢弃证据的压缩策略，要么在代理看不见的层面管理上下文。我们认为这两者都未能解决一个更基本的问题。前沿语言模型对自身上下文是自我感知盲的。仅凭提示，它们无法判断每个块的大小、年龄或使用情况，这些都是做出保留或丢弃决策所需的信号。我们假设，能够的上下文管理已经潜在地存在于有能力的模型中，而缺失的不是学习到的策略，而是一个能够暴露这种状态的接口。我们引入了 VISTA（工具代理的可见内部状态），这是一个无训练、模型无关的层，将工作记忆表示为可类型化、可寻址的块，呈现每个块的令牌使用情况、最近性和访问历史的运行时仪表板，并将块存档为可恢复的全保真负载。在 LOCA-Bench、BrowseComp-Plus 和 GAIA 上，相同的未训练接口在百万级、10万级和1万级轨迹中均可转移。在 LOCA-Bench 上，它改善了四个基础模型，并将 Gemini-3-Flash 的性能从 22.7% 提升至 50.7%。随着上下文压力的增加，提升效果也随之增强，并在不同基础模型之间转移。消融实验进一步确认了仪表板在归档和恢复工具之外的重要性。
cs.CL / 91 / 2606.30009
Node-to-Neighborhood Semantic Consistency: Text-Topology Alignment for TAGs Anomaly Detection
节点与邻域语义一致性：文本-拓扑对齐用于TAGs异常检测
Abstract
Graph anomaly detection (GAD) on text-attributed graphs (TAGs) is vital for applications such as fraud detection and academic integrity verification. Existing approaches generally fall into two paradigms. GNN-based methods effectively capture structural patterns but struggle to capture fine-grained textual semantics. Methods integrating LLMs with graphs improve semantic understanding yet fail to fully comprehend topological relationships among neighboring nodes. Moreover, both paradigms overlook the correspondence between textual semantics and graph topological relationships, limiting their ability to identify nodes whose semantics are inconsistent with their neighborhoods. In this paper, we formalize TAG anomaly detection as a node-to-neighborhood semantic consistency problem, where anomalies may arise from either textual semantic mismatch or topological deviation between a node and its neighbors. We propose N2NSC (Node-to-Neighborhood Semantic Consistency), a framework that captures the correspondence between graph topology and textual semantics through two complementary fusion paths. The two pathways work synergistically, enabling the LLM to fully leverage both textual and structural neighborhood information for anomaly detection. Extensive experiments across eight datasets demonstrate that N2NSC consistently outperforms current state-of-the-art methods.
Chinese Translation
在文本属性图（TAGs）上的图异常检测（GAD）对于欺诈检测和学术诚信验证等应用至关重要。现有方法通常分为两种范式。基于图神经网络（GNN）的方法有效捕捉结构模式，但难以捕捉细粒度的文本语义。将大型语言模型（LLMs）与图结合的方法提高了语义理解，但未能充分理解邻近节点之间的拓扑关系。此外，这两种范式都忽视了文本语义与图拓扑关系之间的对应关系，限制了它们识别语义与邻域不一致节点的能力。本文将TAG异常检测形式化为节点与邻域语义一致性问题，其中异常可能源于节点与其邻居之间的文本语义不匹配或拓扑偏差。我们提出了N2NSC（节点与邻域语义一致性），这是一个通过两条互补的融合路径捕捉图拓扑与文本语义之间对应关系的框架。这两条路径协同工作，使得LLM能够充分利用文本和结构邻域信息进行异常检测。在八个数据集上的大量实验表明，N2NSC始终优于当前最先进的方法。
cs.CL / 92 / 2606.30015
Parametric Skills
参数化技能
Abstract
Since intelligence fundamentally relies on efficient skill acquisition (Chollet, 2019), the ability to leverage skills is critical. For LLMs, skills, manually authored or extracted from task trajectories, are textual recipes encoding mature problem-solving experience and are critical to agentic capabilities. Despite widespread deployment, their utility is limited by the model's ability to comprehend and follow skill instructions, especially under complex and long-context scenarios, where key instructions are difficult to locate and adhere to. To address this limitation, we propose ParametricSkills, a framework that can convert free-form textual skills into parameters at test time, enabling context-free skill exploitation. Specifically, we first construct a large-scale, high-quality skill library, and synthesize single-turn and multi-turn skill exploitation trajectories built around these skills with OpenCode. Using these data, we then train a hypernetwork that parameterizes both the skill content and the test-time exploitation methodology by receiving textual skills and converting them into LoRA adapters. Experimental results on six complex software engineering (SWE) subtasks demonstrate that, the proposed ParametricSkills averagely outperforms in-context learning by 6.44 points as judged by DeepSeek-V4-Flash, while also achieving significantly higher BERT Score and F1 score, confirming its effectiveness. Beyond performance, we further find that parametric skills, being inherently accumulative, offer a preliminary yet promising avenue toward test-time continual learning.
Chinese Translation
由于智能在根本上依赖于高效的技能获取（Chollet, 2019），因此利用技能的能力至关重要。对于大型语言模型（LLMs）而言，技能是手动编写或从任务轨迹中提取的文本配方，编码了成熟的问题解决经验，并对自主能力至关重要。尽管广泛部署，但其效用受到模型理解和遵循技能指令能力的限制，尤其是在复杂和长上下文场景中，关键指令难以定位和遵循。为了解决这一限制，我们提出了参数化技能（ParametricSkills）框架，该框架能够在测试时将自由形式的文本技能转换为参数，从而实现无上下文的技能利用。具体而言，我们首先构建了一个大规模高质量的技能库，并利用OpenCode合成围绕这些技能构建的单轮和多轮技能利用轨迹。利用这些数据，我们训练了一个超网络，该网络通过接收文本技能并将其转换为LoRA适配器，来参数化技能内容和测试时的利用方法。针对六个复杂的软件工程（SWE）子任务的实验结果表明，所提出的参数化技能在DeepSeek-V4-Flash的评判下，平均超越了上下文学习6.44分，同时在BERT Score和F1分数上也显著更高，确认了其有效性。除了性能之外，我们进一步发现，参数化技能本质上是累积的，为测试时的持续学习提供了一条初步但有前景的途径。
cs.CL / 93 / 2606.30062
Little Brains, Big Feats: Exploring Compact Language Models
小模型，大成就：探索紧凑型语言模型
Abstract
While large language models have been dominating the research landscape recently, small language models remain highly relevant across various domains; yet, they receive far less attention. In this study, we investigate how smaller language models perform during the generation stage within a Retrieval-Augmented Generation (RAG) system. To benchmark these models effectively, we utilised both open-source and proprietary datasets covering diverse subject areas and question types. Our findings demonstrate that a RAG system with small language models can be executed directly on-device without requiring any GPU hardware within a reasonable time. The experimental code and links to the supplementary materials can be accessed through the GitHub repository: https://github.com/SibNN/SLM-RAG-EVAL.
Chinese Translation
尽管大型语言模型最近在研究领域占据主导地位，但小型语言模型在各个领域仍然具有高度相关性；然而，它们却受到的关注远远不够。在本研究中，我们探讨了小型语言模型在检索增强生成（Retrieval-Augmented Generation, RAG）系统生成阶段的表现。为了有效地对这些模型进行基准测试，我们利用了涵盖多种主题领域和问题类型的开源和专有数据集。我们的研究结果表明，使用小型语言模型的RAG系统可以直接在设备上运行，而无需任何GPU硬件，并且在合理的时间内完成。实验代码及补充材料的链接可以通过GitHub仓库访问：https://github.com/SibNN/SLM-RAG-EVAL。
cs.CL / 94 / 2606.30085
Not-quite-human tastes: the stylized omnivorousness of LLM survey surrogates
不完全人类的品味：大型语言模型调查代理的风格化杂食性
Abstract
Large-language models have proven to be remarkable if inconsistent parrots of public attitudes and opinions. The extent to which LLMs are able to produce reasonable approximations of cultural taste remains an open empirical question that becomes more urgent by the day, with market research companies already offering provisional `synthetic' survey panels and the contamination of standard survey data from LLM-generated responses. In this study, we build on past work on silicon sampling by extending considerations of its algorithmic fidelity and alignment to the domain of cultural consumption. We use large-language models from OpenAI, Anthropic, and DeepSeek to each produce 277,470 (30x9249) silicon surrogates of survey respondents from the Survey of Public Participation in the Arts (SPPA). We find these silicon surrogates' tastes to be highly stylized facsimiles of human tastes. (1) Silicon samples have a systematic postive-bias for liking, resulting in inflated ecological estimates of tastes. The individual-level bias of silicon samples are not well-explained by the WEIRD-bias often discussed in the literature. (2) The complex relationality in real taste structures is completely lost among silicon samples. (3) Finally, very little of the known cultural alignment between tastes and social space are preserved. Silicon samples attenuate age-taste associations, resurrect anachronistic class-taste associations, caricaturize gender- and race-taste associations.
