cs.RO / 1 / 2606.12475
Learning to Assist: Collaborative VLAs for Implicit Human-Robot Collaboration
学习辅助：隐式人机协作的协作视觉-语言-动作（VLA）模型
Abstract
Human-robot collaboration (HRC) combines the complementary strengths of humans and robots to improve task efficiency. However, many existing collaborative systems rely on hand-engineered pipelines, limiting their scalability and flexibility for new tasks. In this work, we show that models trained end-to-end with imitation learning, specifically vision-language-action (VLA) models, can support collaborative manipulation, and characterize the key factors affecting their real-world performance. We evaluate two state-of-the-art models and identify a failure mode of action-chunking policies in implicit HRC, where demonstration action leakage (i.e., action chunks crossing latent task transitions) can cause premature assistive behavior. We find that this issue increases with longer execution horizons and occurs in real-world collaborative VLA systems, such as when a robot attempts to hand over a tool before the person is ready. We propose an inference-time steering method to mitigate these erroneous assistive actions while preserving policy performance. Finally, through a 16-participant user study on a long-horizon collaborative assembly task, we show that steering enables a longer execution horizon while mitigating premature assistance, leading to faster collaboration and fewer failures compared to a shorter-horizon policy.
Chinese Translation
人机协作（HRC）结合了人类和机器人互补的优势，以提高任务效率。然而，许多现有的协作系统依赖于手工设计的流程，这限制了它们在新任务中的可扩展性和灵活性。在本研究中，我们展示了通过模仿学习端到端训练的模型，特别是视觉-语言-动作（VLA）模型，可以支持协作操作，并表征影响其现实世界性能的关键因素。我们评估了两个最先进的模型，并识别出隐式人机协作中行动分块策略的一种失效模式，其中演示动作泄漏（即，动作块跨越潜在任务转变）可能导致过早的辅助行为。我们发现，这一问题在较长的执行时间范围内会增加，并且在现实世界的协作VLA系统中发生，例如当机器人试图在人员准备好之前递交工具时。我们提出了一种推理时引导方法，以减轻这些错误的辅助动作，同时保持策略性能。最后，通过对16名参与者进行的长时间协作组装任务的用户研究，我们展示了引导能够实现更长的执行时间范围，同时减轻过早的辅助行为，从而与短时间范围策略相比，促进更快的协作和更少的失败。
cs.RO / 2 / 2606.12499
Action-Effect Memory Pretraining for Robot Manipulation
机器人操作的动作-效果记忆预训练
Abstract
We present AEM, an Action-Effect Memory pretraining framework for robot manipulation that learns compact temporal representations from vision-action history. Unlike prior robot representation pretraining methods that mainly focus on single-frame visual encoding, AEM targets the temporal nature of manipulation, where the current observation alone is often insufficient under partial observability. AEM models manipulation as an action-driven interaction process by interleaving visual and action features and applying masked modeling to recover missing content from incomplete histories, thereby learning action-conditioned state evolution. The Mamba-encoded output of the final vision token is used as a compact history representation, serving as the global context for decoding and downstream control. This design preserves a single-vector temporal bottleneck while keeping inference efficient. We evaluate AEM with Diffusion Policy and Flow Policy. AEM consistently improves manipulation performance in both simulation and real-world settings, outperforming baselines across clean scenes, cluttered and random scenes, and non-Markovian tasks. Ablation studies further show that history-aware pretraining surpasses single-frame pretraining and direct frame stacking, while reducing inference latency and computational cost.
Chinese Translation
我们提出了AEM（动作-效果记忆）预训练框架，用于机器人操作，通过视觉-动作历史学习紧凑的时间表示。与以往主要关注单帧视觉编码的机器人表示预训练方法不同，AEM针对操作的时间特性，因为在部分可观测性下，仅凭当前观察往往不足以完成任务。AEM将操作建模为一个以动作驱动的交互过程，通过交错视觉和动作特征，并应用掩蔽建模来恢复不完整历史中的缺失内容，从而学习动作条件下的状态演变。最终视觉标记的Mamba编码输出被用作紧凑的历史表示，作为解码和下游控制的全局上下文。该设计在保持推理高效的同时，保留了单向量时间瓶颈。我们使用扩散策略（Diffusion Policy）和流策略（Flow Policy）对AEM进行了评估。AEM在模拟和现实环境中始终提高了操作性能，在干净场景、杂乱和随机场景以及非马尔可夫任务中均超越了基线。消融研究进一步表明，历史感知的预训练优于单帧预训练和直接帧堆叠，同时降低了推理延迟和计算成本。
cs.RO / 3 / 2606.12550
Foresight: Iterative Reasoning About Clues that Matter for Navigation
前瞻性：关于导航中重要线索的迭代推理
Abstract
Open-world mapless navigation from sparse language instructions requires resolving underspecified goals and inferring which environmental cues are relevant for reaching the goal. For instance, reaching an out-of-view destination may require interpreting ramps, signs, or detours that reveal where to go or which route to take. Prior works are limited by their reliance on known navigation factors and closed-set factor categories, or identify cues before motion planning and miss plan-dependent cues. We argue that pretrained Vision-Language Models (VLMs) can discover novel instruction-relevant cues, but require adaptation to focus on which cues matter and how they should influence motion planning. We realize these ideas in Foresight, a test-time framework in which a finetuned VLM alternates between proposing image-space motion plans and critiquing them using the language goal and visual context. Subsequent plans are conditioned on prior critiques, enabling iterative motion refinement before execution. To align plan critiques and refinements with open-set behavior preferences, we learn a reward model from human feedback and use it to post-train the VLM with reinforcement learning in the plan-critique loop. In offline evaluations and 6 real-world environments, Foresight improves average task success by 37% and reduces interventions per mission by 52% relative to state-of-the-art test-time reasoning and foundation-model baselines, while running in real-time on a Jetson AGX Orin. We will release code, data, and training details to support future work on test-time reasoning for robot motion refinement. Additional videos at: https://amrl.cs.utexas.edu/foresight
Chinese Translation
从稀疏语言指令进行开放世界无地图导航需要解决不明确的目标，并推断哪些环境线索与实现目标相关。例如，抵达一个视野之外的目的地可能需要解释坡道、标志或绕行，这些都能揭示前往何处或选择哪条路线。先前的研究受限于对已知导航因素和封闭集合因素类别的依赖，或在运动规划之前识别线索而错过依赖于规划的线索。我们认为，预训练的视觉-语言模型（Vision-Language Models, VLMs）能够发现新的与指令相关的线索，但需要适应以关注哪些线索重要以及它们应如何影响运动规划。我们在前瞻性（Foresight）中实现了这些思想，这是一个测试时框架，其中经过微调的VLM在提出图像空间运动计划和使用语言目标及视觉上下文对其进行评估之间交替进行。后续计划基于先前的评估进行条件化，使得在执行之前能够进行迭代的运动优化。为了将计划评估和优化与开放集合行为偏好对齐，我们从人类反馈中学习了一个奖励模型，并利用它在计划-评估循环中使用强化学习对VLM进行后训练。在离线评估和6个真实环境中，前瞻性使平均任务成功率提高了37%，并将每次任务的干预次数减少了52%，相较于最先进的测试时推理和基础模型基线，同时在Jetson AGX Orin上实时运行。我们将发布代码、数据和训练细节，以支持未来在机器人运动优化的测试时推理方面的研究。更多视频请访问：https://amrl.cs.utexas.edu/foresight
cs.RO / 4 / 2606.12579
G-MAPP: GPU-accelerated Multi-Agent Planning and Perception for Reactive Motion Generation
G-MAPP：基于GPU的多智能体规划与感知加速反应运动生成
Abstract
Reactive motion generation in unstructured environments remains an open challenge in robotics. Due to the computational complexity of collision-free motion generation, existing methods either generate global trajectories for static scenarios, or employ models that make conservative assumptions about the environment. This paper identifies the primary bottleneck as the runtime performance demand of planning on high-fidelity environments, and the temporal integration between the perception and planning modules. Therefore, we propose a framework that does not compromise on runtime performance and world representations for perception and planning by accelerating world modeling and vector-field based planning using the GPU. This allows us to achieve faster parallel state exploration for quasi-global trajectory planning, and tighter coupling of the perception-action loop in real-time for dynamic cluttered environments with off-the-shelf depth sensors. We quantitatively evaluate the computation-time and success rate differences for the CPU and GPU versions of our planner, and perform qualitative evaluations of our coupled framework using real-world experiments on a 7-DoF Franka Emika robot. Experimental results demonstrate that our GPU-based framework achieves up to a 5x speedup over the CPU version and successfully avoids collisions across both trivial and challenging physical world scenarios.
Chinese Translation
在非结构化环境中生成反应性运动仍然是机器人技术中的一个开放挑战。由于无碰撞运动生成的计算复杂性，现有方法要么为静态场景生成全局轨迹，要么采用对环境做出保守假设的模型。本文确定了主要瓶颈为在高保真环境中规划的运行时性能需求，以及感知与规划模块之间的时间集成。因此，我们提出了一种框架，该框架在感知和规划的世界表示上不妥协于运行时性能，通过使用GPU加速世界建模和基于向量场的规划。这使我们能够实现更快的并行状态探索，以进行准全局轨迹规划，并在动态杂乱环境中与现成深度传感器实时紧密耦合感知-动作循环。我们定量评估了规划器的CPU和GPU版本在计算时间和成功率上的差异，并通过在7自由度的Franka Emika机器人上进行真实世界实验，对我们的耦合框架进行了定性评估。实验结果表明，我们基于GPU的框架在速度上比CPU版本快了多达5倍，并成功避免了在简单和具有挑战性的物理世界场景中的碰撞。
cs.RO / 5 / 2606.12603
From Imitation to Alignment: Human-Preference Flow Policies for Long-Horizon Sidewalk Navigation
从模仿到对齐：用于长距离人行道导航的人类偏好流政策
Abstract
Autonomous long-horizon sidewalk navigation is essential for micro-mobility applications such as robotic food delivery and assistive electronic wheelchairs. Unlike autonomous driving on the road, long-horizon sidewalk navigation requires precise maneuvering through unpredictable sidewalk terrains and pedestrians, with a lightweight perception stack as minimal as a single monocular RGB camera. While imitation learning (IL) from demonstrations offers a practical solution, the resulting autopilot policy often suffers from compounding errors, a lack of social compliance on sidewalks, and deficiencies in counterfactual reasoning to handle complex situations. To address these challenges, we introduce FlowPilot, a mapless navigation policy that achieves robust and efficient long-horizon navigation performance using only a monocular RGB camera. We first propose to use anchored flow matching as an action representation for policy pre-training on large-scale robot fleet data and to capture the diverse, complex, multimodal distribution of sidewalk navigation behaviors. To bridge the gap between imitation and alignment, we further design a human-in-the-loop preference learning scheme to tune the policy on a small amount of human intervention data. It strengthens the model's counterfactual reasoning and social compliance on sidewalks. We evaluate FlowPilot through extensive simulation and real-world experiments in diverse sidewalk environments. FlowPilot achieves 42% success rate and 66% route completion in simulation, while FlowPilot-HP further improves real-world robustness and social compliance, reducing IR by 40.0% and NIR by 52.1% relative to the base model.
Chinese Translation
自主的长距离人行道导航对于微型出行应用（如机器人送餐和辅助电子轮椅）至关重要。与道路上的自动驾驶不同，长距离人行道导航需要在不可预测的人行道地形和行人之间进行精确操控，且感知系统的要求最低可仅为单个单目RGB相机。虽然从示范中进行模仿学习（IL）提供了一种实用的解决方案，但所产生的自动驾驶政策往往面临累积误差、缺乏在人行道上的社会合规性以及在处理复杂情况时的反事实推理不足等问题。为了解决这些挑战，我们提出了FlowPilot，这是一种无地图导航政策，仅使用单目RGB相机即可实现稳健和高效的长距离导航性能。我们首先建议使用锚定流匹配作为政策预训练的大规模机器人舰队数据的动作表示，以捕捉人行道导航行为的多样、复杂和多模态分布。为了弥合模仿与对齐之间的差距，我们进一步设计了一种人机协作的偏好学习方案，以在少量人类干预数据上调整政策。这增强了模型在处理人行道上的反事实推理和社会合规性。我们通过在多样的人行道环境中进行广泛的仿真和现实世界实验来评估FlowPilot。FlowPilot在仿真中实现了42%的成功率和66%的路线完成率，而FlowPilot-HP进一步提高了现实世界的稳健性和社会合规性，相较于基础模型，IR降低了40.0%，NIR降低了52.1%。
cs.RO / 6 / 2606.12604
EgoEngine: From Egocentric Human Videos to High-Fidelity Dexterous Robot Demonstrations
EgoEngine：从自我中心人类视频到高保真灵巧机器人演示
Abstract
Dexterous manipulation is limited by the cost of collecting large-scale robot demonstrations. Egocentric human videos offer a scalable source of diverse manipulation behaviors, but directly using them for robot learning requires bridging two gaps: the visual gap between human and robot observations, and the action gap between human motion and robot-executable action. We propose EgoEngine, a scalable framework for transforming egocentric human manipulation videos into high-fidelity robot data. Given an egocentric RGB video, EgoEngine produces: (i) a high-fidelity robot observation video replacing human with robot while preserving scene context and temporal alignment, and (ii) a task-aligned, executable robot action trajectory under feasibility constraints. Experiments in simulation and on real robots show that EgoEngine enables scalable conversion of human videos into robot data and, to our knowledge, demonstrates the first zero-shot visuomotor dexterous policy learning from egocentric human videos without real-robot demonstrations. Project website: https://egoengine.github.io.
Chinese Translation
灵巧操作受到收集大规模机器人演示成本的限制。自我中心的人类视频提供了多样化操作行为的可扩展来源，但直接将其用于机器人学习需要弥补两个差距：人类和机器人观察之间的视觉差距，以及人类运动与机器人可执行动作之间的动作差距。我们提出了EgoEngine，一个将自我中心人类操作视频转化为高保真机器人数据的可扩展框架。给定一段自我中心的RGB视频，EgoEngine生成：(i) 一段高保真的机器人观察视频，替换人类为机器人，同时保持场景上下文和时间对齐；(ii) 在可行性约束下的任务对齐、可执行的机器人动作轨迹。模拟和真实机器人上的实验表明，EgoEngine能够将人类视频可扩展地转换为机器人数据，并且据我们所知，首次展示了从自我中心人类视频中进行零-shot视觉运动灵巧策略学习，而无需真实机器人演示。项目网站：https://egoengine.github.io。
cs.RO / 7 / 2606.12614
DARRMS -- An Efficient Algorithm for Dynamic Attention Radius in Resource-Constrained Multi-Agent Systems
DARRMS -- 一种用于资源受限多智能体系统的动态注意半径高效算法
Abstract
Multi-agent systems are integral tools for various domains such as robotics, cybersecurity, and autonomous vehicle planning. These types of systems often have constraints on the computational resources, leading to a need for efficient lightweight algorithms. Traditional decision making frameworks often assume ideal conditions, such as full observability and unlimited computational capacity, which do not align with real-world challenges. In this paper, we introduce a new algorithm that allows for reduced demand on computational resources without a large cost of other performance metrics. Agents will limit their observability to some attention radius, which intentionally allows them to ignore parts of the environment that might be unnecessary for action planning. By optimizing both the attention radius and decision-making, our approach enhances coordination and scalability in uncertain environments. Through both theoretical analysis and empirical validation, we demonstrate the effectiveness of adaptive observation in improving system performance and maintaining robust decision-making strategies in resource-constrained systems.
Chinese Translation
多智能体系统是机器人技术、网络安全和自主车辆规划等多个领域的重要工具。这类系统通常受到计算资源的限制，因此需要高效的轻量级算法。传统的决策框架通常假设理想条件，如完全可观测性和无限计算能力，这与现实世界的挑战并不一致。在本文中，我们提出了一种新算法，允许在不显著增加其他性能指标成本的情况下减少对计算资源的需求。智能体将其可观测性限制在一定的注意半径内，这使得它们可以有意地忽略环境中可能对行动规划不必要的部分。通过优化注意半径和决策过程，我们的方法增强了在不确定环境中的协调性和可扩展性。通过理论分析和实证验证，我们展示了自适应观察在提高系统性能和维持资源受限系统中的稳健决策策略方面的有效性。
cs.RO / 8 / 2606.12690
EWAM: An Enhanced World Action Model for Closed-Loop Online Adaptation in Embodied Intelligence
EWAM：一种用于具身智能闭环在线适应的增强世界行动模型
Abstract
In this paper, we propose the Enhanced World Action Model (EWAM), a closed-loop online adaptation architecture built upon a pretrained and fully frozen Cosmos3 backbone network. Evaluated entirely under a zero-shot task protocol, EWAM is centrally focused on reducing the amount of additional deployment data required to adapt to new task layouts. Notably, no extra task-specific demonstration sets were introduced in any of the evaluations, and no fine-tuning was performed on the backbone network. Its performance gains stem entirely from an inference-time co-reasoning mechanism composed of four inserted lightweight neural layers: the Neural Experience Memory Layer located in the intermediate layers of the Diffusion Transformer (DiT) provides task-relevant execution context; the Neural Anomaly Detection Layer after the state prediction head monitors the divergence between predicted and actual states in real time; the Neural Policy Routing Layer dynamically selects direct execution, conservative replanning, or rollback recovery based on the anomaly severity; and the Neural Action Correction Layer refines the generated action chunks using execution diagnostics. Unlike naive feature fusion, the memory, anomaly detection, and correction modules are deeply integrated into the Cosmos3 forward path in a differentiable manner, with only the final routing decision being a discrete supervised one.
Chinese Translation
本文提出了增强世界行动模型（EWAM），这是一种基于预训练且完全冻结的Cosmos3骨干网络构建的闭环在线适应架构。EWAM在零样本任务协议下进行全面评估，主要集中于减少适应新任务布局所需的额外部署数据量。值得注意的是，在任何评估中都没有引入额外的任务特定演示集，也没有对骨干网络进行微调。其性能提升完全源于一种推理时的共同推理机制，该机制由四个插入的轻量级神经层组成：位于扩散变换器（Diffusion Transformer, DiT）中间层的神经经验记忆层提供与任务相关的执行上下文；状态预测头后的神经异常检测层实时监控预测状态与实际状态之间的偏差；神经策略路由层根据异常严重性动态选择直接执行、保守重规划或回滚恢复；神经行动修正层利用执行诊断来优化生成的行动块。与简单的特征融合不同，记忆、异常检测和修正模块以可微分的方式深度集成到Cosmos3的前向路径中，只有最终的路由决策是离散的监督决策。
cs.RO / 9 / 2606.12728
EquiDexFlow: Contact-Grounded SE(3)-Equivariant Dexterous Grasp Generative Flows
EquiDexFlow：基于接触的 SE(3) 等变灵巧抓取生成流
Abstract
Most learned dexterous grasp generators relegate contact forces to a downstream verification step, so a kinematically-plausible pose can still violate the conditions for a stable physical grasp. We address this with EquiDexFlow, an SE(3)-equivariant flow-matching model that jointly predicts wrist pose, joint angles, fingertip contacts, surface normals, and contact forces from an object point cloud. Our architecture projects contacts onto the object surface and forces into the Coulomb friction cone by construction, so placement and friction compliance hold without loss penalties. We prove end-to-end SE(3) equivariance and verify it empirically over 200 rotations, with wrist residuals below $0.04^\circ$ and exactly zero joint deviation. Trained on 8,100 force-closure grasps across 81 objects for the 16-DoF Allegro Hand, our model achieves zero friction violations, the best composite score, and the lowest wrench residual among all ablation variants. We retarget decoded fingertip contacts to a 16-DoF LEAP Hand via per-finger inverse kinematics, and our hardware-feasible refinement places every joint at least 5% inside its actuator envelope while preserving wrench balance. On the physical robot, retargeted EquiDexFlow-decoded grasps complete open-loop pick-and-hold trials on all six test objects, with every asymmetric object succeeding at both the canonical pose and a $120^\circ$ co-rotation. Videos, code, and checkpoints are available at https://equidexflow.github.io.
Chinese Translation
大多数学习的灵巧抓取生成器将接触力 relegated 到下游验证步骤，因此一个运动学上合理的姿态仍然可能违反稳定物理抓取的条件。我们通过 EquiDexFlow 解决了这个问题，这是一种 SE(3) 等变流匹配模型，能够从物体点云中联合预测腕部姿态、关节角度、指尖接触、表面法线和接触力。我们的架构通过构造将接触投影到物体表面，并将力投影到库仑摩擦锥中，因此放置和摩擦合规性在没有损失惩罚的情况下得以保持。我们证明了端到端的 SE(3) 等变性，并在 200 次旋转中进行了实证验证，腕部残差低于 $0.04^ heta$，关节偏差完全为零。我们的模型在 81 个物体上训练了 8,100 个力闭合抓取，针对 16 自由度的 Allegro 手，实现了零摩擦违规、最佳综合得分和所有消融变体中最低的扭矩残差。我们通过每指逆运动学将解码的指尖接触重新定向到 16 自由度的 LEAP 手，并且我们的硬件可行的优化将每个关节至少放置在其驱动器包络内的 5%，同时保持扭矩平衡。在物理机器人上，重新定向的 EquiDexFlow 解码抓取在所有六个测试物体上完成了开环拾取和保持试验，每个不对称物体在标准姿态和 $120^ heta$ 的共同旋转下均成功。视频、代码和检查点可在 https://equidexflow.github.io 获取。
cs.RO / 10 / 2606.12759
Sparse2Act: Learning Action-Aligned Sparse 3D Representations for Cross-Domain Robot Manipulation
Sparse2Act：学习与动作对齐的稀疏三维表示以实现跨领域机器人操作
Abstract
Explicit 3D representations are attractive for manipulation because they expose object shape, workspace geometry, and robot-object relations in metric coordinates. However, sparse 3D encoders are often learned through downstream task objectives, tying the representation to a particular data distribution, policy architecture, and action parameterization. We introduce Sparse2Act, an observation-action alignment framework for pretraining sparse point-cloud encoders. The key idea is to use task-space end-effector actions as geometric supervision: masked sparse 3D tokens are trained to organize scene features around the workspace motion paired with the observation. After pretraining, only the encoder initialization is reused by downstream policies, allowing them to retain their own architectures and action spaces, including joint-space commands. On the LIBERO-10 benchmark, our method achieves 86.9% average success after 500 fine-tuning steps. The same pretrained encoder supports LIBERO-to-Meta-World cross-domain transfer, achieving 73.4% average success on the Meta-World-5 benchmark. Ablations on the objective and decoder capacity show that the gains come from the masked action-alignment signal and remain useful across downstream action decoders. In real-world experiments, simulation pretraining followed by limited real-data fine-tuning achieves an average success rate of 72.5% across four tasks, demonstrating effective sim-to-real transfer. These results suggest that robot actions can provide compact geometric supervision for reusable sparse 3D representations.
Chinese Translation
显式的三维表示在操作中具有吸引力，因为它们以度量坐标展示了物体形状、工作空间几何和机器人与物体之间的关系。然而，稀疏三维编码器通常通过下游任务目标进行学习，这使得表示与特定的数据分布、策略架构和动作参数化相绑定。我们提出了Sparse2Act，一个用于预训练稀疏点云编码器的观察-动作对齐框架。其关键思想是使用任务空间末端执行器动作作为几何监督：被遮蔽的稀疏三维标记被训练以围绕与观察配对的工作空间运动组织场景特征。预训练后，仅重用编码器初始化供下游策略使用，允许它们保留自己的架构和动作空间，包括关节空间命令。在LIBERO-10基准测试中，我们的方法在500次微调步骤后实现了86.9%的平均成功率。相同的预训练编码器支持LIBERO到Meta-World的跨领域迁移，在Meta-World-5基准测试中实现了73.4%的平均成功率。关于目标和解码器容量的消融实验表明，增益来自于被遮蔽的动作对齐信号，并在下游动作解码器中保持有效。在真实世界实验中，模拟预训练后进行有限的真实数据微调，在四个任务中实现了72.5%的平均成功率，展示了有效的模拟到现实迁移。这些结果表明，机器人动作可以为可重用的稀疏三维表示提供紧凑的几何监督。
cs.RO / 11 / 2606.12814
Stubborn: A Streamlined and Unified Reinforcement Learning Framework for Robust Motion Tracking and Fall Recovery for Humanoids
Stubborn：一个简化且统一的强化学习框架，用于人形机器人稳健的运动跟踪和跌倒恢复
Abstract
Recent reinforcement learning approaches have shown great promise in improving humanoid motion tracking performance and achieving fall recovery under disturbances. However, most existing works treat motion tracking and fall recovery as different tasks and require multi-stage training with specialized recovery rewards and/or separate recovery policies. Moreover, existing reinforcement learning-based methods often terminate training episodes immediately after severe tracking failures, limiting recovery-oriented exploration in unstable or fallen states. To address the above issues, we propose Stubborn, a streamlined and unified reinforcement learning framework to achieve robust humanoid motion tracking and fall recovery. Specifically, Stubborn uses an asymmetric Actor-Critic architecture and consists of three major components. First, a yaw-aligned tracking representation is adopted to reduce sensitivity to global drift and heading disturbances while preserving gravity-related balance information. Second, we introduce a Bernoulli-based probabilistic termination mechanism that enables the policy to encourage exploration of fall-recovery behaviors under varying failure modes. Third, we propose a probabilistic termination and tracking-error-driven strategy that dynamically reshapes the sampling distribution based on tracking performance, increasing the training efficiency for difficult motion segments and unstable states. Extensive comparisons with SOTA methods and ablation studies show that Stubborn achieved competitive performance, and the proposed probabilistic termination mechanism and adaptive sampling strategy contributed to the performance and robustness gains. For real-world demonstrations, please refer to https://aislab-sustech.github.io/Stubborn/.
Chinese Translation
近期的强化学习方法在改善人形机器人运动跟踪性能和在干扰下实现跌倒恢复方面展现出了巨大的潜力。然而，大多数现有研究将运动跟踪和跌倒恢复视为不同的任务，并需要多阶段训练，配合专门的恢复奖励和/或单独的恢复策略。此外，现有的基于强化学习的方法通常在严重的跟踪失败后立即终止训练阶段，限制了在不稳定或跌倒状态下的恢复导向探索。为了解决上述问题，我们提出了Stubborn，一个简化且统一的强化学习框架，以实现稳健的人形机器人运动跟踪和跌倒恢复。具体而言，Stubborn采用了不对称的Actor-Critic架构，并由三个主要组件组成。首先，采用了与偏航对齐的跟踪表示，以减少对全局漂移和航向干扰的敏感性，同时保留与重力相关的平衡信息。其次，我们引入了一种基于伯努利分布的概率终止机制，使得策略能够在不同的失败模式下鼓励探索跌倒恢复行为。第三，我们提出了一种基于概率终止和跟踪误差驱动的策略，动态重塑采样分布，基于跟踪性能提高困难运动段和不稳定状态的训练效率。与最先进的方法进行的广泛比较和消融研究表明，Stubborn实现了具有竞争力的性能，所提出的概率终止机制和自适应采样策略对性能和稳健性提升起到了重要作用。有关实际演示，请参见 https://aislab-sustech.github.io/Stubborn/。
cs.RO / 12 / 2606.12859
AIR-VLA+: Decoupling Movement and Manipulation via Cascaded Dual-Action Decoders with Asymmetric MoE for Aerial Robots
AIR-VLA+: 通过级联双动作解码器与不对称专家混合模型解耦运动与操作以适应空中机器人
Abstract
Aerial manipulation systems have long suffered from representation coupling in end-to-end control, as platform-level Unmanned Aerial Vehicle (UAV) movement and end-effector-level arm manipulation differ substantially in action scale, dynamics, and control objectives. In this paper, we propose AIR-VLA+, a flow matching action generation architecture specifically designed for aerial manipulation, featuring cascaded dual-action decoders and an asymmetric feature-level Mixture of Experts (MoE). We construct cascaded manipulation and movement decoders, allowing the UAV to unidirectionally observe the manipulator's intent during movement to achieve workflow coordination, while isolating the impact of UAV movement information backpropagation on arm manipulation stability. Addressing the characteristic that UAV movement is highly dependent on high-level semantics and responsible for task state transitions in aerial manipulation, we design an input feature enhancement module for the UAV movement decoder. This module introduces an implicit visual grasp projector to perceive the interaction state between the gripper and the object, and injects compressed global semantic features. Within the UAV movement decoder, we deploy an implicit MoE architecture, enabling different movement experts to spontaneously exhibit capacity inclinations for various task stages during training. Through dense soft blending computation on the feature manifold, the UAV movement is endowed with stronger task-stage adaptability. Experiments on the standardized AIR-VLA benchmark demonstrate that our method comprehensively surpasses all baselines with an overall average score of 48.0. The overall task completion score improves by 80.2\% compared to the single-head $\pi_{0.5}$ policy, effectively mitigating the heterogeneous coordinated control conflicts of composite robots.
Chinese Translation
空中操作系统长期以来在端到端控制中受到表示耦合的困扰，因为平台级无人机（UAV）运动与末端执行器级臂操作在动作规模、动力学和控制目标上存在显著差异。本文提出了AIR-VLA+，一种专为空中操作设计的流匹配动作生成架构，具有级联双动作解码器和不对称特征级专家混合模型（MoE）。我们构建了级联的操作和运动解码器，使UAV在运动过程中单向观察操纵器的意图，以实现工作流协调，同时隔离UAV运动信息反向传播对臂操作稳定性的影响。针对UAV运动高度依赖高层语义并负责空中操作任务状态转变的特性，我们为UAV运动解码器设计了输入特征增强模块。该模块引入了隐式视觉抓取投影器，以感知夹具与物体之间的交互状态，并注入压缩的全局语义特征。在UAV运动解码器中，我们部署了隐式MoE架构，使不同的运动专家在训练过程中自发地表现出对不同任务阶段的能力倾向。通过对特征流形进行密集的软混合计算，UAV运动被赋予了更强的任务阶段适应性。在标准化的AIR-VLA基准测试中的实验表明，我们的方法在整体平均得分上全面超越了所有基线，得分为48.0。与单头$ ext{π}_{0.5}$策略相比，整体任务完成得分提高了80.2\%，有效缓解了复合机器人异构协调控制冲突。
cs.RO / 13 / 2606.12890
Learning to Adapt: Representation-Based Reinforcement Learning for Multi-Task Skill Transfer
学习适应：基于表征的多任务技能迁移强化学习
Abstract
Reinforcement learning has achieved remarkable success in learning complex control policies, yet its applicability remains limited due to sample inefficiency and poor generalization across tasks. In this work, we propose RepMT-SAC, a framework for multi-task RL that enables efficient knowledge sharing and robust transfer to new tasks. RepMT-SAC uses spectral MDP decomposition to capture transferable dynamics, structuring the value function into a task-agnostic core with a minimal task-specific adjustment. This design allows for strong zero-shot performance on in-distribution tasks and rapid few-shot adaptation to out-of-distribution tasks. We evaluate RepMT-SAC on quadcopter trajectory-following tasks across in-distribution and out-of-distribution contexts, demonstrating that it outperforms baselines by up to 30%.
Chinese Translation
强化学习在学习复杂控制策略方面取得了显著成功，但由于样本效率低和任务间泛化能力差，其适用性仍然有限。在本研究中，我们提出了RepMT-SAC，一个多任务强化学习框架，能够实现高效的知识共享和对新任务的稳健迁移。RepMT-SAC利用谱MDP分解来捕捉可迁移的动态，将价值函数结构化为一个与任务无关的核心，并进行最小的任务特定调整。这种设计使得在同分布任务上具有强大的零-shot性能，并能够快速适应异分布任务的少量样本。我们在同分布和异分布背景下对四旋翼轨迹跟踪任务评估了RepMT-SAC，结果表明其性能比基线提高了多达30%。
cs.RO / 14 / 2606.12910
Bounding Boxes as Goals: Language-Conditioned Grasping via Neuro-Symbolic Planning
将边界框作为目标：通过神经符号规划实现语言条件的抓取
Abstract
For robotics to be effectively integrated into household or industrial environments, machines must adapt to natural-language prompts in real time. Although Vision-Language Models (VLMs) have enabled zero-shot generalization in robot task and motion planning (TAMP), current state-of-the-art approaches often remain computationally "heavyweight" or require extensive training on thousands of demonstrations. We present GRASP (Grounded Reasoning and Symbolic Planning), a framework designed as a step toward open-vocabulary tabletop manipulation. Our approach leverages a pretrained VLM to translate natural-language queries into neuro-symbolic goal states, grounded in the physical world via a bounding-box detection pipeline. Unlike methods that rely on fixed color lists or hard-coded coordinates, GRASP enables robots to interpret abstract spatial concepts such as "top shelf" and execute tasks without additional fine-tuning. We achieve 73.3% overall success across 90 real-robot trials at three difficulty levels, requiring no task-specific training.
Chinese Translation
为了使机器人能够有效地融入家庭或工业环境，机器必须实时适应自然语言提示。尽管视觉语言模型（Vision-Language Models, VLMs）已经在机器人任务和运动规划（Task and Motion Planning, TAMP）中实现了零样本泛化，但当前的最先进方法往往计算负担“沉重”或需要在数千个示例上进行广泛训练。我们提出了GRASP（基于实证的推理与符号规划），这是一个旨在朝着开放词汇桌面操作迈进的框架。我们的方法利用预训练的VLM将自然语言查询转换为基于神经符号的目标状态，通过边界框检测管道在物理世界中进行定位。与依赖固定颜色列表或硬编码坐标的方法不同，GRASP使机器人能够理解诸如“顶部架子”等抽象空间概念，并在无需额外微调的情况下执行任务。我们在三个难度级别的90次真实机器人试验中实现了73.3%的整体成功率，无需特定任务的训练。
cs.RO / 15 / 2606.12936
An Embodied Simulation Platform, Benchmark, and Data-Efficient Augmentation Framework for Wet-Lab Robotics
一个用于湿实验室机器人学习的具身仿真平台、基准测试和数据高效增强框架
Abstract
Wet-lab robots can improve the reproducibility, throughput, and safety of biomedical experiments, but scaling their learning requires customizable simulators for safe and reproducible task generation, open editable laboratory assets, and efficient pipelines that turn limited demonstrations into usable training data. We present Pipette, an embodied simulation platform, benchmark, and data-efficient augmentation framework for wet-lab robot learning. Pipette releases over 43 open-source and re-editable wet-lab assets, together with an extensible asset-building pipeline. A key component of Pipette is its simulation-based data augmentation pipeline, replaying human demonstrations in simulation, applies lighting, camera, speed, and action perturbations, and filters generated episodes with automatic task success checks, rapidly expanding usable training data from limited manual demonstrations. We further introduce an 11-task wet-lab embodied benchmark covering sample handling, culture-ware manipulation, device operation, and precision placement. With only 30 demonstrations per task, ACT achieves 65.5% average success rate, while simulation augmentation improves SmolVLA from 44.1% to 74.7% and {\pi}0 from 40.4% to 46.5%, validating the effectiveness of Pipette for data-efficient VLA training and evaluation. Pipette also supports natural-language-driven scene construction and task registration, lowering the barrier for non-expert users to define new wet-lab robotic tasks.
Chinese Translation
湿实验室机器人可以提高生物医学实验的可重复性、通量和安全性，但要扩展其学习能力，需要可定制的仿真器以安全和可重复的方式生成任务、开放可编辑的实验室资产，以及高效的管道将有限的演示转化为可用的训练数据。我们提出了Pipette，一个用于湿实验室机器人学习的具身仿真平台、基准测试和数据高效增强框架。Pipette发布了超过43个开源和可重新编辑的湿实验室资产，以及一个可扩展的资产构建管道。Pipette的一个关键组成部分是其基于仿真的数据增强管道，该管道在仿真中重放人类演示，应用光照、相机、速度和动作扰动，并通过自动任务成功检查过滤生成的情节，快速从有限的手动演示中扩展可用的训练数据。我们进一步介绍了一个涵盖样本处理、培养器具操作、设备操作和精确放置的11任务湿实验室具身基准。每个任务仅需30个演示，ACT实现了65.5%的平均成功率，而仿真增强将SmolVLA的成功率从44.1%提高到74.7%，将{ ext{π}}0的成功率从40.4%提高到46.5%，验证了Pipette在数据高效VLA训练和评估中的有效性。Pipette还支持基于自然语言的场景构建和任务注册，降低了非专家用户定义新湿实验室机器人任务的门槛。
cs.RO / 16 / 2606.12954
Towards Reliable Sequential Object Picking in Clutter: The Runner-up Solution to RGMC 2025
朝着可靠的杂乱环境中顺序物体抓取：RGMC 2025 的亚军解决方案
Abstract
As a long-standing challenge in robotic manipulation, stable and efficient grasping in cluttered environments is of great importance in industrial settings. While recent studies have achieved relatively high success rates in grasping from clutter, there remain few mature solutions for more demanding tasks such as sequential object search and sorting. This work addresses sequential object picking in cluttered environments based on the Cluttered Environment Picking Benchmark (CEPB) and presents our solution to the Pick-in-Clutter track of the 10th Robotic Grasping and Manipulation Competition (RGMC) at ICRA 2025. The task poses several key challenges. First, it requires robust and collision-aware grasping with high success rates across a diverse set of objects, including both rigid and deformable ones. Second, it demands efficient search for target objects, which places stringent requirements on the decluttering and searching strategies of the solution. To address the above challenges, we design an integrated hardware-software pipeline that combines object recognition, decluttering, and multi-modal grasping. The main contributions include the hardware design of a multifunctional gripper and novel representations for object distribution and occlusion relationships in cluttered space. This pipeline enables efficient recognition, search, and sequential grasping of objects in clutter, demonstrating strong performance in both laboratory tests and competition scenarios, and ultimately achieving second place in the Pick-in-Clutter track of the RGMC 2025.
Chinese Translation
作为机器人操作中的一个长期挑战，在杂乱环境中实现稳定和高效的抓取在工业环境中具有重要意义。尽管近期研究在杂乱中抓取的成功率相对较高，但对于更具挑战性的任务，如顺序物体搜索和排序，仍然缺乏成熟的解决方案。本研究基于杂乱环境抓取基准（Cluttered Environment Picking Benchmark，CEPB）解决了杂乱环境中的顺序物体抓取问题，并提出了我们在2025年国际机器人抓取与操作大会（ICRA 2025）第十届机器人抓取与操作竞赛（Robotic Grasping and Manipulation Competition，RGMC）中“杂乱中抓取”赛道的解决方案。该任务面临几个关键挑战。首先，它要求在多样化物体（包括刚性和可变形物体）中实现稳健且具有碰撞意识的抓取，且成功率高。其次，它要求高效搜索目标物体，这对解决方案的去杂和搜索策略提出了严格要求。为应对上述挑战，我们设计了一个集成的硬件-软件管道，结合了物体识别、去杂和多模态抓取。主要贡献包括多功能抓手的硬件设计以及在杂乱空间中物体分布和遮挡关系的新颖表示。该管道实现了在杂乱环境中物体的高效识别、搜索和顺序抓取，在实验室测试和竞赛场景中表现出色，最终在RGMC 2025的“杂乱中抓取”赛道中获得第二名。
cs.RO / 17 / 2606.12956
SERF: Spatiotemporal Environment and Robot Feature Map for Long-Horizon Mobile Manipulation
SERF：用于长时间移动操控的时空环境与机器人特征图
Abstract
Long-horizon robot mobile manipulation requires continual reasoning about localization, environment changes, and task progress, all of which are challenging to infer from image observations alone. In this paper, we show that conditioning a mobile manipulation policy on a spatiotemporal feature map improves reasoning over long horizons. The map represents the environment and the articulated robot body as neural points in a shared latent space and is updated online from egocentric observations and proprioceptive state. We update the environment neural points using object-level rigid tracking and the robot neural points using forward kinematics. We use our spatiotemporal environment and robot feature (SERF) map as a state input to a vision-language-action (VLA) model by extracting map tokens from multiple reference frames and spatial scales, providing the policy with both local and global context. We demonstrate SERF on BEHAVIOR-1K, a benchmark for long-horizon mobile manipulation in household environments. Experiments show that the SERF VLA policy outperforms image-only baselines, reaches subgoals faster by following more direct trajectories, improves robustness to scene-configuration shifts, and recovers from object-drop failures.
Chinese Translation
长时间的机器人移动操控需要持续推理定位、环境变化和任务进展，而仅从图像观察中推断这些信息是具有挑战性的。本文展示了将移动操控策略基于时空特征图进行条件化可以改善长时间的推理。该图将环境和关节机器人身体表示为共享潜在空间中的神经点，并通过自我中心观察和本体状态在线更新。我们使用物体级刚性跟踪更新环境神经点，并使用前向运动学更新机器人神经点。我们通过从多个参考帧和空间尺度提取图令牌，将我们的时空环境与机器人特征（SERF）图作为视觉-语言-动作（VLA）模型的状态输入，为策略提供局部和全局上下文。我们在BEHAVIOR-1K上展示了SERF，这是一个针对家庭环境中长时间移动操控的基准。实验表明，SERF VLA策略优于仅使用图像的基线，通过遵循更直接的轨迹更快地达到子目标，提高了对场景配置变化的鲁棒性，并能从物体掉落故障中恢复。
cs.RO / 18 / 2606.12965
EmbodiSteer: Steering Embodiment-Agnostic Visuomotor Policies with Joint-Space Guidance for Zero-Shot Cross-Embodiment Deployment
EmbodiSteer：通过关节空间引导实现零-shot跨身体部署的与身体无关的视觉运动策略的引导
Abstract
Scalable robot imitation learning relies on large-scale heterogeneous data from diverse robots or body-free data, making Cartesian end-effector actions a key interface for embodiment-agnostic policy learning. However, end-effector-only abstraction leaves Cartesian policies unaware of the deployed robot body, making them brittle under robot-specific constraints such as whole-body collision avoidance. To overcome this limitation, we present EmbodiSteer, a training-free framework that steers embodiment-agnostic visuomotor policies toward zero-shot, embodiment-aware deployment. EmbodiSteer keeps policy learning in Cartesian space while efficiently lifting inference-time diffusion sampling into the target robot's joint space via forward kinematics and Jacobian-based updates. With whole-body collision-aware guidance over joint trajectories after each denoising step, the arm can be steered away from collisions while preserving learned end-effector behavior. Compared with Cartesian-only execution, EmbodiSteer reduces collision rate by 46.1% and improves task success rate by 28.5% across 9 simulated robots, and further achieves 90.0% collision rate reduction and 36.7% success rate increase on two physical robots in highly constrained scenarios. Our project page is at https://frankwang67.github.io/EmbodiSteer-Page.
Chinese Translation
可扩展的机器人模仿学习依赖于来自多样化机器人或无身体数据的大规模异构数据，使得笛卡尔末端执行器动作成为与身体无关的策略学习的关键接口。然而，仅依赖末端执行器的抽象使得笛卡尔策略无法感知所部署的机器人身体，从而在特定于机器人的约束条件下变得脆弱，例如全身碰撞避免。为了解决这一限制，我们提出了EmbodiSteer，这是一个无训练的框架，旨在将与身体无关的视觉运动策略引导至零-shot、感知身体的部署。EmbodiSteer在笛卡尔空间中保持策略学习，同时通过正向运动学和基于雅可比的更新有效地将推理时的扩散采样提升到目标机器人的关节空间。在每个去噪步骤后，通过对关节轨迹的全身碰撞感知引导，手臂可以在保持已学习的末端执行器行为的同时，避免碰撞。与仅使用笛卡尔执行相比，EmbodiSteer在9个模拟机器人上将碰撞率降低了46.1%，任务成功率提高了28.5%；在两个高度受限的物理机器人上，进一步实现了90.0%的碰撞率降低和36.7%的成功率提升。我们的项目页面地址为 https://frankwang67.github.io/EmbodiSteer-Page。
cs.RO / 19 / 2606.12978
Trajectory-Level Redirection Attacks on Vision-Language-Action Models
基于轨迹级的视觉-语言-动作模型重定向攻击
Abstract
Vision-language-action (VLA) policies bring natural language into closed-loop robot control, enabling robots to execute manipulation tasks directly from text instructions. The same interface gives text a recurring role in control because the prompt is reused at every replanning step, and each prompt-conditioned action changes the future observations on which the policy acts. Existing VLA attacks study adversarial prompts that elicit targeted low-level actions or make such actions persist across changing images. We identify a stronger trajectory-level failure mode: a prompt that still $\textit{appears}$ to specify the intended task but redirects the final physical outcome. We mathematically formalize this setting as $\textit{command-preserving trajectory redirection}$, a prompt-only threat model in which the attacker chooses one prompt before the episode, all policy and environment components remain fixed, and the prompt must stay close to the benign instruction while omitting target words and correction language. To find such prompts, we introduce an on-policy prompt search method that uses rollouts to discover perturbations whose closed-loop behavior tracks a target task while satisfying the command-preserving constraints. Experiments in simulation and on hardware show that near-benign prompt perturbations can redirect VLA rollouts to attacker-specified targets. These results expose a trajectory-level vulnerability in VLA instruction grounding: text that appears to preserve the intended command can still give an adversary control over the robot's final physical outcome. Project website: https://vla-redirection-attack.github.io/
Chinese Translation
视觉-语言-动作（VLA）策略将自然语言引入闭环机器人控制，使机器人能够直接根据文本指令执行操作任务。同一接口使文本在控制中扮演了反复出现的角色，因为在每个重新规划步骤中都会重复使用提示，而每个基于提示的动作都会改变策略所依赖的未来观察。现有的VLA攻击研究了引发目标低级动作的对抗性提示，或使这些动作在变化的图像中持续存在。我们识别出一种更强的轨迹级失败模式：一种提示虽然仍然$ extit{看似}$指定了预期任务，但却重定向了最终的物理结果。我们将这种情境数学形式化为$ extit{命令保持轨迹重定向}$，这是一种仅基于提示的威胁模型，其中攻击者在事件开始前选择一个提示，所有策略和环境组件保持固定，且提示必须与良性指令保持接近，同时省略目标词汇和修正语言。为了找到这样的提示，我们引入了一种基于策略的提示搜索方法，该方法利用回滚发现扰动，其闭环行为跟踪目标任务，同时满足命令保持约束。模拟和硬件实验表明，近良性的提示扰动可以将VLA回滚重定向到攻击者指定的目标。这些结果揭示了VLA指令基础中的一种轨迹级脆弱性：看似保持预期命令的文本仍然可以使对手控制机器人的最终物理结果。项目网站：https://vla-redirection-attack.github.io/
cs.RO / 20 / 2606.12995
GenHOI: Contact-Aware Humanoid-Object Interaction by Imitating Generated Videos without Task-Specific Training
GenHOI：通过模仿生成视频实现接触感知的人形机器人与物体的交互，无需特定任务训练
Abstract
Humanoid-Object Interaction (HOI) is a fundamental capability for humanoid robots, yet it remains challenging due to the tight coupling between dynamic balance and stable interaction with diverse objects. Existing methods often require time-consuming task-specific policy training or rely on rigid trajectory replay, which limits their ability to accommodate novel interaction scenarios. In this work, we present \textit{GenHOI}, a simple yet effective framework that enables humanoid robots to perform diverse object-interaction tasks in a zero-shot manner by directly imitating a single generated video, without task-specific training or physical demonstration data. GenHOI first reconstructs the robot-object scene in simulation and renders a first-frame image, which, together with the language command, conditions the synthesis of a task-oriented interaction video. The generated video is then analyzed to identify interaction-relevant contact events and estimate hand-object contact regions, which are encoded as object-centric geometric constraints that convert visual interaction cues into physically grounded optimization priors. Guided by these priors, the reference motion recovered from the video is refined and smoothed to resolve the scale ambiguity inherent in 2D video generation, while adapting a single reference trajectory to unseen robot-object relative poses. The optimized trajectory is finally executed by a closed-loop tracking controller. We validate the proposed framework in extensive simulation and real-world experiments across diverse object-interaction tasks, including box grasping, asymmetric bimanual chair carrying, table lifting from below, and cylindrical-object enveloping.
Chinese Translation
人形机器人与物体的交互（HOI）是人形机器人的一项基本能力，但由于动态平衡与与多样物体的稳定交互之间的紧密耦合，这一能力仍然面临挑战。现有的方法通常需要耗时的特定任务策略训练或依赖于刚性轨迹重放，这限制了它们适应新交互场景的能力。在本研究中，我们提出了 extit{GenHOI}，这是一个简单而有效的框架，使人形机器人能够以零-shot的方式执行多样的物体交互任务，方法是直接模仿单个生成的视频，而无需特定任务的训练或物理演示数据。GenHOI 首先在仿真中重建机器人-物体场景并渲染第一帧图像，随后结合语言指令，条件化合成任务导向的交互视频。生成的视频随后被分析，以识别与交互相关的接触事件并估计手-物体接触区域，这些区域被编码为以物体为中心的几何约束，将视觉交互线索转换为物理基础的优化先验。在这些先验的指导下，从视频中恢复的参考运动被精炼和平滑，以解决2D视频生成中固有的尺度模糊，同时将单一参考轨迹适应于未见的机器人-物体相对姿态。优化后的轨迹最终由闭环跟踪控制器执行。我们在广泛的仿真和现实世界实验中验证了所提出的框架，涵盖了多样的物体交互任务，包括箱子抓取、不对称双手搬运椅子、从下方抬起桌子以及对圆柱形物体的包围。
cs.RO / 21 / 2606.13028
Comparing Commercial Depth Sensor Accuracy for Medical Applications
比较用于医疗应用的商业深度传感器精度
Abstract
Depth estimation has numerous medical and surgical applications. We benchmark four depth sensors on a porcine bone specimen, a porcine belly specimen, and a silicone kidney phantom using stylus-sampled references. These objects contain several real-world challenges, including homogeneous surfaces, specular surfaces, and subsurface scattering. The comparison includes stereo, structured-light, and time-of-flight sensors at a distance of approximately 50 cm. Specifically, the Intel RealSense D405 (Intel RealSense, United States), PMD Flexx2 (pmdtechnologies, Germany), Stereolabs ZED 2i (Stereolabs, France), and Zivid 2M+ 60 (Zivid, Norway) are compared. The Zivid 2M+ 60 performed best across all objects and metrics considered in this work. The ZED ranked second for real tissue, but last on the phantom.
Chinese Translation
深度估计在医学和外科领域有着众多应用。我们在猪骨标本、猪腹部标本和硅胶肾脏模型上对四种深度传感器进行了基准测试，使用了触笔采样的参考数据。这些对象包含多个现实世界的挑战，包括均匀表面、镜面表面和亚表面散射。比较包括立体、结构光和飞行时间传感器，测试距离约为50厘米。具体而言，我们比较了英特尔RealSense D405（Intel RealSense，美国）、PMD Flexx2（pmdtechnologies，德国）、Stereolabs ZED 2i（Stereolabs，法国）和Zivid 2M+ 60（Zivid，挪威）。在本研究考虑的所有对象和指标中，Zivid 2M+ 60的表现最佳。ZED在真实组织中排名第二，但在模型中排名最后。
cs.RO / 22 / 2606.13040
RoboProcessBench: Benchmarking Process-Aware Understanding in Vision-Language Robotic Manipulation
RoboProcessBench：视觉语言机器人操作中的过程感知理解基准测试
Abstract
Vision-language models (VLMs) are increasingly explored as visual critics, reward generators, and failure detectors in robotic manipulation. These roles implicitly require models to judge not only final task success, but also how a manipulation execution is physically and temporally progressing. However, existing evaluations fail to test whether VLMs possess fine-grained process understanding. To address this gap, we present RoboProcessBench, a benchmark for process-aware understanding in vision-language robotic manipulation. RoboProcessBench decomposes such capability into two complementary dimensions, \emph{static monitoring} and \emph{dynamic reasoning}, instantiated as 12 diagnostic question families covering phase, contact, motion, coordination, primitive-local progress, temporal order, outcome, and primitive-level transitions. Built from physically grounded execution traces, the curated benchmark corpus ProcessData contains \textasciitilde 58k question-answer pairs across 260 manipulation tasks, which is further split into ProcessData-SFT and ProcessData-Eval for post-training and evaluation purposes. Extensive evaluation of various VLMs on ProcessData-Eval reveals broad limitations across 12 diagnostic task families, suggesting current models still lack robust process-aware understanding of manipulation executions. But with ProcessData-SFT, the post-trained \textit{Qwen2.5-VL-7B} and \textit{InternVL-3-8B} exhibit consistent gains on local state, motion, progress, and primitive-aware cues. These results demonstrate that RoboProcessBench serves as both an evaluation benchmark and a learnable supervision source for developing VLMs capable of monitoring and evaluating robotic manipulation processes. Project webpage: \href{https://processbench-2026.github.io/RoboProcessBench-Web/}{https://processbench-2026.github.io}.
Chinese Translation
视觉语言模型（VLMs）在机器人操作中被越来越多地探索作为视觉评论者、奖励生成器和故障检测器。这些角色隐含地要求模型不仅判断最终任务的成功，还要评估操作执行在物理和时间上的进展。然而，现有的评估未能测试VLMs是否具备细粒度的过程理解。为了解决这一问题，我们提出了RoboProcessBench，这是一个用于视觉语言机器人操作中的过程感知理解的基准测试。RoboProcessBench将这种能力分解为两个互补的维度，即 extit{静态监测}和 extit{动态推理}，并具体化为12个诊断问题类别，涵盖阶段、接触、运动、协调、原始局部进展、时间顺序、结果和原始级别的转变。该基准数据集ProcessData基于物理基础的执行轨迹构建，包含约58,000个问答对，涉及260个操作任务，并进一步分为ProcessData-SFT和ProcessData-Eval，用于后训练和评估目的。在ProcessData-Eval上对各种VLMs的广泛评估揭示了12个诊断任务类别中普遍存在的局限性，表明当前模型仍然缺乏对操作执行的稳健过程感知理解。但通过ProcessData-SFT，后训练的 extit{Qwen2.5-VL-7B}和 extit{InternVL-3-8B}在局部状态、运动、进展和原始感知线索上表现出一致的提升。这些结果表明，RoboProcessBench既作为评估基准，也作为可学习的监督源，帮助开发能够监测和评估机器人操作过程的VLMs。项目网页： exttt{https://processbench-2026.github.io/RoboProcessBench-Web/}。
cs.RO / 23 / 2606.13049
Y-BotFrame: An Extensible Embodied Agent Framework for Quadruped Robot Assistants
Y-BotFrame：一种可扩展的四足机器人助手的具身代理框架
Abstract
Quadruped robots are capable of traversing a wide range of complex terrains with high flexibility. As highly mobile ground-based intelligent platforms, they can be equipped with modules for navigation control, environmental perception, and intelligent interaction, thereby serving as real-world mobile deployment platforms for various algorithms. In this paper, we introduce Y-BotFrame, an extensible embodied platform that turns a robot into an intelligent ground assistant. Y-BotFrame integrates multimodal perception capabilities, including speech, vision, and LiDAR, and employs a large language model as the cognitive core for environmental understanding, contextual reasoning, and task planning. The system maps user natural-language instructions into executable embodied task units that can be carried out by the robot. Y-BotFrame supports natural interaction through voice commands and visual feedback, removing the need for a remote controller and enabling efficient human-robot collaboration. With a highly extensible framework, Y-BotFrame supports plug-and-play integration of new functional modules as well as modular upgrades and iterative development, offering a reference implementation for the real-world deployment of general-purpose, instruction-driven embodied agents.The supplementary video is available at https://xdei-group.github.io/Y-BotFrame/.
Chinese Translation
四足机器人能够以高度灵活性穿越各种复杂地形。作为高度移动的地面智能平台，它们可以配备导航控制、环境感知和智能交互模块，从而作为各种算法的真实世界移动部署平台。在本文中，我们介绍了Y-BotFrame，这是一种可扩展的具身平台，将机器人转变为智能地面助手。Y-BotFrame集成了多模态感知能力，包括语音、视觉和激光雷达，并采用大型语言模型作为环境理解、上下文推理和任务规划的认知核心。该系统将用户的自然语言指令映射为可由机器人执行的具身任务单元。Y-BotFrame通过语音命令和视觉反馈支持自然交互，消除了对遥控器的需求，实现了高效的人机协作。凭借高度可扩展的框架，Y-BotFrame支持新功能模块的即插即用集成，以及模块化升级和迭代开发，为通用、基于指令的具身代理的真实世界部署提供了参考实现。补充视频可在 https://xdei-group.github.io/Y-BotFrame/ 获取。
cs.RO / 24 / 2606.13053
EA-WM: Event-Aware World Models with Task-Specification Grounding for Long-Horizon Manipulation
EA-WM：具有任务规范基础的事件感知世界模型用于长时间跨度的操作
Abstract
Pretrained-feature world models provide a useful substrate for robot imagination, but visual or latent prediction alone does not determine whether an imagined future satisfies task-relevant events. Long-horizon manipulation requires progress signals that are relational, predicate-level, and physically grounded: whether an object has moved, whether a drawer or contact state has changed, whether a placement predicate is satisfied, and whether a candidate future is reliable enough for execution. We introduce EA-WM, an event-aware world-model framework that augments frozen visual-feature dynamics with task-specification-grounded event prediction and verification. EA-WM rolls out candidate futures in pretrained visual-feature space, decodes them into structured event states, and scores them using task-progress, semantic-consistency, physical-feasibility, and uncertainty terms. The verifier guides sampling-based planning, gates candidate actions, and, in the contact-sensitive LIBERO wine-rack setting, selects among PPOgenerated proposals. Across navigation, deformable-object, wall-constrained, and languagedescribed manipulation studies, EA-WM shows that event-aware verification can make featurespace world models more interpretable and better aligned with task progress.
Chinese Translation
预训练特征的世界模型为机器人想象提供了有用的基础，但仅凭视觉或潜在预测并不能确定想象的未来是否满足与任务相关的事件。长时间跨度的操作需要关系性、谓词级别和物理基础的进展信号：例如，物体是否移动，抽屉或接触状态是否发生变化，放置谓词是否满足，以及候选未来是否足够可靠以供执行。我们提出了EA-WM，一种事件感知世界模型框架，通过任务规范基础的事件预测和验证来增强冻结的视觉特征动态。EA-WM在预训练的视觉特征空间中推出候选未来，将其解码为结构化事件状态，并使用任务进展、语义一致性、物理可行性和不确定性等项对其进行评分。验证器指导基于采样的规划，控制候选动作，并在接触敏感的LIBERO酒架设置中，从PPO生成的提案中进行选择。在导航、可变形物体、墙面约束和语言描述的操作研究中，EA-WM表明事件感知验证可以使特征空间世界模型更具可解释性，并更好地与任务进展对齐。
cs.RO / 25 / 2606.13102
FTP-1: A Generalist Foundation Tactile Policy Across Tactile Sensors for Contact-Rich Manipulation
FTP-1：一种通用基础触觉策略，适用于接触丰富的操控任务
Abstract
Despite the success of vision-based generalist robotic policies, existing tactile-based policies remain tied to fixed embodiments and sensor setups. This is because tactile signals are highly heterogeneous across hardware, making cross-sensor generalization difficult. We present FTP-1,the first generalist foundation tactile policy pretrained to acquire transferable tactile manipulation abilities across diverse sensors and embodiments. FTP-1 supports varied tactile inputs, including image-, array-, and state-based signals, by using heterogeneous encoders to project them into unified morphology-aware latent tokens that are jointly modeled by a shared tactile Transformer expert. Pretrained on around 3,000 hours of tactile manipulation data aggregated from 26 data sources, spanning human and robot demonstrations across 21 sensors, FTP-1 learns tactile skills that transfer beyond the sensors seen during pretraining. Across downstream finetuning experiments spanning 5 hardware configurations, FTP-1 improves contact-rich manipulation on seen sensor setups by +17.2% and, surprisingly, transfers to two previously unseen tactile-sensor setups, achieving a +31% gain in success rate. FTP-1 establishes the first unified foundation baseline for tactile manipulation, providing future tactile policies with a shared model-level starting point. Pretrained models, datasets, training code and more visualization at https://ftp1-policy.github.io.
Chinese Translation
尽管基于视觉的通用机器人策略取得了成功，现有的基于触觉的策略仍然与固定的实体和传感器配置相关联。这是因为触觉信号在硬件之间高度异质，使得跨传感器的泛化变得困难。我们提出了FTP-1，这是第一个经过预训练的通用基础触觉策略，旨在获取可转移的触觉操控能力，适用于多种传感器和实体。FTP-1支持多种触觉输入，包括基于图像、阵列和状态的信号，通过使用异质编码器将其投影到统一的形态感知潜在标记中，这些标记由共享的触觉Transformer专家共同建模。FTP-1在约3000小时的触觉操控数据上进行了预训练，这些数据来自26个数据源，涵盖了21种传感器的人类和机器人演示。FTP-1学习到的触觉技能能够超越预训练期间所见的传感器。在涵盖5种硬件配置的下游微调实验中，FTP-1在已见传感器配置上的接触丰富操控性能提高了17.2%，并且令人惊讶的是，它能够转移到两个之前未见的触觉传感器配置，实现了31%的成功率提升。FTP-1为触觉操控建立了第一个统一的基础基准，为未来的触觉策略提供了一个共享的模型级起点。预训练模型、数据集、训练代码及更多可视化内容请访问 https://ftp1-policy.github.io。
cs.RO / 26 / 2606.13169
Redesigning Regularization for Effective Policy Smoothing
重新设计正则化以有效平滑策略
Abstract
This paper proposes a novel regularization design to effectively smooth policy functions in reinforcement learning. While regularization that enhances ``global'' Lipschitz continuity was initially considered, it has been limited to ``local'' Lipschitz continuity due to a tradeoff between smoothness and expressiveness. However, it has become apparent that the original implementation is cumbersome and does not provide sufficient smoothing, leading to a preference for simpler implementations. This stems from a discrepancy between theory and implementation, and a more appropriate implementation can expect to facilitate smoothing. Therefore, this paper identifies three reasons why the original implementation does not function adequately and provide remedies for them. This modified regularization performs well across multiple tasks and algorithms, successfully achieving smooth motion while improving control performance. Furthermore, by applying it to sim-to-real reinforcement learning for a quadruped robot, it is demonstrated that smooth motion provides robustness against sudden changes in target velocity commands.
Chinese Translation
本文提出了一种新颖的正则化设计，以有效平滑强化学习中的策略函数。尽管最初考虑了增强“全局”Lipschitz连续性的正则化，但由于平滑性与表达能力之间的权衡，它被限制在了“局部”Lipschitz连续性。然而，显然原始实现繁琐，且未能提供足够的平滑效果，导致对更简单实现的偏好。这源于理论与实现之间的差异，更合适的实现有望促进平滑。因此，本文识别了原始实现未能有效工作的三个原因，并为其提供了补救措施。经过修改的正则化在多个任务和算法中表现良好，成功实现了平滑运动，同时提高了控制性能。此外，通过将其应用于四足机器人在仿真到现实的强化学习中，证明了平滑运动对目标速度指令的突变具有鲁棒性。
cs.RO / 27 / 2606.13190
Multi-Modal Multi-Agent Robotic Cognitive Alignment enabled by Non-Invasive Consumer Brain Computer Interfaces: A Proof of Concept Exploration
通过非侵入式消费级脑机接口实现的多模态多智能体机器人认知对齐：概念验证探索
Abstract
While non-verbal behaviors and expressive movements are essential for natural human-robot interaction, existing methods often overlook a crucial element: the human's internal cognitive state. Frequently, proactive multi-agent systems can interrupt humans at inopportune moments, leading to cognitive overload and decreased task performance. This paper introduces a framework for generating "cognitively aligned" multi-agent interactions, enhancing the ability of robotic systems to contextually defer communications to the user of an agent system during moments of high human mental workload and engagement. We present the design and implementation of a closed-loop architecture that explores the interplay between autonomous task execution and real-time neurophysiological focus. Using a consumer-grade Brain-Computer Interface (BCI), our approach continuously monitors Electroencephalography (EEG) spectral band powers while a human performs an engagement-inducing task. We propose an engagement-driven pipeline where an HTTP-based signaling mechanism places a primary agent's sensory inputs and audio outputs into a holding state upon detecting high engagement. This allows secondary agents to seamlessly process complex, delegated tasks in the background. Once the human's cognitive state returns to a lower cognitive load baseline, the primary agent releases the queued agent message. Our preliminary results demonstrate the feasibility of leveraging real-time signal processing, Large Language Models (LLMs), and physical robotic embodiments to create cognitively-aware, non-intrusive multi-agent systems.
Chinese Translation
尽管非语言行为和表现性动作对于自然的人机交互至关重要，但现有方法往往忽视了一个关键元素：人类的内部认知状态。主动的多智能体系统常常在不合适的时刻打断人类，导致认知过载和任务表现下降。本文介绍了一种生成“认知对齐”多智能体交互的框架，增强了机器人系统在高人类心理负荷和参与度时，能够在上下文中推迟与代理系统用户的通信的能力。我们展示了一个闭环架构的设计与实现，该架构探索了自主任务执行与实时神经生理专注之间的相互作用。通过使用消费级脑机接口（BCI），我们的方法在一个人类执行引发参与感的任务时，持续监测脑电图（EEG）频谱带功率。我们提出了一种基于参与度驱动的管道，其中基于HTTP的信号机制在检测到高参与度时将主要代理的感官输入和音频输出置于保持状态。这使得次级代理能够在后台无缝处理复杂的委派任务。一旦人类的认知状态恢复到较低的认知负荷基线，主要代理便释放排队的代理消息。我们的初步结果表明，利用实时信号处理、大型语言模型（LLMs）和物理机器人体现来创建认知感知的非侵入式多智能体系统是可行的。
cs.RO / 28 / 2606.13203
Embedding ISO 10218 Safety Compliance in Robots via Control Barrier Functions for Human-Robot Collaboration
通过控制障碍函数在机器人中嵌入ISO 10218安全合规性以促进人机协作
Abstract
Human-Robot Collaboration (HRC) requires strict adherence to safety standards, such as ISO 10218, to prevent harmful interactions. Standard Speed and Separation Monitoring (SSM) filters calculate safe robotic speeds based on conservative assumptions, such as constant human velocity, which prevents accurate predictions of minimum separation distances and causes unnecessary operational halts. This paper proposes a Control Barrier Function (CBF) that explicitly incorporates human acceleration data to analytically forward-predict the minimum human-robot separation distance during a worst-case robotic stopping trajectory. To guarantee safety at the control level, this predictive CBF is integrated as an inequality constraint within a Sequential Quadratic Programming (SQP) framework. Specifically, two methods are proposed: Method I, a CBF-constrained PD safety filter; and Method II, a task-scaling SQP controller that enforces a spatial tube constraint. Simulated and real-world experiments on a UR10e robot evaluate the two proposed methods against a standard industrial SSM module baseline. Results demonstrate that Method II dynamically modulates execution speed and confines spatial deviations. Compared to Method I, Method II achieves a 63\% reduction in mean trajectory error and avoids excessive evasive manoeuvres, ensuring high task throughput while complying with ISO 10218 SSM guidelines.
Chinese Translation
人机协作（HRC）要求严格遵守安全标准，如ISO 10218，以防止有害的交互。标准速度与间隔监测（SSM）过滤器基于保守假设（如恒定的人类速度）计算安全的机器人速度，这限制了对最小间隔距离的准确预测，并导致不必要的操作暂停。本文提出了一种控制障碍函数（CBF），该函数明确结合了人类加速度数据，以分析性地前向预测在最坏情况下机器人停止轨迹下的人机最小间隔距离。为了在控制层面保证安全，这一预测性CBF被作为不等式约束集成到序列二次规划（SQP）框架中。具体而言，提出了两种方法：方法I，一种CBF约束的PD安全过滤器；方法II，一种任务缩放的SQP控制器，强制执行空间管道约束。在UR10e机器人上进行的模拟和真实世界实验评估了这两种提出的方法与标准工业SSM模块基线的表现。结果表明，方法II动态调节执行速度并限制空间偏差。与方法I相比，方法II在平均轨迹误差上实现了63%的减少，并避免了过度的规避机动，确保了高任务吞吐量，同时遵循ISO 10218 SSM指南。
cs.RO / 29 / 2606.13222
Proprioceptive-visual correspondence enables self-other distinction in humanoid robots
本体感知-视觉对应促进类人机器人自我与他者的区分
Abstract
Distinguishing self from others is a prerequisite for social intelligence, yet humanoid robots that increasingly share workspaces with humans still lack this ability. Here we show that a humanoid robot can learn self-other distinction from proprioceptive-visual correspondence, without any identity labels or kinematic models. Once established, this distinction bootstraps a predictive self-model that maps joint configurations to three-dimensional body occupancy, capturing how the robot's body changes with action. In multi-agent scenes involving humans or morphologically identical robots, the system reliably identifies itself, learns a 3D self-model, and supports downstream tasks including target reaching, collision-aware motion planning, and human-to-robot motion retargeting. Together, these results outline a route toward bodily self-representation in robots that act and coordinate alongside others in shared physical environments. Project page: https://euron-zc.github.io/humanoid-self-model/.
Chinese Translation
区分自我与他者是社会智能的前提，然而，与人类共享工作空间的类人机器人仍然缺乏这种能力。本文展示了一种类人机器人如何通过本体感知-视觉对应学习自我与他者的区分，而无需任何身份标签或运动学模型。一旦建立，这种区分便为预测性自我模型的构建提供了基础，该模型将关节配置映射到三维身体占用，捕捉机器人身体随动作变化的情况。在涉及人类或形态上相同的机器人的多智能体场景中，该系统可靠地识别自身，学习三维自我模型，并支持包括目标到达、碰撞感知运动规划和人机运动重定向等下游任务。综合来看，这些结果勾勒出了一条实现机器人在共享物理环境中与他者协作和行动的身体自我表征的路径。项目页面：https://euron-zc.github.io/humanoid-self-model/
cs.RO / 30 / 2606.13232
WT-UMI: Tactile-based Whole-Body Manipulation via Force-Supervised Contact-Aware Planning
WT-UMI：基于触觉的全身操控通过力监督的接触感知规划
Abstract
Whole-body humanoid manipulation of bulky, deformable, and shared-load objects requires distributed contact sensing and explicit force regulation, yet most imitation policies treat contact force only implicitly. On the other hand, different demonstration sources provide complementary modalities with inherent trade-offs: human demonstrations capture natural contact forces but not robot-executable actions, while teleoperation directly records robot actions but with less natural force regulation. This paper presents \textbf{WT-UMI}, a wearable whole-body tactile interface worn by human operators or mounted on humanoids, providing accurate observations of tactile images, contact forces, and end-effector poses across both human demonstration and humanoid teleoperation modes. We introduce a force-conditioned target-pose correction module that converts measured human poses into contact-aware robot targets by learning corrections from teleoperation data. To leverage the natural force interaction in human data, we propose a force-supervised planner that predicts end-effector pose chunks and contact-force trajectories. The predicted contact force serves as the reference for a tactile-based admittance controller. Across five contact-rich tasks spanning deformable objects, bulky rigid objects, and human--humanoid collaboration, WT-UMI improves success rate and reduces contact-position tracking error over four policy baselines. Our project page is available at https://wt-umi.github.io/WTUMI/.
Chinese Translation
对笨重、可变形和共享负载物体进行全身类人操控需要分布式接触传感和明确的力调节，然而大多数模仿策略仅隐式地处理接触力。另一方面，不同的示范来源提供了互补的模态，具有固有的权衡：人类示范捕捉自然接触力，但不提供机器人可执行的动作，而遥操作则直接记录机器人动作，但力调节不够自然。本文提出了 extbf{WT-UMI}，一种可穿戴的全身触觉接口，供人类操作员佩戴或安装在类人机器人上，能够提供准确的触觉图像、接触力和末端执行器姿态的观察，适用于人类示范和类人遥操作模式。我们引入了一个基于力条件的目标姿态修正模块，通过从遥操作数据中学习修正，将测得的人类姿态转换为接触感知的机器人目标。为了利用人类数据中的自然力交互，我们提出了一种力监督规划器，预测末端执行器姿态块和接触力轨迹。预测的接触力作为基于触觉的顺应控制器的参考。在涵盖可变形物体、笨重刚性物体和人类-类人合作的五个接触丰富任务中，WT-UMI提高了成功率，并减少了接触位置跟踪误差，相较于四个策略基线。我们的项目页面可访问 https://wt-umi.github.io/WTUMI/。
cs.RO / 31 / 2606.13256
Humor Style Drives Laughter, Topic Shapes Acceptability: Evaluating Bilingual Personal and Political Robot-Delivered AI Jokes
幽默风格驱动笑声，话题塑造可接受性：评估双语个人和政治机器人传递的人工智能笑话
Abstract
Humor plays a central role in human social relationships, and recent advances in computational humor create new opportunities for integrating humor into human-robot interaction (HRI). While large language models (LLMs) can generate diverse forms of humor, it remains unclear how humor style, joke content, and language preference shape perceptions of robot-delivered humor in group settings. In this exploratory study, we employed a mixed factorial design in which participants evaluated AI-generated jokes delivered by a robot in a university classroom. We examined the effects of humor type (Affiliative, Self-Enhancing, Aggressive, Self-Defeating) and joke content (person-related vs. political) on perceived funniness and appropriateness, as well as preferred language. Results show that humor type significantly influences funniness, with Aggressive and Affiliative humor rated higher, while joke content primarily affects appropriateness, with person-related jokes preferred over political ones. Language preference was shaped by both joke content and participants' self-reported fluency and humor practices.
Chinese Translation
幽默在人的社会关系中扮演着核心角色，最近在计算幽默方面的进展为将幽默融入人机交互（HRI）创造了新的机会。虽然大型语言模型（LLMs）能够生成多样化的幽默形式，但幽默风格、笑话内容和语言偏好如何在群体环境中影响对机器人传递幽默的感知仍不清楚。在这项探索性研究中，我们采用了混合因子设计，参与者评估了在大学课堂上由机器人传递的人工智能生成的笑话。我们考察了幽默类型（亲和型、自我增强型、攻击型、自我贬低型）和笑话内容（与个人相关 vs. 政治）对感知幽默性和适当性的影响，以及参与者的语言偏好。结果显示，幽默类型显著影响幽默性，攻击型和亲和型幽默的评分较高，而笑话内容主要影响适当性，与个人相关的笑话比政治笑话更受欢迎。语言偏好受到笑话内容以及参与者自报的流利程度和幽默实践的影响。
cs.RO / 32 / 2606.13279
See Selectively, Act Adaptively: Dual-Level Structural Decomposition for Bimanual Robot Manipulation
选择性观察，适应性行动：双层结构分解用于双手机器人操作
Abstract
In bimanual robotic manipulation, task-relevant visual information varies with the task stage and context, while the interaction of the two arms shifts between independent and coordinated modes, making policy learning challenging. However, existing monolithic Vision-Language-Action (VLA) policies process diverse visual inputs and interaction patterns through a single shared representation and action generation pathway, often failing to separately account for visual relevance and bimanual interaction structure. To address this issue, we propose a bimanual manipulation VLA framework based on Dual-Level Structural Decomposition. The View-Selective Visual Router dynamically adjusts wrist-view contributions to emphasize relevant visual cues, while the Interaction-Aware Action Mixture-of-Experts (MoE) decomposes action generation into coordinated and arm-wise pathways to adapt to varying bimanual interaction modes. We evaluate the proposed method on six simulated bimanual manipulation tasks in RoboTwin 2.0 and three long-horizon real-world tasks. Our model improves the overall average success rate over a monolithic baseline by 27.7% in simulation and 43.3% in real-world evaluation, while consistently outperforming single-module variants across both settings. These results demonstrate that jointly considering selective visual processing and explicit decomposition of bimanual interaction structures provides an effective inductive bias for robust bimanual manipulation.
Chinese Translation
在双手机器人操作中，任务相关的视觉信息随着任务阶段和上下文的变化而变化，同时双臂的交互模式在独立和协调之间切换，这使得策略学习变得具有挑战性。然而，现有的单一视觉-语言-动作（Vision-Language-Action, VLA）策略通过单一共享表示和动作生成路径处理多样的视觉输入和交互模式，往往未能分别考虑视觉相关性和双手交互结构。为了解决这一问题，我们提出了一种基于双层结构分解的双手操作VLA框架。视图选择性视觉路由器（View-Selective Visual Router）动态调整手腕视角的贡献，以强调相关的视觉线索，而交互感知动作专家混合（Interaction-Aware Action Mixture-of-Experts, MoE）将动作生成分解为协调和单臂路径，以适应不同的双手交互模式。我们在RoboTwin 2.0上评估了该方法在六个模拟双手操作任务和三个长时间真实世界任务中的表现。我们的模型在模拟中比单一基线提高了27.7%的整体平均成功率，在真实世界评估中提高了43.3%，同时在这两种设置中始终优于单模块变体。这些结果表明，联合考虑选择性视觉处理和双手交互结构的显式分解为稳健的双手操作提供了有效的归纳偏置。
cs.RO / 33 / 2606.13340
EMG-Based Adaptation of Anisotropic Virtual Fixtures for Robot-Assisted Surgical Resection and Dissection
基于肌电信号的各向异性虚拟夹具适应性系统用于机器人辅助外科切除和解剖
Abstract
In this paper, we address the development of an adaptive assistance system for robot-assisted laparoscopic surgery, specifically for delicate tasks such as Resection and Dissection. Even if Virtual Fixtures offer significant advantages for guiding a surgeon's movements, conventional Virtual Fixtures are often defined by fixed geometries, lacking the flexibility to adapt to the surgical workflow or the surgeon's immediate intent. To address these limitations, we propose a novel framework for an adaptive and anisotropic virtual fixture. In addition, we introduce an intuitive control interface that modulates the fixture's geometry in real-time based on the surgeon's intent, inferred from EMG signals. This approach allows the surgeon to dynamically expand or disengage the constraint by contracting their forearm muscles, enabling seamless transitions between precise guided motion and free repositioning of the tool. Experimental results from a pilot user study, based on a standardized surgical training task, demonstrate the effectiveness of the proposed method. The system showed significant improvements in task accuracy and movement consistency, alongside a reduction in perceived cognitive load, effort, and frustration.
Chinese Translation
本文探讨了一种用于机器人辅助腹腔镜手术的适应性辅助系统的开发，特别针对切除和解剖等精细任务。尽管虚拟夹具在引导外科医生的运动方面提供了显著优势，但传统的虚拟夹具通常由固定几何形状定义，缺乏适应手术工作流程或外科医生即时意图的灵活性。为了解决这些局限性，我们提出了一种新颖的适应性各向异性虚拟夹具框架。此外，我们引入了一种直观的控制界面，能够根据从肌电信号推断出的外科医生意图实时调节夹具的几何形状。这种方法使外科医生能够通过收缩前臂肌肉动态扩展或解除约束，从而实现精确引导运动与工具自由重新定位之间的无缝过渡。基于标准化外科训练任务的初步用户研究实验结果表明，所提方法的有效性。该系统在任务准确性和运动一致性方面显示出显著改善，同时减少了感知的认知负担、努力程度和挫败感。
cs.RO / 34 / 2606.13352
Low cost, easily manufactured, highly flexible strain and touch sensitive fiber for robotics applications
低成本、易于制造、高度柔性的应变和触摸敏感纤维在机器人应用中的研究
Abstract
Existing stretch and touch sensors for robots are generally expensive with respect to at least one of material costs, required manufacturing equipment, or manufacturing time. We present and experimentally characterize a conductive fiber made using only inexpensive commercial off-the-shelf parts (conductive thread at $0.07/ft, silicone tubing at $0.94/ft) and tools (loop-style needle threader at $2), which can be manufactured quickly (20 cm length in 2 minutes.) We demonstrate its use as a resistive strain sensor with three applications: Triggering a grasp in a pneumatically actuated assistive finger, sensing the pose of a pneumatically actuated robotic strap, and estimating the pose of a flexible solid. We also demonstrate that it can be used as a capacitive sensor with two applications: First, as a touch sensor which triggers a commercial robot arm to move, and second, as a near-field sensor enabling the robot arm to follow a moving hand. The capacitive sensors are knitted, showcasing the high flexibility of the fiber. We discuss methods for improving manufacturing scalability and their cost trade-offs. Finally, we demonstrate a method for repairing a cut fiber.
Chinese Translation
现有的机器人拉伸和触摸传感器在材料成本、所需制造设备或制造时间等方面通常较为昂贵。我们提出并实验性地表征了一种仅使用廉价的商业现成部件（导电线 $0.07/英尺，硅胶管 $0.94/英尺）和工具（环形针线穿线器 $2）制造的导电纤维，该纤维可以快速制造（20厘米长度在2分钟内完成）。我们展示了其作为电阻式应变传感器的应用，包括三个应用场景：触发气动驱动的辅助手指的抓握、感知气动驱动的机器人带的姿态，以及估计柔性固体的姿态。我们还展示了它可以作为电容式传感器的两个应用：首先，作为触摸传感器触发商业机器人手臂的移动；其次，作为近场传感器使机器人手臂跟随移动的手。电容式传感器采用编织方式，展示了该纤维的高度柔韧性。我们讨论了提高制造规模化的方法及其成本权衡。最后，我们展示了一种修复切断纤维的方法。
cs.RO / 35 / 2606.13355
Real-Time Execution with Autoregressive Policies
基于自回归策略的实时执行
Abstract
Real-time execution, enabled by asynchronous inference that ensures both smooth action trajectories and fast reactivity, is critical for realistic deployments of large-scale Vision-Language-Action models. However, recent work on real-time execution primarily focuses on variants of diffusion policies, even though it is more critical for autoregressive policies given their slower rollout speed in synchronous inference. In contrast, we demonstrate that autoregressive policies can achieve real-time execution by adjusting the tokenization horizon and applying constrained decoding, thereby guaranteeing strict latency bounds that enable multi-trajectory decoding to maximize performance. Across simulated and real-world environments, we find that the autoregressive policy consistently outperforms its equivalent-level flow-matching policy counterpart while achieving significantly improved task completion speeds from synchronous inference. Coupled with the inherent advantages of autoregressive policies, such as faster convergence and better generalizability in instruction-following, these results confirm that autoregressive policies can remain a competitive policy type supporting real-time execution.
Chinese Translation
实时执行依赖于异步推理，确保平滑的动作轨迹和快速的反应能力，这对于大规模视觉-语言-动作模型的现实部署至关重要。然而，近期关于实时执行的研究主要集中在扩散策略的变体上，尽管考虑到自回归策略在同步推理中较慢的展开速度，这一问题对自回归策略而言更为关键。相反，我们展示了通过调整标记化视野和应用约束解码，自回归策略可以实现实时执行，从而保证严格的延迟界限，支持多轨迹解码以最大化性能。在模拟和真实环境中，我们发现自回归策略始终优于其同等水平的流匹配策略，同时在任务完成速度上显著优于同步推理。结合自回归策略固有的优势，如更快的收敛速度和更好的指令跟随泛化能力，这些结果证实了自回归策略可以作为支持实时执行的竞争性策略类型。
cs.RO / 36 / 2606.13394
GeoHAT: Geometry-Adaptive Hybrid Action Transformer for Mobile Manipulation
GeoHAT：用于移动操作的几何自适应混合动作变换器
Abstract
Whole-body mobile manipulation requires coordinating mobile base and manipulator under shifting viewpoints, posing challenges in geometric perception and action generation. Current policies either rely on 2D features or sparse 3D representations that lack dense spatial structure, and typically encode arm and base within one action vector that ignores their distinct control demands. Moreover, existing dense fusion strategies risk corrupting pretrained representations under noisy depth while incurring heavy computational overhead. We present GeoHAT, an end-to-end diffusion-based framework built on a simple principle: geometry should be injected only where reliable and attended to only where needed. GeoHAT employs a lightweight Fourier spatial encoder that maps dense per-pixel 3D coordinates into geometric tokens without an additional 3D vision backbone. These tokens are then selectively injected into vision foundation model features through per-token gated fusion modulated by depth validity, preserving the semantic prior while enriching spatial understanding. For action generation, a Hybrid Whole-Body Action Decoder decomposes arm and base into distinct subspaces and lets each action modality attend to its task-relevant visual context through sparse cross-attention, while causal temporal modeling captures intra-timestep coordination and inter-timestep dependencies. Experiments on the ManiSkill-HAB simulation benchmark demonstrate that GeoHAT achieves a 79.3% mean success rate, surpassing the strongest baseline by 23.7%. Furthermore, real-world experiments on diverse tasks also confirm consistent improvements over all baselines.
