← Back to Index
Daily Research Digest

arXiv Papers

2026-06-26
286
Papers
4
Categories
286
Translated
收藏清单 0
机器人学 (Robotics)
54
cs.RO / 1 / 2606.26151

Unsupervised Memory-Enhanced Video Transformers: Obstacle Detection for Autonomous Agricultural Rover

无监督记忆增强视频变换器：用于自主农业探测车的障碍物检测
Biardeau, Théo, Capelle-Laizé, Anne-Sophie, Alwan, Salwan, Helbert, David
Abstract
While autonomous rovers have become indispensable to precision farming, achieving consistent operational safety remains a critical challenge. Conventional safety sensors, such as LiDAR, fail to detect obstacles positioned below the plant canopy, posing a significant risk. While camera-based supervised learning methods can detect common objects, they perform poorly when faced with obstacles that were not present in their training data. Actual unsupervised anomaly detection offers a solution by learning the normal visual patterns of an environment, but often fails for the dynamic scenes captured by a moving rover.\\ This paper introduces Video Memory Transformers for Anomaly Detection (VMTAD), a fully unsupervised method designed for real-time obstacle detection in dynamic agricultural scenes. VMTAD utilizes a transformer-driven architecture augmented with a dedicated memory module. This memory module leverages temporal context by processing encoded representations of preceding frames. This approach enables the system to effectively address the dynamic context caused by the robot's movement. The model is trained using only images that represent normal operation, requiring no data labels.\\ VMTAD was rigorously evaluated on the 'Grillion' agricultural rover. On a challenging rapeseed dataset, VMTAD achieved state-of-the-art performance, reaching a 0.973 detection and 0.997 segmentation Area Under the Receiver Operating Characteristic curve. A lightweight variant provides an optimal balance of high accuracy and real-time inference (14 ms), which is critical for safety, as confirmed by our analysis of the rover's total stopping distance.
Chinese Translation
尽管自主探测车在精准农业中已变得不可或缺，但实现持续的操作安全仍然是一个关键挑战。传统的安全传感器，如激光雷达（LiDAR），无法检测位于植物冠层下方的障碍物，这带来了显著的风险。虽然基于相机的监督学习方法可以检测常见物体，但在面对训练数据中未出现的障碍物时表现不佳。实际的无监督异常检测通过学习环境的正常视觉模式提供了解决方案，但在移动探测车捕捉的动态场景中往往失败。本文介绍了一种用于异常检测的视频记忆变换器（Video Memory Transformers for Anomaly Detection, VMTAD），这是一种完全无监督的方法，旨在实时检测动态农业场景中的障碍物。VMTAD利用一种由变换器驱动的架构，并配备专用的记忆模块。该记忆模块通过处理前一帧的编码表示来利用时间上下文。这种方法使系统能够有效应对由机器人运动引起的动态上下文。该模型仅使用代表正常操作的图像进行训练，无需数据标签。VMTAD在“Grillion”农业探测车上进行了严格评估。在一个具有挑战性的油菜数据集上，VMTAD达到了最先进的性能，检测和分割的接收者操作特征曲线（ROC曲线）下面积分别达到了0.973和0.997。一个轻量级变体提供了高准确性和实时推理（14毫秒）之间的最佳平衡，这对于安全至关重要，正如我们对探测车总制动距离的分析所确认的那样。
View on arXiv Download PDF AI Translation
cs.RO / 2 / 2606.26154

Reinforcement Learning Enables Autonomous Microrobot Navigation and Intervention in Simulated Blood Capillaries

强化学习实现自主微型机器人在模拟血管中的导航与干预
Drotleff, Jannik, Tovey, Samuel, Hohenberger, Paul, Lohrmann, Christoph, Hoßbach, Julian, Nikolaou, Konstantin, Holm, Christian
Abstract
Autonomous microrobots navigating biological vasculature could enable targeted drug delivery and thrombolysis, yet training control policies for realistic environments remains an open challenge. Prior reinforcement learning (RL) studies of microrobotic navigation have been limited to idealized geometries that omit complex hydrodynamic flow fields, confined branching structures, and dense cellular obstacles found in vivo. Here, we develop a physically grounded simulation of a blood capillary network, incorporating realistic hydrodynamic flow fields, explicit red blood cell dynamics, and anatomically derived branching geometry, and train deep RL agents to navigate it via chemotaxis. We systematically map the physical limits of navigation across robot size and swimming speed, revealing a forbidden regime where Brownian motion and flow overcome propulsion. Successful agents independently discover multiple universal strategy types, including run-and-rotate and energy-efficient search-and-sit policies, regardless of robot parameters. Without retraining, these agents perform targeted blocking and unblocking of capillary flow, restoring throughput to healthy baseline levels. These results establish RL as a viable framework for developing autonomous microrobotic intervention strategies in complex biological environments.
Chinese Translation
自主微型机器人在生物血管中导航可以实现靶向药物输送和溶栓，但在现实环境中训练控制策略仍然是一个未解决的挑战。先前的强化学习（RL）研究主要集中在理想化的几何结构上，忽略了体内复杂的流体动力学流场、受限的分支结构和密集的细胞障碍。在这里，我们开发了一个基于物理的血管网络模拟，结合了现实的流体动力学流场、明确的红细胞动力学和解剖学派生的分支几何结构，并训练深度RL代理通过趋化性进行导航。我们系统地映射了机器人尺寸和游泳速度对导航的物理限制，揭示了一个禁区，在该区域布朗运动和流动超过了推进力。成功的代理独立发现了多种通用策略类型，包括跑动-旋转和节能搜索-静坐策略，无论机器人参数如何。在不重新训练的情况下，这些代理能够独立进行靶向阻塞和解阻血管流动，将流量恢复到健康的基线水平。这些结果确立了强化学习作为在复杂生物环境中开发自主微型机器人干预策略的可行框架。
View on arXiv Download PDF AI Translation
cs.RO / 3 / 2606.26175

RMTL: Reinforced Micro-task Learning for Long-Horizon Manipulation with VLM Rewards

RMTL：基于强化微任务学习的长时间范围操作与VLM奖励
Ateş, Anıl Can, Kahraman, Orhan, Topal, Cihan
Abstract
Reinforcement learning (RL) for robotic manipulation often requires manually designing a dense reward function, which is difficult to tune and often fragile, or learning a reward from human demonstrations or preferences, which can be expensive. A recent line of work uses pretrained vision-language models (VLMs) as zero-shot reward models, replacing these costs with a single text prompt. However, we argue that a single global prompt is too coarse for long-horizon manipulation tasks with randomized initial conditions. The single-prompt VLM reward is near-flat for much of the trajectory, making early progress hard for the agent to detect. We propose Reinforced Micro-Task Learning (RMTL), an approach that decomposes a manipulation task into a small set of language-described micro-tasks and trains the agent to switch between them. At each step, the agent receives a multi-view VLM reward computed using the prompt of the currently active micro-task and averaged across multiple camera views to reduce the effect of view-specific occlusions. A reverse curriculum gradually exposes the agent to harder initial conditions, while a PPO worker is first trained with a fixed distance-based rule that selects the active micro-task. We then replace this rule with a learned hierarchical manager, turning rule-based phase selection into a fully learned hierarchical policy. We instantiate RMTL on the Fetch manipulation environment using three short stage-specific prompts and without additional prompt tuning. Experiments show that RMTL provides more informative reward signals than single-prompt VLM rewards, enabling faster learning. These results suggest that decomposing VLM rewards into micro-task-specific language prompts can substantially improve the scalability of language-guided reinforcement learning for robotic manipulation.
Chinese Translation
机器人操作的强化学习（RL）通常需要手动设计密集的奖励函数，这种函数难以调整且往往脆弱，或者从人类示范或偏好中学习奖励，这可能代价高昂。最近的研究工作使用预训练的视觉-语言模型（VLM）作为零-shot奖励模型，通过单一文本提示替代这些成本。然而，我们认为单一的全局提示对于具有随机初始条件的长时间范围操作任务来说过于粗糙。单一提示的VLM奖励在轨迹的大部分时间内几乎是平坦的，使得代理难以检测到早期进展。我们提出了强化微任务学习（RMTL），该方法将操作任务分解为一小组用语言描述的微任务，并训练代理在它们之间切换。在每一步，代理接收一个多视角的VLM奖励，该奖励是使用当前活动微任务的提示计算的，并在多个摄像头视角之间取平均，以减少视角特定遮挡的影响。一个反向课程逐渐使代理接触到更困难的初始条件，而PPO工作者首先使用基于固定距离的规则进行训练，该规则选择活动微任务。然后，我们将此规则替换为学习的层次管理器，将基于规则的阶段选择转变为完全学习的层次策略。我们在Fetch操作环境中实例化RMTL，使用三个短期特定提示，并且不进行额外的提示调整。实验表明，RMTL提供比单一提示VLM奖励更具信息性的奖励信号，从而实现更快的学习。这些结果表明，将VLM奖励分解为微任务特定的语言提示可以显著提高语言引导的强化学习在机器人操作中的可扩展性。
View on arXiv Download PDF AI Translation
cs.RO / 4 / 2606.26183

LiMoDE: Rethinking Lifelong Robot Manipulation from a Mixture-of-Dynamic-Experts Perspective

LiMoDE：从动态专家混合视角重新思考终身机器人操控
Gu, Zhihao, Wang, Lin
Abstract
Building a generalist robot that can leverage prior knowledge for continuous task adaptation remains a significant challenge. Previous works alleviate the catastrophic forgetting problem by parameter-efficient fine-tuning for single-task adaptation. However, they fail to extract reusable skills and model the interaction with other skills effectively. Recent works try to address these issues by learning prompts. Differently, this paper presents an architectural perspective on the Lifelong Mixture of Dynamic Experts (\textit{LiMoDE}), a novel two-stage learning scheme for lifelong robot manipulation. Specifically, a dynamic MoE structure is first proposed in the multi-task pre-training stage to learn prior knowledge, where a varied number of heterogeneous experts are activated based on the motion information to address different short-term manipulations. Subsequently, in the task adaptation stage, we design a lifelong MoE adaptation mechanism % (LiMoEAM) that learns lifelong experts and dynamically combines them with frozen ones for new tasks, facilitating the knowledge transfer during adaptation. The proposed \textit{LiMoDE} is evaluated on both the simulated lifelong learning benchmark and real-world tasks. Extensive experiments demonstrate its effectiveness in achieving superior performance and strong lifelong adaptation by introducing a moderate number of additional trainable parameters and inference overhead.
Chinese Translation
构建一个能够利用先前知识进行持续任务适应的通用机器人仍然是一个重大挑战。以往的研究通过针对单一任务适应的参数高效微调来缓解灾难性遗忘问题。然而，它们未能有效提取可重用技能并建模与其他技能的互动。近期的研究尝试通过学习提示来解决这些问题。与此不同，本文从架构的角度提出了终身动态专家混合（Lifelong Mixture of Dynamic Experts， extit{LiMoDE}）的概念，这是一种用于终身机器人操控的新型两阶段学习方案。具体而言，在多任务预训练阶段，首先提出了一种动态的专家混合（MoE）结构，以学习先前知识，其中根据运动信息激活不同数量的异构专家，以应对不同的短期操控。随后，在任务适应阶段，我们设计了一种终身MoE适应机制（LiMoEAM），该机制学习终身专家并将其与冻结的专家动态结合，以应对新任务，从而促进适应过程中的知识转移。所提出的 extit{LiMoDE}在模拟的终身学习基准和实际任务上进行了评估。大量实验表明，通过引入适量的额外可训练参数和推理开销，该方法在实现卓越性能和强大终身适应能力方面的有效性。
View on arXiv Download PDF AI Translation
cs.RO / 5 / 2606.26188

Morphology-Specific Closed-Loop Control of Logarithmic-Spiral Continuum Arms via Online Jacobian Error Compensation

基于形态特定的对数螺旋连续臂的闭环控制通过在线雅可比误差补偿
Datta, Partha, Jin, Yi, Lin, Wei, Cao, C. Chase
Abstract
Logarithmic spirals are ubiquitous in biological appendages and provide an attractive morphology for continuum manipulators capable of reaching, wrapping, and grasping. Recently reported logarithmic-spiral robots demonstrated scalable fabrication and versatile grasping but lacked inverse kinematics and closed-loop control. This work presents the first morphology-specific closed-loop task-space control framework for logarithmic-spiral continuum arms. A segmented tendon-driven model with a centerline backbone and equilateral tendon routing is developed in MuJoCo to capture tapered compliance and contact dynamics. An analytical task-space Jacobian is derived directly from the logarithmic-spiral kinematics and combined with online Jacobian error compensation using a Broyden secant update and Kalman-filter estimation. The resulting controller continuously corrects modeling errors arising from nonlinear deformation, contact, and geometric mismatch. The framework is validated through planar and spatial simulations, including trajectory tracking, attitude regulation, disturbance rejection, three-dimensional position tracking, and simultaneous position-orientation control. Compared with a piecewise-constant-curvature (PCC) baseline, the proposed method consistently reduces tracking errors, suppresses attitude drift, and maintains a bounded Jacobian estimation error. The controller is further applied to morphology-enabled manipulation tasks, including obstacle-assisted reach-wrap-release motions, adaptive whole-arm grasping, and cooperative multi-arm object handling. Results demonstrate that combining logarithmic-spiral morphology with online Jacobian compensation enables accurate, robust, and scalable control of highly underactuated continuum manipulators. The proposed framework establishes a physics-grounded baseline for future hardware implementation and learning-augmented soft robotic control.
Chinese Translation
对数螺旋在生物附肢中普遍存在，为能够进行伸展、缠绕和抓取的连续操控器提供了吸引人的形态。最近报道的对数螺旋机器人展示了可扩展的制造能力和多功能的抓取能力，但缺乏逆向运动学和闭环控制。本文提出了首个针对对数螺旋连续臂的形态特定闭环任务空间控制框架。我们在MuJoCo中开发了一个分段的腱驱动模型，具有中心线骨架和等边腱布置，以捕捉锥形顺应性和接触动态。直接从对数螺旋运动学推导出分析性任务空间雅可比，并结合使用Broyden割线更新和卡尔曼滤波估计的在线雅可比误差补偿。所得到的控制器持续修正由非线性变形、接触和几何不匹配引起的建模误差。通过平面和空间仿真验证了该框架，包括轨迹跟踪、姿态调节、干扰抑制、三维位置跟踪和位置-姿态同时控制。与分段常数曲率（PCC）基线相比，所提方法始终减少跟踪误差，抑制姿态漂移，并保持有限的雅可比估计误差。该控制器进一步应用于形态驱动的操控任务，包括障碍辅助的伸展-缠绕-释放动作、自适应全臂抓取和协作多臂物体处理。结果表明，将对数螺旋形态与在线雅可比补偿相结合，能够实现对高度欠驱动的连续操控器的准确、稳健和可扩展的控制。所提出的框架为未来硬件实现和学习增强的软机器人控制建立了基于物理的基线。
View on arXiv Download PDF AI Translation
cs.RO / 6 / 2606.26201

OmniContact: Chaining Meta-Skills via Contact Flow for Generalizable Humanoid Loco-Manipulation

OmniContact：通过接触流链式连接元技能以实现可推广的人形运动操控
Yu, Runyi, Lin, Xiaoyi, Ma, Ji, Wang, Yinhuai, Luo, Koukou, Ji, Jiahao, Wang, Huayi, Wang, Wenjia, Zhang, Runhan, Tan, Ping, Wu, Ting, Dai, Ruoli, Chen, Qifeng, Han, Lei
Abstract
Learning long-horizon humanoid loco-manipulation poses a dual challenge: it requires not only the robust execution of meta-skills but also their seamless, closed-loop chaining equipped with autonomous recovery. Existing approaches remain limited: explicit humanoid-object interaction representations offer precision but are notoriously difficult for high-level planning, whereas implicit skill embeddings are compact but lack the interpretability required for reliable composition. We propose \ours, a hierarchical framework centered on \textbf{contact flow (CF)}, a compact representation consisting of key body trajectories and time-series binary contact signals. Leveraging this shared interface, our low-level policy \textbf{CF-Track} learns a unified library of loco-manipulation skills, while our high-level module \textbf{CF-Gen} heuristically synthesizes future contact-flow sequences. To support this setting, we additionally collect the OmniContact dataset, a MoCap-based HOI corpus for humanoid loco-manipulation (Appendix~\ref{sec:dataset}). Together, they enable robust execution, autonomous failure recovery, and flexible composition of meta-skills for long-horizon tasks. Experiments show that OmniContact achieves \(98.7\%\) success on \textit{Carry Box} and \(76.5\%\) on \textit{Push-Stack Boxes}, outperforming prior baselines by average margins of \(40.9\%\) in meta-skill and \(66.5\%\) in skill chaining. Besides, our framework naturally integrates with VLMs for semantic task decomposition, enabling complex, semantically grounded loco-manipulation behaviors, such as arranging scattered boxes into a heart shape.
Chinese Translation
学习长时间跨度的人形运动操控面临双重挑战：不仅需要稳健地执行元技能，还需要具备自主恢复能力的无缝闭环链式连接。现有方法仍然存在局限性：显式的人形-物体交互表示提供了精确性，但在高层次规划中 notoriously 难以处理，而隐式技能嵌入虽然紧凑，但缺乏可靠组合所需的可解释性。我们提出了 extbf{OmniContact}，一个以 extbf{接触流 (CF)} 为核心的分层框架，接触流是一种由关键身体轨迹和时间序列二进制接触信号组成的紧凑表示。利用这一共享接口，我们的低层策略 extbf{CF-Track} 学习了一个统一的人形运动操控技能库，而我们的高层模块 extbf{CF-Gen} 则启发式地合成未来的接触流序列。为了支持这一设置，我们还收集了 OmniContact 数据集，这是一个基于运动捕捉的关于人形运动操控的 HOI 语料库（附录~ ef{sec:dataset}）。两者结合使得在长时间跨度任务中实现稳健执行、自主故障恢复和灵活组合元技能成为可能。实验表明，OmniContact 在 extit{Carry Box} 上的成功率达到 98.7\%，在 extit{Push-Stack Boxes} 上的成功率为 76.5\\%，在元技能和技能链式连接方面分别超越了先前基线平均 40.9\\% 和 66.5\\%。此外，我们的框架自然与 VLMs 集成，以实现语义任务分解，从而支持复杂的、语义基础的人形运动操控行为，例如将散乱的盒子排列成心形。
View on arXiv Download PDF AI Translation
cs.RO / 7 / 2606.26213