Chinese Translation
大型语言模型已被证明是公众态度和意见的显著但不一致的模仿者。大型语言模型在多大程度上能够产生合理的文化品味近似值仍然是一个开放的实证问题，随着市场研究公司已经提供临时的“合成”调查小组，以及标准调查数据因大型语言模型生成的回应而受到污染，这一问题变得愈发紧迫。在本研究中，我们在以往关于硅样本的研究基础上，扩展了对其算法忠实度和与文化消费领域对齐的考虑。我们使用来自OpenAI、Anthropic和DeepSeek的大型语言模型，分别生成277,470个（30x9249）来自艺术公共参与调查（SPPA）的调查回应者的硅代理。我们发现这些硅代理的品味是人类品味的高度风格化的仿制品。(1) 硅样本对喜欢的系统性正偏差导致了对品味的生态估计膨胀。硅样本的个体层面偏差并不能很好地用文献中常讨论的WEIRD偏差来解释。(2) 真实品味结构中的复杂关系在硅样本中完全丧失。(3) 最后，已知的品味与社会空间之间的文化对齐几乎没有得到保留。硅样本削弱了年龄与品味的关联，复活了过时的阶级与品味的关联，并对性别和种族与品味的关联进行了夸张化。
cs.CL / 95 / 2606.30093
Efficient Retrieval-Augmented Generation via Token Co-occurrence Graphs
基于标记共现图的高效检索增强生成
Abstract
Retrieval-Augmented Generation (RAG) mitigates hallucinations in Large Language Models (LLMs) by grounding the generation process on external knowledge. However, standard RAG approaches struggle with multi-hop reasoning. While recent graph-based RAG methods improve the retrieval of interconnected chunks, they often rely on computationally expensive and error-prone LLM-based extraction pipelines. To address these issues, we propose TIGRAG (Token-Induced GraphRAG), an efficient graph-augmented RAG framework based on a token co-occurrence Knowledge Graph. TIGRAG directly models topological relationships between tokens using sliding-window co-occurrence statistics, thus enabling scalable graph construction. During inference, it combines graph-based semantic expansion and neural reranking to retrieve interconnected evidence for multi-hop reasoning. Specifically, it introduces an iterative entity-driven retrieval strategy that progressively expands the query using bridging entities extracted from previously retrieved contexts. We evaluated TIGRAG on three widely adopted multi-hop Question Answering (QA) benchmarks. Experimental results demonstrated that our framework consistently outperforms dense retrieval and graph-based RAG methods in both retrieval and downstream QA tasks, while substantially reducing indexing time, inference latency, and prompt footprint.
Chinese Translation
检索增强生成（Retrieval-Augmented Generation, RAG）通过将生成过程基于外部知识来减轻大型语言模型（Large Language Models, LLMs）的幻觉问题。然而，标准的 RAG 方法在多跳推理方面表现不佳。尽管最近的基于图的 RAG 方法改善了互联块的检索，但它们通常依赖于计算成本高且容易出错的基于 LLM 的提取管道。为了解决这些问题，我们提出了 TIGRAG（Token-Induced GraphRAG），一种基于标记共现知识图的高效图增强 RAG 框架。TIGRAG 直接使用滑动窗口共现统计建模标记之间的拓扑关系，从而实现可扩展的图构建。在推理过程中，它结合了基于图的语义扩展和神经重排序，以检索互联证据以进行多跳推理。具体而言，它引入了一种迭代的实体驱动检索策略，通过从先前检索的上下文中提取桥接实体，逐步扩展查询。我们在三个广泛采用的多跳问答（Question Answering, QA）基准上评估了 TIGRAG。实验结果表明，我们的框架在检索和下游 QA 任务中始终优于密集检索和基于图的 RAG 方法，同时显著减少了索引时间、推理延迟和提示占用空间。
cs.CL / 96 / 2606.30096
Information Dynamics of Language Communication
语言交流的信息动态
Abstract
Quantifying how meaning propagates through communicative exchanges remains underdeveloped in computational linguistics. Here we introduce an information-theoretic framework that quantifies the directed flow of semantic content between interlocutors and decomposes multi-source contributions into redundant, unique, and synergistic components. Our approach leverages large language models as probabilistic estimators of natural language to compute two measures: semantic transfer entropy (STE), which captures directed predictive influence between speakers, and semantic partial information decomposition (SPID), which resolves how multiple sources jointly shape a target's language. Across four experiments we show that the framework detects reduced information flow in cognitively rigid dialogue, captures the dominant role of persuaders in shaping discourse, distinguishes high- from low-quality psychotherapy by the directionality of therapist-client information exchange, and reveals synergistic premise contributions in argumentative essays. This framework opens new avenues for studying information dynamics in digital discourse, pedagogical interactions, clinical dialogues, and any domain in which the structure of linguistic exchange is of research relevance.
Chinese Translation
在计算语言学中，量化意义如何在交流中传播仍然处于发展阶段。在此，我们引入了一个信息论框架，该框架量化了交谈者之间语义内容的定向流动，并将多源贡献分解为冗余、独特和协同成分。我们的方法利用大型语言模型作为自然语言的概率估计器来计算两个指标：语义传递熵（semantic transfer entropy, STE），它捕捉了说话者之间的定向预测影响；以及语义部分信息分解（semantic partial information decomposition, SPID），它解析了多个来源如何共同塑造目标的语言。在四个实验中，我们展示了该框架能够检测到在认知僵化对话中信息流的减少，捕捉到劝说者在塑造话语中的主导作用，通过治疗师与客户信息交换的定向性区分高质量与低质量心理治疗，并揭示了论证性文章中的协同前提贡献。该框架为研究数字话语、教学互动、临床对话以及任何语言交流结构具有研究相关性的领域中的信息动态开辟了新的途径。
cs.CL / 97 / 2606.30152
Estimating Grammatical Gender Directions in Contextual Embeddings under Controlled and Natural Contexts
在受控和自然语境下估计上下文嵌入中的语法性别方向
Abstract
Contextual language models conflate grammatical gender and social semantic bias in gendered languages such as Spanish. Existing gender debiasing approaches only operate on static word embeddings leaving contextual representations unexplored for this two dimensional gender disentanglement. To address the this issue, we make the first attempt to disentangle grammatical gender from semantic contamination for contextual embeddings. We construct both controlled templates and natural Wikipedia contexts to build balanced datasets of inanimate nouns, and design a framework equipped with centroid, Support Vector Machine (SVM) and Linear Discriminant Analysis (LDA) gender direction estimators as well as contamination-aware weighting strategies. A set of dual-objective evaluation metrics is proposed to balance the suppression of grammatical gender leakage on inanimate nouns and the preservation of semantic gender distinctions for occupation terms. The results reveal that unweighted controlled contexts yield the purest grammatical gender direction, and the centroid estimator achieves better performance than discriminative baselines.