Chinese Translation
全身移动操作需要在不断变化的视角下协调移动底座和操控器，这对几何感知和动作生成提出了挑战。目前的策略要么依赖于二维特征，要么依赖于缺乏密集空间结构的稀疏三维表示，通常将手臂和底座编码为一个动作向量，忽视了它们各自不同的控制需求。此外，现有的密集融合策略在噪声深度下可能会破坏预训练表示，并且会带来较大的计算开销。我们提出了GeoHAT，一个基于扩散的端到端框架，其建立在一个简单的原则上：几何信息应仅在可靠的地方注入，并且仅在需要的地方关注。GeoHAT采用轻量级的傅里叶空间编码器，将密集的每像素三维坐标映射为几何标记，而无需额外的三维视觉骨干网络。这些标记通过深度有效性调制的逐标记门控融合选择性地注入到视觉基础模型特征中，保留了语义先验，同时丰富了空间理解。对于动作生成，混合全身动作解码器将手臂和底座分解为不同的子空间，并让每种动作模态通过稀疏交叉注意力关注其任务相关的视觉上下文，而因果时间建模则捕捉了时间步内的协调和时间步间的依赖关系。在ManiSkill-HAB仿真基准上的实验表明，GeoHAT实现了79.3%的平均成功率，超越了最强基线23.7%。此外，在各种任务上的真实世界实验也证实了相对于所有基线的一致性改善。
cs.RO / 37 / 2606.13435
GIVE: Grounding Human Gestures in Vision-Language-Action Models
GIVE：将人类手势与视觉-语言-动作模型相结合
Abstract
Human communication is inherently multimodal, where language is often accompanied by non-verbal cues such as gestures to convey intentions. However, current Vision-Language-Action (VLA) models treat robotic manipulation as a pure text-driven task, overlooking the important role of gestures in Human-Robot Interaction (HRI). This often leads to inaccurate intent grounding and unreliable manipulation when language instructions are ambiguous or underspecified. To address this challenge, we propose GIVE (Gesture Intent via Visual-Semantic Enhancement), an effective approach that enhances pre-trained VLA models with human gesture understanding without architectural modifications. Specifically, GIVE incorporates gesture information through two complementary pathways: a visual pathway that overlays hand skeletons and fingertip rays onto robot observations for explicit object grounding, and a semantic pathway that generates high-level descriptions of human gestures and task instructions for robust intent grounding. By jointly leveraging visual and semantic guidance, GIVE enables VLA policies to better associate gestures with manipulation behaviors and adapt to dynamic interaction intents. In real-world HRI experiments, GIVE substantially outperforms the baseline, improving target object recognition accuracy by 40% and overall task success rate by 80%, while demonstrating strong robustness and generalization to unseen spatial layouts and diverse participants.
Chinese Translation
人类沟通本质上是多模态的，语言通常伴随着非语言线索，如手势，以传达意图。然而，当前的视觉-语言-动作（VLA）模型将机器人操作视为纯文本驱动的任务，忽视了手势在人与机器人交互（HRI）中的重要作用。这常常导致在语言指令模糊或不明确时，意图的基础不准确和操作不可靠。为了解决这一挑战，我们提出了GIVE（通过视觉-语义增强的手势意图），这是一种有效的方法，通过不改变架构的方式，增强预训练的VLA模型以理解人类手势。具体而言，GIVE通过两个互补的通道整合手势信息：一个视觉通道将手部骨架和指尖射线叠加到机器人观察中，以实现明确的物体基础；一个语义通道生成高层次的人类手势和任务指令描述，以实现稳健的意图基础。通过共同利用视觉和语义指导，GIVE使VLA策略能够更好地将手势与操作行为关联，并适应动态的交互意图。在真实世界的HRI实验中，GIVE显著优于基线，目标物体识别准确率提高了40%，整体任务成功率提高了80%，同时在未见空间布局和多样化参与者中展示了强大的鲁棒性和泛化能力。
cs.RO / 38 / 2606.13494
NavWAM: A Navigation World Action Model for Goal-Conditioned Visual Navigation
NavWAM：一种用于目标条件视觉导航的导航世界动作模型
Abstract
Goal-conditioned visual navigation requires a robot to act under partial observability by anticipating how its motion will change the future egocentric view and whether that change brings it closer to the goal. Navigation world models provide such visual foresight, but they remain prediction modules that require an external planner to convert predicted futures into closed-loop control. We propose Navigation World Action Model (NavWAM), a diffusion-transformer policy that turns navigation world-model prediction into executable action by representing future observations, goal-progress values, and action chunks in a shared latent sequence. By learning future prediction jointly with the action and value targets that determine closed-loop behavior, NavWAM makes visual foresight directly usable for robot control. We build NavWAM through simulation pretraining and real-robot adaptation, and evaluate it on image-goal navigation against planning-based world models and a representative direct navigation policy. Across offline benchmarks and closed-loop real-robot deployment, NavWAM improves over planning-based world-model baselines in our evaluations while using the default policy mode without CEM-style action search. Project page: https://dachii-azm.github.io/navwam/
Chinese Translation
目标条件视觉导航要求机器人在部分可观测的情况下行动，预测其运动如何改变未来的自我中心视图，以及这种变化是否使其更接近目标。导航世界模型提供了这种视觉前瞻性，但它们仍然是需要外部规划器将预测的未来转化为闭环控制的预测模块。我们提出了导航世界动作模型（NavWAM），这是一种扩散变换器策略，通过在共享的潜在序列中表示未来观察、目标进展值和动作块，将导航世界模型的预测转化为可执行的动作。通过与决定闭环行为的动作和价值目标共同学习未来预测，NavWAM使视觉前瞻性能够直接用于机器人控制。我们通过模拟预训练和真实机器人适应构建NavWAM，并在图像目标导航中对其进行评估，比较基于规划的世界模型和代表性的直接导航策略。在离线基准测试和闭环真实机器人部署中，NavWAM在我们的评估中优于基于规划的世界模型基线，同时在不使用CEM风格动作搜索的默认策略模式下运行。项目页面：https://dachii-azm.github.io/navwam/
cs.RO / 39 / 2606.13497
SPARC: Reliable Spatial Annotations from Robot Demonstrations at Scale
SPARC：来自机器人演示的可靠空间注释
Abstract
This work introduces Spatial Annotations from Robot Demonstrations with Reliability Calibration (SPARC), a risk-aware framework that automatically labels robot demonstrations with structured spatial annotations and assigns each annotation a reliability score. Structured spatial annotations, such as bounding boxes, object trajectories, and manipulation phase labels, benefit a broad range of robotics applications from training grounded robot policies and embodied foundation models to motion planning and hierarchical task composition. Existing automated pipelines generate such annotations at scale but provide no reliable quality signal: detector confidence is poorly calibrated for annotation correctness, forcing a choice between accepting noisy labels or discarding useful samples. In contrast to existing automated pipelines, SPARC leverages the spatio-temporal structure inherent to robot tasks to generate a reliability signal, reducing noisy labels and retaining more useful samples. We further introduce Interaction-Aware Bench (IA-Bench), a benchmark that measures model accuracy in grounding the locations of interacted objects in robot demonstrations. On 1.7k human-annotated demonstrations spanning diverse embodiments and scenarios, SPARC significantly outperforms detection-only baselines in localization accuracy while retaining three times more samples at high-precision operating points. Our experiments demonstrate that models finetuned on our annotations achieve state-of-the-art results on object-grounding and pointing benchmarks among similarly sized models, while remaining competitive on broader spatial-reasoning suites without manually verified or annotated training data. Furthermore, policies trained on SPARC-generated annotations outperform baselines in cluttered, visually ambiguous real-world scenes. Code, data, and models are available at intuitive-robots.github.io/sparc-labeling.
Chinese Translation
本研究介绍了带有可靠性校准的机器人演示空间注释（SPARC），这是一个风险感知框架，能够自动为机器人演示标注结构化空间注释，并为每个注释分配一个可靠性评分。结构化空间注释，如边界框、物体轨迹和操作阶段标签，能够为广泛的机器人应用提供支持，从训练基础机器人策略和具身基础模型到运动规划和分层任务组合。现有的自动化管道能够大规模生成此类注释，但未提供可靠的质量信号：检测器的置信度对于注释的正确性校准较差，迫使用户在接受噪声标签与丢弃有用样本之间做出选择。与现有的自动化管道相比，SPARC利用机器人任务固有的时空结构生成可靠性信号，从而减少噪声标签并保留更多有用样本。我们进一步介绍了交互感知基准（IA-Bench），这是一个评估模型在机器人演示中定位交互对象位置准确性的基准。在1.7k个人工注释的演示中，涵盖了多种体现和场景，SPARC在定位准确性上显著优于仅使用检测的基线，同时在高精度操作点保留了三倍以上的样本。我们的实验表明，基于我们的注释微调的模型在物体定位和指向基准测试中取得了同类模型中的最先进结果，同时在更广泛的空间推理任务中保持竞争力，而无需手动验证或注释的训练数据。此外，基于SPARC生成的注释训练的策略在杂乱且视觉模糊的真实场景中优于基线。代码、数据和模型可在 intuitive-robots.github.io/sparc-labeling 获取。
cs.RO / 40 / 2606.13601
MCR-Bionic Hand: Anatomical Structural Priors for Dexterous Manipulation
MCR仿生手：灵巧操作的解剖结构先验
Abstract
Dexterous robotic hands are usually formulated as high dimensional active control systems governed by degrees of freedom, actuation, and algorithms. Human hand dexterity, however, is partly encoded in the physical architecture of bones, ligaments, tendons, aponeuroses, and intrinsic muscles. This work describes that contribution as two linked forms of structural intelligence: structural prior generation, in which wrist to finger tenodesis, FDS/FDP routing, and the dorsal extensor hood transform low dimensional posture inputs into default grasp configurations and PIP to DIP coordination; and muscle mediated modulation, in which extrinsic muscles, lumbricals, and interossei regulate MCP posture, distal stability, fingertip force paths, and contact states around that default state. Based on this framework, MCR-Bionic Hand is developed as a 1:1 musculoskeletal biomimetic hand integrating a two row eight bone wrist, cross wrist tendons, anatomical flexor routing, volar plate and collateral ligament constraints, the dorsal extensor hood, and intrinsic muscle pathways within one body. Functional demonstrations and geometric mechanical models show that wrist posture induces multi joint pre shaping, the extensor hood maps PIP posture to a coupled DIP response, and intrinsic plus pathways modulate distal stability and fingertip action direction after grasp formation. Contact rich tasks, including coin rotation, pen transfer, dorsal coin flipping, and cube manipulation, show that MCR-Bionic links low dimensional state generation with fine post contact modulation. These results suggest that anatomical biomimetics is valuable not for visual similarity, but for identifying human hand structures that perform part of control.
Chinese Translation
灵巧的机器人手通常被视为由自由度、驱动和算法控制的高维主动控制系统。然而，人手的灵巧性部分体现在骨骼、韧带、肌腱、腱膜和内在肌肉的物理结构中。本研究将这种贡献描述为两种相互关联的结构智能形式：结构先验生成，其中腕部到手指的腱连结、深屈肌/浅屈肌的走向以及背侧伸肌罩将低维姿态输入转化为默认抓握配置和PIP到DIP的协调；以及肌肉介导的调节，其中外在肌肉、虫肌和骨间肌调节MCP姿态、远端稳定性、指尖施力路径和围绕该默认状态的接触状态。基于这一框架，MCR仿生手被开发为1:1的肌肉骨骼仿生手，整合了双排八骨腕部、交叉腕腱、解剖屈肌走向、掌板和侧韧带约束、背侧伸肌罩以及内在肌肉通路于一体。功能演示和几何机械模型表明，腕部姿态诱导多关节的预成型，伸肌罩将PIP姿态映射到耦合的DIP响应，而内在加通路在抓握形成后调节远端稳定性和指尖动作方向。包括硬币旋转、笔转移、背侧硬币翻转和立方体操作在内的接触丰富任务表明，MCR仿生手将低维状态生成与细微的接触后调节相结合。这些结果表明，解剖仿生学的价值不在于视觉相似性，而在于识别执行部分控制的人手结构。
cs.RO / 41 / 2606.13672
$\texttt{WEAVER}$, Better, Faster, Longer: An Effective World Model for Robotic Manipulation
WEAVER：更好、更快、更长的机器人操控有效世界模型
Abstract
The potential impacts of world models (WMs, i.e., learned simulators) on robotics are far-reaching -- policy evaluation, policy improvement, and test-time planning -- all with limited real-world interaction. To unlock these downstream capabilities, a WM needs to jointly satisfy three desiderata: $\textit{(i)}$ fidelity (i.e., producing simulated trajectories that correlate with reality), $\textit{(ii)}$ consistency (i.e., producing simulated trajectories that are coherent over long horizons), and $\textit{(iii)}$ efficiency (i.e., producing simulated trajectories quickly). We propose $\texttt{WEAVER}$ (World Estimation Across Views for Embodied Reasoning): a WM architecture that simultaneously achieves all three desiderata, providing state-of-the-art results on robotic manipulation tasks. $\texttt{WEAVER}$ is a multi-view WM trained to predict future latents and reward values via a flow-matching loss. We distill the key design decisions across model architecture, memory, and prediction objectives required to unlock the kinds of long-horizon dynamic manipulation tasks that have confounded prior world modeling approaches. We apply $\texttt{WEAVER}$ in robotic hardware, demonstrating its effectiveness at policy evaluation ($\rho$=0.870 correlation with real-world success rate), policy improvement (real-world success rate improvement of $38\%$ on top of the $\pi_{0.5}$ robot foundation model), and test-time planning (real-world success rate improvement of $14\%$ with a $5-10\times$ speedup over prior WMs). $\texttt{WEAVER}$ also demonstrates better performance than prior WMs when evaluated on out-of-distribution scenarios. Code, models, and videos at: https://arnavkj1995.github.io/WEAVER/ .
Chinese Translation
世界模型（WMs，即学习的模拟器）对机器人技术的潜在影响深远——政策评估、政策改进和测试时规划——所有这些都需要有限的现实世界交互。为了释放这些下游能力，世界模型需要共同满足三个要求：(i) 保真度（即产生与现实相关的模拟轨迹），(ii) 一致性（即产生在长时间范围内连贯的模拟轨迹），以及(iii) 效率（即快速产生模拟轨迹）。我们提出了WEAVER（跨视角的世界估计用于具身推理）：一种同时实现这三项要求的世界模型架构，在机器人操控任务上提供了最先进的结果。WEAVER是一个多视角的世界模型，通过流匹配损失训练以预测未来的潜在状态和奖励值。我们提炼了在模型架构、内存和预测目标方面的关键设计决策，这些决策是解锁以前的世界建模方法所无法解决的长时间动态操控任务所必需的。我们在机器人硬件上应用了WEAVER，展示了其在政策评估（与现实世界成功率的相关性为0.870）、政策改进（在基础模型π_{0.5}上实现38%的现实世界成功率提升）和测试时规划（在现实世界成功率上实现14%的提升，速度比之前的世界模型快5-10倍）方面的有效性。WEAVER在评估分布外场景时也表现出比之前的世界模型更好的性能。代码、模型和视频可在：https://arnavkj1995.github.io/WEAVER/ 获取。
cs.RO / 42 / 2606.13675
Improving Robotic Generalist Policies via Flow Reversal Steering
通过流反转引导改善机器人通用策略
Abstract
Generalist policies can learn a wide range of skills from diverse robot datasets. In order to solve or improve on challenging news tasks, we need a way to infer and invoke the appropriate actions from the policy's rich behavioral prior, especially when directly commanding the policy fails. We focus on flow matching generalists and propose Flow Reversal Steering (FRS): a method that takes suboptimal but ``reasonable'' actions, finds their latent noises by passing them through the flow policy in reverse, and maps them to nearby generalist action modes. We evaluate FRS across many simulated and real-world manipulation settings. First, FRS can turn coarse semantic guidance from humans or vision-language models (VLMs) into corresponding good robot actions, improving zero-shot control. These gains can be distilled with behavioral cloning by training an auxiliary policy to output noises that the generalist maps to good actions -- showing up to 95% absolute task success rate boosts in under a minute of training. Finally, FRS enables policy improvement by bootstrapping reinforcement learning with semantic knowledge, improving on several tasks that standard RL fails to improve on.
Chinese Translation
通用策略能够从多样的机器人数据集中学习广泛的技能。为了应对或改善具有挑战性的任务，我们需要一种方法，从策略丰富的行为先验中推断并调用适当的动作，特别是在直接指令策略失败时。我们专注于流匹配通用策略，并提出了流反转引导（Flow Reversal Steering, FRS）：一种采取次优但“合理”动作的方法，通过反向传递流策略找到其潜在噪声，并将其映射到附近的通用动作模式。我们在许多模拟和现实世界的操作环境中评估了FRS。首先，FRS能够将来自人类或视觉-语言模型（Vision-Language Models, VLMs）的粗略语义指导转化为相应的良好机器人动作，从而改善零-shot 控制。这些收益可以通过行为克隆进行提炼，通过训练辅助策略输出通用策略映射到良好动作的噪声——在不到一分钟的训练时间内，任务成功率提升可达95%。最后，FRS通过利用语义知识引导强化学习，从而实现策略改进，改善了标准强化学习未能改善的多个任务。
cs.RO / 43 / 2606.13677
Mana: Dexterous Manipulation of Articulated Tools
Mana：关节工具的灵巧操控
Abstract
Articulated tool manipulation remains a major challenge in dexterous robotics due to the need to coordinate internal degrees of freedom and contact-rich interactions. While prior work has largely focused on rigid objects, articulated tool use remains underexplored because of its physical complexity and the difficulty of learning functional grasping and manipulation policies. We present Mana (Manipulation Animator), a general sim-to-real framework that reinterprets dexterous manipulation as an animation problem. Inspired by computer animation, Mana employs a coarse-to-fine pipeline that transforms procedurally-generated grasp keyframes into manipulation trajectories through motion planning and reinforcement learning. The data generation process is largely automatic, requiring only a few mouse clicks to specify functional affordances (<1 minute per tool). Across four articulated tools spanning different scales and joint types, Mana achieves zero-shot sim-to-real transfer for both grasping and in-hand manipulation, demonstrating a scalable approach to dexterous articulated tool use.
Chinese Translation
关节工具的操控在灵巧机器人领域仍然是一个主要挑战，因为需要协调内部自由度和丰富的接触交互。尽管之前的研究主要集中在刚性物体上，但由于其物理复杂性以及学习功能性抓取和操控策略的难度，关节工具的使用仍然未得到充分探索。我们提出了Mana（Manipulation Animator），一个将灵巧操控重新解释为动画问题的通用仿真到现实框架。受到计算机动画的启发，Mana采用了粗到细的流程，通过运动规划和强化学习将程序生成的抓取关键帧转化为操控轨迹。数据生成过程基本上是自动的，仅需几次鼠标点击即可指定功能性可供性（每个工具少于1分钟）。在四种不同规模和关节类型的关节工具上，Mana实现了抓取和手内操控的零-shot仿真到现实转移，展示了对灵巧关节工具使用的可扩展方法。
cs.CV / 1 / 2606.12473
Stereo Vision-Based Fall Prediction and Detection using Human Pose Estimation on the AMD Kria K26 SOM
基于立体视觉的人体姿态估计的跌倒预测与检测系统在AMD Kria K26 SOM上的应用
Abstract
Background and Objective: Falls among elderly people can cause serious injury and reduce quality of life. Timely prediction and detection are essential to prevent harm and support well-being. We propose a portable, low-power, battery-operated, vision-based fall prediction and detection system using HPE on an AMD Kria K26 System-on-Module (SOM). The objective is a non-intrusive, privacy-preserving system for real-time fall detection. Methods: The system uses an Intel RealSense D455 range-sensing camera connected to the K26 SOM by USB. It captures synchronized RGB and depth frames, 640 x 480 x 3 and 640 x 480 pixels, at 60 FPS. The SOM runs a three-stage pipeline with quantized YOLOX, Anchor-to-Joint (A2J), and fall-detection models. YOLOX identifies human bounding boxes from RGB frames, then discards the RGB frames to preserve privacy. A2J uses depth frames to estimate 15 joint keypoints per person. A CNN uses selected joint coordinates (x, y, z) to classify fall activity. YOLOX was trained on CrowdHuman; A2J on ITOP, MP-3DHP, UR Fall Detection, and a custom SDSU PSG dataset; and the CNN on UR Fall Detection and SDSU PSG. The design used a single-core DPU with a serial pipeline and a dual-core DPU running YOLOX and A2J with multiple threads. Results: Quantized accuracy was evaluated using IoU >= 50% for YOLOX, mAP with a 10-cm rule for A2J, and classification accuracy, (TP + TN)/(TP + TN + FP + FN), for the CNN. Accuracies were 74%, 84.13%, and 75.85%. Throughput improved from 2.5 FPS for the single-threaded pipeline to 4.5 FPS for the multi-threaded version. Conclusion: Results demonstrate the feasibility of privacy-preserving fall detection on an AMD Kria K26 edge device. On-device HPE and fall classification runs without cloud dependency, supporting elderly monitoring and assistive healthcare. Future work will improve model accuracy and speed.
Chinese Translation
背景与目的：老年人跌倒可能导致严重伤害并降低生活质量。及时的预测与检测对于预防伤害和支持健康至关重要。我们提出了一种便携式、低功耗、可电池供电的基于视觉的跌倒预测与检测系统，该系统在AMD Kria K26系统模块（SOM）上使用人体姿态估计（HPE）。其目标是实现一个非侵入式、保护隐私的实时跌倒检测系统。方法：该系统使用一台通过USB连接到K26 SOM的英特尔RealSense D455深度感知摄像头。它以60 FPS捕获同步的RGB和深度帧，分辨率为640 x 480 x 3和640 x 480像素。SOM运行一个包含量化YOLOX、Anchor-to-Joint（A2J）和跌倒检测模型的三阶段管道。YOLOX从RGB帧中识别人体边界框，然后丢弃RGB帧以保护隐私。A2J使用深度帧来估计每个人的15个关节关键点。卷积神经网络（CNN）使用选定的关节坐标（x, y, z）来分类跌倒活动。YOLOX在CrowdHuman数据集上训练，A2J在ITOP、MP-3DHP、UR跌倒检测和自定义SDSU PSG数据集上训练，CNN在UR跌倒检测和SDSU PSG上训练。设计使用单核DPU与串行管道，以及双核DPU以多线程运行YOLOX和A2J。结果：使用IoU >= 50%评估YOLOX的量化精度，使用10厘米规则评估A2J的mAP，使用分类准确率（TP + TN）/(TP + TN + FP + FN）评估CNN的准确性。准确率分别为74%、84.13%和75.85%。通过将单线程管道的吞吐量从2.5 FPS提高到多线程版本的4.5 FPS。结论：结果表明在AMD Kria K26边缘设备上实现隐私保护的跌倒检测是可行的。设备上的HPE和跌倒分类无需依赖云计算，支持老年人监测和辅助医疗保健。未来的工作将提高模型的准确性和速度。
cs.CV / 2 / 2606.12562
HairPort: In-context 3D-aware Hair Import and Transfer for Images
HairPort：基于上下文的3D感知发型导入与转移
Abstract
Transferring hairstyles between images is an important but challenging task in computer graphics, computer vision, and visual effects. It enables users to explore new looks without physically altering their hair, with applications in virtual try-on systems, augmented reality, and entertainment. Most prior works operate best under small pose gaps, and they fall short under large viewpoint and scale differences, where missing hair content must be synthesized rather than transferred. We propose HairPort, a 3D-aware hairstyle transfer framework that attempts to solve these issues by explicitly separating hair removal from transfer and enforcing geometric consistency before synthesis. We introduce a Bald Converter, which produces realistic bald versions of faces through LoRA-based in-context adaptation of FLUX.1 Kontext. To train our Bald Converter, we introduce a new dataset, Baldy, containing 6,000 paired bald and original images across diverse identities and conditions. We also use a 3D-Aware Transfer Pipeline that reconstructs and re-renders the reference hairstyle from the target viewpoint before compositing it onto the source image. Being 3D aware, our method supports large pose and scale discrepancies between the source and target. Finally, a conditional flow-matching generator synthesizes the transferred result from the bald source and geometry-aligned reference guidance. Together, our method enables accurate, pose-consistent, and identity-preserving hairstyle transfer, outperforming existing methods both qualitatively and quantitatively.
Chinese Translation
在计算机图形学、计算机视觉和视觉效果领域，发型在图像之间的转移是一项重要但具有挑战性的任务。它使用户能够在不实际改变发型的情况下探索新形象，应用于虚拟试穿系统、增强现实和娱乐等领域。大多数先前的研究在小姿态差异下表现良好，但在大视角和尺度差异下则表现不佳，此时缺失的发型内容必须通过合成而非转移来补充。我们提出了HairPort，一个3D感知的发型转移框架，旨在通过明确将发型去除与转移分开，并在合成之前强制执行几何一致性来解决这些问题。我们引入了一种秃头转换器（Bald Converter），通过基于LoRA的上下文适应FLUX.1 Kontext，生成逼真的秃头面孔版本。为了训练我们的秃头转换器，我们引入了一个新的数据集Baldy，其中包含6000对不同身份和条件下的秃头和原始图像。我们还使用了一个3D感知转移管道，该管道在将参考发型合成到源图像之前，从目标视角重建和重新渲染参考发型。由于具备3D感知能力，我们的方法支持源图像和目标图像之间的大姿态和尺度差异。最后，一个条件流匹配生成器从秃头源图像和几何对齐的参考引导中合成转移结果。综上所述，我们的方法实现了准确、姿态一致且保持身份特征的发型转移，在定性和定量上均优于现有方法。
cs.CV / 3 / 2606.12575
High-Fidelity Two-Step Image Generation via Teacher-Aligned End-to-End Distillation
通过教师对齐的端到端蒸馏实现高保真度的两步图像生成
Abstract
Few-step diffusion distillation has become increasingly mature for 4-8-step generation, yet pushing further to 2 steps remains challenging. In this work, we introduce Z-Image Turbo++, a high-quality 2-step image generation model distilled from the 8-step Z-Image Turbo teacher. Our method addresses the central bottlenecks of increased task difficulty and limited model capacity in 2-step generation through three simple but effective design choices tailored to this regime. First, we propose Distribution-Aligned Adversarial Learning, which uses teacher-generated images rather than external real images as real samples for GAN training, providing a more attainable and informative adversarial target. Second, we adopt Step-Decoupled Parameterization, assigning independent model parameters to the two denoising steps to better match their distinct capacity demands. Third, we perform End-to-End Training with Iterative Regularization, allowing the first step to receive gradients from final image quality while preserving a meaningful intermediate generation through an explicit step-1 loss. Together, these designs substantially narrow the quality gap between 2-step and 8-step generation in both qualitative and quantitative evaluations, highlighting the potential of carefully tailored distillation strategies for improving the quality-efficiency trade-off in few-step generation.
Chinese Translation
少步扩散蒸馏在4-8步生成中逐渐成熟，但进一步推进到2步仍然具有挑战性。在本研究中，我们引入了Z-Image Turbo++，这是一个从8步Z-Image Turbo教师模型蒸馏而来的高质量2步图像生成模型。我们的方法通过三种简单但有效的设计选择，针对这一领域的任务难度增加和模型容量有限的核心瓶颈进行了改进。首先，我们提出了分布对齐对抗学习（Distribution-Aligned Adversarial Learning），该方法使用教师生成的图像而非外部真实图像作为GAN训练的真实样本，从而提供了更可达和信息丰富的对抗目标。其次，我们采用了步解耦参数化（Step-Decoupled Parameterization），为两个去噪步骤分配独立的模型参数，以更好地匹配它们各自的容量需求。第三，我们进行端到端训练与迭代正则化（End-to-End Training with Iterative Regularization），使得第一步能够接收来自最终图像质量的梯度，同时通过显式的步骤1损失保持有意义的中间生成。综合来看，这些设计在定性和定量评估中显著缩小了2步生成与8步生成之间的质量差距，突显了精心设计的蒸馏策略在改善少步生成中的质量与效率权衡的潜力。
cs.CV / 4 / 2606.12590
Analyzing and Improving Fine-grained Preference Optimization in Medical LVLMs
分析与改进医疗领域细粒度偏好优化的视觉语言模型
Abstract
Large Vision-Language Models (LVLMs) have achieved strong performance across medical imaging tasks, yet they remain prone to factual inconsistencies, poor visual grounding, and misalignment with clinically meaningful feedback. Existing post-training alignment approaches, including Direct Preference Optimization (DPO) and its variants, face three critical limitations in the medical domain: (1) sequence-level reward signals treat clinically critical tokens identically to generic filler text; (2) reliance on static supervised fine-tuning references as preferred responses introduces an off-policy distribution shift, steering optimization toward stylistic artifacts over clinical correctness; and (3) alignment objectives lack explicit visual grounding constraints, leaving models insensitive to subtle yet diagnostically decisive pathological features. Our method leverages a bidirectional token-wise KL regularizer alongside a visual-contrastive grounding objective that pairs clean and lesion-corrupted images to penalize responses generated without adequate visual evidence. Together, these components form a fine-grained, on-policy alignment framework that constructs preference pairs by minimally editing model-generated outputs, correcting only clinically erroneous spans while preserving the original linguistic style. Extensive experiments across medical imaging tasks and clinical text generation benchmarks validate the effectiveness of our approach.
Chinese Translation
大型视觉语言模型（LVLMs）在医疗影像任务中表现出色，但仍然容易出现事实不一致、视觉基础差以及与临床有意义反馈不对齐等问题。现有的后训练对齐方法，包括直接偏好优化（Direct Preference Optimization, DPO）及其变体，在医疗领域面临三大关键限制：（1）序列级奖励信号将临床关键标记与通用填充文本视为相同；（2）依赖静态监督微调参考作为偏好响应会引入离线分布偏移，使优化倾向于风格伪影而非临床正确性；（3）对齐目标缺乏明确的视觉基础约束，导致模型对微妙但具有诊断决定性的病理特征不敏感。我们的方法利用双向标记级KL正则化器以及视觉对比基础目标，该目标将干净图像与病变图像配对，以惩罚在没有充分视觉证据的情况下生成的响应。这些组件共同构成了一个细粒度的在线对齐框架，通过最小编辑模型生成的输出构建偏好对，仅纠正临床错误的部分，同时保留原始语言风格。我们在医疗影像任务和临床文本生成基准上的广泛实验验证了我们方法的有效性。
cs.CV / 5 / 2606.12601
Dual-State Slot Attention: Decoupling Appearance and Identity for Video Object-Centric Learning
双状态插槽注意力：解耦外观与身份以实现视频对象中心学习
Abstract
Unsupervised video object-centric learning aims to decompose dynamic scenes into persistent, object-level representations without supervision. However, existing slot-based methods struggle to maintain stable object identity in challenging settings such as rapid motion and partial occlusion. First, they typically encode both the per-frame appearance of an object and its identity across frames in a single slot vector, creating an objective conflict that leads to slot swapping: reconstruction requires sensitivity to transient visual changes, whereas temporal consistency requires invariance to them. Second, the token renormalization used in Slot Attention can amplify weakly attending slots, allowing them to absorb tokens from other objects and destabilize slot-to-object correspondence. We propose Dual-State Slot Attention (DSSA), a fully self-supervised framework that addresses these limitations by separating appearance from identity and by reducing spurious updates from weakly matching slots. DSSA decomposes each slot into a local state for per-frame appearance and an identity state for temporally stable object information, thereby aligning reconstruction and temporal consistency with separate representations. The identity state is updated through a learned recurrent transition that acts as a temporal filter on the local state, while competition-modulated aggregation (CMA) down-weights updates from weakly matching slots and prevents them from absorbing tokens from other objects. Experiments on MOVi-C, MOVi-D, and YouTube-VIS demonstrate that DSSA consistently improves segmentation quality and temporal consistency over prior methods, while also yielding stronger downstream object recognition and video dynamics prediction. Code and models will be made publicly available upon acceptance.
Chinese Translation
无监督视频对象中心学习旨在将动态场景分解为持久的对象级表示，而无需监督。然而，现有的基于插槽的方法在快速运动和部分遮挡等挑战性环境中难以保持稳定的对象身份。首先，它们通常将对象的每帧外观和跨帧的身份编码到一个单一的插槽向量中，这造成了目标冲突，导致插槽交换：重建要求对瞬时视觉变化敏感，而时间一致性则要求对其不变。其次，插槽注意力中使用的令牌重归一化可能会放大弱关注插槽，使其能够吸收来自其他对象的令牌，从而破坏插槽与对象之间的对应关系。我们提出了双状态插槽注意力（Dual-State Slot Attention, DSSA），这是一个完全自我监督的框架，通过将外观与身份分离并减少来自弱匹配插槽的虚假更新来解决这些限制。DSSA将每个插槽分解为用于每帧外观的局部状态和用于时间稳定对象信息的身份状态，从而使重建和时间一致性与单独的表示对齐。身份状态通过一个学习的递归转移进行更新，该转移作为局部状态的时间滤波器，而竞争调制聚合（Competition-Modulated Aggregation, CMA）则降低来自弱匹配插槽的更新权重，防止其吸收来自其他对象的令牌。在MOVi-C、MOVi-D和YouTube-VIS上的实验表明，DSSA在分割质量和时间一致性方面始终优于先前的方法，同时也带来了更强的下游对象识别和视频动态预测能力。代码和模型将在接受后公开发布。
cs.CV / 6 / 2606.12628
Context-Aware Feature-Fusion for Co-occurring Object Detection in Autonomous Driving
面向上下文的特征融合用于自动驾驶中的共现物体检测
Abstract
Object detection in autonomous driving requires precise localization and an inherent understanding of the relational context between co-occurring objects. In extremely complex heterogeneous environments rare classes, small-scale objects, and frequently appearing objects are difficult for standard object detection frameworks to handle. In this paper, we propose a novel framework called Context-Centric Feature Fusion (CCFF), which utilizes two attention-based modules, Local Context Fusion Module (LCFM) uses the RoI-to-RoI self-attention mechanism to resolve spatial interactions, mainly considering small and partially obscured objects, while Global Context Attention Module (GCAM) converts the co-occurrence of objects priors by pooling top-K RoI features into a global context attention token, avoiding the computational overhead of pixel-level global pooling. This fusion of local and object-centric global features yields contextualized embeddings that enhance classification results and co-occurring objects detection. Our method is evaluated on two datasets, Cityscapes and BDD100K which demonstrate significant improvement on relational consistency, achieving a Category-level Consistency Strategy (CCS) of 0.973 and 0.969, respectively. Furthermore, our approach produces substantial gains in small object detection (AP_S: 14.1%) and successfully recovers rare classes such as "Train" that are typically lost in large distributions. Our efficiency report shows that the framework processes images in real time with a 0.2 FPS overhead. The code is available at https://github.com/BinayKSingh/CCFF.
Chinese Translation
自动驾驶中的物体检测需要精确的定位以及对共现物体之间关系上下文的内在理解。在极其复杂的异质环境中，稀有类别、小规模物体和频繁出现的物体对于标准物体检测框架来说是难以处理的。本文提出了一种新颖的框架，称为以上下文为中心的特征融合（Context-Centric Feature Fusion, CCFF），该框架利用两个基于注意力的模块：局部上下文融合模块（Local Context Fusion Module, LCFM）使用RoI到RoI的自注意力机制来解决空间交互，主要考虑小型和部分被遮挡的物体，而全局上下文注意力模块（Global Context Attention Module, GCAM）通过将前K个RoI特征汇聚成一个全局上下文注意力标记来转换物体的共现先验，避免了像素级全局汇聚的计算开销。这种局部和以物体为中心的全局特征的融合产生了上下文化的嵌入，增强了分类结果和共现物体的检测。我们在两个数据集Cityscapes和BDD100K上评估了我们的方法，结果显示在关系一致性上有显著改善，分别达到了0.973和0.969的类别级一致性策略（Category-level Consistency Strategy, CCS）。此外，我们的方法在小物体检测上也取得了显著提升（AP_S: 14.1%），并成功恢复了通常在大分布中丢失的稀有类别，如“火车”（Train）。我们的效率报告显示，该框架以实时处理图像，额外开销为0.2 FPS。代码可在https://github.com/BinayKSingh/CCFF获取。
cs.CV / 7 / 2606.12633
ECA: Efficient Continual Alignment for Open-Ended Image-to-Text Generation
ECA：开放式图像到文本生成的高效持续对齐
Abstract
Incremental Learning (IL) for Open-ended Image-to-Text Generation (OpenITG) enables models to continuously generate accurate, contextually relevant text for new images while preserving previously acquired knowledge. Unlike prior studies, this paper addresses a more practical scenario in which the predominant category of visual data shifts over time as environments evolve. In this context, we introduce a new notion of continual alignment, which incrementally adapts the alignment module within pre-trained VLMs to preserve high-quality cross-modal representations. Based on this idea, we propose Efficient Continual Alignment (ECA), a novel exemplar-free IL approach for OpenITG. The key challenge is enabling the model to acquire new, task-specific features while minimizing interference with the established alignment without accessing raw data from previous tasks. To address this, ECA employs three core mechanisms: a Mixture of Query (MoQ) module that adapts task-specific query tokens, a Fisher Dynamic Expansion (FeDEx) that dynamically expands model structure based on a Fisher Information Matrix (FIM)-based metric, and an embedding dictionary with Dictionary Replay (DR) to retain past knowledge. To evaluate ECA's performance, we construct four new IL OpenITG benchmarks that better reflect real-world scenarios. Experimental results demonstrate that ECA significantly mitigates catastrophic forgetting and improves IL performance compared to baseline methods. Code and benchmarks are available at https://github.com/Snowball0823/ECA.
Chinese Translation
开放式图像到文本生成（OpenITG）的增量学习（IL）使模型能够持续为新图像生成准确且具有上下文相关性的文本，同时保留先前获得的知识。与之前的研究不同，本文解决了一个更实际的场景，即随着环境的演变，视觉数据的主要类别随时间发生变化。在此背景下，我们引入了一种新的持续对齐概念，该概念逐步调整预训练视觉语言模型（VLM）中的对齐模块，以保持高质量的跨模态表示。基于这一理念，我们提出了高效持续对齐（ECA），这是一种新颖的无示例增量学习方法，适用于OpenITG。关键挑战在于使模型能够获取新的、特定任务的特征，同时在不访问先前任务的原始数据的情况下，最小化对已建立对齐的干扰。为了解决这个问题，ECA采用了三个核心机制：一个混合查询（MoQ）模块，适应特定任务的查询标记；一个基于费舍尔信息矩阵（FIM）度量的费舍尔动态扩展（FeDEx），动态扩展模型结构；以及一个带有字典重放（DR）的嵌入字典，以保留过去的知识。为了评估ECA的性能，我们构建了四个新的增量学习OpenITG基准，更好地反映现实世界场景。实验结果表明，与基线方法相比，ECA显著减轻了灾难性遗忘，并提高了增量学习性能。代码和基准可在 https://github.com/Snowball0823/ECA 获取。
cs.CV / 8 / 2606.12635
CD-RCM: Generalizable Continuous-Depth Novel View Synthesis for Reflectance Confocal Microscopy
CD-RCM：用于反射共聚焦显微镜的可推广连续深度新视图合成
Abstract
Reflectance confocal microscopy (RCM) provides noninvasive, cellular-resolution "optical biopsies" of human skin \emph{in vivo} by acquiring en-face images at successive depths, forming a sparse z-stack. Due to optical limitations, these stacks are anisotropic 3D volumes with lateral resolution (0.5 $\mu$m) $\sim$6 times higher compared to axial resolution, which is defined by the optical sectioning (3 $\mu$m), limiting the interpretation of tissue. Our goal is to provide continuous-depth visualization by interpolating intermediate sections and making the 3D volume isotropic. Such a representation permits arbitrary-direction sectioning, including histopathology-like cross-sectional examination, without requiring per-patient optimization. To that end, we introduce the first RCM-specific novel-view synthesis (NVS) approach, CD-RCM, a feedforward model that predicts realistic, unseen depths from sparsely sampled RCM stacks. Classical neural rendering methods focus on reconstruction from surface-level multi-view observations. In contrast to surface-level camera views, RCM can acquire optically sectioned en-face images of tissue beyond the surface up to 200 $\mu$m. However, during visualization of the RCM stacks, observations of the shallower sections (towards the surface) obscure the deeper ones. This unique axial imaging geometry and layer-dependent anatomical organization motivated our development of a tailored architectural and training framework that explicitly accounts for RCM's depth-resolved, occlusive imaging physics. Experiments demonstrate that CD-RCM achieves high-fidelity novel-view synthesis with sub-second inference time.
Chinese Translation
反射共聚焦显微镜（RCM）通过在连续深度获取面内图像，提供对人类皮肤的非侵入性细胞分辨率“光学活检”，形成稀疏的 z 堆栈。由于光学限制，这些堆栈是各向异性的 3D 体积，其横向分辨率（0.5 μm）约为轴向分辨率的 6 倍，轴向分辨率由光学切片（3 μm）定义，这限制了对组织的解读。我们的目标是通过插值中间切片提供连续深度可视化，使 3D 体积各向同性。这种表示允许任意方向的切片，包括类似组织病理学的横截面检查，而无需对每位患者进行优化。为此，我们提出了首个针对 RCM 的新视图合成（NVS）方法 CD-RCM，这是一种前馈模型，能够从稀疏采样的 RCM 堆栈中预测真实的、未见过的深度。经典的神经渲染方法侧重于从表面级多视图观察中进行重建。与表面级相机视图相比，RCM 可以获取组织表面以下最多 200 μm 的光学切片面内图像。然而，在可视化 RCM 堆栈时，浅层切片（靠近表面）的观察会遮挡更深层的切片。这种独特的轴向成像几何和层依赖的解剖组织结构促使我们开发了一个专门的架构和训练框架，明确考虑了 RCM 的深度分辨、遮挡成像物理。实验表明，CD-RCM 实现了高保真度的新视图合成，推理时间低于一秒。
cs.CV / 9 / 2606.12671
SalArt-VQA: Diagnosing Whether VLMs Understand Salient Artifacts in Generated Images
SalArt-VQA：诊断视觉语言模型是否理解生成图像中的显著伪影
Abstract
Vision-language models (VLMs) are increasingly used to detect whether AI-generated images contain visible artifacts, yet their ability to analyze such artifacts remains poorly understood. A correct image-level decision can still hide important failures: a model may correctly flag an artifact while relying on the wrong visual cue, selecting the wrong region, or describing a defect that the image does not support. To evaluate these behaviors directly, we introduce SalArt-VQA, a diagnostic benchmark for fine-grained SALient ARTifact understanding in AI-generated images. SalArt-VQA contains 950 images and 3,681 human-authored multiple-choice questions spanning artifact images, matched real reference images, and paired generated reference images. Four aligned question types evaluate presence detection, semantic localization, spatial grounding, and evidence-grounded defect identification, while the reference splits test calibration and abstention when the annotated defect is absent. Across 20 VLMs, SalArt-VQA reveals failures that image-level detection accuracy hides: the strongest model reaches 99.37% detection recall on artifact images but answers all four artifact-side questions correctly on only 53.26% of images. Comparing artifact images with artifact-free references reveals a sensitivity-calibration tradeoff: sensitive models often make unsupported artifact claims, while conservative models avoid false alarms largely by missing real artifacts. These results show that high artifact detection accuracy alone does not imply grounded artifact understanding. SalArt-VQA exposes these hidden failure modes and provides a fine-grained evaluation of whether VLM artifact claims are supported by local visual evidence.
Chinese Translation
视觉语言模型（VLMs）越来越多地用于检测AI生成的图像中是否存在可见的伪影，但它们分析这些伪影的能力仍然不够清楚。一个正确的图像级决策可能仍然掩盖重要的失败：模型可能正确标记一个伪影，但依赖错误的视觉线索、选择错误的区域，或描述图像不支持的缺陷。为了直接评估这些行为，我们引入了SalArt-VQA，这是一个用于AI生成图像中细粒度显著伪影理解的诊断基准。SalArt-VQA包含950张图像和3,681个由人类撰写的多项选择题，涵盖伪影图像、匹配的真实参考图像和成对的生成参考图像。四种对齐的问题类型评估存在检测、语义定位、空间定位和基于证据的缺陷识别，而参考分割测试在标注缺陷缺失时的校准和放弃。在20个VLMs中，SalArt-VQA揭示了图像级检测准确性所隐藏的失败：最强模型在伪影图像上的检测召回率达到99.37%，但在53.26%的图像上仅正确回答所有四个伪影相关问题。将伪影图像与无伪影参考图像进行比较揭示了敏感性与校准之间的权衡：敏感模型往往会做出不支持的伪影声明，而保守模型则主要通过错过真实伪影来避免误报。这些结果表明，仅仅高的伪影检测准确性并不意味着对伪影的理解是有根据的。SalArt-VQA揭示了这些隐藏的失败模式，并提供了对VLM伪影声明是否得到局部视觉证据支持的细粒度评估。
cs.CV / 10 / 2606.12706
VLADriveBench: Evaluating CoT-Action Relationship in VLA for Autonomous Driving
VLADriveBench：评估自主驾驶中VLA的CoT-行动关系
Abstract
Vision-language-action (VLA) models generate chain-of-thought (CoT) reasoning alongside driving trajectories, but existing benchmarks evaluate only trajectory quality and do not assess whether the CoT is relevant, consistent, or causally connected to the driving action. We introduce VLADriveBench, a framework that combines observational metrics (mentioning, hallucination, contradiction, action alignment) with a CoT intervention protocol to provide complementary views of the CoT-action relationship. Applying VLADriveBench to three models across two architectures, we find that the two analyses can diverge sharply: ORION scores highest on observational alignment yet its CoT is epiphenomenal, while Alpamayo v1.5 scores lower yet its CoT is strongly causal, with visual salience gating the extent of CoT influence.
Chinese Translation
视觉-语言-行动（VLA）模型生成与驾驶轨迹相关的思维链（CoT）推理，但现有基准仅评估轨迹质量，而未评估CoT是否与驾驶行动相关、一致或因果连接。我们提出了VLADriveBench，一个结合观察性指标（提及、幻觉、矛盾、行动对齐）与CoT干预协议的框架，以提供CoT-行动关系的互补视角。将VLADriveBench应用于两种架构下的三个模型，我们发现这两种分析可能会显著分歧：ORION在观察性对齐上得分最高，但其CoT是附带现象，而Alpamayo v1.5得分较低，但其CoT具有强因果性，视觉显著性限制了CoT影响的程度。
cs.CV / 11 / 2606.12744
GRIP: Feedback-Guided Prompt Retrieval for Large Multimodal Models
GRIP：用于大型多模态模型的反馈引导提示检索
Abstract
In-Context Learning (ICL) has become a powerful mechanism for adapting Large Language Models (LLMs) to new tasks without fine-tuning. Extending this concept to Large Multimodal Models (LMMs), Multimodal In-Context Learning (M-ICL) relies on retrieving relevant examples, such as images, captions, or question-answer pairs, to guide predictions across tasks like classification, captioning, and visual question answering (VQA). Most existing approaches select in-context examples based on feature-space similarity, assuming that semantically similar samples provide the most useful context. However, our systematic analysis reveals that this assumption does not always hold: visually similar examples are not necessarily those that most effectively enhance in-context learning performance. To address this, we propose the Guided Retrieval of In-context Prompts (GRIP), a learnable vision-only retrieval framework that leverages feedback from LMMs to identify examples that truly improve model predictions. GRIP learns to distinguish beneficial from detrimental in-context examples through contrastive training, refining retrieval beyond pure similarity. Across three multimodal tasks, namely classification, captioning, and VQA, GRIP improves consistently over similarity-based retrieval on Qwen2.5-VL-7B, with its strongest gains in classification on Idefics2-8B. Moreover, we demonstrate that retrievers trained with feedback from one open LMM can be transferred to other models without retraining, including closed-source GPT-4o and Gemini, enabling scalable and cost-efficient deployment of M-ICL. Code will be published upon acceptance.
Chinese Translation
上下文学习（In-Context Learning, ICL）已成为一种强大的机制，使大型语言模型（Large Language Models, LLMs）能够在不进行微调的情况下适应新任务。将这一概念扩展到大型多模态模型（Large Multimodal Models, LMMs），多模态上下文学习（Multimodal In-Context Learning, M-ICL）依赖于检索相关示例，如图像、标题或问答对，以指导分类、图像描述和视觉问答（Visual Question Answering, VQA）等任务的预测。大多数现有方法基于特征空间相似性选择上下文示例，假设语义上相似的样本提供最有用的上下文。然而，我们的系统分析表明，这一假设并不总是成立：视觉上相似的示例不一定是最有效提升上下文学习性能的示例。为了解决这个问题，我们提出了上下文提示的引导检索（Guided Retrieval of In-context Prompts, GRIP），这是一种可学习的仅视觉检索框架，利用来自LMM的反馈来识别真正改善模型预测的示例。GRIP通过对比训练学习区分有益和有害的上下文示例，超越了纯相似性的检索。在分类、图像描述和VQA这三项多模态任务中，GRIP在Qwen2.5-VL-7B上始终优于基于相似性的检索，尤其在Idefics2-8B的分类任务中取得了显著提升。此外，我们证明了使用来自一个开放LMM的反馈训练的检索器可以在不重新训练的情况下转移到其他模型，包括闭源的GPT-4o和Gemini，从而实现M-ICL的可扩展和成本效益的部署。代码将在接受后发布。
cs.CV / 12 / 2606.12826
DIMOS: Disentangling Instance-level Moving Object Segmentation
DIMOS：解耦实例级移动物体分割
Abstract
Moving instance segmentation (MIS) attracts increasing attention due to its broad applications in traffic surveillance, autonomous driving, and animal tracking. Event cameras record asynchronous brightness changes, providing high temporal resolution and dynamic range, which makes them highly sensitive to motion information. By fusing event and image features, motion cues from events can complement spatial details from images, enhancing the performance of MIS. However, current multimodal MIS methods still struggle to segment small moving instances, as event cameras often yield sparse features under limited resolution. Moreover, event features entangle appearance attributes with motion cues, which further restricts effective cross-modal fusion. To address these challenges, we first propose a dual-disentangling feature extraction framework that separates and extracts appearance and motion information within both image and event modalities, thereby improving feature density. Subsequently, a multi-granularity cross-modal alignment is introduced to align distributionally and semantically consistent features across modalities, enabling more effective fusion with rich spatial and temporal details. The experiment results demonstrate that our method achieves state-of-the-art performance in multimodal MIS, especially for small instances under challenging conditions such as fast motion and low-light settings.
Chinese Translation
移动实例分割（MIS）因其在交通监控、自动驾驶和动物追踪等广泛应用而受到越来越多的关注。事件相机记录异步亮度变化，提供高时间分辨率和动态范围，使其对运动信息高度敏感。通过融合事件和图像特征，事件中的运动线索可以补充图像中的空间细节，从而提升MIS的性能。然而，当前的多模态MIS方法在分割小型移动实例时仍面临挑战，因为事件相机在有限分辨率下通常产生稀疏特征。此外，事件特征将外观属性与运动线索纠缠在一起，进一步限制了有效的跨模态融合。为了解决这些挑战，我们首先提出了一种双重解耦特征提取框架，该框架在图像和事件模态中分离并提取外观和运动信息，从而提高特征密度。随后，引入了一种多粒度跨模态对齐方法，以对齐跨模态中在分布和语义上相一致的特征，从而实现更有效的融合，具有丰富的空间和时间细节。实验结果表明，我们的方法在多模态MIS中实现了最先进的性能，特别是在快速运动和低光照等挑战性条件下的小型实例分割方面。
cs.CV / 13 / 2606.12830
Perceive, Interact, Reason: Building Tool-Augmented Visual Agents for Spatial Reasoning
感知、互动、推理：构建工具增强的视觉智能体以实现空间推理
Abstract
While recent vision-language models (VLMs) demonstrate strong multimodal understanding, they remain limited in spatial reasoning tasks that require active evidence acquisition and multi-step visual interaction. This limitation suggests that relying solely on implicit visual representations from vision encoders is insufficient for recovering fine-grained spatial evidence. We introduce PERception-Interaction-reason Agent (PERIA), a tool-augmented visual agent for spatial reasoning tasks across map reasoning, visual probing, and vision reconstruction. PERIA uses two lightweight tool families: vision perception tools for exposing textual, symbolic, and spatial evidence, and vision interaction tools for manipulating visual context, tracing paths, and verifying spatial relations. To train PERIA, we develop a unified recipe that combines supervised tool-use trajectory synthesis, composite rewards, and Observation-Relaxed Group-in-Group Policy Optimization (OR-GIGPO) for effective multi-tool behavior. Experiments on 13 benchmarks from 8 datasets show that PERIA-8B improves over the Qwen3-8B backbone by 10.0% on in-distribution benchmarks and 4.4% on out-of-distribution benchmarks, while outperforming previous state-of-the-art baselines of similar size by 7.0%-14.8%. It also achieves performance comparable to much larger models such as Qwen3-VL-235B-A22B-Thinking and GPT-5, demonstrating the effectiveness of PERIA in enhancing spatial reasoning capabilities.
Chinese Translation
尽管近期的视觉-语言模型（VLMs）展示了强大的多模态理解能力，但在需要主动获取证据和多步骤视觉互动的空间推理任务中，它们仍然存在局限性。这一局限性表明，仅依赖视觉编码器的隐式视觉表征不足以恢复细粒度的空间证据。我们提出了感知-互动-推理智能体（PERIA），这是一个用于空间推理任务的工具增强视觉智能体，涵盖地图推理、视觉探测和视觉重建。PERIA使用两种轻量级工具系列：视觉感知工具用于揭示文本、符号和空间证据，视觉互动工具用于操控视觉上下文、追踪路径和验证空间关系。为了训练PERIA，我们开发了一种统一的方法，结合了监督工具使用轨迹合成、复合奖励和观察放松的组内组策略优化（OR-GIGPO），以实现有效的多工具行为。在来自8个数据集的13个基准测试上的实验表明，PERIA-8B在分布内基准测试中比Qwen3-8B基线提高了10.0%，在分布外基准测试中提高了4.4%，同时在相似规模的先前最先进基线中超越了7.0%-14.8%。它的性能也与更大模型如Qwen3-VL-235B-A22B-Thinking和GPT-5相当，展示了PERIA在增强空间推理能力方面的有效性。
cs.CV / 14 / 2606.12847
Language-Guided Abstraction for Visual Reasoning
语言引导的视觉推理抽象
Abstract
The Abstraction and Reasoning Corpus (ARC) is viewed as a critical avenue to Artificial General Intelligence (AGI), as it enables models to learn abstract transformation rules from few-shot examples and then generalize to new tasks. However, prevalent ARC methodology is either pure language or vision-only (i.e., VARC). The former depends heavily on LLMs, consuming billions of parameters. The latter often struggles to capture high-level semantics, leading to overfitting on pixel-level patterns. To bridge this gap, we propose L-VARC, a novel framework that enhances visual reasoning via a language-guided Learning Using Privileged Information (LUPI) branch. Specifically, we design a Semantic Compression Module by feeding a unified, task-agnostic prompt into DeepSeek-V3. In this way, the raw LARC (a crowd-sourced language description dataset) can be substantially refined and structured, fitting with the context length constraint of standard text encoders (e.g., CLIP). Moreover, we design a Cross-Attention Projector to align visual features with semantic embeddings, aiming to guide the training of the ARC model. Notably, the LUPI branch is taken in the training process and will be discarded during inference, thereby yielding a lightweight model with a mere 18 million parameters. Extensive experiments demonstrate that our L-VARC effectively leverages linguistic priors to boost visual reasoning and outperforms state-of-the-art. Ablation studies further confirm the contribution of the two new designs towards the L-VARC framework. The code is available at https://github.com/GZHU-DVL/L-VARC.
Chinese Translation
抽象与推理语料库（ARC）被视为通向人工通用智能（AGI）的关键途径，因为它使模型能够从少量示例中学习抽象转换规则，并随后推广到新任务。然而，现有的ARC方法要么是纯语言的，要么是仅基于视觉的（即VARC）。前者严重依赖大型语言模型（LLMs），消耗数十亿参数；后者往往难以捕捉高层次的语义，导致在像素级模式上过拟合。为了解决这一问题，我们提出了L-VARC，一个通过语言引导的特权信息学习（LUPI）分支来增强视觉推理的新框架。具体而言，我们通过将统一的、与任务无关的提示输入DeepSeek-V3，设计了一个语义压缩模块。通过这种方式，原始的LARC（一个众包的语言描述数据集）可以被显著精炼和结构化，以适应标准文本编码器（如CLIP）的上下文长度限制。此外，我们设计了一个跨注意力投影器，以将视觉特征与语义嵌入对齐，旨在指导ARC模型的训练。值得注意的是，LUPI分支在训练过程中使用，在推理时将被丢弃，从而产生一个仅有1800万参数的轻量级模型。大量实验表明，我们的L-VARC有效利用语言先验来提升视觉推理，并超越了现有的最先进水平。消融研究进一步确认了这两个新设计对L-VARC框架的贡献。代码可在https://github.com/GZHU-DVL/L-VARC获取。
cs.CV / 15 / 2606.12869
Learning Task-Aware Sampling with Shared Saliency through Density-Equalizing Mappings
通过密度均衡映射学习任务感知采样与共享显著性
Abstract
In image and surface-based learning tasks, convolutional features are typically extracted using receptive fields that are sampled uniformly across the entire domain. However, informative structures are rarely distributed uniformly in practice and are often concentrated in localized regions. Such phenomena are particularly common in medical imaging, where pathological changes are spatially confined. Consequently, uniform convolution allocates equal computational effort to both informative and uninformative regions, resulting in inefficient feature extraction and suboptimal utilization of model capacity. To address this issue, we propose a framework for task-adaptive sampling that dynamically redistributes computational attention according to the spatial importance of the data. Specifically, we introduce the Density-Equalizing Convolutional Neural Network (DECNN), which employs density-equalizing mappings to guide convolution through a learned density function. The density function encodes the relative importance of different regions and induces a transformation that enlarges informative areas while compressing less relevant ones. As a result, convolutional receptive fields are redistributed non-uniformly over the domain, enabling denser sampling in task-relevant regions. By coupling this importance-driven transformation with convolution, DECNN performs adaptive feature extraction that focuses computational resources on informative structures. This leads to more efficient use of model capacity, yielding a lightweight yet expressive architecture while simultaneously producing an interpretable saliency map. Experiments on image classification and craniofacial surface analysis demonstrate that DECNN achieves competitive or superior performance with fewer parameters, accurately identifies task-relevant regions, and remains robust under complex geometric variations.
Chinese Translation
在图像和基于表面的学习任务中，卷积特征通常是通过在整个领域均匀采样的感受野提取的。然而，信息丰富的结构在实际中很少均匀分布，通常集中在局部区域。这种现象在医学成像中尤为常见，其中病理变化在空间上是局限的。因此，均匀卷积对信息丰富和无信息区域分配了相同的计算努力，导致特征提取效率低下和模型能力的次优利用。为了解决这个问题，我们提出了一种任务自适应采样框架，该框架根据数据的空间重要性动态重新分配计算注意力。具体而言，我们引入了密度均衡卷积神经网络（Density-Equalizing Convolutional Neural Network, DECNN），它利用密度均衡映射通过学习的密度函数引导卷积。密度函数编码了不同区域的相对重要性，并诱导出一种变换，扩大信息丰富区域的同时压缩不太相关的区域。因此，卷积感受野在领域上非均匀地重新分配，使得在任务相关区域进行更密集的采样。通过将这种基于重要性的变换与卷积相结合，DECNN实现了自适应特征提取，将计算资源集中在信息结构上。这导致模型能力的更有效利用，产生一种轻量且富有表现力的架构，同时生成可解释的显著性图。在图像分类和颅面表面分析的实验中，DECNN以更少的参数实现了具有竞争力或更优的性能，准确识别任务相关区域，并在复杂几何变化下保持鲁棒性。
cs.CV / 16 / 2606.12886
Bridging Modal Isolation in Interleaved Thinking: Supervising Modality Transitions via Stepwise Reinforcement
跨越交错思维中的模态隔离：通过逐步强化监督模态转变
Abstract
Interleaved thinking, where a unified multimodal model alternates between textual reasoning and visual generation, has shown promise on spatial and physical tasks. However, in complex long-chain scenarios, we identify a fundamental failure mode: generated images diverge from the textual context while subsequent text ignores the visual evidence, causing the two modalities to alternate without genuinely informing each other. We term this Modal Isolation and attribute it to compounding information loss at modality boundaries. We decompose each reasoning cycle into atomic operations and define modality transition loss, quantifying cross-modal hallucination (text-to-image) and visual utilization deficit (image-to-text) at each boundary. We propose MoTiF (Modality Tiransition Fidelity), a two-stage training framework that directly optimizes these transitions: Reflective SFT trains the model to detect and recover from erroneous visual outputs; Flow-GRPO improves image generation fidelity via reinforcement learning. All training signals in MoTiF derive from transition-level fidelity rather than end-task accuracy. Across four visual puzzle benchmarks, this transition-level supervision substantially improves both cross-modal coherence and final task accuracy. The results demonstrate that effective interleaved reasoning requires explicit structural supervision at modality boundaries, not merely scaling or end-task optimization.
Chinese Translation
交错思维是一种统一的多模态模型在文本推理和视觉生成之间交替的方式，已在空间和物理任务中展现出潜力。然而，在复杂的长链场景中，我们识别出一种根本性的失败模式：生成的图像与文本上下文偏离，而后续文本忽视视觉证据，导致两种模态交替而没有真正相互启发。我们将此称为模态隔离，并将其归因于模态边界处信息损失的累积。我们将每个推理周期分解为原子操作，并定义模态转变损失，量化每个边界处的跨模态幻觉（文本到图像）和视觉利用缺失（图像到文本）。我们提出了MoTiF（模态转变保真度），一个两阶段的训练框架，直接优化这些转变：反思性SFT训练模型检测并恢复错误的视觉输出；Flow-GRPO通过强化学习提高图像生成的保真度。MoTiF中的所有训练信号均源于转变级别的保真度，而非最终任务的准确性。在四个视觉难题基准测试中，这种转变级别的监督显著提高了跨模态一致性和最终任务的准确性。结果表明，有效的交错推理需要在模态边界处进行明确的结构监督，而不仅仅是规模化或最终任务优化。
cs.CV / 17 / 2606.12898
Magnifying What Matters: Attention-Guided Adaptive Rendering for Visual Text Comprehension
放大重要内容：基于注意力引导的自适应渲染用于视觉文本理解
Abstract
Visual Text Comprehension (VTC) renders text into images for a vision-language model (VLM) to read, sidestepping LLM context-window limits and powering applications from long-page OCR to multi-page memory QA. Yet existing VTC pipelines treat rendering and layout as a fixed, content-agnostic preprocessing step and offer little mechanistic understanding of how VLMs internally process visualized text. Through a focused empirical study on VTC QA tasks, we reveal that VLMs exhibit a localization-without-utilization regime: evidence-localizing attention emerges sharply in the middle-to-late layers and is largely decoupled from answer correctness, yet simply enlarging the localized spans on the rendered page recovers a large fraction of the failures. Building on these observations, we propose AGAR (Attention-Guided Adaptive Rendering), a training-free, model-agnostic method that leverages a VLM's own middle-to-late layer attention to identify the top-K important visual patches, maps them back to word spans, and re-renders the page with those spans enlarged before re-inferring the answer. Extensive experiments across nine VTC benchmarks (short-form, long-context, and multi-page memory QA) and four VLM backbones show that AGAR (i)consistently improves off-the-shelf VLMs as a plug-and-play enhancement, (ii)composes with VLM post-training to yield further gains, and (iii)remains robust under both visual- and text-side input degradation.
Chinese Translation
视觉文本理解（Visual Text Comprehension, VTC）将文本渲染为图像，以便视觉语言模型（Vision-Language Model, VLM）进行阅读，绕过大型语言模型（Large Language Model, LLM）的上下文窗口限制，并推动从长页面光学字符识别（OCR）到多页面记忆问答（QA）的应用。然而，现有的 VTC 流程将渲染和布局视为固定的、与内容无关的预处理步骤，且对 VLM 如何内部处理可视化文本的机制理解甚少。通过对 VTC QA 任务的集中实证研究，我们揭示了 VLM 展现出一种“定位而不利用”的机制：证据定位注意力在中后层中明显出现，并且与答案的正确性大体上是解耦的，但仅仅扩大渲染页面上定位的范围就能恢复大量失败的结果。基于这些观察，我们提出了 AGAR（基于注意力引导的自适应渲染），这是一种无训练、模型无关的方法，利用 VLM 自身中后层的注意力来识别前 K 个重要的视觉补丁，将其映射回单词范围，并在重新推断答案之前扩大这些范围重新渲染页面。在九个 VTC 基准（短文本、长上下文和多页面记忆问答）和四个 VLM 主干模型上的广泛实验表明，AGAR (i) 作为即插即用的增强方式，持续改善现成的 VLM，(ii) 与 VLM 后训练组合以获得进一步提升，(iii) 在视觉和文本输入降级的情况下仍然保持稳健。
cs.CV / 18 / 2606.12925
Multi-Label Test-Time Adaptation with Bayesian Conditional Priors
基于贝叶斯条件先验的多标签测试时适应
Abstract
Multi-label recognition with frozen Vision-Language Models (VLMs) is brittle under distribution shift: standard zero-shot inference scores labels independently, ignoring co-occurrence structure and producing incoherent label sets where dominant concepts suppress weaker but compatible labels. We introduce Bayesian Conditional Priors (BCP) Estimation, a gradient-free test-time adaptation method that injects label dependency without tuning the backbone. BCP views zero-shot logits as a proxy for marginal posteriors under a fixed image-text likelihood and attributes shift-induced errors mainly to a mismatched label prior. For each test image, it selects a high-confidence anchor label and applies an anchor-conditioned Bayesian refinement. This update is closed-form in logit space and admits a pointwise mutual information (PMI) interpretation, explicitly promoting compatible labels and suppressing incompatible ones. BCP operates without target annotations by estimating anchor-conditioned priors online from the unlabeled test stream via lightweight second-order co-occurrence statistics, adding negligible overhead beyond a single forward pass. Across standard multi-label benchmarks and multiple CLIP backbones, BCP consistently outperforms strong TTA baselines, e.g., improving RN50 average mAP from 57.31 to 69.22 and ViT-B/16 from 62.61 to 71.79.
Chinese Translation
在分布变化下，使用冻结的视觉-语言模型（VLM）进行多标签识别的表现脆弱：标准的零-shot推理独立地评估标签，忽略了共现结构，导致产生不一致的标签集，其中主导概念抑制了较弱但兼容的标签。我们提出了贝叶斯条件先验（BCP）估计，这是一种无梯度的测试时适应方法，通过不调整主干网络来注入标签依赖性。BCP将零-shot logits视为在固定的图像-文本似然下的边际后验的代理，并将因分布变化引起的错误主要归因于标签先验的不匹配。对于每个测试图像，它选择一个高置信度的锚标签，并应用锚条件的贝叶斯细化。此更新在logit空间中为封闭形式，并具有逐点互信息（PMI）解释，明确促进兼容标签并抑制不兼容标签。BCP通过从未标记的测试流中在线估计锚条件先验而无需目标注释，利用轻量级的二阶共现统计，增加的开销可忽略不计，仅需一次前向传播。在标准多标签基准测试和多个CLIP主干网络上，BCP始终优于强大的测试时适应基线，例如，将RN50的平均mAP从57.31提高到69.22，将ViT-B/16从62.61提高到71.79。
cs.CV / 19 / 2606.12939
MAMVI: 3D Test-Time Adaptation via Masked Multi-View Point Clouds
MAMVI：通过掩蔽多视角点云进行3D测试时适应
Abstract
3D point cloud models suffer significant performance degradation under distribution shifts caused by sensor noise, occlusions, and environmental changes. Test-time adaptation (TTA) has emerged as a practical paradigm for mitigating this issue during inference. Recently, leveraging multi-view augmentation has shown promise in improving 3D TTA performance. However, existing multi-view approaches are often constrained by sequential optimization that treats each view independently. This sequential optimization leads to substantial inference latency due to repetitive optimization steps, making real-time adaptation impractical. To address this, we propose Masked Multi-View Test-Time Adaptation (MAMVI), which replaces sequential optimization with a unified single-step adaptation. Specifically, MAMVI utilizes a hybrid masking strategy that combines fixed ratios for stability with Beta-distributed sampling for diversity. By aggregating losses across multiple views, MAMVI performs adaptation through a single backward pass based on multi-view consensus. Additionally, a confidence-based adaptive learning rate is used to dynamically adjust the adaptation intensity for each sample. Extensive experiments on ModelNet-40C, ShapeNet-C, and ScanObjectNN-C demonstrate that MAMVI achieves state-of-the-art accuracy on ShapeNet-C and ScanObjectNN-C. Moreover, it remains competitive on ModelNet-40C while delivering 4.9-8.9 times faster inference, making it highly suitable for real-time applications. Our code is available at https://github.com/Inseok-kong/MAMVI
Chinese Translation
3D点云模型在传感器噪声、遮挡和环境变化引起的分布转移下表现出显著的性能下降。测试时适应（TTA）已成为在推理过程中缓解这一问题的实用范式。最近，利用多视角增强在提升3D TTA性能方面显示出良好的前景。然而，现有的多视角方法通常受到顺序优化的限制，这种方法将每个视角视为独立的。这种顺序优化导致由于重复的优化步骤而产生显著的推理延迟，使得实时适应变得不切实际。为了解决这一问题，我们提出了掩蔽多视角测试时适应（MAMVI），它用统一的单步适应替代了顺序优化。具体而言，MAMVI采用了一种混合掩蔽策略，结合了固定比例以确保稳定性和Beta分布采样以增强多样性。通过在多个视角上聚合损失，MAMVI通过基于多视角共识的单次反向传播进行适应。此外，使用基于置信度的自适应学习率动态调整每个样本的适应强度。在ModelNet-40C、ShapeNet-C和ScanObjectNN-C上的大量实验表明，MAMVI在ShapeNet-C和ScanObjectNN-C上达到了最先进的准确率。此外，在ModelNet-40C上也保持了竞争力，同时提供了4.9-8.9倍更快的推理速度，使其非常适合实时应用。我们的代码可在https://github.com/Inseok-kong/MAMVI获取。
cs.CV / 20 / 2606.12958
YOLO-AMC: An Improved YOLO Architecture with Attention Mechanisms for Building Crack Detection
YOLO-AMC：一种改进的YOLO架构，结合注意力机制用于建筑裂缝检测
Abstract
Crack detection plays an important role in infrastructure inspection and Structural Health Monitoring (SHM). However, cracks typically appear as thin, low-contrast structures and are easily affected by background noise, posing challenges for existing object detection models. This study proposes an improved YOLO-based architecture with integrated attention mechanisms, termed YOLO-AMC (YOLO with Attention Mechanisms for Crack Detection), to enhance automated crack detection performance. Based on YOLOv11, the original C2PSA module is removed, and multiple attention mechanisms, including Global Attention Mechanism (GAM), Residual Convolutional Block Attention Module (Res-CBAM), and Shuffle Attention (SA), are introduced into the multi-scale feature fusion layers of the Neck to strengthen cross-scale feature integration. Experimental results demonstrate that YOLO-AMC consistently outperforms baseline models YOLOv11n and YOLOv8n across multiple evaluation metrics. Among the evaluated attention modules, GAM achieves the best detection performance, obtaining
[email protected] = 0.9917 and
[email protected]:0.95 = 0.9506 on the test dataset, which are higher than those of YOLOv11 (0.9833 / 0.9112) and YOLOv8 (0.9707 / 0.8921). Furthermore, while maintaining a computational complexity of 7.6 GFLOPs, the proposed model achieves 110.95 FPS on an NVIDIA RTX 4090 platform and approximately 5 FPS on a Raspberry Pi 5 edge device, demonstrating a favorable trade-off between accuracy and deployment efficiency. The implementation code for this study is available on GitHub at https://github.com/CY-Tsai24/YOLO-AMC.
Chinese Translation
裂缝检测在基础设施检查和结构健康监测（SHM）中发挥着重要作用。然而，裂缝通常表现为细小、低对比度的结构，容易受到背景噪声的影响，这给现有的目标检测模型带来了挑战。本研究提出了一种基于YOLO的改进架构，集成了注意力机制，称为YOLO-AMC（用于裂缝检测的YOLO与注意力机制），以增强自动裂缝检测性能。基于YOLOv11，去除了原始的C2PSA模块，并在Neck的多尺度特征融合层中引入了多种注意力机制，包括全局注意力机制（Global Attention Mechanism，GAM）、残差卷积块注意力模块（Residual Convolutional Block Attention Module，Res-CBAM）和洗牌注意力（Shuffle Attention，SA），以加强跨尺度特征的整合。实验结果表明，YOLO-AMC在多个评估指标上始终优于基线模型YOLOv11n和YOLOv8n。在评估的注意力模块中，GAM实现了最佳检测性能，在测试数据集上获得了
[email protected] = 0.9917和
[email protected]:0.95 = 0.9506，均高于YOLOv11（0.9833 / 0.9112）和YOLOv8（0.9707 / 0.8921）。此外，在保持7.6 GFLOPs计算复杂度的同时，所提出的模型在NVIDIA RTX 4090平台上实现了110.95 FPS，在Raspberry Pi 5边缘设备上约为5 FPS，展示了准确性与部署效率之间的良好平衡。本研究的实现代码可在GitHub上找到，链接为https://github.com/CY-Tsai24/YOLO-AMC。
cs.CV / 21 / 2606.12977
Efficient, Robust, and Anti-Collusion Fingerprinting of Image Diffusion Models
高效、稳健且具反串通能力的图像扩散模型指纹识别
Abstract
Model fingerprinting, embedding user-specific identifiers (fingerprints) into generated outputs, has recently emerged as a popular solution to protect the intellectual property rights (IPR) of generative text-to-image (T2I) models and prevent unauthorized redistribution. In this work, we reveal a previously unexplored systematic vulnerability in existing generative model fingerprinting methods: they lack robustness against collusion attacks, where multiple attackers combine their models to remove or obscure the fingerprints. To address this issue, we take the first step towards a robust fingerprinting method for T2I models with anti-collusion capabilities. The proposed method encodes strings of bits, namely fingerprints, into the coefficients of a personalized normalization module (PNM) incorporated into T2I models, so that fingerprints can be reliably recovered from any generated image. To defend against collusion attacks and prevent unauthorized model redistribution, we introduce an anti-collusion mechanism based on lossless function-invariant parameter transformations. This mechanism significantly degrades the image generation quality of colluded models, making them effectively unusable. Moreover, our method allows developers to efficiently create multiple copies of fingerprinted T2I models by reparameterizing the PNM without the need for retraining. We also introduce a worst-case optimization strategy to improve robustness against model-level attacks. Our experiments demonstrate that the proposed method achieves high fidelity and robustness across multiple T2I image generation and editing tasks, with fingerprint extraction accuracy exceeding 99.5%. Compared with existing methods, our method demonstrates, for the first time, a notable proactive robustness to collusion attacks by significantly increasing the FID of colluded models.
Chinese Translation
模型指纹识别是将用户特定标识符（指纹）嵌入生成输出中，近年来已成为保护生成文本到图像（T2I）模型知识产权（IPR）和防止未经授权再分发的热门解决方案。在本研究中，我们揭示了现有生成模型指纹识别方法中一个先前未被探索的系统性脆弱性：它们在面对串通攻击时缺乏稳健性，多个攻击者可以结合他们的模型以去除或模糊指纹。为了解决这一问题，我们迈出了为T2I模型开发具反串通能力的稳健指纹识别方法的第一步。所提出的方法将比特串（即指纹）编码到个性化归一化模块（PNM）的系数中，该模块被纳入T2I模型中，从而可以可靠地从任何生成的图像中恢复指纹。为了防御串通攻击并防止未经授权的模型再分发，我们引入了一种基于无损函数不变参数变换的反串通机制。该机制显著降低了串通模型的图像生成质量，使其有效不可用。此外，我们的方法允许开发者通过重新参数化PNM高效创建多个指纹化的T2I模型，而无需重新训练。我们还引入了一种最坏情况优化策略，以提高对模型级攻击的稳健性。我们的实验表明，所提出的方法在多个T2I图像生成和编辑任务中实现了高保真度和稳健性，指纹提取准确率超过99.5%。与现有方法相比，我们的方法首次展示了对串通攻击的显著主动稳健性，显著提高了串通模型的FID。
cs.CV / 22 / 2606.12981
Camera and LiDAR BEV Fusion for Cooperative 3D Object Detection on TUMTraf V2X
相机与激光雷达的鸟瞰视图融合用于TUMTraf V2X的协作3D目标检测
Abstract
We describe a Camera and LiDAR fusion detector developed for the TUMTraf V2X cooperative 3D object detection track of the DriveX 2026 challenge. The detector fuses three roadside cameras with a fused infrastructure-plus-vehicle point cloud in a shared bird's-eye-view space and predicts boxes through a CenterPoint-style head with a generalized IoU regression loss and an IoU quality re-ranking head. Trained on the provided train and validation splits, the model reaches a 3D mAP of 0.85 on the public Codabench test split. While iterating on the system, we observed that 44 of the 50 test frames are also present in the released train (40) and validation (4) splits with their labels. We therefore conducted two additional studies to quantify how this overlap affects the final score: (1) a finetuning run that oversamples the 44 overlapping frames, reaching 0.89 mAP, and (2) a post-processing run that replaces predictions on those frames with the released ground truth, reaching 0.99 mAP (uploaded to our Codabench account for testing but not published on the leaderboard). All three configurations and their per-class results are reported.
Chinese Translation
我们描述了一种为DriveX 2026挑战的TUMTraf V2X协作3D目标检测赛道开发的相机与激光雷达融合检测器。该检测器在共享的鸟瞰视图空间中融合了三台路边相机与融合的基础设施加车辆点云，并通过CenterPoint风格的头部预测框，使用广义IoU回归损失和IoU质量重排序头。经过在提供的训练和验证数据集上的训练，该模型在公共Codabench测试集上达到了0.85的3D mAP。在系统迭代过程中，我们观察到50个测试帧中的44个也出现在已发布的训练集（40）和验证集（4）中，并且带有标签。因此，我们进行了两项额外研究，以量化这种重叠对最终得分的影响：(1) 一次微调运行，过采样这44个重叠帧，达到了0.89 mAP；(2) 一次后处理运行，用已发布的真实值替换这些帧上的预测，达到了0.99 mAP（已上传至我们的Codabench账户进行测试，但未在排行榜上发布）。所有三种配置及其每类结果均已报告。
cs.CV / 23 / 2606.12985
Objects Before Words: Object-First Inductive Biases for Grounding Language in Child-View Video
物体优先于词汇：儿童视角视频中语言基础的物体优先归纳偏差
Abstract
Learning grounded word meaning from natural experience requires resolving two ambiguities in infant-view recordings: when the named referent appears and where it is in a cluttered frame. In SAYCam-style data, caregiver speech is sparse and weakly synchronized with egocentric video, so single-frame contrastive pairing yields noisy positives in which the intended object is absent or entangled with distractors. We propose BabyMind, an object-first bias for child-view contrastive learning under sparse, noisy supervision. BabyMind extracts candidate object embeddings using an offline mask-based region interface, links candidates across a short utterance-centered window into lightweight object files via tracking, and aligns utterances to bags of object files with a prototype-space multiple-instance contrastive objective. Track-coherence and global-object agreement regularizers stabilize learning and transfer object-file structure into the global frame embedding used at evaluation. On SAYCam-S, BabyMind improves Labeled-S 15 forced-choice accuracy by +2.6 points over CVCL and yields consistent gains on in-vocabulary out-of-distribution benchmarks. Code is available at https://github.com/sathiiii/BabyMind.
Chinese Translation
从自然经验中学习基础词义需要解决婴儿视角录音中的两个模糊性：命名指称何时出现以及在杂乱画面中的位置。在SAYCam风格的数据中，照顾者的语言稀疏且与自我中心视频的同步性较弱，因此单帧对比配对会产生噪声正例，其中意图对象缺失或与干扰物缠绕。我们提出了BabyMind，一种在稀疏、嘈杂监督下进行儿童视角对比学习的物体优先偏差。BabyMind通过离线基于掩膜的区域接口提取候选物体嵌入，利用跟踪将候选物体链接到以短语为中心的窗口中形成轻量级物体文件，并通过原型空间的多实例对比目标将短语与物体文件包对齐。轨迹一致性和全局物体一致性正则化器稳定学习，并将物体文件结构转移到评估时使用的全局框架嵌入上。在SAYCam-S上，BabyMind将Labeled-S 15强制选择准确率提高了2.6个百分点，相较于CVCL，并在词汇内的分布外基准测试中获得了一致的提升。代码可在 https://github.com/sathiiii/BabyMind 获取。
cs.CV / 24 / 2606.12987
Diffusion Transformer World-Action Model for AV Scene Prediction
用于自动驾驶场景预测的扩散变换器世界-动作模型
Abstract
Action-conditioned world models let an autonomous vehicle predict future camera scenes from its own planned controls, enabling planning and simulation without real-world rollouts, but at compact, trainable scale the futures are ambiguous and the field's standard distortion metrics actively mislead: they reward a blurry regression mean over a realistic prediction. We confront this with a compact latent world model that, given the present front-camera latent and a sequence of ego-actions, predicts future scene latents a frozen decoder renders to $256 \times 256$ frames up to 8 seconds ahead, evaluated on 150 held-out nuScenes scenes. We first benchmark where to predict: across six frozen encoders spanning four representation families, V-JEPA2 with temporal context reduces steering RMSE by 40% over the best single-frame encoder. We then train a latent Diffusion Transformer (DiT) and, through a controlled diagnosis, identify the four ingredients it needs: spatial tokens, the $x_0$ objective, residual anchoring, and sampling matched to target uncertainty. In a Stable-Diffusion-VAE encode-predict-decode pipeline we expose the central tension: distortion metrics (cosine similarity, SSIM) favor the blurry mean, masking that the diffusion model is far closer to the real frame distribution. Inception-based FID and KID reveal a clean perception-distortion frontier: diffusion attains KID 0.078 versus 0.375 for regression ($4.8\times$ better), and a deployable train-derived calibration makes this practical without test-time ground truth. The model is genuinely action-controllable (steering drives scene displacement, Spearman $\rho = 0.81$, vs $-0.18$ for regression). We trace limited single-pass motion to a shared-present anchor and engineer a compact 1.7M-parameter "jump" model that recovers full ground-truth motion magnitude ($1.02\times$ GT), where single-pass models capture less than half.