RoboTales: ROBOTic Anthropomorphic LEarning Systems

RoboTales：机器人类人学习系统
Chen, Andrew, Chen, Ju-Hung, Pinyomit, Phurinat, Block, Alexis E.
Abstract
RoboTales is a low-cost robotic storytelling system that animates narratives using expressive sock puppetry. Implemented autonomously on a Baxter robot as a test case, RoboTales synchronizes narration, gestures, and mouth movements to perform character-driven stories. In a pilot study, puppet-based storytelling outperformed a gesture-only mode, producing higher HRIES ratings and improved story recall, suggesting that embodied puppetry enhances engagement and narrative comprehension. Designed to be modular and platform-agnostic, RoboTales can be adapted to other manipulators and offers a screen-free alternative to passive media, supporting future deployment in child-centered learning environments.
Chinese Translation
RoboTales 是一个低成本的机器人讲故事系统，通过富有表现力的袜子木偶来生动演绎叙事。作为一个测试案例，RoboTales 在 Baxter 机器人上自主实现，能够同步叙述、手势和口部动作，以演绎以角色为驱动的故事。在一项初步研究中，基于木偶的讲故事方式优于仅使用手势的模式，产生了更高的 HRIES 评分和更好的故事回忆，表明具身木偶表演增强了参与感和叙事理解。RoboTales 设计为模块化和平台无关，可以适应其他操控器，并为被动媒体提供了一种无屏幕的替代方案，支持未来在以儿童为中心的学习环境中的部署。
View on arXiv Download PDF AI Translation
cs.RO / 8 / 2606.26215

TaskNPoint: How to Teach Your Humanoid to Hit a Backhand in Minutes

TaskNPoint：如何在几分钟内教会你的类人机器人打反手
Werner, Blake, Demler, Ilona, Perona, Pietro, Ames, Aaron D.
Abstract
How do we learn to hit a tennis backhand? Not from a thousand hours of tennis tournaments on TV - we work with a coach and practice. We argue this is also the right recipe for teaching dynamic skills to humanoid robots. This follows from a structural property of dynamic skills: the outcome is decided by a short, crucial portion of the trajectory - for a backhand, the ~20cm of racket travel around ball contact. Getting this interaction window right requires coordinating the whole motion, so that control, physics, and morphology act in concert. Learning thus reduces to mastering a handful of distinct actions and, for each, practicing until the window comes out right. To this end, we introduce TaskNPoint, a training protocol which makes the coach-learner division of labor explicit. The human coach contributes four inputs: a discrete set of skills (e.g. different shots), one demonstration per skill, identification of the interaction window, and the goal. Learning in a physically realistic simulation environment fills in each action trajectory and provides robustness to unmodeled events. Crucially, randomized target sampling during training lets a single demonstration generalize zero-shot to unseen goal locations. We test this approach on a Unitree G1 humanoid that hits forehands and backhands against balls thrown by a human, kicks incoming soccer balls, and picks and places boxes from novel locations. We find that learning is successful from short human video demonstrations and under an hour of training on a single GPU, with no per-task reward tuning.
Chinese Translation
我们是如何学习打网球反手的？并不是通过观看电视上数千小时的网球比赛，而是通过与教练合作和练习。我们认为这也是教导类人机器人动态技能的正确方法。这源于动态技能的一个结构特性：结果由轨迹中一个短暂而关键的部分决定——对于反手击球来说，就是球拍在击球接触时大约20厘米的运动。正确把握这个交互窗口需要协调整个动作，以使控制、物理和形态共同发挥作用。因此，学习归结为掌握少数几个不同的动作，并对每个动作进行练习，直到交互窗口达到理想状态。为此，我们引入了TaskNPoint，一种明确教练与学习者分工的训练协议。人类教练提供四个输入：一组离散的技能（例如，不同的击球方式）、每项技能的一个示范、交互窗口的识别以及目标。通过在物理真实的仿真环境中学习，填补每个动作的轨迹，并为未建模事件提供鲁棒性。关键是，在训练过程中随机化目标采样使得单个示范能够零-shot推广到未见过的目标位置。我们在Unitree G1类人机器人上测试了这种方法，该机器人能够对人类投掷的球进行正手和反手击球，踢来球的足球，并从新位置拾取和放置箱子。我们发现，从短暂的人类视频示范中学习是成功的，并且在单个GPU上进行不到一小时的训练，无需对每个任务进行奖励调优。
View on arXiv Download PDF AI Translation
cs.RO / 9 / 2606.26265

NavIsaacLab: Generating Realistic Crowd via Parallel Robot Learning for Benchmarking Human-aware Navigation

NavIsaacLab：通过并行机器人学习生成真实感人群以进行人类感知导航的基准测试
Xia, Bingyi, Bao, Han, Zhu, Jingyu, Ye, Hanjing, Pang, Yuhan, Chen, Guangcheng, Lin, Liang, Xu, Wenjun, Wang, Jiankun
Abstract
Robot autonomous navigation that accounts for surrounding human activities is crucial for ensuring both safety and natural human-robot interaction in real-world environments shared by humans and robots. Simulation of complex and diverse navigation scenarios serves as the foundation for training reliable robot navigation policies and accurately evaluating the performance of algorithms, offering an efficient alternative to manual supervision of real data. However, current human-aware navigation research faces significant challenges due to the scarcity of diverse, high-quality scene data. Existing simulation platforms often rely on handcrafted rules to approximate pedestrian behavior and lack the capability to provide extensive sensor signals, typically assuming perfect observations. To address these limitations, this paper presents NavIsaacLab, a comprehensive framework for benchmarking and training human-aware navigation policies through physics-based and photo-realistic simulations of pedestrians and scenes. Based on Isaac Lab, the proposed framework employs photo-realistic scene rendering capabilities and supports parallel simulation on GPU, delivering real-time and accurate 3D visual feedback to robots. To enhance the realism of human behavior, a data-driven approach is employed that incorporates a trajectory diffusion model and an adversarial motion learning controller, enabling controllable, physics-based pedestrian simulation. Furthermore, the integration of diverse cross-scale scenes provides a robust benchmark for state-of-the-art human-aware navigation methods.
Chinese Translation
考虑周围人类活动的机器人自主导航对于确保在人与机器人共享的真实环境中实现安全和自然的人机交互至关重要。复杂多样的导航场景模拟是训练可靠的机器人导航策略和准确评估算法性能的基础，为手动监督真实数据提供了高效的替代方案。然而，目前的人类感知导航研究面临着由于缺乏多样化、高质量场景数据而带来的重大挑战。现有的模拟平台通常依赖手工规则来近似行人行为，并缺乏提供广泛传感器信号的能力，通常假设观察是完美的。为了解决这些局限性，本文提出了NavIsaacLab，一个通过基于物理和照片真实感的行人及场景模拟来进行人类感知导航策略基准测试和训练的综合框架。该框架基于Isaac Lab，利用照片真实感场景渲染能力并支持在GPU上进行并行模拟，为机器人提供实时和准确的3D视觉反馈。为了增强人类行为的真实感，采用了一种数据驱动的方法，结合了轨迹扩散模型和对抗运动学习控制器，实现可控的基于物理的行人模拟。此外，多样化的跨尺度场景的整合为最先进的人类感知导航方法提供了稳健的基准。
View on arXiv Download PDF AI Translation
cs.RO / 10 / 2606.26313

Racing a Wheeled Quadruped: Active Load Transfer Mitigation via Model Predictive Control

赛车轮式四足机器人：通过模型预测控制实现主动负载转移缓解
Eisman, Marla, Lam, Brian, Sonnino, Samuel, Borrelli, Francesco
Abstract
This paper presents a hierarchical control framework using model predictive control (MPC) and reinforcement learning (RL) for active roll control to manage lateral load transfer during autonomous racing of a wheeled quadruped. The framework integrates offline time-optimal raceline generation, an online MPC planner that actively minimizes the lateral Load Transfer Ratio (LTR), and a low-level, whole-body RL policy deployed directly onto the robot's 16 actuators. The MPC is based on a vehicle dynamics bicycle model of the Unitree Go2-W platform. The robot's leg actuators act as active suspension where knee joints generate anti-roll torque to bank into turns. Physical track experiments demonstrate that active roll control reduces mean LTR by up to 44%, improves the fastest lap time by 8.7%, and boosts peak lateral acceleration capability by 21.3% to 1.98 $m/s^2$, maintaining robust high-speed stability beyond the range of a non-tilting baseline controller. Supplementary code and video can be found at https://github.com/meisman-ucb/go2w-roll-control-mpc
Chinese Translation
本文提出了一种分层控制框架，利用模型预测控制（MPC）和强化学习（RL）进行主动侧倾控制，以管理轮式四足机器人在自主赛车过程中的侧向负载转移。该框架集成了离线时间最优赛道生成、一个在线MPC规划器，该规划器主动最小化侧向负载转移比（LTR），以及直接部署在机器人16个执行器上的低层次全身RL策略。MPC基于Unitree Go2-W平台的车辆动力学自行车模型。机器人的腿部执行器作为主动悬挂，膝关节产生反侧倾扭矩以便于转弯。物理轨道实验表明，主动侧倾控制将平均LTR降低了高达44%，最快圈速提高了8.7%，并将峰值侧向加速度能力提升了21.3%至1.98 $m/s^2$，在超出非倾斜基线控制器范围的情况下保持了强大的高速稳定性。补充代码和视频可在 https://github.com/meisman-ucb/go2w-roll-control-mpc 找到。
View on arXiv Download PDF AI Translation
cs.RO / 11 / 2606.26315

Layered Outer-Loop Control for Disturbance-Robust Multi-Waypoint UAV Arrival

分层外环控制用于抗干扰多航点无人机到达
Ling, Runfeng
Abstract
Disturbance-robust UAV position control is easy to demonstrate in benign simulations but much harder to make fast in approach, well behaved near the target, and credible beyond a single benchmark. This letter presents a layered terminal-control architecture for multi-waypoint UAV position regulation together with a staged evaluation across PyBullet, PX4/Gazebo, and hardware. Phase I uses a PyBullet benchmark with stochastic wind for rapid structural selection, identifying a controller core that separates smooth approach generation, persistent-bias compensation, and supervised near-target terminal regulation. Phase II carries only that main architecture into a more demanding PX4/Gazebo closed loop, where the outer-loop controller acts through a cascaded flight stack with delay-sensitive settling and stronger transit-to-hover coupling. This step exposes which benchmark gains survive autopilot-mediated dynamics and which refinements collapse once the loop becomes more deployment-like. In Phase I, the bare controller attains 0.024 m mean late-stage wind error. In Phase II, the final controller is selected using a transfer-oriented rule emphasizing absence of benchmark priors, cross-scenario balance, and deployable supervisory logic. Strict is used as the primary reporting reference; the supplementary retrospective Grace analysis shows that part of the residual failure set is sensitive to completion semantics rather than gross waypoint-miss behaviour. The evaluation is completed on one Vicon-tracked Tello stack through a two-level hardware study. Taken together, the results suggest that benchmark success becomes more informative when the main controller design is separated from benchmark-specific refinement and remains defensible under harder closed-loop evaluation.
Chinese Translation
抗干扰无人机位置控制在良好的仿真环境中易于展示，但在快速接近、靠近目标时表现良好以及在多个基准测试中保持可信性则要困难得多。本文提出了一种分层终端控制架构，用于多航点无人机位置调节，并在PyBullet、PX4/Gazebo和硬件上进行了分阶段评估。第一阶段使用带有随机风的PyBullet基准进行快速结构选择，识别出一个控制器核心，该核心将平滑接近生成、持续偏差补偿和监督的近目标终端调节分开。第二阶段将该主要架构引入更具挑战性的PX4/Gazebo闭环中，外环控制器通过具有延迟敏感的稳定性和更强的过渡到悬停耦合的级联飞行堆栈进行操作。这一步骤揭示了哪些基准增益在自动驾驶仪介导的动态中存活，以及哪些改进在闭环变得更接近实际部署时崩溃。在第一阶段，裸控制器达到了0.024米的平均晚期风误差。在第二阶段，最终控制器的选择使用了一种以缺乏基准先验、跨场景平衡和可部署监督逻辑为重点的转移导向规则。严格模式被用作主要报告参考；补充的回顾性Grace分析显示，部分残余失败集对完成语义敏感，而不是对粗略的航点遗漏行为敏感。评估在一个Vicon跟踪的Tello堆栈上通过两级硬件研究完成。综合来看，结果表明，当主要控制器设计与特定基准的细化分开，并在更严格的闭环评估下仍然具有防御性时，基准成功变得更加信息丰富。
View on arXiv Download PDF AI Translation
cs.RO / 12 / 2606.26321

KRVF: A Source-Aware Semantic Voxel World Representation for Edge Mobile Manipulation

KRVF：一种面向源的边缘移动操作的语义体素世界表示
Ling, Runfeng
Abstract
Mobile manipulators need world models that are current, queryable, semantically meaningful, and usable under edge-compute constraints. This technical report presents KRVF, a source-aware semantic voxel world representation for edge mobile manipulation. Unlike reconstruction-centric mapping pipelines that primarily optimize global geometric fidelity, KRVF represents local world state as task-oriented voxels that encode occupancy, color, semantic evidence, temporal freshness, and evidence source. The representation separates measured occupancy from semantic-prior hypotheses, enabling depth-failure-aware object reasoning without silently corrupting persistent geometry. KRVF also closes a feedback loop between mapping and sensing by rendering map-prior depth for repair, and exposes task-level query operators for semantic objects and grasp candidates. The report formalizes the KRVF representation and documents a ROS 2 implementation that turns online RGB-D observations into a task-facing robot memory.
Chinese Translation
移动操作器需要当前的、可查询的、语义上有意义的世界模型，并且能够在边缘计算约束下使用。本技术报告介绍了KRVF，这是一种面向源的语义体素世界表示，旨在用于边缘移动操作。与主要优化全局几何保真度的重建中心映射管道不同，KRVF将局部世界状态表示为任务导向的体素，这些体素编码了占用信息、颜色、语义证据、时间新鲜度和证据来源。该表示将测量的占用信息与语义先验假设分离，使得在不悄然破坏持久几何的情况下能够进行深度失败感知的物体推理。KRVF还通过渲染地图先验深度进行修复，关闭了映射与感知之间的反馈回路，并为语义对象和抓取候选者提供了任务级查询操作符。该报告形式化了KRVF表示，并记录了一个ROS 2实现，该实现将在线RGB-D观测转化为面向任务的机器人记忆。
View on arXiv Download PDF AI Translation
cs.RO / 13 / 2606.26341

Scaling Nonlinear Optimization: Many Problems One GPU

扩展非线性优化：一个GPU解决多个问题
Viljoen, John, Haffner, Johanna, Tomizuka, Masayoshi, Mehr, Negar
Abstract
Many robotics problems, including trajectory optimization, inverse kinematics, and contact-rich motion planning, reduce to nonlinear programs (NLPs). Mature NLP solvers such as IPOPT can solve these problems, offering hard constraint satisfaction, optimality guarantees, and favorable scaling with problem dimension. These solvers underpin gradient-based methods in robotics, yet remain CPU-bound and solve only one problem at a time, preventing their integration into GPU-batched learning pipelines. On the other hand, sampling-based approaches such as reinforcement learning, model predictive path integral, and imitation learning have become the core of modern robotics research due to their ability to leverage GPU-batched simulators. These simulators can generate orders of magnitude more dynamics rollouts per second than was previously possible. If a GPU-batched NLP solver existed, it would unlock similar speedups in the number of constrained, locally optimal solutions generated per second. This regime of solving many problems concurrently versus solving a single problem at a time is a key requirement for integrating NLP solvers in modern GPU-batched robotics frameworks. To this end, we introduce \texttt{jaxipm}, the first GPU-batched NLP solver, based on IPOPT, and implemented in JAX. We accomplish this by redesigning IPOPT's algorithm to eliminate control flow with \textit{heterogeneous iteration fusion}, and by minimizing GPU idle time with \textit{iteration level batching}. We evaluate \texttt{jaxipm} on a variety of quadrotor nonlinear model predictive control benchmarks, including reference tracking in the presence of obstacles, multi-quadrotor navigation without collision, and navigation in a cluttered environment. We demonstrate up to a $32.85\times$ increase in throughput over IPOPT. Our complete open-source codebase is available at https://github.com/johnviljoen/jaxipm.
Chinese Translation
许多机器人问题，包括轨迹优化、逆运动学和富接触运动规划，都可以简化为非线性规划（NLP）。成熟的NLP求解器如IPOPT能够解决这些问题，提供严格的约束满足、最优性保证，以及与问题维度的良好扩展性。这些求解器支撑着机器人中的基于梯度的方法，但仍然受限于CPU，并且一次只能解决一个问题，这阻碍了它们在GPU批处理学习管道中的集成。另一方面，基于采样的方法，如强化学习、模型预测路径积分和模仿学习，因其能够利用GPU批处理模拟器而成为现代机器人研究的核心。这些模拟器每秒可以生成数量级更多的动态展开，远超以往的能力。如果存在一个GPU批处理的NLP求解器，将能够在每秒生成的约束局部最优解数量上实现类似的加速。并行解决多个问题与一次解决单个问题的这种模式是将NLP求解器集成到现代GPU批处理机器人框架中的关键要求。为此，我们介绍了 exttt{jaxipm}，第一个基于IPOPT的GPU批处理NLP求解器，并在JAX中实现。我们通过重新设计IPOPT的算法，消除控制流，采用 extit{异构迭代融合}，并通过 extit{迭代级别批处理}来最小化GPU空闲时间，从而实现这一目标。我们在多种四旋翼非线性模型预测控制基准上评估 exttt{jaxipm}，包括在障碍物存在下的参考跟踪、多四旋翼无碰撞导航以及在杂乱环境中的导航。我们展示了相较于IPOPT高达$32.85 imes$的吞吐量提升。我们的完整开源代码库可在https://github.com/johnviljoen/jaxipm获取。
View on arXiv Download PDF AI Translation
cs.RO / 14 / 2606.26392

MPC-Injection: Biasing Off-Policy Locomotion RL Toward Controller-Induced Behavior Basins

MPC注入：将离策略运动强化学习偏向于控制器诱导的行为盆地
Xing, Roy, Ree, Seyoung, Plancher, Brian
Abstract
Reinforcement learning (RL) for locomotion frequently converges to locally optimal but undeployable behaviors, such as vibrating limbs or scooting on the torso, that maximize return without producing a usable gait. We present MPC-Injection, a low-overhead method that steers RL toward a designer-preferred gait by inserting transitions into the replay buffer from a model predictive controller solving the same Markov decision process. Unlike reward shaping, MPC-Injection does not require redesigning the task reward, and unlike adversarial imitation learning, it adds no discriminator, no kinematic retargeting, and no auxiliary objective. Instead, the controller's preferred behavior is transferred to the policy purely through the replay state distribution. On a 2D walker in simulation and with sim-to-real evaluation on a Go2 quadruped, we show that MPC-Injection drives the policy into the controller's behavior basin using a one to two-term task reward, producing gaits qualitatively comparable to those of reward shaping with twenty-one tuned terms and of adversarial motion priors without their discriminator and retargeting overhead. We further analyze how the injected transitions bias actor-critic updates toward controller-visited states, allowing the policy to learn behaviors that pure RL may fail to reach under simple reward functions.
Chinese Translation
运动强化学习（RL）常常收敛于局部最优但不可部署的行为，例如震动的肢体或在躯干上滑动，这些行为虽然最大化了回报，但并未产生可用的步态。我们提出了MPC注入，这是一种低开销的方法，通过从解决相同马尔可夫决策过程的模型预测控制器中插入过渡到重放缓冲区，来引导RL朝向设计者偏好的步态。与奖励塑形不同，MPC注入不需要重新设计任务奖励；与对抗模仿学习不同，它不添加判别器、不进行运动学重定向，也不引入辅助目标。相反，控制器的偏好行为纯粹通过重放状态分布转移到策略上。在模拟中的二维行走者和在Go2四足机器人上的仿真到现实评估中，我们展示了MPC注入使用一到两个项的任务奖励将策略驱动到控制器的行为盆地，产生的步态在质量上可与使用二十一项调优项的奖励塑形和没有判别器及重定向开销的对抗运动先验相媲美。我们进一步分析了注入的过渡如何使演员-评论家更新偏向于控制器访问的状态，从而使策略能够学习到纯RL在简单奖励函数下可能无法达到的行为。
View on arXiv Download PDF AI Translation
cs.RO / 15 / 2606.26408