Chinese Translation
上下文语言模型在西班牙语等性别语言中将语法性别与社会语义偏见混为一谈。现有的性别去偏见方法仅在静态词嵌入上操作，未能探索上下文表示在这两个维度的性别解耦。为了解决这一问题，我们首次尝试将语法性别与上下文嵌入中的语义污染解耦。我们构建了受控模板和自然维基百科语境，以建立平衡的无生命名词数据集，并设计了一个框架，配备了质心、支持向量机（SVM）和线性判别分析（LDA）性别方向估计器，以及考虑污染的加权策略。提出了一组双目标评估指标，以平衡对无生命名词的语法性别泄漏的抑制和对职业术语语义性别区分的保留。结果表明，无加权的受控语境产生了最纯粹的语法性别方向，而质心估计器的表现优于判别基线。
cs.CL / 98 / 2606.30175
CORTEX: High-Quality Cross-Domain Organization of Web-Scale Corpora through Ontological Corpus Graph
CORTEX：通过本体语料图实现高质量跨域的大规模网络语料组织
Abstract
The continuous evolution of large language models drives escalating demands on data scale and quality, and as different training stages impose increasingly tailored data requirements, systematic organization of high-quality corpora becomes indispensable. Existing corpus construction pipelines confine the resulting corpora to flat, undifferentiated document collections, universally lacking systematic knowledge organization. We present Cortex, to our knowledge the first framework that elevates web-scale corpus construction from flat document filtering to structured knowledge organization through an Ontological Corpus Graph (OCG), a three-layer heterogeneous structure unifying a quality-refined content layer, a hierarchical lightweight ontology layer via LLM-driven automated evolution, and a cross-domain alignment layer enabling inter-domain association at arbitrary taxonomic resolution. Comprehensive experiments confirm the effectiveness of Cortex. In particular, we leverage the OCG to synthesize CortexBench, a cross-domain search-and-reasoning benchmark whose evaluation across eight frontier LLMs validates the effectiveness of quality refinement, domain organization, and cross-domain data synthesis. We will publicly release the complete codebase, a 24.14B-token refined corpus with its OCG, and CortexBench.
Chinese Translation
大型语言模型的持续演进推动了对数据规模和质量的日益增长的需求，而不同的训练阶段对数据的要求越来越具针对性，因此系统化组织高质量语料变得不可或缺。现有的语料构建流程将生成的语料限制为平面、无差异的文档集合，普遍缺乏系统的知识组织。我们提出了Cortex，作为我们所知的第一个框架，它通过本体语料图（Ontological Corpus Graph, OCG）将网络规模的语料构建从平面文档过滤提升到结构化知识组织。OCG是一个三层异构结构，统一了经过质量精炼的内容层、通过大型语言模型（LLM）驱动的自动演进形成的分层轻量本体层，以及支持任意分类分辨率的跨域对齐层，实现了跨域关联。全面的实验验证了Cortex的有效性。特别地，我们利用OCG合成了CortexBench，这是一个跨域搜索与推理基准，其在八个前沿大型语言模型上的评估验证了质量精炼、领域组织和跨域数据合成的有效性。我们将公开发布完整的代码库、一个包含241.4亿标记的精炼语料及其OCG，以及CortexBench。
cs.CL / 99 / 2606.30189
DAIN: Dynamic Agent-Based Interaction Network for Efficient and Collaborative Multimodal Reasoning
DAIN：用于高效协作多模态推理的动态基于代理的交互网络
Abstract
Current multimodal fusion approaches, particularly those based on static Mixture-of-Experts (MoE) architectures, often struggle to provide the adaptive and efficient collaborative reasoning required by complex real-world applications. We introduce the Dynamic Agent-based Interaction Network (DAIN), which reconceptualizes multimodal fusion as a dynamic, multi-agent collaborative process. DAIN employs a context-aware Meta-Controller that dynamically schedules sparse activation of specialized interaction agents and orchestrates compressed inter-agent communication for consensus-building. The framework is guided by a multi-objective loss function that jointly optimizes task accuracy, agent specialization, and operational efficiency through sparse activation and communication regularization. Comprehensive evaluations across five diverse benchmarks -- ADNI, MIMIC-IV, MM-IMDB, CMU-MOSI, and ENRICO -- establish DAIN as a new state-of-the-art, delivering significant performance improvements including a 2.6\% accuracy gain on ADNI. Ablation studies verify the critical roles of both dynamic scheduling and agent communication. Furthermore, DAIN offers enhanced interpretability by exposing context-dependent agent roles and collaboration patterns while maintaining computational efficiency through sample-wise sparse agent activation. Our work demonstrates the promise of dynamic, agent-based paradigms for multimodal reasoning.
Chinese Translation
当前的多模态融合方法，特别是基于静态专家混合（Mixture-of-Experts, MoE）架构的方法，往往难以提供复杂现实应用所需的自适应和高效的协作推理。我们提出了动态基于代理的交互网络（Dynamic Agent-based Interaction Network, DAIN），将多模态融合重新概念化为一个动态的多代理协作过程。DAIN采用上下文感知的元控制器（Meta-Controller），动态调度稀疏激活的专业交互代理，并协调压缩的代理间通信以达成共识。该框架由一个多目标损失函数指导，联合优化任务准确性、代理专业化和通过稀疏激活与通信正则化实现的操作效率。通过对五个不同基准（ADNI、MIMIC-IV、MM-IMDB、CMU-MOSI和ENRICO）的全面评估，确立了DAIN作为一种新的最先进方法，显著提升了性能，包括在ADNI上提高了2.6%的准确率。消融研究验证了动态调度和代理通信的关键作用。此外，DAIN通过揭示上下文依赖的代理角色和协作模式，提供了更好的可解释性，同时通过样本级稀疏代理激活保持计算效率。我们的工作展示了动态基于代理的范式在多模态推理中的潜力。
cs.CL / 100 / 2606.30196
Forewarned is Forearmed: When Non-Sequential Embedding Turns Into an Anomaly Detector
未雨绸缪：非顺序嵌入如何转变为异常检测器
Abstract
This paper offers an in-depth analysis of non-sequential multimodal sentence-level embeddings, with a particular focus on the SONAR model. We demonstrate that certain embedding dimensions are sensitive to perturbations and can serve as indicators of decoding anomalies. By leveraging the consistency between successive encoding and decoding, we successfully build an accurate detector. Additionally, we explore modifying specific dimensions of interest to attempt to correct them. This work underscores the importance of understanding and analyzing the embeddings themselves to enhance the reliability of multimodal representations.
Chinese Translation
本文对非顺序多模态句子级嵌入进行了深入分析，特别关注SONAR模型。我们证明了某些嵌入维度对扰动敏感，并可以作为解码异常的指示器。通过利用连续编码和解码之间的一致性，我们成功构建了一个准确的检测器。此外，我们还探索了修改特定感兴趣维度以尝试纠正它们的可能性。这项工作强调了理解和分析嵌入本身的重要性，以增强多模态表示的可靠性。
cs.CL / 101 / 2606.30217
Before Thinking, Learn to Decide: Proactive Routing for Efficient Visual Reasoning
在思考之前，学会决策：高效视觉推理的主动路由
Abstract
Large multimodal models have achieved strong reasoning on complex visual tasks, but their inference efficiency is often restricted by long chains of thought. A promising solution is to pair a small draft model with a large target model, enabling cooperative inference employing a routing signal that adaptively routes queries to either the draft or target model based on their difficulties for optimal efficiency and accuracy. Yet, the remaining bottleneck is to establish a reliable query difficulty signal under multimodal settings. Existing approaches designed for language models either rely on post-hoc token probabilities, which fall short in multimodal scenarios, or depend on supervised fine-tuning, which is a data-sensitive strategy. Both paradigms perform routing only after a complete output, and ignore whether the target model can actually solve the routed instances. To address this, we propose PRP, a Proactive Routing Paradigm that enables early decision-making by jointly evaluating the competence of both the draft and target models. Our Draft Rating Learning (DRL) equips the draft model with an internal confidence estimator, while Joint Rating Learning (JRL) predicts how well the target model can handle a given query, thereby prioritizing the allocation of samples it excels at rather than the hardest ones. These ratings enable fine-grained, instance-level \textbf{Proactive Routing} and substantially accelerate inference without compromising overall performance. Extensive experiments across multiple multimodal reasoning benchmarks validate our effectiveness and efficiency.