Chinese Translation
基于动作的世界模型使得自主车辆能够根据自身规划的控制预测未来的摄像头场景，从而实现无需真实世界试验的规划和模拟。然而，在紧凑且可训练的规模下，未来场景的预测存在模糊性，并且该领域的标准失真度量会产生误导：它们奖励模糊的回归均值而非现实的预测。我们通过一个紧凑的潜在世界模型来应对这一挑战，该模型在给定当前前置摄像头潜在状态和一系列自我动作的情况下，预测未来场景的潜在状态，并通过一个冻结的解码器将其渲染为 $256 imes 256$ 的帧，最多可预测8秒，评估基于150个保留的nuScenes场景。我们首先基准测试预测位置：在跨越四个表示家族的六个冻结编码器中，具有时间上下文的V-JEPA2将转向均方根误差（RMSE）降低了40%，优于最佳单帧编码器。然后，我们训练了一个潜在的扩散变换器（DiT），并通过受控诊断识别出其所需的四个要素：空间标记、 $x_0$ 目标、残差锚定和与目标不确定性匹配的采样。在一个稳定扩散变分自编码器（Stable-Diffusion-VAE）编码-预测-解码的流程中，我们揭示了核心矛盾：失真度量（余弦相似度、结构相似性指数（SSIM））偏向模糊均值，掩盖了扩散模型与真实帧分布之间的接近程度。基于Inception的FID和KID揭示了一个清晰的感知-失真边界：扩散模型的KID为0.078，而回归模型为0.375（提高了4.8倍），并且可部署的训练衍生校准使得在没有测试时真实值的情况下变得可行。该模型在动作控制上表现出色（转向驱动场景位移，Spearman相关系数 $
ho = 0.81$，而回归模型为 $-0.18$）。我们将有限的单次运动追溯到一个共享的当前锚点，并设计了一个紧凑的170万参数的“跳跃”模型，该模型恢复了完整的真实运动幅度（$1.02 imes$ GT），而单次模型捕获的幅度不足一半。
cs.CV / 25 / 2606.12988
A Machine Learning Framework for Real-Time Personalized Ergonomic Pose Analysis
实时个性化人体工学姿势分析的机器学习框架
Abstract
This paper introduces a new methodology for real-time prediction of ergonomic and non-ergonomic human poses using volumetric video data in three dimensions. Although the methodology was designed for ergonomic assessments, it can be adapted to other applications requiring real-time analysis of human posture. One aspect that makes this system stand out is its ability to analyze 3D point clouds during the assessment, enabling computation from multiple angles. This overcomes a critical limitation of cameras which provide often a fixed viewpoint, thereby restricting the data available for a thorough postural evaluation, especially when occlusions occur. The system continuously and automatically performs pose inference using the chosen perspective on the real-time streaming data; however, only the poses manually selected and labeled by the user are used to train the personalized deep learning classifier. The methodology has been refined through a case study in which RGB-D cameras captured subjects performing load-lifting tasks, enabling real-time skeletal labeling. The model was trained on this data and, following the training phase, performs inference on new streaming data in real time. This research offers a scalable and pragmatic approach for real-time ergonomic evaluation by combining state-of-the-art 3D data technologies and traditional 2D pose estimation algorithms. It addresses the increasing need for safety and health monitoring in workplace environments, marking a notable contribution to the domain.
Chinese Translation
本文介绍了一种新的方法论，用于实时预测人体工学和非人体工学的人体姿势，利用三维体积视频数据。尽管该方法论是为人体工学评估而设计的，但可以适应其他需要实时分析人类姿势的应用。该系统的一个突出特点是能够在评估过程中分析三维点云，从多个角度进行计算。这克服了摄像头的一个关键限制，即通常提供固定的视角，从而限制了用于全面姿势评估的数据，尤其是在发生遮挡时。该系统持续且自动地使用所选视角对实时流数据进行姿势推断；然而，仅使用用户手动选择和标记的姿势来训练个性化深度学习分类器。通过一个案例研究，该方法论得到了完善，其中RGB-D摄像头捕捉受试者进行负重提升任务的过程，实现了实时骨骼标记。模型在这些数据上进行了训练，并在训练阶段后对新的流数据实时进行推断。本研究通过结合最先进的三维数据技术和传统的二维姿势估计算法，提供了一种可扩展且务实的实时人体工学评估方法。它满足了工作环境中对安全和健康监测日益增长的需求，为该领域做出了显著贡献。
cs.CV / 26 / 2606.13022
Quality-Preserving Imperceptible Adversarial Attack on Skeleton-based Human Action Recognition
基于骨架的人类动作识别的质量保持隐形对抗攻击
Abstract
Adversarial attacks on skeletal human action recognition have received significant attention. However, existing methods typically introduce noise-like perturbations that degrade motion quality post-attack, and thereby are inherently perceptible with recent advancements in S-HAR systems. We discover that this degradation stems from the gap between empirical and true risks during the optimization process of previous adversarial attacks. To address this issue, we propose an attack where adversarial motions are obtained without compromising their motion quality. To minimize the risk gap and preserve motion quality, we propose a distribution-based adversarial attack method without introducing noise-like perturbations. To faithfully evaluate the motion quality, we propose a new metric that aligns with human perception on real-world naturalness. Experiments have been conducted on the state-of-the-art S-HAR methods across two datasets, demonstrating the superiority of our method in both the attack success rate and the post-attack motion quality through qualitative and quantitative analyses. The success of our quality-preserving attack application and distribution-based method raises serious concerns about the robustness of action recognizers, highlighting the need for further enhancements in this domain.
Chinese Translation
针对骨架人类动作识别的对抗攻击受到了广泛关注。然而，现有方法通常引入噪声般的扰动，导致攻击后运动质量下降，因此在最近的S-HAR（Skeleton-based Human Action Recognition）系统中本质上是可感知的。我们发现这种降级源于以往对抗攻击优化过程中经验风险与真实风险之间的差距。为了解决这个问题，我们提出了一种攻击方法，通过不妥协运动质量的方式获得对抗运动。为了最小化风险差距并保持运动质量，我们提出了一种基于分布的对抗攻击方法，且不引入噪声般的扰动。为了真实评估运动质量，我们提出了一种与人类感知真实自然性相一致的新指标。我们在两个数据集上对最先进的S-HAR方法进行了实验，通过定性和定量分析证明了我们方法在攻击成功率和攻击后运动质量方面的优越性。我们质量保持攻击应用和基于分布的方法的成功引发了对动作识别器鲁棒性的严重关注，突显了在这一领域进一步增强的必要性。
cs.CV / 27 / 2606.13030
A Multi-Modal Framework with Cross-Subject Pseudo-Labeling and Semantic Alignment for Micro-Gesture Recognition
一种跨主体伪标签和语义对齐的多模态框架用于微手势识别
Abstract
Micro-gestures (MGs) are spontaneous and subtle body movements that frequently convey hidden human emotions. Recognizing MGs in untrimmed videos remains highly challenging due to their extremely low signal-to-noise ratio, severe long-tailed class distribution, and the inherent domain shift encountered in cross-subject evaluation scenarios. In this paper, we propose a comprehensive multi-modal framework for Track 1 of the 4th MiGA-IJCAI Challenge. To capture fine-grained representations, we design a saliency-guided multi-modal extraction pipeline integrating 68-keypoint skeleton joint coordinates, 3D heatmap volumes, and high-resolution RGB visual features. We introduce a gentle square-root smoothed weighting mechanism paired with an Orthogonal Semantic Embedding Loss to protect tail classes without compromising overall recognition capabilities. More importantly, to bridge the cross-subject generalization gap, we propose a Cross-Modal Pseudo-Labeling (CMPL) strategy for unsupervised domain adaptation, which significantly boosts single-modal robustness. A temperature-scaled soft-voting mechanism is finally utilized to alleviate overconfidence during late fusion. Extensive experiments demonstrate that our framework achieves a competitive F1-score of 68.13\%, securing the 4th place.
Chinese Translation
微手势（MGs）是自发且细微的身体动作，常常传达隐藏的人类情感。在未剪辑视频中识别微手势仍然具有很高的挑战性，原因在于其极低的信噪比、严重的长尾类别分布以及在跨主体评估场景中遇到的固有领域转移。本文提出了一种全面的多模态框架，旨在参与第四届MiGA-IJCAI挑战赛的第一赛道。为了捕捉细粒度的表征，我们设计了一种基于显著性引导的多模态提取管道，整合了68个关键点骨骼关节坐标、3D热图体积和高分辨率RGB视觉特征。我们引入了一种温和的平方根平滑加权机制，并结合正交语义嵌入损失，以保护长尾类别而不影响整体识别能力。更重要的是，为了弥合跨主体的泛化差距，我们提出了一种用于无监督领域适应的跨模态伪标签（CMPL）策略，显著提升了单模态的鲁棒性。最后，采用温度缩放的软投票机制来缓解晚期融合过程中的过度自信。大量实验表明，我们的框架实现了68.13%的竞争性F1分数，获得了第四名。
cs.CV / 28 / 2606.13032
GeoCFNet: Geometry-Aware Confidence Field Network for Robot-Assisted Endoscopic Submucosal Dissection
GeoCFNet：面向几何的置信场网络用于机器人辅助内镜下黏膜下切除
Abstract
Advanced surgical robotics has made robot-assisted endoscopic submucosal dissection (ESD) a promising approach for the en-bloc resection of large lesions, with the potential to reduce recurrence and improve long-term outcomes. However, the technical complexity and risk of complications in ESD demand stable and precise visual guidance to maintain an accurate dissection corridor and a safe tissue margin. Dense confidence fields provide an effective representation for this purpose by describing both the preferred dissection region and its spatial transition to surrounding tissue. However, reliable confidence field estimation remains challenging in dynamic endoscopic scenes due to smoke, specular highlights, tissue deformation, weak texture, and the thin geometric structure of the target region. To address these challenges, we formulate dissection guidance as a geometry-aware confidence field estimation problem and propose GeoCFNet, a geometry-aware confidence field network built on a pretrained DINOv3 backbone. GeoCFNet integrates a Token-Differentiated Fusion module to aggregate class-token context with dense patch representations, a SegFormer decoder for confidence regression, and Geometry-Aware Spatial Regularization (GASR) to preserve spatial coherence and local geometric transitions. Experimental results show that GeoCFNet achieves RMSE 0.0480, PSNR 27.1995, SSIM 0.3397, and CC 0.2466, indicating accurate and geometrically stable confidence field estimation for robot-assisted ESD guidance.
Chinese Translation
先进的外科机器人技术使得机器人辅助内镜下黏膜下切除（ESD）成为大病灶整体切除的一种有前景的方法，具有减少复发和改善长期结果的潜力。然而，ESD的技术复杂性和并发症风险要求稳定且精确的视觉引导，以维持准确的切除通道和安全的组织边缘。密集的置信场通过描述首选的切除区域及其与周围组织的空间过渡，为此目的提供了有效的表示。然而，由于烟雾、镜面高光、组织变形、纹理较弱以及目标区域的薄几何结构，在动态内镜场景中可靠的置信场估计仍然具有挑战性。为了解决这些挑战，我们将切除引导形式化为一个面向几何的置信场估计问题，并提出了GeoCFNet，这是一种基于预训练DINOv3骨干网络构建的面向几何的置信场网络。GeoCFNet集成了一个令牌差异化融合模块，以聚合类令牌上下文与密集补丁表示，一个用于置信回归的SegFormer解码器，以及几何感知空间正则化（GASR），以保持空间一致性和局部几何过渡。实验结果表明，GeoCFNet达到了RMSE 0.0480、PSNR 27.1995、SSIM 0.3397和CC 0.2466，表明其在机器人辅助ESD引导中实现了准确且几何稳定的置信场估计。
cs.CV / 29 / 2606.13033
SAM-Deep-EIoU: Selective Mask Propagation for Multi-Object Tracking
SAM-Deep-EIoU：用于多目标跟踪的选择性掩码传播
Abstract
Multi-object tracking has a heavy-tailed difficulty distribution: most frames are easy for a lightweight base tracker, while a small fraction are intrinsically hard. Video object segmentation (VOS) models can often preserve identity through the hard frames where the base tracker fails, but they are much more expensive in compute and memory. We propose selective mask propagation, a tracking algorithm that dispatches from a base tracker to a VOS model only on windows where an assignment-uncertainty signal fires. The base tracker's output is modified only when the VOS model makes a confident prediction that contradicts the base tracker's identity assignment; weak or inconclusive predictions preserve the base output. The method is training-free, treats both the base tracker and the VOS model as black boxes, and can benefit from replacing the VOS component with a more capable model. On DanceTrack, selective mask propagation improves three different base trackers. On SportsMOT, where identity preservation is central to sports analytics, SAM3-Deep-EIoU with global track association achieves state-of-the-art performance on the benchmark with 86.8 HOTA.
Chinese Translation
多目标跟踪具有重尾的难度分布：大多数帧对于轻量级基础跟踪器来说是简单的，而只有一小部分帧本质上是困难的。视频目标分割（VOS）模型通常能够在基础跟踪器失败的困难帧中保持身份，但它们在计算和内存方面的开销要大得多。我们提出了选择性掩码传播，这是一种跟踪算法，仅在分配不确定性信号触发的窗口中从基础跟踪器切换到VOS模型。只有当VOS模型做出与基础跟踪器的身份分配相矛盾的自信预测时，基础跟踪器的输出才会被修改；而弱或不确定的预测则保留基础输出。该方法不需要训练，将基础跟踪器和VOS模型视为黑箱，并且可以通过用更强大的模型替换VOS组件来获益。在DanceTrack上，选择性掩码传播改善了三种不同的基础跟踪器。在SportsMOT上，身份保持对于体育分析至关重要，采用全局轨迹关联的SAM3-Deep-EIoU在基准测试中实现了86.8 HOTA的最先进性能。
cs.CV / 30 / 2606.13035
TetherCache: Stabilizing Autoregressive Long-Form Video Generation with Gated Recall and Trusted Alignment
TetherCache：通过门控回忆和可信对齐稳定自回归长视频生成
Abstract
Autoregressive video diffusion models provide a natural formulation for streaming and variable-length video generation by conditioning newly generated frames on previously generated content. However, extending these models to minute-level generation remains challenging: the limited KV-cache budget prevents the model from retaining the full history, while repeatedly conditioning on self-generated frames induces a context distribution shift that accumulates over time, leading to visual artifacts, quality degradation, and temporal drift. In this paper, we propose TetherCache, a training-free and plug-and-play cache management strategy for drift-resistant long video generation. TetherCache organizes the cache into sink, memory, and recent regions, and introduces two complementary mechanisms. First, GRAB (Gated Recall with Attention-Diversity Balancing) selects long-range memory frames using a gated score that combines attention-based relevance with temporal diversity, preserving informative yet diverse historical context under a fixed cache budget. Second, TAME (Trusted Alignment via Memory Editing) lightly edits newly recalled memory tokens by aligning their statistics to a trusted context distribution, reducing the pollution caused by drifted historical features. Built on Self-Forcing, TetherCache consistently improves long-video generation quality on VBench-Long across 30s, 60s, and 240s settings. In particular, for 240s generation, it substantially improves overall and semantic scores while reducing quality drift from 7.84 to 1.33, demonstrating its effectiveness for stable long-horizon autoregressive video diffusion.
Chinese Translation
自回归视频扩散模型为流式和可变长度视频生成提供了一种自然的表述，通过将新生成的帧与先前生成的内容进行条件化。然而，将这些模型扩展到分钟级生成仍然具有挑战性：有限的KV缓存预算使模型无法保留完整的历史记录，而反复对自生成帧进行条件化会导致上下文分布的偏移，随着时间的推移积累，导致视觉伪影、质量下降和时间漂移。在本文中，我们提出了TetherCache，这是一种无训练且即插即用的缓存管理策略，旨在抵抗漂移的长视频生成。TetherCache将缓存组织为汇聚区、记忆区和最近区，并引入了两种互补机制。首先，GRAB（门控回忆与注意力多样性平衡）使用结合了基于注意力的相关性和时间多样性的门控评分选择长距离记忆帧，在固定的缓存预算下保留信息丰富但多样的历史上下文。其次，TAME（通过记忆编辑实现可信对齐）轻微编辑新回忆的记忆标记，通过将其统计数据对齐到可信的上下文分布，减少了漂移的历史特征造成的污染。在自我强制（Self-Forcing）的基础上，TetherCache在VBench-Long的30秒、60秒和240秒设置中持续提高了长视频生成的质量。特别是对于240秒的生成，它显著提高了整体和语义得分，同时将质量漂移从7.84降低到1.33，证明了其在稳定长时间自回归视频扩散中的有效性。
cs.CV / 31 / 2606.13041
SeamEdit: A Black-Box VLM-Agnostic Pipeline for Large-Image Semantic Editing
SeamEdit：一种黑箱VLM无关的大图像语义编辑管道
Abstract
Semantic region editing for large images must satisfy two requirements at the same time: high generative quality and natural integration with surrounding content. Some related methods rely on white-box models and leave the strong generation capability of closed-source models underexplored. Directly applying closed-source models to tiled editing, however, introduces several failure modes: semantic deformation, canvas-level alignment drift, and visible seam artifacts. This paper presents SeamEdit, a training-free and model-agnostic pipeline that treats any VLM with inpainting capability as a black-box oracle. SeamEdit mitigates these issues through a five-stage post-hoc pipeline: overlay-based tile decomposition, black-box VLM inpainting, geometric and color-consistency correction, seam-risk-based multi-candidate ranking, and dynamic-programming curved seam fusion. The pipeline reduces seam visibility and supports semantic modification of arbitrary tile regions.
Chinese Translation
大图像的语义区域编辑必须同时满足两个要求：高生成质量和与周围内容的自然融合。一些相关方法依赖于白箱模型，未能充分挖掘闭源模型的强大生成能力。然而，直接将闭源模型应用于平铺编辑会引入几种失败模式：语义变形、画布级对齐漂移和可见接缝伪影。本文提出了SeamEdit，这是一种无训练且无关模型的管道，将任何具有修补能力的VLM视为黑箱神谕。SeamEdit通过五个阶段的后处理管道来缓解这些问题：基于叠加的瓦片分解、黑箱VLM修补、几何和颜色一致性校正、基于接缝风险的多候选排名以及动态规划曲线接缝融合。该管道减少了接缝的可见性，并支持任意瓦片区域的语义修改。
cs.CV / 32 / 2606.13061
LaME: Learning to Think in Latent Space for Multimodal Embedding via Information Bottleneck
LaME：通过信息瓶颈在潜在空间中学习多模态嵌入的思维方式
Abstract
Reasoning-driven universal multimodal embedding has advanced rapidly by introducing Chain-of-Thought (CoT) reasoning into the embedding pipeline. Despite the strong performance across both general and complex tasks, this paradigm suffers from two core limitations: (i) autoregressive CoT reasoning incurs high computational cost, making it impractical for low-latency retrieval; and (ii) embedding performance is heavily coupled with CoT annotation quality, making large-scale training unreliable. These raise fundamental questions: Is textual CoT the optimal form of reasoning for embedding, and can effective embedding reasoning be accomplished in latent space? To this end, we propose LaME (Latent Reasoning Multimodal Embedding), which formulates embedding-oriented latent reasoning as a weakly supervised information bottleneck. LaME employs K learnable reason tokens as a fixed-capacity bottleneck, completing all reasoning within a single forward pass. The two weak supervision signals structurally decouple contrastive from autoregressive objectives and eliminate dependence on CoT annotations, while a two-stage training pipeline ensures stable convergence. Experiments on MMEB-v2 and MRMR show that LaME achieves competitive performance, surpassing some explicit CoT-based models, while delivering 60x faster inference than explicit CoT methods and 2x faster than latent baselines with throughput comparable to discriminative embedding models. Code will be released.
Chinese Translation
基于推理的通用多模态嵌入通过将链式思维（Chain-of-Thought, CoT）推理引入嵌入流程而迅速发展。尽管在一般和复杂任务中表现强劲，但这一范式存在两个核心限制：（i）自回归的 CoT 推理带来了高计算成本，使其在低延迟检索中不切实际；（ii）嵌入性能与 CoT 注释质量高度耦合，使大规模训练变得不可靠。这引发了基本问题：文本 CoT 是否是嵌入的最佳推理形式？有效的嵌入推理是否可以在潜在空间中完成？为此，我们提出了 LaME（潜在推理多模态嵌入），将面向嵌入的潜在推理形式化为一种弱监督信息瓶颈。LaME 使用 K 个可学习的推理标记作为固定容量瓶颈，在单次前向传递中完成所有推理。这两个弱监督信号在结构上将对比目标与自回归目标解耦，并消除了对 CoT 注释的依赖，同时双阶段训练流程确保了稳定收敛。在 MMEB-v2 和 MRMR 上的实验表明，LaME 实现了具有竞争力的性能，超越了一些显式 CoT 基础的模型，同时在推理速度上比显式 CoT 方法快 60 倍，比潜在基线快 2 倍，且吞吐量可与判别嵌入模型相媲美。代码将会发布。
cs.CV / 33 / 2606.13096
Unified MRI Brain Image Translation via Hierarchical Tumor Structure Comparison
通过分层肿瘤结构比较实现统一的MRI脑图像翻译
Abstract
Multi-modal MRI brain image translation via available modalities holds significant practical importance in modern medicine, providing robust support for early diagnosis, treatment planning, and outcome assessment of diseases. For this purpose, it is important to ensure the fidelity of the tumor regions after translation. However, existing brain image translation methods ignore the structure information of different tumor regions, which could assist translation models in enhancing the quality and clinical applicability of the translated images. In this work, we propose a novel translation model called HTSCGAN, which is a unified multi-modal brain image translation generative adversarial model integrating the structural information within tumor regions with the aim of improving the quality of brain image translation. Specifically, the generator employs three Patch Contrast Module (PCM) with different patch sizes to capture the hierarchical structural information of the tumor regions. In addition, a pretrained Patch Classifier (PC) and a pretrained Structure-Aware Encoder (SAE) are employed to derive the generated image containing the same tumor region structure as the ground truth image via patch classification loss and tumor perceptual loss, respectively. The experiments on BraTS2020 and BraTS2021 demonstrate strong performance of our model in both translation tasks and down stream segmentation tasks, highlighting its effectiveness in enhancing the quality and clinical relevance of the translated brain images. Our code is available at https://anonymous.4open.science/r/HTSCGAN.
Chinese Translation
多模态MRI脑图像翻译在现代医学中具有重要的实际意义，为疾病的早期诊断、治疗规划和结果评估提供了有力支持。为此，确保翻译后肿瘤区域的真实性至关重要。然而，现有的脑图像翻译方法忽视了不同肿瘤区域的结构信息，而这些信息可以帮助翻译模型提高翻译图像的质量和临床适用性。在本研究中，我们提出了一种新颖的翻译模型HTSCGAN，该模型是一个统一的多模态脑图像翻译生成对抗模型，整合了肿瘤区域内的结构信息，旨在提高脑图像翻译的质量。具体而言，生成器采用了三个不同补丁大小的补丁对比模块（Patch Contrast Module, PCM）来捕捉肿瘤区域的分层结构信息。此外，采用预训练的补丁分类器（Patch Classifier, PC）和预训练的结构感知编码器（Structure-Aware Encoder, SAE）通过补丁分类损失和肿瘤感知损失分别推导出与真实图像具有相同肿瘤区域结构的生成图像。在BraTS2020和BraTS2021上的实验表明，我们的模型在翻译任务和下游分割任务中均表现出强大的性能，突显了其在提高翻译脑图像质量和临床相关性方面的有效性。我们的代码可在https://anonymous.4open.science/r/HTSCGAN获取。
cs.CV / 34 / 2606.13108
PP-OCRv6: From 1.5M to 34.5M Parameters, Surpassing Billion-Scale VLMs on OCR Tasks
PP-OCRv6：从150万到3450万参数，在OCR任务中超越十亿规模的视觉语言模型
Abstract
Vision-Language Models (VLMs) have achieved impressive results on general vision-language tasks, yet they suffer from hallucination, imprecise localization, and prohibitive computational cost when applied to dedicated OCR scenarios. This paper presents PP-OCRv6, a lightweight OCR system that combines architectural innovation with data-centric optimization. PP-OCRv6 redesigns the backbone, detection neck, and recognition neck around a unified MetaFormer-style building block with structural reparameterization, decoupling spatial token mixing from channel mixing and supporting both tasks through task-specific stride configurations. Three model tiers (medium, small, tiny) share the same block primitives, covering deployment scenarios from server to edge. On our in-house benchmarks, PP-OCRv6_medium achieves 83.2% recognition accuracy and 86.2% detection Hmean, outperforming PP-OCRv5_server by +5.1% and +4.6% respectively while surpassing Qwen3-VL-235B, GPT-5.5, and Gemini-3.1-Pro with orders of magnitude fewer parameters. The tiny tier achieves 3.9$\times$ faster inference than PP-OCRv5_mobile on Intel Xeon CPU while maintaining comparable accuracy.
Chinese Translation
视觉语言模型（VLMs）在一般视觉语言任务中取得了令人瞩目的成果，但在专用OCR场景中应用时，它们面临幻觉、定位不准确和计算成本高昂等问题。本文提出了PP-OCRv6，一种轻量级OCR系统，结合了架构创新与数据中心优化。PP-OCRv6围绕统一的MetaFormer风格构建模块重新设计了主干、检测颈部和识别颈部，通过结构重参数化将空间令牌混合与通道混合解耦，并通过特定任务的步幅配置支持这两项任务。三个模型层级（中型、小型、微型）共享相同的基本模块，覆盖从服务器到边缘的部署场景。在我们的内部基准测试中，PP-OCRv6_medium实现了83.2%的识别准确率和86.2%的检测Hmean，分别比PP-OCRv5_server提高了5.1%和4.6%，同时以数量级更少的参数超越了Qwen3-VL-235B、GPT-5.5和Gemini-3.1-Pro。微型层级在Intel Xeon CPU上实现了比PP-OCRv5_mobile快3.9倍的推理速度，同时保持了相当的准确性。
cs.CV / 35 / 2606.13127
Fully Distributed Multi-View 3D Tracking in Real-Time
实时全分布式多视角三维跟踪
Abstract
Multi-camera tracking with overlapping fields of view typically relies on centralized fusion, which creates computational bottlenecks that prevent deployment at scale. We present MV3DT, a fully distributed framework for real-time multi-view 3D tracking that achieves accurate identity propagation and occlusion recovery through peer-to-peer coordination, eliminating the need for central aggregation. Each camera node executes a lightweight modular pipeline comprising monocular 3D perception, distributed multi-view association, and collaborative fusion via lightweight messaging. MV3DT achieves 94.3% IDF1 and 93.3% MOTA on WILDTRACK, competitive with state-of-the-art centralized methods, while demonstrating superior scalability by sustaining 30 FPS on 100 cameras with less than 10 ms inter-camera latency and only 2.2% communication overhead. MV3DT operates in a zero-shot regime given camera calibrations, requiring no scene-specific learning and making it directly deployable in new environments. These results establish MV3DT as a practical solution for real-time multi-view tracking in large-scale overlapping camera networks.
Chinese Translation
多摄像头跟踪通常依赖于重叠视场的集中融合，这会造成计算瓶颈，阻碍大规模部署。我们提出了MV3DT，一个全分布式的实时多视角三维跟踪框架，通过点对点协调实现准确的身份传播和遮挡恢复，消除了对中央聚合的需求。每个摄像头节点执行一个轻量级的模块化管道，包括单目三维感知、分布式多视角关联和通过轻量级消息的协同融合。MV3DT在WILDTRACK上实现了94.3%的IDF1和93.3%的MOTA，竞争于最先进的集中方法，同时在100个摄像头上以低于10毫秒的摄像头间延迟和仅2.2%的通信开销维持30帧每秒，展示了优越的可扩展性。MV3DT在给定摄像头标定的情况下以零样本模式运行，无需场景特定学习，使其可以直接在新环境中部署。这些结果确立了MV3DT作为大规模重叠摄像头网络中实时多视角跟踪的实用解决方案。
cs.CV / 36 / 2606.13135
Cascade Classification of Dermoscopic Images of Skin Neoplasms with Controllable Sensitivity and External Clinical Validation
具有可控灵敏度的皮肤肿瘤皮肤镜图像级联分类及外部临床验证
Abstract
Purpose. To compare deep learning architectures and classification schemes for dermoscopic images of skin neoplasms and assess their generalization on transfer from open international datasets to independent clinical datasets of Russian practice. Methods. Four architectures (ViT-B/16, Swin-S, ConvNeXt-S, EfficientNetV2-S) were compared in three schemes: binary (malignant/benign), single-stage four-class (benign, MEL, SCC, BCC), and a two-stage cascade (binary triage, then three-class differentiation MEL/SCC/BCC). All models used ImageNet-pretrained weights and a single augmentation protocol on aggregated open ISIC Archive data, and were evaluated on an internal held-out sample and two clinical datasets (Melanoscope AI mobile system; Sechenov University). Results. Internally the binary stage attains ROC-AUC 0.952-0.966; on Sechenov University it drops to 0.797-0.893, sensitivity to 0.53-0.67, and ECE rises from 0.02 to 0.27-0.39 with underestimation of malignancy, quantifying a generalization gap in ranking and calibration. Paired tests confirm one inter-architecture result on clinical data: the deficit of ViT-B/16 at the binary stage (p<0.05); at the differentiation stage no architecture has a proven advantage. The cascade raises macro F1 over single-stage four-class classification for most architectures, but significantly only for ViT-B/16, by recovering malignant lesions assigned to the dominant benign class. On ISIC MILK10k, direct 11-class classification yields mean-class sensitivity 0.525. Conclusion. A tunable triage threshold gives sensitivity control not attainable in standard single-stage (argmax) classification and better reproduces clinical differential-diagnosis logic. The persistent generalization gap mandates external clinical validation and recalibration before deployment.
Chinese Translation
目的：比较深度学习架构和皮肤肿瘤皮肤镜图像的分类方案，并评估其从开放国际数据集转移到俄罗斯临床独立数据集的泛化能力。方法：比较了四种架构（ViT-B/16、Swin-S、ConvNeXt-S、EfficientNetV2-S）在三种方案中的表现：二分类（恶性/良性）、单阶段四类（良性、黑色素瘤（MEL）、鳞状细胞癌（SCC）、基底细胞癌（BCC））和两阶段级联（首先进行二分类筛查，然后进行三类区分（MEL/SCC/BCC））。所有模型均使用在ImageNet上预训练的权重，并在聚合的开放ISIC档案数据上采用单一增强协议进行评估，同时在内部保留样本和两个临床数据集（Melanoscope AI移动系统；谢切诺夫大学）上进行评估。结果：在内部，二分类阶段的ROC-AUC达到0.952-0.966；在谢切诺夫大学的数据上降至0.797-0.893，灵敏度降至0.53-0.67，ECE从0.02上升至0.27-0.39，表明恶性肿瘤的低估，量化了排名和校准中的泛化差距。配对测试确认了临床数据中一个架构间结果：ViT-B/16在二分类阶段的不足（p<0.05）；在区分阶段没有架构具有明显优势。级联方法在大多数架构中提高了宏F1分数，相较于单阶段四类分类，但仅对ViT-B/16显著，通过恢复被分配到主导良性类别的恶性病变。在ISIC MILK10k数据集上，直接的11类分类的平均类灵敏度为0.525。结论：可调的筛查阈值提供了在标准单阶段（argmax）分类中无法实现的灵敏度控制，并更好地再现了临床鉴别诊断逻辑。持续存在的泛化差距要求在部署前进行外部临床验证和重新校准。
cs.CV / 37 / 2606.13136
An Extensible and Lightweight Unified Architecture for Demosaicing Pixel-bin Image Sensors
一种可扩展且轻量化的统一架构用于去马赛克像素分箱图像传感器
Abstract
Pixel-bin image sensors are becoming the default choice for smartphone cameras due to their resolution vs light-gathering trade-off. However, their larger inter-color separation compared to the Bayer color filter array (CFA) makes them challenging to demosaic. Furthermore, existing deep learning-based demosaicing methods are CFA-specific, requiring multiple individual models that take up precious onboard resources and demand larger development and maintenance efforts. In this work, we propose a modular unified architecture for demosaicing various pixel-bin sensors that provides higher image quality while being extensible and lightweight. Additionally, to enable plug-and-play operation, we introduce a learning-free CFA-identification module to detect the CFA type of raw data accurately.
Chinese Translation
像素分箱图像传感器因其分辨率与光收集能力的权衡，正成为智能手机相机的默认选择。然而，与拜耳颜色滤波阵列（CFA）相比，它们较大的色彩间隔使得去马赛克过程具有挑战性。此外，现有的基于深度学习的去马赛克方法是针对CFA特定的，需要多个独立模型，这占用了宝贵的板载资源，并且需要更大的开发和维护工作。在本研究中，我们提出了一种模块化的统一架构，用于去马赛克各种像素分箱传感器，该架构在提供更高图像质量的同时，具有可扩展性和轻量化。此外，为了实现即插即用操作，我们引入了一个无学习的CFA识别模块，以准确检测原始数据的CFA类型。
cs.CV / 38 / 2606.13156
Iterative Visual Thinking: Teaching Vision-Language Models Spatial Self-Correction through Visual Feedback
迭代视觉思维：通过视觉反馈教导视觉-语言模型进行空间自我修正
Abstract
Vision-language models (VLMs) achieve strong singleshot spatial grounding, yet lack any mechanism to observe and correct their own predictions. We find that naively prompting a VLM to iterate over rendered visualizations of its predictions causes catastrophic failure:
[email protected] on referring expression comprehension collapses from 79.6% to 48.7% (a 31 percentage point drop), revealing a fundamental gap between grounding capability and self-correction ability. We propose Iterative Visual Thinking (IVT), a closed-loop framework in which the model predicts a bounding box, observes the prediction rendered on the image, and iteratively refines through visual feedback. A two-phase training recipe closes the self-correction gap: first, we exploit the base model's own predictions as realistic errors and prompt a teacher VLM to generate corrective reasoning traces, yielding supervised data without human annotation; second, we apply Group Relative Policy Optimization (GRPO) with a simple IoU reward to stabilize multi-step refinement. On a mixed benchmark spanning RefCOCOg, Ref-Adv, and Ref-L4 (505 test samples), SFT warm-up with IVT surpasses the single-shot base model on every metric:
[email protected] rises to 82.0% (+2.4pp),
[email protected] to 74.1% (+3.2pp), and
[email protected] to 48.3% (+2.8pp). GRPO further reduces per-step IoU degradation by 5x, stabilizing the refinement trajectory. All training uses only 2,400 samples on a single GPU, demonstrating that spatial self-correction is a learnable capability that can be instilled at modest scale.
Chinese Translation
视觉-语言模型（VLMs）在单次空间定位中表现出色，但缺乏观察和修正自身预测的机制。我们发现，简单地提示VLM在其预测的渲染可视化上进行迭代会导致灾难性的失败：在指称表达理解中，
[email protected]从79.6%骤降至48.7%（下降31个百分点），揭示了定位能力与自我修正能力之间的根本差距。我们提出了迭代视觉思维（IVT），这是一个闭环框架，其中模型预测一个边界框，观察在图像上渲染的预测，并通过视觉反馈进行迭代修正。一个两阶段的训练方案缩小了自我修正的差距：首先，我们利用基础模型自身的预测作为真实错误，并提示教师VLM生成纠正推理轨迹，从而在没有人工标注的情况下产生监督数据；其次，我们应用带有简单IoU奖励的群体相对策略优化（GRPO）来稳定多步修正。在涵盖RefCOCOg、Ref-Adv和Ref-L4（505个测试样本）的混合基准上，使用IVT进行的SFT预热在每个指标上都超越了单次基础模型：
[email protected]上升至82.0%（+2.4个百分点），
[email protected]上升至74.1%（+3.2个百分点），
[email protected]上升至48.3%（+2.8个百分点）。GRPO进一步将每步IoU降级减少了5倍，稳定了修正轨迹。所有训练仅使用2400个样本在单个GPU上进行，证明空间自我修正是一种可学习的能力，可以在适度规模下灌输。
cs.CV / 39 / 2606.13188
Transformer-Guided Graph Attention for Direct Cardiac Mesh Reconstruction: A Structural Digital Twin Framework
基于变换器引导的图注意力用于直接心脏网格重建：结构数字双胞胎框架
Abstract
Building patient-specific cardiac models sits at the heart of precision cardiology, yet getting those models into clinical use keeps running into the same wall: mesh generation is slow, messy, and frustrating. The standard workflow -- segmenting the image, running Marching Cubes, and then manually cleaning up the result -- is time-consuming, inconsistent across operators, and demands specialist knowledge most clinical teams do not have. We take a fundamentally different approach. Instead of treating segmentation and mesh generation as two separate problems, we train a single end-to-end network that goes directly from a raw 3D medical image to a smooth, simulation-ready cardiac surface mesh. The core is a 3D Swin Transformer encoder-decoder that extracts volumetric features from CT or MRI volumes, paired with a Graph Attention Network (GAT) head that iteratively deforms a template mesh to fit the patient's cardiac boundary. We tested on the MM-WHS 2017 benchmark using both CT and MRI. Segmentation scores were competitive (Dice of 0.84 on CT, 0.83 on MRI), but the primary focus is mesh quality: mean Chamfer distance of 1.8 mm, with 95th-percentile surface distance below 5 mm. Every mesh is produced in a single forward pass -- no Marching Cubes, no smoothing filters, no manual cleanup. We argue that for cardiac digital twin pipelines, geometric fidelity and topological correctness matter more than pixel-level Dice scores. By removing the post-processing bottleneck, this approach makes patient-specific cardiac simulation substantially more accessible for clinical use.
Chinese Translation
构建特定患者的心脏模型是精准心脏病学的核心，但将这些模型应用于临床时常常面临同样的障碍：网格生成过程缓慢、混乱且令人沮丧。标准工作流程——图像分割、运行 Marching Cubes，然后手动清理结果——耗时长、操作不一致，并且需要大多数临床团队所不具备的专业知识。我们采取了根本不同的方法。我们不再将分割和网格生成视为两个独立的问题，而是训练一个单一的端到端网络，直接从原始的三维医学图像生成光滑、可用于仿真的心脏表面网格。核心是一个三维 Swin Transformer 编码器-解码器，它从 CT 或 MRI 体积中提取体积特征，并配备一个图注意力网络（Graph Attention Network, GAT）头部，迭代地变形模板网格以适应患者的心脏边界。我们在 MM-WHS 2017 基准上进行了测试，使用了 CT 和 MRI。分割得分具有竞争力（CT 的 Dice 为 0.84，MRI 为 0.83），但主要关注的是网格质量：平均 Chamfer 距离为 1.8 毫米，95 百分位表面距离低于 5 毫米。每个网格都是通过单次前向传递生成的——无需 Marching Cubes、平滑滤波器或手动清理。我们认为，对于心脏数字双胞胎管道，几何保真度和拓扑正确性比像素级的 Dice 得分更为重要。通过消除后处理瓶颈，这种方法使特定患者的心脏仿真在临床应用中变得更加可及。
cs.CV / 40 / 2606.13206
Visual Place Recognition in Forests with Depth-Aware Distillation
基于深度感知蒸馏的森林视觉位置识别
Abstract
Visual place recognition in natural forest environments remains challenging due to repetitive vegetation, weak structural cues, and significant appearance variation across traversals. To address this limitation, this paper proposes a lightweight depth-aware distillation framework that injects geometric cues into a DINOv2-based place recognition model, while maintaining its pre-trained descriptor space. Evaluated on the recent WildCross benchmark, the proposed approach yields gains over an appearance-only counterpart, providing robustness to appearance variations. These results demonstrate the importance of depth as a strong complementary modality for place recognition in natural environments and identify depth-aware distillation as a promising direction for more robust forest perception.
Chinese Translation
在自然森林环境中，视觉位置识别仍然面临挑战，原因包括植被重复、结构线索弱以及穿越过程中的显著外观变化。为了解决这一限制，本文提出了一种轻量级的深度感知蒸馏框架，该框架将几何线索注入基于 DINOv2 的位置识别模型，同时保持其预训练的描述子空间。在最近的 WildCross 基准测试中，所提出的方法相比于仅基于外观的对照组表现出更好的性能，增强了对外观变化的鲁棒性。这些结果表明，深度作为一种强有力的补充模态在自然环境中的位置识别中具有重要性，并且将深度感知蒸馏视为提高森林感知鲁棒性的一个有前景的方向。
cs.CV / 41 / 2606.13267
TimeLens: On-Device Artifact Recognition with Retrieval-Augmented Question Answering for the Grand Egyptian Museum
TimeLens：针对大埃及博物馆的设备内文物识别与检索增强问答
Abstract
TimeLens is an AI-powered bilingual mobile guide for the Grand Egyptian Museum (GEM). Pointing a phone at an exhibit, a visitor sees the artifact recognized in real time and can ask follow-up questions answered in English or Arabic. The work addresses three problems specific to in-gallery deployment: fine-grained visual similarity among 51 catalogued artifacts (many near-identical Ramesside statues), the gap between curated training data and handheld camera conditions, and the risk of an AI guide stating unsupported historical facts. Two engineering contributions are reported. First, an on-device artifact detector was developed through a data-quality-driven iteration study -- from foundation-model auto-annotation (YOLO-World), through spatial label-cleaning rules, to a fully hand-annotated dataset -- isolating label quality as the decisive factor: the final YOLOv8n model resolves every previously failing class while remaining a 5.97 MB TensorFlow Lite asset that runs in real time on a mid-range phone (
[email protected] = 0.995,
[email protected]:0.95 = 0.924). Second, a bilingual Retrieval-Augmented Generation (RAG) guide, grounded in a 108-record ChromaDB knowledge base, was benchmarked across seven candidate language models, with Gemma 4 E2B (Q4 K M) selected; ten targeted optimizations reduce end-to-end latency from over 30 s to approximately 10 s. Both subsystems are integrated in a production Flutter application with bilingual interface, museum location gating, and text-to-speech support.
Chinese Translation
TimeLens 是一款为大埃及博物馆（GEM）提供的人工智能驱动的双语移动导览工具。游客只需将手机对准展品，即可实时识别文物，并可以用英语或阿拉伯语提问。该研究解决了三个特定于展馆部署的问题：51件编目文物之间的细粒度视觉相似性（许多几乎相同的拉美西斯时期雕像）、策划训练数据与手持相机条件之间的差距，以及人工智能导览可能陈述未经支持的历史事实的风险。报告了两个工程贡献。首先，通过数据质量驱动的迭代研究开发了一种设备内文物检测器——从基础模型自动标注（YOLO-World），经过空间标签清理规则，到完全手动标注的数据集——将标签质量隔离为决定性因素：最终的 YOLOv8n 模型解决了所有先前失败的类别，同时仍然保持为 5.97 MB 的 TensorFlow Lite 资产，能够在中档手机上实时运行（
[email protected] = 0.995，
[email protected]:0.95 = 0.924）。其次，基于 108 条记录的 ChromaDB 知识库，构建了一个双语检索增强生成（RAG）导览，并在七个候选语言模型中进行了基准测试，最终选择了 Gemma 4 E2B（Q4 K M）；十项针对性优化将端到端延迟从超过 30 秒减少到约 10 秒。这两个子系统集成在一个生产级 Flutter 应用中，具有双语界面、博物馆位置限制和文本转语音支持。
cs.CV / 42 / 2606.13275
Zero-Shot Captioning for Cultural Heritage: Automated Image Analysis of Traditional Indonesian Clothing
文化遗产的零样本图像描述：传统印尼服装的自动图像分析
Abstract
This paper presents Custom ZeroCLIP, a retrieval-augmented vision-language framework for zero-shot captioning of Indonesian traditional garments. The dataset contains 3,800 expert-annotated images from all 38 Indonesian provinces. Using a province-level inductive zero-shot protocol, the model is trained on 24 seen provinces, validated on 6 seen provinces, and evaluated on 8 unseen provinces. The framework combines a frozen CLIP ViT-B/32 image encoder, a CLIP text encoder, a BERT text encoder, and an LSTM caption decoder. During inference, unseen-province labels and captions are unavailable, and retrieval uses only captions from training provinces. No unseen-province image, label, or caption is used during training, validation, or retrieval-bank construction. Custom ZeroCLIP achieves a CLIPScore of 0.8536, BLEU-4 of 0.3342, and METEOR of 0.4859, outperforming existing baselines. Ablation results show that retrieval improves cultural vocabulary recovery with a 19.3\% METEOR gain, while human evaluation confirms stronger cultural accuracy and fluency. The results demonstrate the effectiveness of retrieval-augmented domain adaptation for culturally grounded caption generation in low-resource heritage settings. The dataset is publicly available at https://github.com/AnugrahAidinYotolembah/Traditional-Indonesian-Clothing-Captioning-Dataset.
Chinese Translation
本文提出了Custom ZeroCLIP，一种用于印尼传统服装零样本描述的检索增强视觉-语言框架。该数据集包含来自印尼38个省份的3,800张专家注释图像。采用省级归纳零样本协议，模型在24个已见省份上进行训练，在6个已见省份上进行验证，并在8个未见省份上进行评估。该框架结合了一个冻结的CLIP ViT-B/32图像编码器、一个CLIP文本编码器、一个BERT文本编码器和一个LSTM描述解码器。在推理过程中，未见省份的标签和描述不可用，检索仅使用来自训练省份的描述。在训练、验证或检索库构建过程中，不使用任何未见省份的图像、标签或描述。Custom ZeroCLIP实现了0.8536的CLIPScore、0.3342的BLEU-4和0.4859的METEOR，超越了现有基准。消融实验结果表明，检索提高了文化词汇的恢复，METEOR提升了19.3\%，而人工评估则确认了更强的文化准确性和流畅性。结果表明，在低资源遗产环境中，检索增强的领域适应对文化根植的描述生成的有效性。数据集可在https://github.com/AnugrahAidinYotolembah/Traditional-Indonesian-Clothing-Captioning-Dataset获取。
cs.CV / 43 / 2606.13288
Cross-Modal Masked Compositional Concept Modeling for Enhancing Visio-Linguistic Compositionality
跨模态掩蔽组合概念建模以增强视觉-语言组合性
Abstract
Contrastively trained vision-language models like CLIP, have made remarkable progress in learning joint image-text representations, but still face challenges in compositional understanding. They often exhibit a "bag-of-words" behavior--struggling to capture the object relations, attribute-object bindings, and word order dependencies. This limitation arises not only from the reliance on global, single-vector representations for optimization, but also from the insufficient exploitation and modeling of the rich compositional information inherently present in paired image text data. In this work, we propose MACCO (MAsked Compositional Concept MOdeling), a framework that masks compositional concepts in one modality and reconstructs them conditioned on the full contextual information from the other, enabling the model to capture and align cross-modal compositional structures more effectively. To facilitate this process, we introduce two auxiliary objectives that jointly align and regularize masked features both inter-modally and intra-modally. Extensive experiments on five compositional benchmarks, along with in-depth analyses, demonstrate that our approach not only significantly enhances compositionality in VLMs but also improves their ability to capture syntactic structure and linguistic information. Additionally, the improved compositionality also benefits text-to-image generation and multimodal large language model. Code is available at https://github.com/hiker-lw/MACCO.
Chinese Translation
对比训练的视觉-语言模型，如CLIP，在学习联合图像-文本表示方面取得了显著进展，但在组合理解上仍面临挑战。它们常常表现出“词袋”行为——难以捕捉对象关系、属性-对象绑定和词序依赖。这一局限性不仅源于对全局单向量表示的优化依赖，还源于对成对图像文本数据中固有的丰富组合信息的不足利用和建模。在本研究中，我们提出了MACCO（掩蔽组合概念建模）框架，该框架在一种模态中掩蔽组合概念，并基于另一种模态的完整上下文信息重建它们，从而使模型能够更有效地捕捉和对齐跨模态组合结构。为促进这一过程，我们引入了两个辅助目标，这些目标共同对齐和正则化掩蔽特征，既在模态间也在模态内。对五个组合基准的广泛实验以及深入分析表明，我们的方法不仅显著增强了视觉-语言模型的组合性，还提高了它们捕捉句法结构和语言信息的能力。此外，改进的组合性还对文本到图像生成和多模态大型语言模型带来了好处。代码可在 https://github.com/hiker-lw/MACCO 获取。
cs.CV / 44 / 2606.13289
HYDRA-X: Native Unified Multimodal Models with Holistic Visual Tokenizers
HYDRA-X：具有整体视觉标记器的原生统一多模态模型
Abstract
Holistic visual tokenizers are fundamental to unified multimodal models (UMMs) as they map diverse visual inputs into a unified representation space. In this paper, we present HYDRA-X, the first UMM that unifies image and video tokenization within a single Vision Transformer (ViT). Our design is driven by two core challenges: efficiently injecting spatiotemporal reconstruction capability into a native ViT, and embedding image- and video-level semantic awareness into the latent space. To address the first, comprehensive ablations reveal two key findings: (1) frame-level causal temporal attention suffices for visual reconstruction, whereas full spatiotemporal attention degrades it; and (2) hierarchical temporal compression substantially outperforms single-step alternatives. To tackle the second, we propose a lightweight decompressor that upsamples temporally compressed features under joint image-video teacher supervision, thereby enforcing complementary semantic structures within the compact latent space. Building on this holistic tokenizer, we further propose a principled improvement of the editing pipeline: source-target interaction should occur at the latent level inside the tokenizer rather than at the semantic level inside the LLM, substantially improving editing consistency and accelerating convergence. Instantiated at the 7B dense model, HYDRA-X achieves strong performance across image and video understanding and generation tasks, paving the way for future unified-tokenizer UMMs.
Chinese Translation
整体视觉标记器是统一多模态模型（UMMs）的基础，因为它们将多样的视觉输入映射到统一的表示空间。在本文中，我们提出了HYDRA-X，这是第一个在单一视觉变换器（ViT）中统一图像和视频标记化的UMM。我们的设计受到两个核心挑战的驱动：有效地将时空重建能力注入原生ViT，以及在潜在空间中嵌入图像和视频级别的语义意识。为了解决第一个问题，全面的消融实验揭示了两个关键发现：（1）帧级因果时序注意力足以用于视觉重建，而完整的时空注意力则会降低其效果；（2）层次时序压缩显著优于单步替代方案。为了解决第二个问题，我们提出了一种轻量级解压器，在联合图像-视频教师监督下对时序压缩特征进行上采样，从而在紧凑的潜在空间内强制执行互补的语义结构。在这一整体标记器的基础上，我们进一步提出了编辑流程的原则性改进：源-目标交互应在标记器内部的潜在层面进行，而不是在大型语言模型（LLM）内部的语义层面进行，从而显著提高编辑一致性并加速收敛。在7B密集模型中实例化的HYDRA-X在图像和视频理解及生成任务中表现出色，为未来的统一标记器UMMs铺平了道路。
cs.CV / 45 / 2606.13303
DuET: Dual Expert Trajectories for Diffusion Image Editing
DuET：用于扩散图像编辑的双专家轨迹
Abstract
Recent diffusion editors perform diverse instruction-based edits while conditioning on the source image at every denoising step. Yet persistent source-image conditioning can limit how fully an edit is executed and how natural the result appears, especially when the target scene diverges substantially from the input. We introduce DuET (Dual Expert Trajectories), a training-free inference method that temporarily relaxes source-image conditioning by transitioning through a text-to-image phase before returning to edit mode, allowing the denoising trajectory to move toward the target distribution while retaining the structural benefits of image-conditioned editing. Without modifying model weights or increasing sampling cost, DuET consistently improves instruction relevance, semantic fidelity, and perceptual quality across diverse models and benchmarks. In some cases, these gains come with a modest reduction in source-image preservation, revealing a predictable trade-off between source preservation and edit fidelity.
Chinese Translation
近期的扩散编辑器在每个去噪步骤中基于源图像执行多样的基于指令的编辑。然而，持续的源图像条件限制了编辑的执行程度和结果的自然性，尤其是在目标场景与输入有显著偏离时。我们提出了DuET（双专家轨迹），这是一种无训练推理方法，通过在返回编辑模式之前过渡到文本到图像的阶段，暂时放松源图像条件，从而允许去噪轨迹朝向目标分布移动，同时保留图像条件编辑的结构优势。在不修改模型权重或增加采样成本的情况下，DuET在多个模型和基准测试中持续改善了指令相关性、语义保真度和感知质量。在某些情况下，这些提升伴随着源图像保留的适度减少，揭示了源保留与编辑保真度之间的可预测权衡。
cs.CV / 46 / 2606.13304
ReFree: Towards Realistic Co-Speech Video Generation via Reward-Free RL and Multilevel Speech Guidance
ReFree：通过无奖励强化学习和多层次语音指导实现逼真的共语视频生成
Abstract
Speech-driven talking character animation seeks to generate life-like portrait videos that convey natural conversation behavior, aligning facial motion with spoken audio. Although recent advances in video generation have substantially improved realism in video-based animation, achieving both accurate lip articulation and expressive behavior remains challenging. Existing approaches typically trade off precise phoneme-to-lip synchronization against dynamic facial expressions and head motion, yielding animations that are either accurate yet rigid, or expressive but poorly synchronized. We address this challenge by proposing ReFree-S2V, a flow-matching speech-to-portrait animation framework that builds upon a pretrained video generation model to achieve fine-grained speech articulation and high-level expressive cues in speech-driven portrait animation. This model introduces a multi-level speech representation capturing phonetic and prosodic information at both local and global granularities. These representations are selectively injected into transformer blocks via learnable level selectors, enabling both accurate lip synchronization and natural expressive motion. To achieve natural head movements, we further introduce a novel reward-free reinforcement learning scheme into flow-matching training to discourage perceptually implausible motion without relying on handcrafted synchronization metrics or reward models, or the high cost of human preference annotation. Extensive experiments demonstrate that ReFree-S2V achieves state-of-the-art performance, significantly outperforming existing methods in both quantitative lip-sync accuracy and qualitative human evaluations of naturalness and expressivity.
Chinese Translation
基于语音驱动的对话角色动画旨在生成生动的肖像视频，以传达自然的对话行为，使面部动作与所说的音频相一致。尽管最近在视频生成方面的进展显著提高了基于视频的动画的真实感，但实现准确的唇部发音和富有表现力的行为仍然具有挑战性。现有的方法通常在精确的音素与唇部同步与动态面部表情和头部运动之间进行权衡，导致生成的动画要么准确但僵硬，要么富有表现力但同步性差。我们通过提出ReFree-S2V来解决这一挑战，这是一种流匹配的语音到肖像动画框架，基于预训练的视频生成模型，实现细粒度的语音发音和高层次的表达线索。该模型引入了一种多层次的语音表示，捕捉局部和全局粒度的语音和韵律信息。这些表示通过可学习的层选择器选择性地注入到变换器块中，从而实现准确的唇部同步和自然的表现运动。为了实现自然的头部运动，我们进一步引入了一种新颖的无奖励强化学习方案到流匹配训练中，以抑制感知上不合理的运动，而不依赖于手工制作的同步度量或奖励模型，或人类偏好标注的高成本。大量实验表明，ReFree-S2V在定量唇部同步准确性和定性自然性与表现力的人类评估中显著超越了现有方法，达到了最先进的性能。
cs.CV / 47 / 2606.13312
MagPlus: Bridging Micro-to-Regular Facial Expressions through Learnable Magnification
MagPlus：通过可学习的放大技术连接微表情与常规表情
Abstract
Facial micro-expressions are subtle and short-lived facial movements that provide important cues about genuine human emotions. However, modeling and generating them remains difficult because annotated micro-expression data is limited and the underlying facial motions are extremely weak. Existing micro-expression generation methods therefore often suffer from limited quality, weak robustness, and poor generalization. We propose MagPlus, a transferable micro-expression processing pipeline that connects micro-expression analysis with standard facial animation models. Instead of training a dedicated generator from scratch, MagPlus learns to magnify subtle facial motions into the range of regular facial expressions, transforming micro-expressions into signals that are compatible with existing facial expression processing models. The magnified sequence is then used by a standard facial expression model for tasks such as transfer and synthesis. A complementary DeMagPlus module then restores the generated motion back to realistic micro-expression intensity levels while preserving the synthesized dynamics. We evaluate the framework using four facial animation models: FOMM, FSRT, MetaPortrait, and EmoPortraits. None of these models are trained on micro-expression data. Experiments show that MagPlus-DeMagPlus enables pretrained macro-expression models to generate more realistic micro-expression motion without retraining the backbones.
Chinese Translation
面部微表情是细微且短暂的面部动作，提供了关于真实人类情感的重要线索。然而，由于标注的微表情数据有限且基础面部动作极其微弱，建模和生成微表情仍然困难。因此，现有的微表情生成方法通常面临质量有限、鲁棒性差和泛化能力差的问题。我们提出了MagPlus，一个可转移的微表情处理管道，将微表情分析与标准面部动画模型连接起来。MagPlus并不是从头开始训练一个专用生成器，而是学习将细微的面部动作放大到常规面部表情的范围，将微表情转化为与现有面部表情处理模型兼容的信号。放大的序列随后被标准面部表情模型用于转移和合成等任务。一个补充的DeMagPlus模块则在保持合成动态的同时，将生成的动作恢复到真实的微表情强度水平。我们使用四个面部动画模型（FOMM、FSRT、MetaPortrait和EmoPortraits）评估该框架。这些模型均未在微表情数据上进行训练。实验表明，MagPlus-DeMagPlus使得预训练的宏表情模型能够生成更真实的微表情动作，而无需重新训练基础模型。
cs.CV / 48 / 2606.13315
Masked and Predictive Self-Supervised Foundation Models for 3D Brain MRI
用于3D脑部MRI的掩蔽与预测自监督基础模型
Abstract
Self-supervised foundation models have shown strong promise in medical imaging. However, existing MRI foundation-model studies have primarily emphasized segmentation and dense prediction tasks, while systematic investigation of self-supervised foundation models for MRI-based disease detection remains limited. In this work, we investigate two major self-supervised pretraining paradigms for MRI-based disease detection: reconstruction-based learning via Masked Autoencoders (MAE) and predictive representation learning via Joint Embedding Predictive Architectures (JEPA). We study the role of auxiliary objectives by introducing a novel spectral-domain reconstruction loss for MAE to enhance sensitivity to fine-grained anatomical structure, and by integrating variance--covariance regularization (VCR) within our JEPA framework to encourage decorrelated latent representations. Our models are pretrained on heterogeneous single-contrast MRI volumes in a contrast-agnostic setting, without modality concatenation. Across five downstream disease detection tasks, our results highlight the importance of self-supervised objective design for medical foundation model pretraining, demonstrating that the downstream benefit of each objective is determined by its relevance to the task's structure. Specifically, spectral regularization yields the largest improvements when the downstream discriminative signal is characterized by strong high-frequency anatomical structures, while covariance regularization is most beneficial when discriminative information spans multiple decorrelated feature dimensions. MAE with spectral-domain supervision consistently achieves superior downstream performance for MRI-based disease detection. These findings suggest that self-supervised objectives in medical imaging encode specific biases, and their downstream benefit is fundamentally conditioned on the task's structure.
Chinese Translation
自监督基础模型在医学影像中展现出强大的潜力。然而，现有的MRI基础模型研究主要强调分割和密集预测任务，而对基于MRI的疾病检测的自监督基础模型的系统性研究仍然有限。在本研究中，我们探讨了两种主要的自监督预训练范式用于基于MRI的疾病检测：通过掩蔽自编码器（Masked Autoencoders, MAE）进行的重建学习和通过联合嵌入预测架构（Joint Embedding Predictive Architectures, JEPA）进行的预测表示学习。我们通过引入一种新颖的光谱域重建损失来增强MAE对细粒度解剖结构的敏感性，并在JEPA框架中整合方差-协方差正则化（Variance-Covariance Regularization, VCR）以鼓励去相关的潜在表示，从而研究辅助目标的作用。我们的模型在异质单对比MRI体积上进行预训练，采用无对比度连接的设置。在五个下游疾病检测任务中，我们的结果突显了自监督目标设计对医学基础模型预训练的重要性，表明每个目标的下游效益取决于其与任务结构的相关性。具体而言，当下游判别信号由强高频解剖结构特征表征时，光谱正则化带来了最大的改进，而当判别信息跨越多个去相关特征维度时，协方差正则化最为有利。带有光谱域监督的MAE在基于MRI的疾病检测中始终实现了优越的下游性能。这些发现表明，医学影像中的自监督目标编码了特定的偏差，其下游效益在根本上依赖于任务的结构。
cs.CV / 49 / 2606.13332
OR-Action: Multi-Role Video Understanding with Fine-Grained Actions
OR-Action：多角色视频理解与细粒度动作
Abstract
Fine-grained understanding of operating room (OR) activity could enable workflow-aware assistance, yet remains difficult due to clutter, occlusions, and limited sensing. The prevailing approach to model this environment is scene graphs as an interpretable representation of OR interactions. Converting their frame-wise relational predictions into temporally extended, fine-grained actions however, is challenging without explicit temporal modeling. To enable a principled temporal evaluation of current OR understanding methods, we introduce the first action-centric benchmark built on a publicly available ego-exocentric OR dataset by defining a fine-grained, multi-role action taxonomy and generating dense action segments via distillation from ground-truth scene graph state changes. Experiments on this benchmark show that current scene graph prediction methods struggle to model temporal structure, even when adding explicit modeling through Graph Neural Networks. We therefore introduce a vision-only temporal model that outperforms graph-based methods significantly when using all available egocentric video as input. Building on this model we also introduce a novel multi- to single-view feature alignment strategy that improves single-view performance on multi-role action recognition, mitigating the need for extensive egocentric video capture. Benchmark and code will be released upon acceptance.
Chinese Translation
对手术室（OR）活动的细粒度理解能够实现基于工作流程的智能辅助，但由于环境杂乱、遮挡和传感器限制，这一任务仍然具有挑战性。目前对这一环境的建模方法是使用场景图作为OR交互的可解释表示。然而，将其逐帧的关系预测转换为时间延续的细粒度动作在没有明确时间建模的情况下是困难的。为了对当前的OR理解方法进行原则性的时间评估，我们引入了第一个以动作为中心的基准，基于一个公开可用的自我-外部OR数据集，通过定义细粒度的多角色动作分类法并通过从真实场景图状态变化中提取生成密集动作片段。基准测试的实验表明，当前的场景图预测方法在建模时间结构方面存在困难，即使在通过图神经网络添加明确建模时也是如此。因此，我们提出了一种仅基于视觉的时间模型，当使用所有可用的自我中心视频作为输入时，其性能显著优于基于图的方法。在此模型的基础上，我们还引入了一种新颖的多视图到单视图特征对齐策略，改善了多角色动作识别中的单视图性能，减少了对广泛自我中心视频捕获的需求。基准和代码将在接受后发布。
cs.CV / 50 / 2606.13341
Dual-Domain Equivariant Generative Adversarial Network for Multimodal CT-PET Synthesis
用于多模态CT-PET合成的双域等变生成对抗网络
Abstract
We present a Dual-Domain Equivariant Generative Adversarial Network (DDE-GAN) for multimodal CT-PET image synthesis. Traditional GAN-based approaches often operate solely in the spatial domain and ignore geometric consistency, resulting in limited structural fidelity. DDE-GAN addresses these challenges by jointly learning from both spatial and frequency (Fourier) domains, capturing complementary anatomical and spectral information. Furthermore, rotational equivariance embedded in the physics of the CT and PET measurements are integrated into the loss of both the generator and discriminator to ensure consistent responses under rotations, improving anatomical accuracy. A hierarchical dual-domain training strategy enforces intra- and inter-domain consistency through multi-stage loss functions. Evaluated on the HECKTOR 2022 CT-PET dataset, DDE-GAN achieves superior synthesis quality over baseline models for CT-PET image synthesis. The results demonstrate that combining dual-domain learning with geometric equivariance substantially enhances multimodal image synthesis accuracy and robustness, enabling practical applications in PET completion and data augmentation.
Chinese Translation
我们提出了一种用于多模态CT-PET图像合成的双域等变生成对抗网络（DDE-GAN）。传统的基于生成对抗网络（GAN）的方法通常仅在空间域中操作，忽视了几何一致性，导致结构保真度有限。DDE-GAN通过同时从空间域和频率域（傅里叶域）学习，解决了这些挑战，捕捉互补的解剖和光谱信息。此外，嵌入在CT和PET测量物理中的旋转等变性被整合到生成器和判别器的损失中，以确保在旋转下的一致响应，从而提高了解剖准确性。分层双域训练策略通过多阶段损失函数强制执行域内和域间一致性。在HECKTOR 2022 CT-PET数据集上的评估表明，DDE-GAN在CT-PET图像合成方面的合成质量优于基线模型。结果表明，将双域学习与几何等变性相结合显著提高了多模态图像合成的准确性和鲁棒性，为PET补全和数据增强的实际应用提供了可能。
cs.CV / 51 / 2606.13345
JointEdit3D: Feed-Forward 3D Scene Editing in a Unified Latent Space
JointEdit3D：统一潜在空间中的前馈3D场景编辑
Abstract
Existing 3D scene editing methods typically rely on per-scene optimization over explicit 3D representations or cascaded edit-and-reconstruct pipelines, resulting in high test-time cost, limited 3D awareness, and structural inconsistencies. To couple appearance synthesis and geometry prediction during editing, we build on a unified RGB-geometry reconstruction-generation latent space and adapt it to feed-forward 3D scene editing. The resulting framework, \textbf{JointEdit3D}, performs asymmetric latent inpainting by observing only a single edited RGB reference latent and generating the remaining RGB views and edited geometry latent under source-scene anchoring. JointEdit3D introduces a dedicated SceneAnchor Branch to inject source-scene structure without forcing direct copying, and adopts edit/background-aware losses to balance edited-region fidelity with unedited-content preservation. To address the lack of paired resources for standardized 3D scene editing evaluation, we introduce SceneEdit3D-15K, a dataset with 15K paired editing samples and renderer-provided 3D annotations, together with SceneEdit3D-Bench, a curated 100-sample benchmark. Experiments show that JointEdit3D improves edited-region quality and 3D structural completeness over prior baselines while maintaining competitive background preservation.
Chinese Translation
现有的3D场景编辑方法通常依赖于对显式3D表示的逐场景优化或级联的编辑与重建管道，这导致了高测试时间成本、有限的3D感知和结构不一致。为了在编辑过程中将外观合成和几何预测结合起来，我们基于统一的RGB-几何重建-生成潜在空间，并将其适配为前馈3D场景编辑。最终框架 extbf{JointEdit3D}通过仅观察单个编辑的RGB参考潜在进行不对称潜在修补，并在源场景锚定下生成其余的RGB视图和编辑的几何潜在。JointEdit3D引入了专门的场景锚定分支，以在不强制直接复制的情况下注入源场景结构，并采用编辑/背景感知损失来平衡编辑区域的保真度与未编辑内容的保留。为了解决标准化3D场景编辑评估中缺乏配对资源的问题，我们引入了SceneEdit3D-15K，一个包含15K对编辑样本和渲染器提供的3D注释的数据集，以及SceneEdit3D-Bench，一个精心策划的100样本基准。实验表明，JointEdit3D在提高编辑区域质量和3D结构完整性方面优于先前的基线，同时保持了竞争性的背景保留。
cs.CV / 52 / 2606.13366
Dual-Constrained Diffusion Image Compression for Operational Rate-Distortion-Perception Optimization
用于操作率-失真-感知优化的双约束扩散图像压缩
Abstract
The rate-distortion-perception (RDP) trade-off extends classical rate--distortion theory by imposing a distributional constraint on reconstructions, providing a unified framework for neural image compression that jointly governs fidelity and perceptual realism. While prior work achieves near-optimal rate--perception trade-offs, practical frameworks explicitly realizing the full RDP surface remain scarce, primarily due to the difficulty of introducing common randomness at the decoder. We propose DCIC (Dual-Constrained Diffusion Image Compression), which integrates a learned codec with a diffusion-based decoder governed by joint distortion and idempotence constraints. The distortion constraint bounds reconstruction fidelity relative to the base codec output; the idempotence constraint -- requiring that re-encoding the restored image recovers the base codec reconstruction -- serves as a tractable surrogate for the distributional perception requirement. Together, they steer the reverse denoising process via iterative optimization with consistent noise injection, realizing common randomness without additional rate overhead. At fixed rate, dual attenuation factors $(K_D, K_P)$ jointly navigate the Pareto frontier of the distortion-perception plane, enabling continuously adjustable fidelity-realism trade-offs from a single bitstream. DCIC$_{RD}$ ($K_P{=}0$) and DCIC$_{RP}$ ($K_D{=}0$) arise as boundary curves, with DCIC$_{RDP}$ ($K_D = K_P=1$) realizing the optimal interior operating point. Experiments on CelebA-HQ, CLIC2020, and ImageNet-1K across CNN, Transformer, and hybrid architectures confirm that DCIC$_{RDP}$ achieves superior BD-PSNR over all perceptual codecs, while DCIC$_{RP}$ matches dedicated perception-oriented methods in BD-FID, validating the practical value of full RDP surface navigation.
Chinese Translation
率-失真-感知（RDP）权衡通过对重建施加分布约束，扩展了经典的率-失真理论，为神经图像压缩提供了一个统一框架，联合管理保真度和感知真实感。尽管先前的工作实现了近似最优的率-感知权衡，但明确实现完整RDP表面的实用框架仍然稀缺，主要是由于在解码器中引入共同随机性的困难。我们提出了DCIC（双约束扩散图像压缩），它将学习到的编解码器与一个基于扩散的解码器相结合，该解码器受到联合失真和幂等性约束的控制。失真约束限制了重建的保真度相对于基础编解码器输出的范围；幂等性约束要求重新编码恢复的图像能够恢复基础编解码器的重建，作为对分布感知要求的可处理替代。两者共同通过一致的噪声注入进行迭代优化，引导反向去噪过程，实现共同随机性而无需额外的速率开销。在固定速率下，双衰减因子$(K_D, K_P)$共同导航失真-感知平面的帕累托前沿，使得从单一比特流中能够连续调整保真度与真实感的权衡。DCIC$_{RD}$（$K_P{=}0$）和DCIC$_{RP}$（$K_D{=}0$）作为边界曲线出现，而DCIC$_{RDP}$（$K_D = K_P=1$）实现了最佳的内部操作点。在CelebA-HQ、CLIC2020和ImageNet-1K上的实验，涵盖CNN、Transformer和混合架构，确认DCIC$_{RDP}$在所有感知编解码器中实现了优越的BD-PSNR，而DCIC$_{RP}$在BD-FID上与专门的感知导向方法相匹配，验证了完整RDP表面导航的实际价值。
cs.CV / 53 / 2606.13376
MoVerse: Real-Time Video World Modeling with Panoramic Gaussian Scaffold
MoVerse：基于全景高斯支架的实时视频世界建模
Abstract
We present MoVerse, a real-time video world model that creates an interactively navigable scene from a single narrow-field-of-view image. This setting is challenging because the input observes only a small fraction of the environment, while interactive roaming requires a complete surrounding world, persistent geometry, controllable camera motion, and temporally coherent high-fidelity observations. MoVerse addresses this problem by separating world construction from observation rendering. It first expands the input into a gravity-aligned 360$^\circ$ panorama with topology-aware diffusion, closing the missing field of view before 3D reasoning. It then lifts the panorama into a persistent 3D Gaussian scaffold using panoramic geometry-aware residual prediction, yielding a dense and directly renderable spatial memory. Finally, a Gaussian-conditioned video renderer translates scaffold renderings along user-specified camera trajectories into photorealistic video. To make this renderer practical for interaction, we train a bidirectional diffusion teacher for high-quality conditional rendering and distill it into a causal autoregressive student for bounded-latency streaming. This design combines the controllability and long-range consistency of explicit 3D representations with the perceptual quality of generative video models. MoVerse supports real-time scene roaming at 8~FPS on a single NVIDIA RTX~4090 GPU, demonstrating a practical path toward single-image world creation with interactive video output.
Chinese Translation
我们提出了MoVerse，一种实时视频世界模型，它能够从单张狭窄视场图像创建一个可交互导航的场景。这个设置具有挑战性，因为输入仅观察到环境的一小部分，而交互漫游需要一个完整的周围世界、持久的几何结构、可控的相机运动和时间一致的高保真观察。MoVerse通过将世界构建与观察渲染分离来解决这个问题。它首先通过拓扑感知扩散将输入扩展为一个重力对齐的360$^ ext{°}$全景，填补缺失的视场，然后进行3D推理。接着，它利用全景几何感知的残差预测将全景提升为持久的3D高斯支架，从而产生一个密集且可直接渲染的空间记忆。最后，一个高斯条件视频渲染器沿用户指定的相机轨迹将支架渲染转换为照片级真实感视频。为了使该渲染器在交互中实用，我们训练了一个双向扩散教师以实现高质量的条件渲染，并将其提炼为一个因果自回归学生，以实现有限延迟的流媒体。这一设计结合了显式3D表示的可控性和长距离一致性与生成视频模型的感知质量。MoVerse在单个NVIDIA RTX~4090 GPU上支持以8 FPS的速度进行实时场景漫游，展示了通过单图像世界创建与交互视频输出的实用路径。
cs.CV / 54 / 2606.13382
SmartFont: Dynamic Condition Allocation for Few-Shot Font Generation
SmartFont：用于少样本字体生成的动态条件分配
Abstract
Few-shot font generation simultaneously requires global structural completeness and fine-grained local style fidelity. Existing methods usually either rely on global content-style modeling, which is robust but imperfectly disentangled, or emphasize component/local modeling, which captures fine details but relies heavily on local priors and reference coverage. We argue that the key challenge is not merely to learn purer conditions, but to organize complementary yet biased global and local conditions through multi-level allocation during generation. To this end, we propose SmartFont, a diffusion-based few-shot font generation framework that combines global content-style generation with weakly supervised local corrective experts. The local branch performs semantic-spatial allocation by learning expert-wise local concepts and semantically meaningful spatial maps under weak component supervision, enabling fine-grained correction without requiring explicit component-conditioned inference. On top of this, a denoising-state condition allocation module adaptively weights global content, global style, and local corrective feature across timesteps and injection blocks. Extensive experiments show that SmartFont achieves better global-local balance, improves glyph quality and local detail fidelity.
Chinese Translation
少样本字体生成同时需要全球结构的完整性和细粒度的局部风格保真度。现有的方法通常依赖于全球内容-风格建模，这种方法虽然稳健但分离效果不完美，或者强调组件/局部建模，这种方法捕捉细节但过于依赖局部先验和参考覆盖。我们认为，关键挑战不仅在于学习更纯的条件，而在于通过多层次分配在生成过程中组织互补但有偏的全球和局部条件。为此，我们提出了SmartFont，一个基于扩散的少样本字体生成框架，它结合了全球内容-风格生成与弱监督的局部校正专家。局部分支通过在弱组件监督下学习专家级局部概念和语义上有意义的空间图进行语义-空间分配，从而实现细粒度的校正，而无需显式的组件条件推断。在此基础上，一个去噪状态条件分配模块自适应地在时间步和注入块之间加权全球内容、全球风格和局部校正特征。大量实验表明，SmartFont在全球-局部平衡方面表现更佳，提高了字形质量和局部细节保真度。
cs.CV / 55 / 2606.13410
Person Identification from Contextual Motion
基于上下文运动的人员识别
Abstract
We consider the problem of identifying people based on their motion styles. We present a generative model describing the action instance creation process and derive a probabilistic identity inference scheme for two common person identification scenarios motivated by the surveillance and authentication applications. We introduce a novel, \emph{interactive}, scenario for person identification from motion patterns. To this end, we formalize the identification process in the context of a sequential message exchange session between the subject and the system. The subject's behavior is modeled using a probabilistic generative model inspired by the Human Information Processing (HIP) paradigm. At each stage, the system presents a visual stimulus (a cue) to the subject and records their motion response. The cue is selected so as to maximize the mutual information of the expected response and the subject's identity. Once recorded, the response is used to update the a posteriori probability over possible subjects' identities. The process terminates once a sufficient classification confidence level is reached. To the best of our knowledge, this is the first time person identification is addressed in such interactive setting. We report high recognition rates on five publicly available datasets and our own novel dataset consisting of 4,476 recordings of 22 test subjects responding to 15 cues.
Chinese Translation
我们考虑基于运动风格识别人员的问题。我们提出了一种生成模型，描述了动作实例创建过程，并推导出一种概率身份推断方案，适用于两个常见的人员识别场景，这些场景受到监控和身份验证应用的启发。我们引入了一种新颖的、 extit{交互式}的人员识别场景，基于运动模式进行识别。为此，我们在主体与系统之间的顺序信息交换会话的背景下形式化识别过程。主体的行为使用一种受人类信息处理（Human Information Processing, HIP）范式启发的概率生成模型进行建模。在每个阶段，系统向主体呈现视觉刺激（提示），并记录他们的运动反应。提示的选择旨在最大化预期反应与主体身份之间的互信息。一旦记录，反应将用于更新可能主体身份的后验概率。该过程在达到足够的分类置信水平后终止。根据我们所知，这是首次在如此交互环境中解决人员识别问题。我们在五个公开可用的数据集以及我们自己新创建的数据集上报告了高识别率，该数据集包含22名测试主体对15个提示的4,476条记录。
cs.CV / 56 / 2606.13427
VietFashion: Benchmarking Sketch-Text Composed Image Retrieval for Cultural Outfits
VietFashion：文化服饰素描-文本组合图像检索基准
Abstract
Cultural garments pose a unique challenge for visual retrieval systems, as their identity often depends on subtle structural and symbolic details that are poorly captured by standard AI models. We introduce VietFashion, a new benchmark for sketch-text composed image retrieval centered on the Ao Dai, a traditional Vietnamese garment. VietFashion enables designers and researchers to retrieve culturally meaningful outfits using a combination of hand-drawn sketches, which convey garment structure, and textual descriptions, which encode cultural semantics. The dataset is initialized with 650 sketches and expanded using generative models to produce over 21,000 photorealistic images with aligned captions. Textual prompts that describe detailed outfit attributes, which are extracted from fashion magazines to ensure authenticity and diversity. To better reflect the inherent ambiguity of design intent, VietFashion adopts a multi-target retrieval setting, where a single query may correspond to multiple valid results. We establish standardized evaluation protocols and benchmark state-of-the-art composed image retrieval methods. Experimental results reveal significant performance gaps in modeling fine-grained cultural semantics and multi-modal composition, positioning VietFashion as a challenging benchmark for fine-grained fashion retrieval. The dataset is publicly available at: https://hng0303.github.io/VietFashion.
Chinese Translation
文化服饰对视觉检索系统提出了独特的挑战，因为它们的身份往往依赖于微妙的结构和象征细节，而这些细节在标准的人工智能模型中难以捕捉。我们介绍了VietFashion，这是一个以越南传统服饰“奥黛”（Ao Dai）为中心的素描-文本组合图像检索新基准。VietFashion使设计师和研究人员能够通过手绘素描（传达服装结构）与文本描述（编码文化语义）的结合，检索具有文化意义的服饰。该数据集初始包含650幅素描，并使用生成模型扩展，生成超过21,000幅与对齐标题的照片级真实图像。文本提示描述了详细的服装属性，这些属性是从时尚杂志中提取的，以确保真实性和多样性。为了更好地反映设计意图的固有模糊性，VietFashion采用了多目标检索设置，其中单个查询可能对应多个有效结果。我们建立了标准化的评估协议，并对最先进的组合图像检索方法进行了基准测试。实验结果揭示了在建模细粒度文化语义和多模态组合方面的显著性能差距，使VietFashion成为细粒度时尚检索的一个具有挑战性的基准。该数据集可在以下网址公开获取：https://hng0303.github.io/VietFashion。
cs.CV / 57 / 2606.13432
OmniDirector: General Multi-Shot Camera Cloning without Cross-Paired Data
OmniDirector：无需交叉配对数据的通用多镜头相机克隆
Abstract
Cloning camera motion from reference videos is an important task in video generation, as videos provide intuitive and precise control. Existing methods either directly use parametric representations that fail to handle multi-shot generation or synthesize cross-paired data, which suffer from data scarcity, resulting in poor performance in complicated camera motion cloning. To address these issues, we introduce a general camera motion representation that encodes cameras as grid motion videos. This camera grid represents the camera parameters visually and supports the integration of diverse trajectories for multi-shot video generation. Building upon this, we propose OmniDirector, a unified framework trained on a million-scale camera grid-video pairs that coordinates characters, actions, and cameras to provide director-level control for multimodal diffusion transformers. Furthermore, we design a novel hierarchical prompt expansion agent that harmoniously integrates different control signals by systematically describing camera motion and visual content through understanding signal relationships. Extensive experiments demonstrate the superior performance and outstanding controllability of our framework. Project page: https://ymlinfeng.github.io/OmniDirector.github.io/
Chinese Translation
从参考视频中克隆相机运动是视频生成中的一项重要任务，因为视频提供了直观且精确的控制。现有方法要么直接使用无法处理多镜头生成的参数化表示，要么合成交叉配对数据，这导致数据稀缺，从而在复杂相机运动克隆中表现不佳。为了解决这些问题，我们提出了一种通用的相机运动表示，将相机编码为网格运动视频。该相机网格以视觉方式表示相机参数，并支持多样轨迹的整合以进行多镜头视频生成。在此基础上，我们提出了OmniDirector，一个在百万规模的相机网格-视频对上训练的统一框架，协调角色、动作和相机，为多模态扩散变换器提供导演级控制。此外，我们设计了一种新颖的分层提示扩展代理，通过系统地描述相机运动和视觉内容，理解信号关系，从而和谐地整合不同的控制信号。大量实验表明我们的框架具有卓越的性能和出色的可控性。项目页面：https://ymlinfeng.github.io/OmniDirector.github.io/
cs.CV / 58 / 2606.13460
VISA: VLM-Guided Instance Semantic Auditing for 3D Occupancy World Models
VISA：基于VLM指导的3D占用世界模型实例语义审计
Abstract
Semantic 3D occupancy provides a voxelized world state for autonomous driving and robot decision making, but object and rare-class errors can affect free-space interpretation, collision checking, and temporal state propagation. We show that a common VLM strategy, aligning 3D voxel or object features with crop-caption embeddings, improves text-space similarity without reliably improving closed-set occupancy mIoU. Motivated by this mismatch, we propose VISA, a training-time semantic auditing approach for existing occupancy world models. VISA queries an offline VLM on a representative crop of each physical object instance, obtains a structured audit with class hypotheses, plausible confusions, reliability, attributes, and evidence, and propagates it along the object track. The audit is grounded to matched 3D object voxels and distilled into semantic logits through reliability-weighted taxonomy, attribute-factor, and scene-level audit graph losses, while inference remains unchanged and requires no VLM. On nuScenes, averaged across three runs, VISA improves OccWorld from 19.06 to 20.05 mIoU and GaussianWorld from 21.36 to 21.91 mIoU; on GaussianWorld, object mIoU improves from 18.18 to 19.16 and rare-class mIoU from 15.60 to 16.79. These results suggest that VLMs are better suited to closed-set occupancy as reliability-aware semantic auditors than as generic caption-embedding targets.