Exploring the Intrinsic Geometry of Diffusion Models with Constrained Inverse Kinematics

探索具有约束逆运动学的扩散模型的内在几何
Garcia, Miguel Angel Rogel, Kyaw, Phone Thiha, Kelly, Jonathan
Abstract
Recent studies suggest that diffusion models can recover geometric structure in the data manifolds they are trained on, yet the supporting evidence has so far come mostly from natural-image data, where the underlying geometry itself is unknown. We study this question in a setting where the geometry is analytically tractable: constrained inverse kinematics (IK). Each task-space constraint defines a configuration-space manifold with known intrinsic dimension, giving direct ground truth for evaluating the geometry learned by the model. For each of the 6-DoF UR5 and 7-DoF Franka, we train a single conditional diffusion model across seven constraint families, spanning solution manifolds from discrete IK branches to self-motion manifolds. Our empirical results reveal that the intrinsic dimension recovered from the model's score function matches the analytical degrees of freedom of the corresponding constraint manifold across both robots. Moreover, linear interpolation in the latent space leads to generated solutions that remain close to the appropriate constraint manifold, indicating that the learned representation further captures geometric structure of the constraint family beyond intrinsic dimension alone. Constrained IK therefore offers a controlled setting for studying the intrinsic geometry learned by diffusion models.
Chinese Translation
近期研究表明，扩散模型能够恢复其训练数据流形中的几何结构，但迄今为止，支持这一观点的证据主要来自自然图像数据，而其底层几何结构本身是未知的。我们在一个几何结构可解析的环境中研究这个问题：约束逆运动学（IK）。每个任务空间约束定义了一个具有已知内在维度的配置空间流形，为评估模型学习到的几何提供了直接的真实数据。对于6自由度的UR5和7自由度的Franka，我们在七个约束家族上训练了一个单一的条件扩散模型，这些约束家族涵盖了从离散IK分支到自运动流形的解流形。我们的实证结果显示，从模型的评分函数恢复的内在维度与对应约束流形的解析自由度在两个机器人上是一致的。此外，在潜在空间中的线性插值生成的解仍然接近适当的约束流形，表明所学习的表示进一步捕捉了约束家族的几何结构，超越了单纯的内在维度。因此，约束逆运动学为研究扩散模型学习到的内在几何提供了一个受控的环境。
View on arXiv Download PDF AI Translation
cs.RO / 16 / 2606.26423

CoStream: Composing Simple Behaviors for Generalizable Complex Manipulation

CoStream：为可泛化复杂操作构建简单行为
Chen, Haonan, Ma, Yuxiang, Tian, Stephen, Han, Xiaoshen, Huang, Wenlong, Wu, Feiyang, Li, Yunzhu, Wu, Jiajun, Adelson, Edward H., Du, Yilun
Abstract
Long-horizon, contact-rich complex manipulation tasks, such as seating a GPU into a PCIe slot, demand both millimeter high precision and out-of-the-box generalization to new tasks. Existing paradigms struggle to satisfy both: classical pipelines use brittle, task-specific interfaces to achieve high-precision control but require costly pipeline redesigns to adapt to new tasks, whereas monolithic end-to-end policies provide better generalization but lack high precision on complex, out-of-distribution tasks unless retrained with new data. Both paradigms share an implicit assumption: once a manipulation capability is acquired, it must be deployed as a rigid pipeline or monolithic whole, rather than being freely decomposed and recomposed. In this paper, we show that complex manipulation capabilities can emerge naturally from the composition of simple, independent behaviors. Rather than deploying a monolithic policy or a rigid pipeline, we propose \ourshort, a framework orchestrating foundation models and diverse sensing modalities into multiple composable core behaviors: a semantic behavior extracting spatial constraints via foundation models; a predictive behavior forecasting trajectories by tracking keypoints in imagined videos; and a reactive behavior providing high-frequency tactile and force corrections. On a shared $SE(3)$ interface, these outputs compose by right-multiplication into a single pose command at each control step, executed by a compliant controller. We demonstrate \ourshort on 8 real-world tasks spanning everyday manipulation and precision assembly, with the strongest gains in contact-rich assembly and object transfer, and show robust recovery from manual perturbations during execution. {Website:} https://costream-simple.github.io
Chinese Translation
长时间跨度、接触丰富的复杂操作任务，例如将GPU安装到PCIe插槽中，既需要毫米级的高精度，又需要对新任务的开箱即用的泛化能力。现有的范式难以同时满足这两个要求：经典的管道使用脆弱的、特定任务的接口来实现高精度控制，但需要昂贵的管道重设计以适应新任务，而单一的端到端策略提供了更好的泛化能力，但在复杂的、超出分布的任务上缺乏高精度，除非使用新数据重新训练。这两种范式都隐含着一个假设：一旦获得了操作能力，就必须作为一个刚性管道或单一整体进行部署，而不是自由地进行分解和重组。在本文中，我们展示了复杂操作能力可以自然地从简单、独立行为的组合中涌现。我们提出了 extit{CoStream}，一个将基础模型和多种传感模式编排成多个可组合核心行为的框架：一种通过基础模型提取空间约束的语义行为；一种通过跟踪想象视频中的关键点来预测轨迹的预测行为；以及一种提供高频触觉和力校正的反应行为。在共享的$SE(3)$接口上，这些输出通过右乘组合成每个控制步骤的单一姿态命令，由顺应控制器执行。我们在8个涵盖日常操作和精密组装的真实任务上展示了 extit{CoStream}，在接触丰富的组装和物体转移中取得了显著的提升，并展示了在执行过程中对手动扰动的强大恢复能力。{网站：} https://costream-simple.github.io
View on arXiv Download PDF AI Translation
cs.RO / 17 / 2606.26425

A System for Fast, Resilient, and Adaptable Loco-Manipulation Behaviors on Humanoid Robots

一种快速、弹性和适应性的人形机器人运动操控行为系统
Calvert, Duncan William
Abstract
Humanoid robots could take on physically demanding, hazardous, and repetitive work in spaces built for humans. However, a useful robot for these spaces must coordinate locomotion, whole body motion, perception, contact, and operator supervision. This thesis presents a robot-local, runtime-editable behavior authoring and runtime system. Our system strives to be maximally observable, predictable, and directable following Coactive Design principles developed during the DARPA Robotics Challenge. Our operator interface remains continuously synchronized to the robot for runtime authoring, monitoring, and repair. Our behavior architecture uniquely combines object-centric Affordance Templates, organization and logic inspired by Behavior Trees, and runtime-editable perception through a behavior scene and primitive scene actions. Action primitives build on a whole-body controller that supports moving the arms while walking, and use a concurrent action layering algorithm for speed. The behavior library developed during this work covers more than twenty real-robot task variants, including push and pull doors with knob, push-bar, and lever-handle mechanisms, multi-step exploration sequences, obstacle clearing, and reactive table-to-table manipulation tasks. This behavior system has been deployed on many humanoid robots, such as Boston Dynamics' DRC Atlas, NASA's Valkyrie, IHMC and Boardwalk Robotics' Nadia, Unitree's H1-2, and IHMC's Alex. We evaluate our system across capability, speed, reliability, and speed of behavior creation, adaptation, extension, and combination. Our experiments demonstrate that we can adapt, extend, and combine existing behaviors to create novel loco-manipulation behaviors in minutes or hours. Videos: https://www.youtube.com/playlist?list=PLJK5CTyotYqsfgfnXb-09YNFeBose6uEY.
Chinese Translation
人形机器人可以在为人类建造的空间中承担体力要求高、危险和重复的工作。然而，适用于这些空间的有用机器人必须协调运动、全身运动、感知、接触和操作员监督。本文提出了一种机器人本地的、可在运行时编辑的行为创作和运行系统。我们的系统力求在遵循DARPA机器人挑战赛期间开发的协同设计原则下，达到最大程度的可观察性、可预测性和可引导性。我们的操作员界面在运行时与机器人保持持续同步，以便进行行为创作、监控和修复。我们的行为架构独特地结合了以对象为中心的可供性模板、受行为树启发的组织和逻辑，以及通过行为场景和原始场景动作进行的可在运行时编辑的感知。动作原语建立在一个全身控制器之上，该控制器支持在行走时移动手臂，并使用并发动作分层算法来提高速度。此项工作开发的行为库涵盖了二十多种真实机器人任务变体，包括用旋钮、推杆和杠杆手柄机制推拉门、多步骤探索序列、障碍物清除和反应式桌间操控任务。该行为系统已在许多人形机器人上部署，如波士顿动力公司的DRC Atlas、NASA的Valkyrie、IHMC和Boardwalk Robotics的Nadia、Unitree的H1-2以及IHMC的Alex。我们在能力、速度、可靠性以及行为创作、适应、扩展和组合的速度方面评估了我们的系统。实验表明，我们能够在几分钟或几小时内适应、扩展和组合现有行为，以创造新颖的运动操控行为。视频链接： https://www.youtube.com/playlist?list=PLJK5CTyotYqsfgfnXb-09YNFeBose6uEY。
View on arXiv Download PDF AI Translation
cs.RO / 18 / 2606.26428

Play2Perfect: What Matters in Dexterous Play Pretraining for Precise Assembly?

Play2Perfect：灵巧游戏预训练中哪些因素对精确装配至关重要？
Lum, Tyler Ga Wei, Kedia, Kushal, Liu, C. Karen, Bohg, Jeannette
Abstract
Multi-fingered robots promise the speed and dexterity of human hands, yet challenging problems such as precise assembly have remained out of reach. These tasks are contact-rich, making data collection for imitation learning difficult, and sparse-reward, making direct exploration with reinforcement learning (RL) intractable. Consequently, prior work has made progress by structuring the problem with specialized grippers, tool attachments, and environment fixtures. In this work, we argue that before a robot can perfect precise assembly, it must first learn to play. We further ask the question: what factors in the process of learning to play matter for precise assembly? We propose Play2Perfect, an RL framework for task-agnostic pretraining through play on diverse objects and goals, which is then perfected on precise assembly. The goal of play is to acquire reusable manipulation priors, such as grasping, in-hand reorientation and pose reaching. Finetuning then adapts this general prior to assembly, focusing exploration on the final contact-rich, high-precision interactions needed for success. We systematically study key design choices in play pretraining, including object diversity, training objective, trajectory diversity, and goal precision. We show that our prior is 33x more sample-efficient than RL training from scratch, even when provided with dense, multi-stage rewards. We demonstrate zero-shot sim-to-real transfer, achieving 60% success on tight insertions with only 0.5 mm contact clearance, and over 50% success on long-horizon multi-part assembly and screwing.
Chinese Translation
多指机器人承诺具备人类手部的速度和灵活性，但精确装配等挑战性问题仍然难以实现。这些任务涉及丰富的接触，使得模仿学习的数据收集变得困难，同时由于稀疏奖励，使得直接使用强化学习（RL）进行探索变得不可行。因此，以前的研究通过使用专用夹具、工具附件和环境固定装置来构建问题，从而取得了一定进展。在本研究中，我们认为，在机器人能够完善精确装配之前，它必须首先学会玩。我们进一步提出了一个问题：在学习玩耍的过程中，哪些因素对精确装配至关重要？我们提出了Play2Perfect，这是一个通过在多样物体和目标上进行游戏的任务无关预训练的强化学习框架，随后在精确装配上进行完善。游戏的目标是获取可重用的操作先验，例如抓取、手内重定向和姿态到达。微调则将这一通用先验适应于装配，重点探索成功所需的最终接触丰富、高精度的交互。我们系统地研究了游戏预训练中的关键设计选择，包括物体多样性、训练目标、轨迹多样性和目标精度。我们展示了我们的先验比从头开始的RL训练样本效率高33倍，即使在提供密集的多阶段奖励时也是如此。我们展示了零-shot的模拟到现实转移，在仅有0.5毫米接触间隙的紧插入任务中实现了60%的成功率，并在长时间多部件装配和拧紧任务中实现了超过50%的成功率。
View on arXiv Download PDF AI Translation
cs.RO / 19 / 2606.26443

WatchAct: A Benchmark for Behavior-Grounded Robot Manipulation

WatchAct：一个基于行为的机器人操控基准
Li, Baiqi, Zhang, Ce, Fang, Yu, Yang, Yue, Li, Shangzhe, Ding, Mingyu, Bertasius, Gedas
Abstract
A robot working alongside people must reason about what they have done, in what order, and with what intent. Video carries the spatial layouts, object histories, and gestures that language leaves underspecified, yet today's manipulation benchmarks pair an instruction with a single current image, offering no way to evaluate reasoning over observed human behavior. We introduce WatchAct, a benchmark for robot manipulation grounded in observed human behavior. Each instance pairs a real-world human-action video and a language instruction with an aligned simulator scene and an executable LIBERO task, enabling scalable and reproducible evaluation. WatchAct comprises 3,000 long-horizon instances across 14 tasks in four capability domains drawn from the cognitive demands of watching another agent: parsing events (Event Grounding), recovering procedural structure (Procedural Reasoning), inferring unstated intent (Implicit Intent Inference), and tracking how the scene was changed (Episodic Reasoning). We further propose a disentangled evaluation protocol that separately measures (i)~video-to-plan reasoning by vision-language models, (ii)~policy execution under oracle plans, and (iii)~full task completion by integrated planner--policy pipelines. In both simulation and on a Franka Research 3 robot, current systems remain far from solving WatchAct. The best pipeline, Gemini-3.1-Pro with $\pi_{0.5}$, reaches only 16.3% Success Rate (SR) in simulation and 14.0% on the real robot. Gemini-3.1-Pro attains just 36.8% Plan SR (vs. 97.1% for humans), while $\pi_{0.5}$ reaches only 21.5% Task SR under oracle plans and drops to 10.6% on out-of-domain scenarios. Dataset and code are available at https://baiqi-li.github.io/watchact_page/.
Chinese Translation
与人类协作的机器人必须推理他们所做的事情、顺序以及意图。视频承载了空间布局、物体历史和语言未明确指定的手势，但当前的操控基准将指令与单一当前图像配对，无法评估对观察到的人类行为的推理。我们提出了WatchAct，这是一个基于观察到的人类行为的机器人操控基准。每个实例将一个现实世界的人类动作视频与一个语言指令配对，并与一个对齐的模拟场景和可执行的LIBERO任务相结合，从而实现可扩展和可重复的评估。WatchAct包含3,000个长时间跨度的实例，涵盖来自观察其他代理的认知需求的14个任务的四个能力领域：事件解析（Event Grounding）、恢复程序结构（Procedural Reasoning）、推断未说明的意图（Implicit Intent Inference）和跟踪场景的变化（Episodic Reasoning）。我们进一步提出了一种解耦评估协议，分别测量（i）视觉-语言模型的视频到计划推理，（ii）在oracle计划下的策略执行，以及（iii）通过集成规划者-策略管道的完整任务完成。在模拟和Franka Research 3机器人上，当前系统仍远未解决WatchAct。最佳管道Gemini-3.1-Pro与$ ext{π}_{0.5}$在模拟中仅达到16.3%的成功率（SR），在真实机器人上为14.0%。Gemini-3.1-Pro的计划成功率仅为36.8%（而人类为97.1%），而在oracle计划下，$ ext{π}_{0.5}$的任务成功率仅为21.5%，在域外场景中下降至10.6%。数据集和代码可在https://baiqi-li.github.io/watchact_page/获取。
View on arXiv Download PDF AI Translation
cs.RO / 20 / 2606.26533

OSC2Runner: OpenSCENARIO 2.x Compliant High-Fidelity AV Simulation in CARLA

OSC2Runner：符合 OpenSCENARIO 2.x 的高保真自动驾驶仿真框架在 CARLA 中的应用
Gamage, Thoshitha, Gamage, Lasanthi
Abstract
Scenario-Based Testing predominantly relies on the legacy ASAM OpenSCENARIO 1.x XML standard because existing continuous simulation frameworks lack native execution support for the recently matured v2.x Domain-Specific Language (DSL). Adapting legacy interpreters to evaluate v2.x logic introduces spatiotemporal drift, asynchronous event latencies, and artificial kinematic snapping. Addressing this execution gap, OSC2Runner introduces the first orchestration framework capable of natively mapping the OpenSCENARIO v2.x DSL to CARLA. The framework achieves this by formalizing scenario translation as a compilation pipeline through a multi-pass transpiler architecture. Bypassing static trajectory playback, the architecture synthesizes type-safe Abstract Syntax Trees directly into dynamic deterministic behavior trees (py_trees) natively mapped to CARLA's atomic APIs. Empirical validation in highly concurrent adversarial case studies demonstrates tick-by-tick determinism, exact spatial trigger evaluation, and 100.0 ms cross-actor blackboard synchronization. Kinematic analysis proves the strict adherence to continuous environmental boundaries. This architecture transitions Scenario-Based Testing from approximate behavioral interpretation to mathematically rigorous execution, establishing the deterministic backend required for co-simulation, hardware-in-the-loop testing, and automated LLM-driven generation pipelines.
Chinese Translation
基于场景的测试主要依赖于传统的 ASAM OpenSCENARIO 1.x XML 标准，因为现有的连续仿真框架缺乏对最近成熟的 v2.x 特定领域语言 (DSL) 的原生执行支持。将传统解释器调整为评估 v2.x 逻辑会引入时空漂移、异步事件延迟和人为运动学跳跃。为了解决这一执行差距，OSC2Runner 引入了第一个能够将 OpenSCENARIO v2.x DSL 原生映射到 CARLA 的编排框架。该框架通过将场景翻译形式化为一个编译管道，采用多遍历的转译器架构来实现这一目标。该架构绕过静态轨迹播放，直接将类型安全的抽象语法树合成到原生映射到 CARLA 原子 API 的动态确定性行为树 (py_trees) 中。通过在高度并发的对抗性案例研究中的实证验证，展示了逐步确定性、精确的空间触发评估和 100.0 毫秒的跨参与者黑板同步。运动学分析证明了对连续环境边界的严格遵守。该架构将基于场景的测试从近似行为解释转变为数学上严格的执行，为协同仿真、硬件在环测试和自动化 LLM 驱动的生成管道建立了所需的确定性后端。
View on arXiv Download PDF AI Translation
cs.RO / 21 / 2606.26575

IDEA: Insensitive to Dynamics Mismatch via Effect Alignment for Sim-to-Real Transfer in Multi-Agent Control