Chinese Translation
大型多模态模型在复杂视觉任务上取得了强大的推理能力，但它们的推理效率常常受到长时间思考链的限制。一个有前景的解决方案是将一个小型草稿模型与一个大型目标模型配对，利用路由信号进行协同推理，该信号根据查询的难度自适应地将查询路由到草稿模型或目标模型，以实现最佳的效率和准确性。然而，剩下的瓶颈是如何在多模态环境下建立一个可靠的查询难度信号。现有针对语言模型的方法要么依赖于事后生成的标记概率，这在多模态场景中效果不佳，要么依赖于监督微调，这是一种对数据敏感的策略。这两种范式仅在完整输出后进行路由，忽视了目标模型是否能够实际解决被路由的实例。为了解决这个问题，我们提出了PRP（主动路由范式），通过共同评估草稿模型和目标模型的能力，实现早期决策。我们的草稿评分学习（Draft Rating Learning, DRL）为草稿模型配备了内部置信度估计器，而联合评分学习（Joint Rating Learning, JRL）则预测目标模型处理给定查询的能力，从而优先分配其擅长的样本，而不是最难的样本。这些评分实现了细粒度的实例级主动路由，并在不妥协整体性能的情况下显著加速推理。我们在多个多模态推理基准上的广泛实验验证了我们的有效性和效率。
cs.CL / 102 / 2606.30236
CaresAI at CT-DEB26: Detecting Dosing Errors In Clinical Trials Using Domain-Specific Transformer Embeddings and Classification Models
CaresAI 在 CT-DEB26：使用领域特定的变换器嵌入和分类模型检测临床试验中的剂量错误
Abstract
Medication errors, particularly dosing errors in clinical trials (CT), can lead to patient harm, adverse drug events and worse patient outcomes. Dosing errors are preventable, and early identification can improve trial integrity and mitigate subsequent clinical and financial burden. This study aims to detect dosing errors within CT protocols by evaluating text representations of trial information using transformer-based language models trained on biomedical corpora. CT textual data was encoded using several models, including ClinicalBERT, PubMedBERT, BioBERT, and MedCPT, and integrated with categorical features. These text embeddings were used as input to classical machine learning models and neural network architectures within an experimental framework. Performance was primarily assessed using ROC-AUC with respect to predicting dosage error. Under a logistic regression baseline, BioBERT consistently outperformed alternative encoders, achieving an ROC-AUC of 0.794, a 3.95% improvement over the ClinicalBERT baseline. Combining multiple embeddings did not yield improvements, indicating that domain alignment outweighs representational stacking. Gradient boosting models, support vector classifiers, logistic regression, and residual neural networks achieved the strongest performance for predicting dosage error, achieving ROC-AUCs: 0.821 to 0.853. Overall, the integration of domain-specific transformer embeddings with structured metadata enables discrimination of trials meeting a predefined elevated dosing error risk criterion, advancing safety monitoring and supporting informed regulatory decision-making.
Chinese Translation
药物错误，特别是在临床试验（CT）中的剂量错误，可能导致患者伤害、不良药物事件以及更差的患者结果。剂量错误是可以预防的，早期识别可以改善试验的完整性并减轻随后的临床和财务负担。本研究旨在通过评估使用在生物医学语料库上训练的基于变换器的语言模型的试验信息文本表示，检测 CT 协议中的剂量错误。CT 文本数据使用多个模型进行编码，包括 ClinicalBERT、PubMedBERT、BioBERT 和 MedCPT，并与分类特征集成。这些文本嵌入被用作经典机器学习模型和神经网络架构的输入，置于实验框架内。性能主要通过 ROC-AUC 来评估，以预测剂量错误。在逻辑回归基线下，BioBERT 始终优于其他编码器，ROC-AUC 达到 0.794，比 ClinicalBERT 基线提高了 3.95%。组合多个嵌入未能带来改进，表明领域对齐的优先性超过了表示堆叠。梯度提升模型、支持向量分类器、逻辑回归和残差神经网络在预测剂量错误方面表现最强，ROC-AUC 达到 0.821 至 0.853。总体而言，领域特定的变换器嵌入与结构化元数据的结合使得能够区分符合预定义的较高剂量错误风险标准的试验，从而推动安全监测并支持知情的监管决策。
cs.CL / 103 / 2606.30237
Comparing Human and Automatic Recognition of Dutch Dysarthric Continuous Speech: A Case Study
比较人类与自动识别荷兰语构音障碍连续语音的案例研究
Abstract
In our goal to develop personalised dysarthric speech recognition (DSR) models, this study compared the recognition performances of human listeners and those of three state-of-the-art, off-the-shelf ASR systems (Whisper-large-V3, Google Chirp 3, and Omnilingual) on the recognition of Dutch continuous read and spontaneous speech from a single speaker with severe dysarthria. Results showed that both humans listeners and the three off-the-shelf ASR systems exhibit word error rates (WER) exceeding 70% on average, indicating that DSR is highly challenging for both humans and ASR systems. Fine-tuning on the dysarthric speech significantly reduced WER. Although overall WERs are still quite high (>23%), the personalised DSR models outperformed the human listeners, and performance is getting closer to being useful for supporting day-to-day communication of dysarthric speakers. Future research should focus on improving personalized DSR on spontaneous speech and longer utterances in the case of read speech, with a specific focus on particular phonemes.
Chinese Translation
在我们开发个性化构音障碍语音识别（DSR）模型的目标下，本研究比较了人类听众与三种先进的现成自动语音识别（ASR）系统（Whisper-large-V3、Google Chirp 3 和 Omnilingual）在识别来自一位重度构音障碍说话者的荷兰语连续朗读和自发语音时的表现。结果显示，人类听众和这三种现成的ASR系统的平均词错误率（WER）均超过70%，表明DSR对人类和ASR系统来说都是极具挑战性的。对构音障碍语音的微调显著降低了WER。尽管总体WER仍然相当高（>23%），个性化的DSR模型的表现优于人类听众，且其性能越来越接近于支持构音障碍说话者的日常交流。未来的研究应集中于改善自发语音和朗读语音中较长发音的个性化DSR，特别关注特定音素。
cs.CL / 104 / 2606.30247
Grounding LLM Reasoning under Incomplete Graph Evidence
在不完整图证据下的LLM推理基础
Abstract
Knowledge graphs can guide large language models (LLMs) reasoning, but the graph seen by a system is usually a retrieved, linked, temporally scoped, and incomplete evidence state rather than a complete account of truth. We develop a theoretical perspective on grounding observable LLM trajectories under such incomplete graph evidence.The evidence state induces entity anchors, typed relation residuals, path energies, and support regions, while the language model supplies a prior over candidate trajectories. We show that, under open-world incompleteness, no hard rule based only on the observed state can both reject every false unsupported trajectory and retain every true-but-unobserved one.We then characterize soft grounding as a KL-regularized deformation of the LLM prior: finite slack preserves support for unsupported but non-contradicted trajectories, whereas hard conditioning appears as an infinite-penalty limit.The framework also yields stability bounds under evidence perturbations and clarifies the constraint regimes appropriate for GraphRAG, KGQA, graph agents, constrained decoding, and faithful generation. The claims are evidence-relative: KG compatibility is treated as declared support, not factual truth.
Chinese Translation
知识图谱可以指导大型语言模型（LLMs）的推理，但系统所看到的图通常是一个检索到的、链接的、时间范围内的、不完整的证据状态，而不是完整的真相叙述。我们在这种不完整图证据下发展了一个理论视角，以基础可观察的LLM轨迹。证据状态引发实体锚点、类型关系残差、路径能量和支持区域，而语言模型则提供候选轨迹的先验。我们表明，在开放世界的不完整性下，基于观察状态的任何硬性规则都无法同时拒绝每一个虚假的不支持轨迹和保留每一个真实但未观察到的轨迹。然后，我们将软基础特征化为LLM先验的KL正则化变形：有限的松弛保留对不支持但不矛盾轨迹的支持，而硬性条件则表现为无限惩罚极限。该框架还在证据扰动下提供了稳定性界限，并阐明了适用于GraphRAG、KGQA、图代理、受限解码和真实生成的约束范围。这些主张是相对证据的：KG兼容性被视为声明支持，而非事实真相。
cs.CL / 105 / 2606.30259
Multi-Agentic System Leveraging Open-Source LLMs to Mitigate Disinformation Threats
利用开源大型语言模型的多智能体系统以减轻虚假信息威胁
Abstract
In contemporary societies, the threat of disinformation has reached alarming levels, exacerbated by the proliferation of electronic communication, social media, and advancements in artificial intelligence. As a result, there is an urgent need to develop effective countermeasures to mitigate this menace. However, the sheer scale of the problem renders manual fact-checking and human-based verification inadequate, underscoring the necessity for automated methods to detect and debunk disinformation. This article proposes a novel approach based on a multi-agent system that emulates the decision-making processes of human annotators engaged in disinformation detection tasks. By incorporating a consensus mechanism, diversity in cognition and diversity in knowledge, and also hierarchical structure, inspired by human annotators' behavior, the proposed method achieves superior results compared to individual Large Language Models (LLMs), including GPT 4 and GPT 3.5. The system leverages open models (e.g., LLaMA, Kimi, Qwen, Deepseek and LLaMA-Nemotron) to ensure greater transparency. The evaluation of the proposed method encompasses datasets in languages with varying resource availability, including English (high-resource), Polish (medium-resource), Slovak (low-resource) and Bulgarian (low-resource). Experiments were conducted on tasks such as direct disinformation detection, identification of texts worthy of verification, and detection of texts containing verifiable factual claims.