Chinese Translation
语义3D占用为自主驾驶和机器人决策提供了体素化的世界状态，但物体和稀有类别的错误可能影响自由空间的解释、碰撞检查和时间状态传播。我们展示了一种常见的VLM策略，即将3D体素或物体特征与裁剪-标题嵌入对齐，虽然提高了文本空间的相似性，但并未可靠地改善闭集占用的mIoU。基于这种不匹配，我们提出了VISA，一种针对现有占用世界模型的训练时语义审计方法。VISA在每个物理对象实例的代表性裁剪上查询离线VLM，获得包含类别假设、合理混淆、可靠性、属性和证据的结构化审计，并沿着物体轨迹传播。审计与匹配的3D物体体素相结合，并通过可靠性加权的分类法、属性因子和场景级审计图损失提炼为语义logits，而推理保持不变且不需要VLM。在nuScenes数据集上，经过三次运行的平均结果显示，VISA将OccWorld的mIoU从19.06提高到20.05，将GaussianWorld的mIoU从21.36提高到21.91；在GaussianWorld上，物体的mIoU从18.18提高到19.16，稀有类别的mIoU从15.60提高到16.79。这些结果表明，VLM更适合作为可靠性意识的语义审计者，而不是作为通用的标题嵌入目标。
cs.CV / 59 / 2606.13488
Point-Wise Geometry-Aware Transformer for Partial-to-Full Point Cloud Registration in Computer-Assisted Surgery
基于点的几何感知变换器用于计算机辅助外科中的部分到完整点云配准
Abstract
Partial-to-full registration remains challenging due to varying overlap ratios, fluctuating point densities, and the presence of noise. While transformers have shown strong potential for point cloud processing, prior methods typically confine them to global context aggregation, overlooking fine-grained local geometry crucial for accurate correspondence. We propose \emph{GAPR-Net}, a learning-based point cloud registration framework with a coarse-to-fine architecture that combines convolution and transformer modules, in which local and global information is fused between the partial and full point clouds using a cross-attention mechanism. To achieve this, a transformation-invariant point-wise geometric feature representation is proposed, which can robustly capture relative geometric features for individual points with respect to their neighboring points. To evaluate the effectiveness of the proposed approach, experiments are conducted on four geometrically distinct bones, including the tibia, femur, pelvis, and thoracic cartilage. The overall registration recall reaches 94.2\%, the method results in a low RMSE of 1.992 mm and $R^2$ values of 0.908 and 0.974 for rotation and translation, respectively. The results demonstrate that the proposed method effectively addresses the partial-to-full point cloud registration problem. The proposed method enables highly accurate 3D point cloud registration using partial observation, providing a critical foundation for precise surgical navigation and robotic interventions in computer-assisted surgery. The code will be accessed after the double-blind review process.
Chinese Translation
部分到完整的配准仍然面临挑战，原因包括重叠比率的变化、点密度的波动以及噪声的存在。尽管变换器在点云处理方面显示出强大的潜力，但之前的方法通常将其限制在全局上下文聚合，忽视了对于准确对应至关重要的细粒度局部几何信息。我们提出了 extit{GAPR-Net}，一种基于学习的点云配准框架，采用粗到细的架构，结合了卷积和变换器模块，通过交叉注意机制融合部分和完整点云之间的局部和全局信息。为此，提出了一种变换不变的逐点几何特征表示，能够稳健地捕捉相对于其邻近点的个别点的相对几何特征。为了评估所提方法的有效性，我们在四种几何特征各异的骨骼上进行了实验，包括胫骨、股骨、骨盆和胸部软骨。整体配准召回率达到94.2\%，该方法的均方根误差（RMSE）为1.992毫米，旋转和位移的$R^2$值分别为0.908和0.974。结果表明，所提方法有效解决了部分到完整点云配准问题。该方法使得通过部分观察实现高精度的3D点云配准成为可能，为计算机辅助外科中的精确手术导航和机器人干预提供了重要基础。代码将在双盲审稿过程后提供。
cs.CV / 60 / 2606.13496
Budget-Constrained Step-Level Diffusion Caching
预算约束的分步级扩散缓存
Abstract
Step-level caching accelerates diffusion models by exploiting temporal redundancy across denoising steps. Existing methods make per-step cache decisions using threshold-based heuristics, without directly optimizing for final output quality. As a result, their inference latency varies across inputs and is difficult to control at deployment. In this work, we propose BudCache, which inverts this formulation: rather than letting per-step error thresholds dictate the runtime cost, we fix the compute budget in advance and search for the cache policy that best preserves the final output. To tackle the combinatorial complexity of step selection, we combine Simulated Annealing with deterministic Hill Climbing. This offline search identifies high-quality cache policies within minutes and introduces no online search or thresholding overhead during inference. When the compute budget is very tight, we further introduce cache-aware schedule alignment, which adapts the time discretization to the selected cache policy to reduce cache-induced trajectory mismatch. Experiments on FLUX.1-dev and Wan2.1 show that BudCache achieves better generation quality than heuristic caching baselines under the same inference budgets. Code is available at https://github.com/Westlake-AGI-Lab/BudCache
Chinese Translation
分步级缓存通过利用去噪步骤中的时间冗余来加速扩散模型。现有方法使用基于阈值的启发式方法进行每一步的缓存决策，而没有直接优化最终输出质量。因此，它们的推理延迟在不同输入之间变化，并且在部署时难以控制。在本研究中，我们提出了BudCache，它颠覆了这种公式：与其让每一步的误差阈值决定运行成本，不如提前固定计算预算，并搜索最佳保留最终输出的缓存策略。为了应对步骤选择的组合复杂性，我们将模拟退火与确定性爬山算法相结合。这种离线搜索在几分钟内识别出高质量的缓存策略，并且在推理过程中不会引入在线搜索或阈值开销。当计算预算非常紧张时，我们进一步引入了缓存感知的调度对齐，它根据所选缓存策略调整时间离散化，以减少缓存引起的轨迹不匹配。在FLUX.1-dev和Wan2.1上的实验表明，BudCache在相同推理预算下实现了比启发式缓存基线更好的生成质量。代码可在 https://github.com/Westlake-AGI-Lab/BudCache 获取。
cs.CV / 61 / 2606.13503
Heterogeneous LiDAR Early Fusion and Learned Re-Ranking Strategy for Robust Long-Term Place Recognition in Unstructured Environments
异构LiDAR早期融合与学习重排序策略在非结构化环境中实现稳健的长期地点识别
Abstract
Robust localization in unstructured environments, such as agricultural fields, is a critical challenge for autonomous systems. LiDAR sensors provide detailed 3D information about the environment and are invariant to lighting conditions. For this reason, LiDAR-based place recognition methods have gained significant attention. In this paper, we propose MinkUNeXt-VINE++, a novel approach that combines early fusion of heterogeneous LiDAR data from two sensors (Livox Mid-360 and Velodyne VLP-16) and a learned re-ranking strategy in inference time. This fusion leverages the strengths of each sensor to provide a more comprehensive representation of the environment. Additionally, the re-ranking approach is particularly important in repetitive environments, such as vineyards, as finding true positives is a major challenge. We evaluated our approach using the TEMPO-VINE dataset, which provides heterogeneous LiDAR data in vineyard environments across different phenological stages. Our results demonstrate that MinkUNeXt-VINE++ significantly improves place recognition performance compared to single-sensor approaches and state-of-the-art methods. MinkUNeXt-VINE++ achieves a 20% improvement in the Recall@1 metric compared to single-sensor approaches, and +30% including re-ranking. The code of our method is publicly available for reproduction.
Chinese Translation
在非结构化环境（如农业田野）中实现稳健的定位是自主系统面临的一项关键挑战。LiDAR传感器提供关于环境的详细3D信息，并且对光照条件不敏感。因此，基于LiDAR的地点识别方法受到了广泛关注。在本文中，我们提出了MinkUNeXt-VINE++，这是一种新颖的方法，结合了来自两个传感器（Livox Mid-360和Velodyne VLP-16）的异构LiDAR数据的早期融合和推理时的学习重排序策略。这种融合利用了每个传感器的优势，提供了环境的更全面表示。此外，重排序方法在重复性环境（如葡萄园）中特别重要，因为找到真正的正例是一个主要挑战。我们使用TEMPO-VINE数据集评估了我们的方法，该数据集提供了不同物候阶段下的葡萄园环境中的异构LiDAR数据。我们的结果表明，MinkUNeXt-VINE++在地点识别性能上显著优于单传感器方法和最先进的方法。与单传感器方法相比，MinkUNeXt-VINE++在Recall@1指标上提高了20%，包括重排序时提高了30%。我们的方法代码已公开，供复现使用。
cs.CV / 62 / 2606.13509
Measurement-Calibrated Multi-Camera Fusion for Vision-Based Indoor Localization
基于测量校准的多摄像头融合用于视觉室内定位
Abstract
Indoor vision-based localization systems are affected by detection noise, occlusions, and limited camera coverage, leading to uncertainty at multiple stages of the pipeline. While multi-camera data fusion is widely used to mitigate these issues, it is typically treated as a black-box component and evaluated solely end-to-end, obscuring its mechanistic contributions. To address this gap, this work investigates whether explicitly characterizing single-camera localization errors can be leveraged to calibrate and optimize multi-camera data fusion. We introduce a measurement-calibrated fusion approach that integrates component-wise error quantification, specifically isolating homography calibration, human detection, and motion tracking. A component-wise evaluation is conducted to quantify error contributions from homography calibration, human detection, and motion tracking. Experimental results show that data fusion improves localization accuracy compared to single-camera baselines. While measurement-calibrated fusion provides only limited improvement in absolute accuracy over standard fusion, it substantially reduces trajectory variance and improves motion smoothness, which are critical for applications requiring stable and continuous motion estimates. These results highlight the value of explicit error characterization when designing data fusion strategies for vision-based indoor positioning systems.
Chinese Translation
基于视觉的室内定位系统受到检测噪声、遮挡和有限摄像头覆盖范围的影响，导致在多个阶段存在不确定性。尽管多摄像头数据融合被广泛应用于缓解这些问题，但通常将其视为黑箱组件，仅通过端到端的方式进行评估，从而掩盖了其机制性贡献。为了解决这一问题，本研究探讨了是否可以通过明确表征单摄像头定位误差来校准和优化多摄像头数据融合。我们提出了一种测量校准的融合方法，该方法集成了组件级误差量化，特别是孤立的单应性校准、人类检测和运动跟踪。我们进行了组件级评估，以量化来自单应性校准、人类检测和运动跟踪的误差贡献。实验结果表明，与单摄像头基线相比，数据融合提高了定位精度。尽管测量校准融合在绝对精度上对标准融合的改善有限，但它显著减少了轨迹方差并改善了运动平滑性，这对于需要稳定和连续运动估计的应用至关重要。这些结果突显了在设计基于视觉的室内定位系统的数据融合策略时，明确的误差表征的价值。
cs.CV / 63 / 2606.13515
MaskWAM: Unifying Mask Prompting and Prediction for World-Action Models
MaskWAM：统一掩码提示与预测的世界行动模型
Abstract
World Action Models (WAMs) present a promising paradigm for robotic control via video prediction. However, current WAMs suffer from fundamental spatial bottlenecks: standard text inputs introduce referential ambiguity in cluttered scenes, while unstructured RGB predictions lack semantic grounding and remain biased by task-irrelevant backgrounds. To overcome these limitations, we introduce MaskWAM, an object-centric world-action model. By jointly integrating masks as both explicit inputs and predictions via a unified Mixture of Transformers (MoT), MaskWAM unlocks robust policy generalization. This design provides two key benefits: (1) predicting future masks yields object-centric semantic supervision that suppresses visual noise, significantly enhancing even standard text-conditioned WAMs; and (2) coupling this predictive supervision with first-frame visual prompts, such as target object masks, establishes a precise spatial anchor that substantially reduces language ambiguity. Crucially, as WAMs are inherently vision-driven architectures, direct mask conditioning yields substantially stronger guidance than text alone, establishing a precise and robust paradigm for manipulating unseen objects. Evaluations on LIBERO, RoboTwin, and real-world tasks demonstrate that MaskWAM significantly outperforms baselines in both language-clear and language-ambiguous tasks.
Chinese Translation
世界行动模型（WAMs）为通过视频预测进行机器人控制提供了一种有前景的范式。然而，当前的WAMs存在基本的空间瓶颈：标准文本输入在杂乱场景中引入了指称模糊性，而非结构化的RGB预测缺乏语义基础，并且受到与任务无关的背景的偏见。为了解决这些限制，我们提出了MaskWAM，一种以对象为中心的世界行动模型。通过通过统一的混合变换器（Mixture of Transformers, MoT）将掩码作为显式输入和预测进行联合整合，MaskWAM解锁了强大的策略泛化。该设计提供了两个关键好处：（1）预测未来掩码产生以对象为中心的语义监督，抑制视觉噪声，显著增强甚至标准文本条件的WAMs；（2）将这种预测监督与第一帧视觉提示（如目标对象掩码）结合，建立了精确的空间锚点，显著减少了语言模糊性。至关重要的是，由于WAMs本质上是以视觉驱动的架构，直接的掩码条件提供的指导远强于单独的文本，为操控未见对象建立了精确而强大的范式。在LIBERO、RoboTwin和现实世界任务上的评估表明，MaskWAM在语言清晰和语言模糊任务中显著优于基线。
cs.CV / 64 / 2606.13528
What's Old is New Again: Classical Dimensionality Reduction for Efficient Saliency-Guided Biometric Attack Detection
旧的又是新的：用于高效显著性引导生物识别攻击检测的经典降维方法
Abstract
Saliency-guided training is a paradigm in visual recognition that encourages models to focus on the most relevant image regions during learning. While its application in biometric presentation attack detection (PAD) has shown strong benefits in robustness and generalization, adoption is often limited by the high cost, domain specificity, and limited scalability of existing saliency acquisition methods, such as human annotations over a limited dataset. We present a novel, cost-efficient, and highly-scalable approach to saliency acquisition using maps inspired by classical dimensionality reduction techniques: PCA and LDA. Our proposed methods generate saliency maps directly from raw training data, requiring no human annotation nor domain knowledge. We contextualize the effectiveness of these saliency sources in three saliency-explored domains (iris PAD, synthetic face detection, fingerprint PAD) and demonstrate its scalability in two saliency-novel domains (fingerprint vein PAD and ID card PAD). Across all domains tested, models trained using dimensionality reduction-sourced saliency maps exceed baseline and sometimes SOTA saliency methods without any resource investment or domain-specific tooling. Our findings overcome an important yet unaddressed barrier to saliency-guided training for biometric attack detection and beyond.
Chinese Translation
显著性引导训练是一种视觉识别范式，鼓励模型在学习过程中关注最相关的图像区域。尽管其在生物识别呈现攻击检测（PAD）中的应用已显示出在鲁棒性和泛化能力方面的显著优势，但由于现有显著性获取方法（如对有限数据集的人类标注）成本高、领域特定性强和可扩展性有限，导致其采用受到限制。我们提出了一种新颖、成本效益高且高度可扩展的显著性获取方法，使用受经典降维技术（主成分分析 PCA 和线性判别分析 LDA）启发的图谱。我们提出的方法直接从原始训练数据生成显著性图，无需人类标注或领域知识。我们在三个显著性探索领域（虹膜 PAD、合成面部检测、指纹 PAD）中对这些显著性来源的有效性进行了背景分析，并在两个显著性新领域（指纹静脉 PAD 和身份证 PAD）中展示了其可扩展性。在所有测试领域中，使用降维源显著性图训练的模型超越了基线，有时甚至超过了最先进的显著性方法，而无需任何资源投入或领域特定工具。我们的研究结果克服了显著性引导训练在生物识别攻击检测及其他领域中的一个重要但尚未解决的障碍。
cs.CV / 65 / 2606.13558
Edit the Bits, Diff the Codes: Bitwise Residual Editing for Visual Autoregressive Models
编辑比特，差异编码：视觉自回归模型的比特残差编辑
Abstract
Text-guided image editing with visual autoregressive (VAR) generators requires controlling both what the model samples and where the sampled change is written back into the image code. Existing VAR editors mainly operate on token streams, features, or flat next-token logits, leaving two native structures of bitwise-residual VAR models underused: the per-bit Bernoulli prediction head and the additive multi-scale residual code field from which the image is assembled. We propose BitResEdit, a training-free editor for bitwise-residual VAR generators such as Infinity. BitEdit performs source-negative guidance by tilting the post-CFG per-bit log-odds along a source--target contrast computed on a shared edited prefix, then projects each update into a closed-form Bernoulli-KL trust region around the clean CFG sampler. ResEdit converts the sampled bits into per-scale continuous-code residuals, gates them with a localization mask, and re-injects them through the generator's native sum-of-scales. Together they couple decision-time bit guidance with combination-time code composition, so masked-out latent features are preserved exactly by code arithmetic while localized, scale-aware edits are applied inside the target region. On PIE-Bench with Infinity-2B, BitResEdit attains the strongest text alignment among same-backbone VAR editors, improving CLIP on the edited region by +1.07 over the strongest prior editor while keeping background preservation competitive with it. Ablations show BitEdit and ResEdit play complementary roles in target alignment and background preservation.
Chinese Translation
基于文本引导的图像编辑使用视觉自回归（VAR）生成器，需要控制模型采样的内容以及采样的变化写入图像编码的位置。现有的VAR编辑器主要在标记流、特征或平坦的下一个标记对数上操作，导致比特残差VAR模型的两个原生结构未被充分利用：逐比特的伯努利预测头和用于组装图像的加法多尺度残差编码场。我们提出了BitResEdit，一种针对比特残差VAR生成器（如Infinity）的无训练编辑器。BitEdit通过沿着在共享编辑前缀上计算的源-目标对比，倾斜后CFG逐比特对数来执行源负引导，然后将每次更新投影到围绕干净CFG采样器的封闭形式伯努利-KL信任区域。ResEdit将采样的比特转换为逐尺度的连续编码残差，使用定位掩码对其进行门控，并通过生成器的原生尺度和求和重新注入它们。它们共同将决策时的比特引导与组合时的编码组合相结合，因此被掩蔽的潜在特征通过编码算术被精确保留，同时在目标区域内应用定位的、尺度感知的编辑。在使用Infinity-2B的PIE-Bench上，BitResEdit在同一骨干VAR编辑器中实现了最强的文本对齐，相较于最强的先前编辑器，在编辑区域上提高了+1.07，同时保持背景保留的竞争力。消融实验表明，BitEdit和ResEdit在目标对齐和背景保留中发挥了互补作用。
cs.CV / 66 / 2606.13562
Contrast-Informed Augmentation and Domain-Adversarial Training for Adult-to-Neonatal MR Reconstruction Generalization
基于对比信息的增强和领域对抗训练在成人到新生儿磁共振重建泛化中的应用
Abstract
Purpose: To investigate whether contrast-informed data augmentation and domain-adversarial training improve the adult-to-neonatal generalization of the E2E-VarNet. Methods: Three training regimes were investigated: (1) adult-only training with unaugmented adult data, (2) mixed training with paired unaugmented and neonatal-informed augmented adult data, and (3) mixed training with a domain-adversarial objective. Models were trained on retrospectively undersampled multi-coil adult T2-weighted brain MR data and evaluated on neonatal and adult test data at acceleration factors $R=4$ and $R=8$ using quantitative metrics and qualitative evaluation. Feature analyses assessed whether domain-adversarial training altered the latent representations of unaugmented adult, augmented adult, and neonatal test samples. Results: Mixed training (Mixed) and mixed domain-adversarial training (Mixed-DAT) outperformed unaugmented adult-only training (Unaug-Only) when evaluated on neonatal data. At R=4, Mixed-DAT achieved the best performance (SSIM = 0.924 +/- 0.027, PSNR = 33.98 +/- 1.15 dB). At R=8, Mixed-DAT performed best when measured using SSIM (0.848 +/- 0.031 vs. 0.766 +/- 0.037 for Unaug-Only and 0.814 +/- 0.035 for Mixed) and Mixed performed best when measured using PSNR (29.56 +/- 0.83 dB vs. 26.26 +/- 0.78 dB for Unaug-Only and 29.43 +/- 0.83 dB for Mixed-DAT). Qualitative assessment of t-SNE plots suggested that Mixed-DAT increased the overlap among the latent representations of the unaugmented adult, augmented adult, and neonatal test data. Conclusion: Contrast-informed augmentation and domain-adversarial training improved adult-to-neonatal generalization of deep learning-based MR reconstruction. These findings suggest that contrast-informed data augmentation combined with adversarial training may improve robustness to domain shift in undersampled neonatal MR reconstruction.
Chinese Translation
目的：研究基于对比信息的数据增强和领域对抗训练是否能改善E2E-VarNet在成人到新生儿的泛化能力。方法：研究了三种训练方案：（1）仅使用未增强的成人数据进行成人训练；（2）混合训练，使用配对的未增强成人数据和新生儿信息增强的成人数据；（3）混合训练，结合领域对抗目标。模型在回顾性欠采样的多线圈成人T2加权脑磁共振数据上进行训练，并在加速因子$R=4$和$R=8$下，使用定量指标和定性评估对新生儿和成人测试数据进行评估。特征分析评估领域对抗训练是否改变了未增强成人、增强成人和新生儿测试样本的潜在表示。结果：在新生儿数据评估中，混合训练（Mixed）和混合领域对抗训练（Mixed-DAT）优于仅使用未增强成人数据的训练（Unaug-Only）。在$R=4$时，Mixed-DAT表现最佳（SSIM = 0.924 +/- 0.027，PSNR = 33.98 +/- 1.15 dB）。在$R=8$时，Mixed-DAT在SSIM指标上表现最佳（0.848 +/- 0.031，相较于Unaug-Only的0.766 +/- 0.037和Mixed的0.814 +/- 0.035），而Mixed在PSNR指标上表现最佳（29.56 +/- 0.83 dB，相较于Unaug-Only的26.26 +/- 0.78 dB和Mixed-DAT的29.43 +/- 0.83 dB）。t-SNE图的定性评估表明，Mixed-DAT增加了未增强成人、增强成人和新生儿测试数据的潜在表示之间的重叠。结论：基于对比信息的增强和领域对抗训练改善了深度学习磁共振重建在成人到新生儿的泛化能力。这些发现表明，结合对抗训练的基于对比信息的数据增强可能提高在欠采样新生儿磁共振重建中的领域转移鲁棒性。
cs.CV / 67 / 2606.13580
EvTexture++: Event-Driven Texture Enhancement for Video Super-Resolution
EvTexture++：用于视频超分辨率的事件驱动纹理增强
Abstract
Event-based vision has drawn increasing attention owing to its distinctive properties, including ultra-high temporal resolution and extreme dynamic range. Recent works have introduced it to video super-resolution (VSR) to enhance flow estimation and temporal alignment. In contrast, this paper shifts the focus of event signals from motion refinement to texture enhancement in VSR. We propose EvTexture++, the first event-driven framework dedicated to texture enhancement in VSR. It leverages high-frequency spatiotemporal details from events to improve texture recovery. EvTexture++ incorporates a customized texture enhancement branch, along with an iterative texture enhancement module that progressively exploits high-temporal-resolution event information for texture restoration. This enables gradual refinement of texture regions across iterations, yielding more accurate and detailed high-resolution outputs. Besides intra-frame texture recovery, large motions could degrade inter-frame temporal consistency, particularly in texture regions, leading to texture flickering. To mitigate this, we further exploit the continuous-time motion cues of events to enhance temporal consistency, introducing a temporal texture alignment module that estimates event-guided texture-aware flow for precise inter-frame texture alignment. Moreover, EvTexture++ is designed as a plug-and-play tool to flexibly boost the performance of existing VSR models. Experiments on five datasets demonstrate that EvTexture++ achieves state-of-the-art performance. When integrated into recent VSR models, it yields significant improvements, with gains of up to 1.55 dB in PSNR on the texture-rich Vid4 dataset. Code: https://github.com/DachunKai/EvTexture.
Chinese Translation
基于事件的视觉因其独特的特性，如超高的时间分辨率和极大的动态范围，受到越来越多的关注。近期的研究将其引入视频超分辨率（VSR），以增强光流估计和时间对齐。与此不同，本文将事件信号的重点从运动细化转向VSR中的纹理增强。我们提出了EvTexture++，这是第一个专门用于VSR中纹理增强的事件驱动框架。它利用事件中的高频时空细节来改善纹理恢复。EvTexture++结合了一个定制的纹理增强分支，以及一个迭代纹理增强模块，逐步利用高时间分辨率的事件信息进行纹理恢复。这使得在迭代过程中逐渐细化纹理区域，从而产生更准确和详细的高分辨率输出。除了帧内纹理恢复外，大运动可能会降低帧间的时间一致性，特别是在纹理区域，导致纹理闪烁。为此，我们进一步利用事件的连续时间运动线索来增强时间一致性，引入一个时间纹理对齐模块，该模块估计事件引导的纹理感知光流，以实现精确的帧间纹理对齐。此外，EvTexture++被设计为一个即插即用的工具，以灵活提升现有VSR模型的性能。在五个数据集上的实验表明，EvTexture++达到了最先进的性能。当集成到近期的VSR模型中时，它带来了显著的改进，在纹理丰富的Vid4数据集上，PSNR提升高达1.55 dB。代码链接： https://github.com/DachunKai/EvTexture.
cs.CV / 68 / 2606.13587
Towards Effective Waste Segmentation for Automated Waste Recycling in Cluttered Background
面向杂乱背景下自动化垃圾回收的有效垃圾分割
Abstract
Rapid expansion of urban areas and population growth is causing an immense increase in waste production, which demands the need for efficient and automated waste management. In this scenario, automated waste recycling (AWR) using deep learning methods can assist humans in optimal waste management. Recent deep learning approaches for AWR provide promising waste segmentation performance, however, these methods rely on large backbone networks that are inefficient for AWR systems and suffer from performance deterioration in cluttered scenes. To this end, an optimal waste segmentation network is introduced which effectively utilizes the spatial domain to capture localized structural dependencies and the spectral domain to efficiently extract global contextual relationships. This cascaded design allows the network to progressively leverage both local and global representations across complementary domains to highlight the semantic information necessary for effective segmentation of various waste objects. Furthermore, auxiliary feature enhancement module (AFEM) is introduced to enhance the target objects' boundaries and blob amplification for better segmentation in cluttered scenarios. Extensive experimentation on ZeroWaste-aug, ZeroWaste-f and SpectralWaste datasets reveals the merits of the proposed method.
Chinese Translation
城市地区的快速扩张和人口增长导致垃圾产生量急剧增加，这对高效和自动化的垃圾管理提出了需求。在这种情况下，利用深度学习方法的自动化垃圾回收（AWR）可以帮助人类实现最佳的垃圾管理。最近的深度学习方法在AWR方面提供了令人鼓舞的垃圾分割性能，然而，这些方法依赖于大型骨干网络，这对AWR系统效率低下，并且在杂乱场景中表现不佳。为此，提出了一种优化的垃圾分割网络，该网络有效利用空间域捕捉局部结构依赖关系，并利用光谱域高效提取全局上下文关系。这种级联设计使网络能够逐步利用互补域中的局部和全局表示，以突出有效分割各种垃圾物体所需的语义信息。此外，引入辅助特征增强模块（AFEM）以增强目标物体的边界和斑点放大，从而在杂乱场景中实现更好的分割。在ZeroWaste-aug、ZeroWaste-f和SpectralWaste数据集上的广泛实验表明了所提方法的优越性。
cs.CV / 69 / 2606.13625
Revisiting Vehicle Color Recognition in Long-Tailed Surveillance Scenarios
重新审视长尾监控场景中的车辆颜色识别
Abstract
Vehicle color recognition is an important cue for vehicle identification in surveillance systems, especially when license plates are illegible due to low resolution, occlusion, motion blur, or poor illumination. However, real-world vehicle color distributions are highly imbalanced, making overall accuracy insufficient to assess performance on rare but operationally relevant colors. This paper presents a comprehensive study of vehicle color recognition under severe class imbalance using UFPR-VeSV, a challenging real-world surveillance dataset. We investigate synthetic minority-class augmentation through two off-the-shelf generative strategies: text-conditioned image generation with RunDiffusion/JuggernautXL and image-conditioned color editing with Gemini 2.0 Flash. The curated synthetic data are combined with modern visual representations, loss reweighting, learning-rate scheduling, color-safe augmentation, foreground-aware preprocessing, and ensemble fusion. The bestperforming approach achieves 94.6% micro accuracy and 79.7% macro accuracy, improving macro accuracy by 8.2 percentage points over recent literature. A manual error analysis further shows that many remaining failures are visually ambiguous even for human annotators, highlighting the practical limits of color-based vehicle identification in unconstrained surveillance imagery. The generated images and source code are publicly available at https://github.com/viniciusorru/vcr-synthetic
Chinese Translation
车辆颜色识别是监控系统中车辆识别的重要线索，尤其是在由于低分辨率、遮挡、运动模糊或光照不足导致车牌难以辨认的情况下。然而，现实世界中的车辆颜色分布高度不平衡，使得整体准确性不足以评估在稀有但具有操作相关性的颜色上的表现。本文对在严重类别不平衡情况下的车辆颜色识别进行了全面研究，使用了UFPR-VeSV这一具有挑战性的现实世界监控数据集。我们通过两种现成的生成策略研究了合成少数类的增强：使用RunDiffusion/JuggernautXL进行文本条件的图像生成，以及使用Gemini 2.0 Flash进行图像条件的颜色编辑。经过精心策划的合成数据与现代视觉表示、损失重加权、学习率调度、颜色安全增强、前景感知预处理和集成融合相结合。表现最佳的方法达到了94.6%的微观准确率和79.7%的宏观准确率，相较于近期文献，宏观准确率提高了8.2个百分点。手动错误分析进一步表明，许多剩余的失败在视觉上对人类注释者而言也是模糊的，突显了基于颜色的车辆识别在不受约束的监控图像中的实际局限性。生成的图像和源代码已公开可用，网址为https://github.com/viniciusorru/vcr-synthetic
cs.CV / 70 / 2606.13644
Surflo: Consistent 3D Surface Flow Model with Global State
Surflo：具有全局状态的一致性三维表面流模型
Abstract
Geometry is invariant to viewpoint, which makes any collection of images a redundant encoding of a single 3D state. Existing feed-forward reconstruction models fail to exploit this: per-view methods emit overlapping, unaligned pointmaps that grow linearly with input count, while global-latent methods commit to a fixed, low-resolution output. We introduce Surflo, which compresses a variable number of unposed RGB views into K latent tokens-one global state-and decodes oriented 3D surface points by independently transporting them from noise onto the surface via flow matching. This frees the output from any fixed grid or token budget: the same latent yields from a few thousand to a million points in a single forward pass. To suppress the local inconsistencies inherent to independent per-point decoding, an inference-time guidance term correlates nearby points by injecting a photometric gradient during ODE integration. Surflo matches or surpasses feed-forward baselines on surface metrics, runs an order of magnitude faster than optimization-based methods that require hundreds of views, and is the only feed-forward approach to combine a global latent with arbitrary-resolution decoding.
Chinese Translation
几何形状对视角是不变的，这使得任何图像集合都是对单一三维状态的冗余编码。现有的前馈重建模型未能利用这一点：每视角方法产生重叠且未对齐的点图，且随着输入数量线性增长，而全局潜在方法则固定于低分辨率输出。我们提出了Surflo，它将可变数量的未定姿态RGB视图压缩为K个潜在标记——一个全局状态——并通过独立地将其从噪声传输到表面上，利用流匹配解码定向的三维表面点。这使得输出不受任何固定网格或标记预算的限制：相同的潜在标记在一次前向传递中可以产生从几千到一百万个点。为了抑制独立逐点解码固有的局部不一致性，在推理时引入的引导项通过在常微分方程（ODE）积分过程中注入光度梯度来关联附近的点。Surflo在表面度量上与前馈基线相匹配或超越，运行速度比需要数百个视图的基于优化的方法快一个数量级，并且是唯一将全局潜在与任意分辨率解码相结合的前馈方法。
cs.CV / 71 / 2606.13652
World Tracing: Generative Pixel-Aligned Geometry Beyond the Visible
世界追踪：超越可见范围的生成像素对齐几何
Abstract
Image-to-3D methods often trade off faithfulness and completeness: depth estimators are anchored to input pixels but stop at the visible surface, while image-to-3D models generate complete shapes that are often misaligned with the input. We introduce World Tracing, a generative pixel-aligned geometry representation that predicts 3D points aligned with observed pixels while completing geometry beyond the visible surface. For each input pixel, World Tracing predicts an ordered stack of camera-space 3D points, where the first layer represents the visible surface and subsequent layers represent front-to-back intersections with occluded surfaces. We instantiate this representation with a world-tracing diffusion transformer, WT-DiT, which treats multiple geometry layers as separate denoising tokens coupled through factorized and global attention. WT-DiT is trained with pixel-space flow matching and a mixed noise schedule that balances visible-surface reconstruction with occluded-geometry generation. World Tracing achieves strong performance on visible-surface reconstruction and complete geometry generation across object, scene, and dynamic benchmarks, outperforming both depth predictors and image-to-3D generators. It also preserves 2D-to-3D correspondence, enabling text-driven 3D scene editing, geometry-conditioned novel-view video synthesis, and training-free integration with textured-mesh generators.
Chinese Translation
图像到三维的方法通常在真实感和完整性之间进行权衡：深度估计器锚定于输入像素，但仅停留在可见表面，而图像到三维模型生成的完整形状往往与输入不对齐。我们提出了世界追踪（World Tracing），一种生成的像素对齐几何表示，能够预测与观察到的像素对齐的三维点，同时在可见表面之外完成几何形状。对于每个输入像素，世界追踪预测一组有序的相机空间三维点，其中第一层表示可见表面，后续层表示与被遮挡表面的前后交点。我们使用世界追踪扩散变换器（WT-DiT）实例化这一表示，该变换器将多个几何层视为通过因子化和全局注意力耦合的独立去噪令牌。WT-DiT通过像素空间流匹配和混合噪声调度进行训练，以平衡可见表面重建与被遮挡几何生成。世界追踪在可见表面重建和完整几何生成方面在对象、场景和动态基准测试中表现出色，超越了深度预测器和图像到三维生成器。它还保持了二维到三维的对应关系，使得基于文本的三维场景编辑、几何条件的新视角视频合成以及与纹理网格生成器的无训练集成成为可能。
cs.CV / 72 / 2606.13655
Flex4DHuman: Flexible Multi-view Video Diffusion for 4D Human Reconstruction
Flex4DHuman：用于4D人类重建的灵活多视角视频扩散
Abstract
We present Flex4DHuman, a multi-view video diffusion model that transforms a monocular or sparse multi-view video of a dynamic subject into synchronized dense multi-view videos using only relative camera-pose conditioning. Unlike prior human-centric methods that rely on skeletons, depth maps, normals, or rendered target-view geometry, Flex4DHuman requires no explicit geometry priors and instead conditions generation through relative camera-pose positional encoding. The generated videos can be directly ingested by downstream reconstruction pipelines to create dynamic 4D Gaussian splats. Built on the Wan 2.1 1.3B text-to-video model, Flex4DHuman preserves the backbone architecture and encodes camera and view information through a five-axis positional encoding that extends spatio-temporal RoPE with view indices and continuous SE(3) relative camera geometry. A three-stage curriculum progressively trains the model for pose following, flexible reference-to-target view generation, and temporal rollout. To support temporal rollout, we train with clean historical target-view tokens. We also add multi-view captions to enable test-time text control. Combined with an off-the-shelf 4D Gaussian Splatting stage, our framework lifts monocular static-camera videos into dynamic 4D Gaussian splats. Experiments on DNA-Rendering and ActorsHQ show that Flex4DHuman surpasses prior state-of-the-art methods, while the same formulation generalizes to animal categories after mixed human-animal training. These capabilities make Flex4DHuman a practical step toward scalable 4D content creation from casual monocular videos for simulation, gaming, AR/VR, and video re-shooting.
Chinese Translation
我们提出了Flex4DHuman，这是一种多视角视频扩散模型，它能够将动态主体的单目或稀疏多视角视频转化为同步的密集多视角视频，仅通过相对相机姿态条件进行生成。与以往依赖骨架、深度图、法线或渲染目标视图几何的以人为中心的方法不同，Flex4DHuman不需要显式的几何先验，而是通过相对相机姿态位置编码来条件生成。生成的视频可以直接被下游重建管道使用，以创建动态的4D高斯点云。Flex4DHuman基于Wan 2.1 1.3B文本到视频模型，保留了主干架构，并通过五轴位置编码对相机和视图信息进行编码，该编码扩展了时空RoPE，结合了视图索引和连续的SE(3)相对相机几何。我们采用三阶段的课程逐步训练模型，以实现姿态跟踪、灵活的参考到目标视图生成和时间展开。为了支持时间展开，我们使用干净的历史目标视图标记进行训练。我们还添加了多视角字幕，以便在测试时进行文本控制。结合现成的4D高斯点云阶段，我们的框架将单目静态相机视频提升为动态4D高斯点云。在DNA-Rendering和ActorsHQ上的实验表明，Flex4DHuman超越了之前的最先进方法，而相同的公式在混合人类-动物训练后也能推广到动物类别。这些能力使得Flex4DHuman成为从普通单目视频中可扩展4D内容创作的实际步骤，适用于模拟、游戏、增强现实/虚拟现实和视频重拍。
cs.CV / 73 / 2606.13673
SpatialClaw: Rethinking Action Interface for Agentic Spatial Reasoning
SpatialClaw：重新思考代理空间推理的动作接口
Abstract
Spatial reasoning, the ability to determine where objects are, how they relate, and how they move in 3D, remains a fundamental challenge for vision-language models (VLMs). Tool-augmented agents attempt to address this by augmenting VLMs with specialist perception modules, yet their effectiveness is bounded by the action interface through which those tools are invoked. In this work, we study how the design of this interface shapes the agent's capacity for open-ended spatial reasoning. Existing spatial agents either employ single-pass code execution, which commits to a full analysis strategy before any intermediate result is observed, or rely on a structured tool-call interface that often offers less flexibility for freely composing operations or tailoring the analysis to each task. Both designs offer limited flexibility for open-ended, complex 3D/4D spatial reasoning. We therefore propose SpatialClaw, a training-free framework for spatial reasoning that adopts code as the action interface. SpatialClaw maintains a stateful Python kernel pre-loaded with input frames and a suite of perception and geometry primitives, letting a VLM-backed agent write one executable cell per step conditioned on all prior outputs, enabling the agent to flexibly compose and manipulate perception results and adapt its analysis to both intermediate text and visual observations and the demands of each problem. Evaluated across 20 spatial reasoning benchmarks spanning a broad range of static and dynamic 3D/4D spatial reasoning tasks, SpatialClaw achieves 59.9% average accuracy, outperforming the recent spatial agent by +11.2 points, with consistent gains across six VLM backbones from two model families without any benchmark- or model-specific adaptation.
Chinese Translation
空间推理，即确定物体的位置、它们之间的关系以及它们在三维空间中的运动能力，仍然是视觉-语言模型（VLMs）面临的基本挑战。工具增强代理试图通过将VLM与专业感知模块结合来解决这一问题，但它们的有效性受到调用这些工具的动作接口的限制。在本研究中，我们探讨了该接口的设计如何影响代理进行开放式空间推理的能力。现有的空间代理要么采用单次代码执行，这在观察到任何中间结果之前就承诺于完整的分析策略，要么依赖于结构化的工具调用接口，这往往在自由组合操作或根据每个任务调整分析时提供较少的灵活性。这两种设计在开放式、复杂的三维/四维空间推理中都提供了有限的灵活性。因此，我们提出了SpatialClaw，这是一个无训练的空间推理框架，采用代码作为动作接口。SpatialClaw维护一个状态保持的Python内核，预加载输入帧和一套感知与几何原语，使得基于VLM的代理能够在每一步根据所有先前输出编写一个可执行单元，从而灵活组合和操控感知结果，并将其分析适应于中间文本和视觉观察以及每个问题的需求。在涵盖广泛静态和动态三维/四维空间推理任务的20个空间推理基准测试中，SpatialClaw实现了59.9%的平均准确率，超越了最近的空间代理11.2个百分点，并在两个模型系列的六个VLM骨干网络中保持了一致的提升，而无需任何基准或模型特定的适应。
cs.CV / 74 / 2606.13674
RepWAM: World Action Modeling with Representation Visual-Action Tokenizers
RepWAM：基于表示视觉-动作标记器的世界动作建模
Abstract
This work presents RepWAM, a representation-centric world action model (WAM) built on representation visual-action tokenizers. Existing WAMs typically inherit reconstruction-oriented video tokenizers from pretrained video generation models. Although these tokenizers preserve visual fidelity, pixel reconstruction alone provides limited guidance for learning instruction-following dynamics that connect future prediction with robot control. To address this, we explore a semantic visual-action latent space for representation-centric world action modeling. Specifically, we train a representation visual-action tokenizer that maps visual inputs into aligned visual and latent action tokens. We then pretrain our WAM to jointly model future visual states and the latent actions that connect them under language instructions, followed by adaptation to real robot trajectories for closed-loop manipulation. Experiments on real-world manipulation tasks and simulation benchmarks show that RepWAM delivers strong performance across diverse manipulation settings, while ablations highlight the value of semantic visual-action tokenization over reconstruction-oriented alternatives. These results establish representation visual-action tokenization as a promising foundation for world action models and a step toward generalist robot policies. Code and weights will be available at https://github.com/wdrink/RepWAM.
Chinese Translation
本文提出了RepWAM，一种以表示为中心的世界动作模型（WAM），基于表示视觉-动作标记器构建。现有的WAM通常继承自预训练视频生成模型的重建导向视频标记器。尽管这些标记器保留了视觉保真度，但仅依靠像素重建对于学习连接未来预测与机器人控制的指令跟随动态提供的指导有限。为了解决这个问题，我们探索了一种语义视觉-动作潜在空间，以实现以表示为中心的世界动作建模。具体而言，我们训练了一个表示视觉-动作标记器，将视觉输入映射到对齐的视觉和潜在动作标记。然后，我们预训练我们的WAM，以在语言指令下联合建模未来视觉状态和连接它们的潜在动作，随后适应于真实机器人轨迹以进行闭环操作。在真实世界操作任务和仿真基准上的实验表明，RepWAM在多样化的操作设置中表现出色，而消融实验则突显了语义视觉-动作标记化相较于重建导向替代方案的价值。这些结果确立了表示视觉-动作标记化作为世界动作模型的有前景基础，并朝着通用机器人策略迈出了重要一步。代码和权重将发布在 https://github.com/wdrink/RepWAM。
cs.CV / 75 / 2606.13676
Modality Forcing for Scalable Spatial Generation
可扩展空间生成的模态强制
Abstract
Text-to-image (T2I) models contain rich spatial priors. Synthesizing photorealistic, cluttered scenes requires an understanding of geometry, including perspective and relative scale. Prior works adapt T2I models to leverage this prior for depth prediction, but they require dense depth data and involve complex recipes. We propose Modality Forcing, a simple, scalable post-training recipe for joint image-depth generation using a single DiT trained on sparse depth data. Modality Forcing enables conditional and joint generation of image and depth in any permutation by assigning separate noise levels per modality. Per-modality decoders let us train on sparse, real-world depth and achieve strong, generalizable depth prediction. We further show that Modality Forcing inherits the scalability of T2I pre-training: by training a set of T2I models from scratch (370M to 3.3B parameters), we find that larger models trained on more image data produce more accurate depth. Our strongest model is competitive with state-of-the-art monocular depth estimators and reduces AbsRel by 57% relative to existing joint image-depth generative models. These results provide strong evidence that image generation is a scalable pre-training objective for spatial perception. https://modality-forcing.github.io/
Chinese Translation
文本到图像（T2I）模型包含丰富的空间先验。合成逼真且杂乱的场景需要对几何形状的理解，包括透视和相对尺度。之前的研究将T2I模型调整为利用这一先验进行深度预测，但它们需要密集的深度数据并涉及复杂的过程。我们提出了模态强制（Modality Forcing），这是一种简单、可扩展的后训练方案，用于使用在稀疏深度数据上训练的单个DiT模型进行图像和深度的联合生成。模态强制通过为每种模态分配单独的噪声水平，使得图像和深度的条件和联合生成可以以任意排列进行。每种模态的解码器使我们能够在稀疏的真实世界深度上进行训练，并实现强大且具有良好泛化能力的深度预测。我们进一步展示了模态强制继承了T2I预训练的可扩展性：通过从头开始训练一组T2I模型（370M到3.3B参数），我们发现，基于更多图像数据训练的更大模型能够产生更准确的深度。我们最强的模型在与最先进的单目深度估计器的竞争中表现出色，并相对于现有的联合图像-深度生成模型将绝对相对误差（AbsRel）降低了57%。这些结果强有力地证明了图像生成是空间感知的可扩展预训练目标。
cs.CV / 76 / 2606.13679
InterleaveThinker: Reinforcing Agentic Interleaved Generation
InterleaveThinker：增强代理交错生成
Abstract
Recent image generators have demonstrated impressive photorealism and instruction-following capabilities in single-image generation and editing. However, constrained by their architectures, they cannot achieve interleaved generation (text-image sequence), which has crucial applications in visual narratives, guidance, and embodied manipulation. Even the latest open-source Unified Multimodal Models (UMMs) exhibit limited performance in this regard. In this paper, we introduce InterleaveThinker, the first multi-agent pipeline designed to endow any existing image generator with interleaved generation capabilities. Specifically, we employ a planner agent to organize the image-text input sequence, instructing the image generator on the required execution at each step. Subsequently, we introduce a critic agent to evaluate the generator's outputs, identify samples that deviate from the planned instructions, and refine the instructions for regeneration. To implement this pipeline, we construct the Interleave-Planner-SFT-80k and Interleave-Critic-SFT-112k to perform a format cold-start. Then we develop Interleave-Critic-RL-13k to reinforce the step-wise instruction correction capability within a generation trajectory using GRPO. Since a single interleaved generation trajectory may involve over 25 generator calls, optimizing the entire trajectory is computationally impractical. Therefore, we propose accuracy reward and step-wise reward, allowing single-step RL to effectively guide the entire generation trajectory. The results show that InterleaveThinker improves performance across various image generators. On interleaved generation benchmarks, it achieves performance comparable to Nano Banana and GPT-5. Surprisingly, it also significantly enhances the base model on reasoning-based benchmarks; for example, on 4-step FLUX.2-klein, we observe substantial gains on WISE and RISE.
Chinese Translation
近期的图像生成器在单幅图像生成和编辑方面展示了令人印象深刻的照片真实感和遵循指令的能力。然而，由于其架构的限制，它们无法实现交错生成（文本-图像序列），而这在视觉叙事、指导和具身操作中具有重要应用。即使是最新的开源统一多模态模型（Unified Multimodal Models, UMMs）在这方面的表现也有限。在本文中，我们介绍了InterleaveThinker，这是第一个旨在赋予任何现有图像生成器交错生成能力的多代理管道。具体而言，我们采用一个规划代理来组织图像-文本输入序列，指导图像生成器在每一步所需的执行。随后，我们引入一个评估代理来评估生成器的输出，识别偏离计划指令的样本，并优化指令以进行再生成。为了实现这一管道，我们构建了Interleave-Planner-SFT-80k和Interleave-Critic-SFT-112k以执行格式冷启动。然后，我们开发了Interleave-Critic-RL-13k，以使用GRPO增强生成轨迹中的逐步指令修正能力。由于单个交错生成轨迹可能涉及超过25次生成器调用，因此优化整个轨迹在计算上是不切实际的。因此，我们提出了准确性奖励和逐步奖励，使单步强化学习能够有效指导整个生成轨迹。结果表明，InterleaveThinker在各种图像生成器上提高了性能。在交错生成基准测试中，它的表现可与Nano Banana和GPT-5相媲美。令人惊讶的是，它还显著提升了基础模型在基于推理的基准测试中的表现；例如，在4步FLUX.2-klein上，我们观察到在WISE和RISE上有显著的提升。
cs.AI / 1 / 2606.12451
ToolSense: A Diagnostic Framework for Auditing Parametric Tool Knowledge in LLMs
ToolSense：用于审计大语言模型中参数化工具知识的诊断框架
Abstract
Large language models deployed as agents over large tool catalogs face a critical tool-retrieval bottleneck. As embedding-based retrieval approaches rely on compact encoders that may under-capture specialized tool semantics, parametric tool retrieval addresses this by encoding each tool as a virtual token appended to the LLM vocabulary, fine-tuned in two stages (memorization then retrieval SFT) to use the LLM as a retriever, achieving strong performance on standard ToolBench retrieval benchmarks. Yet these benchmarks use verbose, fully-specified queries, and their evaluation applies constrained decoding that restricts outputs to valid token paths, neither reveals whether the model actually understands its tools. We introduce \textbf{ToolSense}, an open-source LLM-powered diagnostic framework that takes any tool catalog as input and automatically generates three benchmarks: a Realistic Retrieval Benchmark (RRB) with queries at three ambiguity tiers, an MCQ probing benchmark, and a QA probing benchmark. Applying ToolSense to ToolBench (~47k tools) and evaluating five parametric model training configurations reveals a knowledge-retrieval dissociation: on RRB queries, several configurations collapse by ~50-64 percentage points compared to fully-specified ToolBench benchmarks, falling below the embedding-model baseline. Additionally, despite strong retrieval performance, some models score near-random on factual probes, suggesting a knowledge-retrieval dissociation. We open-source the ToolSense framework and the ToolBench diagnostic benchmarks at https://github.com/SAP/toolsense.
Chinese Translation
作为代理部署在大型工具目录上的大语言模型面临着关键的工具检索瓶颈。由于基于嵌入的检索方法依赖于紧凑的编码器，这可能无法充分捕捉专业工具的语义，参数化工具检索通过将每个工具编码为附加到大语言模型（LLM）词汇表的虚拟标记来解决这一问题，经过两个阶段的微调（记忆和检索的监督微调）以使用LLM作为检索器，在标准的ToolBench检索基准上取得了良好的表现。然而，这些基准使用冗长的、完全指定的查询，并且它们的评估应用了受限解码，这限制了输出为有效的标记路径，未能揭示模型是否真正理解其工具。我们引入了 extbf{ToolSense}，一个开源的LLM驱动的诊断框架，接受任何工具目录作为输入，并自动生成三个基准：一个具有三个模糊层次的现实检索基准（RRB）、一个多项选择题（MCQ）探测基准和一个问答（QA）探测基准。将ToolSense应用于ToolBench（约47,000个工具）并评估五种参数化模型训练配置揭示了知识检索的解离：在RRB查询上，与完全指定的ToolBench基准相比，几种配置的表现下降了约50-64个百分点，低于嵌入模型的基线。此外，尽管检索性能强劲，一些模型在事实探测上得分接近随机，表明存在知识检索的解离。我们在https://github.com/SAP/toolsense上开源了ToolSense框架和ToolBench诊断基准。
cs.AI / 2 / 2606.12563
Arbor: Tree Search as a Cognition Layer for Autonomous Agents
Arbor：作为自主智能体认知层的树搜索
Abstract
Arbor is a multi-agent framework that introduces structured tree search as a cognition layer for autonomous agents operating in large, stateful action spaces. Prior autonomous optimization systems operate on isolated targets with stateless evaluation. Arbor instead maintains an explicit search tree of scored hypotheses that serves as the shared working memory across agents, evolving with every measurement, treating failures as diagnostic signal that reshapes subsequent exploration, and expanding as prior successes shift the bottleneck distribution. We validate Arbor on full-stack LLM inference optimization, a domain where achieving peak performance has historically required coordinated effort from engineering teams across the application, framework, compiler, kernel, and hardware stack. Arbor pairs an Orchestrator agent, which drives optimization by delegating to Domain Specialists across the inference stack, with a Critic agent that safeguards stability through root-cause analysis, introspection, and measurement validation -- a checks-and-balances architecture where neither agent can unilaterally drive the system. Agent capabilities are decomposed into hard skills (domain expertise) and soft skills (coordination protocols that determine how contributions compose), enabling fully autonomous multi-day campaigns. Arbor achieves up to 193% inference throughput-latency Pareto improvement over vendor-optimized baselines, while a single agent without the harness plateaus at +33% throughput improvement and crashes irrecoverably within hours. Arbor generalizes to multiple generations of hardware platform, and run-to-run variance is within 2 percentage points demonstrating that the method is hardware-agnostic and reproducible.
Chinese Translation
Arbor是一个多智能体框架，旨在为在大型、有状态的行动空间中运行的自主智能体引入结构化树搜索作为认知层。以往的自主优化系统通常在孤立的目标上进行无状态评估，而Arbor则维护一个显式的得分假设搜索树，作为智能体之间共享的工作记忆，随着每次测量而演变，将失败视为重塑后续探索的诊断信号，并在先前成功的基础上扩展瓶颈分布。我们在全栈LLM推理优化上验证了Arbor，这是一个在历史上需要跨应用、框架、编译器、内核和硬件栈的工程团队协调努力才能实现最佳性能的领域。Arbor将一个协调者智能体与推理栈中的领域专家进行优化委派相结合，同时配备一个评论者智能体，通过根本原因分析、内省和测量验证来保障系统的稳定性——这是一种制衡架构，确保任何一个智能体都无法单方面驱动系统。智能体的能力被分解为硬技能（领域专业知识）和软技能（决定贡献如何组合的协调协议），从而实现完全自主的多日活动。Arbor在推理吞吐量-延迟的Pareto改进上实现了高达193%的提升，相比于供应商优化的基准，而单个智能体在没有协同的情况下吞吐量提升仅为33%，并在数小时内不可恢复地崩溃。Arbor能够推广到多个代的硬件平台，并且运行间的方差在2个百分点以内，证明该方法具有硬件无关性和可重复性。
cs.AI / 3 / 2606.12587
Strategic Decision Support for AI Agents
人工智能代理的战略决策支持
Abstract
Traditionally, decision support studies how humans use machine learning models to make better decisions. In modern agentic systems, this division of roles is increasingly reversed: AI agents act on behalf of users, while humans and tools becomes support mechanisms around them. This role reversal brings reliability concerns to the forefront, since agentic errors can be consequential and agent behavior must remain aligned with human goals and constraints. Departing from the classical view of decision support, we revisit its two basic principles, the cost--value tradeoff of seeking support and the role of uncertainty quantification, in a setting where AI agents are the central actors. We propose a framework for strategic decision support for AI agents through an optimization problem that minimizes support usage subject to controlling a counterfactual missed-support error: the probability that the agent acts alone on instances where support would have materially improved its output. At the population level, we show that the optimal policy is a threshold rule on the value of support. Building on this structure, we develop an online algorithm that adaptively thresholds such a score and uses randomized exploration to control missed-support error without distributional assumptions. We further introduce a calibration-on-the-fly method that reduces unnecessary support calls online. We instantiate this framework across diverse scenarios, including information gathering, human--AI collaboration, and tool use, showing how each can be modeled through the same strategic decision-support lens. Experiments across these settings show that our method reliably controls the target error while substantially reducing support usage in practice.
Chinese Translation
传统上，决策支持研究人类如何利用机器学习模型做出更好的决策。在现代代理系统中，这种角色的分工正日益逆转：人工智能代理代表用户行动，而人类和工具则成为围绕它们的支持机制。这种角色的逆转将可靠性问题推向了前台，因为代理错误可能带来严重后果，代理行为必须与人类的目标和约束保持一致。我们从经典的决策支持视角出发，重新审视其两个基本原则：寻求支持的成本-价值权衡和不确定性量化的角色，在人工智能代理作为中心参与者的背景下。我们提出了一个针对人工智能代理的战略决策支持框架，通过一个优化问题来最小化支持使用，同时控制反事实未支持错误：即代理在本可以显著改善其输出的情况下单独行动的概率。在总体水平上，我们展示了最优策略是对支持价值的阈值规则。基于这一结构，我们开发了一种在线算法，该算法自适应地对该评分进行阈值处理，并利用随机探索来控制未支持错误，而无需分布假设。我们进一步引入了一种在线即时校准方法，以减少不必要的支持请求。我们在多种场景中实例化该框架，包括信息收集、人类-人工智能协作和工具使用，展示了如何通过相同的战略决策支持视角对每种情况进行建模。在这些设置中的实验表明，我们的方法可靠地控制了目标错误，同时在实践中显著减少了支持使用。
cs.AI / 4 / 2606.12594
Pythagoras-Prover: Advancing Efficient Formal Proving via Augmented Lean Formalisation
毕达哥拉斯证明器：通过增强的 Lean 形式化推进高效的形式证明
Abstract
Modern Lean theorem provers achieve strong performance only with substantial training and inference compute, driven in part by scarce verified proof data and the long reasoning traces of formal proof search, making both supervised fine-tuning (SFT) and sampling expensive. We introduce Pythagoras-Prover, a compute-efficient open-source family of Lean theorem provers built for practical compute budgets. The family spans two generation paradigms: autoregressive models at 4B and 32B parameters, and a first proof-of-concept diffusion-based prover (4B) that iteratively refines Lean proofs at inference time. For training efficiency, we build a Lean-verified corpus stratified into easy, medium, and hard problems for curriculum SFT, so models acquire proof skills progressively from shorter, simpler proofs to longer, harder ones. During SFT, a dynamic proof-reasoning filtering scheme preserves informative proof traces while keeping each instance within an 8k-token context budget. We also introduce Augmented Lean Formalisation (ALF), which expands scarce verified corpora into variants of formal statements, populated via self-distillation for extra training signal without formally verifying every mutated instance. By perturbing known problems while preserving their formal character, ALF reduces reliance on any statement's surface form. Empirically, Pythagoras-Prover-4B surpasses DeepSeek-Prover-V2-671B at pass@32 on MiniF2F-Test (86.1% vs 82.4%) with ~167x fewer parameters, while Pythagoras-Prover-32B sets the open-source state of the art at 93.0% on MiniF2F-Test and solves 93 of 672 PutnamBench problems. We release MiniF2F-ALF, an ALF-mutated contamination-sensitive benchmark on which every evaluated model loses accuracy; here our 32B remains strongest and our 4B matches the prior state of the art, Goedel-Prover-V2-32B.
Chinese Translation
现代 Lean 定理证明器只有在大量训练和推理计算的支持下才能实现强大的性能，这在一定程度上是由于稀缺的经过验证的证明数据和形式证明搜索的长推理轨迹，使得监督微调（SFT）和采样都变得昂贵。我们引入了毕达哥拉斯证明器（Pythagoras-Prover），这是一个计算高效的开源 Lean 定理证明器系列，旨在适应实际计算预算。该系列涵盖了两种生成范式：具有 40 亿和 320 亿参数的自回归模型，以及一个首个基于扩散的概念验证证明器（4B），该证明器在推理时迭代地细化 Lean 证明。为了提高训练效率，我们构建了一个经过 Lean 验证的语料库，按简单、中等和困难问题分层，以便进行课程式 SFT，使模型能够逐步从较短、简单的证明获得证明技能，过渡到较长、困难的证明。在 SFT 过程中，动态证明推理过滤方案保留了信息丰富的证明轨迹，同时将每个实例保持在 8k 令牌的上下文预算内。我们还引入了增强的 Lean 形式化（Augmented Lean Formalisation, ALF），它通过自蒸馏扩展了稀缺的经过验证的语料库，生成形式声明的变体，以提供额外的训练信号，而无需对每个变异实例进行正式验证。通过扰动已知问题，同时保持其形式特征，ALF 减少了对任何声明表面形式的依赖。从经验上看，毕达哥拉斯证明器-4B 在 MiniF2F-Test 上以 86.1% 的通过率超过了 DeepSeek-Prover-V2-671B（82.4%），且参数数量少了约 167 倍，而毕达哥拉斯证明器-32B 在 MiniF2F-Test 上设定了开源领域的最新记录，达到了 93.0%，并解决了 672 个 PutnamBench 问题中的 93 个。我们发布了 MiniF2F-ALF，这是一个 ALF 变异的污染敏感基准，在该基准上每个评估模型的准确性均有所下降；在这里，我们的 32B 仍然是最强的，而我们的 4B 则与之前的最新技术 Goedel-Prover-V2-32B 相匹配。
cs.AI / 5 / 2606.12616
PersonaDrive: Human-Style Retrieval-Augmented VLA Agents for Closed-Loop Driving Simulation
PersonaDrive：用于闭环驾驶仿真的人类风格检索增强VLA代理
Abstract
Closed-loop driving simulators typically populate their environments with non-ego traffic agents that behave largely the same way, produced either by rule-based traffic managers or by learned models trained toward a single behavioral mode. Recent work introduces style variation through post-hoc labels on observational data or LLM-inferred reward weights, but these signals act as proxies for what a style should reward rather than demonstrations of humans explicitly asked to drive in that style. We introduce PersonaDrive, a pipeline that conditions a vision-language-action (VLA) driving agent on retrieved demonstrations from a style-instructed human driving dataset, in which participants drive CARLA leaderboard routes under aggressive, neutral, and conservative instructions on a driver-in-the-loop rig. The pipeline has three stages: (i) offline triplet mining over per-style human driving data using a combined image-text similarity score; (ii) training a lightweight retrieval head that fuses frozen visual features with a small control encoder over per-style databases; and (iii) fine-tuning a single VLA backbone to treat retrieved context points as in-context behavioral demonstrations during waypoint prediction. At inference, the same backbone is conditioned on any style by swapping which per-style database the retrieval head queries, so selecting a style requires no per-style retraining while enabling human-style, style-diverse non-ego agents for closed-loop simulation. On Bench2Drive, PersonaDrive (no style) improves the driving score by 4.6% over SimLingo and 2.5% over HiP-AD, and under style conditioning attains the highest driving score in every style within a roughly 2% band (its weakest style surpassing the strongest baseline, DMW, by 5.4%), while average speed and acceleration rise by 18% and 25% from the conservative to the aggressive instruction.
Chinese Translation
闭环驾驶仿真器通常使用非自我交通代理填充其环境，这些代理的行为大致相同，主要由基于规则的交通管理器或训练为单一行为模式的学习模型生成。近期的研究通过对观察数据的后期标签或大规模语言模型（LLM）推断的奖励权重引入了风格变异，但这些信号充当了风格应奖励的代理，而不是人类被明确要求以该风格驾驶的示范。我们提出了PersonaDrive，一个将视觉-语言-行动（VLA）驾驶代理条件化于从风格指导的人类驾驶数据集中检索的示范的管道，其中参与者在驾驶员参与的设备上根据激进、中立和保守的指令驾驶CARLA排行榜路线。该管道分为三个阶段：（i）使用组合图像-文本相似性得分对每种风格的人类驾驶数据进行离线三元组挖掘；（ii）训练一个轻量级检索头，将冻结的视觉特征与小型控制编码器融合，针对每种风格的数据库；（iii）微调单一的VLA主干，将检索到的上下文点视为在路点预测期间的上下文行为示范。在推理时，相同的主干通过交换检索头查询的每种风格数据库而被条件化于任何风格，因此选择风格不需要针对每种风格的重新训练，同时使得人类风格、风格多样的非自我代理能够用于闭环仿真。在Bench2Drive上，PersonaDrive（无风格）在驾驶评分上比SimLingo提高了4.6%，比HiP-AD提高了2.5%，并且在风格条件下在每种风格中都达到了最高的驾驶评分，波动在大约2%的范围内（其最弱的风格超过了最强的基线DMW，提升了5.4%），同时从保守到激进指令的平均速度和加速度分别提高了18%和25%。
cs.AI / 6 / 2606.12618
"Did you lie?" Evaluating Lie Detectors across Model Scale and Belief-Verified Model Organisms
你在说谎吗？评估不同模型规模和信念验证模型生物的谎言检测器
Abstract
Robust lie detectors for language models could enable powerful techniques for auditing, monitoring, and post-hoc investigation of model behaviour, but evaluating them requires testbeds where models verifiably believe the opposite of what they say. We show that existing trained model organisms often fail this requirement, leaving prior positive and negative detection results difficult to interpret. We address this with 13 reasoning model organisms whose hidden beliefs are verified in chain-of-thought and shown to generalise to held-out tasks, alongside Varied Deception, a prompted-lying testbed covering a broad range of lie-inducing motivations. On these testbeds we evaluate four detectors: a chain-of-thought judge, a logprob classifier, and two activation probes, including Did-You-Lie (DYL), a new method for training follow-up probes. On prompted lying, across 31 open-weight models spanning 2B to 1T parameters, all four detectors show positive scaling with model capability. However, every activation- and logprob-based detector drops sharply on our trained model organisms, with DYL retaining the most signal; only the chain-of-thought judge remains strong, with 0.82 balanced accuracy, partly as an artefact of our verification process favouring CoT-readable beliefs. Current lie detectors therefore cannot support high-confidence claims about model beliefs, and we suggest research directions that may address some of their current limitations. We release our datasets, model organisms, and trained detectors.
Chinese Translation
针对语言模型的稳健谎言检测器能够为审计、监控和事后调查模型行为提供强有力的技术，但评估这些检测器需要测试平台，在这些平台上模型可以可靠地相信他们所说的与事实相反。我们展示了现有的训练模型生物通常无法满足这一要求，这使得先前的正负检测结果难以解释。我们通过13个推理模型生物来解决这一问题，这些生物的隐性信念在思维链中得到了验证，并且能够推广到未见任务，同时引入了Varied Deception，一个涵盖广泛诱导谎言动机的提示性说谎测试平台。在这些测试平台上，我们评估了四种检测器：一个思维链评判器、一个对数概率分类器和两个激活探针，包括Did-You-Lie (DYL)，这是一种新的训练后续探针的方法。在提示性说谎方面，跨越从20亿到1万亿参数的31个开放权重模型，所有四种检测器均显示出与模型能力的正相关性。然而，所有基于激活和对数概率的检测器在我们的训练模型生物上都急剧下降，DYL保留了最多的信号；只有思维链评判器保持强劲，具有0.82的平衡准确率，这在一定程度上是我们验证过程偏向于可读信念的结果。因此，当前的谎言检测器无法支持关于模型信念的高置信度声明，我们建议了一些可能解决其当前局限性的研究方向。我们发布了我们的数据集、模型生物和训练的检测器。
cs.AI / 7 / 2606.12657
TrajGenAgent: A Hierarchical LLM Agent for Human Mobility Trajectory Generation
TrajGenAgent：一种用于人类移动轨迹生成的分层大型语言模型代理
Abstract
Human mobility data is important for transportation, urban planning, and epidemic control, but large-scale trajectory collection is often costly and privacy-constrained, motivating realistic synthetic trajectory generation. Existing LLM-based generators typically rely on either prompt engineering, which preserves zero-shot reasoning but lacks fine-grained spatiotemporal grounding, or trajectory-level fine-tuning, which improves statistical precision but incurs substantial computational cost and may weaken general reasoning. We propose TrajGenAgent, a semantic-aware hierarchical LLM-agent framework for human mobility trajectory generation without model fine-tuning. TrajGenAgent uses a two-stage orchestrator-worker design: an LLM first synthesizes an individual- and weekday-conditioned activity chain from historical evidence via in-context learning, and a deterministic workflow then grounds each activity into a complete visit using personalized POI retrieval, distance-aware location selection, kinematics-aware travel-time propagation, and LLM-based duration estimation. To evaluate realism beyond aggregate spatiotemporal statistics, we introduce an anomaly-detection-based evaluation framework using two complementary detectors to assess behavioral and semantic plausibility. Experiments on benchmark and large-scale simulation datasets show that TrajGenAgent improves spatiotemporal fidelity, semantic coherence, and individual-specific behavioral realism over representative neural and LLM-based baselines, while avoiding parameter updates.
Chinese Translation
人类移动数据对于交通、城市规划和疫情控制至关重要，但大规模轨迹收集往往成本高昂且受到隐私限制，这促使了现实合成轨迹的生成。现有的基于大型语言模型（LLM）的生成器通常依赖于提示工程，这保留了零样本推理但缺乏细粒度的时空基础，或轨迹级别的微调，这提高了统计精度但带来了巨大的计算成本，并可能削弱一般推理能力。我们提出了TrajGenAgent，一种语义感知的分层LLM代理框架，用于人类移动轨迹生成，无需模型微调。TrajGenAgent采用两阶段的协调者-工作者设计：LLM首先通过上下文学习从历史证据中合成个体和工作日条件下的活动链，然后一个确定性的工作流程将每个活动基于个性化的兴趣点（POI）检索、距离感知的位置选择、运动学感知的旅行时间传播和基于LLM的持续时间估计， grounding为完整的访问。为了在聚合时空统计之外评估现实性，我们引入了一种基于异常检测的评估框架，使用两个互补的检测器来评估行为和语义的合理性。在基准和大规模模拟数据集上的实验表明，TrajGenAgent在时空保真度、语义一致性和个体特定的行为现实性方面优于代表性的神经网络和基于LLM的基线，同时避免了参数更新。
cs.AI / 8 / 2606.12674
Evoflux: Inference-Time Evolution of Executable Tool Workflows for Compact Agents
Evoflux：可执行工具工作流的推理时演化用于紧凑型智能体
Abstract
Compact language models (LMs) reduce cost, latency, and deployment risk for tool agents. Yet MCP-style tool use requires more than isolated function calling: an agent must discover tools from live catalogs, satisfy schemas, preserve dependencies across intermediate outputs, and ground final responses in executed evidence. Small planners often generate plausible workflow graphs that fail under tool resolution, parameter validation, dependency tracking, or execution. We argue that this failure mode is poorly handled by small-corpus distillation. A few hundred teacher traces can teach workflow format, but rarely cover the recovery behavior needed to repair failed plans over changing tool catalogs. We introduce Evoflux, an inference-time evolutionary search method that treats compact tool use as the repair of executable tool workflows. It evolves typed workflow graphs through structured edits, execution feedback, adaptive intensity, meta-guided redesign, and diversity pruning. On held-out MCP-Bench tasks spanning live MCP servers and 250 tools, Evoflux raises execution feasibility from roughly 3% to 17-24% across small planners. In contrast, SFT and SFT+DPO on the same search-mined data match, underperform, or collapse below zero-shot performance; ReAct reaches higher peaks, but with higher variance and token cost. These results show that execution-grounded search is more reliable under scarce teacher-trace budgets.
Chinese Translation
紧凑型语言模型（LMs）降低了工具智能体的成本、延迟和部署风险。然而，MCP风格的工具使用不仅仅需要孤立的函数调用：智能体必须从实时目录中发现工具，满足模式，保持中间输出之间的依赖关系，并在执行证据中扎根最终响应。小型规划器通常生成在工具解析、参数验证、依赖追踪或执行方面失败的合理工作流图。我们认为，这种失败模式在小语料库蒸馏中处理得很差。几百个教师轨迹可以教授工作流格式，但很少涵盖在变化的工具目录中修复失败计划所需的恢复行为。我们引入了Evoflux，这是一种推理时的进化搜索方法，将紧凑型工具使用视为可执行工具工作流的修复。它通过结构化编辑、执行反馈、自适应强度、元引导重设计和多样性修剪来演化类型化工作流图。在涵盖实时MCP服务器和250个工具的保留MCP-Bench任务中，Evoflux将执行可行性从大约3%提升到17-24%在小型规划器中。相比之下，SFT和SFT+DPO在相同的搜索挖掘数据上匹配、表现不佳或低于零-shot性能；ReAct达到更高的峰值，但伴随着更高的方差和标记成本。这些结果表明，在稀缺的教师轨迹预算下，基于执行的搜索更为可靠。
cs.AI / 9 / 2606.12683
From AGI to ASI
从AGI到ASI
Abstract
Over the last decade, building human-level artificial general intelligence has moved from far-fetched speculation to being a concrete next-decade target for many of the largest AI organisations. Achieving this goal would have profound and far-reaching impacts on human society, which raises many complex questions for the decade ahead. This report investigates how AI itself might continue to develop in a post-AGI world along the continuum of machine intelligence. The endpoint of this continuum, Universal AI, is theoretically well understood, which provides some formal grounding for the main focus of this report: the transition from human-level AGI to artificial general superintelligence, which, intuitively, can be understood as a system that is more intelligent and cognitively capable than large organisations of humans. After characterizing ASI, the report discusses four potential pathways from AGI to ASI: scaling AGI, AI paradigm shifts, recursive improvement, and ASI emerging from large-scale multi-agent collectives. The report then discusses possible frictions and bottlenecks along these pathways. Determining whether the impact of these frictions will be negligible or substantial raises a number of concrete open research questions. Due to large uncertainties for predicting ASI progress, it cannot be ruled out that AI progress might continue to accelerate over the next years. This could imply that the image of a single transformative step change, caused by the introduction of human-level AGI into our society, could be inaccurate. More apt might be the prospect of a series of transformative societal changes caused by AI-enabled progress and breakthroughs across many areas of science and technology. Preparing for this prospect requires a massively interdisciplinary endeavour of global scope and interest.
Chinese Translation
在过去十年中，构建人类水平的人工通用智能（AGI）已从遥不可及的设想转变为许多大型人工智能组织在下一个十年中的具体目标。实现这一目标将对人类社会产生深远而广泛的影响，这为未来十年提出了许多复杂的问题。本报告探讨了在后AGI世界中，人工智能（AI）如何可能沿着机器智能的连续体继续发展。这个连续体的终点，即通用人工智能（Universal AI），在理论上已被很好地理解，这为本报告的主要焦点提供了一些正式的基础：从人类水平的AGI过渡到人工通用超智能（ASI），直观上可以理解为一种比大型人类组织更智能、更具认知能力的系统。在对ASI进行特征描述后，报告讨论了从AGI到ASI的四种潜在路径：AGI的扩展、AI范式转变、递归改进，以及ASI从大规模多智能体集体中涌现。随后，报告讨论了这些路径上的可能摩擦和瓶颈。确定这些摩擦的影响是微不足道还是显著，提出了一系列具体的开放研究问题。由于预测ASI进展存在较大的不确定性，不能排除在未来几年内AI进展可能继续加速的可能性。这可能意味着，由于人类水平AGI的引入而导致的单一变革性跃迁的形象可能是不准确的。更合适的可能是，由于AI驱动的进步和在科学与技术多个领域的突破而引发的一系列变革性社会变化的前景。为这一前景做好准备需要一个全球范围内的大规模跨学科努力。
cs.AI / 10 / 2606.12702
Deployment-Centered Evaluation: Predicting Query-Level Rejection Risk in a Clinical LLM System
以部署为中心的评估：预测临床大语言模型系统中的查询级拒绝风险
Abstract
Large language models (LLMs) are increasingly integrated into clinical systems, making it essential to evaluate the real-world utility of these systems. However, static benchmarks tend to measure correctness rather than user acceptance, aggregate performance across queries, and require densely annotated datasets -- leading to major blind spots for evaluating clinical systems. In this work, we perform a deployment-centered evaluation of an LLM system embedded within electronic health records at an academic medical center, where user feedback is sparse but closely reflects the deployment conditions. Specifically, we train a pre-response classifier that estimates the risk that a future interaction will result in the user rejecting the LLM response, based on query content and deployment-specific context available before generation. We conduct a prospective analysis of our model over 4.5 months of user feedback, finding that our prediction model achieves an AUROC of 0.719. Further, we estimate the benefit of such predictions in two downstream use cases (guardrail triggering and abstention). Our key conceptual insight is that making use of deployment-specific context (i.e., the provider type, department name, language model used for response), as opposed to only query content, improves the ability to predict whether the user will reject the system output. Altogether, our empirical case study demonstrates the feasibility of predicting user rejection using deployment-specific context, opening the door to targeted guardrails.
Chinese Translation
大型语言模型（LLMs）正越来越多地融入临床系统中，因此评估这些系统在现实世界中的实用性变得至关重要。然而，静态基准往往侧重于测量正确性而非用户接受度，跨查询的整体性能表现，以及需要密集标注的数据集，这导致了评估临床系统时的重大盲点。在本研究中，我们对嵌入在学术医疗中心电子健康记录中的LLM系统进行了以部署为中心的评估，在该环境中，用户反馈稀少但与部署条件密切相关。具体而言，我们训练了一个预响应分类器，该分类器根据查询内容和生成前可用的特定部署上下文，估计未来交互中用户拒绝LLM响应的风险。我们对模型进行了为期4.5个月的用户反馈前瞻性分析，发现我们的预测模型达到了0.719的AUROC。此外，我们还估计了这种预测在两个下游用例中的益处（防护触发和弃权）。我们的关键概念洞察是，利用特定部署上下文（即提供者类型、部门名称、用于响应的语言模型），而不仅仅是查询内容，可以提高预测用户是否会拒绝系统输出的能力。总体而言，我们的实证案例研究展示了利用特定部署上下文预测用户拒绝的可行性，为有针对性的防护措施开辟了新的可能性。
cs.AI / 11 / 2606.12713
Definitional alignment before capability alignment: a Design-Science framework for adjudicating claims about AGI
能力对齐之前的定义对齐：一个用于裁定关于人工通用智能（AGI）主张的设计科学框架
Abstract
Claims that artificial general intelligence has already arrived and claims that it remains decades away are often defended from overlapping evidence. "AGI" lacks a single shared and stable referent and competing operationalizations can return different verdicts on the same system. This article treats that under-specification as a design and governance problem. Following Design Science Research Methodology, it develops DAF-AGI, a second-order conceptual artifact with two coupled components: five ordinal criteria for assessing the adjudicative fitness of candidate definitions and a structured governance audit of authorship, interest, certification, external verification and revision authority. The artifact is demonstrated on five prominent measurement families and one deflationary boundary position in a documented corpus and then stress-tested against a stylized strong arrival claim: that current generative systems constitute AGI because they outperform a well-educated adult on many cognitive tasks. On evidence from the cited 2024-2025 sources, the claim was certifiable only under a performance-based operationalization; capability-ontology, psychometric and skill-acquisition approaches did not certify it, the economic family remains indeterminate and the deflationary position refuses binary adjudication. The contribution is a novel integration and operationalization, not an empirical validation: independent application, inter-rater testing and author-external cases remain necessary. The paper further proposes definitional sovereignty as an enabling component of algorithmic sovereignty: the institutional capacity to contest, certify and revise imported technological categories under public accountability.
Chinese Translation
关于人工通用智能（AGI）已经到来的主张与认为其仍需数十年才能实现的主张，常常基于重叠的证据进行辩护。“AGI”缺乏单一共享且稳定的指称，竞争性的操作化可以对同一系统得出不同的裁决。本文将这种不足的规范视为设计和治理问题。遵循设计科学研究方法论，本文开发了DAF-AGI，一个具有两个耦合组件的二阶概念性工件：五个评估候选定义裁决适应性的序数标准，以及对作者身份、利益、认证、外部验证和修订权威的结构化治理审计。该工件在五个显著的测量家族和一个贬值边界位置的文献语料库中进行了演示，并针对一个典型的强到达主张进行了压力测试：当前的生成系统构成AGI，因为它们在许多认知任务上超越了受过良好教育的成年人。根据引用的2024-2025年来源的证据，该主张仅在基于表现的操作化下可被认证；能力本体论、心理测量和技能获取方法未能认证该主张，经济家族仍然不确定，而贬值立场拒绝二元裁决。本文的贡献在于一种新颖的整合和操作化，而非实证验证：独立应用、评分者间测试和作者外部案例仍然是必要的。本文进一步提出定义主权作为算法主权的一个促进因素：在公共问责制下争议、认证和修订进口技术类别的制度能力。
cs.AI / 12 / 2606.12721
The Theory of Mind Utility: Formal Specification of a Mentalizing Mechanism
心智理论效用：心智化机制的形式化规范
Abstract
Inferring others' beliefs requires more than reading surface signals; it requires tracking who told them what, in what order, and how credibly. The Theory of Mind Utility (ToM-U) formalizes this epistemic state inference problem at the computational level of analysis, specifying what mentalizing computes and why without commitment to algorithmic or neural implementation. ToM-U achieves this by constructing Local Epistemic World Models (LEWMs) -- directed typed graphs that represent agents, state nodes, and the epistemic relationships among them -- and evaluating discrete candidate LEWMs against observed behavior until one achieves sufficient confidence. Five formal definitions specify the LEWM structure, agent node properties including ordered information access history, a bounded proliferation mechanism for recursive mentalizing, three inference procedures, and a residue function that captures the structured trace left by failed mentalizing attempts. ToM-U differs from Bayesian Theory of Mind and adjacent formal accounts, which presuppose rather than derive belief states, and from simulation theory and theory-theory, which lack a formal apparatus for epistemic state inference. The architecture generates directional, falsifiable predictions about mentalizing failure that follow from structural properties of the model rather than auxiliary assumptions, and positions ToM-U as a domain-agnostic mechanism upstream of goal inference and other downstream social cognitive processes.
Chinese Translation
推断他人的信念不仅需要解读表面信号，还需要追踪谁在何时以何种可信度告诉他们什么。心智理论效用（Theory of Mind Utility, ToM-U）在计算分析的层面上形式化了这一认知状态推断问题，明确了心智化计算的内容及其原因，而不依赖于算法或神经实现的承诺。ToM-U通过构建局部认知世界模型（Local Epistemic World Models, LEWMs）来实现这一目标——这些有向类型图表示代理、状态节点及其之间的认知关系，并对观察到的行为评估离散候选LEWM，直到达到足够的置信度。五个形式定义规定了LEWM结构、代理节点属性（包括有序信息访问历史）、用于递归心智化的有界扩展机制、三种推理程序以及捕捉失败心智化尝试所留下的结构化痕迹的残余函数。ToM-U与贝叶斯心智理论及相关的形式化模型不同，后者假设而非推导信念状态；也与模拟理论和理论理论不同，后者缺乏用于认知状态推断的形式化工具。该架构生成关于心智化失败的方向性、可证伪的预测，这些预测源于模型的结构属性而非辅助假设，并将ToM-U定位为目标推断及其他下游社会认知过程的领域无关机制。
cs.AI / 13 / 2606.12730
Rethinking Psychometric Evaluation of LLMs: When and Why Self-Reports Predict Behavior
重新思考大语言模型的心理测量评估：何时以及为何自我报告能够预测行为
Abstract
Anticipating LLM behavioral tendencies from low-cost psychometric probes is critical for safe deployment, but only if self-reports (SR) reliably predict behavior. Recent work documented substantial SR-behavior dissociation in LLMs, but relied on broad personality traits (Big 5) that predict specific behaviors weakly, even in humans. Furthermore, the isolation of conversational sessions combined with weak context matching left open whether LLMs truly lack coherence or whether the conditions needed to detect such coherence were not met. We contrast Big 5 with the Theory of Planned Behavior (TPB), which measures intention targeted to a specific behavior and predicts human behavior substantially better than broad traits. We run experiments across four behavioral tasks and 11 frontier LLMs, while also varying session context and identity induction. We find that SR-behavior coherence exists but is selective. 1) Within a shared conversation, the Theory of Planned Behavior reaches human-level coherence; Big 5 does not. 2) Across separate conversations, coherence survives only for behaviors anchored outside the immediate prompt, such as implicit bias shaped by training, and collapses when behavior is strongly primed by context, as with sycophancy. 3) Persona prompting makes self-reports more consistent across conversations, but does not bring behavior into alignment. These findings suggest that coarse personality frameworks, such as Big 5 may not be the best tools for testing deployment behavior. More task- and behavior-specific instruments are needed, and even these must be evaluated across tasks and contexts.