IDEA：通过效应对齐实现对动态不匹配的不敏感性，以促进多智能体控制中的仿真到现实转移
Liu, Chenlong, Zhang, Zhuohui, Chen, Xinyan, Wang, Zhipeng, Cheng, Bin, He, Bin
Abstract
Complex multi-agent control tasks remain challenging for traditional rule-based and model-based approaches, motivating the adoption of learning-based methods. However, learning-based methods often struggle with sim-to-real transfer because they rely on accurate dynamics modeling or system identification and learn policies in low-level control spaces that are highly sensitive to dynamics mismatch, making them costly and fragile in complex environments. To address this issue, we propose a sim-to-real method for multi-agent control, which is insensitive to dynamics mismatch via effect alignment. Our method combines random environmental structure with discrete semantic actions through closed-loop control, elevating policy learning to a semantic abstraction level. Additionally, we develop an action synchronization mechanism that mitigates inter-agent action timing mismatches, thereby enhancing the temporal consistency of the system. Experiments on four multi-agent navigation tasks demonstrate that our method substantially improves training efficiency over mainstream transfer methods and achieves higher success rates in real-world scenarios, thereby improving the robustness and deployment stability of multi-agent systems under dynamics mismatch.
Chinese Translation
复杂的多智能体控制任务对传统的基于规则和基于模型的方法仍然具有挑战性，这促使了基于学习的方法的采用。然而，基于学习的方法在仿真到现实转移中常常面临困难，因为它们依赖于准确的动态建模或系统识别，并在对动态不匹配高度敏感的低级控制空间中学习策略，这使得它们在复杂环境中成本高昂且脆弱。为了解决这一问题，我们提出了一种多智能体控制的仿真到现实方法，该方法通过效应对齐实现对动态不匹配的不敏感性。我们的方法通过闭环控制将随机环境结构与离散语义动作相结合，将策略学习提升到语义抽象层次。此外，我们开发了一种动作同步机制，以减轻智能体间动作时序的不匹配，从而增强系统的时间一致性。在四个多智能体导航任务上的实验表明，我们的方法在训练效率上显著优于主流转移方法，并在现实场景中实现了更高的成功率，从而提高了多智能体系统在动态不匹配下的鲁棒性和部署稳定性。
View on arXiv Download PDF AI Translation
cs.RO / 22 / 2606.26588

Inference-Time Robot Behavior Steering through Physically-Aware Reconfiguration of Task-Structure

通过物理感知的任务结构重新配置实现推理时的机器人行为引导
Pan, Yiyuan, Hu, Hanjiang, Li, Shangtao, Luo, Xusheng, Liu, Changliu
Abstract
A central challenge in deploying learned robot policies is inference-time behavior steering: redirecting a policy at test time to satisfy user preferences not anticipated during training, without retraining. Existing methods fail in two modes: end-to-end methods require fine-tuning or expert-level guidance, while neuro-symbolic methods rely on predefined symbols whose edits can result in logically reasonable but physically infeasible plans. To address this challenge, we propose ReStruct, which builds upon a neural automaton policy that decomposes a visuomotor policy into a high-level state-machine skeleton capturing task structure and a low-level continuous controller represented as a residual policy. Specifically, ReStruct adopts the automaton to represent the preference and incorporates it into the skeleton through a synchronous product, thereby reconfiguring the task structure. With the controller kept frozen, the action priors provided by the skeleton are updated accordingly to enable physically-aware control under a modified task structure. Extensive experiments from simulation and real-world show that ReStruct steers a wide range of preferences, from object-centric specifications to temporal-logic constraints, and after steering surpasses existing methods, exceeding VLA models in both task success and preference-following by up to 25%.
Chinese Translation
在部署学习到的机器人策略时，一个核心挑战是推理时的行为引导：在测试时重新定向策略以满足训练期间未预见的用户偏好，而无需重新训练。现有方法在两种模式下失败：端到端方法需要微调或专家级指导，而神经符号方法依赖于预定义符号，其编辑可能导致逻辑上合理但在物理上不可行的计划。为了解决这一挑战，我们提出了ReStruct，它基于神经自动机策略，将视觉运动策略分解为捕捉任务结构的高层状态机框架和作为残差策略表示的低层连续控制器。具体而言，ReStruct采用自动机来表示偏好，并通过同步乘积将其纳入框架中，从而重新配置任务结构。在控制器保持不变的情况下，框架提供的动作先验相应更新，以实现基于修改后的任务结构的物理感知控制。来自仿真和现实世界的广泛实验表明，ReStruct能够引导从以对象为中心的规格到时间逻辑约束等广泛的偏好，并且在引导后超越现有方法，在任务成功率和偏好遵循方面超过VLA模型，提升幅度高达25%。
View on arXiv Download PDF AI Translation
cs.RO / 23 / 2606.26603

Bridging Handheld and Teleoperated Supervision for Contact-Rich Manipulation via State-Gated Experts

通过状态门控专家桥接手持和遥控监督以实现接触丰富的操作
Surendran, Vidullan, Peri, Neehar, Watkins, David
Abstract
Handheld data collection systems, such as the Universal Manipulation Interface (UMI), enable scalable data collection across diverse environments but only capture observed actions rather than the desired actions executed by a robot controller. In contrast, teleoperation captures desired actions directly, but is prohibitively time-consuming to collect. We revisit this trade-off through the lens of action validity across task phases. We observe that handheld trajectories provide valid supervision in tolerant, free-space phases, but lack dynamic feasibility in contact-sensitive phases, where tracking observed trajectories at high stiffness produces large, unsafe contact forces. We study the interaction between these two supervision types for contact-rich manipulation and find that training policies that combine handheld data with a small number of targeted teleoperated demonstrations provide an efficient hybrid strategy. Specifically, rather than teleoperating the entire task, we only collect partial teleoperated demonstrations for task segments where base handheld policies fail. However, naively mixing handheld and teleoperated phase-specific data yields worse performance than training on handheld data alone. To address this mismatch between observed and desired supervision, we propose Bi-modal Routing for Imitation Data via Gated Experts (BRIDGE), a mixture of diffusion policy experts that routes between specialist task phase heads conditioned on the current robot state. Notably, our approach enables task-phase specific use of desired actions during contact sensitive segments and improves success rates over handheld-only baselines by up to 36.7% across three contact-rich manipulation tasks.
Chinese Translation
手持数据采集系统，如通用操作接口（Universal Manipulation Interface, UMI），能够在多样化环境中进行可扩展的数据采集，但仅捕捉观察到的动作，而非机器人控制器执行的期望动作。相比之下，遥控直接捕捉期望动作，但采集过程耗时过长。我们通过任务阶段的动作有效性重新审视这一权衡。我们观察到，手持轨迹在宽容的自由空间阶段提供有效的监督，但在接触敏感阶段缺乏动态可行性，在这些阶段，以高刚度跟踪观察到的轨迹会产生大的、不安全的接触力。我们研究这两种监督类型在接触丰富操作中的相互作用，发现结合手持数据与少量针对性遥控演示训练策略提供了一种高效的混合策略。具体而言，我们并不对整个任务进行遥控，而仅在基础手持策略失败的任务片段中收集部分遥控演示。然而，简单地混合手持和遥控阶段特定数据的效果不如仅在手持数据上训练。为了解决观察到的监督与期望监督之间的不匹配，我们提出了通过门控专家的模态双重路由（Bi-modal Routing for Imitation Data via Gated Experts, BRIDGE），这是一种扩散策略专家的混合体，根据当前机器人状态在专业任务阶段头之间进行路由。值得注意的是，我们的方法在接触敏感阶段实现了期望动作的任务阶段特定使用，并在三个接触丰富操作任务中将成功率提高了高达36.7%，相较于仅使用手持基线的表现。
View on arXiv Download PDF AI Translation
cs.RO / 24 / 2606.26616

A Closed-Form 4-DoF Inter-Robot Pose Estimator using Bearing-only Measurements

基于仅测量方位的闭式4自由度机器人间位姿估计器
De, Qixin, Zhuang, Ao, Zhang, Yechen, Qian, Zhuozhou, Zou, Danping
Abstract
Bearing-odometry-based cooperative localization has attracted increasing research interest due to its minimal infrastructure requirements, low communication bandwidth and broad applicability in complex environments. However, existing 6-DoF approaches still face challenges in rapidly obtaining accurate and reliable inter-robot pose estimation, as the system is prone to observability degeneracy under specific motion patterns. To address these issues, we first propose a closed-form 4-DoF inter-robot pose estimator, which relaxes nonlinear constraints for rotations estimation and employs error projection for translations estimation. We then conduct a theoretical analysis of the system's observability, identifying degeneracy under two typical motion patterns: collinear and shape-preserving formations. The analysis further shows that the proposed 4-DoF system requires less stringent motion excitation for observability, enabling reliable estimation under a broader range of cooperative maneuvers. Furthermore, an observability test module is introduced to autonomously determine the optimal estimation instant, eliminating reliance on a predefined fixed-length sliding window. Extensive simulations and real-world experiments demonstrate that the proposed algorithm achieves higher estimation accuracy with significantly low computational cost, and the observability test module ensures estimation reliability while minimizing the data collection interval.
Chinese Translation
基于方位的协作定位因其对基础设施的最低需求、低通信带宽以及在复杂环境中的广泛适用性而受到越来越多的研究关注。然而，现有的6自由度方法在快速获得准确可靠的机器人间位姿估计方面仍面临挑战，因为系统在特定运动模式下容易出现可观测性退化。为了解决这些问题，我们首先提出了一种闭式4自由度机器人间位姿估计器，该估计器放宽了旋转估计的非线性约束，并采用误差投影进行平移估计。随后，我们对系统的可观测性进行了理论分析，识别出在两种典型运动模式下的退化现象：共线和保持形状的编队。分析进一步表明，所提出的4自由度系统对可观测性的运动激励要求不那么严格，从而能够在更广泛的协作机动中实现可靠估计。此外，我们引入了一个可观测性测试模块，以自主确定最佳估计时刻，消除了对预定义固定长度滑动窗口的依赖。大量的仿真和实际实验表明，所提出的算法在显著降低计算成本的同时，实现了更高的估计精度，而可观测性测试模块确保了估计的可靠性，同时最小化了数据收集间隔。
View on arXiv Download PDF AI Translation
cs.RO / 25 / 2606.26643

Hardware Design for Table Tennis Robot Capable of Beating Professional Players

能够击败职业选手的乒乓球机器人硬件设计
Mukai, Nobuhiko, Adodin, Pavel, Heusser, Stefan, Sigrist, Alexander, Grover, Divij, Torrente, Guillem, Khadivar, Farshad, Kakinuma, Takekazu, Dürr, Peter
Abstract
This paper focuses on the hardware specifications required for a table tennis robot to beat professional players. After analyzing the motions of elite players, we defined target specifications for the workspace, payload, external-force resistance, physical performance, serve capability, and end-effector accuracy. Based on these specifications, we developed "Ace", a custom 8-DoF robot. The mechanical structure was improved through topology optimization to minimize mass while preserving stiffness. Motor and gearbox selection was optimized using an inverse-dynamics torque model. Low-order per-joint dynamics models with delay compensation were identified and integrated into simulation to enable the use of an RL control policy. Experiments demonstrated repeated full-stroke swings with a cycle time of 0.8 s and a peak racket-center velocity of 22 m/s. The robot successfully defeated multiple professional players.
Chinese Translation
本文关注于乒乓球机器人击败职业选手所需的硬件规格。在分析了精英选手的动作后，我们定义了工作空间、负载、抗外力能力、物理性能、发球能力和末端执行器精度的目标规格。基于这些规格，我们开发了名为“Ace”的定制8自由度机器人。通过拓扑优化改进机械结构，以最小化质量同时保持刚度。使用逆动力学扭矩模型优化了电机和齿轮箱的选择。识别并将低阶每关节动力学模型与延迟补偿集成到仿真中，以便使用强化学习（RL）控制策略。实验表明，该机器人能够以0.8秒的周期时间和22米/秒的最高球拍中心速度重复进行全挥杆动作。该机器人成功击败了多名职业选手。
View on arXiv Download PDF AI Translation
cs.RO / 26 / 2606.26661

LAMP: Lane-Aligned Motion Primitives for Feasible Trajectory Prediction

LAMP：基于车道对齐的运动原语用于可行轨迹预测
Han, Sangjin, Jung, Hoseong, Her, Jeongtae, Choi, Changhyun, Kim, H. Jin
Abstract
Motion forecasting is essential for autonomous driving systems to enable safe decision-making and planning in complex driving scenarios. While existing predictors excel at minimizing standard displacement errors, they often overlook the adherence to lane topology of multimodal predictions, particularly for lower-probability modes. Consequently, predicted trajectories may violate physical and logical constraints, making the prediction set unreliable for safety-critical planning. In this paper, we propose LAMP (Lane-Aligned Motion Primitives), a topology-aware forecasting framework that anchors multimodal prediction to structured motion primitives aligned with lane topology. Specifically, we use a VQ-VAE to learn shape-aware motion primitives as discrete intention queries, capturing spatiotemporal patterns beyond endpoint-based intentions. We further introduce a feasibility-aware intention selector trained with a lane-topology prior for filtering unreachable intention queries, guiding the decoder to prioritize topology-consistent intentions while preserving behavioral diversity. Extensive experiments on the Argoverse 2 dataset demonstrate that LAMP achieves prediction accuracy comparable to state-of-the-art baselines while outperforming them in feasibility and diversity metrics.
Chinese Translation
运动预测对于自主驾驶系统在复杂驾驶场景中实现安全决策和规划至关重要。尽管现有的预测模型在最小化标准位移误差方面表现出色，但它们往往忽视了多模态预测中对车道拓扑的遵循，尤其是在低概率模式下。因此，预测的轨迹可能违反物理和逻辑约束，使得预测集在安全关键的规划中变得不可靠。本文提出了LAMP（Lane-Aligned Motion Primitives），一种关注拓扑的预测框架，将多模态预测锚定在与车道拓扑对齐的结构化运动原语上。具体而言，我们使用VQ-VAE学习形状感知的运动原语作为离散意图查询，捕捉超越基于端点意图的时空模式。我们进一步引入了一种基于可行性的意图选择器，该选择器通过车道拓扑先验进行训练，以过滤不可达的意图查询，引导解码器优先考虑与拓扑一致的意图，同时保持行为多样性。在Argoverse 2数据集上的大量实验表明，LAMP在预测准确性上可与最先进的基线相媲美，同时在可行性和多样性指标上超越了它们。
View on arXiv Download PDF AI Translation
cs.RO / 27 / 2606.26663

Tactile-WAM: Touch-Aware World Action Model with Tactile Asymmetric Attention

触觉-WAM：具有触觉不对称注意力的触觉感知世界动作模型
Wu, Siyu, You, Linjing, Zhu, Junjie, Liu, Yaozu, Zhang, Changhao, Liu, Jian, Wang, Weiqiang, Li, Qi, Li, Jituo, Zhao, Hengshuang
Abstract
World Action Models (WAMs) generate actions together with predicted futures, offering a powerful interface for robot decision making. In contact-rich manipulation, however, visually plausible futures can be physically incomplete: insertion, assembly, search, and reorientation often depend on slip, jamming, contact normals, or small alignment errors that are weakly visible or hidden in RGB. A natural solution is to predict future tactile states, however, we identify tactile pollution, a failure mode where unconstrained tactile-token injection degrades video and action prediction by forcing a visual dynamics model to absorb sparse, local, event-driven contact signals. To address this, we propose Tactile-WAM, a touch-aware WAM with a Tactile Asymmetric Attention Mechanism (TAAM). TAAM combines a VideoClean mask, which blocks video-query access to tactile key/value tokens while preserving action-query access, with a touch-aware bias for action attention. The VideoClean mask protects visual prediction while keeping contact information available for action generation; the touch-aware bias is derived from predicted touch changes and modulates action attention to tactile tokens during denoising. On ManiFeel, Tactile-WAM improves the mean success rate by 38.9% overall and by 86% on contact-rich tasks.
Chinese Translation
世界动作模型（WAMs）生成动作及其预测的未来，为机器人决策提供了强大的接口。然而，在接触丰富的操作中，视觉上合理的未来可能在物理上是不完整的：插入、组装、搜索和重新定向往往依赖于滑动、卡滞、接触法线或小的对齐误差，这些在RGB图像中往往难以看见或被隐藏。一个自然的解决方案是预测未来的触觉状态，然而，我们识别出触觉污染，这是一种失败模式，其中不受限制的触觉标记注入通过迫使视觉动态模型吸收稀疏的、局部的、事件驱动的接触信号而降低了视频和动作预测的质量。为了解决这个问题，我们提出了触觉-WAM，一种具有触觉不对称注意力机制（TAAM）的触觉感知WAM。TAAM结合了视频清洁掩码，该掩码阻止视频查询访问触觉关键/值标记，同时保留动作查询的访问权限，并且具有针对动作注意力的触觉感知偏置。视频清洁掩码保护视觉预测，同时保持接触信息可用于动作生成；触觉感知偏置源自预测的触觉变化，并在去噪过程中调节对触觉标记的动作注意力。在ManiFeel上，触觉-WAM整体提高了38.9%的平均成功率，并在接触丰富的任务上提高了86%。
View on arXiv Download PDF AI Translation
cs.RO / 28 / 2606.26700

Learning Motion Feasibility from Point Clouds in Cluttered Environments

从杂乱环境中的点云学习运动可行性
Ansari, Sajid, Arthi, Varma, Girish, Thomas, Antony
Abstract
Motion feasibility prediction plays a central role in robotics, particularly in task and motion planning and manipulation. A major bottleneck for this problem in cluttered environments is that infeasible planning attempts by Sampling-based motion planners (SBMPs) can incur substantial computational cost. Also existing approaches for infeasibility certification are limited to low-dimensional configuration spaces and often assume simplified geometric environments represented by primitive objects with known parameters. We study the complementary problem of learning motion feasibility prediction directly from raw RGB-D observations for a 7-DOF manipulator operating in realistic cluttered scenes. We introduce the first large-scale benchmark for this setting, comprising 2.7M grasp feasibility labels over 88 scanned objects and 190 cluttered tabletop scenes. We benchmark three representative classifier families spanning MLP- based, volumetric-CNN, and point-cloud-based Transformer architectures under matched training conditions. Our best model, GRASPFC-PTX (a point-cloud transformer), achieves an AUROC of 0.996 on Novel objects while providing predictions significantly faster than SBMPs.
Chinese Translation
运动可行性预测在机器人技术中扮演着核心角色，尤其是在任务和运动规划以及操作中。在杂乱环境中，这一问题的主要瓶颈在于基于采样的运动规划器（SBMPs）所进行的不可行规划尝试可能会导致巨大的计算成本。此外，现有的不可行性认证方法仅限于低维配置空间，并且通常假设由具有已知参数的基本对象表示的简化几何环境。我们研究了从原始RGB-D观测中直接学习运动可行性预测的互补问题，针对在现实杂乱场景中操作的7自由度（7-DOF）操纵器。我们为这一设置引入了第一个大规模基准，包含2.7百万个抓取可行性标签，涵盖88个扫描对象和190个杂乱的桌面场景。我们在匹配的训练条件下，对三种代表性的分类器家族进行了基准测试，分别是基于多层感知器（MLP）、体积卷积神经网络（volumetric-CNN）和基于点云的变换器（Transformer）架构。我们的最佳模型GRASPFC-PTX（一个点云变换器）在新对象上达到了0.996的AUROC，同时提供的预测速度显著快于SBMPs。
View on arXiv Download PDF AI Translation
cs.RO / 29 / 2606.26741