Chinese Translation
在当代社会，虚假信息的威胁已达到令人担忧的程度，这一问题因电子通信、社交媒体的普及以及人工智能的进步而加剧。因此，迫切需要开发有效的对策来减轻这一威胁。然而，问题的规模使得人工事实核查和基于人类的验证显得不足，强调了自动化方法在检测和揭穿虚假信息中的必要性。本文提出了一种基于多智能体系统的新方法，该系统模拟参与虚假信息检测任务的人类标注者的决策过程。通过引入共识机制、认知多样性和知识多样性，以及受到人类标注者行为启发的层次结构，所提出的方法在效果上优于单一的大型语言模型（LLMs），包括GPT 4和GPT 3.5。该系统利用开放模型（如LLaMA、Kimi、Qwen、Deepseek和LLaMA-Nemotron）以确保更大的透明度。所提方法的评估涵盖了不同资源可用性的语言数据集，包括英语（高资源）、波兰语（中资源）、斯洛伐克语（低资源）和保加利亚语（低资源）。实验涉及直接虚假信息检测、值得验证的文本识别以及包含可验证事实声明的文本检测等任务。
cs.CL / 106 / 2606.30312
DialogPII: A multilingual dataset of synthetic dialog transcripts to detect personal information
DialogPII：用于检测个人信息的多语言合成对话记录数据集
Abstract
Conversational data collected in domains such as healthcare or social sciences is a valuable resource for research and automated analysis. However, responsible data sharing requires the detection and removal of personally identifiable and sensitive information to protect individual privacy. To support the development and evaluation of automatic de-identification systems, we present DialogPII, a multilingual dataset of synthetic dialogs and speech-derived transcripts for personal information detection. DialogPII covers eight interaction scenarios (emergency calls, medical anamnesis interviews, therapy sessions, insurance communication, customer support, clinical interviews regarding an AI-supported dashboard, police reports, and group therapy discussions), 19 entity types, and 11 languages (English, Arabic, Finnish, French, German, Hindi, Italian, Polish, Portuguese, Spanish, and Turkish). Dialogs were generated semi-automatically using large language models, manually curated for plausibility and diversity, and localized to country- and city-specific contexts. All dialogs were additionally converted to speech via text-to-speech synthesis, transcribed with Whisper, and annotated through automatic projection and manual correction, yielding aligned written and speech-derived resources across all languages. We further release baseline multilingual named entity recognition models and provide technical validation through inter-annotator agreement analysis, translation quality evaluation, annotation projection assessment, and benchmark experiments with transformer-based sequence labeling models.
Chinese Translation
在医疗保健或社会科学等领域收集的对话数据是研究和自动化分析的宝贵资源。然而，负责任的数据共享需要检测和移除个人可识别信息和敏感信息，以保护个人隐私。为了支持自动去标识化系统的开发和评估，我们提出了DialogPII，这是一个用于个人信息检测的多语言合成对话和语音衍生记录的数据集。DialogPII涵盖八种交互场景（紧急电话、医疗病史访谈、治疗会议、保险沟通、客户支持、关于AI支持的仪表板的临床访谈、警察报告和小组治疗讨论）、19种实体类型和11种语言（英语、阿拉伯语、芬兰语、法语、德语、印地语、意大利语、波兰语、葡萄牙语、西班牙语和土耳其语）。对话是通过大型语言模型半自动生成的，经过手动策划以确保合理性和多样性，并本地化到特定国家和城市的背景。所有对话还通过文本到语音合成转换为语音，使用Whisper转录，并通过自动投影和手动校正进行注释，从而在所有语言中生成对齐的书面和语音衍生资源。我们还发布了基准多语言命名实体识别模型，并通过标注者间一致性分析、翻译质量评估、注释投影评估以及基于变换器的序列标注模型的基准实验提供技术验证。
cs.CL / 107 / 2606.30339
REAR: Test-time Preference Realignment through Reward Decomposition
REAR：通过奖励分解实现测试时偏好重新对齐
Abstract
Aligning large language models (LLMs) with diverse user preferences is a critical yet challenging task. While post-training methods can adapt models to specific needs, they often require costly data curation and additional training. Test-time scaling (TTS) presents an efficient, training-free alternative, but its application has been largely limited to verifiable domains like mathematics and coding, where response correctness is easily judged. To extend TTS to preference alignment, we introduce a novel framework that models the task as a realignment problem, since the base model often fails to sufficiently align with the stated preference. Our key insight is to decompose the underlying reward function into two components: one related to the question and the other to preference information. This allows us to derive a REAlignment Reward (REAR) that selectively rescales the proportions of these two reward terms. We then show that REAR can be formulated as a linear combination of token-level policy log-probabilities, making it computationally efficient and easy to integrate with various TTS algorithms such as best-of-$N$ sampling and tree search. Experiments show that compared to other test-time baselines, REAR not only enables scalable test-time realignment for preference alignment tasks under diverse user requirements, but also generalizes to mathematical and visual tasks under appropriate preference settings.
Chinese Translation
将大型语言模型（LLMs）与多样化的用户偏好对齐是一项关键但具有挑战性的任务。尽管后训练方法可以使模型适应特定需求，但通常需要昂贵的数据整理和额外的训练。测试时缩放（TTS）提供了一种高效的、无训练的替代方案，但其应用主要限于数学和编程等可验证领域，在这些领域中，响应的正确性易于判断。为了将TTS扩展到偏好对齐，我们引入了一种新颖的框架，将任务建模为重新对齐问题，因为基础模型通常无法充分对齐所述的偏好。我们的关键见解是将基础奖励函数分解为两个组成部分：一个与问题相关，另一个与偏好信息相关。这使我们能够推导出一种重新对齐奖励（REAlignment Reward，REAR），该奖励选择性地重新调整这两个奖励项的比例。然后，我们展示了REAR可以被表述为令牌级策略对数概率的线性组合，使其在计算上高效，并且易于与各种TTS算法（如最佳$N$采样和树搜索）集成。实验表明，与其他测试时基线相比，REAR不仅能够在多样化用户需求下实现可扩展的测试时偏好对齐任务，而且在适当的偏好设置下也能推广到数学和视觉任务。
cs.CL / 108 / 2606.30356
OLIVE: View-Augmented Latent Prediction with Waveform Reconstruction for Speech SSL
OLIVE：基于波形重建的视图增强潜在预测用于语音自监督学习
Abstract
We propose Online Latent prediction with Invariant Views and rEconstruction (OLIVE), a self-supervised speech representation learning framework that jointly optimizes analysis and synthesis objectives. OLIVE combines view-augmented masked latent prediction with waveform reconstruction under a unified objective. Reconstruction constrains early encoder features to retain signal-level information, while masked latent prediction shapes later contextual representations toward invariance for robust downstream performance. We show that these objectives enable representations that support a broad range of tasks. In particular, OLIVE improves results on generation and speaker tasks, maintains competitive performance on recognition and semantic tasks, and improves waveform reconstruction.