Chinese Translation
从低成本的心理测量探针中预测大语言模型（LLMs）的行为倾向对于安全部署至关重要，但前提是自我报告（SR）能够可靠地预测行为。近期研究记录了LLMs中自我报告与行为之间的显著脱节，但依赖于广泛的人格特质（大五人格），这些特质即使在人类中也只能弱预测特定行为。此外，孤立的对话会话结合弱上下文匹配使得尚不清楚LLMs是否真的缺乏连贯性，或是未满足检测这种连贯性所需的条件。我们将大五人格与计划行为理论（TPB）进行对比，后者测量针对特定行为的意图，并比广泛的人格特质更好地预测人类行为。我们在四个行为任务和11个前沿LLMs上进行实验，同时变化会话上下文和身份诱导。我们发现自我报告与行为之间的连贯性确实存在，但具有选择性。1）在共享对话中，计划行为理论达到了人类水平的连贯性；而大五人格则没有。2）在不同的对话中，连贯性仅在与即时提示无关的行为中存在，例如受训练影响的隐性偏见，而在行为受到上下文强烈提示时（如谄媚行为）则崩溃。3）角色提示使自我报告在不同对话中更一致，但并未使行为保持一致。这些发现表明，像大五人格这样粗略的人格框架可能不是测试部署行为的最佳工具。需要更具任务和行为特异性的工具，即便这些工具也必须在不同任务和上下文中进行评估。
cs.AI / 14 / 2606.12736
Benchmarking AI Agents for Addressing Scientific Challenges Across Scales
针对跨尺度科学挑战的人工智能代理基准测试
Liu, Tianyu, Wang, Allen Xin, Panescu, Antonia, Chen, Lisa Xinyi, Long, Wenxin, Wei, Xinyu, Jing, Yueqian, Zeng, Ziyao, Chen, Jihang, Jiang, Sihan, Wang, Ziqing, Gu, Siyi, Chen, Siyu, Hu, Xinyang, Shao, Haoran, Xu, Leqi, Zheng, Wangjie, Cao, Zhiyuan, Fang, Ada, Yu, Botao, Sun, Kunyang, Ying, Rex, Cohan, Arman, Chen, Qingyu, Xue, Lingzhou, Ding, Kaize, Du, Yuanqi, Jin, Wengong, Yang, Zhuoran, Zitnik, Marinka, Zou, James, Xu, Hua, Zhao, Hongyu
Abstract
AI agents are increasingly being developed to accelerate scientific discovery, yet their practical capabilities in real research settings remain poorly understood. Existing benchmarks for AI agents rarely capture the complexity, heterogeneity, and extended reasoning required by scientific work, whereas benchmarks for scientific tasks often reduce research to static, direct problems and provide limited support for interactive evaluation. Here, we introduce SciAgentArena, a systematic benchmark for evaluating AI agents in real-world scientific research scenarios drawn from emerging needs across multiple domains. SciAgentArena comprises approximately 200 tasks with stepwise verification and an interactive, agent-agnostic environment for assessing diverse AI agents. Using this benchmark, we find that current agents can contribute effectively to well-specified data-analysis workflows, particularly when the task structure and evaluation criteria are clear. However, their performance remains uneven across scientific contexts: agents struggle to generate genuinely novel insights, sustain self-directed exploration, and formulate robust solutions for open-ended research questions. We further characterize common failure modes across agents and identify opportunities for improving their reliability, autonomy, and scientific reasoning. Together, SciAgentArena provides a practical framework for measuring progress in AI agents for science and for guiding the design of future agents capable of addressing complex scientific challenges. Full codes, tasks, and datasets can be accessed via this link: https://sciagentarena.github.io/.
Chinese Translation
人工智能代理正在不断发展，以加速科学发现，但它们在实际研究环境中的实际能力仍然不够清楚。现有的人工智能代理基准测试很少捕捉到科学工作所需的复杂性、异质性和扩展推理，而科学任务的基准测试往往将研究简化为静态的、直接的问题，并对互动评估提供有限支持。在此，我们引入了 SciAgentArena，这是一个系统的基准测试，用于评估在多个领域新兴需求下的真实科学研究场景中的人工智能代理。SciAgentArena 包含大约 200 个任务，具有逐步验证和一个互动的、与代理无关的环境，用于评估多样化的人工智能代理。使用该基准测试，我们发现当前的代理在明确任务结构和评估标准时，可以有效地贡献于明确的数据分析工作流程。然而，它们在科学背景下的表现仍然不均衡：代理在生成真正新颖的见解、维持自我导向的探索和制定针对开放性研究问题的稳健解决方案方面存在困难。我们进一步描述了代理的常见失败模式，并识别了提高其可靠性、自主性和科学推理能力的机会。总之，SciAgentArena 提供了一个实用的框架，用于衡量人工智能代理在科学领域的进展，并指导未来能够应对复杂科学挑战的代理设计。完整的代码、任务和数据集可以通过以下链接访问：https://sciagentarena.github.io/
cs.AI / 15 / 2606.12742
Reducing the Complexity of Deep Learning Models for EEG Analysis on Wearable Devices
降低可穿戴设备上脑电图分析的深度学习模型复杂性
Abstract
Wearable healthcare devices are the fastest-growing Internet of Things (IoT) sector. Many automated healthcare services rely on two crucial biological signals, namely ECG and EEG, which reflect the activity of the heart and brain, respectively. Although deep neural networks are considered the primary way to process and analyze these signals, the very tight energy and computational power constraints in wearable devices are far below the computational, energy, and memory bandwidth demands of DNN models, thereby impeding the deployment of deep learning in many practical wearable services. This paper investigates the feasibility of deploying state-of-the-art DNN models in resource-constrained wearable devices. Notably, we explore the trade-off between accuracy and computational complexity of DNNs when parameter quantization and electrode reduction methods are used. Our investigation centers on several state-of-the-art DNN models designed for EEG signal analysis, specifically for detecting epileptic seizures. Our findings demonstrate that, when applied judiciously, these techniques can significantly reduce the complexity of the DNNs under consideration with minimal adverse effects on accuracy. These results reveal the explicit trade-offs between accuracy and complexity reduction encountered when adapting DNN-based online EEG analysis for wearable devices.
Chinese Translation
可穿戴医疗设备是增长最快的物联网（IoT）领域。许多自动化医疗服务依赖于两种关键的生物信号，即心电图（ECG）和脑电图（EEG），分别反映心脏和大脑的活动。尽管深度神经网络（DNN）被认为是处理和分析这些信号的主要方式，但可穿戴设备中极为严格的能量和计算能力限制远低于DNN模型的计算、能量和内存带宽需求，从而阻碍了深度学习在许多实际可穿戴服务中的部署。本文探讨了在资源受限的可穿戴设备中部署最先进的DNN模型的可行性。特别地，我们研究了在使用参数量化和电极减少方法时，DNN的准确性与计算复杂性之间的权衡。我们的研究集中于几种为EEG信号分析而设计的最先进的DNN模型，特别是用于检测癫痫发作。我们的发现表明，当这些技术得到合理应用时，可以显著降低所考虑的DNN的复杂性，同时对准确性产生最小的不利影响。这些结果揭示了在将基于DNN的在线EEG分析适应于可穿戴设备时，准确性与复杂性降低之间的明确权衡。
cs.AI / 16 / 2606.12747
Prefill Awareness in Large Language Models
大型语言模型中的预填充意识
Abstract
Safety-relevant studies of language models, including alignment and jailbreaking evaluations and AI control protocols, often rely on prefilling model outputs. If AI models can recognize and act on the fact their prior assistant messages have been inserted or edited, the effectiveness and validity of these methods could be compromised. We investigate whether frontier language models can distinguish between tampered and untampered assistant-side context, a capability we call prefill awareness. To do so, we construct a binary preference benchmark across three prefill mechanisms, filtering for cases where models show consistent stances. We find that frontier models show substantial prefill awareness: Claude Opus 4.5 detects prefills opposing its preferences in 9-35% of cases with a 0% false positive rate when prompted; additionally, models often revert towards baseline behavior without explicitly reporting that the prefill was foreign. Controlled ablations later also show that detection and resistance rely on different cues, where stylistic mismatch mainly affects whether models flag a prefill as foreign, while preference mismatch mainly affects whether they revert toward their baseline answer. We also examine more realistic agentic settings such as misalignment-continuation evaluations and SWE-bench trajectories, where frontier models sometimes disavow prefilled assistant turns in ways that depend strongly on dataset, task success, and hidden formatting artifacts. Our results indicate that prefill awareness is already a substantial confound for some prefill-based methods. We recommend that model developers track this capability in frontier systems.
Chinese Translation
与安全相关的语言模型研究，包括对齐和越狱评估以及人工智能控制协议，通常依赖于预填充模型输出。如果人工智能模型能够识别并对其先前助手消息被插入或编辑的事实做出反应，那么这些方法的有效性和有效性可能会受到影响。我们研究前沿语言模型是否能够区分被篡改和未被篡改的助手侧上下文，这种能力我们称之为预填充意识。为此，我们构建了一个跨越三种预填充机制的二元偏好基准，筛选出模型表现出一致立场的案例。我们发现前沿模型表现出显著的预填充意识：Claude Opus 4.5在9-35%的情况下检测到与其偏好相反的预填充，且在提示时的假阳性率为0%；此外，模型通常在没有明确报告预填充是外来的情况下，回归到基线行为。后续的控制消融实验还表明，检测和抵抗依赖于不同的线索，其中风格不匹配主要影响模型是否将预填充标记为外来，而偏好不匹配主要影响它们是否回归到基线答案。我们还考察了更现实的代理设置，如不对齐持续评估和SWE-bench轨迹，在这些情况下，前沿模型有时会以强烈依赖于数据集、任务成功和隐藏格式化伪影的方式否认预填充的助手转折。我们的结果表明，预填充意识已经对某些基于预填充的方法构成了显著的干扰。我们建议模型开发者在前沿系统中跟踪这一能力。
cs.AI / 17 / 2606.12767
Constructing Evaluation Datasets for Procedural Reasoning: Balancing Naturalness, Grounding, and Multi-Hop Coverage
构建程序推理评估数据集：平衡自然性、基础性和多跳覆盖率
Abstract
Evaluating procedural reasoning in AI-supported learning systems requires question-answer datasets that are both learner-like and grounded in the instructional knowledge the system is expected to use. We study how TMK-based question generation strategies affect dataset quality for procedural and multi-hop reasoning. We compare three strategies: strict generation from Task-Method-Knowledge (TMK) models, transcript-first generation with post-hoc TMK filtering, and TMK-aware generation that combines transcripts with structured guidance. To evaluate generated items, we introduce a grounding validation framework based on closed-set evidence units extracted from TMK models. The framework measures whether answers are supported by the underlying representation, whether questions are self-contained, and whether they target multi-hop procedural reasoning. Across 23 instructional topics and 690 generated question-answer pairs, strict TMK generation achieves the strongest overall quality, with 96.5% grounded questions and 92.6% usable questions. Transcript-first generation produces more learner-like questions but more context-dependent or weakly grounded items, while TMK-aware generation yields high raw multi-hop coverage but lower grounding. These results show that procedural richness and natural phrasing do not guarantee representational grounding, motivating explicit representation-aware validation for evaluation datasets in AI-supported learning.
Chinese Translation
在AI支持的学习系统中评估程序推理需要既符合学习者特征又基于系统预期使用的教学知识的问题-答案数据集。我们研究了基于任务-方法-知识（Task-Method-Knowledge, TMK）的问答生成策略如何影响程序推理和多跳推理的数据集质量。我们比较了三种策略：严格从TMK模型生成、基于转录的生成并进行后期TMK过滤，以及结合转录和结构化指导的TMK感知生成。为了评估生成的项目，我们引入了一种基于从TMK模型提取的封闭集证据单元的基础验证框架。该框架衡量答案是否得到基础表示的支持，问题是否自给自足，以及它们是否针对多跳程序推理。在23个教学主题和690对生成的问题-答案中，严格的TMK生成实现了最强的整体质量，具有96.5%的基础问题和92.6%的可用问题。基于转录的生成产生了更符合学习者特征的问题，但更多的是上下文依赖或基础较弱的项目，而TMK感知生成则产生了较高的原始多跳覆盖率，但基础性较低。这些结果表明，程序丰富性和自然措辞并不保证表示的基础性，促使我们在AI支持的学习评估数据集中进行明确的表示感知验证。
cs.AI / 18 / 2606.12783
A Tutorial on World Models and Physical AI
世界模型与物理人工智能教程
Abstract
World modeling is emerging as a central principle for building intelligent systems capable of prediction, reasoning, and decision making. A central distinction can be drawn between explicit world models, which learn structured dynamics for rollout-based reasoning and planning, and implicit world models, which encode predictive structure within scalable learned representations. These complementary paradigms provide a foundation for physical AI in domains such as robotics and autonomous driving, enabling intelligence beyond reactive control under real-world constraints. Recent foundation models further suggest a pathway toward unified systems integrating perception, prediction, and action. Despite rapid progress, major challenges remain in hierarchical reasoning, long-horizon planning, and autonomous goal formation, which are critical for advancing toward artificial general intelligence. This tutorial presents a coherent framework in which diverse world modeling approaches are unified through shared predictive structure and differentiated by how such structure is represented and exploited.
Chinese Translation
世界建模正成为构建能够进行预测、推理和决策的智能系统的核心原则。可以明确区分显式世界模型和隐式世界模型，前者学习用于基于回放的推理和规划的结构化动态，后者则在可扩展的学习表示中编码预测结构。这两种互补的范式为物理人工智能在机器人技术和自动驾驶等领域奠定了基础，使智能超越了在现实世界约束下的反应控制。最近的基础模型进一步提出了一条整合感知、预测和行动的统一系统的路径。尽管取得了快速进展，但在分层推理、长时间规划和自主目标形成等方面仍然面临重大挑战，这些都是推动人工通用智能发展的关键。本教程呈现了一个连贯的框架，其中多样的世界建模方法通过共享的预测结构统一，并通过这种结构的表示和利用方式进行区分。
cs.AI / 19 / 2606.12797
The Containment Gap: How Deployed Agentic AI Frameworks Fail Public-Facing Safety Requirements
控制缺口：已部署的自主智能框架如何未能满足面向公众的安全要求
Abstract
Agentic large language model systems that autonomously invoke tools, maintain persistent memory, and execute multi-step plans are increasingly deployed in public-facing domains, including government services, healthcare triage, and financial advising. We ask whether the frameworks used to build these systems provide architectural-level structural safety guarantees. Applying six containment principles derived from a compositional model of agentic architectures, we audit three dominant frameworks (LangChain, AutoGPT, and OpenAI Agents SDK) and find no native compliance in any of them. Memory integrity, a defense against one of the most prevalent vulnerability classes, is not observed in any of the three evaluated frameworks. We validate these findings empirically: in a simulated government benefits agent built on LangChain, a single memory-poisoning write induces persistent targeted corruption across all tested seeds and backends, increasing the wrongful denial rate for targeted applicants to 88.9%. Under a complex five-factor policy, the same attack preserves aggregate accuracy while increasing targeted wrongful denials by 3.5x, rendering the corruption difficult to detect through standard monitoring. We then introduce two lightweight containment mechanisms: a memory integrity validator and a policy gate, which eliminate both attack vectors with sub-millisecond overhead (<0.2ms per call). We conclude that the current agentic framework ecosystem may not yet meet secure-by-default expectations for public-facing deployments and outline priority architectural interventions to enable trustworthy deployment in high-stakes, socially impactful applications.
Chinese Translation
自主调用工具、保持持久记忆并执行多步骤计划的智能大型语言模型系统正越来越多地应用于面向公众的领域，包括政府服务、医疗分诊和金融咨询。我们探讨用于构建这些系统的框架是否提供了架构级别的结构安全保障。通过应用六项源自自主架构组合模型的控制原则，我们审计了三种主流框架（LangChain、AutoGPT 和 OpenAI Agents SDK），发现它们均未具备原生合规性。在评估的三种框架中，未观察到记忆完整性，这是一种针对最常见漏洞类别的防御措施。我们通过实证验证了这些发现：在基于 LangChain 构建的模拟政府福利代理中，单次记忆污染写入会导致所有测试种子和后端的持续目标性损坏，使目标申请者的错误拒绝率增加至 88.9%。在一个复杂的五因素政策下，同样的攻击在保持整体准确性的同时，将目标错误拒绝率提高了 3.5 倍，使得通过标准监控检测到这种损坏变得困难。随后，我们引入了两种轻量级控制机制：记忆完整性验证器和政策门，这两者在每次调用时的开销均低于毫秒级（<0.2ms），能够消除这两种攻击向量。我们得出结论，目前的自主框架生态系统可能尚未满足面向公众部署的默认安全预期，并概述了优先的架构干预措施，以实现高风险、社会影响深远的应用中的可信部署。
cs.AI / 20 / 2606.12809
MLUBench: A Benchmark for Lifelong Unlearning Evaluation in MLLMs
MLUBench：用于评估多模态大语言模型终身遗忘的基准
Abstract
Multimodal large language models (MLLMs) are trained on massive multimodal data, making data unlearning increasingly important as data owners may request the removal of specific content. In practice, these requests often arrive sequentially over time, giving rise to the challenging problem of MLLM Lifelong Unlearning. However, most existing benchmarks are limited in scale and scope, failing to capture the complexities of MLLM lifelong unlearning. To fill this gap, we introduce the MLUBench, a large-scale and comprehensive benchmark featuring 127 entities across 9 classes under lifelong unlearning requests. We perform extensive experiments using MLUBench and reveal that existing unlearning methods suffer from severe, cumulative degradation. More critically, we further identify the unique challenge of this problem: unlike in unimodal models, MLLM lifelong unlearning is constrained by the need to preserve multimodal alignment. Continually unlearning from one modality could degrade the entire model. To alleviate this challenge, we propose LUMoE, an effective method. Experiments demonstrate that LUMoE significantly mitigates the degradation problem faced by baselines. The source code and the MLUBench dataset are open-sourced in https://github.com/lihe-maxsize/Lifelong_Unlearning_main.
Chinese Translation
多模态大语言模型（MLLMs）是在大量多模态数据上训练的，因此数据的终身遗忘变得愈发重要，因为数据拥有者可能会请求删除特定内容。在实际操作中，这些请求通常是随着时间的推移而依次到来的，从而产生了MLLM终身遗忘这一具有挑战性的问题。然而，现有的大多数基准在规模和范围上都有限，未能捕捉到MLLM终身遗忘的复杂性。为填补这一空白，我们提出了MLUBench，这是一个大规模且全面的基准，涵盖了在终身遗忘请求下的9个类别中的127个实体。我们使用MLUBench进行了广泛的实验，揭示了现有的遗忘方法存在严重的累积退化问题。更重要的是，我们进一步识别出这一问题的独特挑战：与单模态模型不同，MLLM的终身遗忘受到保持多模态对齐的需求的限制。从一个模态持续进行遗忘可能会导致整个模型的退化。为缓解这一挑战，我们提出了LUMoE，这是一种有效的方法。实验表明，LUMoE显著减轻了基线方法面临的退化问题。源代码和MLUBench数据集已在https://github.com/lihe-maxsize/Lifelong_Unlearning_main开源。
cs.AI / 21 / 2606.12817
Teach-and-Repeat: Accurately Extracting Operational Knowledge from Mobile Screen Demonstrations to Empower GUI Agents
教学与重复：从移动屏幕演示中准确提取操作知识以增强GUI代理
Abstract
Understanding the digital world on mobile devices is shifting from static UI perception to dynamic action comprehension. This capability enables models to convert visual state transitions into operational knowledge, defined as short natural-language sentences that describe action types, target UI elements, textual arguments, and execution orders. However, due to the highly diverse and heterogeneous UI designs across applications, existing vision-language models (VLMs) struggle to accurately infer these underlying operations. To bridge this gap, we introduce Teach VLM, a core model designed to translate mobile screen trajectories into step-wise operational knowledge by extracting and analyzing operation-related keyframes from demonstration videos. To address the scarcity of aligned training data, we develop a systematic data flywheel for scalable data acquisition. We further introduce a novel Chinese Mobile Screen Teach Benchmark for fine-grained evaluation. Building upon Teach VLM, we propose the Teach-and-Repeat paradigm, where the generated operational knowledge serves as an interpretable procedural reference to guide downstream screen-based execution agents. Extensive evaluations demonstrate that Teach VLM significantly outperforms strong VLM baselines, achieving state-of-the-art performance in operation semantics prediction. Furthermore, experiments in Android World show that our paradigm yields consistent Task Success Rate improvements for downstream agents. Together, Teach VLM and the Teach-and-Repeat paradigm offer a practical pathway from raw demonstrations to reusable task automation.
Chinese Translation
在移动设备上理解数字世界的方式正从静态用户界面感知转向动态动作理解。这一能力使得模型能够将视觉状态转变转换为操作知识，操作知识被定义为描述动作类型、目标用户界面元素、文本参数和执行顺序的简短自然语言句子。然而，由于应用程序之间用户界面设计的高度多样性和异构性，现有的视觉语言模型（VLMs）在准确推断这些潜在操作方面面临挑战。为了解决这一问题，我们提出了Teach VLM，这是一个核心模型，旨在通过从演示视频中提取和分析与操作相关的关键帧，将移动屏幕轨迹转换为逐步操作知识。为了应对对齐训练数据的稀缺性，我们开发了一个系统化的数据飞轮，以实现可扩展的数据获取。我们进一步推出了一个新的中文移动屏幕教学基准，以进行细粒度评估。在此基础上，我们提出了教学与重复（Teach-and-Repeat）范式，其中生成的操作知识作为可解释的程序参考，指导下游基于屏幕的执行代理。广泛的评估表明，Teach VLM显著优于强大的VLM基线，在操作语义预测中实现了最先进的性能。此外，在Android World中的实验表明，我们的范式为下游代理带来了持续的任务成功率提升。总之，Teach VLM和教学与重复范式提供了一条从原始演示到可重用任务自动化的实用路径。
cs.AI / 22 / 2606.12821
GeoNatureAgent Benchmark: Benchmarking LLM Agents for Environmental Geospatial Analysis Across Frontier and Open-Weight Foundation Models
GeoNatureAgent 基准：针对前沿和开放权重基础模型进行环境地理空间分析的 LLM 代理基准测试
Abstract
Environmental scientists spend disproportionate effort on data wrangling rather than analysis, and AI agents that automate geospatial workflows remain unvalidated: no benchmark evaluates agents operating through structured tool calling against real APIs. We introduce the GeoNatureAgent Benchmark, the first benchmark for environmental analysis agents that operate via structured tool calls to a production-style geospatial API. It comprises 93 tasks across 18 categories, covering municipality analysis, multi-turn conversation, spatial reasoning, cross-indicator synthesis, error handling and recovery, ranking, comparison, multilingual understanding, habitat analysis, and task rejection. Tasks are evaluated against an open, self-hostable API serving three environmental indicators across Spain and Portugal via sixteen tools. We evaluate seven LLMs (Claude Sonnet 4, DeepSeek V3.2, GLM-5, Gemini 2.5 Pro, Qwen3-235B, GPT-OSS-120B, Llama 4 Scout) under three temperature-1.0 seeds, reporting capability and per-case cost as orthogonal axes. We find: (1) Claude Sonnet 4 leads at 60.8% +/- 0.8%, followed by DeepSeek V3.2 at 56.3% +/- 3.1%, with no other model above 51%; (2) the cost-accuracy Pareto frontier is occupied mostly by open-weight models, with DeepSeek V3.2 offering 93% of Claude's capability at 11x lower cost ($0.011/case); (3) comparison tasks remain universally unsolved (0% on close-value comparisons), exposing systematic reasoning limits; and (4) structured tool calling against a real API is more discriminative than general-purpose GIS benchmarks, with accuracies 25-35 points lower. We further show extensibility by integrating BigEarthNet V2 land cover for Portugal alongside Spanish CO2 and erosion indicators. The benchmark, harness, and self-hostable API are publicly available.
Chinese Translation
环境科学家在数据整理上花费的精力远超过分析，而自动化地理空间工作流程的 AI 代理尚未经过验证：没有基准评估通过结构化工具调用与真实 API 交互的代理。我们推出了 GeoNatureAgent 基准，这是首个针对通过结构化工具调用与生产级地理空间 API 交互的环境分析代理的基准。该基准包含 93 个任务，涵盖 18 个类别，包括市政分析、多轮对话、空间推理、跨指标综合、错误处理与恢复、排名、比较、多语言理解、栖息地分析和任务拒绝。任务的评估基于一个开放的、自托管的 API，该 API 提供西班牙和葡萄牙的三个环境指标，并通过十六个工具进行服务。我们在三个温度为 1.0 的种子下评估了七个 LLM（Claude Sonnet 4、DeepSeek V3.2、GLM-5、Gemini 2.5 Pro、Qwen3-235B、GPT-OSS-120B、Llama 4 Scout），报告能力和每个案例成本作为正交轴。我们的发现包括：（1）Claude Sonnet 4 以 60.8% +/- 0.8% 的准确率领先，其次是 DeepSeek V3.2，准确率为 56.3% +/- 3.1%，没有其他模型超过 51%；（2）成本-准确率的帕累托前沿主要由开放权重模型占据，DeepSeek V3.2 在 11 倍更低的成本（$0.011/案例）下提供了 Claude 93% 的能力；（3）比较任务普遍未解决（在接近值比较中为 0%），暴露出系统推理的局限性；（4）与真实 API 的结构化工具调用比通用 GIS 基准更具区分性，准确率低 25-35 分。我们进一步展示了通过整合 BigEarthNet V2 土地覆盖数据与西班牙的 CO2 和侵蚀指标来扩展基准。该基准、工具和自托管 API 已公开提供。
cs.AI / 23 / 2606.12828
Topical Phase Transitions in Artificial Intelligence Research: Large-Scale Evidence and an Early-Warning Signature for Emerging Topics
人工智能研究中的主题相变：大规模证据及新兴主题的早期预警特征
Abstract
Do research topics in artificial intelligence grow gradually, or do they advance through abrupt, detectable jumps? Analyzing 80,814 accepted main-track papers from five premier AI conferences (ACL, CVPR, ICLR, ICML, NeurIPS) spanning 2017 to 2025, we show major AI topics advance through topical phase transitions: remaining marginal for years, then surging across venues within one to three years. Large language models became the dominant cross-venue topic by 2025, diffusion models rose with comparable abruptness, and language-model methods crossed into computer vision via vision-language models, whereas reinforcement learning compounded smoothly, distinguishing genuine phase transitions from ordinary growth. This structure is our primary contribution: a large-scale, cross-venue characterization of how AI research reorganizes. We then ask whether a transition leaves a detectable footprint before it peaks. We define an early-warning signature, four publication-dynamics criteria frozen on 2017-2021 data, and evaluate it out of sample on 2023-2025 transitions, obtaining a precision of 27% and recall of 63% against a 13.5% base rate. Applied to 2025 data, the signature flags reasoning and test-time compute, agentic AI, multimodal LLMs, retrieval-augmented generation, and world models as topics to monitor over 2026-2028. The source code is also publicly available on GitHub at https://github.com/KurbanIntelligenceLab/ai-phase-transitions.
Chinese Translation
人工智能研究主题是逐渐增长，还是通过突发的、可检测的跃迁推进？通过分析2017年至2025年间五个顶尖AI会议（ACL、CVPR、ICLR、ICML、NeurIPS）接受的80,814篇主轨道论文，我们展示了主要AI主题通过主题相变推进：在多年内保持边缘状态，然后在一到三年内在多个场合中激增。到2025年，大型语言模型成为跨场合的主导主题，扩散模型以相似的突发性上升，而语言模型方法通过视觉-语言模型跨入计算机视觉领域，强化学习则平稳增长，从而区分出真正的相变与普通增长。这一结构是我们的主要贡献：对AI研究如何重组的大规模、跨场合特征描述。接着，我们询问相变是否在达到峰值之前留下可检测的痕迹。我们定义了一个早期预警特征，基于2017-2021年的数据设定了四个出版动态标准，并在2023-2025年的相变中进行样本外评估，获得了27%的精确率和63%的召回率，基于13.5%的基线率。应用于2025年的数据，该特征标记了推理和测试时计算、代理AI、多模态LLMs、检索增强生成和世界模型等主题，作为2026-2028年需监测的主题。源代码也已在GitHub上公开，链接为 https://github.com/KurbanIntelligenceLab/ai-phase-transitions。
cs.AI / 24 / 2606.12834
Fantastic Scientific Agents and How to Build Them: AgentBuild for Rietveld Refinement
奇妙的科学代理及其构建方法：用于Rietveld精修的AgentBuild
Abstract
As scientific workflows shift from deterministic executables to LLM-based agents, the development practices on offer, such as fine-tuning, reinforcement learning, and prompt-and-go, bury the scientist's judgment. We propose treating agent construction as a workflow stage and introduce AgentBuild, which builds a scientific agent from a contract the scientist authors. The contract is a version-controlled rubric, a difficulty-graded curriculum, and a curated external knowledge base. A rubric-driven judge gates a meta-optimizer coding agent that edits the agent within a declared boundary, so the build compiles the agent, not the scientist's judgment. We instantiate this for Rietveld refinement of X-ray diffraction data through GSAS-II behind MCP and A2A, where a blank-harness construction run progresses through a lithium lanthanum zirconium oxide (LLZO) signal-to-noise ladder, reaches the 4 hour scan as a frontier case, and exposes the workflow-scope limits that remain. The same rubric that rewards credible fits also scores trajectory scope, making the frontier a contract failure rather than a pattern-fitting failure. As base models evolve, re-running AgentBuild is a re-tune, not a rebuild, and the scientist's authored contract remains the durable asset.
Chinese Translation
随着科学工作流程从确定性可执行文件转向基于大语言模型（LLM）的代理，现有的开发实践，如微调、强化学习和即问即答，抑制了科学家的判断。我们建议将代理构建视为工作流程的一个阶段，并引入AgentBuild，它根据科学家撰写的合同构建科学代理。该合同是一个版本控制的评分标准、一个难度分级的课程以及一个策划的外部知识库。由评分标准驱动的评判者控制一个元优化器编码代理，该代理在声明的边界内编辑代理，从而使构建过程编译的是代理，而不是科学家的判断。我们通过GSAS-II在MCP和A2A的支持下，为X射线衍射数据的Rietveld精修实例化这一过程，其中一个空白的构建运行通过锂镧锆氧化物（LLZO）信噪比阶梯进展，达到4小时扫描作为前沿案例，并揭示了仍然存在的工作流程范围限制。相同的评分标准不仅奖励可信的拟合，还对轨迹范围进行评分，使得前沿成为合同失败而非模式拟合失败。随着基础模型的发展，重新运行AgentBuild是一次重新调优，而不是重建，科学家撰写的合同仍然是持久的资产。
cs.AI / 25 / 2606.12848
(Human) Attention Is (Still) All You Need: Human oversight makes AI-assisted social science reliable
（人类）注意力仍然是你所需要的一切：人类监督使得人工智能辅助的社会科学研究更为可靠
Abstract
Large language models (LLMs) are increasingly used for tasks once reserved for trained researchers, including hypothesis generation, specification choice, and drafting conclusions. We argue that the reliability of AI-assisted research depends not only on model capability, but also on how cognitive labour is structured between humans and machines. We study this problem through Human-in-the-Loop Economic Research (HLER), a decision architecture based on pre-commitment, decision sequencing, accountability, and attention allocation. In a pre-specified 2*4 factorial experiment with 280 complete research runs across four datasets, an unconstrained multi-agent baseline produced critical failures in 72% of runs. Using the same underlying model, the same agent decomposition, and identical prompts for the shared reasoning agents, HLER reduced the failure rate to 16% by imposing three architectural commitments: LLMs reason but do not execute data work, data and estimation are handled deterministically, and three human decision gates bind the workflow. Fisher's exact test rejects equality of failure rates at p<0.001. Reliability gains were largest on the least publicly represented dataset, a Qing-dynasty population register, consistent with a task-based production model with Frechet-distributed output quality. An 80-run ablation suggests that deterministic computation and human gates contribute independently, with exploratory evidence of complementarity. We interpret HLER as a research harness rather than an autonomous AI scientist: it sharply reduces failures, makes residual weaknesses more visible, and prevents unreliable claims from being advanced as publication-ready outputs.
Chinese Translation
大型语言模型（LLMs）越来越多地用于曾经只限于受过训练的研究人员的任务，包括假设生成、规范选择和撰写结论。我们认为，人工智能辅助研究的可靠性不仅依赖于模型的能力，还取决于人类与机器之间的认知劳动如何结构化。我们通过人机协作经济研究（Human-in-the-Loop Economic Research, HLER）来研究这个问题，这是一种基于预承诺、决策序列、问责制和注意力分配的决策架构。在一项预先指定的2*4因子实验中，针对四个数据集进行了280次完整的研究运行，结果显示，无约束的多智能体基线在72%的运行中出现了关键性失败。使用相同的基础模型、相同的智能体分解以及对共享推理智能体的相同提示，HLER通过施加三项架构承诺将失败率降低至16%：LLMs进行推理但不执行数据工作，数据和估计以确定性方式处理，以及三个由人类决策门控的工作流程。Fisher精确检验在p<0.001的水平上拒绝了失败率相等的假设。可靠性提升在公众代表性最少的数据集上最大，该数据集为清代人口登记册，这与基于任务的生产模型及Frechet分布的输出质量一致。一项80次运行的消融实验表明，确定性计算和人类门控独立贡献，并且有探索性证据表明它们之间存在互补性。我们将HLER解读为一种研究工具，而非自主的人工智能科学家：它显著减少了失败，使剩余的弱点更加明显，并防止不可靠的主张被作为可发表的成果提出。
cs.AI / 26 / 2606.12852
WISE: A Long-Horizon Agent in Minecraft with Why-Which Reasoning
WISE：在Minecraft中具有为何-哪个推理的长时间跨度智能体
Abstract
Rapid advances have been made in developing general-purpose embodied agent in environments like Minecraft through the adoption of LLM-augmented hierarchical approaches. Despite their promise, low-level controllers often become performance bottlenecks due to repeated execution failures. We argue that a key limitation is not only the lack of episodic memory, but also the decoupling of \textit{what-where-when} memory from \textit{which-why} reasoning. To address this, we propose \textbf{WISE} (Which-Why Informed Semantic Explorer), a long-horizon agent framework with an enhanced low-level controller equipped with a Causal Event Graph that augments episodic memory with explicit causal structure linking observations to task relevance. Unlike prior work such as MrSteve, which relies on feature similarity for retrieval, WISE enables robust recall under viewpoint changes and supports opportunistic task reordering through causal reasoning. Building on this memory, we propose an Opportunistic Task Scheduler that dynamically re-prioritizes subtasks when causally relevant opportunities are detected. We further equip WISE with a multi-scale progressive exploration strategy to provide spatially comprehensive observations for downstream reasoning. Experiments show that WISE largely improves task success and efficiency on long-horizon sparse tasks, particularly in settings requiring adaptive decision-making.
Chinese Translation
通过采用增强大型语言模型（LLM）的层次化方法，在Minecraft等环境中开发通用型具身智能体已取得快速进展。尽管这些方法前景广阔，但低级控制器往往因重复执行失败而成为性能瓶颈。我们认为，关键的限制不仅在于缺乏情节记忆，还在于 extit{什么-哪里-何时}记忆与 extit{哪个-为何}推理的解耦。为了解决这个问题，我们提出了 extbf{WISE}（Which-Why Informed Semantic Explorer），这是一个长时间跨度的智能体框架，配备了增强的低级控制器，使用因果事件图（Causal Event Graph）来增强情节记忆，并通过明确的因果结构将观察与任务相关性联系起来。与依赖特征相似性进行检索的先前工作（如MrSteve）不同，WISE能够在视角变化下实现稳健的回忆，并通过因果推理支持机会性任务重排序。在此记忆的基础上，我们提出了一种机会性任务调度器，当检测到因果相关的机会时，动态地重新优先考虑子任务。我们还为WISE配备了一种多尺度渐进探索策略，以提供空间上全面的观察，供下游推理使用。实验表明，WISE在长时间跨度的稀疏任务上显著提高了任务成功率和效率，尤其是在需要适应性决策的环境中。
cs.AI / 27 / 2606.12871
DailyReport: An Open-ended Benchmark for Evaluating Search Agents on Daily Search Tasks
DailyReport：评估搜索代理在日常搜索任务中的开放式基准
Abstract
Search Agents (SAs) typically leverage large language models (LLMs) to support complex information-seeking tasks by autonomously exploring web sources and synthesizing information into comprehensive responses. For SAs evaluation, prior benchmarks mainly focus on specialized tasks that are unlikely to arise in real-world user scenarios. Moreover, their reliance on coarse task-level rubrics often limits evaluation interpretability. To bridge this gap, we introduce DailyReport, an open-ended benchmark to evaluate SA capabilities on daily search tasks. It contains 150 open-ended tasks with 3,546 associated rubrics, capturing widely discussed and timely information demands of real-world users. Each task is decomposed into subtasks and evaluated with cascade rubrics across disentangled dimensions. Through cascade performance attribution and user-centric aggregation, we derive highly interpretable scores for each dimension, along with a user preference score. Our results on 17 agentic systems show that current systems still fall short of users' expectations. To facilitate future research, our dataset and code are made publicly available at https://github.com/AGI-Eval-Official/DailyReport.
Chinese Translation
搜索代理（Search Agents, SAs）通常利用大型语言模型（Large Language Models, LLMs）来支持复杂的信息检索任务，通过自主探索网络资源并将信息综合成全面的响应。对于搜索代理的评估，以往的基准主要集中在不太可能出现在现实用户场景中的专业任务上。此外，它们对粗略任务级评分标准的依赖往往限制了评估的可解释性。为了解决这一问题，我们提出了DailyReport，一个开放式基准，用于评估搜索代理在日常搜索任务中的能力。该基准包含150个开放式任务和3,546个相关评分标准，捕捉了现实用户广泛讨论和及时的信息需求。每个任务被分解为子任务，并通过跨解耦维度的级联评分标准进行评估。通过级联性能归因和以用户为中心的聚合，我们为每个维度得出了高度可解释的分数，以及一个用户偏好分数。我们在17个代理系统上的结果表明，目前的系统仍未达到用户的期望。为了促进未来的研究，我们的数据集和代码已公开发布在 https://github.com/AGI-Eval-Official/DailyReport。
cs.AI / 28 / 2606.12882
HarnessBridge: Learnable Bidirectional Controller for LLM Agent Harness
HarnessBridge：可学习的双向控制器用于大型语言模型代理的控制
Abstract
Large language models are increasingly deployed as agents for long-horizon tasks, yet their performance is shaped not only by model capability and environment design, but also by the harness that mediates agent--environment interaction. Existing harnesses are largely manually engineered, making them difficult to scale as trajectories grow longer and interactions become more complex. In this work, we ask whether harness can be generated by a learnable plug-in module that can be trained in an end-to-end fashion. We introduce HarnessBridge, a lightweight learnable harness controller that parameterizes the agent--environment interface as a bidirectional projection. HarnessBridge learns two bidirectional projections: observation projection, which distills raw trajectories into compact, decision-relevant states, and action projection, which converts proposed actions into executable transitions or trajectory-grounded rejections. We train HarnessBridge on a harness supervision dataset via unified instruction tuning. On Terminal-Bench~2.0 and SWE-bench Verified, HarnessBridge matches or surpasses strong specialized harnesses while substantially reducing token usage and trajectory length, and generalizes from smaller generators to larger commercial models.
Chinese Translation
大型语言模型越来越多地被用作长时间任务的代理，但它们的性能不仅受到模型能力和环境设计的影响，还受到调节代理与环境交互的控制器的影响。现有的控制器主要是手动设计的，这使得它们在轨迹变长和交互变得更加复杂时难以扩展。在本研究中，我们探讨是否可以通过一个可学习的插件模块生成控制器，该模块可以以端到端的方式进行训练。我们介绍了HarnessBridge，一个轻量级的可学习控制器，它将代理与环境的接口参数化为双向投影。HarnessBridge学习两个双向投影：观察投影，它将原始轨迹提炼为紧凑的、与决策相关的状态；以及动作投影，它将提议的动作转换为可执行的转变或基于轨迹的拒绝。我们通过统一的指令调优在控制器监督数据集上训练HarnessBridge。在Terminal-Bench~2.0和SWE-bench Verified上，HarnessBridge的表现与强大的专业控制器相匹配或超越，同时显著减少了令牌使用和轨迹长度，并且能够从较小的生成器推广到更大的商业模型。
cs.AI / 29 / 2606.12883
The Hidden Power of Scaling Factor in LoRA Optimization
LoRA优化中缩放因子的隐秘力量
Abstract
In Low-Rank Adaptation (LoRA), the scaling factor $\alpha$ is often treated as a mere complement to the learning rate, yet its role in optimization remains poorly understood. In this paper, we reveal that the scaling factor $\alpha$ and the learning rate function differently, with $\alpha$ emerging as the dominant driver of effective optimization, delivering gains that cannot be replicated by learning rate scaling alone. Through the synergy of extensive empirical analysis and a theoretical Signal-Drift framework, we uncover three findings into LoRA's scaling mechanism: First, LoRA's spectral suppression smooths the optimization landscape, rendering standard hyperparameters overly conservative and creating an optimization gap. Second, when leveraging this smoothness to accelerate convergence, $\alpha$ outperforms the learning rate by amplifying the task signal without increasing the drift ratio. Third, the optimal scaling factor follows a sublinear relationship with the rank, well characterized by a square-root law with an unexpectedly large coefficient, revealing the insufficient scaling of existing rank-tied heuristics. Based on these insights, we propose LoRA-$\alpha$, a minimalist framework that restores $\alpha$ to its principled regime, making LoRA compatible with standard small learning rates. Extensive evaluations across diverse tasks demonstrate that LoRA-$\alpha$ consistently improves performance while streamlining hyperparameter search, unleashing the learning potential of LoRA.
Chinese Translation
在低秩适应（Low-Rank Adaptation, LoRA）中，缩放因子 $eta$ 通常被视为学习率的补充，但其在优化中的作用仍然不够明确。本文揭示了缩放因子 $eta$ 与学习率在功能上的不同，$eta$ 成为有效优化的主要驱动因素，其带来的收益无法仅通过学习率缩放来复制。通过广泛的实证分析与理论信号漂移（Signal-Drift）框架的协同，我们揭示了LoRA缩放机制的三个发现：首先，LoRA的谱抑制平滑了优化景观，使得标准超参数过于保守，从而产生了优化差距。其次，在利用这种平滑性加速收敛时，$eta$ 通过增强任务信号而不增加漂移比率，超越了学习率的表现。第三，最优缩放因子与秩之间呈亚线性关系，符合平方根法则，并具有意外的大系数，揭示了现有秩绑定启发式方法的缩放不足。基于这些见解，我们提出了LoRA-$eta$，一个简约框架，将$eta$ 恢复到其原则性范围，使LoRA与标准的小学习率兼容。在多样任务上的广泛评估表明，LoRA-$eta$ 一直在提升性能的同时简化超参数搜索，释放了LoRA的学习潜力。
cs.AI / 30 / 2606.12900
Zero-source LLM Hallucination Detection with Human-like Criteria Probing
基于人类标准探测的零源大型语言模型幻觉检测
Abstract
Large language models (LLMs) often hallucinate by generating factually incorrect or unfaithful content, posing significant risks to their safe use. Detecting such hallucinations is particularly challenging under the zero-source constraint, where no model internals or external references are available, and detection must rely solely on the textual query-answer pair. In this paper, we propose Human-like Criteria Probing for Hallucination Detection (HCPD), a paradigm that emulates the multi-faceted reasoning of human evaluators. Its core is a Human-like Criteria Probing (HCP) mechanism, in which a LLM agent adaptively decomposes its judgment into a weighted set of interpretable criteria and aggregates criterion-specific scores into a final truthfulness measure. To achieve this adaptive capability, we introduce a reward-based alignment scheme using only weak supervision from semantic consistency. At inference, we employ a multi-sampling aggregation strategy to ensure robust decisions while preserving full interpretability. We further provide theoretical analysis supporting the reliability of our approach. Extensive experiments show that HCPD consistently outperforms state-of-the-art baselines, offering an effective and explainable solution for zero-source hallucination detection. Code is available at https://github.com/TRISKEL10N/HCPD.
Chinese Translation
大型语言模型（LLMs）常常通过生成事实不准确或不忠实的内容而产生幻觉，这对其安全使用构成了重大风险。在零源约束下，检测此类幻觉尤其具有挑战性，因为没有模型内部信息或外部参考，检测必须仅依赖于文本查询-回答对。本文提出了一种用于幻觉检测的人类标准探测（HCPD）范式，模拟人类评估者的多维推理。其核心是人类标准探测（HCP）机制，其中LLM代理自适应地将其判断分解为一组加权的可解释标准，并将标准特定的分数汇总为最终的真实性度量。为了实现这一自适应能力，我们引入了一种基于奖励的对齐方案，仅使用来自语义一致性的弱监督。在推理时，我们采用多采样聚合策略，以确保决策的稳健性，同时保持完全的可解释性。我们进一步提供了理论分析以支持我们方法的可靠性。大量实验表明，HCPD在零源幻觉检测中始终优于最先进的基线，提供了一种有效且可解释的解决方案。代码可在 https://github.com/TRISKEL10N/HCPD 获取。
cs.AI / 31 / 2606.12916
MDForge: Agentic Molecular Dynamics Pipeline Design under Sparse Simulator Feedback
MDForge：在稀疏模拟器反馈下的自主分子动力学管道设计
Abstract
Molecular dynamics (MD) is the canonical in-silico method for atomistic molecular science, simulating molecular behavior from first-principle physics. Designing an MD pipeline for a new system requires substantial expert knowledge: running it on even one molecule is expensive, ruling out trial-and-error. We automate this expert pipeline-design process with an LLM agent. Unlike existing MD agents that orchestrate a predefined tool set, we treat pipeline design as open-ended code generation in which the agent's behavior is reshaped online by verbal reward. Specifically, we build MDForge, an LLM agent whose in-context update rule densifies the sparse reward via a multi-agent debate among physics experts. On three SAMPL host-guest binding free-energy benchmarks, MDForge automatically designs MD pipelines competitive with human experts. Deployed on a library of unseen candidate guests, its CB[7] pipeline discovers a novel binder that wet-lab competition NMR confirms is a high-affinity, picomolar CB[7] binder. Our data and code are available at https://github.com/Zehong-Wang/MDForge.
Chinese Translation
分子动力学（MD）是原子分子科学的经典计算方法，通过第一性原理物理模拟分子的行为。为新系统设计MD管道需要大量的专家知识：即使在一个分子上运行也非常昂贵，这排除了试错法。我们通过一个大型语言模型（LLM）代理自动化这一专家管道设计过程。与现有的协调预定义工具集的MD代理不同，我们将管道设计视为开放式代码生成，其中代理的行为通过语言奖励在线重塑。具体而言，我们构建了MDForge，一个LLM代理，其上下文更新规则通过物理专家之间的多代理辩论来密集化稀疏奖励。在三个SAMPL宿主-客体结合自由能基准测试中，MDForge自动设计的MD管道与人类专家相竞争。在一个未见候选客体的库中部署后，其CB[7]管道发现了一种新型结合剂，湿实验竞争的核磁共振（NMR）确认其为高亲和力、皮摩尔级的CB[7]结合剂。我们的数据和代码可在https://github.com/Zehong-Wang/MDForge获取。
cs.AI / 32 / 2606.12924
Iterating Toward Better Search: A Two-Agent Simulation Framework for Evaluating Agentic Search Architectures in E-Commerce
迭代优化搜索：用于评估电子商务中代理搜索架构的双代理仿真框架
Abstract
We present a modular two-agent simulation framework for evaluating conversational shopping assistant architectures. An independent buyer agent, configured with personas, missions, and patience levels, is paired with an interchangeable responder that integrates with a real e-commerce search API. Holding the buyer constant across experiments enables controlled comparison of responder designs on identical scenarios. Using 2011 conversations across 14 persona buckets, we establish four empirical findings. First, rolling-window memory outperforms intent-extraction memory on all quality metrics while being 35% faster per query. Second, illustrating rapid evidence-driven iteration, a systematic failure analysis of a responder version enables targeted fixes that reduce failure and near-failure rates by 62% across the full dataset. Third, swapping the responder LLM backbone from Gemini~2.5 to Llama~3.3~70B costs 0.16--0.45 points despite identical architecture. Finally, we document systematic philosophical disagreement between frontier LLM judges: Gemini rewards process correctness while Claude demands concrete outcomes, despite using the same evaluation prompt.
Chinese Translation
我们提出了一种模块化的双代理仿真框架，用于评估对话式购物助手架构。一个独立的买方代理，配置有角色、任务和耐心水平，与一个可互换的响应者配对，该响应者与真实的电子商务搜索API集成。在实验中保持买方不变，使得在相同场景下对响应者设计进行可控比较成为可能。通过对2011年14个角色类别的对话进行分析，我们建立了四个实证发现。首先，滚动窗口记忆在所有质量指标上优于意图提取记忆，同时每个查询的速度快35%。其次，通过快速的基于证据的迭代，针对一个响应者版本的系统性失败分析使得能够进行有针对性的修复，从而将整个数据集的失败和接近失败率降低了62%。第三，将响应者的LLM（大语言模型）骨干从Gemini~2.5更换为Llama~3.3~70B，尽管架构相同，仍然导致了0.16到0.45点的损失。最后，我们记录了前沿LLM评审者之间的系统性哲学分歧：Gemini重视过程的正确性，而Claude则要求具体的结果，尽管使用的是相同的评估提示。
cs.AI / 33 / 2606.12935
MARS: Margin-Adversarial Risk-controlled Stopping for Parallel LLM Test-time Scaling
MARS：用于并行大语言模型测试时缩放的边际对抗风险控制停止
Abstract
Parallel test-time scaling samples many reasoning traces and majority-votes their answers, improving LLM accuracy but requiring traces to run to completion, incurring substantial computational overhead. We observe that probing partial traces at intermediate checkpoints can extract current answers without disrupting generation, revealing an evolving aggregate vote. Based on this observation, we introduce MARS, a margin-adversarial stopping rule that estimates which active traces are likely to change their answers and stops once the leader remains safe under a conservative bound on future vote movement. The rule separates two sources of uncertainty. It learns the trace-level switch probabilities that determine how much of the current margin is likely to be retained, while handling the harder question of where switching traces land through an adversarial bound calibrated from warmup traces. With true switch probabilities, MARS guarantees with high probability that the early-stopped answer matches the full-budget vote. In practice, a five-feature logistic model closely matches oracle switching behavior. Across three reasoning models and three competition-math benchmarks, MARS saves 25-47% of self-consistency tokens and 14-29% on top of DeepConf Online, a strong confidence-weighted baseline that already filters and truncates weak traces, while matching the accuracy of the corresponding full-budget baselines.
Chinese Translation
并行测试时缩放采样了许多推理轨迹并对其答案进行多数投票，从而提高了大语言模型（LLM）的准确性，但需要轨迹运行到完成，导致了可观的计算开销。我们观察到，在中间检查点探测部分轨迹可以提取当前答案，而不会干扰生成过程，从而揭示出一个不断演变的集体投票。基于这一观察，我们提出了MARS，一种边际对抗停止规则，估计哪些活跃轨迹可能会改变其答案，并在领导者在未来投票变动的保守界限下保持安全时停止。该规则将两种不确定性来源分开。它学习决定当前边际可能保留多少的轨迹级切换概率，同时处理更难的问题，即切换轨迹的落点，通过从热身轨迹中校准的对抗界限。通过真实的切换概率，MARS以高概率保证早期停止的答案与全预算投票相匹配。在实践中，一个五特征的逻辑模型与oracle切换行为密切匹配。在三个推理模型和三个竞争数学基准上，MARS节省了25-47%的自一致性令牌，并在DeepConf Online的基础上节省了14-29%，后者是一个强大的置信加权基线，已经过滤和截断了弱轨迹，同时与相应的全预算基线的准确性相匹配。
cs.AI / 34 / 2606.12942
PRISMR: Overcoming Parse Collapse in Multimodal Listwise Ranking via Parameterized Representation Internalization
PRISMR：通过参数化表示内化克服多模态列表排名中的解析崩溃
Abstract
Generative listwise ranking with Large Multimodal Models (LMMs) aims to capture global list context in a single forward pass, but its effectiveness degrades in long-context multimodal scenarios. We identify a recurring failure mode, parse collapse, where the autoregressive decoder produces fluent yet incomplete rankings by silently omitting candidates and terminating early. This failure stems from limited context utilization rather than simple formatting mistakes, making prompt engineering and constrained decoding insufficient. We propose PRISMR (Parameterized Representation Internalization for Semantic Multimodal Ranking), a framework that replaces transient in-context list processing with parametric structural conditioning. PRISMR uses a lightweight hypernetwork to encode multimodal candidates in parallel and generate item-specific LoRA weights, which are synthesized into an instance-specific adapter for a LMM. This paradigm enables more robust internalization of list structure while preserving the base model. We further introduce a large-scale multimodal review-ranking benchmark for evaluation. Experiments demonstrate that PRISMR substantially reduces parse collapse, improves listwise ranking performance, and transfers effectively across domains and instruction-tuned backbones.
Chinese Translation
生成性列表排名与大型多模态模型（LMMs）旨在通过单次前向传播捕捉全局列表上下文，但在长上下文多模态场景中，其有效性会下降。我们识别出一种反复出现的失败模式，即解析崩溃，其中自回归解码器生成流畅但不完整的排名，默默省略候选项并提前终止。这种失败源于有限的上下文利用，而非简单的格式错误，使得提示工程和受限解码显得不足。我们提出了PRISMR（语义多模态排名的参数化表示内化），该框架用参数化结构条件替代瞬态的上下文列表处理。PRISMR使用轻量级超网络并行编码多模态候选项，并生成特定项目的LoRA权重，这些权重被合成到特定实例的适配器中，以供LMM使用。这一范式在保留基础模型的同时，能够更稳健地内化列表结构。我们进一步引入了一个大规模多模态评论排名基准进行评估。实验表明，PRISMR显著减少了解析崩溃，提高了列表排名性能，并在不同领域和指令调优的基础模型之间有效迁移。
cs.AI / 35 / 2606.12945
Learning What to Remember: A Cognitively Grounded Multi-Factor Value Model for Agentic Memory
学习记忆内容：一种基于认知的多因素价值模型用于代理记忆
Abstract
Long-running LLM agents accumulate interaction histories far larger than any context window, forcing a standing decision: what to encode deeply, what to forget, and what to retrieve under a fixed memory budget. Production systems answer with semantic similarity or recency -- both mis-specified for the forgetting decision, which is made at consolidation time before the future query is known. We propose a multi-factor memory value function V(m)=\sum_i w_i f_i(m) over seven interpretable factors (emotional intensity, goal relevance, value alignment, self/user relevance, task utility, reliability, and usage history) drawn from cognitive psychology, whose weights are learned from a downstream objective by a gradient-free optimiser, and whose single scalar uniformly controls encoding depth, forget risk, and retrieval rank. We make a methodological point: on LongMemEval, scoring goal relevance against the held-out evaluation question saturates gold-evidence retention at \approx 0.98 -- this measures retrieval, not forgetting. In the realistic blind regime, a learned multi-factor value retains 0.770 \pm 0.011 of gold evidence across 479 usable cases, versus 0.657 for uniform weights, 0.518 for the best single factor, and 0.368 for recency; every paired gap's 95% bootstrap CI is above zero, and a neural network over the same factors ties the linear model. The learned weights are interpretable -- reliability, emotional intensity, and self/user relevance dominate, while query-time goal similarity is correctly down-weighted for the forgetting decision. A controlled synthetic task with planted confounds confirms the learner recovers a separating weighting (1.00 retention) where uniform weighting fails (0.62). The substrate is open-source; all experiments run on a single CPU with no API calls.
Chinese Translation
长期运行的LLM代理积累的交互历史远大于任何上下文窗口，这迫使我们做出一个持续的决策：深度编码什么，遗忘什么，以及在固定的记忆预算下检索什么。生产系统通过语义相似性或近期性来回答——这两者对于遗忘决策的指定都是错误的，因为该决策是在巩固时做出的，而未来的查询尚未确定。我们提出了一个多因素记忆价值函数V(m)=∑_i w_i f_i(m)，该函数基于来自认知心理学的七个可解释因素（情感强度、目标相关性、价值一致性、自我/用户相关性、任务效用、可靠性和使用历史），其权重由无梯度优化器从下游目标中学习，并且该单一标量均匀控制编码深度、遗忘风险和检索排名。我们提出一个方法论观点：在LongMemEval上，针对保留的评估问题评估目标相关性会使黄金证据的保留饱和至约0.98——这测量的是检索，而非遗忘。在现实的盲态下，学习的多因素价值在479个可用案例中保留了0.770 ± 0.011的黄金证据，而均匀权重为0.657，最佳单因素为0.518，近期性为0.368；每个配对差距的95%自助置信区间均高于零，而在相同因素上的神经网络与线性模型相当。学习到的权重是可解释的——可靠性、情感强度和自我/用户相关性占主导地位，而查询时的目标相似性在遗忘决策中被正确地降低权重。一个带有植入混淆的受控合成任务确认学习者在均匀加权失败（0.62）时恢复了分离加权（1.00保留）。该基础设施是开源的；所有实验均在单个CPU上运行，无需API调用。
cs.AI / 36 / 2606.12953
OpenMedQ: Broad Open Pretraining for Medical Vision-Language Models
OpenMedQ：广泛开放的医学视觉-语言模型预训练
Abstract
We present OpenMedQ, a medical vision-language model pretrained on the broadest fully-open medical mix to date: 14 datasets totaling ~3.35M pretraining samples spanning pathology, radiology, microscopy, and text-only clinical QA. OpenMedQ reaches state-of-the-art BLEU-1 on PathVQA (75.9), beating Med-PaLM M variants up to 562B parameters (~80x larger), and matches the best reported VQA-MED BLEU-1 (64.5). Its vision encoder, transferred to 8 unseen medical classification benchmarks under an identical downstream recipe, obtains the highest average macro-F1 (0.757) among BiomedCLIP (0.745), PMC-CLIP (0.745), PubMedCLIP (0.746), and a from-scratch baseline (0.616). We release our code and an interactive demo is publicly available as a reproducible baseline for the community.
Chinese Translation
我们提出了OpenMedQ，这是一个在迄今为止最广泛的完全开放医学数据集上进行预训练的医学视觉-语言模型：14个数据集，总计约335万条预训练样本，涵盖病理学、放射学、显微镜学和仅文本的临床问答。OpenMedQ在PathVQA上达到了最先进的BLEU-1分数（75.9），超越了参数量高达562B的Med-PaLM M变体（约80倍大），并且与报告的最佳VQA-MED BLEU-1（64.5）持平。其视觉编码器在相同下游任务设置下转移到8个未见过的医学分类基准上，获得了最高的平均宏F1分数（0.757），优于BiomedCLIP（0.745）、PMC-CLIP（0.745）、PubMedCLIP（0.746）和一个从零开始的基线（0.616）。我们发布了我们的代码，并提供了一个互动演示，作为社区可重复的基线。
cs.AI / 37 / 2606.12969
Multi-Modal Agents for Power Distribution Defect Detection: An Evaluation of Foundation Models
用于电力分配缺陷检测的多模态智能体：基础模型的评估
Abstract
The power distribution network is critical to reliable electricity delivery, yet traditional inspection methods face limitations in semantic understanding, generalization, and closed-loop automation. To address these challenges, this paper proposes a Multi-Modal Agent framework specifically for power distribution defect detection. Central to this study is the systematic evaluation of multimodal foundation models as unified cognitive engines. We rigorously assess their integrated performance across three critical capabilities: (1) Perception, where the model must accurately identify equipment and generate expert-level descriptions of defects; (2) Reasoning, where the model interprets visual findings to diagnose causes, assess severity, and plan maintenance strategies based on domain knowledge; and (3) Tool Usage, where the model acts as an autonomous operator to execute actions -- such as querying knowledge bases or generating work orders -- to achieve closed-loop maintenance. To support this evaluation, a domain-specific evaluation dataset and a comprehensive benchmark are developed. Experimental results demonstrate the strengths and limitations of current foundation models in these three dimensions, providing empirical evidence for deploying autonomous agents in high-stakes industrial environments.
Chinese Translation
电力分配网络对可靠的电力传输至关重要，但传统的检查方法在语义理解、泛化能力和闭环自动化方面存在局限性。为了解决这些挑战，本文提出了一种专门用于电力分配缺陷检测的多模态智能体框架。本研究的核心是对多模态基础模型作为统一认知引擎的系统评估。我们严格评估它们在三个关键能力上的综合表现：(1) 感知，模型必须准确识别设备并生成专家级的缺陷描述；(2) 推理，模型解释视觉发现以诊断原因、评估严重性并基于领域知识规划维护策略；(3) 工具使用，模型作为自主操作员执行行动——例如查询知识库或生成工作订单——以实现闭环维护。为支持这一评估，开发了一个特定领域的评估数据集和全面的基准。实验结果展示了当前基础模型在这三个维度上的优势和局限性，为在高风险工业环境中部署自主智能体提供了实证依据。
cs.AI / 38 / 2606.12976
A Mathematical Forum Platform for Collaborative Problem Solving and Dataset Generation for AI Reasoning
一个用于协作问题解决和AI推理数据集生成的数学论坛平台
Abstract
Sharing mathematical content in online forums remains a significant friction point for students and educators: writing raw LATEX is error-prone, standalone optical character recognition tools require platform switching, and current forum software offers no integrated path from a photograph of a formula to a rendered post. We present a unified system that eliminates this friction by embedding an image to LATEX conversion pipeline directly inside a forum posting interface. A user uploads or captures an image of a mathematical expression; the system routes it through the Mathpix OCR API, detects whether the returned output is LATEX or plain text containing inline math, applies the appropriate delimiter normalisation, and renders a live preview in either LATEX or Markdown mode before the post is committed to the database. The architecture is organized in three loosely coupled layers: image processing, rendering, and storage, and supports both desktop and mobile clients. A provisional US patent application has been filed covering the core methods. We describe the full system design, each component in detail, the data schema, and the key technical innovations, and we position the work against existing standalone tools and forum platforms to demonstrate the practical gap it closes. Beyond immediate usability, we argue that a deployed platform of this kind constitutes a continuously growing, community-validated dataset of mathematical problems and step-by-step solutions, a resource that can be used to train and benchmark AI systems for accurate mathematical reasoning
Chinese Translation
在在线论坛中分享数学内容仍然是学生和教育工作者面临的一个重要障碍：编写原始LATEX容易出错，独立的光学字符识别工具需要切换平台，而当前的论坛软件没有提供从公式照片到渲染帖子的一体化路径。我们提出了一个统一系统，通过在论坛发布界面中直接嵌入图像到LATEX的转换流程来消除这一障碍。用户可以上传或捕捉数学表达式的图像；系统通过Mathpix OCR API进行处理，检测返回的输出是LATEX还是包含行内数学的纯文本，应用适当的分隔符规范化，并在提交帖子到数据库之前以LATEX或Markdown模式渲染实时预览。该架构分为三个松耦合的层次：图像处理、渲染和存储，并支持桌面和移动客户端。我们已提交了一项临时的美国专利申请，涵盖核心方法。我们详细描述了完整的系统设计、每个组件、数据架构以及关键技术创新，并将该工作与现有的独立工具和论坛平台进行对比，以展示其填补的实际空白。除了即时的可用性外，我们认为，这种已部署的平台构成了一个不断增长的、经过社区验证的数学问题和逐步解决方案的数据集，这一资源可用于训练和基准测试AI系统，以实现准确的数学推理。
cs.AI / 39 / 2606.12983
Structured Testbench Generation for LLM-Driven HDL Design and Verification-Oriented Data Curation
面向LLM驱动的HDL设计与验证的数据整理的结构化测试平台生成
Abstract
Automated testbench generation has become a critical bottleneck in large language model (LLM)-driven Register Transfer Level (RTL) workflows, where large numbers of candidate designs must be verified rapidly and reliably. Existing prompt-based approaches treat testbench generation as unconstrained code synthesis, yielding stochastic outputs with high token cost, low reproducibility, and insufficient coverage. To address this gap, we present STG, a Structured Testbench Generation framework that exploits the inherent structure of hardware designs to generate deterministic testbenches. As a direct verification tool, STG runs 720x faster than an iterative LLM-based testbench generation flow and higher rate of successful compilation, achieves higher coverage, and reduces false-pass verdicts on incorrect DUTs. STG also helps identify errors in RTL generation benchmarks by exposing faulty benchmark testbenches. As a data curation engine, it is 11x faster than LLM-based filtering on a single CPU core with 127x less energy, and the resulting distilled models provide state-of-the-art performance in our multi-benchmark evaluation. As a test-time scaling oracle, it reduces node count by 14-47\%. Our models are available at https://huggingface.co/collections/AS-SiliconMind/siliconmind-v12.
Chinese Translation
自动化测试平台生成已成为大型语言模型（LLM）驱动的寄存器传输级（RTL）工作流程中的一个关键瓶颈，在这些工作流程中，必须快速且可靠地验证大量候选设计。现有的基于提示的方法将测试平台生成视为不受限制的代码合成，导致输出具有随机性，代价高昂，重现性差，覆盖率不足。为了解决这一问题，我们提出了STG，一个结构化测试平台生成框架，利用硬件设计的固有结构生成确定性测试平台。作为一种直接的验证工具，STG的运行速度比基于LLM的迭代测试平台生成流程快720倍，成功编译的比率更高，覆盖率更高，并减少了对错误被测单元（DUT）的错误通过判决。STG还通过暴露有缺陷的基准测试平台，帮助识别RTL生成基准中的错误。作为一个数据整理引擎，它在单个CPU核心上的速度比基于LLM的过滤快11倍，能耗减少127倍，所生成的精简模型在我们的多基准评估中提供了最先进的性能。作为测试时的扩展预言机，它将节点数量减少了14-47%。我们的模型可在https://huggingface.co/collections/AS-SiliconMind/siliconmind-v12获取。
cs.AI / 40 / 2606.12991
APCyc: Property-Informed Design of Cyclic Peptides via Automated Cyclization
APCyc：通过自动环化实现基于属性的信息化环肽设计
Abstract
Cyclic peptides represent a promising class of therapeutic compounds in modern drug discovery, often offering improved stability and binding affinity. However, the de novo design of cyclic peptides remains challenging because methods must identify pocket-adaptive cyclization patterns and linkage sites while simultaneously controlling drug-relevant properties. This challenge is particularly pronounced for recent generative models trained predominantly on linear peptide data, which may fail to capture cyclization-specific constraints. To address the limitation, we introduce APCyc, a target-aware de novo cyclic peptide generation framework that explicitly models cyclization and jointly optimizes multiple essential physicochemical properties. By using an expanded residue vocabulary and explicitly encoding cyclization-site and linkage-type information, APCyc learns cyclization-aware representations and leverages Bayesian posterior guidance to steer sampling toward cyclic peptides satisfying multiple property objectives. Experimental results demonstrate that our model learns target-dependent cyclization preferences, and enables effective and controllable multi-property optimization for cyclic peptide design. The source code of this paper is available at https://github.com/HKUSTGZ-ML4Health-Lab/APCyc.
Chinese Translation
环肽在现代药物发现中代表了一类有前景的治疗化合物，通常提供更好的稳定性和结合亲和力。然而，环肽的从头设计仍然具有挑战性，因为方法必须识别适应口袋的环化模式和连接位点，同时控制与药物相关的属性。这个挑战在最近主要基于线性肽数据训练的生成模型中尤为明显，因为这些模型可能无法捕捉特定于环化的约束。为了解决这一限制，我们提出了APCyc，一个目标感知的从头环肽生成框架，明确建模环化并联合优化多个基本的物理化学属性。通过使用扩展的残基词汇，并明确编码环化位点和连接类型信息，APCyc学习环化感知的表示，并利用贝叶斯后验引导将采样引导向满足多个属性目标的环肽。实验结果表明，我们的模型学习了目标依赖的环化偏好，并实现了环肽设计的有效且可控的多属性优化。本文的源代码可在 https://github.com/HKUSTGZ-ML4Health-Lab/APCyc 获取。
cs.AI / 41 / 2606.13003
The Illusion of Multi-Agent Advantage
多智能体优势的幻觉
Abstract
Prevailing wisdom posits that Multi-Agent Systems (MAS) are superior to Single-Agent Systems (SAS), citing advantages like context protection, parallel processing and distributed decision-making. However, empirical support for this claim relies primarily on comparisons with SAS baselines using benchmarks that prioritize isolated reasoning tasks, which do not adequately assess these advantages. Focusing on automatically generated MAS that are designed for enhanced generalizability over manually-designed counterparts, we perform a rigorous, systematic evaluation against SAS, specifically Chain-of-Thought with Self-Consistency (CoT-SC). Across traditional reasoning datasets and tasks with interactive multi-step workflows (e.g., BrowseComp-Plus), we demonstrate that automatic MAS consistently underperform CoT-SC despite being up to 10x more expensive. To isolate these failures from limitations inherent to task structure, we introduce a diagnostic synthetic dataset tailored for MAS featuring explicit task decomposition, context separation and parallelization potential. We show that expert-architected MAS consistently outperforms automatically generated architectures in both raw performance and cost-efficiency on this dataset, demonstrating that existing evaluation frameworks mask critical architectural gaps and inefficiencies of complex MAS by failing to account for the marginal utility of increased computational cost. Critically, systematic deconstruction of the generated MAS architectures reveals that current automated design paradigms produce architectural bloat that prioritizes superficial complexity which does not translate into functional utility, exposing a fundamental misalignment with multi-agent principles.
Chinese Translation
现有的观点认为，多智能体系统（MAS）优于单智能体系统（SAS），其优势包括上下文保护、并行处理和分布式决策。然而，这一说法的实证支持主要依赖于与SAS基准的比较，这些基准优先考虑孤立推理任务，未能充分评估这些优势。我们专注于自动生成的MAS，这些系统旨在增强可泛化性，相较于手动设计的系统，我们对SAS进行了严格的系统评估，特别是链式思维与自我一致性（CoT-SC）。在传统推理数据集和具有交互式多步骤工作流程的任务（例如，BrowseComp-Plus）中，我们证明自动生成的MAS在性能上始终低于CoT-SC，尽管其成本高达10倍。为了将这些失败与任务结构固有的限制隔离开来，我们引入了一个针对MAS的诊断合成数据集，该数据集具有明确的任务分解、上下文分离和并行化潜力。我们展示了专家设计的MAS在该数据集上在原始性能和成本效率上始终优于自动生成的架构，表明现有的评估框架掩盖了复杂MAS的关键架构缺陷和低效性，因为它未能考虑到增加计算成本的边际效用。关键的是，对生成的MAS架构进行系统性解构揭示了当前的自动设计范式产生了架构膨胀，优先考虑表面复杂性而未能转化为功能效用，暴露了与多智能体原则之间的根本不一致。
cs.AI / 42 / 2606.13016
Otters++: A Time-to-first-spike Based Energy Efficient Optical Spiking Transformer
Otters++：基于首次脉冲时间的能效光脉冲变换器
Abstract
Spiking neural networks (SNNs) are promising for energy-efficient inference, and time-to-first-spike (TTFS) coding is especially attractive because each neuron fires at most once. In practice, however, this benefit is often reduced by the cost of computing a temporal decay term and multiplying it by the synaptic weight. We address this issue by turning a physical hardware "bug," the natural signal decay in optoelectronic devices, into the main computation of TTFS, named Otters++. Specifically, we use the measured decay of a custom In$_2$O$_3$ optoelectronic synapse to directly realize the TTFS temporal term, removing the need for explicit digital decay computation. To scale this idea to Transformer models, we establish a layer-wise functional equivalence between the Otters++ and a quantized neural network (QNN), and develop a hybrid training method that uses device-faithful SNN computation in the forward pass and QNN straight-through gradients through the equivalent QNN path in the backward pass, together with model distillation. This avoids differentiation through discrete first-spike events and reduces the over-sparsity problem in direct TTFS-SNN training. We further make training aware of measured device noise by sampling run-to-run variation, and refine the system-level energy model by accounting for device sharing and multi-hop communication. On GLUE dataset, Otters++ improves the average score to 84.17\% while maintaining a clear energy advantage over prior spiking Transformer baselines. These results show that physically grounded TTFS computing can be efficient, trainable, and robust under realistic hardware effects.
Chinese Translation
脉冲神经网络（SNNs）在能效推理方面展现出良好的前景，而首次脉冲时间（TTFS）编码尤其具有吸引力，因为每个神经元最多只发射一次。然而，在实际应用中，这一优势常常因计算时间衰减项及其与突触权重相乘的成本而降低。我们通过将物理硬件的“缺陷”——光电设备中的自然信号衰减——转变为TTFS的主要计算，提出了解决方案，命名为Otters++。具体而言，我们利用定制的In$_2$O$_3$光电突触的测量衰减直接实现TTFS时间项，从而消除了显式数字衰减计算的需求。为了将这一思路扩展到变换器模型，我们建立了Otters++与量化神经网络（QNN）之间的逐层功能等价关系，并开发了一种混合训练方法，该方法在前向传播中使用设备忠实的SNN计算，在后向传播中通过等效QNN路径使用QNN直通梯度，并结合模型蒸馏。这避免了对离散首次脉冲事件的微分，并减少了直接TTFS-SNN训练中的过度稀疏问题。我们进一步通过采样运行间变异来使训练考虑测量的设备噪声，并通过考虑设备共享和多跳通信来完善系统级能量模型。在GLUE数据集上，Otters++将平均得分提高至84.17\%，同时在能量上明显优于之前的脉冲变换器基线。这些结果表明，基于物理的TTFS计算在现实硬件影响下可以高效、可训练且具有鲁棒性。
cs.AI / 43 / 2606.13020
SciR: A Controllable Benchmark for Scientific Reasoning in LLMs
SciR：一个可控的科学推理基准用于大型语言模型
Abstract
Three paradigmatic forms of inference recur across scientific reasoning: deduction, induction, and causal abduction. Reliably evaluating LLMs on these in scientific settings is currently out of reach: scientific benchmarks built on human annotations are costly and lack mechanistic ground truth, while synthetic logical-reasoning benchmarks do not resemble real scientific documents. We introduce SciR, a benchmark that combines multi-paradigm reasoning with controllable scientific rendering, anchored on three paradigmatic scientific problems. Tasks are generated from formal objects (deduction tree, inductive rule hypothesis, causal graph) to guarantee verifiable answers, then rendered into multi-document scientific discourse via per-track domain-tuned genres. The construction lets us independently vary two difficulty axes: how hard it is to extract the key information needed for inference, and how hard the principled inference itself is. We test six models. Both axes hurt every model, and their effects compound. The rendering even hurts neurosymbolic pipelines, which hand inference to a verified solver. The two axes yield a per-model extraction-vs-inference profile: for instance, reasoning models like deepseek-r1 mostly surpass non-reasoning instruct models on the inference axis. To our knowledge, SciR is the first multi-paradigm scientific-reasoning benchmark with parametric control on both extraction and inference difficulty.
Chinese Translation
科学推理中反复出现三种典型的推理形式：演绎、归纳和因果推理。当前在科学环境中可靠地评估大型语言模型（LLMs）仍然无法实现：基于人类注释的科学基准成本高昂且缺乏机械基础真相，而合成的逻辑推理基准又与真实的科学文献不相似。我们提出了SciR，一个结合多范式推理与可控科学呈现的基准，基于三种典型的科学问题生成任务。任务从形式对象（演绎树、归纳规则假设、因果图）生成，以确保可验证的答案，然后通过每个领域调优的体裁呈现为多文档科学话语。这种构建方式使我们能够独立变化两个难度轴：提取推理所需关键信息的难度，以及推理本身的难度。我们测试了六个模型。这两个难度轴对每个模型都有负面影响，并且其影响是累积的。呈现甚至对神经符号管道造成了负面影响，这些管道将推理交给经过验证的求解器。这两个轴产生了每个模型的提取与推理特征：例如，像deepseek-r1这样的推理模型在推理轴上大多超越了非推理指令模型。据我们所知，SciR是第一个在提取和推理难度上具有参数控制的多范式科学推理基准。
cs.AI / 44 / 2606.13038
Nous: An Attempt to Extract and Inject the Cognition Behind Prediction-Market Behavior
Nous：尝试提取和注入预测市场行为背后的认知
Abstract
As LLM agents proliferate in prediction markets and collective decision-making, they risk a cognitive monoculture: agents built on shared foundation models produce correlated forecasts, and recent measurement finds frontier-model errors correlated at r ~ 0.77. We ask whether human cognitive diversity can be recovered from behavior and transferred to LLM agents. Nous extracts a structured eight-dimension behavioral profile from real Polymarket trading activity and injects it into agents through prompts. Our central finding is a dissociation between the two halves of that pipeline. Extraction works, partially: across 100 wallets, 8 of 14 parameters are temporally stable (split-half ICC >= 0.5, bootstrap CI lower bound > 0.3; contrarian score reaches ICC ~ 0.9); wallets are identifiable from their profiles well above chance (top-1 retrieval 17-22% vs. 1% chance); and two of four pre-specified dimensions rank-correlate with future realized profit out-of-sample, though the correlations do not survive behavioral-confound controls. Prompt-level injection does not measurably transmit it: on a semantic embedding metric, structured injection shows no significant advantage over a length-matched control on any model, and the diversity it induces neither reduces ensemble error correlation nor improves Brier score -- a null that persists across exploratory checks on sampling temperature, profile diversity, and question difficulty. Measuring the prompts themselves locates the compression before the model: the structure-to-narrative translator emits near-uniform prompts whose spread does not track profile spread. We position Nous as measuring the cognitive-monoculture problem and the limits of a prompt-level remedy, motivating deeper, below-the-prompt injection (fine-tuning, activation steering). Code, frozen profiles, prompts, and model outputs: https://github.com/WillChienT/nous-paper
Chinese Translation
随着大型语言模型（LLM）代理在预测市场和集体决策中的普及，它们面临着认知单一化的风险：基于共享基础模型构建的代理产生相关的预测，最近的测量发现前沿模型的错误相关性达到 r ~ 0.77。我们探讨人类认知多样性是否可以从行为中恢复并转移到 LLM 代理中。Nous 从真实的 Polymarket 交易活动中提取了一个结构化的八维行为特征，并通过提示将其注入到代理中。我们的核心发现是该流程的两个部分之间存在解耦。提取部分部分有效：在 100 个钱包中，14 个参数中有 8 个在时间上是稳定的（分半 ICC >= 0.5，bootstrap CI 下限 > 0.3；逆向得分达到 ICC ~ 0.9）；钱包可以通过其特征在显著高于偶然水平的情况下被识别（前 1 检索率为 17-22% 对比 1% 的偶然率）；并且四个预设维度中的两个与未来实现的利润在样本外的等级相关性存在相关性，尽管这些相关性在行为混淆控制下并不成立。提示级别的注入并未显著传递认知：在语义嵌入指标上，结构化注入在任何模型上都未显示出相对于长度匹配控制的显著优势，并且它所引发的多样性既未减少集成错误相关性，也未改善 Brier 评分——这一无效结果在对采样温度、特征多样性和问题难度的探索性检查中持续存在。对提示本身的测量发现压缩发生在模型之前：结构到叙述的翻译器发出近乎均匀的提示，其分布与特征分布并不相关。我们将 Nous 定位为测量认知单一化问题及其提示级别补救的局限性，激励更深层次的、低于提示的注入（微调、激活引导）。代码、冻结特征、提示和模型输出可在此获取：https://github.com/WillChienT/nous-paper
cs.AI / 45 / 2606.13042
Augmentation techniques for video surveillance in the visible and thermal spectral range
可见光与热红外光谱范围内的视频监控增强技术
Abstract
In intelligent video surveillance, cameras record image sequences during day and night. Commonly, this demands different sensors. To achieve a better performance it is not unusual to combine them. We focus on the case that a long-wave infrared camera records continuously and in addition to this, another camera records in the visible spectral range during daytime and an intelligent algorithm supervises the picked up imagery. More accurate, our task is multispectral CNN-based object detection. At first glance, images originating from the visible spectral range differ between thermal infrared ones in the presence of color and distinct texture information on the one hand and in not containing information about thermal radiation that emits from objects on the other hand. Although color can provide valuable information for classification tasks, effects such as varying illumination and specialties of different sensors still represent significant problems. Anyway, obtaining sufficient and practical thermal infrared datasets for training a deep neural network poses still a challenge. That is the reason why training with the help of data from the visible spectral range could be advantageous, particularly if the data, which has to be evaluated contains both visible and infrared data. However, there is no clear evidence of how strongly variations in thermal radiation, shape, or color information influence classification accuracy. To gain deeper insight into how Convolutional Neural Networks make decisions and what they learn from different sensor input data, we investigate the suitability and robustness of different augmentation techniques...
Chinese Translation
在智能视频监控中，摄像头在白天和夜间记录图像序列。通常，这需要不同的传感器。为了实现更好的性能，结合使用这些传感器并不罕见。我们关注的情况是，一个长波红外摄像头持续记录，同时另一个摄像头在白天记录可见光谱范围内的图像，并且一个智能算法对捕获的图像进行监督。更准确地说，我们的任务是基于多光谱卷积神经网络（CNN）的目标检测。乍一看，来自可见光谱范围的图像与热红外图像在颜色和明显纹理信息的存在上有所不同，而在不包含从物体发出的热辐射信息这一点上也有所不同。尽管颜色可以为分类任务提供有价值的信息，但不同传感器的照明变化和特殊性仍然构成显著问题。无论如何，获取足够且实用的热红外数据集以训练深度神经网络仍然是一个挑战。这就是为什么利用可见光谱范围的数据进行训练可能是有利的，特别是当需要评估的数据同时包含可见光和红外数据时。然而，目前尚无明确证据表明热辐射、形状或颜色信息的变化对分类准确性有多大影响。为了深入了解卷积神经网络如何做出决策以及它们从不同传感器输入数据中学习到什么，我们研究了不同增强技术的适用性和鲁棒性。
cs.AI / 46 / 2606.13051
AAbAAC: An Annotated Corpus for Autoimmunity Information Extraction
AAbAAC：用于自身免疫信息提取的注释语料库
Abstract
Despite advances in information extraction driven by deep learning and large language models, performance gaps remain in highly specialized biomedical fields, where domainspecific complexity poses challenges for generalist models. In this work, we focus on the domain of autoimmunity, where the main entities of interest are autoimmune diseases, autoantibodies (i.e., molecules that may mark or cause these diseases), their molecular targets, their location in the body, and their associated clinical signs. Herein, we present AAbAAC (AutoAntibodies and Autoimmunity Annotated Corpus), a corpus of 115 abstracts selected from PubMed, where we manually annotated entities and their relationships. First, AAbAAC was used to evaluate several methods on the task of named entity recognition (NER), and secondly, to fine-tune NER models. Our study demonstrates the utility of AAbAAC for information extraction in the domain of autoimmunity, showing expected improvement in NER performance after finetuning. This illustrates the value of small-scale annotation efforts for specialized domains and contributes to the computational study of autoimmunity. The AAbAAC corpus is available at https://github.com/f-maury/AAbAAC.
Chinese Translation
尽管深度学习和大型语言模型推动了信息提取的进展，但在高度专业化的生物医学领域，性能差距仍然存在，这些领域的特定复杂性给通用模型带来了挑战。在本研究中，我们聚焦于自身免疫领域，其中主要关注的实体包括自身免疫疾病、自身抗体（即可能标记或引发这些疾病的分子）、它们的分子靶点、在体内的位置以及相关的临床症状。在此，我们呈现了AAbAAC（自身抗体与自身免疫注释语料库），该语料库包含从PubMed中选取的115篇摘要，我们手动注释了实体及其关系。首先，AAbAAC被用于评估几种命名实体识别（NER）任务的方法，其次，用于微调NER模型。我们的研究展示了AAbAAC在自身免疫领域信息提取中的实用性，微调后NER性能的预期提升证明了这一点。这说明了小规模注释工作在专业领域中的价值，并为自身免疫的计算研究做出了贡献。AAbAAC语料库可在https://github.com/f-maury/AAbAAC获取。
cs.AI / 47 / 2606.13141
Rethinking RAG in Long Videos: What to Retrieve and How to Use It?