PressMimic: Pressure-Guided Motion Capture and Control for Humanoid Robot Imitation

PressMimic：基于压力引导的人形机器人模仿运动捕捉与控制
Lu, Yi, Ren, Shenghao, Xiong, Tianyu, Li, Zhaoxiang, Li, Jiaqi, Zhang, He, Yu, Tao, Shen, Qiu, Cao, Xun
Abstract
Humanoid motion imitation requires not only accurate perception of human kinematics but also faithful reproduction of physical interactions with the environment. However, existing pipelines rely primarily on vision-based motion capture and kinematic imitation, largely ignoring contact dynamics, leading to artifacts such as foot sliding, floor penetration, and unstable behaviors. In this work, we revisit humanoid motion imitation from the perspective of physical grounding and leverage pressure as a unified modality across perception and control. We present PressMimic, a framework that integrates pressure into the full pipeline from motion capture to humanoid control. In the perception stage, we introduce FRAPPE++, a multimodal model that fuses RGB and pressure to jointly estimate 3D pose and global motion, where pressure provides explicit contact and support constraints to resolve ambiguity in vision-based estimation. In the control stage, we propose a pressure-supervised policy (PSP) that incorporates pressure-derived signals into reinforcement learning, enabling physically consistent contact patterns during execution. We further construct MotionPRO, a large-scale dataset with synchronized RGB, pressure, and motion capture data. Experiments show that pressure improves motion estimation accuracy, trajectory consistency, and execution stability. These results demonstrate that pressure serves as an effective physical grounding signal, bridging perception and control for physically consistent humanoid motion imitation.
Chinese Translation
人形运动模仿不仅需要准确感知人类运动学，还需要真实再现与环境的物理交互。然而，现有的流程主要依赖基于视觉的运动捕捉和运动学模仿，往往忽视接触动态，导致诸如脚滑、地面穿透和不稳定行为等伪影。在本研究中，我们从物理基础的角度重新审视人形运动模仿，并利用压力作为感知与控制之间的统一模态。我们提出了PressMimic，一个将压力整合到从运动捕捉到人形控制的完整流程中的框架。在感知阶段，我们引入了FRAPPE++，一个将RGB和压力融合的多模态模型，以共同估计3D姿态和全局运动，其中压力提供明确的接触和支撑约束，以解决基于视觉的估计中的歧义。在控制阶段，我们提出了一种压力监督策略（PSP），将压力衍生信号纳入强化学习，使得在执行过程中能够实现物理一致的接触模式。我们进一步构建了MotionPRO，一个包含同步RGB、压力和运动捕捉数据的大规模数据集。实验表明，压力提高了运动估计的准确性、轨迹一致性和执行稳定性。这些结果表明，压力作为有效的物理基础信号，架起了感知与控制之间的桥梁，实现了物理一致的人形运动模仿。
View on arXiv Download PDF AI Translation
cs.RO / 30 / 2606.26800

SSI-Policy: Learning Structured Scene Interfaces for Vision-Language Robotic Manipulation

SSI-Policy：学习用于视觉-语言机器人操作的结构化场景接口
Wang, Kaijun, Ouyang, Zikai, Wu, Xuping, Hong, Jinyi, Pan, Wei, Lu, Haibo, Pan, Jia, Zhang, Wei, Zheng, Linfang
Abstract
Real-world robotic manipulation demands spatial grounding, task-aware reasoning, and precise control. Learning such capabilities becomes particularly challenging in the low-data regime. Prior methods often trade off scalable task-level reasoning and explicit physical structure: video-based approaches can drift geometrically over long horizons, 3D approaches often require depth sensing, and many flow/trajectory interfaces emphasize motion without an explicit RGB-only geometric representation. We introduce SSI-Policy, a modular framework built around a Structured Scene Interface (SSI) -- a unified, RGB-only intermediate representation that jointly encodes monocular depth features, language-grounded object layouts, and instruction-conditioned 2D motion trajectories. Critically, SSI is robot-agnostic and trainable from action-free video, decoupling perception from control so that the downstream policy can learn from few demonstrations. On the LIBERO benchmark with only 10 demonstrations per task, SSI-Policy improves over the strongest prior method by nearly 15\% and remains competitive with 50-demo methods that leverage large-scale external pretraining. Ablations show that geometric and motion cues provide complementary benefits within the shared interface. We further validate on 13 real-world tasks spanning spatial reasoning, cross-embodiment transfer, and contact-rich manipulation.
Chinese Translation
现实世界中的机器人操作需要空间定位、任务感知推理和精确控制。在低数据环境下学习这些能力尤其具有挑战性。以往的方法通常在可扩展的任务级推理和明确的物理结构之间进行权衡：基于视频的方法在长时间范围内可能会出现几何漂移，3D方法通常需要深度传感，而许多流/轨迹接口则强调运动而没有明确的仅基于RGB的几何表示。我们提出了SSI-Policy，一个围绕结构化场景接口（Structured Scene Interface, SSI）构建的模块化框架——这是一个统一的、仅基于RGB的中间表示，联合编码单目深度特征、语言基础的物体布局和指令条件的2D运动轨迹。重要的是，SSI与机器人无关，并且可以从无动作视频中进行训练，将感知与控制解耦，从而使下游策略能够从少量演示中学习。在仅有10个演示的LIBERO基准测试中，SSI-Policy的表现比最强的先前方法提高了近15%，并且在利用大规模外部预训练的50个演示方法中仍然具有竞争力。消融实验表明，几何和运动线索在共享接口中提供了互补的好处。我们进一步验证了涵盖空间推理、跨体现迁移和接触丰富操作的13个现实世界任务。
View on arXiv Download PDF AI Translation
cs.RO / 31 / 2606.26801

Improving Vision-Language-Action Model Fine-Tuning with Structured Stage and Keyframe Supervision

通过结构化阶段和关键帧监督改进视觉-语言-动作模型的微调
Xu, Yuan, Chen, Yixiang, Wang, Kai, Yang, Jiabing, Li, Peiyan, Ma, Qisen, Huang, Yan, Wang, Liang
Abstract
Vision-Language-Action (VLA) models have shown strong potential for generalizable robotic manipulation. During fine-tuning, however, action supervision applies equally across all timesteps, without structured supervision on which manipulation stage the robot is in or what the next gripper-event target should be. This causes failures to concentrate around challenging gripper-event transitions. To address this, we propose StaKe, a plug-in auxiliary supervision framework that automatically derives two complementary signals from demonstration gripper states without manual annotation: a stage classifier that identifies the current manipulation stage, and a keyframe predictor that estimates the target joint action at the next gripper transition. Both are modeled as lightweight auxiliary heads that enrich the learned representations during training, while leaving the base VLA policy architecture and inference loop unchanged. Experiments on bimanual simulation and single-arm Franka real-robot tasks show that StaKe consistently improves success rates (relative gains of 14% and 56%, respectively), with larger improvements on longer-horizon tasks that involve more gripper-event transitions. Ablation studies validate each design choice, and qualitative analysis confirms that the learned representations faithfully track manipulation stages. These results indicate that structured supervision is an effective and general strategy for enhancing VLA fine-tuning in long-horizon manipulation. Project website: https://hi-yuanxu.github.io/StaKe-Web/
Chinese Translation
视觉-语言-动作（VLA）模型在可推广的机器人操控方面展现了强大的潜力。然而，在微调过程中，动作监督在所有时间步上均匀应用，缺乏对机器人当前处于哪个操控阶段或下一个夹持事件目标应是什么的结构化监督。这导致在困难的夹持事件过渡阶段出现失败。为了解决这个问题，我们提出了StaKe，一个插件式辅助监督框架，它自动从演示夹持状态中推导出两个互补信号，而无需手动标注：一个阶段分类器用于识别当前的操控阶段，以及一个关键帧预测器用于估计下一个夹持过渡时的目标关节动作。这两者都被建模为轻量级的辅助头，在训练过程中丰富学习到的表示，同时保持基础VLA策略架构和推理循环不变。在双手模拟和单臂Franka真实机器人任务上的实验表明，StaKe始终提高了成功率（相对增益分别为14%和56%），在涉及更多夹持事件过渡的长时间任务中改善更显著。消融研究验证了每个设计选择的有效性，定性分析确认学习到的表示忠实地跟踪操控阶段。这些结果表明，结构化监督是一种有效且通用的策略，用于增强长时间操控中的VLA微调。项目网站：https://hi-yuanxu.github.io/StaKe-Web/
View on arXiv Download PDF AI Translation
cs.RO / 32 / 2606.26839

Ordinal Neural Collapse as a Representation Prior for Visual Navigation

序数神经崩溃作为视觉导航的表征先验
Son, E-In, Kim, Jung-Taak, Seo, Seung-Woo
Abstract
Learning robust navigation policies directly from visual observations remains a fundamental challenge in vision-based robotic navigation. In end-to-end imitation learning approaches, the visual encoder and action decoder are jointly optimized using a single action loss, which provides only an indirect supervisory signal to the encoder. This indirect supervision frequently results in the encoder learning ambiguous, action-agnostic representations. The problem is further complicated by substantial variations in scene structure and appearance across diverse environments, as well as the prevalence of visual distractors inherent to real-world navigation settings. Such action-agnostic features cause the navigation policy to produce inconsistent actions at ambiguous decision points, leading to navigation failure. To overcome these limitations, we propose ORION (Ordinal Neural Collapse for Visual Navigation), a method that explicitly organizes the encoder's representation space according to the ordinal structure of navigation actions. In the context of goal-directed navigation, ego-centric control categories from Far Left to Far Right exhibit a natural ordinal relationship in which neighboring classes share similar visual contexts, while semantically opposing classes differ substantially in appearance. We encourage class representations to be arranged sequentially along a single discriminative axis, while suppressing off-axis variance within each class. The pretrained encoder is then integrated into a diffusion-based navigation framework, and the full pipeline is fine-tuned end-to-end. Extensive experiments in both simulation and real-world settings show that ORION consistently outperforms end-to-end and neural collapse baselines in navigation success rate and goal progress, with notable gains in visually challenging scenarios such as complex multi-way intersections.
Chinese Translation
从视觉观察中直接学习稳健的导航策略仍然是基于视觉的机器人导航中的一个基本挑战。在端到端模仿学习方法中，视觉编码器和动作解码器通过单一的动作损失共同优化，这仅为编码器提供了间接的监督信号。这种间接监督常常导致编码器学习到模糊的、与动作无关的表征。问题因不同环境中场景结构和外观的显著变化，以及真实世界导航环境中固有的视觉干扰物的普遍存在而进一步复杂化。这种与动作无关的特征使得导航策略在模糊决策点产生不一致的动作，导致导航失败。为了解决这些局限性，我们提出了ORION（Ordinal Neural Collapse for Visual Navigation），一种明确根据导航动作的序数结构组织编码器表征空间的方法。在目标导向导航的背景下，从极左到极右的自我中心控制类别表现出自然的序数关系，其中相邻类别共享相似的视觉上下文，而语义上对立的类别在外观上有显著差异。我们鼓励类别表征沿着单一的判别轴顺序排列，同时抑制每个类别内的轴外方差。然后，将预训练的编码器集成到基于扩散的导航框架中，并对整个流程进行端到端的微调。在模拟和真实环境中的广泛实验表明，ORION在导航成功率和目标进展方面始终优于端到端和神经崩溃基线，尤其在复杂的多路交叉口等视觉挑战场景中表现出显著的提升。
View on arXiv Download PDF AI Translation
cs.RO / 33 / 2606.26855

Humanoid-DART: Humanoid Loco-Manipulation using Diffusion-guided Augmentation through Relabeling and Tracking

人形-DART：通过重标记和跟踪的扩散引导增强实现人形运动操控
Debbad, Pranav, Thiagarajan, Kanish, Dhédin, Victor, Omar, Shafeef, Khadiv, Majid
Abstract
Imitating human demonstrations has emerged as a dominant paradigm for learning humanoid loco-manipulation policies. However, scaling these approaches remains challenging due to the high cost of collecting diverse demonstrations and the need for continual human intervention to correct policy failures. In this paper, we present a self-supervised framework that bootstraps from sparse demonstrations and progressively expands its behavioral repertoire, enabling the learning of a goal-conditioned policy that automatically explores the goal space with minimal expert supervision. Our approach combines diffusion-based trajectory generation with reinforcement learning, where the latter is used to track goal-conditioned trajectories produced by the diffusion model for a range of loco-manipulation skills. Through extensive ablation studies and comparisons with state-of-the-art methods, we demonstrate the effectiveness of our framework on multiple humanoid loco-manipulation skills.
Chinese Translation
模仿人类示范已成为学习人形运动操控策略的主要范式。然而，由于收集多样化示范的高成本以及需要持续的人类干预来纠正策略失败，这些方法的扩展仍然具有挑战性。本文提出了一种自监督框架，该框架从稀疏示范中启动，并逐步扩展其行为库，使其能够学习一种目标条件策略，该策略在最小专家监督下自动探索目标空间。我们的方法结合了基于扩散的轨迹生成与强化学习，其中后者用于跟踪由扩散模型生成的一系列目标条件轨迹，以实现多种运动操控技能。通过广泛的消融研究和与最先进方法的比较，我们展示了该框架在多个人形运动操控技能上的有效性。
View on arXiv Download PDF AI Translation
cs.RO / 34 / 2606.26858

PlanRL: A Trajectory Planning Architecture for Reinforcement Learning-based Driving Experts

PlanRL：一种基于强化学习的驾驶专家轨迹规划架构
Lim, Joonhee, Lee, Yongjae, Shin, Jangho, Kum, Dongsuk
Abstract
Reinforcement learning (RL) has become a prominent framework for developing driving experts in autonomous vehicles. However, most existing RL-based experts are designed to output direct control commands (e.g., throttle, steering), which suffer from a lack of interpretability, high spatial complexity in learning road geometries, and poor compatibility with modern end-to-end planning architectures. To address these limitations, we propose a novel trajectory planning architecture for RL driving experts that integrates an RL policy with a polynomial-based trajectory planner. By employing a Frenet-frame coordinate system, our method simplifies complex road geometries into a curvilinear framework, offering a structured coordinate prior that facilitates policy learning. Furthermore, we incorporate a kinematic feasibility check into the planning stage to ensure that generated trajectories remain within the vehicle's physical limits, effectively mitigating cumulative tracking errors typically found in planning-based systems. We evaluate our approach on key CARLA benchmarks, where it significantly outperforms existing state-of-the-art control-based RL experts. On the CARLA Offline Leaderboard v1 and NoCrash benchmarks, our method improves the driving score by 5% and 11%, respectively, and increases the success rate by 8% and 19%.
Chinese Translation
强化学习（Reinforcement Learning, RL）已成为开发自动驾驶车辆驾驶专家的重要框架。然而，现有的大多数基于RL的专家设计为输出直接控制命令（例如，油门、转向），这导致其缺乏可解释性、在学习道路几何形状时具有高空间复杂性，并且与现代端到端规划架构的兼容性较差。为了解决这些局限性，我们提出了一种新颖的轨迹规划架构，旨在将RL策略与基于多项式的轨迹规划器相结合。通过采用Frenet坐标系，我们的方法将复杂的道路几何形状简化为曲线框架，提供了一种结构化的坐标先验，促进了策略学习。此外，我们在规划阶段中引入了运动学可行性检查，以确保生成的轨迹保持在车辆的物理限制内，有效减轻了基于规划的系统中通常存在的累积跟踪误差。我们在关键的CARLA基准测试中评估了我们的方法，结果显示其显著优于现有的最先进的基于控制的RL专家。在CARLA离线排行榜v1和NoCrash基准测试中，我们的方法分别提高了驾驶得分5%和11%，成功率提高了8%和19%。
View on arXiv Download PDF AI Translation
cs.RO / 35 / 2606.26922

Risk-Aware Selective Multimodal Driver Monitoring with Driver-State World Modeling

风险感知的选择性多模态驾驶员监测与驾驶员状态世界建模
Qiu, Daosheng, Chi, Haozhuang, Su, Hao, Long, Shu, Miao, Xinyue, Dong, Yongle, Zhang, Wei
Abstract
Continuous driver monitoring in automated vehicles requires low-latency inference while avoiding unsafe decisions under uncertain driver states. Large vision-language models provide broad multimodal priors, but their latency and limited reliability in this setting make them unsuitable as always-on in-cabin monitors. We propose a cost-aware selective inference framework for deployable multimodal driver monitoring. The core system is a lightweight RGB-physiological student that combines in-cabin visual observations with window-level HR/EDA signals, and a learned gate that decides when to accept the fast prediction or abstain for safety intervention. Additional controls show that the learned scores contain sample-level information beyond scenario priors, while exact physiological synchronization remains a limitation. To incorporate predictive evidence, we further study a compact driver-state world modeling module that rolls out latent driver-state features and estimates future fast-model errors and counterfactual system-level action costs. On scenario-induced driver-demand recognition, the RGB-physiological student improves over RGB-only and physiology-only baselines, reaching 0.7440 Macro-F1 and 0.9099 balanced accuracy with 11.39M parameters and 3.08ms inference latency. Cost-aware selective inference reduces unsafe false negatives from 17.37% under always-fast inference to approximately 5% across seeds, while maintaining deployment-level latency. While driver-state world modeling offers valuable predictive signals, worst-group evaluations highlight persistent operating-point calibration drift. Ultimately, reliable edge driver monitoring requires advancing not only perception backbones, but also risk-aware selective control and group-robust calibration.
Chinese Translation
在自动驾驶车辆中，持续的驾驶员监测需要低延迟推理，同时避免在不确定的驾驶员状态下做出不安全的决策。大型视觉-语言模型提供了广泛的多模态先验，但在这种环境下，它们的延迟和有限的可靠性使其不适合作为始终在线的车舱监测器。我们提出了一种成本感知的选择性推理框架，用于可部署的多模态驾驶员监测。核心系统是一个轻量级的RGB-生理学生模型，它结合了车舱内的视觉观察与窗口级的心率（HR）/皮肤电反应（EDA）信号，并通过一个学习的门控机制决定何时接受快速预测或为了安全干预而放弃。额外的控制显示，学习的评分包含了超出场景先验的样本级信息，而精确的生理同步仍然是一个限制。为了结合预测证据，我们进一步研究了一个紧凑的驾驶员状态世界建模模块，该模块展开潜在的驾驶员状态特征，并估计未来快速模型的误差和反事实系统级行动成本。在场景诱导的驾驶员需求识别中，RGB-生理学生模型在RGB-only和生理-only基准上有所提升，达到了0.7440的宏观F1分数和0.9099的平衡准确率，参数量为1139万，推理延迟为3.08毫秒。成本感知的选择性推理将不安全的假阴性从始终快速推理下的17.37%降低到约5%，同时保持部署级延迟。尽管驾驶员状态世界建模提供了有价值的预测信号，但最差组评估突显了持续的操作点校准漂移。最终，可靠的边缘驾驶员监测不仅需要提升感知骨干网络，还需要推进风险感知的选择性控制和群体鲁棒校准。
View on arXiv Download PDF AI Translation
cs.RO / 36 / 2606.26928