Chinese Translation
我们提出了在线潜在预测与不变视图和重建（OLIVE），这是一个自监督语音表示学习框架，旨在联合优化分析和合成目标。OLIVE在统一目标下结合了视图增强的掩蔽潜在预测与波形重建。重建约束早期编码器特征以保留信号级信息，而掩蔽潜在预测则使后期上下文表示朝向不变性发展，以实现稳健的下游性能。我们展示了这些目标使得表示能够支持广泛的任务。特别是，OLIVE在生成和说话人任务上改善了结果，在识别和语义任务上保持了竞争力的表现，并改善了波形重建。
cs.CL / 109 / 2606.30406
MOPD: Multi-Teacher On-Policy Distillation for Capability Integration in LLM Post-Training
MOPD：用于大规模语言模型后训练能力集成的多教师在线蒸馏
Abstract
Modern large language models (LLMs) rely on reinforcement learning during post-training to push specific capabilities, yet integrating multiple capabilities into one model remains hard. Existing methods, such as Off-Policy Finetune and Mix-RL, are either inefficient or lose performance. In this work, we propose Multi-teacher On-Policy Distillation (MOPD), a post-training paradigm for combining the capabilities of multiple domain RL teachers: we first run per-domain specialised RL to obtain a set of domain teachers, then distill these teachers into the student on its own rollouts. This eliminates exposure bias and provides a dense optimization signal. On Qwen3-30B-A3B, MOPD outperforms Mix-RL, Cascade RL, Off-Policy Finetune, and Param-Merge baselines, inheriting nearly all of each teacher's capability. MOPD also enables parallel, independent development of domain teachers, removing the cross-domain coupling typical of multi-domain post-training. MOPD has been deployed in the post-training of MiMo-V2-Flash, an industrial-scale frontier model, demonstrating its practical value for capability integration in frontier-scale LLMs.
Chinese Translation
现代大型语言模型（LLMs）在后训练阶段依赖强化学习以提升特定能力，但将多种能力整合到一个模型中仍然困难。现有方法，如离线微调（Off-Policy Finetune）和混合强化学习（Mix-RL），要么效率低下，要么性能下降。在本研究中，我们提出了多教师在线蒸馏（Multi-teacher On-Policy Distillation，MOPD），这是一种结合多个领域强化学习教师能力的后训练范式：我们首先针对每个领域运行专业化的强化学习，以获得一组领域教师，然后将这些教师蒸馏到学生模型的自我回放中。这消除了暴露偏差，并提供了密集的优化信号。在 Qwen3-30B-A3B 上，MOPD 超越了混合强化学习、级联强化学习、离线微调和参数合并（Param-Merge）基线，几乎继承了每位教师的所有能力。MOPD 还使领域教师的并行独立开发成为可能，消除了多领域后训练中典型的跨领域耦合。MOPD 已在工业规模前沿模型 MiMo-V2-Flash 的后训练中部署，展示了其在前沿规模 LLM 能力集成中的实际价值。
cs.CL / 110 / 2606.30473
Field Order Should Not Matter: Permutation-Invariant Embedding Model Fine-Tuning for Structured Metadata Retrieval
字段顺序不应影响结果：用于结构化元数据检索的置换不变嵌入模型微调
Abstract
We study retrieval over catalogs of structured metadata, where each record is a small schema whose fields answer different kinds of query. Embedding a record with a text encoder first serializes its fields into a string, which forces a choice of field order. We show this choice, usually treated as an implementation detail, silently controls retrieval quality once the encoder is fine-tuned. A standard fine-tune loses 7.4 nDCG@10 points when the index is rebuilt under a different field order, because it reads absolute position instead of the field labels. We propose permutation-invariant fine-tuning ($\textbf{PI-FT}$), which serializes each record under a freshly sampled field order with random field dropout, so meaning binds to the labels rather than to position. The change is about two lines in the data loader; it costs negligible in-distribution accuracy and cuts the order-change penalty to 0.2 points. We study this in the discovery of development statistics, a catalog of nearly 10,000 indicators that should be searchable in many languages by a model small enough to self-host. As AI assistants and agents increasingly mediate access to public data and statistics, this retrieval step decides whether an answer is grounded in the right indicator or series, making discoverability a precondition for disseminating data through AI. Because usage logs cannot provide training signal for indicators no one has searched, we generate the queries instead. $\textbf{DevDataBench}$ is a fully LLM-generated benchmark of grounded, facet-targeted queries across 15 languages, covering every indicator for both training and evaluation. A fine-tuned 118M-parameter CPU encoder outperforms every zero-shot baseline, including $\texttt{text-embedding-3-large}$ (0.707 vs.\ 0.556 nDCG@10), with the largest gains in low-resource languages. We release the benchmark, pipeline, models, and a reusable PI-FT framework.
Chinese Translation
我们研究了对结构化元数据目录的检索，其中每条记录是一个小型模式，其字段回答不同类型的查询。首先，使用文本编码器对记录进行嵌入，将其字段序列化为字符串，这就强制选择字段顺序。我们展示了这一选择，通常被视为实现细节，在编码器微调后默默地控制了检索质量。当索引在不同字段顺序下重建时，标准微调会损失7.4个nDCG@10点，因为它读取的是绝对位置而不是字段标签。我们提出了置换不变微调（Permutation-Invariant Fine-Tuning，$ extbf{PI-FT}$），该方法在随机字段丢弃的情况下，以新采样的字段顺序序列化每条记录，从而使意义与标签绑定，而不是与位置绑定。这个变化仅需在数据加载器中增加大约两行代码；它对分布内准确性的影响微乎其微，并将顺序变化的惩罚降低到0.2点。我们在开发统计数据的发现中研究这一点，该目录包含近10,000个指标，应该能够通过一个足够小以自我托管的模型以多种语言进行检索。随着人工智能助手和代理越来越多地介入公共数据和统计信息的访问，这一步检索决定了答案是否基于正确的指标或系列，使得可发现性成为通过人工智能传播数据的前提条件。由于使用日志无法为无人搜索的指标提供训练信号，我们生成查询作为替代。$ extbf{DevDataBench}$是一个完全由大型语言模型生成的基准，涵盖15种语言的基础、面向特征的查询，涵盖了所有指标用于训练和评估。微调后的118M参数CPU编码器在每个零样本基线中表现优于，包括$ exttt{text-embedding-3-large}$（0.707对0.556 nDCG@10），在低资源语言中获得了最大的提升。我们发布了基准、管道、模型和可重用的PI-FT框架。
cs.CL / 111 / 2606.30491
SIMAX: A Scalable and Interpretable Framework for Multi-Fidelity and Annotated Clinician-Patient Dialogue Simulation
SIMAX：一种可扩展且可解释的多保真度和标注临床医生-患者对话模拟框架
Abstract
Background. The widespread deployment of ambient digital scribes is driving large-scale capture of clinician-patient dialogues. Human coding of clinical communication data remains costly, inconsistent, and difficult to scale, motivating AI-driven communication coding systems. However, evaluating these systems requires real-world dialogues and human-coded labels, both hard to obtain at scale. Methods. We developed SIMAX (Scalable and Interpretable Framework for Multi-Fidelity and Annotated Clinician-Patient Dialogue Simulation), a framework for generating controlled clinical dialogue data with reference behavioral annotations. SIMAX generates clinician-patient dialogues from predefined clinical scenarios, personas and voice conditions, and target communication behaviors. Behaviors are controlled using two codebooks: the Global Codebook for overall communication quality and the WISER Codebook for specific countable behaviors. We evaluated SIMAX using automated and human quality assessments and an example communication coding system. Results. SIMAX generated 3,388 simulated dialogues across three specialties, multiple visit stages, persona characteristics, and accent conditions. Automated assessment showed mean UTMOS and WV-MOS scores of 3.03 and 2.61, WER and CER of 0.07 and 0.05, and CLAP cosine similarity of 0.41, suggesting reasonable speech naturalness, high transcription fidelity, and positive text-audio correspondence. Human evaluation showed a median MOS of 4.67 and a median clinical realism score of 3.00. Downstream evaluation suggests that SIMAX can assess how a communication coding system responds to behavioral targets and reveal insufficient sensitivity in some dimensions. Conclusions. SIMAX generates controlled and reproducible simulated clinician-patient dialogues, providing a data foundation for developing, validating, and refining communication coding systems.