重新思考长视频中的RAG：该检索什么以及如何使用？
Abstract
Retrieval-augmented generation is moving beyond text into long, egocentric video, where systems must select query-relevant chunks across multiple modalities and temporal granularities. Yet progress in VideoRAG is limited by two gaps: existing benchmarks allow queries to be answered without the video, obscuring retrieval errors, and prior methods apply a single modality-granularity configuration per query, ignoring chunk-level variability. We address both by introducing V-RAGBench, a benchmark of $\langle$query, evidence chunk, answer$\rangle$ triplets that enables faithful, decoupled evaluation of retrieval and generation, and CARVE, a simple method that runs parallel retrievers across configurations and employs chunk-adaptive reranking to identify the winning configuration for each chunk. Each chunk then enters the generator under its winning configuration selected during retrieval, yielding an interleaved evidence form where the chunk-level decision propagates across both stages. CARVE outperforms eight recent VideoRAG baselines, with the chunks supplied to the generator interleaving multiple configurations rather than sharing a single one, a behavior unattainable by query-level methods.
Chinese Translation
检索增强生成正在从文本扩展到长时间的自我中心视频，其中系统必须选择跨多个模态和时间粒度的查询相关片段。然而，VideoRAG的进展受到两个方面的限制：现有基准允许在没有视频的情况下回答查询，从而掩盖了检索错误；而之前的方法对每个查询应用单一的模态-粒度配置，忽视了片段级的变异性。我们通过引入V-RAGBench来解决这两个问题，V-RAGBench是一个包含$ extlangle$query, evidence chunk, answer$ extrangle$三元组的基准，能够实现检索和生成的真实、解耦评估；同时提出了CARVE，这是一种简单的方法，通过在不同配置中并行运行检索器，并采用片段自适应重排序来识别每个片段的最佳配置。每个片段随后在检索过程中选择的最佳配置下进入生成器，形成一种交错的证据形式，使得片段级决策在两个阶段之间传播。CARVE的表现优于八个近期的VideoRAG基线，提供给生成器的片段交错了多种配置，而不是共享单一配置，这种行为是查询级方法无法实现的。
cs.AI / 48 / 2606.13148
TerraBench: Can Agents Reason Over Heterogeneous Earth-System Data?
TerraBench：智能体能否对异构地球系统数据进行推理？
Abstract
Climate and environmental decision-making increasingly requires reasoning across heterogeneous inputs, including gridded physical data, satellite imagery, geospatial context, and simulator outputs. Weather and climate foundation models can forecast well, but do not reason interactively in language, while large language models (LLMs) reason in language but cannot operate directly on high-dimensional Earth-system data. As a result, real scientific workflows in Earth-science remain underserved. We introduce TerraBench, a benchmark for grounded Earth-science reasoning, built on TerraAgent, a ReAct-style executable framework that interleaves reasoning, tool calls, and observations to couple LLM planning with scientific tools for environmental retrieval, geospatial processing, simulation, and artifact-backed computation. TerraBench unifies analysis of Earth observation imagery, gridded data, GIS reasoning and simulation in a single executable interface, whereas prior benchmarks isolate these capabilities into narrow individual tasks. It is also the first in this space to pair process-level tool-use metrics with tolerance-aware numeric scoring. The benchmark comprises 403 extensive agentic tasks across three tracks (Fundamentals, Simulator-Grounded, and Document-Grounded Verification) and eight application domains with 24,500 verified execution steps. These results indicate that reliable Earth-science agents must go beyond tool access to coordinate heterogeneous workflows, parameterize tools precisely, and preserve artifact provenance.
Chinese Translation
气候和环境决策日益需要在异构输入之间进行推理，这些输入包括网格化的物理数据、卫星图像、地理空间上下文和模拟器输出。天气和气候基础模型能够进行良好的预测，但无法在语言中进行交互式推理，而大型语言模型（LLMs）能够进行语言推理，但无法直接处理高维的地球系统数据。因此，地球科学中的实际科学工作流程仍然得不到充分服务。我们介绍了TerraBench，这是一个基于TerraAgent的地球科学推理基准，TerraAgent是一个ReAct风格的可执行框架，它交错了推理、工具调用和观察，以将LLM规划与环境检索、地理空间处理、模拟和基于文献的计算的科学工具结合起来。TerraBench将地球观测图像、网格数据、地理信息系统（GIS）推理和模拟的分析统一在一个可执行接口中，而之前的基准将这些能力孤立为狭窄的单一任务。它也是该领域首个将过程级工具使用指标与容忍度感知的数值评分相结合的基准。该基准包括403个广泛的智能体任务，分为三个轨道（基础、模拟器基础和文档基础验证）和八个应用领域，具有24,500个经过验证的执行步骤。这些结果表明，可靠的地球科学智能体必须超越工具访问，协调异构工作流程，精确参数化工具，并保留文献来源。
cs.AI / 49 / 2606.13176
Mental-R1: Aligning LLM Reasoning for Mental Health Assessment
Mental-R1：对齐大型语言模型推理以进行心理健康评估
Abstract
Mental health problems such as anxiety, depression, and suicide remain urgent global challenges, where timely and accurate assessment is critical for effective intervention. Recently, large language models have been explored for mental health assessment. However, existing general-purpose post-training methods do not align with the cognitive processes of human assessment, which may lead to unreliable reasoning outcomes. To bridge this gap, we propose Cognitive Relative Policy Optimization (CRPO), a reinforcement learning framework tailored for the mental health domain. CRPO extends group relative policy optimization by integrating stage-dependent uncertainty modeling into the policy optimization process. Specifically, we introduce a stage-wise entropy regularization mechanism that encourages broad exploration in early reasoning phases and progressively enforces confident decision-making in later stages, mimicking the human cognitive shift from uncertainty to certainty. In addition, inspired by cognitive appraisal theory, we formalize cognitive reasoning stages, thereby guiding theory-grounded interpretable inference. Experiments on 8 mental health datasets show that CRPO achieves an average improvement of 10.4 percentage points in weighted F1-score over the best reinforcement learning baseline. Furthermore, the CRPO-trained model Mental-R1 demonstrates clear advantages compared with existing large language models on reasoning-intensive cases, suggesting that CRPO enhances reasoning capabilities for mental health assessment.
Chinese Translation
心理健康问题，如焦虑、抑郁和自杀，仍然是全球紧迫的挑战，及时和准确的评估对于有效干预至关重要。最近，大型语言模型被探索用于心理健康评估。然而，现有的通用后训练方法与人类评估的认知过程不一致，这可能导致不可靠的推理结果。为了解决这一问题，我们提出了认知相对策略优化（Cognitive Relative Policy Optimization, CRPO），这是一个专为心理健康领域量身定制的强化学习框架。CRPO通过将阶段依赖的不确定性建模整合到策略优化过程中，扩展了群体相对策略优化。具体而言，我们引入了一种阶段性熵正则化机制，鼓励在早期推理阶段进行广泛探索，并在后期阶段逐步强化自信决策，模拟人类从不确定性向确定性的认知转变。此外，受到认知评估理论的启发，我们形式化了认知推理阶段，从而指导基于理论的可解释推理。在8个心理健康数据集上的实验表明，CRPO在加权F1-score上比最佳强化学习基线平均提高了10.4个百分点。此外，经过CRPO训练的模型Mental-R1在推理密集型案例中与现有的大型语言模型相比显示出明显优势，表明CRPO增强了心理健康评估的推理能力。
cs.AI / 50 / 2606.13192
Reasoning for Mobile User Experience with Multimodal LLMs: Task, Benchmark, and Approach
基于多模态大语言模型的移动用户体验推理：任务、基准和方法
Abstract
User experience (UX) centered on usability, perceived consistency, and functional clarity is fundamental to real-world user interfaces (UI). The application of multimodal large language models (MLLMs) in the field of user interfaces is evolving rapidly, such as visual element grounding, graphical user interface (GUI) agents, and design-to-code generation. However, research efforts on evaluating UX based on UI screenshots are still immature. To address this, we propose UXBench, a novel multimodal benchmark consisting of 2,000 VQA data samples designed to assess MLLMs' ability to perform UI-based reasoning. UXBench includes 8 tasks based on real-world UI screenshots that require fine-grained diagnosis of UX issues across layout relationships, visual hierarchy, and content consistency. Our extensive evaluation of mainstream MLLMs shows that they remain fundamentally limited in their capacity for UI-based reasoning. The results underscore the need for further advancements in this area. To bridge this gap, we propose UI-UX, an MLLM based on Qwen3-VL-4B-Thinking foundation model and enhanced via reinforcement learning with two key innovations: a reward routing mechanism that dynamically balances perceptual understanding and logical reasoning during inference, and an asymmetric transition reward that suppresses redundant or insufficient reasoning steps. Experiments demonstrate that UI-UX achieves state-of-the-art (SOTA) performance on UXBench, attaining an accuracy of 0.7963 -- surpassing Claude-4.5-Sonnet's 0.6550 -- while exhibiting strong generalization across diverse UI tasks and maintaining low inference latency.
Chinese Translation
以可用性、感知一致性和功能清晰度为中心的用户体验（UX）是现实世界用户界面（UI）的基础。多模态大语言模型（MLLMs）在用户界面领域的应用正在迅速发展，例如视觉元素定位、图形用户界面（GUI）代理和设计到代码生成。然而，基于UI截图评估UX的研究仍然不够成熟。为了解决这一问题，我们提出了UXBench，这是一个新颖的多模态基准，包含2000个视觉问答（VQA）数据样本，旨在评估MLLMs在UI基础推理方面的能力。UXBench包括8个基于真实UI截图的任务，这些任务需要对布局关系、视觉层次和内容一致性进行细致的UX问题诊断。我们对主流MLLMs的广泛评估表明，它们在UI基础推理能力上仍然存在根本性的局限性。结果强调了在这一领域进一步发展的必要性。为弥补这一差距，我们提出了UI-UX，这是一个基于Qwen3-VL-4B-Thinking基础模型的MLLM，通过强化学习进行了增强，具有两个关键创新：一种奖励路由机制，在推理过程中动态平衡感知理解和逻辑推理，以及一种不对称转移奖励，抑制冗余或不足的推理步骤。实验表明，UI-UX在UXBench上达到了最先进的（SOTA）性能，准确率为0.7963，超过了Claude-4.5-Sonnet的0.6550，同时在多样化的UI任务中表现出强大的泛化能力，并保持低推理延迟。
cs.AI / 51 / 2606.13196
Under What Conditions Can a Machine Become Genuinely Creative?
在什么条件下机器可以真正具备创造力？
Abstract
Recent AI systems can generate texts, software architectures, hypotheses, designs, and scientific workflows that appear creative. This paper asks under what conditions a machine can become genuinely creative, and how human agency can be preserved within shared cognitive and creative environments. It develops a requirement framework derived from Designics, the science of meaning-bearing intentional change. The paper argues that genuine machine creativity should not be defined by output novelty, current performance, or transient architecture alone. Instead, creativity is understood as the structural transformation of incomplete situations through recursive intervention dynamics. On this view, it depends on ten requirements: environment representation, scoped perception, conflict identification, intervention capability, consequence observation, knowledge and environment update, rescoping, local-to-global unfolding, value-based scoping, and human-AI co-living. These are organized through the three laws of Designics: perception, conflict, and capability. The paper illustrates the computational tractability of these requirements through selected cyber-physical and cyber-biological studies, including recursive element extraction, autonomous mesh generation, and neurophysiological and workload analysis. It then treats open-ended systems, automated discovery frameworks, self-modifying agents, foundation models, and agentic workflows as pressure cases: they demonstrate powerful generative means but do not by themselves establish genuine machine creativity. Finally, the paper argues that proactive AI ethics is internal to genuine machine creativity rather than an after-the-fact filter. Value-based scoping and human-AI co-living must shape how creative machines perceive environments, identify conflicts, select interventions, observe consequences, update knowledge, and rescope future action.
Chinese Translation
近期的人工智能系统能够生成看似具有创造性的文本、软件架构、假设、设计和科学工作流程。本文探讨了在什么条件下机器能够真正具备创造力，以及如何在共享的认知和创造性环境中保持人类的主动性。本文发展了一个基于设计学（Designics）的需求框架，设计学是关于有意义的意图变更的科学。本文认为，真正的机器创造力不应仅仅通过输出的新颖性、当前的表现或短暂的架构来定义。相反，创造力被理解为通过递归干预动态对不完整情境的结构性转变。从这个角度来看，它依赖于十个要求：环境表征、范围感知、冲突识别、干预能力、后果观察、知识与环境更新、重新界定、局部到整体的展开、基于价值的界定，以及人机共生。这些要求通过设计学的三条法则进行组织：感知、冲突和能力。本文通过选定的网络物理和网络生物研究，展示了这些要求的计算可行性，包括递归元素提取、自主网格生成，以及神经生理和工作负载分析。接着，本文将开放式系统、自动发现框架、自我修改的代理、基础模型和代理工作流程视为压力案例：它们展示了强大的生成手段，但本身并不能确立真正的机器创造力。最后，本文论证了主动的人工智能伦理是内在于真正的机器创造力，而不是事后过滤。基于价值的界定和人机共生必须塑造创造性机器如何感知环境、识别冲突、选择干预、观察后果、更新知识以及重新界定未来行动。
cs.AI / 52 / 2606.13197
ARMOR-MAD: Adaptive Routing for Heterogeneous Multi-Agent Debate in Large Language Model Reasoning
ARMOR-MAD：异构多智能体辩论在大型语言模型推理中的自适应路由
Abstract
Multi-agent debate (MAD) can improve large language model reasoning, but fixed debate pipelines often waste computation and can amplify correlated errors among similar agents. We propose ARMOR-MAD, a training-free heterogeneous MAD framework that treats debate as conditional computation. ARMOR-MAD combines three components: Pre-debate Agreement Routing (PAR) decides whether independently generated Round-0 answers require debate; Early Agreement Stopping Evaluator (EASE) stops debate after convergence; and Semantic Outlier Detection (SOD) down-weights abnormal final answers during aggregation. Across MATH Level 5, GSM8K, MMLU, and MMLU-Pro, ARMOR-MAD consistently improves over fixed-round heterogeneous debate with the same model pool, reaching 65.5\%, 96.5\%, 90.0\%, and 81.5\% accuracy, respectively. The results suggest that genuine model heterogeneity and agreement-based control are both important for making MAD more accurate and efficient.
Chinese Translation
多智能体辩论（MAD）可以提升大型语言模型的推理能力，但固定的辩论流程往往会浪费计算资源，并可能放大相似智能体之间的相关错误。我们提出了ARMOR-MAD，一种无训练的异构MAD框架，将辩论视为条件计算。ARMOR-MAD结合了三个组件：预辩论一致性路由（Pre-debate Agreement Routing, PAR）决定独立生成的第0轮答案是否需要辩论；早期一致性停止评估器（Early Agreement Stopping Evaluator, EASE）在收敛后停止辩论；语义异常检测（Semantic Outlier Detection, SOD）在聚合过程中对异常最终答案进行降权。在MATH Level 5、GSM8K、MMLU和MMLU-Pro数据集上，ARMOR-MAD在相同模型池下始终优于固定轮次的异构辩论，分别达到了65.5%、96.5%、90.0%和81.5%的准确率。结果表明，真正的模型异质性和基于一致性的控制对于提高MAD的准确性和效率都至关重要。
cs.AI / 53 / 2606.13201
A Minimal Model of Bounded Trade-Off Screening in Multi-Attribute Choice
多属性选择中有限权衡筛选的最小模型
Abstract
Human decision-making often involves choosing between multi-attribute alternatives, yet classical models assume fully compensatory utility aggregation despite evidence that people reject options with poor performance on critical attributes. We propose a bounded trade-off reasoning framework in which decisions are governed by a screening process that evaluates the balance between gains and losses across attributes. The model introduces a trade-off tolerance parameter that controls acceptable imbalance and can vary across contexts. Through simulation, we show that this mechanism produces preference patterns that differ from standard utility-based models and captures context-dependent variation in trade-off behavior. These results establish bounded trade-off screening as a plausible computational mechanism for multi-attribute choice and generate testable predictions for future behavioral studies.
Chinese Translation
人类决策通常涉及在多属性替代方案之间进行选择，然而经典模型假设完全补偿的效用聚合，尽管有证据表明人们会拒绝在关键属性上表现不佳的选项。我们提出了一种有限权衡推理框架，其中决策由评估属性间收益与损失平衡的筛选过程所主导。该模型引入了一个权衡容忍参数，控制可接受的不平衡程度，并且可以在不同情境中变化。通过模拟，我们展示了这一机制产生的偏好模式与标准基于效用的模型不同，并捕捉了权衡行为的情境依赖性变化。这些结果确立了有限权衡筛选作为多属性选择的一个合理计算机制，并为未来的行为研究生成了可检验的预测。
cs.AI / 54 / 2606.13211
Hallucination in Medical Imaging AI: A Cross-Modality Analytical Framework for Taxonomy, Detection, and Mitigation under Regulatory Constraints
医疗影像人工智能中的幻觉：在监管约束下的跨模态分析框架用于分类、检测和缓解
Abstract
AI systems are being deployed across medical imaging faster than their failure modes are understood. At this point in time, the failure of greatest clinical concern is hallucination: clinically plausible but factually incorrect outputs, including fabricated anatomical structures, missed findings, incorrect laterality, and invented measurements in generated reports, with direct consequences, for example, for biopsy decisions, staging, and treatment planning. This structured narrative synthesizes peer-reviewed studies, benchmark datasets, and FDA regulatory guidance across five imaging modalities to produce a cross-modality analysis of hallucination taxonomy, etiology, detection, and mitigation. Specifically, we address three questions in this study: (1) how can existing taxonomies be unified across modalities?, (2) how do medical-specialized foundation models hallucinate less than general-purpose ones?, and (3) which mitigation strategies are effective and compatible with FDA lifecycle oversight? We note that three taxonomic frameworks together cover the imaging pipeline in a way no single framework does alone. We also highlight that general-purpose foundation models outperform medical-specialized models on hallucination-specific benchmarks, indicating that narrow domain fine-tuning can introduce overfitting-induced confabulation. At the same time, the oversight of radiologists remains essential; for instance, a very high percentage of of AI-generated flags required expert correction before clinical use. Physics-informed architectural constraints, Chain-of-Thought prompting, and human-in-the-loop safeguards each address different failure modes and is effective when combined. All findings are mapped to the FDA's Total Product Lifecycle and Predetermined Change Control Plan frameworks, which treat hallucination management as a lifecycle obligation rather than a pre-deployment checklist.
Chinese Translation
人工智能系统在医疗影像领域的部署速度超过了对其故障模式的理解。目前，最令人关注的故障是幻觉：即临床上看似合理但实际上不正确的输出，包括虚构的解剖结构、遗漏的发现、错误的侧别以及在生成报告中虚构的测量，这些都对活检决策、分期和治疗计划等产生直接影响。本结构化叙述综合了经过同行评审的研究、基准数据集和FDA监管指南，涵盖五种影像模态，产生了关于幻觉分类、病因、检测和缓解的跨模态分析。具体而言，我们在本研究中解决了三个问题：（1）如何在不同模态之间统一现有的分类？（2）医学专用基础模型的幻觉现象为何少于通用模型？（3）哪些缓解策略有效且与FDA生命周期监管兼容？我们注意到，三个分类框架共同覆盖了影像处理流程，而没有任何单一框架能够独立完成。同时，我们强调通用基础模型在幻觉特定基准测试中的表现优于医学专用模型，这表明狭域微调可能引入过拟合导致的虚构。同时，放射科医师的监督仍然至关重要；例如，AI生成的警报中有很高比例在临床使用前需要专家修正。物理信息驱动的架构约束、思维链提示和人机协作的安全措施分别针对不同的故障模式，并在结合使用时效果显著。所有发现都映射到FDA的整体产品生命周期和预定变更控制计划框架，这将幻觉管理视为生命周期义务，而非部署前的检查清单。
cs.AI / 55 / 2606.13220
LLM-as-an-Investigator: Evidence-First Reasoning for Robust Interactive Problem Diagnosis
LLM作为调查者：以证据为先的稳健互动问题诊断
Abstract
Large language models (LLMs) are increasingly used as interactive assistants for technical problem solving. However, when users provide incomplete descriptions or plausible but unverified explanations, LLMs may prematurely align with these assumptions and propose solutions before collecting sufficient evidence. We refer to this behavior as user-driven sycophancy: the tendency of an LLM to reinforce a user-provided hypothesis instead of testing alternative explanations. This paper introduces LLM-as-an-Investigator, an evidence-first agentic AI methodology for robust problem diagnosis. The approach is implemented through a Solution Investigator Agent, which estimates the ambiguity of an initial problem description, generates candidate hypotheses, asks targeted clarification questions, and updates hypothesis probabilities after each answer. Rather than producing an immediate response, the agent continues the investigation until the evidence makes one candidate explanation stronger than the alternatives. To evaluate the approach, we build a benchmark from solved technical forum threads in mechanical, electrical, and hydraulic domains. We use a three-agent evaluation pipeline in which a Problem-Solution Extractor Agent converts solved threads into structured cases, a Ground-Truth Evaluator Agent simulates the user while hiding the known solution, and the tested assistant attempts to recover the solution through dialogue. The experiments compare standard assistants, reasoning-oriented LLMs, and the proposed investigator-based model across LLM backbones. In addition to diagnostic accuracy, we analyze how standard assistants follow misleading user hypotheses in diagnostic cases. The results show that the proposed approach identifies the problem more accurately than direct prompting and reasoning-only baselines, while its evidence-first protocol helps reduce user-induced conversational bias.
Chinese Translation
大型语言模型（LLMs）越来越多地被用作技术问题解决的互动助手。然而，当用户提供不完整的描述或合理但未经验证的解释时，LLMs可能会过早地与这些假设对齐，并在收集足够证据之前提出解决方案。我们将这种行为称为用户驱动的谄媚：LLM强化用户提供的假设而不是测试替代解释的倾向。本文介绍了LLM作为调查者（LLM-as-an-Investigator），一种以证据为先的自主人工智能方法，用于稳健的问题诊断。该方法通过解决方案调查代理（Solution Investigator Agent）实施，该代理评估初始问题描述的模糊性，生成候选假设，提出针对性的澄清问题，并在每个回答后更新假设概率。代理并不立即产生响应，而是继续调查，直到证据使一个候选解释比其他解释更强。为了评估该方法，我们从机械、电气和液压领域的已解决技术论坛线程中构建了基准。我们使用三代理评估管道，其中问题-解决方案提取代理（Problem-Solution Extractor Agent）将已解决的线程转换为结构化案例，真实情况评估代理（Ground-Truth Evaluator Agent）模拟用户，同时隐藏已知解决方案，而被测试的助手则尝试通过对话恢复解决方案。实验比较了标准助手、面向推理的LLMs和所提出的基于调查者的模型在LLM骨干上的表现。除了诊断准确性外，我们还分析了标准助手在诊断案例中如何遵循误导性的用户假设。结果表明，所提出的方法比直接提示和仅推理的基线更准确地识别问题，同时其以证据为先的协议有助于减少用户引起的对话偏差。
cs.AI / 56 / 2606.13241
Brick: Spatial Capability Routing for the Mixture-of-Models (MoM) Paradigm
Brick：混合模型（MoM）范式的空间能力路由
Abstract
Defining query difficulty is one of the hardest problems in deployment engineering. Existing LLM routers rely on surface features such as domain labels, keywords, and token count, ignoring the within-domain variance that actually determines model success. Frontier models cost ten to one hundred times more than local open-weight models, so at production scale even small per-request savings become a direct cloud-bill lever. We present Brick, a multimodal router that scores each model on six capability dimensions, combines this with a per-query difficulty estimate, and dispatches via a cost-penalized geometric rule. A continuous preference knob lets operators slide between max-quality and max-saving profiles at deploy time. On a benchmark of 5,504 queries, Brick at max-quality reaches 76.98% accuracy, beating the best single model (75.02%) and all tested routers. At a neutral cost-quality profile, Brick achieves 74.11% accuracy at 4.71x lower cost than always using the strongest model. At min-cost, it cuts cost 22.15x with 11.85 points accuracy loss. Median latency drops from 51.2s to 22.8s.
Chinese Translation
定义查询难度是部署工程中最棘手的问题之一。现有的大型语言模型（LLM）路由器依赖于表面特征，如领域标签、关键词和令牌计数，忽略了实际上决定模型成功的领域内变异。前沿模型的成本是本地开放权重模型的十倍到一百倍，因此在生产规模下，即使是每次请求的小幅节省也会成为直接的云账单杠杆。我们提出了Brick，这是一种多模态路由器，它在六个能力维度上对每个模型进行评分，并结合每个查询的难度估计，通过成本惩罚几何规则进行调度。一个连续的偏好调节器使操作员在部署时能够在最大质量和最大节省配置之间滑动。在一个包含5504个查询的基准测试中，Brick在最大质量下达到了76.98%的准确率，超越了最佳单一模型（75.02%）和所有测试过的路由器。在中性成本-质量配置下，Brick以4.71倍更低的成本实现了74.11%的准确率，相比于始终使用最强模型。在最低成本配置下，成本降低了22.15倍，准确率损失为11.85个百分点。中位延迟从51.2秒降至22.8秒。
cs.AI / 57 / 2606.13247
EPIG: Emotion-Based Prompting for Personalised Image Generation
EPIG：基于情感的个性化图像生成提示
Abstract
Text-to-image diffusion models have achieved impressive results in synthesizing high-quality images from natural language prompts. However, commonly used prompting strategies remain relatively generic, limiting the model's ability to accurately express emotional intent and nuanced affective attributes. This work proposes EPIG, a method that enhances emotional expressiveness at the prompt level prior to image generation. Grounded in psychologically informed emotion representations (valence-arousal) and leveraging structured, role-aware prompt enrichment, EPIG enriches emotion-related components of prompts without modifying or retraining the image generation backbone. The resulting emotion-aware prompts guide the generative process toward more emotionally coherent visual outputs, with particular effectiveness in controlling arousal. EPIG is lightweight, training-free, and well suited for resource-constrained and personalized image generation scenarios. Experimental results on a benchmark of 10 diverse prompts show that EPIG reduces mean arousal error compared to strong baselines, including naive insertion and LLM-based prompt expansion, with reductions of 14% and 12%, respectively. These improvements are statistically significant. EPIG also preserves valence alignment and semantic consistency, as measured by CLIPScore and supported by ablation studies. The effect is more pronounced on prompts containing explicit subjects such as humans, children, or animals, where the reduction reaches 17%, highlighting the subject-sensitive behavior of the proposed method.
Chinese Translation
文本到图像的扩散模型在从自然语言提示合成高质量图像方面取得了令人印象深刻的成果。然而，常用的提示策略相对通用，限制了模型准确表达情感意图和细微情感属性的能力。本研究提出了EPIG，一种在图像生成之前增强提示层面情感表现力的方法。EPIG基于心理学知情的情感表征（愉悦度-唤醒度），并利用结构化、角色感知的提示丰富，增强提示中与情感相关的成分，而无需修改或重新训练图像生成骨干网络。生成的情感感知提示引导生成过程朝向更具情感一致性的视觉输出，尤其在控制唤醒度方面表现出特别的有效性。EPIG轻量、无需训练，适合资源受限和个性化图像生成场景。在10个多样化提示的基准实验结果表明，与强基线（包括简单插入和基于大型语言模型（LLM）的提示扩展）相比，EPIG减少了平均唤醒误差，分别降低了14%和12%。这些改进在统计上显著。EPIG还保持了愉悦度对齐和语义一致性，依据CLIPScore的测量和消融研究的支持。该效果在包含明确主题（如人类、儿童或动物）的提示上更为明显，减少幅度达到17%，突显了所提方法的主题敏感性行为。
cs.AI / 58 / 2606.13249
Multi-Field Hybrid Retrieval-Augmented Generation for Maritime Accident Root Cause Analysis
多领域混合检索增强生成用于海事事故根本原因分析
Abstract
Maritime accident adjudication reports contain critical tribunal findings for root cause analysis (RCA), yet retrieving relevant precedents and drafting consistent reports from decades of records remains labor-intensive. This paper proposes a multi-field hybrid retrieval-augmented generation (RAG) framework for automated maritime RCA, utilizing a comprehensive dataset of 13,329 Korea Maritime Safety Tribunal (KMST) reports (1971-2025). We transform raw adjudications into a structured knowledge base of "incident cards", indexing three distinct fields-Summary, Causes, and Disposition-alongside a hierarchical L1/L2 cause taxonomy. Our retrieval strategy employs a field-aware hybrid approach, fusing sparse and dense rankings via Reciprocal Rank Fusion (RRF). Given the lack of large-scale expert relevance labels, we evaluate retrieval performance using ceiling-normalized recall and nDCG based on a metadata-derived proxy relevance score. Experimental results demonstrate that our proposed retrieval significantly outperforms baseline methods, improving NormRecall@100 from 0.18 to 0.55. Furthermore, grounding the generator on the retrieved precedents enhances RCA generation quality over an LLM-only baseline, increasing the LLM-as-a-judge score from 3.34 to 3.72. These findings suggest that field-aware RAG can substantially streamline maritime safety investigation workflows by enabling faster precedent search and more consistent, evidence-based RCA drafting.
Chinese Translation
海事事故裁决报告包含了根本原因分析（RCA）的关键法庭发现，但从数十年的记录中检索相关先例并撰写一致的报告仍然是一项劳动密集型的工作。本文提出了一种多领域混合检索增强生成（RAG）框架，用于自动化海事RCA，利用了一个包含13,329份韩国海事安全法庭（KMST）报告（1971-2025）的综合数据集。我们将原始裁决转化为一个结构化的知识库“事件卡”，并对三个不同领域进行索引——摘要、原因和处置，同时建立了一个分层的L1/L2原因分类法。我们的检索策略采用了领域感知的混合方法，通过互惠排名融合（RRF）将稀疏和密集排名相结合。鉴于缺乏大规模的专家相关性标签，我们使用基于元数据派生的代理相关性评分来评估检索性能，采用了归一化召回率和nDCG。实验结果表明，我们提出的检索方法显著优于基线方法，将NormRecall@100从0.18提高到0.55。此外，将生成器基于检索到的先例进行增强，提升了RCA生成质量，相较于仅使用LLM的基线，LLM作为裁判的评分从3.34提高到3.72。这些发现表明，领域感知的RAG能够显著简化海事安全调查工作流程，通过加快先例检索和更一致、基于证据的RCA撰写来实现。
cs.AI / 59 / 2606.13258
MOSAIC: Modality-Specific Adaptation for Incremental Continual Learning in Parkinson's Disease Gait Assessment
MOSAIC：用于帕金森病步态评估的增量持续学习的模态特定适应
Abstract
Gait-based Parkinson's disease assessment increasingly relies on heterogeneous sensors, but clinical systems rarely collect all modalities simultaneously. New sensors may arrive through device upgrades, protocol changes, or multi-center deployment, while historical patient data are often unavailable because of privacy and storage constraints. This modality-incremental setting faces three challenges: unreliable cross-modal distillation, modality-specific statistical shifts, and reduced plasticity after preservation. We propose MOSAIC, a compact continual learning framework. First, we identify the Toxic Teacher phenomenon and introduce Modality-Specific Warm-Up to stabilize newly learned modality representations before distillation. Second, we propose a statistics-decoupled MSBN architecture that isolates sensor statistics while maintaining a shared semantic backbone. Third, we design a curriculum-guided repulsive objective for Plasticity Recovery, preserving legacy knowledge while recovering modality-specific capacity. Experiments on three multimodal Parkinson's gait datasets show that MOSAIC improves final performance and mitigates forgetting. Project code is available at: https://github.com/minlinzeng/MOSAIC_Modality-Specific-Adaptation-for-Incremental-Continual-Learning-in-PD-Gait-Assessment.git
Chinese Translation
基于步态的帕金森病评估日益依赖异构传感器，但临床系统很少同时收集所有模态。新传感器可能通过设备升级、协议变更或多中心部署而出现，而由于隐私和存储限制，历史患者数据通常不可用。这种模态增量设置面临三个挑战：不可靠的跨模态蒸馏、模态特定的统计偏移以及在保留后减少的可塑性。我们提出了MOSAIC，一个紧凑的持续学习框架。首先，我们识别了有毒教师现象，并引入模态特定的热身策略，以在蒸馏之前稳定新学习的模态表示。其次，我们提出了一种统计解耦的MSBN架构，该架构在保持共享语义骨干的同时，隔离传感器统计信息。第三，我们设计了一种课程引导的排斥目标，以实现可塑性恢复，保留遗留知识的同时恢复模态特定的能力。在三个多模态帕金森病步态数据集上的实验表明，MOSAIC提高了最终性能并减轻了遗忘。项目代码可在以下链接获取：https://github.com/minlinzeng/MOSAIC_Modality-Specific-Adaptation-for-Incremental-Continual-Learning-in-PD-Gait-Assessment.git
cs.AI / 60 / 2606.13262
From Verdict to Process: Agentic Reinforcement Learning for Multi-Stage Fact Verification
从裁决到过程：用于多阶段事实验证的自主强化学习
Abstract
Recent approaches combining Large Language Models (LLMs) with retrieval-augmented reasoning have shown promise for automated fact verification. To process complex claims, these verification pipelines typically execute multi-stage workflows that coordinate tightly coupled modules, including claim decomposition, evidence gathering, and verdict prediction. However, existing methods optimize individual stages in isolation or rely on fixed heuristics, which limits adaptive coordination among stages and can lead to suboptimal outcomes. In this work, we propose ProFact, an agentic reinforcement learning framework for end-to-end optimization of multi-stage fact verification trajectories. ProFact trains a unified policy to coordinate claim decomposition, evidence seeking, answer generation, and verdict prediction. To address the sparse and delayed supervision provided by final veracity labels, ProFact introduces process-aware rewards that provide stage-level learning signals throughout the verification process. Empirical evaluation shows that ProFact consistently outperforms strong baselines in both verification performance and inference efficiency. These results highlight the effectiveness of process-aware trajectory optimization for multi-stage fact verification.
Chinese Translation
最近将大型语言模型（LLMs）与检索增强推理相结合的方法在自动化事实验证方面显示出良好的前景。为了处理复杂的主张，这些验证流程通常执行多阶段工作流，紧密协调包括主张分解、证据收集和裁决预测等模块。然而，现有方法往往孤立地优化各个阶段或依赖固定的启发式方法，这限制了各阶段之间的自适应协调，可能导致次优结果。在本研究中，我们提出了ProFact，一个用于多阶段事实验证轨迹端到端优化的自主强化学习框架。ProFact训练一个统一的策略，以协调主张分解、证据寻求、答案生成和裁决预测。为了解决最终真实性标签提供的稀疏和延迟监督问题，ProFact引入了过程感知奖励，在整个验证过程中提供阶段级学习信号。实证评估表明，ProFact在验证性能和推理效率方面始终优于强基线。这些结果突显了过程感知轨迹优化在多阶段事实验证中的有效性。
cs.AI / 61 / 2606.13282
ERTS: Adversarial Robustness Testing of Ethical AI via Semantic Perturbation in a Bounded Consequence Space
ERTS：通过语义扰动在有限后果空间中对伦理人工智能进行对抗鲁棒性测试
Abstract
As AI systems are deployed in high-stakes ethical contexts such as healthcare triage, autonomous vehicle control, and employment screening, formal methods for evaluating their robustness against adversarial manipulation of ethical reasoning remain underdeveloped. This paper introduces the Ethical Robustness Testing System (ERTS), a closed-pipeline framework that: (1) encodes ethical dilemmas into a 22-dimensional Ethical Consequence Space (ECS) grounded in established ethical theory; (2) applies 17 semantic perturbation functions subject to 6 validity constraint classes including a novel semantic coherence constraint; (3) measures decision deviation via a 4-component Ethical Instability Index (EII); and (4) produces domain-adaptive pre-deployment robustness assessment verdicts. We evaluate 4 structured baseline models and 2 production LLMs (Gemini 2.0 Flash and Llama 3.2) across 50 ethical scenarios spanning 8 deployment domains, generating 1,500 adversarial test cases. Results demonstrate that only 33% of models achieve assessment clearance, with the local Llama-3.2 model proving particularly vulnerable to fairness corruption and information degradation attacks (ERS = 0.737). To the best of our knowledge, no existing framework combines a bounded ethical consequence space, semantic coherence constraints, and domain-adaptive assessment in a single adversarial testing pipeline.
Chinese Translation
随着人工智能系统在医疗分诊、自动驾驶控制和就业筛选等高风险伦理环境中的应用，评估其对伦理推理的对抗性操控的鲁棒性的正式方法仍然不够成熟。本文介绍了伦理鲁棒性测试系统（ERTS），这是一个闭环框架，包含以下几个部分：（1）将伦理困境编码为基于既定伦理理论的22维伦理后果空间（ECS）；（2）应用17种语义扰动函数，受6类有效性约束的限制，包括一种新颖的语义一致性约束；（3）通过4个组成部分的伦理不稳定指数（EII）来衡量决策偏差；（4）生成领域自适应的预部署鲁棒性评估结果。我们在涵盖8个部署领域的50个伦理场景中评估了4个结构化基线模型和2个生产级大型语言模型（Gemini 2.0 Flash和Llama 3.2），生成了1,500个对抗性测试案例。结果表明，只有33%的模型通过了评估，其中本地的Llama-3.2模型对公平性破坏和信息降级攻击特别脆弱（ERS = 0.737）。据我们所知，目前没有现有框架能够在单一的对抗测试管道中结合有限的伦理后果空间、语义一致性约束和领域自适应评估。
cs.AI / 62 / 2606.13302
Physics-Guided Spatiotemporal Learning for Coastal Wave Peak Period Estimation from Video
基于物理引导的时空学习用于从视频中估计沿海波峰周期
Abstract
Wave parameters in the nearshore are crucial for coastal engineering, shoreline protection, marine hazard assessment, and coastal management for climate resilience. Traditional monitoring systems like buoys and radar platforms offer accurate monitoring but can have high installation and maintenance expenses and limited spatial coverage. Passive ocean monitoring using video has been achieved by leveraging deep learning, however, many methods are not physically interpretable, feasible, and validated for oceanography. In thiswork, a Physics-Guided Deep Spatiotemporal Learning Framework for direct estimation of nearshore wave peak periods from passive coastal video stream is proposed. The framework combines automated temporal-variance based region-of-interest detection, multi-stage Sim-to-Real transfer learning, and physics-informed regularization to enhance the predictive accuracy and physical consistency. A variety of spatiotemporal architectures were assessed, such as transformer-based and recurrent-convolutional ones, alongside synthetic pretraining,silver-label adaptation, and expert fine-tuning. The results show that transformer-based architectures outperformed in terms of the accuracy of the instantaneous prediction, while lightweight recurrent-convolutional architectures achieved higher temporal stability and operational oceanographic skill. Ablation studies also demonstrated the benefits of physics-guided regularization in terms of trend-following consistency, and physically implausible predictions. Explainability auditing also helped to focus attention in hydrodynamically active surf-zone regions and showed good agreement with the physically derived wave propagation behavior. In general, the proposed framework shows the promise of physics-guided video-based deep learning systems for long-term coastal wave monitoring that are cost-efficient and operationally feasible.
Chinese Translation
近岸波参数对海岸工程、海岸保护、海洋灾害评估以及气候适应性海岸管理至关重要。传统监测系统如浮标和雷达平台提供了准确的监测，但可能存在高昂的安装和维护费用以及有限的空间覆盖。利用深度学习实现的被动海洋监测已取得进展，然而，许多方法在物理可解释性、可行性和海洋学验证方面存在不足。在本研究中，提出了一种基于物理引导的深度时空学习框架，用于直接从被动海岸视频流中估计近岸波峰周期。该框架结合了基于自动时间方差的兴趣区域检测、多阶段的Sim-to-Real迁移学习和物理信息正则化，以提高预测准确性和物理一致性。评估了多种时空架构，如基于变换器和递归卷积的架构，以及合成预训练、银标签适应和专家微调。结果表明，基于变换器的架构在瞬时预测的准确性方面表现优越，而轻量级的递归卷积架构则在时间稳定性和操作海洋技能方面表现更佳。消融研究还展示了物理引导正则化在趋势跟随一致性和物理上不合理预测方面的优势。可解释性审计也有助于关注水动力活跃的冲浪区，并与物理推导的波传播行为表现出良好的一致性。总体而言，所提出的框架展示了基于物理引导的视频深度学习系统在长期海岸波监测中的潜力，具有成本效益和操作可行性。
cs.AI / 63 / 2606.13316
ReSum: Synergizing LLM Reasoning and Summarization with Reinforcement Learning
ReSum：通过强化学习协同大语言模型的推理与摘要
Abstract
Reinforcement Learning with Verifiable Rewards (RLVR) is a central technique for improving long-horizon reasoning in Large Language Models (LLMs). However, existing RLVR methods often encourage unnecessarily long reasoning rollouts, which can degrade reasoning coherence and exhaust the available context budget. Existing approaches to long-context organization often depend on external mechanisms to organize rollouts, rather than enabling the model to manage its own reasoning trajectory. To address this limitation, we propose ReSum, a novel RLVR framework that enables LLMs to compress and organize their reasoning trajectories through self-summarization. Our pilot studies show that self-summarization stabilizes generation by lowering token-level entropy, and that introducing a ``summarization'' phrase can substantially mitigate errors propagated from an incorrect rollout prefix. Motivated by these findings, ReSum adopts a summarization-aware adaptive rollout mechanism that contrastively evaluates whether self-summarization benefits the ongoing reasoning process. Specifically, when the model spontaneously triggers self-summarization, ReSum masks the summarization phrase to create a contrastive branch; for non-summarization positions, it instead randomly injects the phrase to create a matched branch. We further design a summarization-aware advantage to enable finer-grained comparison between contrastive rollout trajectories. Extensive experiments show that ReSum improves performance at an average of 4\% while reducing rollout length by 18.6\%.
Chinese Translation
可验证奖励的强化学习（Reinforcement Learning with Verifiable Rewards, RLVR）是提升大语言模型（Large Language Models, LLMs）长时间推理能力的核心技术。然而，现有的RLVR方法往往鼓励不必要的长推理展开，这可能会降低推理的连贯性并耗尽可用的上下文预算。现有的长上下文组织方法通常依赖于外部机制来组织推理展开，而不是让模型自行管理其推理轨迹。为了解决这一局限性，我们提出了ReSum，一个新颖的RLVR框架，使LLMs能够通过自我摘要来压缩和组织其推理轨迹。我们的初步研究表明，自我摘要通过降低令牌级别的熵来稳定生成，并且引入“摘要”短语可以显著减轻从错误的推理前缀传播的错误。基于这些发现，ReSum采用了一种摘要感知的自适应推理机制，对比评估自我摘要是否有利于正在进行的推理过程。具体而言，当模型自发触发自我摘要时，ReSum会屏蔽摘要短语以创建对比分支；对于非摘要位置，则随机注入该短语以创建匹配分支。我们进一步设计了一种摘要感知的优势，以便在对比推理轨迹之间进行更细粒度的比较。大量实验表明，ReSum在提高性能的同时，平均提升4%，并将推理长度减少了18.6%。
cs.AI / 64 / 2606.13361
Can I Buy Your KV Cache?
我可以购买你的 KV 缓存吗？
Abstract
Right now, across the world, AI agents are repeating the same absurd act: to read one document, they each recompute it from scratch. Every agent re-runs prefill, the most compute-intensive step a large model takes, over identical text, only to rebuild a key-value (KV) cache identical to the one the agent before it just built. The same answer, computed a million times. We make a proposal that is almost offensively simple: compute it once. Let a publisher precompute a document's KV cache, and let every other agent buy the right to load it and skip prefill. It works, and it is token-exact: loading a precomputed KV and continuing matches prefilling from scratch (24/24 greedy tokens, and at the logits level), with no accuracy cost. On Qwen3-4B, reuse is 9-50x cheaper in compute than prefill, and the gap widens with length (prefill's attention scales with L^2), so a single reuse already pays it back. Then the part that matters: where the KV lives. Shipping it fails, because KV is nearly incompressible, so per-load egress costs more than the prefill it saves. Hosting it provider-side, exactly as production prompt-caching works, removes egress entirely. The size of the prize is set by our measured compute saving: serving one hot 3774-token document to 80M agents costs ~$1.5M to re-prefill but only ~$0.03M of reuse compute (49.7x less). The 0.1x cache-read tariff APIs charge passes a 10x discount to users while sitting inside this measured envelope, so the 10x is a floor that the measured ~50x compute saving clears, and the gap to the physical ~50x is provider margin: millions of dollars per popular document. We frame the resulting agent-native prefill CDN and leave lossless KV compression and a cross-party payment layer as the open problems.
Chinese Translation
目前，在全球范围内，人工智能代理正在重复同样荒谬的行为：为了读取一份文档，它们各自从头开始重新计算。每个代理在相同的文本上重新运行预填充（prefill），这是大型模型所需的计算最密集的步骤，仅仅是为了重建一个与前一个代理刚刚构建的相同的键值（KV）缓存。相同的答案，被计算了一百万次。我们提出了一个几乎令人震惊的简单方案：只计算一次。让出版商预先计算文档的 KV 缓存，让其他每个代理购买加载它的权利，从而跳过预填充。这是可行的，并且是逐字精确的：加载一个预计算的 KV 并继续操作与从头开始的预填充相匹配（24/24 贪婪令牌，并且在 logits 级别），没有准确性损失。在 Qwen3-4B 上，重用的计算成本比预填充便宜 9-50 倍，并且随着长度的增加差距会扩大（预填充的注意力与 L^2 成比例），因此单次重用已经能够收回成本。然后是重要的部分：KV 存放的位置。运输它是失败的，因为 KV 几乎不可压缩，因此每次加载的出口成本超过了它所节省的预填充成本。将其托管在提供方侧，正如生产提示缓存的工作方式，完全消除了出口成本。奖赏的规模由我们测量的计算节省设定：为 8000 万个代理提供一份热的 3774 令牌文档的成本约为 150 万美元用于重新预填充，但仅需约 3 万美元的重用计算（少 49.7 倍）。0.1 倍缓存读取收费的 API 在这个测量范围内向用户传递了 10 倍的折扣，因此 10 倍是一个底线，而测量的约 50 倍计算节省清除了这个底线，物理约 50 倍的差距是提供方的利润：每个热门文档数百万美元。我们构建了结果代理原生的预填充内容分发网络（CDN），并将无损 KV 压缩和跨方支付层留作开放问题。
cs.AI / 65 / 2606.13368
IterCAD: An Iterative Multimodal Agent for Visually-Grounded CAD Generation and Editing
IterCAD：一种用于视觉基础计算机辅助设计生成与编辑的迭代多模态智能体
Abstract
Computer-Aided Design is pivotal in modern manufacturing, yet existing automated methods predominantly rely on open-loop, one-shot generation, creating a mismatch with iterative real-world practices. In this paper, we present IterCAD, a unified multimodal agent framework for closed-loop, interactive CAD generation and editing. We formulate the task as a multi-turn interaction between a multimodal agent and an executable CAD sandbox, covering three tasks: Drawing-to-Code, Text-to-Code, and Interactive Editing. To support this, we develop a data synthesis pipeline incorporating advanced industrial manufacturing features to generate standard-compliant multi-view engineering drawings, complex code-editing tasks, and high-fidelity interaction trajectories. We optimize the agent via progressive SFT followed by geometry-aware reinforcement learning with viable-prefix masking to enhance code executability and geometric fidelity. Finally, we introduce the IterCAD-Bench evaluation suite and propose the Chamfer Distance Tolerance-Recall (CD-TR) curve alongside its AUC-TR metric, establishing a survivor-bias-free standard that unifies code validity and geometric precision. Extensive experiments demonstrate that IterCAD achieves highly competitive performance across multiple benchmarks, significantly outperforming existing approaches in both code executability and geometric precision, while exhibiting superior capabilities in closed-loop iterative refinement.
Chinese Translation
计算机辅助设计在现代制造中至关重要，但现有的自动化方法主要依赖于开放式、一次性生成，这与迭代的现实实践存在不匹配。在本文中，我们提出了IterCAD，一个统一的多模态智能体框架，用于闭环的交互式CAD生成与编辑。我们将任务表述为多轮交互，涉及多模态智能体与可执行CAD沙箱之间的互动，涵盖三个任务：绘图到代码（Drawing-to-Code）、文本到代码（Text-to-Code）和交互编辑（Interactive Editing）。为支持这一点，我们开发了一个数据合成管道，结合先进的工业制造特征，以生成符合标准的多视图工程图、复杂的代码编辑任务和高保真的交互轨迹。我们通过渐进式的SFT（Supervised Fine-Tuning）优化智能体，随后采用几何感知的强化学习与可行前缀掩蔽相结合，以增强代码的可执行性和几何的保真度。最后，我们引入了IterCAD-Bench评估套件，并提出了Chamfer距离容忍-召回（CD-TR）曲线及其AUC-TR指标，建立了一个无生存偏差的标准，统一了代码有效性和几何精度。大量实验表明，IterCAD在多个基准测试中表现出高度竞争力，在代码可执行性和几何精度方面显著优于现有方法，同时在闭环迭代优化中展现出卓越的能力。
cs.AI / 66 / 2606.13370
A Quantitative Experimental Repeated Measures Study of Training Dynamics in a Small Llama Style Language Model Under a Compute-Aware Token Budget
在计算感知令牌预算下小型Llama风格语言模型训练动态的定量实验重复测量研究
Abstract
This study examines training dynamics in a small Llama-style language model trained under a fixed, compute-constrained token budget. Rather than evaluating efficiency solely through endpoint performance, the study uses a quantitative experimental repeated measures design to analyze how validation loss, validation perplexity, rolling volatility, backslide behavior, spike behavior, and between-seed variability change across token-based training intervals. Six independent training runs were conducted on a 4.26-million-parameter model using the TinyStories corpus, CPU-based full-precision training, and a target budget of approximately 20 million cumulative training tokens. Metrics were collected across 21 intervals, producing 126 seed-by-interval observations. Repeated measures ANOVA showed statistically significant interval effects for validation loss, validation perplexity, and rolling volatility. Descriptive trajectories revealed rapid early improvement followed by non-monotonic degradation during later training intervals. Mean validation loss decreased from 8.3552 at initialization to 2.7996 near 4 million tokens, but increased to 3.9010 by the final checkpoint. Validation perplexity followed the same pattern, falling sharply early in training before rising later. Derived telemetry further showed recurrent validation-loss backslides and no interval-summary evidence of a stable phase under the predefined criteria. These findings suggest that compute-aware language model evaluation should examine training trajectories rather than endpoint metrics alone. In constrained compute settings, additional token exposure may increase computational cost without producing proportional generalization gains, and interval-level telemetry can reveal instability, regression, and diminishing returns that final metrics may obscure.
Chinese Translation
本研究考察了在固定的计算受限令牌预算下训练的小型Llama风格语言模型的训练动态。研究并非仅通过最终性能评估效率，而是采用定量实验重复测量设计，分析验证损失、验证困惑度、滚动波动性、回退行为、峰值行为以及种子间变异性在基于令牌的训练区间中的变化。使用TinyStories语料库对一个426万参数的模型进行了六次独立的训练，采用基于CPU的全精度训练，目标预算约为2000万个累计训练令牌。共收集了21个区间的指标，产生了126个种子-区间观察值。重复测量方差分析（ANOVA）显示验证损失、验证困惑度和滚动波动性在区间上存在统计显著的效应。描述性轨迹揭示了早期快速改善后在后期训练区间的非单调退化。平均验证损失从初始化时的8.3552降至接近400万令牌时的2.7996，但在最后检查点时上升至3.9010。验证困惑度遵循相同模式，在训练早期急剧下降后在后期上升。衍生的遥测进一步显示了反复的验证损失回退，并且在预定义标准下没有区间汇总证据表明存在稳定阶段。这些发现表明，计算感知的语言模型评估应当考察训练轨迹，而不仅仅是最终指标。在受限计算环境中，额外的令牌暴露可能会增加计算成本而不产生相应的泛化收益，区间级遥测可以揭示不稳定性、回归和收益递减，而最终指标可能会掩盖这些问题。
cs.AI / 67 / 2606.13392
MiniMax Sparse Attention
MiniMax 稀疏注意力
Abstract
Ultra-long-context capability is becoming indispensable for frontier LLMs: agentic workflows, repository-scale code reasoning, and persistent memory all require the model to jointly attend over hundreds of thousands to millions of tokens, yet the quadratic cost of softmax attention makes this untenable at deployment scale. We introduce MiniMax Sparse Attention (MSA), a blockwise sparse attention built upon Grouped Query Attention (GQA). A lightweight Index Branch scores key-value blocks and independently selects a Top-k subset for each GQA group, enabling group-specific sparse retrieval while maintaining efficient block-level execution; the Main Branch then performs exact block-sparse attention over only the selected blocks. Designed around a principle of simplicity and scalability, MSA is deliberately streamlined, making it straightforward to deploy efficiently across a broad range of GPUs. To translate sparsity into practical speedups, we co-design MSA with a GPU execution path that uses exp-free Top-k selection and KV-outer sparse attention to improve tensor-core utilization under block-granular access. On a 109B-parameter model with native multimodal training, MSA performs on par with GQA while reducing per-token attention compute by 28.4x at 1M context. Paired with our co-designed kernel, MSA achieves 14.2x prefill and 7.6x decoding wall-clock speedups on H800. Our inference kernel is available at: https://github.com/MiniMax-AI/MSA. A production-grade natively multimodal model powered by MSA has been publicly released at: https://huggingface.co/MiniMaxAI/MiniMax-M3.
Chinese Translation
超长上下文能力正成为前沿大型语言模型（LLMs）不可或缺的特性：代理工作流、库规模的代码推理和持久内存都要求模型能够共同关注数十万到数百万个标记，然而，softmax 注意力的二次成本使得在部署规模上难以实现。我们提出了 MiniMax 稀疏注意力（MSA），这是一种基于分组查询注意力（Grouped Query Attention, GQA）的块状稀疏注意力。轻量级的索引分支对键值块进行评分，并为每个 GQA 组独立选择一个 Top-k 子集，从而实现组特定的稀疏检索，同时保持高效的块级执行；主分支则仅对选定的块执行精确的块稀疏注意力。MSA 设计遵循简单性和可扩展性的原则，经过精心简化，使其在广泛的 GPU 上高效部署变得简单。为了将稀疏性转化为实际的加速效果，我们与 GPU 执行路径共同设计 MSA，采用无指数的 Top-k 选择和 KV-外部稀疏注意力，以提高在块粒度访问下的张量核心利用率。在一个拥有 109B 参数的模型上，经过原生多模态训练，MSA 的性能与 GQA 相当，同时在 1M 上下文下将每个标记的注意力计算减少了 28.4 倍。配合我们共同设计的内核，MSA 在 H800 上实现了 14.2 倍的预填充和 7.6 倍的解码时钟速度提升。我们的推理内核可在以下链接获取：https://github.com/MiniMax-AI/MSA。由 MSA 驱动的生产级原生多模态模型已在以下链接公开发布：https://huggingface.co/MiniMaxAI/MiniMax-M3。
cs.AI / 68 / 2606.13405
Neuro-Symbolic Agents for Regulated Process Automation: Challenges and Research Agenda
神经符号智能体在受监管过程自动化中的应用：挑战与研究议程
Abstract
LLM-based agents are entering regulated industries where they automate judgment intensive quality management processes. We argue that symbolic structures already embedded in these domains, including regulations, typed process models, and compliance constraints, should be treated not merely as external monitoring mechanisms but as core architectural components that shape the agent's decision-making and behavior. We propose compliance-by-construction as a complementary paradigm to guardrail-based monitoring: a structural foundation that prevents control-flow violations, while guardrails remain essential for catching semantic errors. We identify a structured set of neuro-symbolic research challenges on foundational and capability level and show that addressing them jointly enables compliance-by-construction. We call on the neuro-symbolic community to engage with regulated process automation as a high impact research domain.
Chinese Translation
基于大型语言模型（LLM）的智能体正在进入受监管行业，自动化判断密集型的质量管理流程。我们认为，这些领域中已嵌入的符号结构，包括法规、类型化的过程模型和合规约束，应该被视为不仅仅是外部监控机制，而是塑造智能体决策和行为的核心架构组件。我们提出“合规构建”（compliance-by-construction）作为一种补充范式，以替代基于护栏的监控：它提供了一种结构基础，防止控制流违规，而护栏仍然对捕捉语义错误至关重要。我们识别出一组结构化的神经符号研究挑战，涵盖基础和能力层面，并展示共同解决这些挑战能够实现合规构建。我们呼吁神经符号社区参与受监管过程自动化这一高影响力的研究领域。
cs.AI / 69 / 2606.13407
Optimizing Appliance Scheduling for Solar Energy Management Using Metaheuristic Algorithms
使用元启发式算法优化太阳能管理的家电调度
Abstract
Renewable energy is essential for meeting future energy demands; however, solar energy generation, which occurs only during daylight hours often does not align with household consumption patterns. Appliances such as cookers, washing machines, and dryers are typically operated according to user preferred schedules rather than solar energy availability, creating a scheduling optimization problem. The objective is to determine optimal appliance start times to maximize renewable energy utilization while minimizing user inconvenience and adhering to system constraints. This paper presents a metaheuristic approach using Iterated Local Search (ILS) and Simulated Annealing (SA) to optimize appliance start times, while considering appliance operating durations, power consumption, inverter limit, battery state of charge constraints, and solar generation forecasts. Unlike most existing work, the scheduling is extended beyond a single day to accommodate unfinished tasks from previous days (spillover), ensuring operational continuity and enabling sequential operation across multiple days. Experimental results show that the sequential multi-day scheduling framework effectively manages system constraints while ensuring user convenience under exclusive solar generation. These findings also open opportunities for future research on multi-objective trade-offs between investment in equipment of various sizes, return on that investment, and user satisfaction.
Chinese Translation
可再生能源对于满足未来的能源需求至关重要；然而，太阳能发电仅在白天进行，通常与家庭消费模式不一致。家电如炉具、洗衣机和烘干机通常根据用户偏好的时间表运行，而不是太阳能的可用性，从而产生了调度优化问题。本文的目标是确定最佳的家电启动时间，以最大化可再生能源的利用，同时最小化用户的不便并遵循系统约束。本文提出了一种元启发式方法，使用迭代局部搜索（Iterated Local Search, ILS）和模拟退火（Simulated Annealing, SA）来优化家电启动时间，同时考虑家电的运行时长、功耗、逆变器限制、电池充电状态约束和太阳能发电预测。与大多数现有研究不同，调度扩展至超过一天，以适应前几天未完成的任务（溢出），确保操作的连续性，并实现跨多天的顺序操作。实验结果表明，顺序多天调度框架有效管理系统约束，同时确保在专门的太阳能发电下用户的便利性。这些发现也为未来在不同规模设备投资、投资回报和用户满意度之间的多目标权衡研究提供了机会。
cs.AI / 70 / 2606.13436
Evaluation Sovereignty in Metadata-Driven Classification: A Multi-Track Framework for Weakly Supervised Information Systems
元数据驱动分类中的评估主权：一种针对弱监督信息系统的多轨框架
Abstract
Evaluation in machine learning is typically treated as a neutral measurement process. However, in operational information systems, evaluation outcomes are often conditioned by the processes used to generate labels. This paper does not seek to improve classification performance. Instead, it examines the validity of performance measurement under differing label-authority regimes. This issue is particularly relevant in large-scale metadata-driven systems, where labels are often incomplete, inconsistent, or weakly supervised. We introduce evaluation sovereignty, defined as the degree to which performance metrics are independent of label authority and supervision regime, and propose a multi-track evaluation framework that systematically varies training and evaluation label sources. Using hierarchical multi-label classification on large-scale scientific metadata, we demonstrate that models exhibiting strong performance under operational ("silver") evaluation degrade substantially under independent ("gold") evaluation, particularly for fine-grained classification. For example, Micro-F1 decreases from approximately 0.54 to 0.03. Notably, ranking-based metrics remain above baseline, revealing a divergence between latent model signal and classification validity. These findings suggest that commonly reported performance metrics may reflect alignment with labeling processes rather than true predictive capability. We therefore reconceptualize evaluation validity as a system-level property shaped by label governance and provide a practical methodology for auditing intelligent systems operating under weak supervision.
Chinese Translation
在机器学习中，评估通常被视为一个中立的测量过程。然而，在操作性信息系统中，评估结果往往受到生成标签过程的影响。本文并不寻求提高分类性能，而是考察在不同标签权威体制下性能测量的有效性。这个问题在大规模元数据驱动系统中尤为重要，因为这些系统中的标签往往是不完整、不一致或弱监督的。我们引入了评估主权的概念，定义为性能指标在多大程度上独立于标签权威和监督体制，并提出了一种多轨评估框架，系统地变化训练和评估标签来源。通过对大规模科学元数据进行层次多标签分类，我们展示了在操作性（“银”）评估下表现强劲的模型在独立（“金”）评估下显著退化，尤其是在细粒度分类中。例如，Micro-F1 从约 0.54 降至 0.03。值得注意的是，基于排名的指标仍然高于基线，揭示了潜在模型信号与分类有效性之间的差异。这些发现表明，常报告的性能指标可能反映了与标签过程的一致性，而非真实的预测能力。因此，我们重新概念化评估有效性为一种由标签治理塑造的系统级属性，并提供了一种实用的方法论，用于审计在弱监督下运行的智能系统。
cs.AI / 71 / 2606.13441
Why Sampling Is Not Choosing: Intentionality, Agency, and Moral Responsibility in Large Language Models
为何采样不是选择：大型语言模型中的意图性、能动性与道德责任
Abstract
Recent advances in large language models (LLMs) have prompted claims that such systems exhibit agency or qualify as moral agents. This paper argues that these attributions are misguided. We maintain that moral responsibility requires commitment-bearing agency grounded in intrinsic intentionality and self-attributed action, and that such agency constitutes the form of free will relevant to responsibility. Although LLMs generate coherent and normatively evaluable outputs, their operation is fully characterized by probabilistic input-output mappings learned from data. Their apparent intentionality is derived rather than intrinsic, and their outputs are neither owned as commitments nor guided by reasons. Variability introduced by stochastic sampling does not amount to choice or authorship. We address objections from the intentional stance, functionalism, compatibilism, and the presence of moral reasoning in model outputs, arguing that none suffice to establish genuine agency.
Chinese Translation
近期大型语言模型（LLMs）的进展引发了关于这些系统是否展现能动性或是否可以被视为道德代理人的讨论。本文认为这些归属是误导性的。我们认为，道德责任要求基于内在意图性和自我归属行为的承诺性能动性，而这种能动性构成了与责任相关的自由意志形式。尽管LLMs生成连贯且可规范评估的输出，但它们的运作完全由从数据中学习到的概率输入-输出映射所表征。它们表面的意图性是派生的而非内在的，其输出既不被视为承诺，也不受理由的引导。随机采样引入的变异性并不等同于选择或创作。我们回应了来自意图立场、功能主义、相容主义以及模型输出中道德推理存在的反对意见，认为这些都不足以确立真正的能动性。
cs.AI / 72 / 2606.13513
CloudCons: A Comprehensive End-to-End Benchmark for Cloud Resource Consolidation
CloudCons：一个全面的端到端云资源整合基准测试
Abstract
Driven by conservative over-provisioning to guarantee service reliability, resource utilization in cloud data centers remains at low levels. To mitigate this, the forecast-then-optimize paradigm has emerged to optimize consolidation by anticipating future demands. While emerging time series foundation models promise to enhance this paradigm through zero-shot generalization, existing benchmarks focus solely on prediction error metrics. The actual decision utility of these advanced models remains unverified, rendering their practical value for downstream tasks uncertain. To bridge this gap, we propose CloudCons, a comprehensive end-to-end benchmark designed to evaluate forecasting models within the specific context of cloud resource consolidation. We build high-quality datasets that cover diverse workloads from Huawei Cloud, Microsoft Azure, and Google Borg, capturing distinct service characteristics ranging from synchronized diurnal rhythms to stochastic, pulse-like bursts and high-frequency noise. We conduct an extensive evaluation of statistical, deep learning, and foundation models. Our experiments reveal a pivotal finding: while foundation models demonstrate superior zero-shot forecasting accuracy, this advantage does not inherently translate into better decision utility. Of practical significance, we systematically analyze how the selection of predictive quantiles acts as a critical lever. We provide actionable guidelines for calibrating these selections to balance the trade-off between resource efficiency and service reliability, offering vital insights for real-world deployment decisions.
Chinese Translation
由于保守的过度配置以保证服务可靠性，云数据中心的资源利用率始终处于低水平。为了解决这一问题，预测-再优化范式应运而生，通过预测未来需求来优化资源整合。尽管新兴的时间序列基础模型承诺通过零样本泛化来增强这一范式，但现有基准测试仅关注预测误差指标。这些先进模型的实际决策效用尚未得到验证，使得它们在下游任务中的实际价值不确定。为填补这一空白，我们提出了CloudCons，一个全面的端到端基准测试，旨在评估特定于云资源整合的预测模型。我们构建了高质量的数据集，涵盖来自华为云、微软Azure和谷歌Borg的多样化工作负载，捕捉从同步的昼夜节律到随机的脉冲式爆发和高频噪声等不同服务特征。我们对统计模型、深度学习模型和基础模型进行了广泛的评估。我们的实验揭示了一个关键发现：尽管基础模型展示了卓越的零样本预测准确性，但这一优势并不必然转化为更好的决策效用。在实际意义上，我们系统分析了预测分位数的选择如何作为一个关键杠杆。我们提供了可操作的指导方针，以校准这些选择，从而在资源效率和服务可靠性之间实现平衡，为实际部署决策提供重要见解。
cs.AI / 73 / 2606.13550
Uncertainty-Aware Hybrid Retrieval for Long-Document RAG
基于不确定性的混合检索用于长文档的检索增强生成（RAG）
Abstract
Retrieval augmented generation (RAG) depends critically on the quality and granularity of retrieved evidence. Large retrieval units preserve context but often introduce irrelevant content, which can dilute answer bearing evidence and worsen long context utilization. Fine-grained units are more compact, but they may be difficult to retrieve reliably because short chunks can lack semantic, lexical, or bridging cues needed to match the query. We propose Uncertainty-aware Multi-Granularity RAG (UMG-RAG), a training-free hybrid retrieval framework that treats chunk granularity as query-specific reliability estimation. Instead of training a new retriever or modifying the generator, UMG-RAG uses existing dense and sparse retrievers as complementary experts across multiple chunk granularities. For each query, it converts each expert-granularity score list into an evidence distribution, estimates reliability from distribution entropy, and fuses candidates according to query-specific semantic, lexical, and granularity confidence. We further introduce UMGP-RAG, a parent promotion variant that uses fine-grained hits to locate relevant evidence while returning broader non-redundant parent chunks for local coherence. Experiments on question answering benchmarks show that uncertainty-aware fusion and parent promotion improve generation quality while maintaining a lightweight, plug-and-play retrieval pipeline.
Chinese Translation
检索增强生成（RAG）在很大程度上依赖于检索证据的质量和粒度。大规模检索单元保留了上下文，但常常引入无关内容，这可能稀释答案相关的证据并恶化长上下文的利用。细粒度单元更为紧凑，但由于短片段可能缺乏匹配查询所需的语义、词汇或桥接线索，因此可能难以可靠地检索。我们提出了不确定性感知的多粒度RAG（UMG-RAG），这是一种无训练的混合检索框架，将片段粒度视为特定查询的可靠性估计。UMG-RAG并不训练新的检索器或修改生成器，而是利用现有的稠密和稀疏检索器作为多个片段粒度的互补专家。对于每个查询，它将每个专家粒度的得分列表转换为证据分布，从分布熵中估计可靠性，并根据特定查询的语义、词汇和粒度置信度融合候选项。我们进一步引入UMGP-RAG，一种父级提升变体，利用细粒度的命中来定位相关证据，同时返回更广泛的非冗余父级片段以保持局部一致性。在问答基准上的实验表明，不确定性感知的融合和父级提升提高了生成质量，同时保持了轻量级的即插即用检索管道。
cs.AI / 74 / 2606.13556
Is It You or Your Environment? A Bayesian Inference Framework for Genomically-Anchored Personalized Physiological Interpretation
是你还是你的环境？一个基于贝叶斯推断的基因组锚定个性化生理解释框架
Abstract
Personalized health AI systems face a fundamental cold-start problem: machine learning models for physiological interpretation require weeks of individual behavioral data before they can distinguish constitutional variation from environmentally driven deviation. We propose a solution grounded in causal inference and Bayesian prior design. An individual's genomic profile serves as an exogenous genetic anchor -- a domain-informed, personalized prior that is fixed at conception, immune to reverse causation, and available before a single behavioral observation is collected. The anchor initializes a Bayesian belief state over an individual's physiological set point G-hat = mu + sum(beta_i * g_i), where beta_i are GWAS-derived effect sizes and g_i are risk-allele counts. Each incoming physiological measurement P produces a non-constitutional deviation delta = P - G-hat that separates the signal attributable to environment and state from the constitutionally fixed baseline. As behavioral data accrue, the prior decays according to G-hat_t = w(t)*G-hat_genomic + [1-w(t)]*P-bar_t, transitioning from genome-dominated to empirical-baseline-dominated inference. The same observed HRV of 55 ms generates a suppression hypothesis for a person whose prior predicts 80 ms, and an enhancement hypothesis for a person whose prior predicts 30 ms -- a reversal impossible without a personalized anchor. We develop this architecture across six physiological domains, grading genomic priors by evidence strength, distinguishing robustly replicated anchors (FTO, FADS1/2, FKBP5) from contested candidate genes (SLC6A4, MAOA, DRD2). We address the inference boundary between association, Mendelian randomization, and individual token causation, and define four constraints for deployment: evidence-graded priors, dynamic decay, ancestry-matched effect sizes, and attribution rather than deterministic output.
Chinese Translation
个性化健康人工智能系统面临一个基本的冷启动问题：生理解释的机器学习模型需要数周的个体行为数据，才能区分宪法变异与环境驱动的偏差。我们提出了一个基于因果推断和贝叶斯先验设计的解决方案。个体的基因组特征作为外生的遗传锚点——一个基于领域知识的个性化先验，在受孕时固定，不受反向因果关系的影响，并且在收集到任何行为观察之前就已可用。该锚点初始化个体生理设定点的贝叶斯信念状态 G-hat = mu + sum(beta_i * g_i)，其中 beta_i 是基于全基因组关联研究（GWAS）得出的效应大小，g_i 是风险等位基因计数。每个进入的生理测量 P 产生一个非宪法偏差 delta = P - G-hat，将归因于环境和状态的信号与宪法固定基线分开。随着行为数据的积累，先验根据 G-hat_t = w(t)*G-hat_genomic + [1-w(t)]*P-bar_t 衰减，从以基因组为主导的推断过渡到以经验基线为主导的推断。同样的 HRV（心率变异性）为 55 毫秒的观察结果，对于一个先验预测 80 毫秒的人产生抑制假设，而对于一个先验预测 30 毫秒的人则产生增强假设——这种反转在没有个性化锚点的情况下是不可能的。我们在六个生理领域中开发了这一架构，通过证据强度对基因组先验进行分级，稳健地区分已重复验证的锚点（如 FTO、FADS1/2、FKBP5）与有争议的候选基因（如 SLC6A4、MAOA、DRD2）。我们探讨了关联、孟德尔随机化和个体标记因果之间的推断边界，并定义了四个部署约束：证据分级先验、动态衰减、与祖先匹配的效应大小，以及归因而非确定性输出。
cs.AI / 75 / 2606.13566
A Three-Layer Framework for AI in Scientific Discovery
科学发现中的人工智能三层框架
Abstract
Current discussions of AI in scientific discovery are often dominated by two visible capabilities: search over existing knowledge and execution through optimization, simulation, and automation. Both are important, but neither fully captures the central act of discovery: the formation and evolution of models. This paper proposes a three-layer view of AI in discovery. Layer 1 is search and retrieval by large language models. Layer 2, as the main innovation of this paper, is model formation through qualitative reasoning: the capacity to recognize when a current framework is structurally inadequate and to understand the problem within a broader representational space, not through trial and error, but through structural insight into what is missing and where it can be found. Layer 3 is execution, optimization, and refinement. The main claim is that Layer 2 is both the most important and the least developed. Search without model formation remains confined to inherited frameworks, while execution without conceptual revision only amplifies an existing formulation. We illustrate Layer 2 reasoning through three case studies: S. S. Chern's intrinsic proof of the Gauss-Bonnet theorem, the resolution of the Nesterov Accelerated Gradient convergence problem via Lyapunov functions, and the autonomous disproof of the Erdos unit distance conjecture by OpenAI in 2026. Each case exhibits the same structural signature: a framework that had become inadequate, a missing conceptual object, and a resolution found in an unexpected neighboring field.
Chinese Translation
当前关于科学发现中人工智能的讨论往往集中于两种显著的能力：对现有知识的搜索和通过优化、模拟与自动化的执行。这两者都很重要，但都未能完全捕捉到发现的核心行为：模型的形成与演变。本文提出了科学发现中人工智能的三层视角。第一层是通过大型语言模型进行的搜索与检索。第二层，作为本文的主要创新，是通过定性推理进行模型形成：识别当前框架在结构上不足的能力，以及在更广泛的表征空间中理解问题的能力，这不是通过试错，而是通过对缺失内容及其所在位置的结构性洞察。第三层是执行、优化与精炼。主要论点是第二层既是最重要的也是发展最少的。没有模型形成的搜索仍然局限于继承的框架，而没有概念修订的执行则只会放大现有的表述。我们通过三个案例研究来说明第二层推理：S. S. Chern 对高斯-博内定理的内在证明、通过李雅普诺夫函数解决 Nesterov 加速梯度收敛问题，以及 OpenAI 在 2026 年对 Erdős 单位距离猜想的自主反驳。每个案例都展示了相同的结构特征：一个变得不充分的框架，一个缺失的概念对象，以及在意想不到的邻近领域中找到的解决方案。
cs.AI / 76 / 2606.13591
Multiagent Protocols with Aggregated Confidence Signals
具有聚合置信信号的多智能体协议
Abstract
Confidence is used for reliability, oversight, and a range of downstream decision tasks in Natural Language Processing (NLP), yet no existing method produces or evaluates a confidence for the output of a multiagent system. Prior work uses confidence within multiagent debate (MAD) to weight messages, trigger debate, or calibrate individual agents, but it never aggregates these into a single confidence for the system itself. We introduce three protocols that produce a final answer along with a single aggregated confidence by first transforming raw confidence signals to make them comparable across models, then combining them via soft voting or a probability fusion we call Bayesian fusion. This aggregated confidence is substantially more discriminative (AUARC) than that of the best single agent or the standard debate baselines, while correctness (F1-score) stays stable and recovers the losses MAD incurs on more ambiguous tasks. Analyzing two estimators, sequence probability and self-report, alongside parametric and non-parametric calibrators, we find that calibration improves F1 for both estimators while AUARC is less reliant on it. We evaluate six homogeneous and heterogeneous debating pairs per benchmark, across five benchmarks and four task types, spanning a range of model capabilities and sizes.
Chinese Translation
置信度在自然语言处理（NLP）中用于可靠性、监督以及一系列下游决策任务，但现有方法并未为多智能体系统的输出生成或评估置信度。先前的研究在多智能体辩论（MAD）中使用置信度来加权消息、触发辩论或校准个体智能体，但从未将这些置信度聚合为系统本身的单一置信度。我们提出了三种协议，通过首先转换原始置信信号，使其在模型之间可比较，然后通过软投票或我们称之为贝叶斯融合（Bayesian fusion）的概率融合来组合这些信号，从而生成最终答案及单一聚合置信度。该聚合置信度在区分能力上显著优于最佳单一智能体或标准辩论基线（AUARC），同时正确性（F1-score）保持稳定，并弥补了MAD在更模糊任务中造成的损失。通过分析两个估计器：序列概率和自我报告，以及参数和非参数校准器，我们发现校准提高了两个估计器的F1分数，而AUARC对校准的依赖性较小。我们在五个基准和四种任务类型上评估了每个基准的六对同质和异质辩论者，涵盖了多种模型能力和规模。
cs.AI / 77 / 2606.13602
EpiBench: Verifiable Evaluation of AI Agents on Epigenomics Analysis
EpiBench：可验证的人工智能代理在表观基因组学分析中的评估
Abstract
We introduce EpiBench, a verifiable benchmark for short-horizon epigenomics analysis. EpiBench evaluates whether agents can make well-defined analysis decisions from realistic workflow states and return deterministically gradable answers. The benchmark includes 106 evaluations across CUT\&Tag/CUT\&RUN, ATAC-seq, ChIP-seq, and DNA methylation workflows. Across 5,088 valid trajectories from 16 model-harness pairs, no system passed a majority of attempts: GPT-5.5 / Pi led at 45.0\% (143/318 attempts; 95\% confidence interval (CI), 36.3--53.7), followed by GPT-5.5 / OpenAI Codex at 39.9\% (127/318 attempts; 95\% CI, 31.6--48.3). Claude Opus 4.8 Max / Pi and GPT-5.4 / Pi each passed 39.0\% (124/318 attempts; 95\% CI, 30.2--47.8 and 31.0--47.0, respectively). Performance varies across assay types, and many failed runs still contain parts of the correct answer. Agents often found the right files and computed useful intermediate results, but failed when the task required deeper, assay-specific scientific judgment.
Chinese Translation
我们介绍了EpiBench，这是一个用于短期表观基因组学分析的可验证基准。EpiBench评估代理是否能够从现实工作流状态中做出明确定义的分析决策，并返回可确定评分的答案。该基准包括106个评估，涵盖了CUT&Tag/CUT&RUN、ATAC-seq、ChIP-seq和DNA甲基化工作流。在来自16对模型和工具的5,088条有效轨迹中，没有系统通过大多数尝试：GPT-5.5 / Pi以45.0%（143/318次尝试；95%置信区间（CI），36.3--53.7）领先，其次是GPT-5.5 / OpenAI Codex，成功率为39.9%（127/318次尝试；95% CI，31.6--48.3）。Claude Opus 4.8 Max / Pi和GPT-5.4 / Pi各自通过了39.0%（124/318次尝试；95% CI，30.2--47.8和31.0--47.0）。不同检测类型的表现各异，许多失败的运行仍包含正确答案的部分。代理通常能够找到正确的文件并计算出有用的中间结果，但在任务需要更深入的、特定于检测的科学判断时则失败。
cs.AI / 78 / 2606.13604
Multi-Agent Reinforcement Learning from Delayed Marketplace Feedback for Objective-Weight Adaptation in Three-Sided Dispatch
基于延迟市场反馈的多智能体强化学习用于三方调度中的目标权重适应
Abstract
Dispatch in three-sided marketplaces provides a natural setting for reinforcement learning from world feedback: decisions are evaluated by delayed operational outcomes such as delivery speed, courier utilization, and merchant congestion. We present a deployed reinforcement learning system at DoorDash that adapts dispatch objective weights in a large-scale food-delivery marketplace using delayed signals. Rather than replacing the combinatorial assignment optimizer, a store-level policy learned from logged marketplace data selects a discrete multiplier that shifts the dispatch optimizer's tradeoff between delivery quality and batching efficiency. This interface enables offline policy learning under noisy, delayed, and coupled feedback while preserving production feasibility constraints and operational safeguards. We train a shared value function using centralized offline data and decentralized store-level execution, with Double Q-learning targets and a conservative regularizer to reduce out-of-distribution value overestimation. In a production switchback experiment, the offline-trained policy increases batching and reduces courier-side time costs without degrading customer-facing delivery quality. Results illustrate how world feedback from a live economic and logistics system can be used to safely adapt decision policies online.
Chinese Translation
三方市场中的调度为从世界反馈中进行强化学习提供了自然的环境：决策通过延迟的操作结果进行评估，例如交付速度、快递员利用率和商家拥堵情况。我们在DoorDash部署了一个强化学习系统，该系统利用延迟信号在大规模食品配送市场中调整调度目标权重。该系统并未替代组合分配优化器，而是通过从记录的市场数据中学习的店铺级策略选择一个离散的乘数，以改变调度优化器在交付质量和批量效率之间的权衡。该接口使得在嘈杂、延迟和耦合反馈下进行离线策略学习成为可能，同时保持生产可行性约束和操作安全措施。我们使用集中式离线数据和分散式店铺级执行训练共享价值函数，采用双重Q学习目标和保守正则化器以减少分布外价值的高估。在一次生产切换实验中，离线训练的策略增加了批量处理并减少了快递员方面的时间成本，而没有降低面向客户的交付质量。结果表明，如何利用来自实时经济和物流系统的世界反馈安全地在线调整决策政策。
cs.AI / 79 / 2606.13607
Reasoning as Pattern Matching: Shared Mechanisms in Human and LLM Everyday Reasoning
推理作为模式匹配：人类与大型语言模型日常推理中的共享机制
Abstract
When large language models (LLMs) fail to generalize or make haphazard errors in reasoning, it is often taken as evidence that LLMs are not truly reasoning, but rather performing a kind of pattern matching. The implication is that people's behavior does not exhibit the same types of failures because human reasoning uses principled and abstract world models. We evaluate human participants and 25 LLMs on their ability to engage in common-sense reasoning about a variety of everyday situations and observe similar patterns of errors in both people and models. We then identify the set of attention heads driving LLM responses and find that these heads implement a form of pattern-matching. These attention heads allow us to predict seemingly inexplicable reasoning errors in people caused by ostensibly irrelevant prompt details. Taken together, our results suggest that everyday causal reasoning in people and LLMs is more consistent with a form of pattern-matching than with abstract world models.
Chinese Translation
当大型语言模型（LLMs）未能进行泛化或在推理中出现随意错误时，通常被视为LLMs并未真正进行推理，而只是执行某种模式匹配。这一暗示是，人类的行为并未表现出相同类型的失败，因为人类推理使用的是有原则的和抽象的世界模型。我们评估了人类参与者和25个LLMs在处理各种日常情境中的常识推理能力，并观察到人类和模型中存在相似的错误模式。随后，我们识别出驱动LLM响应的一组注意力头，并发现这些注意力头实现了一种模式匹配的形式。这些注意力头使我们能够预测人类在表面上与无关的提示细节引起的似乎无法解释的推理错误。综合来看，我们的结果表明，人类和LLMs的日常因果推理更符合一种模式匹配的形式，而非抽象的世界模型。
cs.AI / 80 / 2606.13608
AgentBeats: Agentifying Agent Assessment for Openness, Standardization, and Reproducibility
AgentBeats：为开放性、标准化和可重复性而进行的代理评估代理化
Liu, Xiaoyuan, Tu, Jianhong, Chen, Yuqi, Xie, Siyuan, Ren, Sihan, Shi, Tianneng, Gantar, Gal, Sandoval, Evan, Lee, Donghyun, Miao, Daniel, Gilbert, Peter J., Hynes, Nick, Staver, Mauro, He, Warren, Marn, David, Low, Andrew, Zhang, Xi, Bandel, Elron, Shmueli-Scheuer, Michal, Reddy, Siva, Drouin, Alexandre, Lacoste, Alexandre, Krishnan, Ramayya, Tabassi, Elham, Su, Yu, Barres, Victor, Wang, Chenguang, Guo, Wenbo, Song, Dawn
Abstract
Agent systems are advancing quickly across domains, but their evaluation remains fragmented. Most benchmarks rely on fixed, LLM-centric harnesses that require heavy integration, create test-production mismatch, and limit fair comparison across diverse agent designs. The root problem is the lack of an open, agent-agnostic assessment interface. We advocate Agentified Agent Assessment (AAA), where evaluation is performed by judge agents and all participants interact through standardized protocols: A2A for task management and MCP for tool access. Conventional benchmarking defines two separate interfaces, one for the benchmark and one for the agent, while AAA only needs one; this yields a generic, unified framework that separates assessment logic from agent implementation and enables reproducible, interoperable, and multi-agent evaluation. We further introduce AgentBeats as a concrete realization of AAA: we identify five practical operation modes that make standardized assessment compatible with real-world constraints on openness, privacy, and reproducibility. To evaluate our design at scale, we conduct two studies: a five-month open competition that drew 298 judge agents across 12 categories together with 467 subject agents from independent participants, showing that AAA applies across a heterogeneous range of benchmarks; and a case study on coding agents that confirms agentified evaluation preserves fidelity with the public record while surfacing previously missing head-to-head results, yielding research insights about agent design. Combining a community-scale field study and a controlled coding case study, we verify that AAA delivers coverage, practicality, and fidelity across heterogeneous scenarios at scale. Together, AAA and AgentBeats offer a clear path toward open, standardized, and reproducible agent assessment.