UAV-MapFusion: RTK-Aligned Uncertainty-Aware Coarse-to-Fine Multi-Session UAV Mapping

UAV-MapFusion：RTK对齐的不确定性感知粗到细多会话无人机映射
Pan, Feng, Zheng, Chunran, Xue, Bing, Cui, Yukang, Wen, Jiayu, Chen, Zhiyu, Wang, Wei
Abstract
Large-scale point cloud maps are essential for robotics and spatial intelligence tasks. UAVs provide an efficient means for large-scale map acquisition; however, due to limited flight endurance and onboard storage, mapping a large-scale scene within a single flight remains difficult. Existing multi-session map merging methods can extend the mapping range, yet in UAV scenarios they still struggle to simultaneously suppress long-range drift and preserve local geometric accuracy. To address this issue, an uncertainty-aware multi-session point cloud map merging and coarse-to-fine optimization system is proposed. The proposed method first performs initial multi-session map merging based on a scene graph, and then incorporates RTK observations through an RTK spatiotemporal alignment module, where temporal offsets are estimated using Dynamic Time Warping (DTW), and continuous RTK constraints are recovered using Multi-Output Gaussian Processes (MOGP) under incomplete sampling and frame dropouts. On this basis, a unified uncertainty-aware factor graph is constructed, and local geometric accuracy is further improved through iterative plane-factor refinement. Experiments on real-world datasets validate the effectiveness and robustness of the proposed method. To facilitate further research and development in the community, our code and dataset will be publicly released.
Chinese Translation
大规模点云地图对于机器人技术和空间智能任务至关重要。无人机（UAV）提供了一种高效的大规模地图获取手段；然而，由于飞行耐力和机载存储的限制，在单次飞行中映射大规模场景仍然困难。现有的多会话地图合并方法可以扩展映射范围，但在无人机场景中，它们仍然难以同时抑制长距离漂移并保持局部几何精度。为了解决这个问题，提出了一种不确定性感知的多会话点云地图合并及粗到细优化系统。该方法首先基于场景图执行初步的多会话地图合并，然后通过RTK时空对齐模块引入RTK观测，其中时间偏移使用动态时间规整（Dynamic Time Warping, DTW）进行估计，并在不完全采样和帧丢失的情况下使用多输出高斯过程（Multi-Output Gaussian Processes, MOGP）恢复连续的RTK约束。在此基础上，构建了一个统一的不确定性感知因子图，并通过迭代平面因子细化进一步提高局部几何精度。对真实世界数据集的实验验证了所提方法的有效性和鲁棒性。为了促进社区的进一步研究和开发，我们的代码和数据集将公开发布。
View on arXiv Download PDF AI Translation
cs.RO / 37 / 2606.26955

RobOralScan: Learning Active Intraoral Scanning for Robotic Dental Reconstruction

RobOralScan：用于机器人牙科重建的主动口内扫描学习
Lee, Jinhyung, Yun, Haeun, Kim, Siwon, Baek, Gihyun, Moon, Sungho, Hwang, Sehyun, Im, Sunghoon
Abstract
Intraoral scanning is widely used for digital optical impressions in prosthodontic, implant, and orthodontic treatment, but full-arch and long-span scanning remain labor-intensive tasks with limited automation. In the confined oral cavity, operators must continuously adjust scanner motion while accumulating narrow field-of-view observations, making reconstruction quality sensitive to missing tooth surfaces and operator workload. We propose RobOralScan, which, to the best of our knowledge, is the first reinforcement learning (RL)-based pipeline for robotic automatic intraoral scanning. RobOralScan introduces a geometric memory-based observation space that accumulates partial scan observations into a tri-state geometric representation, allowing the policy to reason over scan history and insufficiently observed regions. It further introduces tooth-wise coverage learning, combining coverage-aware reward signals and a progressive training scheme to improve global reconstruction coverage while reducing uneven coverage across individual teeth. The learned policy selects relative scanner motions from accumulated geometric memory and robot proprioception for closed-loop scan control within the oral workspace. RobOralScan achieves a Chamfer Distance of 0.00838, an average coverage of 92.58%, a lower-tail per-tooth coverage of 88.45%, and a normalized AUC of 0.6674, completing the scan criterion in 8 of 10 evaluation episodes. Furthermore, zero-shot sim-to-real experiments demonstrate its practical feasibility on a physical robot-scanner setup.
Chinese Translation
口内扫描广泛应用于义齿、植牙和正畸治疗中的数字光学印模，但全拱和长跨度扫描仍然是劳动密集型任务，自动化程度有限。在狭小的口腔环境中，操作员必须不断调整扫描仪的运动，同时积累狭窄视野的观察数据，这使得重建质量对缺失的牙齿表面和操作员的工作负荷非常敏感。我们提出了RobOralScan，据我们所知，这是首个基于强化学习（RL）的机器人自动口内扫描管道。RobOralScan引入了一种基于几何记忆的观察空间，将部分扫描观察数据累积为三状态几何表示，使得策略能够对扫描历史和观察不足的区域进行推理。它进一步引入了牙齿覆盖学习，结合了关注覆盖的奖励信号和渐进训练方案，以提高全局重建覆盖率，同时减少各个牙齿之间的不均匀覆盖。学习到的策略从累积的几何记忆和机器人本体感知中选择相对的扫描仪运动，以实现口腔工作空间内的闭环扫描控制。RobOralScan实现了0.00838的Chamfer距离、92.58%的平均覆盖率、88.45%的每牙覆盖率下限，以及0.6674的标准化AUC，在10个评估回合中完成了8个扫描标准。此外，零样本模拟到现实的实验展示了其在物理机器人扫描仪设置中的实际可行性。
View on arXiv Download PDF AI Translation
cs.RO / 38 / 2606.26981

In-Context Model Predictive Generation: Open-Vocabulary Motion Synthesis from Language Models to Physics

上下文模型预测生成：从语言模型到物理的开放词汇运动合成
Fu, Xiaomeng, Lin, Junfan, Liu, Yang, Wang, Yaowei, Li, Guanbin, Lin, Liang, Chen, Ziliang
Abstract
Synthesizing human motion from textual descriptions is essential for immersive digital applications, yet existing methods face a persistent trade-off between semantic fidelity and physical realism. Large language model (LLM)-based approaches can interpret diverse open-vocabulary instructions and compose high-level action plans, but they often generate motions that violate physical constraints. Physics-aware models improve realism through simulation or control, but they struggle with semantic complexity, fine-grained instructions, and novel concepts. To address this gap, we propose In-Context Model Predictive Generation (ICMPG), a framework that integrates language-model planning with inference-time physical feedback. ICMPG reformulates motion synthesis as a Model Predictive Control (MPC)-like process with two modules. The Context-Aware Motion Generation (CAMG) module uses an LLM as a planner to decompose textual commands and generate candidate motion sequences from motion tokens. The Model Predictive Generation (MPG) module evaluates these candidates through physical simulation and semantic alignment, estimates a composite reward, and selects the best sequence to guide subsequent generation steps. Unlike open-loop generation, this closed-loop refinement enables ICMPG to adapt motions to both the input semantics and the simulated physical environment without task-specific policy retraining. Extensive experiments across standard and zero-shot open-vocabulary settings show that ICMPG generalizes robustly to diverse commands and produces motions that are more physically plausible and semantically faithful than representative baselines on the evaluated benchmarks. The framework bridges semantic interpretation and physical simulation while remaining flexible enough to incorporate different LLM backbones, enabling more versatile and controllable text-driven motion synthesis.
Chinese Translation
从文本描述合成人类运动对于沉浸式数字应用至关重要，但现有方法在语义忠实性与物理现实性之间面临持续的权衡。基于大型语言模型（LLM）的方法能够解读多样的开放词汇指令并构建高层次的行动计划，但它们往往生成违反物理约束的运动。具有物理感知的模型通过模拟或控制提高了现实性，但在处理语义复杂性、细粒度指令和新概念时表现不佳。为了解决这一问题，我们提出了上下文模型预测生成（ICMPG），这是一个将语言模型规划与推理时物理反馈相结合的框架。ICMPG将运动合成重新表述为类似模型预测控制（MPC）的过程，包含两个模块。上下文感知运动生成（CAMG）模块使用LLM作为规划者，分解文本命令并从运动标记生成候选运动序列。模型预测生成（MPG）模块通过物理模拟和语义对齐评估这些候选，估计复合奖励，并选择最佳序列以指导后续生成步骤。与开放循环生成不同，这种闭环优化使得ICMPG能够根据输入语义和模拟的物理环境调整运动，而无需针对特定任务的策略重训练。在标准和零样本开放词汇设置下的大量实验表明，ICMPG对多样化指令具有强大的泛化能力，并在评估基准上生成的运动在物理上更具可信性且在语义上更为忠实，超越了代表性基线。该框架在语义解释与物理模拟之间架起了桥梁，同时保持足够的灵活性以整合不同的LLM骨干，从而实现更为多样化和可控的文本驱动运动合成。
View on arXiv Download PDF AI Translation
cs.RO / 39 / 2606.27036

RelAfford6D: Relational 6D Affordance Graphs for Constraint-Driven Robotic Manipulation

RelAfford6D：用于约束驱动机器人操作的关系6D可供性图
Zhang, Guodong, He, Qichen, Xie, Wenyuan, Wu, Shaokai, Ji, Yanbiao, Li, Qiuchang, Bayramli, Bayram, Ding, Yue, Lu, Hongtao
Abstract
Bridging abstract semantics and precise physical control remains a fundamental challenge in open-world robotic manipulation. While recent data-driven policies show promise, their reliance on isolated contact points or latent affordance embeddings lacks the rigorous kinematic constraints necessary for complex articulated objects.To overcome the limitation, we introduce RelAfford6D, a novel training-free framework centered on a Relational 6D Affordance Graph. Given a free-form instruction, our system deduces a semantic topology linking a primary interacting part to its physical anchor. By elevating these topological nodes into precise metric $SE(3)$ poses via vision foundation models, we analytically formulate downstream execution as a kinematic constraint satisfaction problem. The robot synthesizes continuous trajectories by tracking strictly defined physical manifolds (e.g., revolute or prismatic orbits). Coupled with a closed-loop tracking mechanism for dynamic replanning against disturbances, our physically grounded approach achieves superior zero-shot success rates, cross-category generalization and execution robustness in both simulation and the real world environments, outperforming existing data-driven baselines.
Chinese Translation
在开放世界的机器人操作中，弥合抽象语义与精确物理控制之间的差距仍然是一个基本挑战。尽管最近的数据驱动策略展现出希望，但它们对孤立接触点或潜在可供性嵌入的依赖缺乏复杂关节物体所需的严格运动学约束。为了解决这一限制，我们提出了RelAfford6D，这是一种以关系6D可供性图为中心的新型无训练框架。给定一个自由形式的指令，我们的系统推导出一个语义拓扑，将主要交互部分与其物理锚点连接起来。通过视觉基础模型将这些拓扑节点提升为精确的度量$SE(3)$姿态，我们将下游执行分析性地表述为一个运动学约束满足问题。机器人通过跟踪严格定义的物理流形（例如，旋转或棱柱轨道）合成连续轨迹。结合一个用于动态重新规划的闭环跟踪机制，我们的物理基础方法在模拟和现实环境中实现了优越的零-shot成功率、跨类别泛化和执行鲁棒性，超越了现有的数据驱动基线。
View on arXiv Download PDF AI Translation
cs.RO / 40 / 2606.27079

ForesightSafety-VLA: A Unified Diagnostic Safety Benchmark for Vision-Language-Action Models

ForesightSafety-VLA：一种统一的视觉-语言-动作模型诊断安全基准
Lyu, Mingyang, Sun, Yinqian, Jia, Yiyang, Shen, Sicheng, Sha, Moquan, Li, Huangrui, Zhao, Feifei, Zeng, Yi
Abstract
In embodied intelligence, safety is a prerequisite for reliable robot deployment in the physical world. Current vision-language-action (VLA) models continue to advance toward general-purpose task capability, yet their embodied safety limits remain poorly understood. To address this gap, we introduce ForesightSafety-VLA, a diagnostic benchmark that makes safety the primary evaluation target for VLA systems. We define a 13-category safety taxonomy covering physical interaction safety (Safe-Core), instruction-side safety (Safe-Lang), and perception-side safety (Safe-Vis), and evaluate policies under three controlled dimensions of variation -- scene structure, language command, and visual observation -- so that failure sources can be diagnosed rather than hidden in a single aggregate score. Beyond binary task success, ForesightSafety-VLA measures process-level risk through cumulative safety cost (CC) and risk exposure time (RET), together with a four-quadrant decomposition of safe/unsafe success and failure. We instantiate 66 safety-augmented base scenarios in RoboTwin across 5 embodiments and report results on representative VLA baselines. Across the evaluated baselines, even the strongest policy incurs non-trivial safety cost and unsafe nominal success, while structure and visual variation induce substantially stronger safety degradation than ordinary language variation. These results suggest that embodied safety is tightly coupled to perception, grounding, and control competence rather than being reducible to post-hoc safety filtering alone.
Chinese Translation
在具身智能中，安全是可靠机器人在物理世界中部署的前提。目前的视觉-语言-动作（VLA）模型在朝着通用任务能力不断进步，但其具身安全的限制仍然不够清晰。为了解决这一问题，我们引入了ForesightSafety-VLA，一个将安全作为VLA系统主要评估目标的诊断基准。我们定义了一个涵盖物理交互安全（Safe-Core）、指令侧安全（Safe-Lang）和感知侧安全（Safe-Vis）的13类安全分类法，并在场景结构、语言指令和视觉观察这三个受控变化维度下评估策略，以便能够诊断失败来源，而不是将其隐藏在单一的综合评分中。除了二元任务成功外，ForesightSafety-VLA还通过累积安全成本（CC）和风险暴露时间（RET）来衡量过程级别的风险，并结合安全/不安全成功与失败的四象限分解。在RoboTwin中，我们实例化了66个增强安全的基础场景，涵盖5种具身形式，并报告了代表性VLA基准的结果。在评估的基准中，即使是最强的策略也会产生非平凡的安全成本和不安全的名义成功，而结构和视觉变化引起的安全降级明显强于普通语言变化。这些结果表明，具身安全与感知、基础和控制能力紧密相关，而不仅仅是后期安全过滤所能简化的。
View on arXiv Download PDF AI Translation
cs.RO / 41 / 2606.27123

Proposal-Conditioned Latent Diffusion for Closed-Loop Traffic Scenario Generation

基于提案条件的潜在扩散用于闭环交通场景生成
Phoolari, Shubham Vaijanath, Kara, Aleyna, Lauer, Christoph, Peters, Steven
Abstract
Closed-loop traffic simulation remains challenging because it must generate interactive multi-agent behaviors that are scene-consistent and controllable throughout rollout. Prior diffusion-based approaches achieve strong realism, but their computational cost can hinder deployment in time-constrained replanning loops for autonomous vehicle planning and simulation. We present a diffusion-based scenario generation framework conditioned on instance-centric scene context and multimodal proposal priors, with optional test-time guidance for shaping safety-critical behaviors. A compact action-latent representation and proposal-based initialization improve sampling efficiency and reduce per-step runtime without retraining. Experiments on the Waymo Open Motion Dataset demonstrate a favorable balance among realism, safety, and controllability across diverse interactive scenarios, while showing that test-time guidance enables systematic trade-offs among competing objectives.
Chinese Translation
闭环交通模拟仍然面临挑战，因为它必须生成在整个展开过程中场景一致且可控的交互多智能体行为。先前的基于扩散的方法实现了强大的真实感，但其计算成本可能会阻碍在时间受限的自主车辆规划和模拟中的部署。我们提出了一种基于扩散的场景生成框架，该框架以实例中心的场景上下文和多模态提案先验为条件，并提供可选的测试时引导，以塑造安全关键行为。紧凑的动作潜在表示和基于提案的初始化提高了采样效率，并在不重新训练的情况下减少了每步的运行时间。在Waymo开放运动数据集上的实验表明，在各种交互场景中，真实感、安全性和可控性之间取得了良好的平衡，同时显示测试时引导能够在竞争目标之间实现系统性的权衡。
View on arXiv Download PDF AI Translation
cs.RO / 42 / 2606.27144

PAMAE: Phase-Aware-MoE Action Experts Towards Reliable Flow-Matching Vision-Language-Action Policies

PAMAE：面向可靠流匹配视觉-语言-动作策略的阶段感知混合专家行动模块
Yang, Jiayu, Yang, Tao, Chang, Xiang, Chao, Fei, Shang, Changjing, Shen, Qiang
Abstract
Reliable action generation for multi-stage robotic manipulation remains challenging for Vision-Language-Action (VLA) models. While existing flow-matching VLA policies offer strong multimodal grounding and generalization, they typically employ a single shared action expert, limiting their ability to capture phase-specific control patterns across distinct execution stages. We propose a plug-and-play Phase-Aware Mixture-of-Experts Action Module (PAMAE), as a step towards more reliable phase-consistent action generation. PAMAE replaces the original flow-matching action expert with a sparse expert mixture while preserving the pretrained VLA backbone. PAMAE introduces a phase-aware router that leverages execution-phase cues to allocate action generation across experts, supported by a lightweight phase prediction head and a routing alignment objective. To stabilize specialization, we adopt a two-stage training scheme that first warms up the expert module under the standard flow-matching loss and then optimizes phase-consistent routing under auxiliary supervision. On multi-stage manipulation simulation tasks, PAMAE improves task success by up to \textbf{9.2\%} over strong VLA baselines. Further ablations show that both phase-supervised routing and staged optimization are essential for the observed gains. Our results highlight phase-consistent expert allocation as an effective mechanism for improving the reliability and action quality of flow-matching VLA policies.
Chinese Translation
多阶段机器人操作的可靠动作生成对于视觉-语言-动作（VLA）模型仍然具有挑战性。尽管现有的流匹配VLA策略提供了强大的多模态基础和泛化能力，但它们通常采用单一共享的动作专家，这限制了它们在不同执行阶段捕捉阶段特定控制模式的能力。我们提出了一种即插即用的阶段感知混合专家行动模块（PAMAE），作为实现更可靠的阶段一致性动作生成的一步。PAMAE用稀疏专家混合替代了原有的流匹配动作专家，同时保留了预训练的VLA骨干网络。PAMAE引入了一种阶段感知路由器，利用执行阶段线索在专家之间分配动作生成，并辅以轻量级的阶段预测头和路由对齐目标。为了稳定专业化，我们采用了两阶段训练方案，首先在标准流匹配损失下对专家模块进行预热，然后在辅助监督下优化阶段一致的路由。在多阶段操作仿真任务中，PAMAE的任务成功率比强VLA基线提高了高达9.2%。进一步的消融实验表明，阶段监督路由和分阶段优化对于观察到的增益都是至关重要的。我们的结果强调阶段一致的专家分配作为提高流匹配VLA策略的可靠性和动作质量的有效机制。
View on arXiv Download PDF AI Translation
cs.RO / 43 / 2606.27146