Chinese Translation
背景：环境数字记录员的广泛部署正在推动临床医生-患者对话的大规模捕捉。对临床沟通数据的人为编码仍然成本高昂、不一致且难以扩展，这促使了基于人工智能的沟通编码系统的出现。然而，评估这些系统需要真实世界的对话和人工编码的标签，而这两者在大规模获取上都很困难。方法：我们开发了SIMAX（可扩展且可解释的多保真度和标注临床医生-患者对话模拟框架），这是一个用于生成具有参考行为注释的受控临床对话数据的框架。SIMAX根据预定义的临床场景、角色和语音条件以及目标沟通行为生成临床医生-患者对话。通过两个编码本来控制行为：用于整体沟通质量的全球编码本（Global Codebook）和用于特定可计数行为的WISER编码本（WISER Codebook）。我们使用自动化和人工质量评估以及一个示例沟通编码系统对SIMAX进行了评估。结果：SIMAX生成了3,388个模拟对话，涵盖三个专业、多个就诊阶段、角色特征和口音条件。自动评估显示平均UTMOS和WV-MOS评分分别为3.03和2.61，WER和CER分别为0.07和0.05，CLAP余弦相似度为0.41，表明语音自然性合理、转录保真度高以及文本与音频之间的良好对应关系。人工评估显示中位MOS为4.67，中位临床现实评分为3.00。下游评估表明，SIMAX可以评估沟通编码系统对行为目标的响应，并揭示某些维度的敏感性不足。结论：SIMAX生成受控且可重复的模拟临床医生-患者对话，为开发、验证和完善沟通编码系统提供了数据基础。
cs.CL / 112 / 2606.30518
Regime-Aware Peer Specialization for Robust RAG under Heterogeneous Knowledge Conflicts
考虑体制的同行专业化在异构知识冲突下的鲁棒检索增强生成
Abstract
Retrieval-augmented generation (RAG) improves language models by grounding generation in external context. However, it can be fragile when the retrieved context conflicts with the model's parametric knowledge. Such conflicts span a reliability spectrum, ranging from reliable and partially reliable evidence to adversarial context. Existing remedies often handle such heterogeneous conflicts with regime-agnostic supervision, which can conflate incompatible learning signals across reliability regimes. To disentangle these signals, we propose RAPS-DA, a regime-aware peer specialization framework that addresses conflict at two complementary granularities. At the sample level, conflicts are divided into three regimes, including Grounding, Arbitration, and Resistance, with one same-scale peer specialist trained per regime from a shared base model. Each sample is then hard-routed to its regime-matched peer for on-policy reverse-KL supervision. At the token level, a dual-layer selector uses inter-teacher disagreement, student-teacher divergence, and student entropy to filter uninformative or unstable tokens, upweight confidently misaligned ones, and gradually focus supervision on high-conflict tokens as the student matures. Gains stem from specialization at a fixed model scale, not from a stronger teacher, and the peer specialists exist only during training, so the deployed student requires no regime labels or peer access. Experiments on five conflict scenarios and two out-of-distribution benchmarks show RAPS-DA surpasses all prompting, decoding, fine-tuning, RL, and single-teacher baselines.
Chinese Translation
检索增强生成（RAG）通过将生成过程与外部上下文相结合来改善语言模型。然而，当检索到的上下文与模型的参数知识发生冲突时，它可能会变得脆弱。这种冲突在可靠性谱上呈现，从可靠和部分可靠的证据到对抗性上下文。现有的解决方案通常采用与体制无关的监督来处理这种异构冲突，这可能会混淆不同可靠性体制之间不兼容的学习信号。为了理清这些信号，我们提出了RAPS-DA，一个考虑体制的同行专业化框架，该框架在两个互补的粒度上解决冲突。在样本级别，冲突被划分为三个体制，包括基础、仲裁和抵抗，每个体制从共享的基础模型中训练一个同规模的同行专家。然后，每个样本被硬性路由到其匹配的同行进行政策内反向KL监督。在标记级别，双层选择器利用教师之间的不一致、学生与教师之间的差异以及学生的熵来过滤无信息或不稳定的标记，增加自信地不一致标记的权重，并随着学生的成熟逐渐将监督集中在高冲突标记上。收益来自于在固定模型规模下的专业化，而不是来自更强的教师，并且同行专家仅在训练期间存在，因此部署的学生不需要体制标签或同行访问。在五个冲突场景和两个分布外基准上的实验表明，RAPS-DA超越了所有提示、解码、微调、强化学习和单教师基线。
cs.CL / 113 / 2606.30543
TRACE: Temporal Relationship-Aware Conversational Entrainment Detection in Dyadic Speech
TRACE：基于时间关系的双人对话情感同步检测
Abstract
With the proliferation of speech AI agents, understanding emotional entrainment in conversational interaction has become increasingly important. Emotional entrainment is shaped by social relationships and conversational context, influencing affective coordination over time. We introduce DyadEE, a dataset for emotional entrainment detection in dyadic speech interactions, containing both emotionally entrained conversations and synthetic interactions where entrainment is disrupted through partner swapping and emotion resynthesis. We further propose TRACE, a window-level framework that models dyadic interaction as ordered sequences of acoustic embeddings derived from emotion fine-tuned Whisper representations, treating each sample as an interaction trace rather than pooled utterances. Experimental results on DyadEE show that incorporating conversational context and relationship information improves emotional entrainment detection, with TRACE achieving the best accuracy of 97.01%.
Chinese Translation
随着语音人工智能代理的普及，理解对话互动中的情感同步变得越来越重要。情感同步受到社会关系和对话背景的影响，进而影响情感协调的时间演变。我们引入了DyadEE，一个用于双人对话情感同步检测的数据集，包含了情感同步的对话和通过交换伙伴和情感再合成而中断同步的合成互动。我们进一步提出了TRACE，一个窗口级框架，将双人互动建模为从情感微调的Whisper表示中提取的声学嵌入的有序序列，将每个样本视为互动轨迹，而不是汇总的发言。DyadEE上的实验结果表明，结合对话背景和关系信息能够提高情感同步检测的准确性，TRACE实现了97.01%的最佳准确率。
cs.CL / 114 / 2606.30556
Poller: Are LLMs Suitable for Evaluating the Poetry Understanding Task?
Poller：大语言模型是否适合评估诗歌理解任务？
Abstract
Traditional automatic evaluation methods have been shown to be unsuitable for modern Chinese poetry because of the distinct nature of this literary genre. Human evaluation remains reliable, but is expensive and not applicable to large-scale data. In this paper, we propose Poller (Poetry LLM Evaluator), a novel method leveraging large language models (LLMs) to evaluate the poetry understanding task. Specifically, our method requires LLMs to play the role of a poem's author with detailed information, thereby emulating human evaluation and judgment by adopting the poet's perspective. We conducted comprehensive experiments on multiple LLMs, evaluating the interpretations of poems across eight specialized dimensions. Experimental results demonstrate that our method effectively reduces the evaluation error between LLMs and humans. Especially for specific dimension evaluation, Poller-based LLMs achieve a 94.55% and 89.53% error reduction for rhetorical techniques and defamiliarization, respectively, compared to baseline methods. These performances are unattainable by conventional LLM evaluation methods. Experimental results from multiple LLMs across various dimensions validate the efficacy of our method. This work bridges the gap between automated efficiency and human expertise, establishing a foundation for automated evaluation in poetry-related tasks.
Chinese Translation
传统的自动评估方法已被证明不适用于现代汉语诗歌，因为这一文学体裁具有独特的性质。尽管人工评估仍然可靠，但其成本高昂且不适用于大规模数据。在本文中，我们提出了Poller（诗歌大语言模型评估器），这是一种利用大语言模型（LLMs）评估诗歌理解任务的新方法。具体而言，我们的方法要求LLMs扮演诗歌作者的角色，并提供详细信息，从而通过采用诗人的视角模拟人类的评估和判断。我们对多种LLMs进行了全面实验，评估了诗歌在八个专业维度上的解读。实验结果表明，我们的方法有效减少了LLMs与人类之间的评估误差。特别是在特定维度评估方面，与基线方法相比，基于Poller的LLMs在修辞技巧和陌生化方面分别实现了94.55%和89.53%的误差减少。这些表现是传统LLM评估方法无法达到的。来自多种LLMs在不同维度上的实验结果验证了我们方法的有效性。本研究弥合了自动化效率与人类专业知识之间的差距，为诗歌相关任务的自动评估奠定了基础。
cs.CL / 115 / 2606.30562
Morphing into Hybrid Attention Models
转变为混合注意力模型
Abstract
Hybrid attention models improve long-context efficiency by retaining only a subset of full-attention layers and replacing the remaining layers with linear attention. However, the effectiveness of Transformer-to-hybrid conversion critically depends on which layers preserve full attention. Existing hybrid layer selection methods typically rely on heuristic strategies such as fixed placement patterns or layerwise scoring, implicitly treating layer importance as isolated and overlooking the interdependent layer effect under a global hybrid configuration. In this work, we formulate hybrid layer selection as a budget-constrained subset optimization problem. We further propose FlashMorph (Fast LAyer Selection for Hybrid MORPHing), an effective, efficient and scalable layer selection method for Transformer-to-hybrid conversion. FlashMorph first constructs a morphable model by equipping each full-attention layer with a converted linear-attention branch. It then freezes all model weights and jointly optimizes layerwise gates on synthetic long-context retrieval data, with a linearization regularization that encourages the model to rely on linear attention for efficiency. The learned gates are discretized under a preset full-attention budget to instantiate the hybrid architecture, followed by standard logits distillation and long-context finetuning. Extensive experiments show that FlashMorph discovers more effective hybrid configurations, preserves strong long-context recall and general benchmark performance while substantially reducing layer selection cost compared with existing layer selection methods, demonstrating its effectiveness, efficiency, and scalability.