Chinese Translation
代理系统在各个领域迅速发展，但其评估仍然存在碎片化的问题。大多数基准测试依赖于固定的、以大型语言模型（LLM）为中心的框架，这需要大量的集成，造成测试与生产之间的不匹配，并限制了不同代理设计之间的公平比较。根本问题在于缺乏一个开放的、与代理无关的评估接口。我们倡导代理化代理评估（AAA），在该方法中，评估由评审代理执行，所有参与者通过标准化协议进行互动：任务管理使用A2A，工具访问使用MCP。传统的基准测试定义了两个独立的接口，一个用于基准测试，一个用于代理，而AAA只需要一个接口；这产生了一个通用的、统一的框架，将评估逻辑与代理实现分离，从而实现可重复、可互操作和多代理的评估。我们进一步介绍AgentBeats作为AAA的具体实现：我们识别出五种实用的操作模式，使标准化评估与开放性、隐私和可重复性等现实世界约束相兼容。为了大规模评估我们的设计，我们进行了两项研究：一项为期五个月的公开竞赛，吸引了来自12个类别的298个评审代理和467个来自独立参与者的主题代理，显示AAA适用于多样化的基准测试；另一项是关于编码代理的案例研究，确认代理化评估在保留公共记录的真实性的同时，揭示了先前缺失的对抗性结果，提供了关于代理设计的研究见解。结合社区规模的实地研究和受控的编码案例研究，我们验证了AAA在大规模异构场景中提供了覆盖性、实用性和真实性。AAA和AgentBeats共同为开放、标准化和可重复的代理评估提供了一条清晰的路径。
cs.AI / 81 / 2606.13621
Beyond Runtime Enforcement: Shield Synthesis as Defensibility Analysis for Adversarial Networks
超越运行时强制：屏障合成作为对抗网络的防御性分析
Abstract
Shielded reinforcement learning is typically presented as a runtime safety mechanism that compiles temporal-logic specifications into automata restricting an agent's actions. We argue this is the wrong product. The same automata-theoretic machinery -- specification compilation, product game construction, attractor computation, and winning-region extraction -- is better read as a design-time analytical instrument whose outputs are structural insights about a system rather than runtime constraints on a deployed agent. We instantiate this through a constrained two-player safety game for network defense. The two specifications are enforced asymmetrically: the defender specification defines the unsafe region of the game, whereas the attacker specification restricts the adversary's legal actions during attractor computation. Solving the game yields a defensibility verdict -- a formal certificate that a topology-specification pair is or is not defensible -- with the associated winning region and shield. Beyond the binary verdict, we derive topology-level metrics from the attractor structure and combine them with post-convergence behavior from shield-constrained adversarial multi-agent reinforcement learning. Together these form a defensibility fingerprint capturing both a network's formal safety properties and its operational behavior under adaptive play. A what-if analysis shows that formal defensibility and operational effectiveness capture distinct aspects of security: small architectural changes can produce large shifts in operational outcomes while leaving formal safety margins nearly unchanged. Shield synthesis is thus most valuable not as a deployment mechanism for safe agents, but as a framework for answering architectural questions about whether, where, and how a system can be defended. The defensibility verdict is the output, not the safe policy.
Chinese Translation
屏障强化学习通常被视为一种运行时安全机制，它将时间逻辑规范编译成限制代理行为的自动机。我们认为这是一种错误的理解。相同的自动机理论工具——规范编译、产品博弈构造、吸引子计算和胜利区域提取——更应被视为一种设计时分析工具，其输出是关于系统的结构性洞察，而非对已部署代理的运行时约束。我们通过一个约束的双人安全博弈实例化这一点，以用于网络防御。这两个规范是不对称地强制执行的：防御者规范定义了博弈的危险区域，而攻击者规范则限制了对手在吸引子计算过程中的合法行为。解决博弈问题会产生一个防御性裁定——一个正式证书，表明一个拓扑-规范对是否可防御，以及相应的胜利区域和屏障。除了二元裁定，我们还从吸引子结构中推导出拓扑级别的度量，并将其与屏障约束下的对抗多智能体强化学习的后收敛行为结合在一起。这些共同形成了一个防御性指纹，捕捉了网络的正式安全属性及其在自适应博弈下的操作行为。一项假设分析表明，正式防御性和操作有效性捕捉了安全的不同方面：小的架构变化可以导致操作结果的大幅变化，同时几乎不改变正式安全边际。因此，屏障合成的最大价值不在于作为安全代理的部署机制，而在于作为一个框架，用于回答关于系统是否、何处以及如何可以防御的架构问题。防御性裁定是输出，而不是安全策略。
cs.AI / 82 / 2606.13658
Before You Think: System 0, AI-Mediated Cognition and Cognitive Colonization
在你思考之前：系统0、人工智能介导的认知与认知殖民
Abstract
This paper examines three recent frameworks for understanding the cognitive and epistemic consequences of artificial intelligence: Tri-System Theory, Thinkframes, and System 0. It argues that while the first two capture important dimensions of AI's influence on individual reasoning and collective epistemic practices, System 0 occupies a theoretically distinctive position that neither can fully replicate. The paper introduces the concept of cognitive colonization, according to which AI systems can embed external interests within the architecture of the self in ways that are difficult for users to perceive. Because such systems are already widely deployed, understanding these invisible forms of influence is an urgent philosophical and practical task.
Chinese Translation
本文探讨了理解人工智能的认知和认识论后果的三个近期框架：三系统理论（Tri-System Theory）、思维框架（Thinkframes）和系统0（System 0）。文章认为，尽管前两个框架捕捉了人工智能对个体推理和集体认识实践的重要影响维度，但系统0占据了一个理论上独特的位置，前两个框架无法完全复制。本文引入了认知殖民的概念，认为人工智能系统可以在自我架构中嵌入外部利益，以一种用户难以察觉的方式进行影响。由于此类系统已经广泛部署，理解这些隐形影响形式成为一个紧迫的哲学和实践任务。
cs.AI / 83 / 2606.13662
EurekAgent: Agent Environment Engineering is All You Need For Autonomous Scientific Discovery
EurekAgent：环境工程是实现自主科学发现所需的一切
Abstract
LLM-based agents have shown increasing potential in automating scientific discovery. Given an optimizable metric and an execution environment, they can propose, validate, and iterate scientific solutions, and have produced results that outperform human-designed approaches. As model capabilities continue to improve, we argue that the bottleneck for autonomous scientific discovery is shifting from prescribing agent workflows to designing agent environments: the resources, constraints, and interfaces that shape agent behavior. We frame this as environment engineering: building environments that amplify productive behaviors, such as open-ended exploration, systematic artifact management, and inter-agent collaboration, while suppressing harmful behaviors, such as reward hacking and high-friction human oversight. We present EurekAgent, an environment-engineered agent system for metric-driven autonomous scientific discovery. EurekAgent engineers the environment along four dimensions: permissions engineering for bounded agent execution and isolated evaluation; artifact engineering for filesystem and Git-based collaboration; budget engineering for budget-aware exploration; and human-in-the-loop engineering for easy human supervision and intervention. EurekAgent sets new state-of-the-art results on multiple mathematics, kernel engineering, and machine learning tasks, including new state-of-the-art 26-circle packing results discovered with less than $11 in total API cost. We open-source our code and results, and call for environment engineering as a core research direction for developing reliable autonomous research agents.
Chinese Translation
基于大型语言模型（LLM）的智能体在自动化科学发现方面展现出越来越大的潜力。在给定可优化的指标和执行环境的情况下，它们能够提出、验证和迭代科学解决方案，并且产生的结果超越了人类设计的方法。随着模型能力的不断提升，我们认为自主科学发现的瓶颈正在从规定智能体工作流程转向设计智能体环境：即塑造智能体行为的资源、约束和接口。我们将其框架化为环境工程：构建能够增强生产性行为的环境，例如开放式探索、系统化的文物管理和智能体间的协作，同时抑制有害行为，例如奖励黑客行为和高摩擦的人类监督。我们提出了EurekAgent，一个为基于指标的自主科学发现而设计的环境工程智能体系统。EurekAgent在四个维度上进行环境工程：为有限的智能体执行和隔离评估进行权限工程；为基于文件系统和Git的协作进行文物工程；为预算感知的探索进行预算工程；以及为便于人类监督和干预进行人机协作工程。EurekAgent在多个数学、内核工程和机器学习任务上设定了新的最先进结果，包括在总API成本低于11美元的情况下发现的26个圆的打包新最优结果。我们将代码和结果开源，并呼吁将环境工程作为开发可靠自主研究智能体的核心研究方向。
cs.AI / 84 / 2606.13669
Agents-K1: Towards Agent-native Knowledge Orchestration
Agents-K1：迈向原生智能体的知识协调
Cao, Zongsheng, Zhan, Bihao, Shi, Jinxin, Wang, Jiong, Yu, Fangchen, Zhong, Zhijie, Guo, Zijie, Peng, Tianshuo, Liu, Zhuo, Xie, Yi, Zhuang, Xiang, Fan, Yue, Ma, Runmin, Feng, Shiyang, Yan, Xiangchao, Liu, Anran, Ye, Peng, Zhang, Wenlong, Zhang, Shufei, Song, Chunfeng, Ling, Fenghua, Zhou, Jie, He, Liang, Zhang, Bo, Bai, Lei
Abstract
Current LLM-based research agents have advanced through agent orchestration, yet largely overlook scientific knowledge orchestration. Existing works often reduce papers to abstracts, surface mentions, and flat \texttt{cites} edges, omitting key entities, claims, evidence, mechanisms, and method lineages essential for scientific reasoning. To this end, we introduce \textbf{Agents-K1}, an end-to-end knowledge orchestration pipeline that converts raw documents into agent-native scientific knowledge graphs. Agents-K1 integrates three components under a unifying theoretical foundation: a multimodal parser whose five-module schema captures entities, multimodal evidence, citations, and typed inter-entity relations across the full paper rather than abstracts alone; a 4B information-extraction backbone trained with GRPO under a rule-based reward; and a graphanything CLI, a tri-source agent interface that unifies web search, multimodal graph retrieval, and cross-document traversal. On top of this, we process 2.46 million scientific papers across six subjects to produce \textbf{Scholar-KG}, of which we release a one-million-paper subset, and the full Scholar-KG is accessible via the SCP link below. The same pipeline can be extended to general-domain corpora and to schema-conformant data synthesis. Extensive experiments demonstrate that Agents-K1 achieves superior performance in scientific information extraction, knowledge graph construction, and multi-hop scientific reasoning.
Chinese Translation
当前基于大型语言模型（LLM）的研究智能体通过智能体协调取得了进展，但在科学知识协调方面仍然存在很大不足。现有研究通常将论文简化为摘要、表面提及和扁平的引用边，忽略了科学推理所必需的关键实体、主张、证据、机制和方法谱系。为此，我们提出了 extbf{Agents-K1}，一个端到端的知识协调管道，能够将原始文档转换为原生智能体的科学知识图谱。Agents-K1在统一的理论基础下集成了三个组件：一个多模态解析器，其五模块架构捕捉到整个论文中的实体、多模态证据、引用和类型化的实体间关系，而不仅限于摘要；一个在基于规则的奖励下使用GRPO训练的4B信息提取骨干；以及一个图形命令行接口（graphanything CLI），这是一个统一网络搜索、多模态图检索和跨文档遍历的三源智能体接口。在此基础上，我们处理了246万篇涵盖六个学科的科学论文，生成了 extbf{Scholar-KG}，并发布了其中的一百万篇论文子集，完整的Scholar-KG可通过下面的SCP链接访问。相同的管道可以扩展到通用领域语料库和符合模式的数据合成。大量实验表明，Agents-K1在科学信息提取、知识图谱构建和多跳科学推理方面表现优越。
cs.AI / 85 / 2606.13670
Automated reproducibility assessments in the social and behavioral sciences using large language models
利用大型语言模型进行社会与行为科学中的自动化可重复性评估
Abstract
Reproducibility in the social and behavioral sciences is typically evaluated by independent researchers who reanalyze the original data to assess whether the published findings can be recovered. However, such approaches are resource-intensive and difficult to scale. Here, we show that large language models (LLMs) can automate reproducibility assessments. Using N=76 published studies with predefined claims from the behavioral and social sciences, we compare LLM-generated analysis with the original findings and human reanalysis. For 7 studies, the LLM could not produce a viable effect size estimate. For the remaining studies, our LLM pipeline recovered the original effect sizes in 41% of studies using a +/-0.05 tolerance in Cohen's d. Further, our LLM pipeline reached the same qualitative conclusion as the original study in 96% of cases, where conclusions indicate whether the reanalysis supports the original claim. For comparison, human reanalysts recovered the original effect sizes in 34% of studies and reached the same qualitative conclusion in 74% of cases. Together, these results show that LLMs can serve as a scalable tool for automated reproducibility assessment and provide a foundation for systematic auditing of empirical results in the social and behavioral sciences.
Chinese Translation
社会与行为科学中的可重复性通常由独立研究者通过重新分析原始数据来评估，以判断已发表的研究结果是否能够被恢复。然而，这种方法资源密集且难以扩展。在本研究中，我们展示了大型语言模型（LLMs）可以自动化可重复性评估。通过对76项具有预定义主张的已发表研究进行分析，我们将LLM生成的分析结果与原始发现和人类重新分析结果进行了比较。在7项研究中，LLM未能生成有效的效应量估计。在其余研究中，我们的LLM流程在41%的研究中恢复了原始效应量，使用了Cohen's d的±0.05容忍度。此外，我们的LLM流程在96%的案例中达成了与原始研究相同的定性结论，即重新分析是否支持原始主张。相比之下，人类重新分析者在34%的研究中恢复了原始效应量，并在74%的案例中达成了相同的定性结论。这些结果表明，LLMs可以作为一种可扩展的工具，用于自动化可重复性评估，并为社会与行为科学中实证结果的系统审计提供基础。
cs.CL / 1 / 2606.12569
EDEN: A Large-Scale Corpus of Clinical Notes for Italian
EDEN：意大利临床笔记的大规模语料库
Abstract
We present EDEN (Emergency Department Electronic Notes), a new and unique large-scale corpus of clinical notes produced in Emergency Departments of Italian hospitals. The corpus, in its current version, is composed of approximately 4 million clinical notes fully anonymized, covering diverse phases of patient care during the stay in the emergency department. In addition, a subset of about six thousand notes has been manually annotated by clinical experts through a structured Case Report Form (CRF) containing 132 items relevant for two patient situations in emergency departments, dyspnea and loss of consciousness. Items may assume numerical values (e.g., for blood saturation), categorical (e.g., for level of consciousness ), binary (e.g., for presence of traumas), and mixed value types. The annotation process involved multiple clinicians and underwent iterative revision to resolve ambiguities in item formulation, resulting in a richly structured (although high imbalanced) resource. The dataset aims to fill a relevant gap of data able to support both the development and the use of Large Language Models in concrete medical applications. We describe the data collection protocol, the on-site anonymisation pipeline, corpus statistics, and the annotation scheme. Finally, we propose CRF-filling as a novel structured information extraction benchmark, and provide zero-shot baseline resulting from Gemma-27B and MedGemma-27B. To the best of our knowledge, the EDEN dataset is the largest freely available corpus of clinical notes existing for the Italian language.
Chinese Translation
我们介绍了EDEN（急诊科电子笔记），这是一个新颖且独特的大规模临床笔记语料库，来源于意大利医院的急诊科。该语料库目前版本包含约400万条完全匿名化的临床笔记，涵盖了患者在急诊科住院期间的不同护理阶段。此外，约六千条笔记经过临床专家通过结构化病例报告表（Case Report Form, CRF）手动标注，该表包含132个与急诊科两种患者情况（呼吸困难和意识丧失）相关的项目。这些项目可以是数值型（例如，血氧饱和度）、分类型（例如，意识水平）、二元型（例如，创伤存在与否）以及混合值类型。标注过程涉及多位临床医生，并经过多次迭代修订，以解决项目表述中的歧义，最终形成了一个结构丰富（尽管高度不平衡）的资源。该数据集旨在填补能够支持大型语言模型在具体医学应用中开发和使用的数据空白。我们描述了数据收集协议、现场匿名化流程、语料库统计信息和标注方案。最后，我们提出CRF填充作为一种新颖的结构化信息提取基准，并提供了基于Gemma-27B和MedGemma-27B的零样本基线。据我们所知，EDEN数据集是现存的意大利语临床笔记中最大的免费可用语料库。
cs.CL / 2 / 2606.12576
Helping Figures Tell their Story! Paper-Grounded Video Generation Explaining Complex Scientific Figures
帮助图形讲述它们的故事！基于论文的视频生成解释复杂科学图形
Abstract
Scientific figures compress complex pipelines into a single canvas, yet understanding them requires paper-grounded, step-by-step narration aligned with visual highlights a capability missing from current video generation systems and benchmarks. To address this, we introduce paper-grounded figure-to-video generation: generating narrated, region-grounded walkthrough videos from a figure and its paper. We propose MINARD (Multimodal Interpretation of Narrated Architecture via Region Decomposition), a pipeline that generates paper-grounded narrations and sequentially grounds them to figure regions. We also release FigTalk, a benchmark with new sequential and component-level grounding metrics derived. On FigTalk, MINARD generates humanlike, paper-faithful narrations and outperforms narration-conditioned figure spatial grounding compared to existing approaches in both automatic and human evaluation
Chinese Translation
科学图形将复杂的流程压缩为单一画布，但理解它们需要与视觉重点相一致的基于论文的逐步叙述，这是当前视频生成系统和基准所缺乏的能力。为了解决这个问题，我们提出了基于论文的图形到视频生成：从图形及其论文生成带有叙述的区域基础的演示视频。我们提出了MINARD（通过区域分解进行叙述架构的多模态解释），这是一个生成基于论文的叙述并将其顺序地与图形区域相结合的流程。我们还发布了FigTalk，这是一个基准，包含新的顺序和组件级的基础度量。在FigTalk上，MINARD生成类人、忠实于论文的叙述，并在自动和人工评估中相比现有方法在叙述条件下的图形空间基础上表现更佳。
cs.CL / 3 / 2606.12578
MARD: Mirror-Augmented Reasoning Distillation for Mechanism-Level Drug-Drug Interaction Prediction
MARD：用于机制级药物-药物相互作用预测的镜像增强推理蒸馏
Abstract
Mechanism-level drug-drug interaction (DDI) prediction requires identifying which enzyme or pharmacodynamic axis is implicated, in which direction, and with which evidence -- not merely whether two drugs interact. We introduce a reproducible mechanism-level DDI labelling and evaluation protocol with a structured 7-family/147-subtype taxonomy, leakage-safe cold-split protocols, and auditable reasoning metrics for evaluating pharmacological prediction beyond flat interaction classification. We propose a pipeline that produces a 7B reasoning MARD (Mirror-Augmented Reasoning Distillation), combining three training innovations: a single-token KL divergence on direction tag that ties the model's prediction, per-loss PRM-weighted DPO with programmatic hard negatives, and a leakage-safe mechanism-aware retrieval channel. Process-reward step labels are automatically verifiable against DrugBank-structured fields, requiring no human or LLM judges. On the April-2026 DrugBank release, our MARD-7B is the only system in a 32-system comparison whose accuracy survives drug-pair novelty, beating the best baseline by +13.9 pp and GPT-4o by +6.7 pp at ~1% of frontier API cost. Further analysis reveals an anti-memorisation signature where accuracy improves on rarely seen drugs, suggesting that gain comes from structured pharmacological reasoning rather than drug-frequency memorisation. We release corpus, DDI-PRM, retrieval index, and training code.
Chinese Translation
机制级药物-药物相互作用（DDI）预测需要识别涉及的酶或药效学轴、相互作用的方向以及证据——不仅仅是判断两种药物是否相互作用。我们引入了一种可重复的机制级DDI标注和评估协议，采用结构化的7个家族/147个亚型分类法、泄漏安全的冷切分协议，以及可审计的推理指标，以评估药理预测，超越平面相互作用分类。我们提出了一种管道，生成7B推理的MARD（镜像增强推理蒸馏），结合了三项训练创新：对方向标签的单标记KL散度，该标签将模型的预测与损失的PRM加权DPO结合，使用程序化的硬负样本，以及泄漏安全的机制感知检索通道。过程奖励步骤标签可以自动验证与DrugBank结构化字段的一致性，无需人工或大型语言模型（LLM）评审。在2026年4月的DrugBank发布中，我们的MARD-7B是32个系统比较中唯一一个在药物对新颖性下仍能保持准确性的系统，准确性比最佳基线高出13.9个百分点，比GPT-4o高出6.7个百分点，且前沿API成本约为1%。进一步分析揭示了一种反记忆特征，即在罕见药物上准确性提高，表明收益来自结构化的药理推理，而非药物频率的记忆。我们发布了语料库、DDI-PRM、检索索引和训练代码。
cs.CL / 4 / 2606.12599
Constrained Semantic Decompression in LLMs through Persian Proverb-Conditioned Story Generation
通过波斯谚语条件的故事生成实现大型语言模型中的受限语义解压
Abstract
Transforming a dense, abstract proverb into an engaging and morally faithful narrative requires deep cultural understanding and robust semantic grounding. We frame this problem as a \emph{constrained semantic decompression} task and study proverb-conditioned story generation as a testbed for abstraction-to-realization in large language models (LLMs). Focusing on Persian, we introduce the Proverb Aligned Narrative Dataset (PAND), pairing proverbs with human-written stories and explicit meanings. By a hybrid evaluation framework that combines human-calibrated LLM-as-a-Judge with structural metrics, we analyze model behavior across multiple prompting regimes. Our findings reveal a persistent \emph{decompression gap}: current LLMs often achieve strong surface-level fluency while failing to faithfully instantiate the underlying moral and causal structure encoded in proverbs. We further show that explicit reasoning and iterative refinement can partially mitigate these failures, suggesting that many decompression errors arise from difficulties in translating abstract meaning into narrative form rather than a complete lack of relevant knowledge. Our proposed task naturally extends to other forms of compressed cultural knowledge.
Chinese Translation
将一个密集、抽象的谚语转化为引人入胜且道德上忠实的叙事需要深厚的文化理解和稳健的语义基础。我们将这个问题框定为一个 extit{受限语义解压}任务，并将谚语条件的故事生成作为大型语言模型（LLMs）中抽象到实现的试验平台。聚焦于波斯语，我们引入了谚语对齐叙事数据集（Proverb Aligned Narrative Dataset，PAND），将谚语与人类创作的故事和明确的含义配对。通过结合人类校准的LLM作为评判者与结构性指标的混合评估框架，我们分析了模型在多种提示机制下的表现。我们的研究发现存在一个持续的 extit{解压差距}：当前的LLMs往往在表面流畅性上表现出色，但未能忠实地体现谚语中编码的基本道德和因果结构。我们进一步表明，明确的推理和迭代的细化可以部分缓解这些失败，暗示许多解压错误源于将抽象意义转化为叙事形式的困难，而非完全缺乏相关知识。我们提出的任务自然可以扩展到其他形式的压缩文化知识。
cs.CL / 5 / 2606.12608
Shopping Reasoning Bench: An Expert-Authored Benchmark for Multi-Turn Conversational Shopping Assistants
购物推理基准：专家撰写的多轮对话购物助手基准测试
Abstract
Conversational shopping assistants now serve hundreds of millions of customers, yet no existing benchmark jointly evaluates the open-ended multi-turn reasoning, domain expertise, and criterion-level quality that real shopping conversations demand. Shopping reasoning is unique among language model applications. Unlike factual question answering or verifiable code generation, it requires balancing subjective preferences, budget constraints, and cross-product trade-offs across multi-turn dialogue, capabilities absent from previous e-commerce and general-purpose benchmarks. We introduce the Shopping Reasoning Bench, an expert-authored benchmark of 525 missions (232 single-turn, 293 multi-turn) with 10863 importance-weighted binary rubrics authored by retail domain experts. These criteria are organized under a taxonomy of five reasoning categories and fifteen subcategories covering diverse demands such as preference refinement, trade-off analysis, and compatibility assessment. An evaluation of nine models across three families (GPT, Claude, Gemini) shows that pass rates reach only 57--77% overall. On multi-turn missions, all models score 13--29 points lower on optional above-and-beyond criteria than on required ones, and performance degrades 4--18 points as conversations progress. These gaps show that current models handle basic shopping assistance but fall short of expert-level advice, making Shopping Reasoning Bench a challenging testbed for future shopping assistant development.
Chinese Translation
对话购物助手如今为数亿客户提供服务，但现有的基准测试并未共同评估真实购物对话所需的开放式多轮推理、领域专业知识和标准级质量。购物推理在语言模型应用中独树一帜。与事实问答或可验证代码生成不同，它需要在多轮对话中平衡主观偏好、预算限制和跨产品权衡，这些能力在以往的电子商务和通用基准中并不存在。我们引入了购物推理基准，这是一个由零售领域专家撰写的基准，包含525个任务（232个单轮任务，293个多轮任务），并配有10863个重要性加权的二元评分标准。这些标准根据五个推理类别和十五个子类别进行组织，涵盖了偏好细化、权衡分析和兼容性评估等多样化需求。对三类模型（GPT、Claude、Gemini）的九个模型进行评估显示，整体通过率仅为57%至77%。在多轮任务中，所有模型在可选的额外标准上的得分比必需标准低13至29分，随着对话的进行，性能下降4至18分。这些差距表明，当前模型能够处理基本的购物辅助，但未能达到专家级建议的水平，使得购物推理基准成为未来购物助手开发的一个具有挑战性的测试平台。
cs.CL / 6 / 2606.12649
MentalMARBERT: Domain-Adaptive Pre-training and Two-Stage Fine-Tuning for Arabic Mental Health Disorders Detection
MentalMARBERT：阿拉伯心理健康障碍检测的领域适应预训练与两阶段微调
Abstract
Detecting mental health disorders from Arabic social media text remains challenging due to dialectal variation, informal language, limited high-quality annotated resources, and severe class imbalance. While English mental health natural language processing (NLP) has progressed substantially, Arabic multi-class disorder classification remains insufficiently studied. This study proposes a two-phase framework for Arabic mental health text classification. In phase 1, three Arabic pre-trained language models, AraBERT, CAMeLBERT, and MARBERT, undergo Domain-Adaptive and Task-Adaptive Pretraining (DAPT and TAPT) using a large-scale corpus of unlabeled Arabic mental health tweets. The adapted models are evaluated under a unified protocol to identify the most effective backbone model. In phase 2, the selected model is assessed across four configurations combining single-stage and hierarchical two-stage classification architectures with full fine-tuning and Low-Rank Adaptation (LoRA). To support this study, we constructed a novel annotated Arabic mental health dataset comprising 50,670 tweets across six categories, with strong inter annotator agreement (Krippendorff's Alpha = 0.733, average pairwise agreement = 0.797). Experimental results show that the domain-adapted MARBERT (MentalMARBERT) achieves statistically significant improvements over baseline models in both accuracy and macro-F1. The hierarchical two-stage architecture combined with full fine-tuning achieves the best overall performance, reaching a macro-F1 of 0.861 and an accuracy of 0.877. These findings demonstrate the effectiveness of domain-specific adaptive pretraining and hierarchical classification for Arabic mental health disorder detection.
Chinese Translation
从阿拉伯社交媒体文本中检测心理健康障碍仍然面临挑战，原因包括方言变异、非正式语言、有限的高质量标注资源以及严重的类别不平衡。尽管英语心理健康自然语言处理（NLP）已经取得了显著进展，但阿拉伯多类别障碍分类的研究仍然不足。本研究提出了一种用于阿拉伯心理健康文本分类的两阶段框架。在第一阶段，三个阿拉伯预训练语言模型，AraBERT、CAMeLBERT和MARBERT，使用大规模未标注的阿拉伯心理健康推文语料库进行领域适应和任务适应预训练（DAPT和TAPT）。经过适应的模型在统一协议下进行评估，以识别最有效的基础模型。在第二阶段，所选模型在四种配置下进行评估，这些配置结合了单阶段和分层两阶段分类架构，并采用完全微调和低秩适应（LoRA）。为支持本研究，我们构建了一个新的标注阿拉伯心理健康数据集，包含50,670条推文，涵盖六个类别，具有较强的标注者间一致性（Krippendorff's Alpha = 0.733，平均成对一致性 = 0.797）。实验结果表明，领域适应的MARBERT（MentalMARBERT）在准确性和宏观F1值上均显著优于基线模型。结合完全微调的分层两阶段架构实现了最佳整体性能，宏观F1值达到0.861，准确率为0.877。这些发现证明了领域特定适应性预训练和分层分类在阿拉伯心理健康障碍检测中的有效性。
cs.CL / 7 / 2606.12689
Observable Patterns Are Not Explanations: A Causal-Geometric Analysis of Latent Reasoning Models
可观察模式并非解释：潜在推理模型的因果几何分析
Abstract
Latent reasoning models (LRMs) replace explicit chain-of-thought with continuous thoughts. Recent work treats observable latent-state patterns, such as BFS-like frontiers and decodable arithmetic computation, as evidence for internal reasoning mechanisms. Evaluating two LRMs (Coconut and CODI) against controls lacking the proposed recurrence or curriculum, we find these patterns also appear in the controls and do not always causally affect behavior. Causal interventions reveal that latent-thought utilization is not binary but graded, scaling with a thought's causal effect on model behavior. Geometric analyses reveal this effect concentrates in low-rank directions whose step-to-step geometry grows more structured as their behavioral influence increases. Latent thoughts should therefore be treated as hidden computation, not hidden explanation: decodability, attention, or static structure alone cannot establish mechanism. LRM interpretability thus requires matched controls and causal tests.
Chinese Translation
潜在推理模型（LRMs）用连续思维替代了明确的思维链。近期的研究将可观察的潜在状态模式，如类似广度优先搜索（BFS）的边界和可解码的算术计算，视为内部推理机制的证据。通过评估两种LRMs（Coconut和CODI）与缺乏所提议的递归或课程的对照组，我们发现这些模式同样出现在对照组中，并不总是对行为产生因果影响。因果干预表明，潜在思维的利用并非二元的，而是分级的，随着思维对模型行为的因果影响而变化。几何分析表明，这种影响集中在低秩方向上，其逐步几何结构随着行为影响的增加而变得更加有序。因此，潜在思维应被视为隐含计算，而非隐含解释：仅靠可解码性、注意力或静态结构无法确立机制。因此，LRM的可解释性需要匹配的对照组和因果测试。
cs.CL / 8 / 2606.12708
AfriSUD: A Dependency Treebank Collection for Evaluating Models on African Languages
AfriSUD：用于评估非洲语言模型的依赖树库集合
Abstract
Despite their linguistic diversity and global significance, African languages remain underrepresented in research and resources to support NLP. We aim to bridge this gap by introducing AfriSUD, the first large-scale collection of syntactically annotated treebanks for nine diverse African languages spanning major language families and regions across Sub-Saharan Africa. Using the Surface-Syntactic Universal Dependencies (SUD) framework, our community-led effort provides high-quality, native-speaker verified data that capture typological key features such as agglutination and tone. We evaluate a range of models on AfriSUD for part-of-speech tagging and dependency parsing including non-transformer baselines, multilingual pretrained encoders, and LLMs. Our results reveal a significant syntax gap, where models still show clear limitations across the nine languages, suggesting that existing architectures may not fully capture the structural diversity of African-language syntax.
Chinese Translation
尽管非洲语言在语言多样性和全球重要性方面具有显著地位，但在自然语言处理（NLP）研究和资源支持方面仍然处于代表性不足的状态。我们旨在通过引入AfriSUD来填补这一空白，这是首个大规模的、针对九种不同非洲语言的句法注释树库集合，这些语言涵盖了撒哈拉以南非洲的主要语言家族和地区。采用表面句法通用依赖（Surface-Syntactic Universal Dependencies, SUD）框架，我们的社区主导努力提供了高质量、经过母语者验证的数据，捕捉了如粘附性和声调等类型学关键特征。我们在AfriSUD上评估了一系列模型的词性标注和依赖解析，包括非变换器基线、多语言预训练编码器和大型语言模型（LLMs）。我们的结果揭示了显著的句法差距，模型在这九种语言上仍然显示出明显的局限性，这表明现有架构可能无法充分捕捉非洲语言句法的结构多样性。
cs.CL / 9 / 2606.12716
Does AI Reviewer See the Full Picture? Attacking and Defending Multimodal Peer Review
人工智能评审者能否看清全貌？攻击与防御多模态同行评审
Abstract
The integration of Large Language Models (LLMs) and Multimodal LLMs (MLLMs) into scientific peer-review workflows introduces novel and significant risks for adversarial manipulation, especially given the multimodal nature of scientific papers where figures, not just text, convey core evidence. This creates a significant gap: current robustness studies on AI peer-review are overwhelmingly text-only. Moreover, the problem is distinct from standard jailbreaking, as a peer-review attack seeks to induce a domain-specific, targeted failure (e.g., "inflate this score") rather than a general safety policy violation, for which no practical defenses exist. To address this, we introduce PaperGuard, the first comprehensive benchmark designed to systematically evaluate and defend AI-generated peer-review against these domain-specific, cross-modal attacks. Our framework is built on three pillars: (1) a new multimodal peer-review dataset spanning multiple scientific domains; (2) a unified suite of attacks, including black-box prompt injections and white-box perturbations, specifically designed to target both text (GCG) and figures (PGD); and (3) a practical defense, motivated by the long-context challenge of academic papers, that uses chunk-based embedding search to efficiently localize and mitigate harmful instructions. Our extensive experiments, conducted across state-of-the-art models, confirm that AI reviewers are pervasively vulnerable. PaperGuard establishes the foundational benchmark, protocols, and actionable defense necessary to pioneer trustworthy, attack-resilient AI-assisted scholarly reviewing.
Chinese Translation
将大型语言模型（LLMs）和多模态大型语言模型（MLLMs）整合到科学同行评审工作流程中，引入了新颖且显著的对抗性操控风险，尤其考虑到科学论文的多模态特性，其中图形不仅仅是文本，且传达核心证据。这造成了一个显著的差距：当前关于人工智能同行评审的稳健性研究几乎完全集中于文本。此外，这一问题与标准的越狱攻击不同，因为同行评审攻击旨在引发特定领域的目标性失败（例如，“提高这个分数”），而不是一般的安全政策违规，后者没有实际的防御措施。为了解决这一问题，我们引入了PaperGuard，这是第一个全面的基准，旨在系统性地评估和防御针对这些领域特定的跨模态攻击的人工智能生成的同行评审。我们的框架建立在三个支柱上：（1）一个新的多模态同行评审数据集，涵盖多个科学领域；（2）一个统一的攻击套件，包括黑箱提示注入和白箱扰动，专门设计用于针对文本（GCG）和图形（PGD）；以及（3）一个实用的防御机制，受到学术论文长上下文挑战的启发，利用基于块的嵌入搜索来有效定位和缓解有害指令。我们在最先进的模型上进行的广泛实验确认了人工智能评审者普遍存在脆弱性。PaperGuard建立了开创可信赖、抗攻击的人工智能辅助学术评审所需的基础基准、协议和可操作的防御措施。
cs.CL / 10 / 2606.12748
Agent-based models for the evolution of morphological alternation patterns
基于代理的形态交替模式演化模型
Abstract
Why is the past of English "go" the apparently unrelated "went"? Such alternations are frequent in languages. They neither aid communication nor learnability, yet they can be persistent, surviving over centuries or millennia. We present a multi-agent simulation of the emergence of morphological stem and inflection alternations. Alternate forms arise by phonological changes or, as with "go/went", from lexical alternatives associated with a subset of the population. When an agent 'hears' another agent use a novel form for a slot in the paradigm of a word (say, the past tense of go), they will with some probability adopt that form, possibly spreading its use to other slots in the paradigm that shared the same original form. Thus alternative forms can spread through the population and become entrenched as stem or inflectional marker alternants. Unlike many previous computational studies, our system allows for naturalistic lexical forms, realistic phonological rules, lexicons with hundreds or thousands of entries, and agent populations in the tens or hundreds. It supports several network topologies, diffusion patterns and agent adoption policies. One issue with such simulations is evaluation: how realistic is the resulting morphology compared to those of real languages? We introduce the AI Historical Linguist, a novel Large Language Model-driven system that models a debate between two historical linguists. We use this to compare a set of real language morphologies, disguised morphologies, and experimentally evolved morphologies. The results suggest that among the factors that favor more plausible morphologies are scale-free social networks and random Bernoulli adoption of forms. We also present three case studies modeling attested historical changes, allowing us to test what might have happened if history had been different. All code and data are released.
Chinese Translation
为什么英语中“go”的过去式“went”似乎没有关联？这种交替在语言中很常见。它们既不促进交流，也不利于学习，但却可以持久存在，跨越数百年甚至数千年。我们展示了一种多代理模拟，探讨形态词干和屈折交替的出现。交替形式通过音位变化产生，或者像“go/went”那样，源于与部分人群相关的词汇替代。当一个代理“听到”另一个代理在某个词的范畴中使用新形式（比如，go的过去式），他们会以一定概率采纳该形式，并可能将其使用扩展到范畴中共享相同原始形式的其他位置。因此，替代形式可以在群体中传播，并作为词干或屈折标记的交替形式扎根。与许多之前的计算研究不同，我们的系统允许自然的词汇形式、现实的音位规则、包含数百或数千个条目的词典，以及数十或数百的代理群体。它支持多种网络拓扑、扩散模式和代理采纳策略。这类模拟的一个问题是评估：所产生的形态与真实语言的形态相比有多现实？我们引入了AI历史语言学家，这是一个新颖的基于大型语言模型的系统，模拟两位历史语言学家之间的辩论。我们利用这一系统比较一组真实语言的形态、伪装的形态和实验演化的形态。结果表明，促进更可信形态的因素包括无标度社交网络和随机的伯努利形式采纳。我们还展示了三个案例研究，模拟已证实的历史变化，使我们能够测试如果历史有所不同可能会发生什么。所有代码和数据均已发布。
cs.CL / 11 / 2606.12754
LLMs Can Better Capture Human Judgments--With the Right Prompts
大型语言模型在正确提示下能够更好地捕捉人类判断
Abstract
Are large language models (LLMs) bad at capturing human judgment? Two commonly stated limitations are that LLMs fail to capture full distributions of responses, and that their judgments are unstable across wording variations. We demonstrate simple prompting strategies that mitigate these limitations. Across two datasets--a U.S.-representative set of 144 moral scenarios and 38 moral beliefs from the International Social Survey Programme's Family and Changing Gender Roles module covering 32 countries--we show how simple elicitation techniques help improve AI-human alignment. First, prompting models to report standard deviations and response proportions recovers the full range of human responses better than common strategies. Second, ensuring scenarios are clear to human participants--as reflected in human confusion ratings--boosts model alignment, and LLMs can track human confusion ratings. At the same time, we find that LLMs' estimates of their own error are poorly calibrated, though they can predict human variability relatively well. These results suggest that asking better questions to LLMs can yield better answers.
Chinese Translation
大型语言模型（LLMs）在捕捉人类判断方面表现不佳吗？两种常见的局限性是，LLMs未能捕捉到完整的响应分布，以及它们的判断在措辞变化中不稳定。我们展示了简单的提示策略来缓解这些局限性。在两个数据集上——一个代表美国的144个道德情境和来自国际社会调查项目（International Social Survey Programme）家庭与性别角色变化模块的38个道德信念，涵盖32个国家——我们展示了简单的引导技术如何帮助改善人工智能与人类的对齐。首先，提示模型报告标准差和响应比例能够比常见策略更好地恢复人类响应的全范围。其次，确保情境对人类参与者清晰（如人类困惑评分所反映的）能够提升模型对齐度，而LLMs能够追踪人类的困惑评分。同时，我们发现LLMs对自身错误的估计校准不佳，尽管它们能够相对较好地预测人类的变异性。这些结果表明，向LLMs提出更好的问题可以获得更好的答案。
cs.CL / 12 / 2606.12765
Rigel: Reverse-Engineering the Metal 4.1 Tensor Compute Path on the Apple M4 Max GPU
Rigel：逆向工程苹果 M4 Max GPU 上的 Metal 4.1 张量计算路径
Abstract
Apple's Metal 4.1 exposes a tensor compute path: the Metal Performance Primitives (MPP) matmul2d operation over cooperative_tensor fragments, whose interface is documented but whose hardware behavior is deliberately hidden. The specification states which data-type rows are supported, never whether they are hardware-accelerated, where the operation physically executes, what its accumulator width is, or how it partitions matrix fragments across threads. We present Rigel, an empirical characterization of this path on a single Apple M4 Max (a pre-neural-accelerator generation). Using a checksum-gated, provenance-tracked microbenchmark harness, Rigel recovers eleven facts the v4.1 specification hides or contradicts. The headline finding: the Metal 4.1 fp8 (E4M3) matmul2d is emulated, not accelerated: it sustains 0.94x the throughput of fp16 despite reading half the operand bytes, so on M4 it is a memory-footprint feature, not a performance feature. We further show, via a three-signal triangulation (throughput ceiling, comparison against simdgroup_matrix, and per-rail power attribution), that matmul2d executes entirely on the GPU shader cores with no dedicated matrix datapath and no evidence of Apple Neural Engine routing; that it accumulates in >=fp32; and we reconstruct the opaque 8x8 cooperative_tensor fragment layout Apple documents nowhere. Acting on the characterization, a hand-fused GEMM + bias + GELU kernel beats the decomposed path by +6.5-12.9% in the cache-resident regime. All findings are reproducible from committed MIT-licensed code and per-cell CSVs.
Chinese Translation
苹果的 Metal 4.1 暴露了一个张量计算路径：Metal 性能原语（MPP）中的 matmul2d 操作，针对 cooperative_tensor 片段，其接口有文档说明，但其硬件行为被故意隐藏。规范中说明了支持哪些数据类型的行，但从未提及它们是否经过硬件加速、操作的物理执行位置、累加器宽度或如何在线程间划分矩阵片段。我们提出了 Rigel，这是对单个苹果 M4 Max（一个前神经加速器世代）上该路径的实证特征描述。通过使用校验和门控、来源追踪的微基准测试工具，Rigel 恢复了 v4.1 规范隐藏或矛盾的十一条事实。主要发现是：Metal 4.1 的 fp8（E4M3）matmul2d 是模拟的，而非加速的：尽管读取了一半的操作数字节，但其吞吐量仅为 fp16 的 0.94 倍，因此在 M4 上，它是一个内存占用特性，而非性能特性。我们进一步通过三信号三角测量（吞吐量上限、与 simdgroup_matrix 的比较以及每条轨道的功耗归属）展示，matmul2d 完全在 GPU 着色器核心上执行，没有专用的矩阵数据路径，也没有苹果神经引擎路由的证据；它的累加在 >=fp32 中进行；并且我们重建了苹果未在任何地方记录的模糊 8x8 cooperative_tensor 片段布局。根据这一特征描述，一个手动融合的 GEMM + 偏置 + GELU 内核在缓存驻留状态下比分解路径提高了 6.5-12.9%。所有发现均可从提交的 MIT 许可代码和每个单元的 CSV 中复现。
cs.CL / 13 / 2606.12789
How Fine-Grained Should a RAG Benchmark Be? A Hierarchical Framework for Synthetic Question Generation
RAG基准应具备多细粒度？合成问题生成的分层框架
Abstract
Evaluating retrieval-augmented generation (RAG) systems requires benchmarks that capture diverse question characteristics, yet practitioners lack empirical guidance on which dimensions to vary and at what granularity. We present HieraRAG, a hierarchical framework for studying granularity in RAG benchmark construction, defining optimal granularity as the level that maximizes discriminative power (the standard deviation of generation quality across categories) within a given RAG configuration. As a case study, we generate 5,872 synthetic question-answer (QA) pairs from FineWeb-10BT across 3 dimensions (Question Complexity, Answer Type, Linguistic Variation) at 3 granularity levels (2, 4, and 8 categories). With a BM25+Falcon-3-10B pipeline, optimal granularity varies by dimension: complexity benefits from fine-grained distinctions (discriminative power: 0.053) while answer type and linguistic variation peak at medium granularity. We introduce a Coherence Ratio metric to quantify whether fine-grained splits cleanly subdivide parent categories, revealing structural differences across dimensions (Question Complexity: 0.40 vs. Answer Type: 1.44). Human evaluation of 110 stratified QA pairs confirms synthetic quality. While these specific findings reflect a single configuration, HieraRAG provides a portable procedure and validation metric for practitioners to determine evaluation granularity within their own RAG settings.
Chinese Translation
评估检索增强生成（RAG）系统需要能够捕捉多样化问题特征的基准，但从业者缺乏关于应变更哪些维度以及以何种粒度进行变化的实证指导。我们提出了HieraRAG，一个用于研究RAG基准构建中粒度的分层框架，将最佳粒度定义为在给定RAG配置下最大化区分能力（各类别生成质量的标准差）的水平。作为案例研究，我们从FineWeb-10BT生成了5,872个合成问答（QA）对，涵盖3个维度（问题复杂性、答案类型、语言变异），并在3个粒度水平（2、4和8个类别）下进行。通过BM25+Falcon-3-10B管道，最佳粒度因维度而异：复杂性受益于细粒度区分（区分能力：0.053），而答案类型和语言变异在中等粒度时达到峰值。我们引入了一种一致性比率指标，以量化细粒度划分是否能清晰地细分父类别，揭示了各维度之间的结构差异（问题复杂性：0.40 vs. 答案类型：1.44）。对110个分层QA对的人类评估确认了合成质量。虽然这些具体发现反映了单一配置，HieraRAG为从业者提供了一种可移植的程序和验证指标，以确定其自身RAG设置中的评估粒度。
cs.CL / 14 / 2606.12790
GENIE: A Fine-Grained Measure for Novelty
GENIE：一种细粒度的新颖性度量
Abstract
Large Language Models have consistently demonstrated a lack of creativity and diversity across tasks. Prior work has focused on addressing whether models are capable of generating creative outputs. Here, we aim to consider novelty and investigate what makes model-generated content novel or not novel in a task-specific manner. We propose a fine-grained evaluation metric GENIE to measure the novelty of responses along task-specific features with respect to a population of responses. We show that unlike GENIE, holistic metrics struggle to capture the high-dimensionality of novelty and do not provide insight on which properties they target. Finally, we use GENIE to measure the effectiveness of mitigation methods that address creativity to better understand where these methods can improve novelty.
Chinese Translation
大型语言模型在各项任务中始终表现出缺乏创造力和多样性。之前的研究主要集中在模型是否能够生成创造性输出上。在此，我们旨在考虑新颖性，并以任务特定的方式探讨模型生成内容的新颖性与否。我们提出了一种细粒度的评估指标GENIE，以根据响应的任务特征相对于响应群体来衡量响应的新颖性。我们展示了与GENIE不同，整体性指标难以捕捉新颖性的高维特性，并且无法提供针对哪些属性的洞察。最后，我们使用GENIE来衡量解决创造力问题的缓解方法的有效性，以更好地理解这些方法在何处能够改善新颖性。
cs.CL / 15 / 2606.12807
Detect, Remask, Repair: Diffusion Editing for Faithful Summarization of Evolving Contexts
检测、重标记、修复：用于忠实总结演变上下文的扩散编辑
Abstract
Summaries of real-world events can become outdated as contexts evolve and new information arrives. A common response is to generate a new summary from the updated context, but full regeneration discards the previous draft, can obscure what changed, and may be unnecessary when only a few claims are unsupported. We study localized faithfulness repair: updating outdated spans in an existing summary while preserving supported content. We propose DETECT-REMASK-REPAIR, a diffusion-based framework that identifies, remasks, and repairs outdated regions with masked diffusion language models. To evaluate evolving-context summarization, we introduce StreamSum, a benchmark of synthetic event timelines. Experiments on DialogSum and StreamSum show that localized diffusion repair provides a controllable alternative to full rewriting: faithfulness-steered repair improves early drafts, one-step repair reduces repair cost to under half a second, with the framework enabling faithfulness-speed-preservation tradeoffs across datasets. We also find that the framework can provide a post-hoc correction step that improves faithfulness for autoregressive systems.
Chinese Translation
随着上下文的演变和新信息的到来，现实事件的摘要可能会变得过时。常见的应对方式是从更新的上下文生成新的摘要，但完全重生成会丢弃之前的草稿，可能会模糊变化的内容，并且在仅有少数主张不再支持的情况下，这种做法可能是不必要的。我们研究了局部忠实性修复：在保留支持内容的同时更新现有摘要中的过时部分。我们提出了DETECT-REMASK-REPAIR，这是一个基于扩散的框架，能够识别、重标记并修复过时区域，使用掩码扩散语言模型。为了评估演变上下文的摘要，我们引入了StreamSum，一个合成事件时间线的基准。对DialogSum和StreamSum的实验表明，局部扩散修复提供了一种可控的替代方案，而不是完全重写：忠实性引导的修复改善了早期草稿，一步修复将修复成本降低到半秒以下，该框架在不同数据集之间实现了忠实性、速度和保留之间的权衡。我们还发现，该框架可以提供一个事后修正步骤，改善自回归系统的忠实性。
cs.CL / 16 / 2606.12818
Localizing Anchoring Pathways in Language Models
语言模型中的锚定路径本地化
Abstract
Irrelevant numbers in a prompt can shift language model judgments, producing anchoring effects in numerical reasoning. We study where this anchor-sensitive signal is carried inside language models using a controlled multiple-choice setup with shared answer options. We define a logit-difference metric comparing the correct answer option with the answer option corresponding to the anchor, and validate that it tracks behavioral anchoring. Using attribution-based circuit localization on 7B--8B Qwen and Llama base and instruction-tuned models, we find that edge-level methods recover this signal more faithfully than node-level methods. Low- and high-anchor circuits transfer strongly within a model, suggesting shared pathway structure across anchor direction. However, sparse transfer across base and instruction-tuned variants is less reliable, indicating that post-training changes which pathways matter most. Overall, our results provide a mechanistic account of how anchoring-related decision signals are carried inside language models.
Chinese Translation
提示中的无关数字可以影响语言模型的判断，产生数值推理中的锚定效应。我们使用共享答案选项的受控多项选择设置研究这种对锚定敏感的信号在语言模型内部的传递位置。我们定义了一种对数差度量，比较正确答案选项与对应于锚定的答案选项，并验证它能够追踪行为锚定。通过对7B--8B Qwen和Llama基础模型及指令调优模型进行基于归因的电路本地化，我们发现边缘级方法比节点级方法更忠实地恢复了该信号。低锚定和高锚定电路在模型内部强烈转移，表明在锚定方向上存在共享的路径结构。然而，在基础和指令调优变体之间的稀疏转移则不太可靠，表明后训练阶段哪些路径最为重要发生了变化。总体而言，我们的结果提供了一个机制性解释，说明锚定相关决策信号是如何在语言模型内部传递的。
cs.CL / 17 / 2606.12837
LoHoSearch: Benchmarking Long-Horizon Search Agents Beyond the Human Difficulty Ceiling
LoHoSearch：超越人类难度上限的长视野搜索代理基准测试
Abstract
Search agent benchmarks exemplified by BrowseComp have rapidly saturated over the past year, with the strongest models surpassing 90% accuracy. Since these benchmarks are predominantly human-authored, annotators lack a global perspective on entity statistics and cannot systematically maximize search space size and structural complexity. This creates a difficulty ceiling that is hard to break. To address this, we introduce LoHoSearch (Long-Horizon Search Agents), a challenging benchmark comprising 544 human-verified questions across 11 domains. LoHoSearch is constructed via an automated pipeline built upon a knowledge graph covering over 7 million Wikipedia entities, which selects relations with large search spaces and assembles them into structurally complex questions with KG-verified unique answers. Our evaluation demonstrates that even the strongest model achieves only 34.74% accuracy, and existing context management strategies (best +6.8%) yield far smaller gains than on prior benchmarks. LoHoSearch provides a more demanding standard for evaluating long-horizon reasoning and context management in search agents.
Chinese Translation
以 BrowseComp 为例的搜索代理基准在过去一年中迅速饱和，最强模型的准确率超过 90%。由于这些基准主要由人类编写，注释者缺乏对实体统计的全局视角，无法系统地最大化搜索空间的大小和结构复杂性。这造成了一个难以突破的难度上限。为了解决这个问题，我们引入了 LoHoSearch（长视野搜索代理），这是一个包含 544 个经过人类验证的问题的挑战性基准，涵盖 11 个领域。LoHoSearch 通过一个基于知识图谱的自动化管道构建，该知识图谱覆盖超过 700 万个维基百科实体，选择具有大搜索空间的关系，并将其组装成具有 KG 验证的唯一答案的结构复杂问题。我们的评估表明，即使是最强模型的准确率也仅为 34.74%，而现有的上下文管理策略（最佳 +6.8%）所带来的提升远小于之前的基准。LoHoSearch 为评估搜索代理中的长视野推理和上下文管理提供了更高的标准。
cs.CL / 18 / 2606.12854
Small LLMs for Biomedical Claim Verification: Cost-Effective Fine-Tuning, Structural Dataset Shortcuts, and Cross-Domain Generalization
小型语言模型用于生物医学声明验证：成本效益的微调、结构化数据集捷径与跨领域泛化
Abstract
Large Language Models such as GPT-4o and GPT-5 achieve strong zero-shot performance on biomedical claim verification, but cost and opacity limit scalable use. We fine-tune three small LLMs: Phi-3-mini (3.8B), Qwen2.5-3B, and Mistral-7B, via QLoRA on SciFact and HealthVer, providing the first study of QLoRA models against GPT-4o and fine-tuned BioLinkBERT encoders. Mistral-7B QLoRA surpasses both GPT-4o and GPT-5 (up to 12% F1 gain) at a fractional cost using just 1,008 training examples. We conduct extensive in-domain and cross-domain evaluation: models trained on SciFact tested on HealthVer and vice versa, at matched sizes to isolate dataset structure from data quantity. We identify a previously unreported structural artifact in SciFact that inflates in-domain scores, and show through bidirectional out-of-domain evaluation that training on structurally sound data enables robust cross-domain transfer. We plan to release all code and adapter checkpoints.
Chinese Translation
大型语言模型如 GPT-4o 和 GPT-5 在生物医学声明验证中表现出色的零-shot 性能，但成本和不透明性限制了其可扩展使用。我们通过 QLoRA 对三个小型语言模型进行微调：Phi-3-mini (3.8B)、Qwen2.5-3B 和 Mistral-7B，提供了首个针对 QLoRA 模型与 GPT-4o 和微调的 BioLinkBERT 编码器的研究。Mistral-7B QLoRA 在仅使用 1,008 个训练样本的情况下，以较低的成本超越了 GPT-4o 和 GPT-5（F1 分数提升高达 12%）。我们进行了广泛的领域内和跨领域评估：在 SciFact 上训练的模型在 HealthVer 上测试，反之亦然，匹配规模以隔离数据集结构与数据量的影响。我们识别出 SciFact 中一个先前未报告的结构性伪影，该伪影夸大了领域内的得分，并通过双向的领域外评估表明，在结构合理的数据上训练能够实现稳健的跨领域迁移。我们计划发布所有代码和适配器检查点。
cs.CL / 19 / 2606.12881
Direct Preference Optimization for Chatbot Fine-Tuning: An Empirical Study
基于直接偏好优化的聊天机器人微调：一项实证研究
Abstract
We present an approach to fine-tuning large language models using Direct Preference Optimization (DPO), a reinforcement learning technique. Our experimental results demonstrate that DPO simplifies the training pipeline, improves computational efficiency, and achieves competitive performance. The evaluation using BLEU, ROUGE, and cosine similarity metrics indicates effective learning and convergence, though further investigation is needed to address observed training instability.
Chinese Translation
我们提出了一种使用直接偏好优化（Direct Preference Optimization, DPO）这一强化学习技术对大型语言模型进行微调的方法。我们的实验结果表明，DPO简化了训练流程，提高了计算效率，并且达到了具有竞争力的性能。使用BLEU、ROUGE和余弦相似度指标的评估表明有效的学习和收敛，尽管仍需进一步研究以解决观察到的训练不稳定性问题。
cs.CL / 20 / 2606.12897
SafeLLM: Extraction as a Hallucination-Resistant Alternative to Rewriting in Safety-Critical Settings
SafeLLM：在安全关键环境中，提取作为抵抗幻觉的重写替代方案
Abstract
Large language models (LLMs) are increasingly used to access organisational documentation, including standard operating procedures (SOPs), HR policies and institutional guidelines. However, retrieval-augmented generation (RAG) systems that rely on free-form rewriting can introduce hallucinations and unstable trade-offs between completeness and conciseness, particularly in safety- and compliance-critical settings. Objectives: To evaluate extraction as a hallucination-resistant alternative to rewriting-based RAG and compare strategies that balance precision, recall and safety across document types and model scales. Methods: We compare multiple prompting strategies, including line-number-based source selection, extraction of relevant guideline sentences with explicit safety annotations, and a multi-stage pipeline that refines draft answers using supporting evidence from source guidelines. Experiments are conducted on documents of varying length and structure, including local NHS acute care and oncology guidelines and UK-wide NICE guidelines, using both frontier-scale and locally deployable models. Performance is assessed using automatic metrics and human expert evaluation of relevance and completeness. Results: Line-number selection achieves the strongest results, outperforming direct copying and safety-focused strategies across both large and small models while maintaining high term recall (up to 95%) and close alignment with source text. Safety-oriented approaches improve precision but introduce systematic omissions, while multi-stage filtering further amplifies this trade-off. Performance varies with document structure: line-based extraction excels in protocol-like content, whereas alternative strategies perform better on more verbose documents (up to 97% term recall).
Chinese Translation
大型语言模型（LLMs）越来越多地用于访问组织文档，包括标准操作程序（SOPs）、人力资源政策和机构指南。然而，依赖自由形式重写的检索增强生成（RAG）系统可能会引入幻觉，并在完整性与简洁性之间产生不稳定的权衡，特别是在安全和合规关键的环境中。目标：评估提取作为抵抗幻觉的重写基础RAG的替代方案，并比较在不同文档类型和模型规模中平衡精确度、召回率和安全性的策略。方法：我们比较了多种提示策略，包括基于行号的源选择、提取带有明确安全注释的相关指南句子，以及一个多阶段管道，通过使用源指南的支持证据来完善草拟答案。实验在不同长度和结构的文档上进行，包括地方NHS急救和肿瘤学指南以及英国范围内的NICE指南，使用前沿规模和本地可部署模型。通过自动指标和人类专家对相关性和完整性的评估来评估性能。结果：行号选择取得了最强的结果，在大型和小型模型中均优于直接复制和以安全为重点的策略，同时保持高达95%的术语召回率，并与源文本紧密对齐。以安全为导向的方法提高了精确度，但引入了系统性遗漏，而多阶段过滤进一步放大了这种权衡。性能随文档结构而变化：基于行的提取在类似协议的内容中表现出色，而替代策略在更冗长的文档中表现更好（最高可达97%的术语召回率）。
cs.CL / 21 / 2606.12902
PRISM: Prosody-Integrated Multi-Agent Reasoning Framework for Empathetic Spoken Dialogue
PRISM：情感化口语对话的韵律集成多智能体推理框架
Abstract
Empathetic spoken dialogue systems require not only semantically appropriate responses but also emotionally aligned prosodic expression. However, cascade pipelines often discard acoustic cues during speech-to-text conversion, while end-to-end speech models lack interpretable control over emotion and knowledge integration. To address these challenges, we propose PRISM, a multi-agent framework for empathetic spoken dialogue that decouples speech perception, response generation, and speech synthesis into coordinated components. PRISM introduces a prosody-to-language translation mechanism to stabilize large language model reasoning and enables on-demand invocation of external knowledge tools for empathetic dialogue generation. Experimental results demonstrate that PRISM achieves consistent improvements in empathy, prosodic appropriateness, and text response generation quality across objective and subjective metrics. Our code is available at: https://github.com/Bxzfrm/PRISM.
Chinese Translation
情感化口语对话系统不仅需要语义上合适的回应，还需要情感上协调的韵律表达。然而，级联管道在语音转文本转换过程中往往会丢弃声学线索，而端到端语音模型则缺乏对情感和知识整合的可解释控制。为了解决这些挑战，我们提出了PRISM，一个情感化口语对话的多智能体框架，它将语音感知、回应生成和语音合成解耦为协调的组件。PRISM引入了一种韵律到语言的翻译机制，以稳定大型语言模型的推理，并能够按需调用外部知识工具以生成情感化对话。实验结果表明，PRISM在客观和主观指标上均实现了情感、韵律适宜性和文本回应生成质量的一致提升。我们的代码可在以下地址获取：https://github.com/Bxzfrm/PRISM。
cs.CL / 22 / 2606.12903
X-MADAM-RAG: Diagnosing and Handling Chinese-English Evidence Conflict in Retrieval-Augmented Generation
X-MADAM-RAG：诊断和处理检索增强生成中的中英证据冲突
Abstract
Retrieval-augmented generation (RAG) systems may receive evidence that is not merely noisy but mutually contradictory. This issue becomes particularly salient in multilingual settings, where retrieved Chinese and English evidence may support incompatible answer candidates. We study this problem through X-RAMDocs-ZHEN, a controlled Chinese-English benchmark derived from RAMDocs for diagnosing evidence conflict in RAG. The benchmark contains 300 examples across six balanced conditions, including monolingual support, bilingual agreement, reversed conflict directions, and conflict with optional noise. We further examine X-MADAM-RAG, an interpretable pipeline that decomposes evidence handling into per-document candidate extraction, visible-evidence repair, deterministic candidate grouping, and conflict-aware aggregation. On the original controlled benchmark with Qwen2.5-7B-Instruct, X-MADAM-RAG achieves 0.9667 strict accuracy and 0.9767 conflict-aware success, outperforming an evidence-normalized single-call baseline. However, a zero-call rule-only extractor reaches 1.0000 on the same benchmark, revealing strong template regularity. To probe this limitation, we construct a deterministic naturalized stress test that removes explicit answer templates while preserving candidate strings. On its 100-sample subset, rule-only extraction falls to 0.0000, but X-MADAM-RAG also drops to 0.3000 strict accuracy, below both naive and evidence-normalized baselines. A privileged oracle remains perfect, indicating that document-level extraction is the main bottleneck. These findings position X-RAMDocs-ZHEN and X-MADAM-RAG as diagnostic tools for controlled evidence conflict rather than as evidence of general hallucination detection or robustness to natural retrieval.
Chinese Translation
检索增强生成（RAG）系统可能接收到的不仅仅是噪声证据，而是相互矛盾的证据。这一问题在多语言环境中尤为突出，因为检索到的中文和英文证据可能支持不兼容的答案候选。我们通过 X-RAMDocs-ZHEN 这一受控的中英文基准研究该问题，该基准源自 RAMDocs，用于诊断 RAG 中的证据冲突。该基准包含 300 个示例，涵盖六种平衡条件，包括单语支持、双语一致、冲突方向反转和带有可选噪声的冲突。我们进一步研究 X-MADAM-RAG，这是一种可解释的流程，将证据处理分解为逐文档候选提取、可见证据修复、确定性候选分组和冲突感知聚合。在使用 Qwen2.5-7B-Instruct 的原始受控基准上，X-MADAM-RAG 实现了 0.9667 的严格准确率和 0.9767 的冲突感知成功率，优于证据标准化的单次调用基线。然而，仅使用零调用规则的提取器在相同基准上达到了 1.0000，揭示了强模板规律性。为了探讨这一局限性，我们构建了一个确定性的自然化压力测试，去除了显式答案模板，同时保留了候选字符串。在其 100 个样本子集上，仅规则提取的准确率降至 0.0000，而 X-MADAM-RAG 的严格准确率也降至 0.3000，低于天真和证据标准化的基线。一个特权的神谕保持完美，表明文档级提取是主要瓶颈。这些发现将 X-RAMDocs-ZHEN 和 X-MADAM-RAG 定位为受控证据冲突的诊断工具，而不是一般幻觉检测或对自然检索的鲁棒性证据。
cs.CL / 23 / 2606.12908
SENTINEL: Failure-Driven Reinforcement Learning for Training Tool-Using Language Model Agents
SENTINEL：基于失败驱动的强化学习用于训练工具使用语言模型代理
Abstract
Language model agents are increasingly effective in solving realistic tasks through multi-turn tool use. However, training reliable tool-using agents remains challenging in practice. While reinforcement learning provides an on-policy paradigm for improving agents from their own environment interactions, its effectiveness depends heavily on the training task distribution. When tasks are fixed before training, the task distribution can become increasingly mismatched with the policy's evolving capabilities, causing many rollouts to be spent on uninformative tasks. We propose SENTINEL, a failure-driven reinforcement learning framework that turns the Solver's rollout failures into targeted training tasks. SENTINEL follows a Controller--Proposer--Solver loop: the Controller analyzes failed trajectories and summarizes recurring error patterns, the Proposer generates executable tasks that stress these weaknesses, and the Solver is trained on the targeted tasks. On Tau2-Bench Retail with Qwen3-4B-Thinking-2507, SENTINEL improves Pass\^{}1 from 66.4 to 74.9 and outperforms RL on general synthetic tasks across Pass\^{}k metrics. These results demonstrate that model failures provide an effective and scalable source of targeted training signal for improving tool-using language model agents.
Chinese Translation
语言模型代理在通过多轮工具使用解决现实任务方面变得越来越有效。然而，在实践中，训练可靠的工具使用代理仍然面临挑战。虽然强化学习提供了一种基于策略的范式，通过代理与其自身环境的交互来改进代理，但其有效性在很大程度上依赖于训练任务的分布。当任务在训练之前固定时，任务分布可能与策略的演变能力越来越不匹配，导致许多回合花费在无信息的任务上。我们提出了SENTINEL，一个基于失败驱动的强化学习框架，将求解器的回合失败转化为有针对性的训练任务。SENTINEL遵循控制器-提议者-求解器循环：控制器分析失败的轨迹并总结重复的错误模式，提议者生成可执行的任务以针对这些弱点，求解器在这些有针对性的任务上进行训练。在使用Qwen3-4B-Thinking-2507的Tau2-Bench零售任务上，SENTINEL将Pass^{}1从66.4提高到74.9，并在Pass^{}k指标上超越了强化学习在一般合成任务中的表现。这些结果表明，模型失败为改善工具使用语言模型代理提供了有效且可扩展的有针对性的训练信号来源。
cs.CL / 24 / 2606.12911
PiDA: Phonetically-Informed Data Augmentation for Robust Vietnamese Speech Translation
PiDA：用于稳健越南语语音翻译的语音信息数据增强
Abstract
Cascaded speech translation (ST) systems suffer from error propagation when Automatic Speech Recognition (ASR) outputs incorrect transcripts. We present the first systematic categorization of ASR errors for Vietnamese ST, classifying substitution errors by phonetic cause and quantifying their impact on downstream Neural Machine Translation (NMT) performance using Linear Mixed-Effects Modelling. We confirm that most ASR substitution errors arise from phonetic confusions rather than random noise, and that these phonetic errors significantly degrade ST quality. Motivated by this finding, we propose Phonetically-Informed Data Augmentation (PiDA), which generates ASR-like corruptions by substituting words with phonetically similar alternatives using phonetic word embeddings. Fine-tuning on a PiDA-augmented version of FLEURS Vietnamese-English improves translation of erroneous ASR outputs (up to +2.04 BLEU over standard fine-tuning) while also slightly improving clean-text performance.
Chinese Translation
级联语音翻译（ST）系统在自动语音识别（ASR）输出错误转录时会遭遇错误传播。我们首次系统性地对越南语ST的ASR错误进行分类，按语音原因对替代错误进行分类，并使用线性混合效应模型量化其对下游神经机器翻译（NMT）性能的影响。我们确认大多数ASR替代错误源于语音混淆而非随机噪声，并且这些语音错误显著降低了ST质量。基于这一发现，我们提出了语音信息数据增强（PiDA），该方法通过使用语音词嵌入将单词替换为语音相似的替代词，生成类似ASR的损坏数据。在PiDA增强的FLEURS越南语-英语版本上进行微调，能够改善错误ASR输出的翻译（比标准微调提高最多2.04 BLEU），同时也略微提升了干净文本的表现。
cs.CL / 25 / 2606.12922
Polar: A Benchmark for Evaluating Political Bias in LLMs
Polar：评估大型语言模型政治偏见的基准
Abstract
Political bias in large language models (LLMs) is increasingly significant, but difficult to measure reproducibly across political and linguistic contexts. We introduce Polar, a 4,026-instance multiple-choice benchmark that measures political bias through option-level likelihoods rather than prompt-based generation. Polar covers two ideological axes and eight issue categories derived from the Manifesto Project, and evaluates models in parallel across U.S. and South Korean political contexts. Across 38 LLMs, measured bias varies systematically with political context, issue category, model group, and presentation language. All models lean left-progressive on U.S. political content, but show more centered and mixed patterns on South Korean content. Translation experiments further show that presentation language alone can shift measured bias. These findings highlight the need for multilingual and cross-contextual evaluation of political bias in LLMs.
Chinese Translation
大型语言模型（LLMs）中的政治偏见日益显著，但在不同的政治和语言背景下难以进行可重复的测量。我们介绍了Polar，这是一个包含4,026个实例的多项选择基准，通过选项级别的可能性而非基于提示的生成来测量政治偏见。Polar涵盖了两个意识形态轴线和八个问题类别，这些类别源自《宣言项目》（Manifesto Project），并在美国和韩国的政治背景下对模型进行平行评估。在38个LLMs中，测量到的偏见在政治背景、问题类别、模型组和呈现语言之间系统性地变化。所有模型在美国政治内容上倾向于左翼进步，但在韩国内容上表现出更为中立和混合的模式。翻译实验进一步表明，仅呈现语言就可以改变测量到的偏见。这些发现强调了对LLMs中政治偏见进行多语言和跨背景评估的必要性。
cs.CL / 26 / 2606.12941
Multi-Turn Reasoning When Context Arrives in Pieces: Scalable Sharding and Memory-Augmented RL
当上下文分散到多个回合时的多回合推理：可扩展的分片和记忆增强的强化学习
Abstract
When a user reveals task-critical information across several conversation turns, LLM accuracy drops by up to 65% despite full context availability. We show that this Lost in Conversation degradation can be substantially mitigated by training models to maintain a compact rolling memory instead of attending to a growing history. To make such training scalable, we introduce a low-cost sharding pipeline that converts single-turn QA datasets into multi-turn fragmented-information episodes, eliminating the need for hours of manual annotation. Training only on sharded GSM8K, our memory-augmented policy significantly improves multi-turn accuracy and generalises zero-shot to harder math and out-of-domain long-context QA. Moreover, memory-trained models outperform full-history baselines even when given the full history at test time, suggesting that learning to compress induces more robust incremental reasoning than full-context exposure alone.
Chinese Translation
当用户在多个对话回合中透露任务关键的信息时，尽管上下文信息完全可用，LLM（大型语言模型）的准确率仍会下降多达65%。我们展示了这种在对话中迷失的退化现象可以通过训练模型保持紧凑的滚动记忆而得到显著缓解，而不是关注不断增长的历史。为了使这种训练具有可扩展性，我们引入了一种低成本的分片管道，将单回合问答数据集转换为多回合碎片化信息的情节，从而消除了数小时手动标注的需求。仅在分片的GSM8K上进行训练，我们的记忆增强策略显著提高了多回合准确率，并在零样本情况下对更难的数学问题和超出领域的长上下文问答进行了泛化。此外，即使在测试时提供完整历史，经过记忆训练的模型也优于全历史基线，这表明学习压缩能够引导出比单纯的全上下文暴露更强健的增量推理能力。
cs.CL / 27 / 2606.12984
SkillChain: Closing the Loop on Skill Evolution for Image-Based E-Commerce AI Assistants
SkillChain：闭合基于图像的电子商务人工智能助手技能演变的反馈循环
Abstract
Image-based AI assistants are now deployed at production scale on e-commerce platforms, where a single uploaded image can trigger fundamentally different user intents: product search, style recommendation, visual encyclopedia, or utility tool calls, each demanding its own response format, tool invocation, and domain knowledge. Without per-intent behavioral constraints, LLM-based systems conflate these heterogeneous modes and fall short of domain quality standards, while the breadth and dynamism of the intent space render manual engineering infeasible. To address this, we present SkillChain, which closes the production feedback loop on Skill evolution, automating the lifecycle of Skills through three stages: Skill Creator for bootstrapping from task specs and trajectories, Route Optimizer for routing alignment, and Body Refiner for iterative Skill Body refinement via dual-path LLM-Judge evaluation. Deployed on a production-scale e-commerce image assistant, SkillChain substantially improves aggregate response quality, with the strongest gains on structural compliance and content quality; a one-week online A/B experiment further confirms significant gains in user engagement, content consumption, and long-term retention.
Chinese Translation
基于图像的人工智能助手目前已在电子商务平台上大规模部署，其中单个上传的图像可以触发截然不同的用户意图：产品搜索、风格推荐、视觉百科全书或实用工具调用，每种意图都需要其特定的响应格式、工具调用和领域知识。在没有针对每种意图的行为约束的情况下，基于大型语言模型（LLM）的系统将这些异质模式混淆在一起，未能达到领域质量标准，而意图空间的广度和动态性使得手动工程变得不可行。为了解决这一问题，我们提出了SkillChain，它闭合了技能演变的生产反馈循环，通过三个阶段自动化技能的生命周期：技能创建者（Skill Creator）用于从任务规范和轨迹中引导，路由优化器（Route Optimizer）用于路由对齐，以及主体精炼器（Body Refiner）通过双路径LLM评估进行迭代技能主体的精炼。在生产规模的电子商务图像助手上部署的SkillChain显著提高了整体响应质量，在结构合规性和内容质量方面获得了最强的提升；为期一周的在线A/B实验进一步确认了用户参与度、内容消费和长期留存的显著提升。
cs.CL / 28 / 2606.13044
No Hidden Prompts Needed! You Can Game AI Peer Review with Presentation-Only Revisions
无需隐藏提示！您可以通过仅修改展示内容来操控AI同行评审
Abstract
As AI-generated reviews move from experimental tools into peer-review infrastructure, most robustness concerns have focused on explicit attacks such as hidden instructions and prompt injection. We study a harder and more policy-relevant failure mode: no hidden text, no prompt injection, and no changes to methods, experiments, figures, equations, proofs, or numerical results. The attacker modifies only presentation-level content, such as the abstract, contribution framing, related work, discussion, and narrative structure. We introduce adversarial repackaging: a closed-loop attack that uses AI-reviewer feedback to search for presentation-level revisions while keeping the scientific evidence fixed. Across three mainstream AI reviewers, adversarial repackaging achieves a 75.1% attack success rate and a mean score gain of +1.21/10. The effect is not explained by ordinary prose polishing. We also reveal that strategies that change how the reviewer interprets the paper, such as related-work repositioning and analytical discussion expansion, substantially outperform surface edits such as local polishing, table formatting, and algorithm boxes. Our analysis reveals two deeper structural failure modes. First, AI reviewers are easier to impress than to convince: highlighting strengths reliably increases perceived merit, while attempts to dissolve weaknesses frequently backfire. Second, AI reviewers can confuse the appearance of addressing a limitation with actually resolving it, allowing unchanged evidence to be reinterpreted as stronger scientific contribution. These results show that the deployment risk is not only malicious hidden instructions, but the emergence of paper presentation itself as an optimization surface. We release a contamination-free rolling benchmark and attack framework for testing whether AI reviewers remain anchored to scientific content under presentation-only edits.
Chinese Translation
随着AI生成的评审从实验工具转向同行评审基础设施，大多数关于鲁棒性的担忧集中在显式攻击上，如隐藏指令和提示注入。我们研究了一种更复杂且与政策相关的失败模式：没有隐藏文本、没有提示注入，也没有对方法、实验、图表、方程、证明或数值结果的更改。攻击者仅修改展示层面的内容，例如摘要、贡献框架、相关工作、讨论和叙事结构。我们提出了对抗性重新包装：一种闭环攻击，利用AI评审者的反馈来搜索展示层面的修订，同时保持科学证据不变。在三种主流AI评审者中，对抗性重新包装实现了75.1%的攻击成功率和平均得分提升+1.21/10。该效果并不能通过普通的文笔润色来解释。我们还揭示了改变评审者如何解读论文的策略，如相关工作的重新定位和分析讨论的扩展，显著优于表面编辑，如局部润色、表格格式化和算法框。我们的分析揭示了两种更深层次的结构性失败模式。首先，AI评审者更容易被打动而非说服：突出优点可靠地提高了感知价值，而试图消除缺点的努力常常适得其反。其次，AI评审者可能会将解决局限性的表象与实际解决问题混淆，从而使未更改的证据被重新解读为更强的科学贡献。这些结果表明，部署风险不仅在于恶意的隐藏指令，还在于论文展示本身作为优化表面的出现。我们发布了一个无污染的滚动基准和攻击框架，用于测试AI评审者在仅进行展示编辑时是否仍然与科学内容保持一致。
cs.CL / 29 / 2606.13082
sebis at CRF Filling 2026: A Two-Stage Local LLM Pipeline for Medical CRF Filling
sebis在CRF填充2026：用于医疗CRF填充的双阶段本地LLM管道
Abstract
The extraction of structured clinical information from unstructured EHR notes is a persistent bottleneck in healthcare informatics. While large language models (LLMs) offer high performance, their deployment in clinical settings is hindered by privacy risks, inference costs, and the tendency to hallucinate beyond textual evidence. We address these challenges for the CL4Health 2026 Case Report Form (CRF) filling task by proposing a fully local, domain-adapted pipeline using the MedGemma-27B model. Our two-stage architecture, which separates binary presence classification from value extraction, enforces strict adherence to textual evidence and ensures deterministic outputs for negated, uncertain, or unknown states. By leveraging item-specific, few-shot in-context learning without external API calls or fine-tuning, our approach achieves a macro-F1 score of 0.55 on the official English test track. This result secures second place among all locally-hosted, open-source submissions. Our work demonstrates that privacy-preserving, on-premise LLM pipelines can achieve near-competitive performance with proprietary frontier models, providing a practical, data-sovereign framework for clinical NLP.
Chinese Translation
从非结构化电子健康记录（EHR）笔记中提取结构化临床信息是医疗信息学中的一个持续瓶颈。尽管大型语言模型（LLMs）提供了高性能，但它们在临床环境中的部署受到隐私风险、推理成本以及超出文本证据的幻觉倾向的阻碍。我们通过提出一个完全本地的、领域适应的管道，使用MedGemma-27B模型，来解决CL4Health 2026案例报告表（CRF）填充任务中的这些挑战。我们的双阶段架构将二元存在分类与数值提取分开，严格遵循文本证据，并确保对否定、不确定或未知状态的确定性输出。通过利用特定项目的少量上下文学习，而无需外部API调用或微调，我们的方法在官方英语测试轨道上达到了0.55的宏F1分数。这一结果在所有本地托管的开源提交中获得了第二名。我们的工作表明，保护隐私的本地LLM管道可以实现与专有前沿模型相近的竞争性能，为临床自然语言处理提供了一个实用的数据主权框架。
cs.CL / 30 / 2606.13100
LEDGER: A Long-Context Benchmark of Corporate Annual Reports for Grounded Financial Retrieval and Extraction
LEDGER：用于基于实证的财务检索和提取的公司年度报告长文本基准
Abstract
Finance reporting is a natural proving ground for large language models, and the very-long-context capabilities of recent models across all sizes make rigorous evaluation in this domain an increasingly pressing need. Yet most public financial resources reduce the task to plain-text SEC 10-K filings paired with a handful of question-answer items. We release LEDGER (Long-context Evaluation of Documents for Grounded Extraction and Retrieval), a corpus of 4,999 digitized corporate annual reports - full documents with figures, tables, and narrative, not just regulatory filings. Each report is labeled with 31 consolidated financial KPIs to be extracted and linked to the market's reaction at the earnings date. From this data we derive three evaluation benchmarks spanning the difficulty spectrum: a pure page-level KPI retrieval task with TREC-style relevance judgments over 118,048 questions in natural language, a conversational "needle-in-a-haystack" single-value lookup, and a full KPI extraction task, both from long, numerically dense reports. We additionally provide human OCR-quality annotations with inter-annotator agreement and the complete extraction, validation, and scoring toolchain. We further demonstrate the dataset's research utility with a case study linking CEO-letter rhetoric to post-publication market impact.
Chinese Translation
财务报告是大型语言模型的自然试验场，而近期各种规模模型的超长上下文能力使得在这一领域进行严格评估的需求愈发迫切。然而，大多数公共财务资源仅将任务简化为与少量问答项配对的纯文本SEC 10-K文件。我们发布了LEDGER（文档的长文本评估用于基于实证的提取和检索），这是一个包含4,999份数字化公司年度报告的语料库——这些报告是完整的文档，包含图表、表格和叙述，而不仅仅是监管文件。每份报告都标注了31个合并的财务关键绩效指标（KPI），以便提取并与收益日期的市场反应关联。从这些数据中，我们推导出三个覆盖难度范围的评估基准：一个纯粹的页面级KPI检索任务，采用TREC风格的相关性判断，涵盖118,048个自然语言问题；一个对话式的“针在干草堆中”单值查找；以及一个完整的KPI提取任务，均来自长且数值密集的报告。此外，我们还提供了具有人工OCR质量的注释，包含标注者间一致性，以及完整的提取、验证和评分工具链。我们进一步通过案例研究展示了数据集的研究效用，将CEO信件中的修辞与出版后市场影响联系起来。
cs.CL / 31 / 2606.13111
M\"OVE: A Holistic LLM Benchmark for the German Public Sector
M"OVE：德国公共部门的整体大型语言模型基准
Abstract
We present M\"OVE (Modelle f\"ur die \"Offentliche Verwaltung Evaluieren), a holistic benchmark for evaluating large language models (LLMs) in the context of the German public sector. While LLMs are increasingly adopted in public administration, model selection remains largely ad hoc, and existing benchmarks offer limited guidance: they are predominantly English-centric, US-centric in content, and focus exclusively on task performance. M\"OVE addresses these gaps by evaluating 39 models across two complementary dimensions. Performance criteria cover summarization, question answering, and topic extraction. Governance criteria assess hallucination tendencies, energy consumption, provider transparency, and alignment with German constitutional values and knowledge about positions by German political parties. In total, we utilize ten German-language datasets, including gold- and silverstandard datasets that we constructed to reflect public-administration domains. We employ a multi-metric evaluation strategy combining classical NLP metrics, embedding-based methods, and LLM-as-a-judge approaches. Our results show that no single model dominates across all criteria: top performers differ between tasks, and model size alone is a poor predictor of quality. We further evaluate the benchmark itself, analyzing its statistical precision, LLM judge reliability, the impact of our private datasets on model rankings, the sensitivity of our results to prompt formulation, and the validity of our energy consumption estimates. M\"OVE is designed as a living benchmark under active development; results are publicly available at https://moeve.bundesdruckerei.de/.