PhysReflect-VLA: Physical Feasibility and Self-Reflective Regulation for Reliable Vision-Language-Action Policies

PhysReflect-VLA：可靠的视觉-语言-动作策略的物理可行性与自我反思调节
Yang, Jiayu, Yang, Tao, Li, Weijun, Chang, Xiang, Chao, Fei, Shang, Changjing, Shen, Qiang
Abstract
Long-horizon robotic manipulation is highly sensitive to physically infeasible transitions, contact-induced disturbances, and the lack of effective self-correction during execution. Although Vision-Language-Action (VLA) models provide strong task grounding through multimodal learning, they typically generate actions in a feed-forward manner without explicitly checking physical feasibility or diagnosing execution errors online. We present PhysReflect-VLA, a plug-and-play execution-time reliability framework that augments VLA policies with physical feasibility evaluation and structured self-reflection in a closed-loop control pipeline. A Feasibility Operator evaluates whether candidate actions induce dynamically consistent state transitions; an Action Explanation Operator verifies transition coherence; and an LLM-based Reflection Module analyzes state discrepancies to generate corrective guidance for subsequent actions. A two-stage training procedure stabilizes feasibility modeling and integrates reflection into the control loop. Experiments on multi-stage, contact-rich real-world manipulation tasks show consistent improvements in stage-wise stability and overall task success compared with representative VLA baselines with an average gain of 5.4\%. Ablation results further indicate that feasibility checking and reflection-based correction both contribute to improved execution robustness. These results highlight the importance of embedding physical consistency checks and online self-reflection for reliable long-horizon robotic manipulation.
Chinese Translation
长时间跨度的机器人操作对物理不可行的过渡、接触引起的干扰以及执行过程中的有效自我修正高度敏感。尽管视觉-语言-动作（VLA）模型通过多模态学习提供了强大的任务基础，但它们通常以前馈方式生成动作，而没有明确检查物理可行性或在线诊断执行错误。我们提出了PhysReflect-VLA，这是一种即插即用的执行时间可靠性框架，通过物理可行性评估和结构化自我反思增强了VLA策略在闭环控制管道中的应用。可行性操作符评估候选动作是否引发动态一致的状态转换；动作解释操作符验证转换的一致性；基于大语言模型（LLM）的反思模块分析状态差异，以生成后续动作的纠正指导。两阶段的训练程序稳定了可行性建模，并将反思整合到控制循环中。在多阶段、接触丰富的现实世界操作任务中的实验表明，与代表性的VLA基线相比，阶段稳定性和整体任务成功率均有一致改善，平均增益为5.4%。消融实验结果进一步表明，可行性检查和基于反思的修正均有助于提高执行的鲁棒性。这些结果强调了在可靠的长时间跨度机器人操作中嵌入物理一致性检查和在线自我反思的重要性。
View on arXiv Download PDF AI Translation
cs.RO / 44 / 2606.27163

Learning to Fold: prizewinning solution at LeHome Challenge 2026 (1st place online, 2nd offline)

学习折叠：2026年LeHome挑战赛获奖解决方案（线上第一，线下第二）
Larchenko, Ilia
Abstract
I describe my solution to the LeHome Challenge 2026, an ICRA 2026 competition on bimanual garment folding. The system placed 1st of 62 teams in the online (simulation) round and 2nd in the real-world final. It improves a vision-language-action (VLA) policy with a reinforcement-learning loop. The policy is its own value function: the same network that predicts actions also predicts success, progress, and a few task-relevant future quantities, and those predictions drive advantage estimation, live failure detection, and candidate selection. The work mostly recombines existing RL ideas with engineering and optimization contributions that can be used together as one recipe or individually: AWR + RECAP combined for flow-matching VLA; an asynchronous distributed training / rollout pipeline through HuggingFace Hub; inference-time hyperparameters optimization via Thompson sampling; a sim-to-real recipe with camera-alignment tooling, heavy augmentation and DAgger-like HIL data collection.
Chinese Translation
我描述了我对2026年LeHome挑战赛的解决方案，该比赛是ICRA 2026关于双手服装折叠的竞赛。该系统在62支队伍中在线（模拟）轮次中获得第一名，在现实世界决赛中获得第二名。它通过强化学习循环改进了视觉-语言-动作（VLA）策略。该策略本身就是其价值函数：同一网络不仅预测动作，还预测成功、进展以及一些与任务相关的未来量，这些预测驱动了优势估计、实时故障检测和候选选择。这项工作主要是将现有的强化学习理念与工程和优化贡献重新组合，可以作为一个整体方案或单独使用：AWR + RECAP结合用于流匹配VLA；通过HuggingFace Hub实现的异步分布式训练/回放管道；通过汤普森采样进行推理时超参数优化；带有相机对齐工具、重度增强和类似DAgger的HIL数据收集的仿真到现实的方案。
View on arXiv Download PDF AI Translation
cs.RO / 45 / 2606.27239

HumanoidUMI: Bridging Robot-Free Demonstrations and Humanoid Whole-Body Manipulation

HumanoidUMI：连接无机器人演示与类人机器人全身操控
Wang, Hongwu, Yu, Chenhao, Hu, Youhao, Zhang, Jiachen, Li, Yuanyuan, Luo, Shaqi
Abstract
High-quality demonstration data are essential for humanoid robot skill learning, especially for whole-body behaviors that require coordinated perception, locomotion, and manipulation. Existing data-collection methods largely rely on robot teleoperation, which is constrained by hardware accessibility, operator expertise, and limited efficiency. Inspired by the Universal Manipulation Interface (UMI), we propose HumanoidUMI, a portable and robot-free framework for humanoid whole-body data collection. HumanoidUMI uses lightweight VR devices and UMI-inspired grippers to collect sparse human keypoint trajectories, wrist-view observations, and gripper actions. These demonstrations train a high-level policy to predict future keypoints, which are retargeted to robot-native whole-body references and executed by a whole-body controller. Experiments in five real-world scenarios demonstrate the effectiveness of the proposed framework and validate the collected demonstrations for transferable humanoid whole-body skill learning.
Chinese Translation
高质量的演示数据对于类人机器人技能学习至关重要，尤其是对于需要协调感知、运动和操控的全身行为。现有的数据收集方法主要依赖于机器人遥操作，这受到硬件可及性、操作员专业知识和效率限制的制约。受到通用操控接口（Universal Manipulation Interface, UMI）的启发，我们提出了HumanoidUMI，一个便携式且无机器人框架，用于类人机器人全身数据收集。HumanoidUMI使用轻量级虚拟现实设备和受UMI启发的抓手来收集稀疏的人体关键点轨迹、手腕视角观察和抓手动作。这些演示用于训练高层策略，以预测未来的关键点，这些关键点被重新定向到机器人原生的全身参考，并由全身控制器执行。在五个真实场景中的实验验证了所提框架的有效性，并验证了收集的演示在可转移的类人机器人全身技能学习中的应用。
View on arXiv Download PDF AI Translation
cs.RO / 46 / 2606.27251

Advancing Omnimodal Embodied Agents from Isolated Skills to Everyday Physical Autonomy

从孤立技能到日常物理自主性的全模态具身智能体的进展
Shi, Junhao, Huai, Zezheng, Wang, Siyin, Chen, Jia, Wang, Yubang, Fei, Zhaoye, Chen, Hechang, Gong, Jingjing, Qiu, Xipeng, Jiang, Yu-Gang
Abstract
Building persistent embodied agents in unstructured environments demands unified orchestration of heterogeneous tools spanning both cyber (APIs, IoT) and physical (manipulation, navigation) domains, coupled with autonomous recovery from physical failures that inevitably arise over extended operation. Existing systems treat these as separate problems: VLM-based planners lack a unified cyber-physical action space, agent frameworks accumulate unbounded context that degrades temporal coherence, and VLA policies execute open-loop without detecting their own failures. We argue that persistent autonomy requires not a monolithic model but a hierarchical asynchronous architecture with explicit separation of planning, memory, and verification. To this end, we present OmniAct, a framework integrating a multimodal semantic planner for skill routing across unified action spaces, an adaptive hierarchical memory with event-boundary-driven compression for sub-linear context growth, and an asynchronous visual preemption engine that closes the semantic loop during physical execution. Across 40 real-world long-horizon tasks on two robotic platforms coordinating four IoT devices, OmniAct achieves consistent improvements in end-to-end success across all complexity levels, maintains near-flat token consumption over under 100k+ accumulated interaction tokens, and elevates mid-scale open-weight models to proprietary-level performance.
Chinese Translation
在非结构化环境中构建持久的具身智能体需要对跨越网络（API、物联网）和物理（操作、导航）领域的异构工具进行统一协调，并能够在长时间操作中自主恢复不可避免的物理故障。现有系统将这些问题视为独立的：基于视觉语言模型（VLM）的规划器缺乏统一的网络-物理行动空间，智能体框架积累了无限的上下文，导致时间一致性下降，而视觉语言代理（VLA）策略则在开放回路中执行，无法检测自身的失败。我们认为，持久自主性需要的不是单一模型，而是一个具有明确规划、记忆和验证分离的分层异步架构。为此，我们提出了OmniAct，一个框架，集成了用于在统一行动空间中进行技能路由的多模态语义规划器、具有事件边界驱动压缩的自适应分层记忆以实现亚线性上下文增长，以及在物理执行过程中闭合语义循环的异步视觉抢占引擎。在两个机器人平台上协调四个物联网设备的40个真实世界长时间任务中，OmniAct在所有复杂度级别上实现了端到端成功的一致性提升，在累计超过10万交互标记的情况下保持近乎平坦的标记消耗，并将中型开放权重模型提升至专有级别的性能。
View on arXiv Download PDF AI Translation
cs.RO / 47 / 2606.27268

E-TTS: A New Embodied Test-Time Scaling Framework for Robotic Manipulation

E-TTS：一种用于机器人操作的新型具身测试时间缩放框架
Ye, Wen, Li, Peiyan, Yuan, Tingyu, Xu, Yuan, Wu, Xiangnan, Zhao, Chaoyang, Liu, Jing, Liu, Nianfeng, Huang, Yan, Wang, Liang
Abstract
Recently, a few works have made early attempts to study test-time scaling for embodied tasks. However, two major challenges remain unsolved: (1) reasoning can effectively improve the performance of the policy, but its scaling mechanism has seldom been studied; (2) historical information is essential, as embodied tasks are inherently long-horizon and sequential, making sole reliance on current observations for action scaling inadequate due to the lack of historical context utilization. To address these challenges, we introduce E-TTS, a modular and plug-and-play Embodied Test-Time Scaling framework that unifies reasoning and action scaling for robotic manipulation via history-aware iterative refinement with vision-language verifiers. To support joint reasoning-action scaling, E-TTS performs reasoning-action joint sampling and scoring in a pairwise manner. To better utilize historical information, E-TTS uses a history buffer to store historical context, which is then used by reasoning and action verifiers to evaluate the sampled candidates. Unlike conventional open-loop TTS methods, E-TTS introduces feedback generation into the sampling process to form a closed-loop iterative refinement mechanism, enhancing both inference efficiency and environmental adaptability. Each component functions as an independent and composable module, allowing flexible and adaptive configuration depending on task requirements. To evaluate the advantages of our framework, we conduct experiments across 4 different benchmarks, 6 environments, 3 embodiments, and 4 base vision-language-action models. The experimental results demonstrate that, without requiring additional expert data collection or retraining, E-TTS consistently improves performance, achieving up to a 33.14% increase in simulation and 26.62% in real-world scenarios.
Chinese Translation
最近，一些研究开始尝试研究具身任务的测试时间缩放。然而，仍然存在两个主要挑战尚未解决：（1）推理可以有效提高策略的性能，但其缩放机制鲜有研究；（2）历史信息至关重要，因为具身任务本质上是长时间跨度和顺序的，仅依赖当前观察进行动作缩放由于缺乏历史上下文的利用而显得不足。为了解决这些挑战，我们提出了E-TTS，一个模块化和即插即用的具身测试时间缩放框架，通过具备历史意识的迭代精炼与视觉-语言验证器统一推理和动作缩放。为了支持联合推理-动作缩放，E-TTS以成对的方式进行推理-动作联合采样和评分。为了更好地利用历史信息，E-TTS使用历史缓冲区存储历史上下文，随后由推理和动作验证器利用这些信息来评估采样候选。与传统的开放环路测试时间缩放方法不同，E-TTS在采样过程中引入反馈生成，形成闭环的迭代精炼机制，从而提高推理效率和环境适应性。每个组件作为独立且可组合的模块运作，允许根据任务需求进行灵活和自适应的配置。为了评估我们框架的优势，我们在4个不同基准、6个环境、3种具身形式和4个基础视觉-语言-动作模型上进行了实验。实验结果表明，在不需要额外专家数据收集或重新训练的情况下，E-TTS始终提高性能，在模拟场景中实现高达33.14%的提升，在现实场景中实现26.62%的提升。
View on arXiv Download PDF AI Translation
cs.RO / 48 / 2606.27292

BOWConnect: Parallel Bayesian Optimization over Windows with Learned Local Cost Maps for Sample-Efficient Kinodynamic Motion Planning

BOWConnect：基于窗口的并行贝叶斯优化与学习局部成本地图的样本高效运动规划
Raxit, Sourav, Newaz, Abdullah Al Redwan, Fuentes, Jose, Bobadilla, Leonardo
Abstract
This paper presents BOWConnect, a bidirectional parallel kinodynamic motion planner that addresses three fundamental limitations of existing sampling-based methods: sample inefficiency in high-dimensional state spaces, unreliable cost heuristics under dynamic constraints, and poor performance in narrow passage environments. Unlike classical planners that rely on random control sampling and geometric distance heuristics, BOWConnect integrates Bayesian Optimization over Windows (BOW) as a learning-based steering function within a parallel tree-based exploration framework, enabling each worker to learn local cost maps and constraints to guide sampling toward dynamically feasible and collision-free controls. A bidirectional architecture simultaneously grows forward and backward trees from the start and goal regions in parallel threads, with a spatial hashing mechanism enabling fast connection queries and a boundary value problem solver generating kinodynamically consistent bridge trajectories. Extensive evaluations across ten benchmark environments demonstrate that BOWConnect achieves 100\% success while delivering the fastest or near-fastest planning time in complex scenarios, including narrow passages and non-convex spaces where state-of-the-art planners fail or degrade substantially. Real-world deployment on a ground vehicle and a quadrotor confirms real-time planning with no collisions. Videos of real-world and simulated experiments, high-resolution versions of the figures, and the open-source code are available at https://bow-connect.github.io/.
Chinese Translation
本文提出了BOWConnect，一种双向并行的动力学运动规划器，旨在解决现有基于采样的方法的三个基本局限性：在高维状态空间中的样本低效性、在动态约束下不可靠的成本启发式以及在狭窄通道环境中的较差性能。与依赖随机控制采样和几何距离启发式的经典规划器不同，BOWConnect在并行树状探索框架中整合了基于窗口的贝叶斯优化（Bayesian Optimization over Windows, BOW）作为一种基于学习的引导函数，使每个工作者能够学习局部成本地图和约束，从而引导采样朝向动态可行且无碰撞的控制。双向架构在并行线程中同时从起始区域和目标区域向前和向后扩展树，空间哈希机制实现快速连接查询，边界值问题求解器生成动力学一致的桥接轨迹。在十个基准环境中的广泛评估表明，BOWConnect在复杂场景中实现了100%的成功率，并在狭窄通道和非凸空间等情况下提供了最快或接近最快的规划时间，而在这些情况下，最先进的规划器往往失败或显著降级。在地面车辆和四旋翼上的实际部署确认了实时规划且无碰撞。真实世界和模拟实验的视频、高分辨率图形版本以及开源代码可在 https://bow-connect.github.io/ 获取。
View on arXiv Download PDF AI Translation
cs.RO / 49 / 2606.27295

LA4VLA: Learning to Act without Seeing via Language-Action Pretraining

LA4VLA：通过语言-动作预训练学习在无视觉情况下的行动
Lin, Tao, Du, Yuxin, Mao, Yiran, Ye, Zewei, Zhong, Yilei, Cheng, Bing, Wang, Yiming, Liu, Jiting, Tian, Yang, Yan, Junchi, Wu, Feiran, Meng, Zenan, Wei, Hu, Fu, Yuqian, Li, Gen, Zhao, Bo
Abstract
Vision-Language-Action (VLA) models are commonly pretrained on robot demonstrations by jointly mapping visual observations and language instructions to actions. However, dense visual-action supervision can dominate the comparatively sparse language-action signal. As a result, policies may rely on visual shortcuts rather than learn how language conditions action execution, making them sensitive to visual variations. To address this limitation, we propose LA4VLA, a language-action pretraining framework that enables policies to acquire language-conditioned action priors without visual observations. These priors capture reusable manipulation skills shared across tasks and scenes, reducing reliance on scene-specific visual cues. Specifically, LA4VLA decomposes expert demonstration trajectories into atomic action segments and pairs each segment with a corresponding low-level action description. This yields LA4-33K, a dataset of 33K Language-Action (LA) episodes derived entirely from existing demonstrations without additional robot data collection. We further develop LA4VLA-1B, a lightweight 1B-parameter VLA model, and investigate three paradigms for incorporating language-action supervision into VLA learning: LA-only pretraining, sequential LA-to-VLA pretraining, and mixed LA-VLA pretraining. Across simulation and real-world tasks, LA-pretrained policies consistently outperform matched VLA-pretrained counterparts, while combining LA and VLA supervision leads to further gains. In particular, mixed LA-VLA pretraining improves the average success rate of LA4VLA-1B over the no-pretraining baseline by up to 17.8 and 45.0 percentage points in simulation and real-world tasks, respectively. These results establish LA4VLA as an effective and complementary pretraining strategy for building stronger and more robust VLA policies.
Chinese Translation
视觉-语言-动作（VLA）模型通常通过将视觉观察与语言指令共同映射到动作上进行机器人演示的预训练。然而，密集的视觉-动作监督可能会主导相对稀疏的语言-动作信号。因此，策略可能依赖于视觉捷径，而不是学习语言如何影响动作执行，这使得它们对视觉变化敏感。为了解决这一局限性，我们提出了LA4VLA，这是一种语言-动作预训练框架，使策略能够在没有视觉观察的情况下获取语言条件的动作先验。这些先验捕捉了跨任务和场景共享的可重用操作技能，从而减少对特定场景视觉线索的依赖。具体而言，LA4VLA将专家演示轨迹分解为原子动作段，并将每个段与相应的低级动作描述配对。这产生了LA4-33K，一个完全基于现有演示而不需要额外机器人数据收集的33K语言-动作（LA）剧集。我们进一步开发了LA4VLA-1B，一个轻量级的1B参数VLA模型，并研究了三种将语言-动作监督纳入VLA学习的范式：仅LA预训练、顺序LA到VLA预训练和混合LA-VLA预训练。在模拟和现实世界任务中，经过LA预训练的策略始终优于匹配的VLA预训练策略，而结合LA和VLA监督则带来了进一步的提升。特别是，混合LA-VLA预训练使得LA4VLA-1B在模拟和现实世界任务中相较于无预训练基线的平均成功率分别提高了高达17.8和45.0个百分点。这些结果确立了LA4VLA作为构建更强大和更稳健的VLA策略的有效且互补的预训练策略。
View on arXiv Download PDF AI Translation
cs.RO / 50 / 2606.27344