Chinese Translation
混合注意力模型通过仅保留部分全注意力层并用线性注意力替代其余层来提高长上下文的效率。然而，Transformer到混合模型的转换效果在很大程度上取决于哪些层保留全注意力。现有的混合层选择方法通常依赖于启发式策略，如固定放置模式或逐层评分，隐含地将层的重要性视为孤立的，并忽视了在全局混合配置下层之间的相互依赖效应。在本研究中，我们将混合层选择形式化为一个预算约束的子集优化问题。我们进一步提出了FlashMorph（快速层选择用于混合转变），这是一种有效、高效且可扩展的层选择方法，用于Transformer到混合模型的转换。FlashMorph首先通过为每个全注意力层配备一个转换的线性注意力分支来构建一个可变形模型。然后，它冻结所有模型权重，并在合成的长上下文检索数据上联合优化逐层门控，采用线性化正则化以鼓励模型依赖线性注意力以提高效率。学习到的门控在预设的全注意力预算下被离散化，以实例化混合架构，随后进行标准的logits蒸馏和长上下文微调。大量实验表明，FlashMorph发现了更有效的混合配置，保持了强大的长上下文召回和一般基准性能，同时与现有层选择方法相比，显著降低了层选择成本，证明了其有效性、高效性和可扩展性。
cs.CL / 116 / 2606.30578
Uncertainty-Aware Generation and Decision-Making Under Ambiguity
不确定性感知生成与模糊下的决策制定
Abstract
With rapidly improving capabilities, Large Language Models (LLMs) are increasingly used in many complex real-world tasks. Beyond requiring in-depth knowledge and reasoning skills, many of these tasks exhibit a high degree of subjectivity and require that the outputs of the model can be trusted. While a lot of progress has been made to train better models, decision-making algorithms have received less attention. In this work, we present and evaluate various uncertainty-aware decision-making algorithms based on Bayesian decision theory and risk-averse decision making on the tasks of tutoring and automatic peer reviewing. Concretely, we take uncertainty over tutoring strategies and review scores into account when generating a tutor response or review and use conformal prediction to provide guarantees over strategy and score. We find empirically that these algorithms can improve the utility of the generations but need to be carefully implemented when ambiguity is high. For example, risk-averse rules can degrade performance by optimizing for generic outputs, while Bayesian methods tend to perform better. Our work uses techniques from decision theory to improve LLM-based decision-making and outlines open challenges for the community.
Chinese Translation
随着大型语言模型（LLMs）能力的快速提升，它们在许多复杂的现实任务中得到了越来越多的应用。这些任务不仅需要深入的知识和推理能力，还表现出高度的主观性，要求模型的输出能够被信任。尽管在训练更好的模型方面取得了许多进展，决策算法却受到的关注相对较少。在本研究中，我们提出并评估了基于贝叶斯决策理论和风险规避决策的各种不确定性感知决策算法，应用于辅导和自动同行评审任务。具体而言，我们在生成辅导响应或评审时考虑了辅导策略和评审分数的不确定性，并使用保形预测提供对策略和分数的保证。我们通过实证研究发现，这些算法能够提高生成的效用，但在模糊性较高时需要谨慎实施。例如，风险规避规则可能通过优化通用输出而降低性能，而贝叶斯方法往往表现更好。我们的研究利用决策理论中的技术来改善基于LLM的决策制定，并概述了该领域面临的开放挑战。
cs.CL / 117 / 2606.30616
Scaling the Horizon, Not the Parameters: Reaching Trillion-Parameter Performance with a 35B Agent
扩展视野，而非参数：通过35B智能体实现万亿参数级性能
Bai, Lei, Cao, Zongsheng, Chen, Yang, Cui, Zhiyao, Du, Shangheng, Fan, Yue, Feng, Shiyang, Guo, Zijie, He, Haonan, He, Liang, He, Xiaohan, Hu, Shuyue, Hu, Yusong, Huang, Songtao, Jiang, Yichen, Li, Hao, Li, Xin, Lin, Dahua, Lin, Weihao, Ling, Fenghua, Liu, Dongrui, Liu, Zhuo, Ma, Runmin, Mu, Chunjiang, Peng, Haoyang, Peng, Tianshuo, Shi, Jinxin, Shi, Luohe, Sun, Boyuan, Tan, Zelin, Tang, Shengji, Wang, Qianyi, Wu, Yiming, Xie, Yi, Yan, Xiangchao, Ye, Jingqi, Ye, Peng, Yu, Fangchen, Yuan, Jiakang, Zhan, Bihao, Zhang, Bo, Zhang, Chen, Zhang, Shufei, Zhang, Shuaiyu, Zhang, Wenlong, Zhang, Yiqun, Zhao, Junpeng, Zhong, Zhijie, Zhou, Bowen, Zhou, Yuhao
Abstract
We introduce Agents-A1, a 35B Mixture-of-Experts Agentic Model that reaches trillion-parameter-level performance by scaling the agent horizon. We investigate agent-horizon scaling from two perspectives: scaling long-horizon trajectories and scaling heterogeneous agent abilities. To support this goal, we build a long-horizon knowledge-action infrastructure that connects external knowledge, actions, observations, and verifier outcomes, producing agentic trajectories with an average length of 45K tokens. Based on this, we train Agents-A1 with a three-stage recipe. First, we perform full-domain supervised fine-tuning to align the base model with broad agentic behaviors. Second, we train domain-level teacher models to capture specialized expertise in each domain. Third, we propose a multi-teacher domain-routed on-policy distillation with salient vocabulary alignment to improve knowledge transfer efficiency across different domains, unifying six heterogeneous domains into one deployable student model. Agents-A1 achieves strong and broad performance for long-horizon agent benchmarks. Compared with 1T-parameter model such as Kimi-K2.6 and DeepSeek-V4-pro, Agents-A1 achieves leading results on SEAL-0 (56.4), IFBench (80.6), HiPhO (46.4), FrontierScience-Olympiad (79.0), and MolBench-Bind (56.8), and remains highly competitive on SciCode (44.3), HLE (47.6) and BrowseComp (75.5). We hope this work provides the community with a practical path for scaling the horizon using a 35B agent that can reach or match the performance of 1T models on long-horizon tasks.
Chinese Translation
我们介绍了Agents-A1，这是一种35B混合专家智能模型，通过扩展智能体的视野实现了万亿参数级的性能。我们从两个角度研究智能体视野的扩展：扩展长视野轨迹和扩展异构智能体能力。为了支持这一目标，我们建立了一个长视野知识-行动基础设施，连接外部知识、行动、观察和验证结果，生成平均长度为45K标记的智能轨迹。在此基础上，我们采用三阶段的训练方案来训练Agents-A1。首先，我们进行全领域的监督微调，以使基础模型与广泛的智能行为对齐。其次，我们训练领域级教师模型，以捕捉每个领域的专业知识。第三，我们提出了一种多教师领域路由的在线蒸馏方法，并通过显著词汇对齐来提高不同领域间的知识转移效率，将六个异构领域统一为一个可部署的学生模型。Agents-A1在长视野智能体基准测试中表现出色且广泛。与1T参数模型如Kimi-K2.6和DeepSeek-V4-pro相比，Agents-A1在SEAL-0（56.4）、IFBench（80.6）、HiPhO（46.4）、FrontierScience-Olympiad（79.0）和MolBench-Bind（56.8）上取得了领先结果，并在SciCode（44.3）、HLE（47.6）和BrowseComp（75.5）上保持高度竞争力。我们希望这项工作为社区提供了一条实用的路径，通过使用35B智能体扩展视野，从而在长视野任务上达到或匹配1T模型的性能。