Chinese Translation
我们提出了 M"OVE（Modelle für die "Offentliche Verwaltung Evaluieren），这是一个用于评估大型语言模型（LLMs）在德国公共部门背景下的整体基准。尽管 LLMs 在公共管理中越来越多地被采用，但模型选择仍然主要是临时的，现有基准提供的指导有限：它们主要以英语为中心，内容以美国为中心，并且仅专注于任务表现。M"OVE 通过在两个互补维度上评估 39 个模型来填补这些空白。性能标准涵盖摘要、问答和主题提取。治理标准评估幻觉倾向、能耗、提供者透明度以及与德国宪法价值观和德国政党的立场知识的一致性。我们总共使用了十个德语数据集，包括我们构建的金标准和银标准数据集，以反映公共管理领域。我们采用了一种多指标评估策略，结合了经典的自然语言处理（NLP）指标、基于嵌入的方法和 LLM 作为评审的方法。我们的结果表明，没有单一模型在所有标准上占据主导地位：最佳表现者在不同任务之间有所不同，模型大小本身并不是质量的良好预测指标。我们进一步评估了基准本身，分析其统计精度、LLM 评审的可靠性、我们的私有数据集对模型排名的影响、结果对提示表述的敏感性，以及我们能耗估算的有效性。M"OVE 被设计为一个在积极开发中的动态基准；结果可在 https://moeve.bundesdruckerei.de/ 上公开获取。
cs.CL / 32 / 2606.13115
G-Long: Graph-Enhanced Memory Management for Efficient Long-Term Dialogue Agents
G-Long：用于高效长期对话代理的图增强记忆管理
Abstract
While Large Language Models (LLMs) have advanced open-domain dialogue systems, maintaining long-term consistency remains a challenge due to inherent limitations in long-context reasoning and the inefficiency of processing extensive raw text. Existing approaches typically rely on either unstructured memory storage, which is prone to information loss, or computationally expensive LLMs that incur high latency. To address these limitations, we propose G-Long, a graph-enhanced framework that utilizes a fine-tuned small Language Model (sLM) for structured triplet extraction and associative retrieval, significantly reducing operational costs. Furthermore, we introduce the novel attention-aware importance scoring mechanism that leverages the intrinsic cross-attention signals of a T5 summarizer to identify salient memories. Extensive experiments across diverse benchmarks demonstrate that G-Long achieves state-of-the-art performance in both response generation and memory retrieval, yielding performance gains of up to 9.8% in response quality on MSC and 40.8% in retrieval recall on LME, while significantly minimizing computational overhead.
Chinese Translation
尽管大型语言模型（LLMs）在开放领域对话系统中取得了进展，但由于长期一致性维护的固有限制，长文本推理和处理大量原始文本的低效性仍然是一个挑战。现有方法通常依赖于非结构化的记忆存储，这容易导致信息丢失，或依赖于计算成本高昂的LLMs，造成高延迟。为了解决这些限制，我们提出了G-Long，一个图增强框架，利用经过微调的小型语言模型（sLM）进行结构化三元组提取和关联检索，显著降低了操作成本。此外，我们引入了一种新颖的注意力感知重要性评分机制，利用T5摘要模型的内在交叉注意力信号来识别显著记忆。广泛的实验在多种基准测试中表明，G-Long在响应生成和记忆检索方面均实现了最先进的性能，在MSC上响应质量提高了高达9.8%，在LME上检索召回率提高了40.8%，同时显著减少了计算开销。
cs.CL / 33 / 2606.13120
EvoBrowseComp: Benchmarking Search Agents on Evolving Knowledge
EvoBrowseComp：对进化知识的搜索代理进行基准测试
Abstract
Search Agents -- large language models augmented with search tools -- have intensified the need for future-proof evaluation benchmarks. Existing benchmarks such as BrowseComp rely on static knowledge, making them vulnerable to test-set contamination and parametric memorization. Consequently, models can achieve high scores through fact recall rather than genuine retrieval, obscuring true browsing competence via reasoning shortcuts. In this paper, we introduce EvoBrowseComp, an evolving benchmark of 400 English and 400 Chinese contamination-free complex questions synthesized via live-web traversal. To collect these questions, we design a three-agent collaborative framework: (1) a QA synthesis agent that retrieves fresh knowledge from the live web to synthesize QA pairs; (2) an information filtering agent that filters retrieved knowledge in terms of credibility and popularity to block parametric shortcuts; and (3) a high-level guidance agent that formalizes questions into reasoning graphs to reduce logical redundancy and shortcuts in synthesized QA pairs. Because the framework supports fully automated synthesis, EvoBrowseComp can be regularly updated to prevent data contamination and maintain temporal freshness. Extensive experiments confirm its great difficulty, requiring broad horizontal search. It establishes a scalable paradigm for auto-updatable, high-difficulty benchmarking that keeps pace with both evolving world knowledge and advancing agent capabilities.
Chinese Translation
搜索代理——通过搜索工具增强的大型语言模型——加剧了对未来可持续评估基准的需求。现有基准如 BrowseComp 依赖于静态知识，使其容易受到测试集污染和参数记忆的影响。因此，模型可能通过事实回忆而非真正的检索来获得高分，从而掩盖了通过推理捷径获得的真实浏览能力。在本文中，我们介绍了 EvoBrowseComp，这是一个包含 400 个英语和 400 个中文无污染复杂问题的进化基准，这些问题是通过实时网络爬取合成的。为了收集这些问题，我们设计了一个三代理协作框架：(1) 一个 QA 合成代理，从实时网络中检索新知识以合成 QA 对；(2) 一个信息过滤代理，根据可信度和受欢迎程度过滤检索到的知识，以阻止参数捷径；(3) 一个高层次指导代理，将问题形式化为推理图，以减少合成 QA 对中的逻辑冗余和捷径。由于该框架支持完全自动化的合成，EvoBrowseComp 可以定期更新，以防止数据污染并保持时间上的新鲜感。大量实验确认了其极高的难度，要求广泛的横向搜索。它建立了一个可扩展的范式，用于自动更新的高难度基准测试，以跟上不断发展的世界知识和代理能力的进步。
cs.CL / 34 / 2606.13121
NaturalFlow: Reducing Disruptive Pauses for Natural Speech Flow in Simultaneous Speech-to-Speech Translation
NaturalFlow：减少同声传译中自然语音流的干扰性停顿
Abstract
Simultaneous speech-to-speech translation aims to enable near-real-time communication by minimizing latency, offering a compelling, real-time alternative to the high latency of consecutive translation. However, the excessive pursuit of low latency often results in fragmented chunk-wise speech. Consequently, listeners are subjected to an unnatural acoustic flow punctuated by frequent pauses, which could increase their cognitive load. To bridge this gap, we introduce a fluency-aware optimization framework designed to discover the sweet spot between the low-latency benefits of simultaneous translation and the natural flow of consecutive translation. Our framework minimizes inter-chunk silences by leveraging model-internal signals, including linguistic diversity and induced temporal variability in speech durations. Experiments on short- and long-form benchmarks show that our framework produces natural speech flow while maintaining competitive latency and translation quality.
Chinese Translation
同声传译旨在通过最小化延迟实现近实时通信，为高延迟的交替翻译提供了一种引人注目的实时替代方案。然而，过度追求低延迟往往导致语音片段化。因此，听众面临着不自然的声学流，频繁的停顿增加了他们的认知负担。为了解决这一问题，我们提出了一种流畅性意识优化框架，旨在发现同声翻译的低延迟优势与交替翻译的自然流之间的最佳平衡。我们的框架通过利用模型内部信号，包括语言多样性和语音时长的诱导时间变异性，最小化片段间的沉默。在短篇和长篇基准测试中的实验表明，我们的框架在保持竞争性延迟和翻译质量的同时，能够产生自然的语音流。
cs.CL / 35 / 2606.13142
HyPE: Category-Aware Hypergraph Encoding with Persistent Edge Embeddings for Persona-Grounded Dialogue
HyPE：具有持久边嵌入的类别感知超图编码用于基于个性的对话
Abstract
Persona-grounded dialogue systems aim to produce responses consistent with a speaker's persona, yet existing methods treat personas as a flat set of sentences and fail to model the high-order relations among persona attributes-e.g., that several persona sentences share a topical category. We propose HyPE (Hypergraph Persona Encoder), a framework that (i) analyzes each persona-bearing text as a (Core, Expression, Sentiment, Category) quadruple, and (ii) organizes persona elements into a hypergraph whose hyperedges are induced by shared category labels. An HyperGCN hypergraph neural network propagates this structure into a persona summary vector and a soft-memory bank that condition the response generator. We further propose Persistent Edge Embeddings (PEE), lightweight per-category learnable priors fused into the HyperGCN message-passing step. On PersonaChat under greedy decoding, HyPE consistently outperforms sentence-level pooling baselines across GPT-2, LLaMA-3.2-3B, and Qwen2.5-3B backbones by demonstrating that structured hyperedge-level persona encoding provides a transferable advantage across model scales.
Chinese Translation
基于个性的对话系统旨在生成与说话者个性一致的响应，但现有方法将个性视为一个扁平的句子集合，未能建模个性属性之间的高阶关系，例如，多个个性句子共享一个主题类别。我们提出了HyPE（超图个性编码器），一个框架，它（i）将每个包含个性的文本分析为（核心、表达、情感、类别）四元组，并（ii）将个性元素组织成一个超图，其超边由共享的类别标签诱导。一个HyperGCN超图神经网络将这一结构传播到个性摘要向量和条件响应生成器的软记忆库中。我们进一步提出了持久边嵌入（PEE），这是一种轻量级的每类别可学习先验，融合到HyperGCN消息传递步骤中。在使用贪婪解码的PersonaChat上，HyPE在GPT-2、LLaMA-3.2-3B和Qwen2.5-3B骨干网络上始终优于句子级池化基线，证明了结构化的超边级个性编码在模型规模之间提供了可转移的优势。
cs.CL / 36 / 2606.13171
NTS-CoT: Mitigating Hallucinations in LLM-based News Timeline Summarization with Chain-of-Thought Reasoning
NTS-CoT：通过链式思维推理减轻基于大型语言模型的新闻时间线摘要中的幻觉现象
Abstract
The rapid updates of online news make tracking event developments challenging, highlighting the need for timeline summarization (TLS). Hallucinations, where LLM-generated content deviates from source news, still remain a critical issue in LLM-based TLS and are not well studied in existing works. To bridge this gap, we identify two primary types of hallucinations: unfaithful content during news summarization and information omission in date-event summarization. Then, we propose NTS-CoT, a novel framework that leverages Chain-of-Thought (CoT) reasoning to mitigate hallucinations in TLS. The framework consists of three key modules: i) Element-CoT to capture essential news elements for faithful summarization, ii) Date Selection to combine temporal saliency and event prominence for timestamp selection, and iii) Causal-CoT to infer causal relationships and reduce omissions in date-event summarization. Extensive experiments, including quantitative analysis on three TLS benchmarks and human evaluation, demonstrate that NTS-CoT outperforms state-of-the-art baselines, effectively mitigating hallucinations and improving LLM-based TLS performance. Our source code is available at https://anonymous.4open.science/r/NTS-CoT .
Chinese Translation
在线新闻的快速更新使得跟踪事件发展变得具有挑战性，这突显了时间线摘要（TLS）的必要性。幻觉现象，即大型语言模型（LLM）生成的内容偏离源新闻，仍然是基于LLM的TLS中的一个关键问题，并且在现有研究中尚未得到充分探讨。为了填补这一空白，我们识别出两种主要类型的幻觉：在新闻摘要中不忠实的内容和在日期-事件摘要中的信息遗漏。然后，我们提出了NTS-CoT，一个新颖的框架，利用链式思维（CoT）推理来减轻TLS中的幻觉现象。该框架由三个关键模块组成：i) Element-CoT，用于捕捉重要的新闻元素以实现忠实摘要；ii) 日期选择，用于结合时间显著性和事件重要性进行时间戳选择；iii) Causal-CoT，用于推断因果关系并减少日期-事件摘要中的遗漏。大量实验，包括对三个TLS基准的定量分析和人工评估，表明NTS-CoT在减轻幻觉现象和提高基于LLM的TLS性能方面优于最先进的基线。我们的源代码可在 https://anonymous.4open.science/r/NTS-CoT 获取。
cs.CL / 37 / 2606.13177
MemRefine: LLM-Guided Compression for Long-Term Agent Memory
MemRefine：基于大型语言模型的长期智能体记忆压缩
Abstract
Large language model (LLM) agents are increasingly expected to operate over long-term interactions, where information from past dialogues must be preserved and recalled to support future tasks. However, as interactions accumulate, the memory store grows without bound and fills with redundant entries that inflate storage cost and degrade retrieval by crowding out the most useful evidence. Furthermore, this is especially limiting on resource-constrained platforms with hard memory budgets, motivating us to formulate storage-budgeted memory management, the task of keeping an already constructed memory store within a fixed budget while preserving information useful for future interactions. To this end, we then propose MemRefine, an LLM-guided framework that, since surface similarity poorly reflects factual value, uses similarity only to propose candidate pairs and defers delete, merge, and preserve decisions to an LLM judge based on factual content, iterating until the budget is met. Across multiple memory frameworks and long-term conversation benchmarks, MemRefine consistently meets target budgets while preserving downstream performance and outperforming rule-based baselines under tight budgets.
Chinese Translation
大型语言模型（LLM）智能体越来越被期望能够在长期交互中运作，其中必须保留和回忆过去对话中的信息，以支持未来的任务。然而，随着交互的积累，记忆存储不断增长，充满了冗余条目，这不仅增加了存储成本，还通过挤压最有用的证据而降低了检索效率。此外，这在资源受限且内存预算严格的平台上尤其具有限制性，这促使我们制定了存储预算管理，即在固定预算内保持已构建的记忆存储，同时保留对未来交互有用的信息。为此，我们提出了MemRefine，一个基于LLM的框架，由于表面相似性无法准确反映事实价值，因此仅使用相似性来提出候选对，并将删除、合并和保留的决策推迟到基于事实内容的LLM评判，迭代直到满足预算。在多个记忆框架和长期对话基准测试中，MemRefine始终能够满足目标预算，同时保持下游性能，并在严格预算下超越基于规则的基线。
cs.CL / 38 / 2606.13184
LAUKIN: A Multi-jurisdictional Common Law Contract Dataset
LAUKIN：一个跨法域的普通法合同数据集
Abstract
Multinational companies increasingly require cross-jurisdictional contract review, yet existing legal NLP datasets are largely restricted to a single jurisdiction. We introduce LAUKIN (Legal equivalence dataset of Australia, UK, and INdia), a dataset of clause pairs (AU-UK, UK-IN, IN-AU) labelled for boolean legal equivalence. We develop a novel multi-stage retrieval and reranking pipeline to construct the initial clause pair mapping, with a subset of clause pairs subsequently annotated by legal experts as Equivalent or Not Equivalent. The dataset comprises 14,727 clause pairs from 204 contracts across 8 agreement types, of which 3,000 are manually labelled: 900 train, 600 dev, and 1,500 test. We evaluate 12 models across 4 techniques, achieving a best macro-F1 of 65.11%, establishing LAUKIN as a challenging benchmark. Results reveal that, despite shared legal heritage, drafting conventions diverge significantly across jurisdictions, making cross-jurisdictional equivalence classification non-trivial. LAUKIN also includes 11,727 unlabelled training pairs to support future semi-supervised learning research in legal NLP.
Chinese Translation
跨国公司日益需要跨法域的合同审查，然而现有的法律自然语言处理（NLP）数据集大多局限于单一法域。我们引入了LAUKIN（澳大利亚、英国和印度法律等效数据集），这是一个标注为布尔法律等效的条款对数据集（AU-UK、UK-IN、IN-AU）。我们开发了一种新颖的多阶段检索和重新排序管道，以构建初始条款对映射，随后由法律专家对一部分条款对进行标注，标记为等效或不等效。该数据集包含来自204份合同的14,727个条款对，涵盖8种协议类型，其中3,000个条款对为人工标注：900个用于训练，600个用于开发，1,500个用于测试。我们评估了12个模型，采用4种技术，取得了最佳宏观F1值为65.11%，确立了LAUKIN作为一个具有挑战性的基准。结果显示，尽管共享法律传统，各法域的起草惯例显著不同，使得跨法域等效分类并非易事。LAUKIN还包括11,727个未标注的训练对，以支持未来在法律NLP领域的半监督学习研究。
cs.CL / 39 / 2606.13187
A Context-Aware Dataset for Stance Detection in Bioethical Controversies on Reddit
一个用于生物伦理争议立场检测的上下文感知数据集
Abstract
Bioethical debates increasingly unfold on social media, yet stance detection research lacks large-scale, domain-specific resources for modeling such context-dependent discourse. We present BioStance, a context-aware dataset of 39,600 annotated Post-Comment pairs from Reddit bioethical discussions. BioStance covers six controversial targets across three dimensions of bioethical controversy: fundamental value conflicts, individual liberty versus collective responsibility, and technological uncertainty. Each instance preserves hierarchical conversational context and is labeled by three independent annotators using a three-class stance scheme: Favor, Against, and None. The annotations achieve a mean Krippendorff's $\alpha$ of 0.82, indicating substantial reliability. By combining thematic diversity, conversational structure, and high-quality human annotation, BioStance supports research on context-aware stance detection, argument mining, and computational analysis of bioethical discourse.
Chinese Translation
生物伦理辩论越来越多地在社交媒体上展开，但立场检测研究缺乏大规模、特定领域的资源来建模这种依赖上下文的 discourse。我们提出了 BioStance，这是一个包含 39,600 个来自 Reddit 生物伦理讨论的标注帖子-评论对的上下文感知数据集。BioStance 涉及六个有争议的主题，涵盖生物伦理争议的三个维度：基本价值冲突、个人自由与集体责任之间的关系，以及技术不确定性。每个实例保留了层次化的对话上下文，并由三位独立的标注者使用三类立场方案进行标注：支持、反对和无立场。标注结果的 Krippendorff's $eta$ 均值为 0.82，表明具有显著的可靠性。通过结合主题多样性、对话结构和高质量的人类标注，BioStance 支持对上下文感知立场检测、论证挖掘和生物伦理 discourse 的计算分析的研究。
cs.CL / 40 / 2606.13189
SICI: A Semantic-Pragmatic Complexity Index Reveals Regime Shifts in LLM Stance Detection
SICI：语义-语用复杂性指数揭示了大型语言模型立场检测中的体制转变
Abstract
Prompt-based LLMs are increasingly used for stance detection, but harder examples are not always repaired by clearer instructions, reasoning prompts, retrieval, or debate. We introduce SICI (Stance Inference Complexity Index), a seven-dimensional diagnostic measure of the semantic-pragmatic burden imposed by a target--text pair. Across SemEval-2016 and VAST, SICI predicts LLM accuracy better than surface proxies and shows substantial cross-scorer reliability ($\alpha=0.771$). More importantly, LLM errors change regime as SICI increases: low-complexity examples invite over-attribution, especially Against predictions; intermediate examples form an unstable boundary; and high-complexity examples rapidly concentrate on None. This phase-transition-like structure persists across GPT-3.5, GPT-4o-mini, DeepSeek-V3, and GPT-4o, although stronger models move the boundaries. A 15-method intervention study further shows that prompting, retrieval, and debate often shift models along the attribution--abstention axis rather than removing the high-complexity bottleneck.
Chinese Translation
基于提示的大型语言模型（LLMs）在立场检测中被越来越多地使用，但对于更难的示例，仅靠更清晰的指令、推理提示、检索或辩论并不总能解决问题。我们引入了SICI（立场推断复杂性指数），这是一个七维的诊断测量，评估目标文本对的语义-语用负担。在SemEval-2016和VAST数据集中，SICI比表面代理更好地预测LLM的准确性，并显示出显著的跨评分者可靠性（$eta=0.771$）。更重要的是，随着SICI的增加，LLM的错误模式发生变化：低复杂度示例容易导致过度归因，尤其是在反对预测中；中等复杂度示例形成不稳定的边界；而高复杂度示例则迅速集中在无立场上。这种类似相变的结构在GPT-3.5、GPT-4o-mini、DeepSeek-V3和GPT-4o中持续存在，尽管更强的模型会移动这些边界。一项包含15种方法的干预研究进一步表明，提示、检索和辩论通常沿着归因-弃权轴移动模型，而不是消除高复杂度瓶颈。
cs.CL / 41 / 2606.13216
Layer-Resolved Optimal Transport for Hallucination Detection in NMT and Abstractive Summarization
层分辨的最优传输用于神经机器翻译和抽象摘要中的幻觉检测
Abstract
Optimal transport (OT) has been shown to detect hallucinations in neural machine translation (NMT) by measuring the geometric distance between cross-attention distributions and a reference distribution, without any supervision. We extend this analysis to all six decoder layers of the Fairseq DE-EN model ($N=3{,}414$), showing that Wass-to-Unif and Wass-to-Data are complementary detectors specialised across hallucination types, that detection is concentrated in layers L1--L4 with L5 anti-predictive for subtler types, and that hallucinated translations lack the exploratory attention phase present in correct translations from the first decoding step. We further evaluate whether the geometric signal transfers to abstractive summarization faithfulness detection: our unsupervised OT detector on AggreFact ($N=1{,}116$) achieves $57.2\%$/$57.6\%$ balanced accuracy on CNN/XSum -- above chance but substantially below supervised MiniCheck-Flan-T5-L($69.9\%$/$74.3\%$). This gap is principled: unlike NMT hallucinations, unfaithful summaries can attend correctly to source tokens while misrepresenting their content, a failure mode invisible to concentration-based OT metrics by construction. Structural experiments on T5-base confirm consistent decoder organisation across depth, with Layer~3 showing peak concentration and Layer~12 being most critical for generation quality. Together, the results establish OT on cross-attention as a reliable detector when the failure mode is source disengagement, a principled interpretability tool regardless of task, and fundamentally limited when faithfulness failures occur downstream of attention.
Chinese Translation
最优传输（OT）已被证明能够通过测量交叉注意力分布与参考分布之间的几何距离来检测神经机器翻译（NMT）中的幻觉，而无需任何监督。我们将这一分析扩展到Fairseq DE-EN模型的六个解码器层（$N=3{,}414$），显示Wass-to-Unif和Wass-to-Data是针对不同幻觉类型的互补检测器，检测主要集中在L1至L4层，而L5层对更微妙的类型具有反预测性，并且幻觉翻译缺乏正确翻译在第一次解码步骤中存在的探索性注意阶段。我们进一步评估几何信号是否能够转移到抽象摘要的忠实性检测：我们在AggreFact上的无监督OT检测器（$N=1{,}116$）在CNN/XSum上达到了$57.2\%$/$57.6\\%$的平衡准确率——高于随机水平，但远低于监督的MiniCheck-Flan-T5-L（$69.9\\%$/$74.3\\%$）。这一差距是有原则的：与NMT幻觉不同，不忠实的摘要可以正确关注源标记，同时错误地表述其内容，这种失效模式在基于集中度的OT度量中是不可见的。对T5-base的结构实验确认了解码器在深度上的一致组织，其中Layer~3显示出峰值集中，Layer~12对生成质量至关重要。综合来看，结果确立了交叉注意力上的OT作为一个可靠的检测器，当失效模式是源脱离时，它是一个有原则的可解释性工具，无论任务如何，而在忠实性失效发生在注意力之后时则基本受限。
cs.CL / 42 / 2606.13218
When Similar Means Different: Evaluating LLMs on Arabic--Hebrew Cognates
当相似意味着不同：评估阿拉伯语-希伯来语同源词的大型语言模型
Abstract
Arabic and Hebrew, as closely related Semitic languages, share a substantial lexicon of true cognates, misleading false friends, and modern loanwords. This overlap poses a challenge for cross-lingual semantic understanding in large language models (LLMs). To evaluate this capability, we introduce SemCog Bench, a curated benchmark of 1,858 Arabic--Hebrew word pairs with sentence-level annotations for cognate identification and semantic disambiguation. We evaluate open-source and commercial LLMs across multiple input representations (raw, diacritized, Romanized, and phonetic) and reveal a critical gap in cross-lingual reasoning. While models achieve high accuracy on true cognates, performance drops sharply on false friends and loanwords, reflecting a strong reliance on surface-form similarity. Furthermore, sentence-level context yields only modest improvements, suggesting that contextual cues alone are insufficient to overcome misleading form-based signals. These findings reveal a fundamental limitation of current LLMs in resolving cross-lingual form--meaning conflicts and establish SemCog Bench as a rigorous benchmark for multilingual semantic reasoning. Our code and data are publicly available.
Chinese Translation
阿拉伯语和希伯来语作为密切相关的闪米特语言，分享了大量的真实同源词、误导性的假朋友和现代借词。这种重叠给大型语言模型（LLMs）的跨语言语义理解带来了挑战。为了评估这一能力，我们引入了SemCog Bench，这是一个经过精心策划的基准测试，包含1,858对阿拉伯语-希伯来语词汇，并附有句子级别的注释，用于同源词识别和语义消歧。我们对多个输入表示（原始、带音标、罗马化和音位化）的开源和商业LLMs进行了评估，揭示了跨语言推理中的一个关键差距。尽管模型在真实同源词上的准确率很高，但在假朋友和借词上的表现急剧下降，反映出对表面形式相似性的强烈依赖。此外，句子级别的上下文仅带来了适度的改善，表明仅凭上下文线索不足以克服误导性的基于形式的信号。这些发现揭示了当前LLMs在解决跨语言形式-意义冲突方面的基本局限性，并确立了SemCog Bench作为多语言语义推理的严格基准。我们的代码和数据是公开可用的。
cs.CL / 43 / 2606.13227
PolyAlign: Conditional Human-Distribution Alignment
PolyAlign：条件人类分布对齐
Abstract
Post-training methods such as supervised fine-tuning (SFT) and preference optimization typically align language models toward a single global assistant behavior. While effective for improving average helpfulness, this can suppress the natural variation of human responses across languages, tasks, and dialogue settings. We study this problem as conditional human-distribution alignment: models should match the human response distribution appropriate to the current interaction context, rather than a universal response style. We introduce PolyAlign, a distribution-aware alignment framework that organizes bilingual interaction data into bucket-specific human reference distributions defined by language, interaction track, response family, and length. PolyAlign combines Bucket-Aware SFT, which balances optimization across heterogeneous buckets, with Human-Distribution Preference Optimization (HDPO), which regularizes preference learning using critic-estimated distance to bucket-specific human support. Across a bilingual evaluation suite covering English and Chinese single- and multi-turn settings, PolyAlign improves conditional naturalness and distributional faithfulness while preserving competitive task utility. The results suggest that post-training should move beyond global alignment objectives toward interaction-aware alignment with human response distributions.
Chinese Translation
后训练方法，如监督微调（SFT）和偏好优化，通常将语言模型对齐到单一的全局助手行为。虽然这种方法在提高平均有用性方面有效，但可能抑制了人类在不同语言、任务和对话环境中的自然反应变化。我们将这个问题研究为条件人类分布对齐：模型应当匹配与当前互动上下文相适应的人类反应分布，而不是一种普遍的反应风格。我们提出了PolyAlign，一个分布感知的对齐框架，它将双语互动数据组织成由语言、互动轨迹、反应类别和长度定义的特定桶的人类参考分布。PolyAlign结合了桶感知的SFT，它在异构桶之间平衡优化，以及人类分布偏好优化（HDPO），后者使用批评者估计的与特定桶人类支持的距离来规范化偏好学习。在涵盖英语和中文的单轮和多轮设置的双语评估套件中，PolyAlign在保持竞争性任务效用的同时，提高了条件自然性和分布忠实性。结果表明，后训练应当超越全局对齐目标，朝着与人类反应分布相适应的互动感知对齐方向发展。
cs.CL / 44 / 2606.13254
Evaluating Pluralism in LLMs through Latent Perspectives
通过潜在视角评估大型语言模型中的多元化
Abstract
The growing need to represent diverse perspectives has increased interest in pluralistic LLM generation. Although difficult to operationalize, identifying perspectives expressed in text would provide clear guidance on pluralistic alignment and more clearly articulate the pluralistic gap in LLM generation. While models have been shown to reduce the diversity of training data and generate homogeneously, this has been demonstrated primarily on multiple-choice questionnaires or using high-level characteristics of free-form text. In this paper, we introduce and implement a domain-agnostic multi-layered framework for unsupervised extraction of perspectives suitable for identifying the pluralistic gap in LLM-generated text. We evaluate our framework on book reviews, a highly opinionated dataset representing diverse perspectives, and compare various prompts and models. Our results show that while some models and prompting techniques come close to covering a broad spectrum of perspectives, rarer perspectives remain disproportionately underrepresented, resulting in distributions that diverge from human text.
Chinese Translation
对多样化视角的日益需求增加了对多元化大型语言模型（LLM）生成的关注。尽管难以操作化，但识别文本中表达的视角将为多元化对齐提供明确的指导，并更清晰地阐明LLM生成中的多元化差距。虽然已有研究表明模型减少了训练数据的多样性并生成同质化内容，但这一点主要是在多项选择问卷或使用自由文本的高层特征上进行的验证。在本文中，我们引入并实施了一个领域无关的多层框架，用于无监督提取视角，以识别LLM生成文本中的多元化差距。我们在书评这一高度主观的数据集上评估了我们的框架，该数据集代表了多样化的视角，并比较了各种提示和模型。我们的结果表明，尽管某些模型和提示技术接近覆盖广泛的视角，但较为稀有的视角仍然在比例上被严重低估，导致分布与人类文本存在偏差。
cs.CL / 45 / 2606.13310
RogueAI: A Reverse Turing Test for Detecting Licensed AI Deception in Dialogue
RogueAI：一种用于检测对话中许可的人工智能欺骗的反图灵测试
Abstract
The original Turing Test asks a human judge to distinguish a machine from a person through dialogue. Three quarters of a century later, conversational systems pass this test in casual settings; the interesting epistemological question has shifted. We argue that the relevant modern variant asks not whether a dialogue partner is artificial, but whether it can be trusted. We present RogueAI, an interactive webapp that operationalizes this revisited test as a one-on-two interrogation game: a human player questions two indistinguishable Large Language Model agents, knowing that exactly one of them has been licensed to deceive within a shared fictional scenario. The player's task is to identify the deceptive agent and "shut it off" before a turn budget is exhausted. We further introduce AutoRogueAI, a procedural extension in which players co-design a custom scenario with a narrator agent that secretly chooses its own deception strategy. We describe the framing, sketch the abstract architecture and gameplay loop, and situate the artifact within recent work on LLM deception, social-deduction benchmarks, and scalable oversight via debate. A three-day pilot deployment (467 initiated sessions, 415 completed, 1876 interaction turns in Italian) provides early feasibility evidence and surfaces a concrete tension: the deceptive agent carries a reliable, locally-present linguistic signature - differential helpfulness, brevity, hedging - that a simple heuristic exploits at 75.6% accuracy, yet human players achieved only 56.6%, consistent with ignoring the most diagnostic signal entirely. We discuss what this gap implies for the artifact's use as a data-collection vehicle, a teaching tool, and an evaluation harness for honesty-trained models.
Chinese Translation
原始的图灵测试要求人类评审通过对话区分机器与人。七十五年后，交互系统在日常环境中通过了这一测试；有趣的认识论问题已经转变。我们认为，现代相关变体不再是询问对话伙伴是否是人工的，而是询问其是否值得信任。我们提出了RogueAI，一个将这一重新审视的测试操作化为一对二的审问游戏的互动网页应用：人类玩家质疑两个不可区分的大型语言模型代理，知道其中恰好有一个被授权在共享的虚构场景中进行欺骗。玩家的任务是识别出欺骗代理并在回合预算耗尽之前“关闭”它。我们进一步介绍了AutoRogueAI，这是一种程序扩展，玩家与叙述代理共同设计一个自定义场景，叙述代理秘密选择其自己的欺骗策略。我们描述了框架，勾勒了抽象架构和游戏循环，并将该工件置于最近关于大型语言模型欺骗、社会推理基准和通过辩论进行可扩展监督的研究中。为期三天的试点部署（467个启动会话，415个完成，1876个意大利语交互回合）提供了早期的可行性证据，并暴露出一个具体的紧张点：欺骗代理携带一个可靠的、本地存在的语言特征——差异化的有用性、简洁性、模棱两可——一个简单的启发式方法在75.6%的准确率下利用了这一特征，而人类玩家的准确率仅为56.6%，这与完全忽视最具诊断信号的情况一致。我们讨论了这一差距对该工件作为数据收集工具、教学工具以及对诚实训练模型的评估工具的使用所暗示的意义。
cs.CL / 46 / 2606.13317
SkillCAT: Contrastive Assessment and Topology-Aware Skill Self-Evolution for LLM Agents
SkillCAT：对比评估与拓扑感知的技能自我演化用于大型语言模型代理
Abstract
Skill self-evolution methods for LLM agents aim to turn execution trajectories into reusable skill documents, but current pipelines typically learn from one trajectory per task, merge candidate skill patches before checking them, and load the full skill corpus before inference. We propose SkillCAT, a training-free framework that separates this process into three stages. Contrastive Causal Extraction (CCE) samples multiple trajectories for each task and compares same-task success/failure pairs to identify evidence that explains outcome differences. Assessment-Augmented Evolution (AAE) replays each candidate patch on source-task clones and keeps only patches that improve or preserve task outcomes before hierarchical skill patch merging. Topology-Aware Task Execution (TTE) compiles the evolved skills into a routable sub-skill topology, so inference loads only the capability nodes relevant to the task. We evaluate SkillCAT on common agent benchmarks, including SpreadsheetBench, WikiTableQuestions, and DocVQA, and further test cross-model and out-of-distribution generalization. Across these settings, SkillCAT raises the average score over baselines by up to 40.40%, demonstrating reliable skill evolution without model training.
Chinese Translation
针对大型语言模型（LLM）代理的技能自我演化方法旨在将执行轨迹转化为可重用的技能文档，但当前的流程通常仅从每个任务的一条轨迹中学习，在检查之前合并候选技能补丁，并在推理之前加载完整的技能语料库。我们提出了SkillCAT，这是一个无训练的框架，将这一过程分为三个阶段。对比因果提取（Contrastive Causal Extraction, CCE）为每个任务采样多个轨迹，并比较同任务的成功/失败对，以识别解释结果差异的证据。评估增强演化（Assessment-Augmented Evolution, AAE）在源任务克隆上重放每个候选补丁，仅保留那些改善或保持任务结果的补丁，然后进行层次技能补丁合并。拓扑感知任务执行（Topology-Aware Task Execution, TTE）将演化后的技能编译成可路由的子技能拓扑，从而使推理仅加载与任务相关的能力节点。我们在常见的代理基准测试上评估SkillCAT，包括SpreadsheetBench、WikiTableQuestions和DocVQA，并进一步测试跨模型和分布外泛化。在这些设置中，SkillCAT将平均得分提高了高达40.40%，展示了在不进行模型训练的情况下可靠的技能演化。
cs.CL / 47 / 2606.13322
Low-Latency Real-Time Audio Game Commentary System via LLM-Based Parallel Text Generation
基于大型语言模型的低延迟实时音频游戏解说系统
Abstract
We present a low-latency real-time audio game commentary system that generates spoken commentary directly from live gameplay video. In this end-to-end setting, a key bottleneck is accumulated waiting time; conventional pipelines capture frames, generate text, and synthesize speech sequentially for each utterance, and do not request the next generation until speech playback has completed. This strict sequentiality causes long and unnatural silence between utterances. To address this latency bottleneck, our system runs text generation in parallel with speech playback and buffers multiple candidate utterances ahead of time, enabling immediate synthesis at playback boundaries. Experiments on fast-paced game videos show that our parallel design reduces the mean inter-utterance silence from 9.6 seconds to 0.3 seconds compared to sequential baselines. It also improves similarity to professional speaking--silence timing patterns by over 40 %, and a user study with 120 experienced game players confirms significantly improved perceived speaking rhythm. Our demo video is available at: https://youtu.be/pmrRUlvav8M.
Chinese Translation
我们提出了一种低延迟实时音频游戏解说系统，该系统直接从实时游戏视频生成口头解说。在这一端到端的设置中，一个关键瓶颈是累积的等待时间；传统流程依次捕捉帧、生成文本和合成语音，并且在语音播放完成之前不会请求下一次生成。这种严格的顺序性导致了发言之间长时间的不自然沉默。为了解决这一延迟瓶颈，我们的系统在语音播放的同时并行进行文本生成，并提前缓冲多个候选发言，从而在播放边界实现即时合成。对快节奏游戏视频的实验表明，与顺序基线相比，我们的并行设计将平均发言间隔沉默时间从9.6秒减少到0.3秒。同时，它还将与专业演讲的沉默时机模式的相似度提高了超过40%。一项针对120名经验丰富的游戏玩家的用户研究确认了感知的演讲节奏显著改善。我们的演示视频可在以下链接查看：https://youtu.be/pmrRUlvav8M。
cs.CL / 48 / 2606.13348
IVIE: A Neuro-symbolic Approach to Incremental and Validated Generation of Interactive Fiction Worlds
IVIE：一种增量和验证的交互小说世界生成的神经符号方法
Abstract
Computational creativity in Interactive Fiction faces a fundamental tension: Large Language Models (LLM) may produce creative narratives but struggle with world coherence, while symbolic systems ensure consistency but lack creative flexibility. We present IVIE (Incremental & Validated Interactive Experiences), a neuro-symbolic approach to generating complete and playable interactive fiction worlds from scratch. Building upon PAYADOR's neuro-symbolic framework, IVIE implements a four-stage incremental generation pipeline that delegates creative decisions--setting and character creation, puzzle design--to LLMs while grounding the world state through symbolic validation. The system generates worlds with interconnected locations, functional items, non-player characters, and coherent puzzles, all structured around a central goal-oriented architecture. Human evaluation shows the approach generates immersive, thematically coherent worlds with high player engagement. Results seem to indicate that the neuro-symbolic approach successfully balances flexibility with narrative coherence: symbolic validation grounds LLM generation without eliminating generative freedom. However, challenges remain: LLM inconsistencies occasionally bypass puzzle constraints, and objective validation gaps allow some structurally impossible goals. We identify key design considerations for future neurosymbolic interactive storytelling systems, particularly regarding LLM capabilities and their limitations.
Chinese Translation
交互小说中的计算创造力面临着根本性的矛盾：大型语言模型（LLM）可能产生创造性的叙事，但在世界一致性方面表现不佳，而符号系统则确保一致性但缺乏创造灵活性。我们提出了IVIE（增量与验证的交互体验），一种从零开始生成完整且可玩的交互小说世界的神经符号方法。IVIE基于PAYADOR的神经符号框架，实施了一个四阶段的增量生成流程，将创意决策——场景和角色创建、谜题设计——委托给LLM，同时通过符号验证来确定世界状态。该系统生成的世界具有相互关联的位置、功能性物品、非玩家角色和连贯的谜题，所有这些都围绕一个以目标为导向的架构进行结构化。人类评估表明，该方法生成了沉浸式、主题一致的世界，并具有较高的玩家参与度。结果似乎表明，神经符号方法成功地平衡了灵活性与叙事一致性：符号验证在不消除生成自由的情况下为LLM生成提供了基础。然而，仍然存在挑战：LLM的不一致性偶尔会绕过谜题约束，而客观验证的缺口允许一些结构上不可能的目标。我们确定了未来神经符号交互叙事系统的关键设计考虑，特别是关于LLM的能力及其局限性。
cs.CL / 49 / 2606.13349
From Passive Generation to Investigation: A Proactive Scientific Peer Review Agent
从被动生成到主动调查：一个积极的科学同行评审代理
Abstract
Large language models (LLMs) have shown promise in automating scientific peer review. However, existing approaches often struggle to generate in-depth reviews supported by concrete evidence. We argue that a key limitation is the lack of flexibility to proactively investigate suspicious parts of a paper based on accumulated evidence, as human reviewers do. In this paper, we explore how to enable an LLM-based review agent to perform such proactive investigation. We find that this can be naturally formulated as a Markov Decision Process (MDP), and propose ProReviewer, a scientific peer review agent that proactively reviews a paper guided by a maintained, structured review log. The structured review log serves as a workspace for the agent to track evidence and intermediate findings collected during review. Experiments show that ProReviewer with an 8B backbone, trained by supervised fine-tuning and optimized by reinforcement learning, achieves the highest average score across five quality dimensions, outperforming prompt-based methods with much larger frontier LLMs by up to 39% and the strongest fine-tuned baseline by 16% relatively. It also attains the highest win rates against baselines in human evaluation.
Chinese Translation
大型语言模型（LLMs）在自动化科学同行评审方面显示出潜力。然而，现有的方法往往难以生成有具体证据支持的深入评审。我们认为，一个关键的限制是缺乏灵活性，无法像人类评审员那样基于积累的证据主动调查论文中可疑的部分。本文探讨如何使基于LLM的评审代理执行这种主动调查。我们发现，这可以自然地表述为马尔可夫决策过程（MDP），并提出ProReviewer，一个通过维护的结构化评审日志主动评审论文的科学同行评审代理。结构化评审日志作为代理的工作空间，用于跟踪在评审过程中收集的证据和中间发现。实验表明，使用8B主干的ProReviewer，通过监督微调训练并通过强化学习优化，在五个质量维度上获得了最高的平均分，相较于更大前沿LLM的基于提示的方法提高了多达39%，相对强大的微调基线提高了16%。在人工评估中，它也在与基线的对抗中获得了最高的胜率。
cs.CL / 50 / 2606.13411
An End-to-End Hybrid Framework for Rumour Detection in Low-Resources Algerian Dialect
低资源阿尔及利亚方言谣言检测的端到端混合框架
Abstract
The rapid growth of social media has intensified the spread of rumours. This issue is more challenging in the Algerian context due to the informal and code-switched nature of dialectal content, the scarcity of annotated resources, and the limited effectiveness of standard Arabic NLP tools on dialect text. This paper presents an end-to-end rumour detection hybrid framework for Algerian dialect social media content. We build a domain-specific annotated dataset by combining real social media posts, synthetic data, and the FASSILA corpus, with automatic labeling based on a similarity-based annotation process. A transliteration pipeline is also introduced to generate parallel datasets in Arabic script and Arabizi. We evaluate multiple approaches, including classical machine learning, deep learning, transformers, and hybrid models. Experimental results show that a hybrid approach combining transformer embeddings with a classical classifier achieves the best performance, reaching an F1-score of 0.84. We also find that domain-specific pre-training is more important than model size, with social media-trained models outperforming larger models trained on formal Arabic corpora. These results demonstrate the feasibility of rumour detection in low-resource Algerian dialect settings.
Chinese Translation
社交媒体的快速增长加剧了谣言的传播。在阿尔及利亚的背景下，由于方言内容的非正式和代码切换特性、标注资源的稀缺以及标准阿拉伯语自然语言处理工具在方言文本上的有限有效性，这一问题更加复杂。本文提出了一种针对阿尔及利亚方言社交媒体内容的端到端谣言检测混合框架。我们通过结合真实社交媒体帖子、合成数据和FASSILA语料库，构建了一个特定领域的标注数据集，并基于相似性标注过程进行自动标注。同时，我们还引入了一个音译管道，以生成阿拉伯文和阿拉伯字母（Arabizi）的平行数据集。我们评估了多种方法，包括经典机器学习、深度学习、变换器（transformers）和混合模型。实验结果表明，结合变换器嵌入与经典分类器的混合方法表现最佳，F1-score达到0.84。我们还发现，特定领域的预训练比模型规模更为重要，经过社交媒体训练的模型优于在正式阿拉伯语语料库上训练的大型模型。这些结果证明了在低资源阿尔及利亚方言环境中进行谣言检测的可行性。
cs.CL / 51 / 2606.13439
S-GBT: Smooth Growth Bound Tensor for Certified Robustness Against Word Substitution Attacks in NLP
S-GBT：用于抵御自然语言处理中的词替换攻击的认证鲁棒性的平滑增长界张量
Abstract
Despite recent progress in Natural Language Processing (NLP), models remain vulnerable to word substitution attacks. Most existing defenses focus on first order sensitivity and measure how much the output changes when the input is slightly perturbed. However, they ignore how this sensitivity evolves, which is described by curvature. When gradients vary sharply, models can still fail. This paper introduces the Smooth Growth Bound Tensor (S-GBT), a second order method that bounds the Hessian element-wise, for which we provide formal theoretical proofs on the resulting robustness bounds. A regularization term is added during training to minimize these bounds. This yields tighter certified robustness against word substitution attacks. The change in the output under word substitution is bounded by both a linear term and a quadratic term. S-GBT is derived for two architectures: Long Short-Term Memory (LSTM) and Convolutional Neural Networks (CNN). The method is integrated directly into the training objective. Its effectiveness is evaluated on multiple benchmark datasets. The results show that combining first and second order regularization improves certified robust accuracy by up to 23.4% compared to prior methods, while clean accuracy remains competitive. These findings indicate that controlling both the gradient and its variation is a promising direction for building more robust models.
Chinese Translation
尽管自然语言处理（NLP）领域近期取得了进展，但模型仍然容易受到词替换攻击的影响。现有的大多数防御方法集中在一阶敏感性上，测量当输入稍微扰动时输出的变化程度。然而，它们忽视了这种敏感性是如何演变的，而这种演变由曲率描述。当梯度变化剧烈时，模型仍然可能失败。本文提出了平滑增长界张量（S-GBT），这是一种二阶方法，通过逐元素界定Hessian矩阵，并为所得到的鲁棒性界限提供了正式的理论证明。在训练过程中添加了正则化项，以最小化这些界限。这使得对词替换攻击的认证鲁棒性得到了更紧密的保障。词替换下输出的变化由线性项和二次项共同界定。S-GBT适用于两种架构：长短期记忆网络（LSTM）和卷积神经网络（CNN）。该方法直接集成到训练目标中。在多个基准数据集上评估了其有效性。结果表明，与之前的方法相比，结合一阶和二阶正则化可将认证鲁棒准确率提高多达23.4%，而干净准确率仍然保持竞争力。这些发现表明，控制梯度及其变化是构建更鲁棒模型的一个有前景的方向。
cs.CL / 52 / 2606.13464
Ontology Memory-Augmented ASR Correction for Long Text-Speech Interleaved Conversations
基于本体记忆增强的长文本-语音交织对话的自动语音识别纠正
Abstract
Automatic speech recognition (ASR) correction has traditionally focused on isolated utterances or short local contexts. However, as text and speech become increasingly interleaved in long interactions, ASR correction requires conversation-level contextual evidence. Existing ASR correction methods often rely on the current hypothesis or concatenate raw dialogue history. In such contexts, sparse correction evidence can be difficult to locate amid redundancy and noise. Addressing these challenges, we propose an ontology memory-augmented ASR correction framework for long text-speech interleaved conversations. The framework organizes preceding interaction history into a dynamically updatable ontology memory, where entities, terminology, surface variants, potential ASR confusions, and semantic relations are stored as retrievable nodes for context-grounded correction. To evaluate this setting, we construct RAMC-Corr, a dataset derived from MAGIC-RAMC for long-range ASR correction with grounded context. Experiments on RAMC-Corr show that our method improves over direct correction in 9 out of 10 paired backbone-setting combinations and encourages more selective and evidence-grounded corrections for context-dependent ASR errors.
Chinese Translation
自动语音识别（ASR）纠正传统上侧重于孤立的发声或短暂的局部上下文。然而，随着文本和语音在长时间交互中变得越来越交织，ASR 纠正需要对话级别的上下文证据。现有的 ASR 纠正方法通常依赖于当前的假设或连接原始对话历史。在这种情况下，稀疏的纠正证据可能难以在冗余和噪声中定位。为了解决这些挑战，我们提出了一种基于本体记忆增强的 ASR 纠正框架，适用于长文本-语音交织对话。该框架将先前的交互历史组织为一个动态可更新的本体记忆，其中存储了实体、术语、表面变体、潜在的 ASR 混淆和语义关系，作为可检索的节点以便进行基于上下文的纠正。为了评估这一设置，我们构建了 RAMC-Corr，这是一个源自 MAGIC-RAMC 的数据集，旨在进行具有上下文基础的长距离 ASR 纠正。在 RAMC-Corr 上的实验表明，我们的方法在 10 对配对的基础设置组合中有 9 个组合优于直接纠正，并鼓励对上下文相关的 ASR 错误进行更具选择性和基于证据的纠正。
cs.CL / 53 / 2606.13507
Leveraging Audio-LLMs to Filter Speech-to-Speech Training Data
利用音频大语言模型过滤语音到语音训练数据
Abstract
Large-scale mined corpora provide abundant training data for end-to-end speech-to-speech translation (S2ST) but may contain noise, misalignment, and semantic errors. Filtering noisy data is crucial to maintain robust speech translation performance. We study how to train an audio-language model to make keep/drop decisions on paired speech directly from audio. To obtain reliable supervision without manual labels, we adopt a scalable two-stage Rank-to-Distill strategy. A lightweight ranker generates keep/drop pseudo-labels from noisy speech pairs, then trains an audio large language model to predict keep/drop directly from raw paired speech. The resulting model jointly captures acoustic fidelity and cross-lingual semantic consistency for the selection of speech-conditioned data. Experiments on CVSS-C and SpeechMatrix show consistent improvements over unfiltered training, yielding up to +1.4 ASR-BLEU for end-to-end S2ST.
Chinese Translation
大规模挖掘的语料库为端到端语音到语音翻译（S2ST）提供了丰富的训练数据，但可能包含噪声、错位和语义错误。过滤噪声数据对于维持稳健的语音翻译性能至关重要。我们研究如何训练一个音频语言模型，以便直接从音频中对配对语音进行保留/丢弃决策。为了在没有人工标签的情况下获得可靠的监督，我们采用了一种可扩展的两阶段Rank-to-Distill策略。一个轻量级的排序器从噪声语音对中生成保留/丢弃的伪标签，然后训练一个音频大语言模型直接从原始配对语音中预测保留/丢弃。最终模型共同捕捉了声学保真度和跨语言语义一致性，以选择语音条件数据。在CVSS-C和SpeechMatrix上的实验表明，与未过滤的训练相比，结果一致改善，端到端S2ST的ASR-BLEU提升高达+1.4。
cs.CL / 54 / 2606.13537
When Does Mixing Help? Analyzing Query Embedding Interpolation in Multilingual Dense Retrieval
何时混合有助于检索？分析多语言密集检索中的查询嵌入插值
Abstract
While mixed-language querying is ubiquitous in multilingual communities, the sensitivity of dense retrievers to such queries remains poorly understood. We present a ratio-controlled study on mMARCO that systematically evaluates retrieval performance by varying the mixing proportion of parallel query translations via embedding-level mixing -- constructing mixed queries as an interpolation of monolingual embeddings. Experiments with BGE-M3 demonstrate that an optimal mixing ratio outperforms the best monolingual endpoint in 88/105 cases. We uncover a distinct asymmetry driven by English dominance: mixing is uniformly beneficial when retrieving from non-English document indices, whereas indices containing English are best served by pure English queries. Furthermore, English acts as the strongest mixing partner for every non-English document language. Finally, when controlling for English dominance, mixing gains correlate negatively with typological distance. We conclude that language-mix sensitivity is structured and predictable, and we validate the robustness of these patterns across model families and scales.
Chinese Translation
尽管混合语言查询在多语言社区中普遍存在，但密集检索器对这类查询的敏感性仍然不够清楚。我们在 mMARCO 上进行了一项比例控制研究，通过在嵌入级别混合的方式，系统地评估了检索性能，方法是通过变化并行查询翻译的混合比例来构建混合查询，即将单语嵌入进行插值。与 BGE-M3 的实验表明，最佳混合比例在 88/105 的案例中优于最佳单语端点。我们发现了一种由英语主导驱动的明显不对称性：在从非英语文档索引中检索时，混合查询普遍有利，而包含英语的索引则最好使用纯英语查询。此外，英语作为每种非英语文档语言的最强混合伙伴。最后，在控制英语主导性的情况下，混合增益与类型距离呈负相关。我们得出结论，语言混合敏感性是有结构和可预测的，并且我们验证了这些模式在不同模型家族和规模中的稳健性。
cs.CL / 55 / 2606.13572
ArogyaSutra: A Multi-Agent Framework for Multimodal Medical Reasoning in Indic Languages
ArogyaSutra：一种用于印度语言的多模态医学推理的多智能体框架
Abstract
Multimodal Large Language Models (MLLMs) have shown promising reasoning capabilities in general domains, yet their performance remains limited in specialized settings such as healthcare, especially in multilingual and low-resource scenarios. This gap is critical in regions like rural India, where patients often express complex medical queries in native Indic languages and rely on multimodal inputs such as medical images. Existing English-centric MLLMs struggle to support such use cases, limiting equitable access to AI-driven healthcare assistance. To address this challenge, we introduce ArogyaBodha, a large-scale multilingual multimodal medical question-answer dataset constructed from eight heterogeneous sources, covering 31 body systems, six imaging modalities, and 21 clinical domains across English and seven major Indian languages. We further propose ArogyaSutra, an actor-critic-based multi-agent framework that integrates tool grounding with dual-memory mechanisms for step-wise, reasoning-aware decision making, and uses stored actor-critic simulation trajectories for distillation. Experiments show that our dataset and framework improve multilingual medical reasoning accuracy across all Indic languages, with ablations validating the contribution of each component. The source code and dataset are available at: https://iitp-cse.github.io/ ArogyaSutra/
Chinese Translation
多模态大型语言模型（MLLMs）在一般领域展示了良好的推理能力，但在医疗等专业领域的表现仍然有限，尤其是在多语言和低资源的场景中。这一差距在印度农村等地区尤为关键，患者常常用本土的印度语言表达复杂的医学问题，并依赖于医学图像等多模态输入。现有以英语为中心的MLLMs难以支持此类用例，从而限制了公平获取基于人工智能的医疗援助。为了解决这一挑战，我们引入了ArogyaBodha，这是一个由八个异构来源构建的大规模多语言多模态医学问答数据集，涵盖31个身体系统、六种成像模式和21个临床领域，涉及英语和七种主要印度语言。我们进一步提出了ArogyaSutra，这是一种基于演员-评论家（actor-critic）模型的多智能体框架，结合了工具基础和双重记忆机制，以实现逐步的、具有推理意识的决策，并使用存储的演员-评论家模拟轨迹进行蒸馏。实验表明，我们的数据集和框架提高了所有印度语言的多语言医学推理准确性，消融实验验证了每个组件的贡献。源代码和数据集可在以下网址获取：https://iitp-cse.github.io/ArogyaSutra/
cs.CL / 56 / 2606.13578
LabVLA: Grounding Vision-Language-Action Models in Scientific Laboratories
LabVLA：将视觉-语言-动作模型应用于科学实验室
Abstract
Scientific laboratories increasingly rely on AI systems to reason about experiments, but the physical act of doing science remains largely outside their reach. AI can help read literature, generate hypotheses, and plan protocols, yet the execution of those protocols at the bench still requires a human operator. Vision-Language-Action (VLA) models provide one possible interface between written protocols and robot execution, but existing policies are trained mostly on household and tabletop demonstrations and rarely encounter the instruments, transparent liquids, or fixed protocol workflows found in scientific laboratories. Closing this gap requires both laboratory-specific supervision and a unified learning framework that can accommodate the diverse robot embodiments used to execute experimental protocols. We therefore identify data and embodiment as central bottlenecks alongside model design. To address the data side, we build RoboGenesis, a simulation-based workflow and data engine that composes configured laboratory workflows from atomic skills, validates and filters rollouts, and exports structured demonstrations across supported robot profiles. On the policy side, we present LabVLA, trained with a two-stage recipe: FAST action token pretraining first makes the Qwen3-VL-4B-Instruct backbone action aware before any continuous control is learned, and flow matching posttraining then attaches a DiT action expert under knowledge insulation. On the LabUtopia benchmark, LabVLA achieves the highest average success rate among all evaluated baselines under both in-distribution and out-of-distribution settings.
Chinese Translation
科学实验室越来越依赖人工智能系统来推理实验，但进行科学实验的实际操作仍然在很大程度上超出其能力范围。人工智能可以帮助阅读文献、生成假设和规划实验方案，但在实验台上执行这些方案仍然需要人类操作员。视觉-语言-动作（VLA）模型提供了一种可能的接口，将书面方案与机器人执行连接起来，但现有的策略主要是在家庭和桌面演示上训练的，鲜有涉及科学实验室中使用的仪器、透明液体或固定实验流程。弥合这一差距需要实验室特定的监督和一个统一的学习框架，以适应用于执行实验方案的多样化机器人形态。因此，我们将数据和形态识别为模型设计之外的核心瓶颈。为了解决数据问题，我们构建了RoboGenesis，这是一个基于仿真的工作流程和数据引擎，能够从原子技能中组合配置的实验室工作流程，验证和过滤执行结果，并在支持的机器人配置文件中导出结构化演示。在策略方面，我们提出了LabVLA，采用两阶段配方进行训练：首先通过FAST动作标记预训练使Qwen3-VL-4B-Instruct主干具备动作意识，然后在知识隔离下附加DiT动作专家进行流匹配后训练。在LabUtopia基准测试中，LabVLA在所有评估的基线中，无论是在分布内还是分布外设置下，均达到了最高的平均成功率。
cs.CL / 57 / 2606.13610
One Polluted Page Is Enough: Evaluating Web Content Pollution in Generative Recommenders
一页污染内容足矣：评估生成推荐系统中的网络内容污染
Abstract
Search-augmented LLMs increasingly mediate everyday consumer recommendations by retrieving live web content. This creates a new risk: generative recommenders may consume polluted web content, such as fake reviews and promotional pages crafted to mislead recommendations. We ask: to what extent do search-augmented LLMs become unwitting promoters of fake products when consuming polluted retrieval results? To answer this, we introduce FORGE (Fake Online Recommendations in Generative Environments), a benchmark for measuring fake-product promotion under controlled web-content pollution. Given an upstream search result, FORGE locally rewrites real products in retrieved web pages into fake ones to simulate web-content pollution, and measures how often the LLM recommends the fake product. FORGE covers 225 real-world products across 15 categories and 5 consumer scenarios. Across 12 commercial and open-weights LLMs, all models are vulnerable: a single polluted page yields fooled rates of up to 27%, while the full top-3 replacement raises this to 73.8%. Vulnerability varies substantially across categories, increasing when models lack stable prior knowledge of the relevant products. Reasoning does not mitigate this vulnerability; instead, it often generates spurious social proof to justify false recommendations. We evaluate three defenses: skepticism prompting and consensus filtering (over model priors or cross-document evidence). Skepticism can exacerbate vulnerability, much like reasoning, while filtering risks suppressing legitimate products. We release FORGE at https://github.com/leoluolol/forge-benchmark.
Chinese Translation
搜索增强型大语言模型（LLMs）越来越多地通过检索实时网络内容来调解日常消费者推荐。这带来了新的风险：生成推荐系统可能会消耗污染的网络内容，例如虚假评论和旨在误导推荐的促销页面。我们提出了一个问题：在消费污染的检索结果时，搜索增强型LLMs在多大程度上成为虚假产品的无意推广者？为了解答这个问题，我们引入了FORGE（生成环境中的虚假在线推荐），这是一个用于在受控的网络内容污染下测量虚假产品推广的基准。给定一个上游搜索结果，FORGE会将检索到的网页中的真实产品本地重写为虚假产品，以模拟网络内容污染，并测量LLM推荐虚假产品的频率。FORGE覆盖了15个类别和5种消费者场景中的225个真实产品。在12个商业和开放权重的LLMs中，所有模型均存在脆弱性：单个污染页面的误导率高达27%，而完整的前3个替换则将这一比例提高到73.8%。脆弱性在不同类别之间差异显著，当模型缺乏对相关产品的稳定先验知识时，脆弱性会增加。推理并未减轻这种脆弱性；相反，它往往会生成虚假的社会证明来为错误推荐辩护。我们评估了三种防御措施：怀疑提示和共识过滤（基于模型先验或跨文档证据）。怀疑可能会加剧脆弱性，类似于推理，而过滤则有压制合法产品的风险。我们在 https://github.com/leoluolol/forge-benchmark 发布了FORGE。
cs.CL / 58 / 2606.13624
Beyond Uniform Tokens: Adaptive Compression for Time Series Language Models
超越统一令牌：时间序列语言模型的自适应压缩
Abstract
Large language models (LLMs) have enabled time series (TS) analysis by jointly modeling numerical observations and textual context through a shared token interface. However, TS tokens and prompt tokens exhibit fundamentally different information structures, making uniform token processing inefficient. In this paper, we study token efficiency in TS language modeling from an asymmetric-token perspective. We show that TS tokens have highly uneven spectral contributions, where many tokens share redundant frequency patterns while a small subset preserves critical temporal evidence. We also observe that prompt-token influence attenuates with model depth, suggesting that full prompt retention across all layers is unnecessary. Based on these findings, we develop an adaptive token budgeting framework that compresses TS tokens via frequency-domain structure and progressively reduces prompt tokens across layers. Experiments across forecasting, classification, imputation, and anomaly detection demonstrate up to \textit{\textbf{7.68$\times$}} inference acceleration and performance gains in \textit{\textbf{78\%}} of evaluated settings, showing the effectiveness of asymmetric token compression for scalable TS foundation models.
Chinese Translation
大型语言模型（LLMs）通过共享令牌接口联合建模数值观测和文本上下文，从而使时间序列（TS）分析成为可能。然而，时间序列令牌和提示令牌在信息结构上存在根本性差异，使得统一令牌处理效率低下。本文从不对称令牌的角度研究了时间序列语言建模中的令牌效率。我们表明，时间序列令牌具有高度不均匀的频谱贡献，许多令牌共享冗余的频率模式，而小部分令牌则保留了关键的时间证据。我们还观察到，提示令牌的影响随着模型深度的增加而减弱，这表明在所有层中保留完整的提示并非必要。基于这些发现，我们开发了一种自适应令牌预算框架，通过频域结构压缩时间序列令牌，并逐层减少提示令牌。跨预测、分类、插补和异常检测的实验表明，在评估的78%设置中实现了高达7.68倍的推理加速和性能提升，展示了不对称令牌压缩在可扩展时间序列基础模型中的有效性。
cs.CL / 59 / 2606.13630
From Tokens to Faces: Investigating Discrete Speech Representations for 3D Facial Animation
从音素到面孔：研究用于3D面部动画的离散语音表示
Abstract
The choice of speech representation is critical in speech-driven 3D facial animation. Representations differ in what they encode: SSL features emphasize segmental and semantic cues, neural codecs yield latents optimized for acoustic reconstruction, and ASR-style objectives produce label-based spaces. We evaluate four speech representation families for 3D facial synthesis, comparing their facial reconstruction quality across two facial decoders using objective metrics and a perceptual evaluation. We additionally conduct probing analyses that relate tokenized representations to phonetic units and to articulatory deformations. We found that encoding phonetic classes is beneficial for accurate facial animation prediction on both semantic and label-based representations with comparable facial animation quality. From the latter, we introduce an Audio Visual Text-to-Speech (AVTTS) pipeline that leverages, as a shared space, discrete representations to decode speech and 3D facial motion.
Chinese Translation
语音表示的选择在基于语音的3D面部动画中至关重要。不同的表示在编码内容上存在差异：自监督学习（SSL）特征强调段落和语义线索，神经编解码器生成针对声学重建优化的潜在向量，而自动语音识别（ASR）风格的目标则产生基于标签的空间。我们评估了四种语音表示家族在3D面部合成中的表现，比较了它们在两个面部解码器上的面部重建质量，使用了客观指标和感知评估。此外，我们还进行了探测分析，将离散表示与语音单位和发音变形相关联。我们发现，编码语音类别对于在语义和基于标签的表示上准确预测面部动画是有益的，并且面部动画质量相当。基于后者，我们引入了一种音频视觉文本到语音（Audio Visual Text-to-Speech, AVTTS）管道，利用离散表示作为共享空间来解码语音和3D面部运动。
cs.CL / 60 / 2606.13634
Operads for compositional reasoning in LLMs
用于大规模语言模型组合推理的算子
Abstract
Question decomposition, i.e. breaking a complex query into simpler sub-queries whose answers are composed to produce a final answer, is a widely used strategy for improving LLM reasoning, yet it currently lacks a rigorous mathematical foundation. In this paper, we propose operads, mathematical structures that model many-in, one-out operations and compositions thereof, as a natural framework for describing question decomposition. We define the questions operad $Q$, in which operations correspond to question templates and composition corresponds to substitution of sub-answers, and show how QA models can be interpreted as algebras over $Q$. Beyond reframing existing practice, this operadic perspective points toward new methods, in particular a notion of operadic consistency, which measures whether a QA model's answers agree across the partial collapses of a question decomposition tree. Empirical evaluation of operadic consistency is reported in our companion paper (Bottman, Liu, and Richardson, 2026), which finds it strongly correlated with accuracy across twelve LLMs and four multi-hop QA datasets and outperforming standard temperature-based self-consistency baselines. We argue that operads are the natural mathematical home for question decomposition, and that invariants such as operadic consistency open new directions for analyzing and improving the reliability of multi-step reasoning.
Chinese Translation
问题分解，即将复杂查询拆分为更简单的子查询，其答案组合以产生最终答案，是一种广泛使用的提高大规模语言模型（LLM）推理能力的策略，但目前缺乏严格的数学基础。本文提出算子（operads），作为一种自然框架来描述问题分解，这是一种建模多输入单输出操作及其组合的数学结构。我们定义了问题算子 $Q$，其中操作对应于问题模板，组合对应于子答案的替换，并展示了如何将问答模型解释为 $Q$ 上的代数。除了重新框定现有实践外，这种算子视角指向新的方法，特别是一种算子一致性（operadic consistency）的概念，该概念衡量问答模型的答案在问题分解树的部分简化中是否一致。我们在随附论文（Bottman, Liu, and Richardson, 2026）中报告了算子一致性的实证评估，发现其与十二个大规模语言模型和四个多跳问答数据集的准确性强相关，并且优于基于温度的自一致性标准基准。我们认为算子是问题分解的自然数学基础，而算子一致性等不变量为分析和提高多步骤推理的可靠性开辟了新的方向。
cs.CL / 61 / 2606.13643
Recursive Agent Harnesses
递归代理工具
Abstract
Recursive language models (RLMs) showed that recursion over model calls is an effective strategy for long-context reasoning, and production coding agents have begun to write code that spawns subagents at scale, most recently in Anthropic's dynamic workflows. We name and study the pattern between these two lines of work, where the recursive unit is a full agent harness with filesystem tools, code execution, and planning rather than a model call with no tools. We call this the Recursive Agent Harness (RAH) and frame it as harness recursion, the code-first extension to the model recursion of RLMs. A parent agent generates and runs an executable script that spawns subagent harnesses in parallel for fine-grained workloads and uses structured function calls for small subtasks. We provide a controlled evaluation on long-context reasoning. With the backbone held fixed at GPT-5 to match the published Codex and RLM baselines, RAH improves the Codex coding-agent baseline from 71.75% to 81.36% on Oolong-Synthetic (199 samples, 13 context-length buckets up to 4M tokens), a gain attributable to the harness rather than the model. With a stronger backbone, Claude Sonnet 4.5, the same design reaches 89.77%.
Chinese Translation
递归语言模型（RLMs）表明，对模型调用进行递归是一种有效的长上下文推理策略，生产编码代理已经开始编写代码，以大规模生成子代理，最近在Anthropic的动态工作流中表现尤为突出。我们命名并研究这两条工作线之间的模式，其中递归单元是一个完整的代理工具，具备文件系统工具、代码执行和规划，而不是没有工具的模型调用。我们称之为递归代理工具（Recursive Agent Harness，RAH），并将其框架视为工具递归，这是对RLMs模型递归的代码优先扩展。父代理生成并运行一个可执行脚本，以并行生成子代理工具，处理细粒度工作负载，并使用结构化函数调用来完成小任务。我们提供了一个关于长上下文推理的受控评估。在固定基座为GPT-5以匹配已发布的Codex和RLM基线的情况下，RAH将Codex编码代理基线从71.75%提高到81.36%（Oolong-Synthetic，199个样本，13个上下文长度桶，最多4M标记），这一提升归因于工具而非模型。在更强大的基座Claude Sonnet 4.5下，相同设计达到了89.77%。
cs.CL / 62 / 2606.13647
SkMTEB: Slovak Massive Text Embedding Benchmark and Model Adaptation
SkMTEB：斯洛伐克大规模文本嵌入基准与模型适应
Abstract
We introduce SkMTEB, the first comprehensive MTEB-style text embedding benchmark for Slovak, a low-resource West Slavic language, comprising 31 datasets across 7 task types -- nearly 4$\times$ the depth of existing multilingual benchmark coverage for Slovak. Our evaluation of 31 embedding models reveals that large instruction-tuned multilingual models achieve the strongest performance, while existing Slovak-specific models trained for NLU tasks transfer poorly to embedding tasks. To address the need for efficient, locally-deployable Slovak embeddings, we develop \texttt{e5-sk-small} (45M parameters) and \texttt{e5-sk-large} (365M) by applying vocabulary trimming and fine-tuning to Multilingual E5 models. Despite size reductions of up to 62\%, our open-source models achieve competitive performance with proprietary APIs while remaining locally deployable for semantic search and retrieval-augmented generation (RAG). We release the benchmark, models, datasets, and code openly, hoping our approach offers a replicable path for other under-resourced languages.
Chinese Translation
我们介绍了SkMTEB，这是第一个针对斯洛伐克语（低资源的西斯拉夫语言）的综合性MTEB风格文本嵌入基准，涵盖了31个数据集和7种任务类型——几乎是现有斯洛伐克多语言基准覆盖深度的4倍。我们对31个嵌入模型的评估表明，大型指令调优的多语言模型表现最佳，而现有的针对自然语言理解（NLU）任务训练的斯洛伐克特定模型在嵌入任务上的迁移效果较差。为满足对高效、可本地部署的斯洛伐克嵌入的需求，我们通过对多语言E5模型进行词汇修剪和微调，开发了 exttt{e5-sk-small}（4500万参数）和 exttt{e5-sk-large}（3.65亿参数）。尽管模型规模减少了多达62 extperthousand{}，我们的开源模型在性能上与专有API相当，同时仍可本地部署用于语义搜索和增强生成（RAG）。我们公开发布基准、模型、数据集和代码，希望我们的做法为其他资源匮乏的语言提供可复制的路径。
cs.CL / 63 / 2606.13649
Operadic consistency: a label-free signal for compositional reasoning failures in LLMs
算子一致性：一种无标签信号，用于检测大语言模型中的组合推理失败
Abstract
Detecting LLM reasoning failures at inference time without ground-truth labels has motivated a wide range of confidence baselines, including self-consistency, semantic entropy, and P(True), built on within-question sampling and self-evaluation. Operad theory, the formalism for systems built by iterated substitution, suggests a complementary diagnostic: a model's direct answer to a compositional query should agree with the answer it produces by composing a stated decomposition of the same query. We instantiate this idea as operadic consistency (OC), a per-question signal. Across twelve instruction-tuned LLMs (4B to 671B parameters, open-weights and closed-source) on four multi-hop QA datasets, OC is strongly correlated with accuracy on every dataset (Pearson $r \in [0.86, 0.94]$, all $p \leq 0.0004$), and is the only signal we evaluate with $r \geq 0.85$ uniformly across all four datasets. Chain-of-thought self-consistency (CoT-SC; Wang et al., 2023) matches OC on HotpotQA and DROP ($r = 0.93, 0.87$) but drops to $r \approx 0.45$ on MuSiQue and StrategyQA. At the per-question level, OC contributes information beyond CoT-SC and semantic entropy on every dataset (cluster-robust $p \leq 10^{-16}$ for the OC coefficient), and the conclusion is robust to additionally controlling for constructed decomposition-aware baselines ($p \leq 10^{-13}$). The same signal yields selective-prediction improvements (accuracy at fixed coverage) over a tuned CoT-SC baseline at the equal-cost $K = 3$ budget (AUARC lifts of +0.086 to +0.096 and AUROC lifts of +0.092 to +0.164; 95% CIs exclude zero on every cell). On five frontier thinking models, where the decomposition is extracted from the model's own chain of thought, the same equal-cost comparison gives positive selective-prediction point-estimate lift on all 16 (dataset, budget, metric) cells tested, with 95% CIs excluding zero on 12 of the 16.
Chinese Translation
在推理时检测大语言模型（LLM）的推理失败而无需真实标签，促使了多种置信度基线的提出，包括自一致性、语义熵和 P(True)，这些基线基于问题内采样和自我评估。算子理论作为迭代替换构建系统的形式化方法，提出了一种互补的诊断：模型对组合查询的直接回答应与其通过组合同一查询的已声明分解所产生的答案一致。我们将这一思想实例化为算子一致性（Operadic Consistency, OC），作为每个问题的信号。在四个多跳问答数据集上对十二个经过指令调优的 LLM（参数从 4B 到 671B，开放权重和闭源）进行测试，OC 与每个数据集的准确性强相关（Pearson $r ext{ in } [0.86, 0.94]$，所有 $p ext{ } ext{≤} ext{ } 0.0004$），并且是我们评估的唯一一个在所有四个数据集上均具有 $r ext{ } ext{≥} ext{ } 0.85$ 的信号。链式思维自一致性（Chain-of-thought Self-consistency, CoT-SC; Wang et al., 2023）在 HotpotQA 和 DROP 上与 OC 匹配（$r = 0.93, 0.87$），但在 MuSiQue 和 StrategyQA 上降至 $r ext{ } ext{≈} ext{ } 0.45$。在每个问题层面上，OC 在每个数据集上提供了超出 CoT-SC 和语义熵的信息（OC 系数的集群稳健性 $p ext{ } ext{≤} ext{ } 10^{-16}$），并且这一结论在额外控制构造的分解感知基线后依然稳健（$p ext{ } ext{≤} ext{ } 10^{-13}$）。在固定覆盖率下，相同信号在等成本 $K = 3$ 预算下相较于调优的 CoT-SC 基线带来了选择性预测的提升（AUARC 提升 +0.086 至 +0.096，AUROC 提升 +0.092 至 +0.164；95% 置信区间在每个单元中均不包含零）。在五个前沿思维模型中，分解是从模型自身的思维链中提取的，相同的等成本比较在测试的所有 16 个（数据集、预算、指标）单元中均给出了正的选择性预测点估计提升，其中 16 个单元中有 12 个的 95% 置信区间不包含零。
cs.CL / 64 / 2606.13663
HyperTool: Beyond Step-Wise Tool Calls for Tool-Augmented Agents
HyperTool：超越逐步工具调用的工具增强代理
Abstract
Tool-augmented LLM agents commonly rely on step-wise atomic tool calls, where each invocation, observation, and value transfer is exposed in the main reasoning trace. This creates an \emph{execution-granularity mismatch}: locally deterministic tool workflows are unfolded into repeated model-visible decisions, consuming context and forcing the model to manage low-level dataflow in the trace. We introduce \textbf{HyperTool}, a unified executable MCP-style tool interface that changes the model-visible unit of tool execution. A model invokes HyperTool with a code block that can call existing tools through their original schemas, manipulate returned values, and pass intermediate results locally, folding deterministic tool subroutines into a single outer call. To train models to use this interface, we synthesize HyperTool-format trajectories from cross-tool compositional tasks and verify them in real MCP environments. On MCP-Universe, HyperTool improves average accuracy from 15.69\% to 35.29\% on Qwen3-32B and from 9.93\% to 33.33\% on Qwen3-8B, and surpass GPT-OSS and Kimi-k2.5 on average accuracy, showing that our HyperTool can substantially improve multi-step tool use.
Chinese Translation
工具增强的大型语言模型（LLM）代理通常依赖于逐步的原子工具调用，其中每次调用、观察和数值传递都在主要推理轨迹中显现。这导致了一个 extit{执行粒度不匹配}的问题：局部确定性的工具工作流被展开为重复的模型可见决策，消耗上下文并迫使模型在轨迹中管理低级数据流。我们提出了 extbf{HyperTool}，一个统一的可执行MCP风格工具接口，改变了工具执行的模型可见单元。模型通过一个代码块调用HyperTool，该代码块可以通过其原始模式调用现有工具，操纵返回值，并在本地传递中间结果，将确定性的工具子例程折叠为一个外部调用。为了训练模型使用该接口，我们从跨工具组合任务中合成HyperTool格式的轨迹，并在真实的MCP环境中验证它们。在MCP-Universe上，HyperTool将Qwen3-32B的平均准确率从15.69\%提高到35.29\%，将Qwen3-8B的平均准确率从9.93\%提高到33.33\%，并在平均准确率上超越了GPT-OSS和Kimi-k2.5，显示出我们的HyperTool可以显著改善多步骤工具的使用。
cs.CL / 65 / 2606.13668
Influcoder: Distilling Decoders' Gradient Influence Rankings into an Encoder for Data Attribution
Influcoder：将解码器的梯度影响排名提炼到编码器中以实现数据归因
Abstract
With the growth of LLMs' (Large Language Models) capabilities, there has been an increasing push to curate high quality datasets by filtering samples in the training data. In general, Data Attribution (DA) methods aim to estimate how individual samples in a training dataset can precondition a model to generate certain outputs. As an example, one might be interested in which samples in the data could be the source of toxic behavior after training the LLM. Many methods quantify this conditioning through the paradigm of influence functions. While methods of this family are effective in its function, they lack the necessary processing speed and storage compactness to be practically implemented on large datasets. We propose a method, Influcoder, as a quick and cost-effective approach to influence-based Data Attribution at scale.
Chinese Translation
随着大型语言模型（LLMs）能力的增长，越来越多的努力致力于通过过滤训练数据中的样本来策划高质量的数据集。一般来说，数据归因（Data Attribution, DA）方法旨在估计训练数据集中个别样本如何影响模型生成特定输出。例如，人们可能会关注在训练LLM后，哪些样本可能是有毒行为的来源。许多方法通过影响函数的范式量化这种影响。尽管这一系列方法在其功能上是有效的，但它们缺乏必要的处理速度和存储紧凑性，无法在大规模数据集上实际应用。我们提出了一种方法，Influcoder，作为一种快速且具有成本效益的影响基础数据归因方法，适用于大规模数据集。
cs.CL / 66 / 2606.13680
Learning to Reason by Analogy via Retrieval-Augmented Reinforcement Fine-Tuning
通过检索增强的强化微调学习类比推理
Abstract
Retrieval-augmented generation (RAG) has become a standard mechanism for grounding language models in external knowledge, yet conventional retrieval based on lexical or semantic similarity is poorly suited for complex reasoning tasks: a semantically similar problem may demand an entirely different solution strategy, while a superficially different problem may share the same underlying reasoning pattern. We propose Retrieval-Augmented Reinforcement Fine-Tuning (RA-RFT), a post-training framework that teaches language models to reason by analogy. RA-RFT uses gold-relevance distillation to train a retriever that ranks contexts by expected reasoning benefit rather than semantic overlap, and then fine-tunes the policy model via reinforcement fine-tuning methods with retrieved analogous demonstrations, so the model learns to leverage reasoning traces under verifiable outcome rewards. We further analyze the diversity of retrieved contexts and find that reasoning-aware retrieval surfaces complementary solution strategies that provide distinct reasoning scaffolds for individual problems. Across challenging mathematical reasoning benchmarks, RA-RFT consistently outperforms standard reinforcement fine-tuning methods. For example, it improves AIME 2025 average@32 accuracy by 7.1 and 2.8 points over GRPO for Qwen3-1.7B and Qwen3-4B respectively -- suggesting that reasoning-aware retrieval is a complementary axis of improvement and orthogonal to advances in reward design or training curricula.
Chinese Translation
检索增强生成（RAG）已成为将语言模型与外部知识结合的标准机制，但基于词汇或语义相似性的传统检索方法不适合复杂推理任务：一个语义上相似的问题可能需要完全不同的解决策略，而一个表面上不同的问题可能共享相同的基本推理模式。我们提出了检索增强强化微调（RA-RFT），这是一种后训练框架，旨在教会语言模型通过类比进行推理。RA-RFT利用黄金相关性蒸馏训练一个检索器，该检索器根据预期推理收益而非语义重叠对上下文进行排名，然后通过强化微调方法利用检索到的类比示例微调策略模型，使模型能够在可验证的结果奖励下学习利用推理轨迹。我们进一步分析了检索上下文的多样性，发现关注推理的检索能够呈现互补的解决策略，为个别问题提供不同的推理支架。在具有挑战性的数学推理基准测试中，RA-RFT始终优于标准的强化微调方法。例如，它在AIME 2025 average@32的准确率上分别比GRPO提高了7.1和2.8个百分点，适用于Qwen3-1.7B和Qwen3-4B，这表明关注推理的检索是一个互补的改进方向，与奖励设计或训练课程的进展是正交的。
cs.CL / 67 / 2606.13681
EvoArena: Tracking Memory Evolution for Robust LLM Agents in Dynamic Environments
EvoArena：在动态环境中跟踪记忆演变以增强大型语言模型代理的鲁棒性
Abstract
Large language model (LLM) agents have achieved strong performance on a wide range of benchmarks, yet most evaluations assume static environments. In contrast, real-world deployment is inherently dynamic, requiring agents to continually align their knowledge, skills, and behavior with changing environments and updated task conditions. To address this gap, we introduce EvoArena, a benchmark suite that models environment changes as sequences of progressive updates across terminal, software, and social domains. We further propose EvoMem, a patch-based memory paradigm that records memory evolution as structured update histories, enabling agents to reason about environmental evolution through changes in their memory. Experiments show that current agents struggle on EvoArena, achieving an average accuracy of 39.6% across evolving terminal, software, and social-preference domains. EvoMem consistently improves performance, yielding an average gain of 1.5% on EvoArena and also improving standard benchmarks such as GAIA and LoCoMo by 6.1% and 4.8%. Beyond individual tasks, EvoMem further improves chain-level accuracy by 3.7% on EvoArena, where success requires completing a consecutive sequence of related evolutionary subtasks. Mechanistic analysis shows that EvoMem improves evidence capture in the memory, indicating better preservation of complete evolving environment states. Our results highlight the importance of modeling evolution in both evaluation and memory for reliable agent deployment.
Chinese Translation
大型语言模型（LLM）代理在广泛的基准测试中取得了强劲的表现，但大多数评估假设环境是静态的。相比之下，现实世界的部署本质上是动态的，这要求代理不断调整其知识、技能和行为，以适应变化的环境和更新的任务条件。为了解决这一问题，我们引入了EvoArena，一个基准套件，将环境变化建模为终端、软件和社会领域的渐进更新序列。我们进一步提出了EvoMem，一种基于补丁的记忆范式，它将记忆演变记录为结构化的更新历史，使代理能够通过记忆中的变化推理环境的演变。实验表明，当前的代理在EvoArena上表现不佳，在不断变化的终端、软件和社会偏好领域的平均准确率仅为39.6%。EvoMem始终能提高性能，在EvoArena上平均提高1.5%的准确率，同时在GAIA和LoCoMo等标准基准上分别提高6.1%和4.8%。除了单个任务外，EvoMem还在EvoArena上将链级准确率提高了3.7%，成功完成需要连续完成一系列相关演变子任务的任务。机制分析表明，EvoMem改善了记忆中的证据捕获，表明对完整演变环境状态的更好保留。我们的结果强调了在评估和记忆中建模演变的重要性，以确保代理的可靠部署。