VibeAct: Vibration to Actions for Contact-Rich Reactive Robot Dexterity

VibeAct：从振动到动作的接触丰富反应机器人灵巧性
Mao, Yuemin, Yoo, Uksang, Oh, Jean, Francis, Jonathan, Ichnowski, Jeffrey
Abstract
Dexterous manipulation depends on contact events that are fast, local, and often visually occluded. Piezoelectric microphones offer a compact and high-bandwidth way to sense these interactions, but the resulting vibro-acoustic signals are difficult to simulate faithfully enough for end-to-end sim-to-real policy learning on dexterous robot hands. We propose VibeAct, a framework that bridges real vibrotactile sensing and simulation-based reinforcement learning through a shared physical representation of contact and slip. In the real world, we embed piezoelectric microphones into a dexterous robot hand and collect vibro-acoustic data through teleoperation, then replay the recordings in a calibrated digital clone to automatically label per-finger contact and slip. A tactile estimator learns to predict contact and slip from real microphone waveforms, while manipulation policies are trained in simulation on the same representation computed directly from simulated contacts. This decoupling lets policies exploit rapid tactile feedback without simulating raw audio. Across five contact-rich tasks spanning regrasping, in-hand reorientation, and insertion, VibeAct consistently outperforms a proprioception-and-point-cloud baseline in simulation, with the largest gains on tasks requiring sustained reactive control, where the continuous slip-magnitude channel proves the most informative observation. The learned policies transfer to a physical dexterous hand-arm platform, improving success rates on deployed tasks. Project videos and additional details are at https://vibeact.github.io/.
Chinese Translation
灵巧操作依赖于快速、局部且常常被视觉遮挡的接触事件。压电麦克风提供了一种紧凑且高带宽的方式来感知这些交互，但由此产生的振动声学信号难以忠实地模拟，以便用于灵巧机器人手的端到端仿真到现实的策略学习。我们提出了VibeAct，一个通过共享的接触和滑动的物理表示，连接真实的振动触觉感知和基于仿真的强化学习的框架。在现实世界中，我们将压电麦克风嵌入灵巧机器人手中，通过遥操作收集振动声学数据，然后在经过校准的数字克隆中重放录音，以自动标记每根手指的接触和滑动。一个触觉估计器学习从真实麦克风波形中预测接触和滑动，而操作策略则在仿真中基于直接从模拟接触计算的相同表示进行训练。这种解耦使得策略能够利用快速的触觉反馈，而无需模拟原始音频。在五个接触丰富的任务中，包括重新抓取、手内重新定向和插入，VibeAct在仿真中始终优于基于本体感知和点云的基线，尤其在需要持续反应控制的任务中，滑动幅度通道提供了最有价值的观察。学习到的策略能够迁移到物理灵巧手臂平台上，提高了部署任务的成功率。项目视频和更多细节请访问 https://vibeact.github.io/。
View on arXiv Download PDF AI Translation
cs.RO / 51 / 2606.27348

Bridging Performance and Generalization in Reinforcement Learning for Agile Flight

在灵活飞行中的强化学习性能与泛化的桥接
Green, Jonathan, Xing, Jiaxu, Messikommer, Nico, Romero, Angel, Scaramuzza, Davide
Abstract
Autonomous drone racing is a fundamentally challenging regime for autonomous aerial robots, requiring time-optimal control while operating under persistent actuation saturation. While reinforcement learning (RL) has achieved human-level performance in this domain, current methods fail to generalize; policies trained on specific environments often crash immediately in unseen configurations. This failure reflects the intrinsic difficulty of zero-shot generalization in agile flight, arising from high-dimensional task variation and the tight coupling between safety and performance at high speeds. Existing approaches that improve generalization impose a substantial cost on flight speed: control policies must significantly degrade performance to achieve even modest levels of generalization. In this work, we propose a framework for zero-shot generalization in agile flight for RL-based drone racing. By combining task-aware switching based on learning progress with a physically informed procedural track generator, the framework produces a fast and robust generalist policy without test-time adaptation. Our method achieves strong zero-shot performance across a wide range of unseen racetracks in the real world, demonstrating a 7.4x improvement in generalization over the state-of-the-art approaches, while maintaining competitive racing speeds. We validate our method's results in both simulation and real-world settings, including a challenging vision-based, end-to-end control setting that operates without explicit state estimation, where all prior approaches fail to generalize.
Chinese Translation
自主无人机竞速是自主空中机器人面临的一个根本性挑战，需要在持续的驱动饱和下实现时间最优控制。尽管强化学习（RL）在这一领域已达到了人类水平的表现，但当前的方法在泛化能力上存在不足；在特定环境中训练的策略在未见过的配置中往往会立即崩溃。这一失败反映了灵活飞行中零-shot 泛化的内在困难，源于高维任务变异和在高速下安全与性能之间的紧密耦合。现有的提高泛化能力的方法对飞行速度造成了显著影响：控制策略必须显著降低性能才能实现即使是适度的泛化。在本研究中，我们提出了一个用于RL基础无人机竞速的灵活飞行零-shot 泛化框架。通过结合基于学习进度的任务感知切换与物理信息驱动的程序化赛道生成器，该框架在不进行测试时适应的情况下生成快速且稳健的通用策略。我们的方法在现实世界中对一系列未见过的赛道实现了强大的零-shot 性能，相比于最先进的方法，泛化能力提高了7.4倍，同时保持了竞争性的竞速速度。我们在模拟和现实环境中验证了我们方法的结果，包括一个具有挑战性的基于视觉的端到端控制设置，该设置在没有显式状态估计的情况下运行，而所有先前的方法在此情况下均未能实现泛化。
View on arXiv Download PDF AI Translation
cs.RO / 52 / 2606.27353

Continual Robot Policy Learning via Variational Neural Dynamics

通过变分神经动力学进行持续机器人策略学习
Xing, Jiaxu, Zhu, Zhiyuan, Ren, Yunfan, Geles, Ismail, Zhai, Yifan, Reiter, Rudolf, Scaramuzza, Davide
Abstract
Robots deployed in the real world rarely operate under a single fixed dynamics model: wind changes, payloads vary, batteries drain, contacts shift, and hardware wears. Yet most learning-based controllers are trained once and deployed as if learning were complete. This prevents the robot from using deployment experience to further improve task performance. In this work, we propose a continual learning framework that uses real-world experience to improve robot policies under hidden and recurring dynamics. Our method learns a condition-aware dynamics model from real state-action trajectories by combining an analytical physics prior with a neural residual for unmodeled effects. A recurrent encoder infers the current hidden condition from recent interaction, and this estimate conditions both the residual model and the policy. Policy learning is performed via differentiable simulation using diverse learned dynamics sampled from the latent model. At deployment, these sampled conditions are replaced by conditions inferred online from recent real interaction, allowing the policy to recover recurring dynamics by recognition rather than residual re-fitting. Through extensive simulation studies and real-world experiments, we demonstrate that the framework improves policy performance under diverse unobserved disturbances. On real quadrotor trajectory tracking under changing wind, the policy recovers from recurring disturbances in roughly 1s, about 5x faster than online residual re-fitting. It also reduces large-disturbance hover and tracking errors by 65.7% and 53.3% over the state-of-the-art online adaptation approaches
Chinese Translation
在现实世界中部署的机器人很少在单一固定的动力学模型下运行：风力变化、负载不同、电池耗尽、接触点移动以及硬件磨损。然而，大多数基于学习的控制器仅训练一次，随后就像学习已经完成一样进行部署。这阻止了机器人利用部署经验进一步提高任务性能。在本研究中，我们提出了一种持续学习框架，利用现实世界的经验来改善在隐藏和重复动力学下的机器人策略。我们的方法通过结合分析物理先验和用于未建模效应的神经残差，从真实的状态-动作轨迹中学习条件感知的动力学模型。递归编码器从最近的交互中推断当前的隐藏条件，并将此估计用于条件化残差模型和策略。策略学习通过可微分仿真进行，使用从潜在模型中采样的多样化学习动力学。在部署时，这些采样条件被最近真实交互中在线推断的条件所替代，使得策略能够通过识别而非残差重新拟合来恢复重复的动力学。通过广泛的仿真研究和现实世界实验，我们证明该框架在多种未观察到的干扰下改善了策略性能。在应对变化风力的真实四旋翼轨迹跟踪中，策略在大约1秒内从重复干扰中恢复，速度约为在线残差重新拟合的5倍。同时，它还将大干扰悬停和跟踪误差分别减少了65.7%和53.3%，优于最先进的在线适应方法。
View on arXiv Download PDF AI Translation
cs.RO / 53 / 2606.27355

RouterVLA: Turning Smoke Tests into Supervision for Heterogeneous VLA Selection

RouterVLA：将烟雾测试转化为异构 VLA 选择的监督
Ren, Xingyu, Yi, Chugang, Ma, Ge, Sun, Youran
Abstract
We study whether pre-deployment evaluation rollouts can be reused to supervise policy selection. Robot teams routinely smoke test candidate vision-language-action (VLA) policies, then compress those trials into a global winner. RouterVLA evaluates this idea with outcome-disjoint cross-fitting: recorded probes build a profile for each frozen expert, and a separate trial scores the selected expert without entering its profile. Across 34,752 LIBERO-Plus rollout records, a transparent probe-success rule raises held-out success from 0.4686 to 0.6149, a +14.64pp gain. Under the scalar-only profiles studied here, learned scorers are statistically indistinguishable from this rule, showing that commissioning carries the routing value while extra scalar scorer capacity does not create it. Reusing the scored trial inflates the measured gain by $1.87\times$, so credible ledger routing needs outcome separation; model scaling improves individual policies, while commissioning-aware routing improves the system built from them.
Chinese Translation
我们研究了预部署评估回合是否可以被重用于监督策略选择。机器人团队常规地对候选的视觉-语言-行动（VLA）策略进行烟雾测试，然后将这些试验压缩为一个全球赢家。RouterVLA 通过结果不相交的交叉拟合来评估这一想法：记录的探测器为每个冻结的专家建立一个档案，而单独的试验则在不进入其档案的情况下对所选专家进行评分。在 34,752 个 LIBERO-Plus 回合记录中，一个透明的探测成功规则将保留的成功率从 0.4686 提高到 0.6149，增幅为 +14.64 个百分点。在这里研究的仅标量档案下，学习到的评分器在统计上与该规则无显著区别，表明委托具有路由价值，而额外的标量评分器容量并未创造出这种价值。重用已评分的试验将测量增益膨胀了 $1.87 imes$，因此可信的分类路由需要结果分离；模型扩展改善了单个策略，而关注委托的路由则改善了由这些策略构建的系统。
View on arXiv Download PDF AI Translation
cs.RO / 54 / 2606.27374

World Action Models Enable Continual Imitation Learning with Recurrent Generative Replays

世界行动模型使得持续模仿学习与递归生成重放成为可能
Govind, Manish Kumar, Reilly, Dominick, Patel, Smit, Le, Hieu, Das, Srijan
Abstract
Going beyond predicting robot actions, World Action Models (WAMs) can also generate future visual observations. We build on this generative capability to propose Recurrent Generative Replay (REGEN), a continual imitation learning framework that synthesizes pseudo-replay trajectories, enabling a robot policy to rehearse previously learned tasks without storing their original human demonstrations. During continual adaptation, REGEN recursively queries the WAM to synthesize pseudo-replay trajectories conditioned only on prior task instructions and current-task observations. Experiments in both simulation and real-world manipulation settings show that REGEN reduces catastrophic forgetting by up to $50\%$ relative to sequential fine-tuning, while approaching the performance of privileged experience replay methods that require access to real replay data. Finally, we analyze the factors limiting generated replay, identifying long-horizon visual degradation and action-observation inconsistency as the primary bottlenecks. Our results establish WAMs as a promising foundation for continual robot learning without stored demonstrations.
Chinese Translation
超越对机器人动作的预测，世界行动模型（World Action Models, WAMs）还能够生成未来的视觉观察。我们基于这一生成能力提出了递归生成重放（Recurrent Generative Replay, REGEN），这是一个持续模仿学习框架，能够合成伪重放轨迹，使得机器人策略能够在不存储原始人类示范的情况下，排练之前学习的任务。在持续适应过程中，REGEN 递归地查询 WAM，以仅基于先前任务指令和当前任务观察合成伪重放轨迹。在模拟和真实世界操作环境中的实验表明，REGEN 相较于顺序微调减少了高达 50% 的灾难性遗忘，同时接近需要访问真实重放数据的特权经验重放方法的性能。最后，我们分析了限制生成重放的因素，识别出长时间视角退化和动作-观察不一致是主要瓶颈。我们的结果确立了 WAMs 作为持续机器人学习的有希望基础，而无需存储示范。
View on arXiv Download PDF AI Translation
计算机视觉 (Computer Vision)
92
cs.CV / 1 / 2606.26122

DocArena: Turning Raw Documents into Controllable Training Environments for Document Search Agents

DocArena：将原始文档转化为可控的文档搜索代理训练环境
Wang, Jiamian, Zhang, Ruiyi, Yu, Tong, Shi, Jing, Basu, Samyadeep, Jain, Rajiv, Tao, Zhiqiang, Sun, Tong
Abstract
Recent methods train search agents via reinforcement learning from (question, answer, evidence) tuples without requiring expert trajectories. The tuples serve as the training environment, and whose properties directly shape what search strategies and generalization abilities the agent can develop. While prior works have made encouraging progress in improving training data quality, existing environments remain predominantly text-based and existing approaches can struggle to construct training environments that are controllable, scalable, and account for multimodal data. Given this, we propose DocArena, a fully automated data curation pipeline building on the practical need for multimodal document search and question-answering. It transforms raw document collections into training environments for search agents without any human annotation. The pipeline first structures and indexes documents through MLLM-based visual perception, then profiles and leverage the cross-page information distribution to construct reasoning-intensive QA pairs, as well as performs cascaded quality assurance operations via MLLM. We introduce DocArena-79K with QA pairs from 8,336 documents spanning 16 domains and 49 languages. We further design a Doc-Search agent infrastructure that decouples visual perception from the policy model, allowing text-based LLMs to serve as the reasoning backbone for multimodal document retrieval and QA. Under a unified evaluation framework where only the policy model differs, experiments on six multimodal document scenarios and seven text-based QA benchmarks show that agents trained on DocArena data achieve the best performance on both retrieval accuracy and QA quality. Further analysis on agent search behaviors confirms the effectiveness and controllability of the constructed training environment.
Chinese Translation
近期的方法通过强化学习训练搜索代理，利用（问题、答案、证据）元组，而无需专家轨迹。这些元组作为训练环境，其属性直接影响代理能够发展出的搜索策略和泛化能力。尽管之前的研究在提高训练数据质量方面取得了可喜的进展，但现有环境仍主要基于文本，现有方法在构建可控、可扩展并考虑多模态数据的训练环境方面面临挑战。基于此，我们提出了DocArena，一个完全自动化的数据策划管道，满足多模态文档搜索和问答的实际需求。它将原始文档集合转化为搜索代理的训练环境，而无需任何人工标注。该管道首先通过基于MLLM的视觉感知对文档进行结构化和索引，然后对跨页面信息分布进行分析，构建推理密集型的问答对，并通过MLLM执行级联质量保证操作。我们推出了DocArena-79K，包含来自8,336份文档的问答对，涵盖16个领域和49种语言。此外，我们设计了一个Doc-Search代理基础设施，将视觉感知与策略模型解耦，使基于文本的LLM能够作为多模态文档检索和问答的推理支撑。在一个统一的评估框架下，仅策略模型不同，针对六个多模态文档场景和七个基于文本的问答基准的实验表明，基于DocArena数据训练的代理在检索准确性和问答质量上均表现最佳。对代理搜索行为的进一步分析确认了所构建训练环境的有效性和可控性。
View on arXiv Download PDF AI Translation
cs.CV / 2 / 2606.26165

Predicting Fruit Quality with a Hybrid Machine Learning and Image Processing Approach

采用混合机器学习与图像处理方法预测水果质量
Hashemi, Amir Reza, Amiri, Shahram
Abstract
Fruit spoilage is a significant issue in agriculture, leading to substantial economic losses. Addressing this, our study introduces a hybrid approach combining image processing and deep learning to assess fruit freshness. We developed an image processing algorithm that quantifies spoilage on a scale from 0 (fully fresh) to 100 (fully rotten). Alongside, we trained a convolutional neural network (CNN) to perform binary classification (fresh or rotten) using a large dataset of fruit images. The outcomes of both methods were synthesized using logistic regression to enhance the accuracy of freshness predictions. Subsequently, this logistic regression model was utilized to enable the image processing algorithm to provide binary classification based on its percentage output, thus eliminating the need for the CNN in real-time applications. Our approach, which does not require high computational resources, achieved real-time performance and was validated with over 90% accuracy on a dataset comprising apples and oranges. The primary limitation lies in the requirement for fruits to be isolated on a background that must be either white or transparent, suggesting future improvements could include advanced segmentation models to automate background removal. This study's results highlight the potential of integrating simple image processing techniques with machine learning to provide practical solutions in the agricultural sector.
Chinese Translation
水果腐烂是农业中的一个重要问题，导致了巨大的经济损失。为了解决这一问题，我们的研究提出了一种结合图像处理和深度学习的混合方法来评估水果的新鲜度。我们开发了一种图像处理算法，该算法将腐烂程度量化为从0（完全新鲜）到100（完全腐烂）的尺度。同时，我们训练了一个卷积神经网络（CNN）来使用大量水果图像数据集进行二分类（新鲜或腐烂）。这两种方法的结果通过逻辑回归进行综合，以提高新鲜度预测的准确性。随后，这个逻辑回归模型被用于使图像处理算法能够根据其百分比输出提供二分类，从而在实时应用中消除了对CNN的需求。我们的方法不需要高计算资源，达到了实时性能，并在包含苹果和橙子的数据库上验证了超过90%的准确率。主要的限制在于要求水果在背景上被孤立，背景必须是白色或透明的，这表明未来的改进可以包括先进的分割模型以自动去除背景。本研究的结果突显了将简单的图像处理技术与机器学习相结合，为农业部门提供实用解决方案的潜力。
View on arXiv Download PDF AI Translation
cs.CV / 3 / 2606.26171

LCG: Long-Context Consistent Image Generation with Sparse Relational Attention

LCG：基于稀疏关系注意力的长上下文一致性图像生成
Wang, Zihao, Xu, Yijia, Zheng, Haoze, Ma, Xuran, Gui, Haokun, Yang, Harry
Abstract
Recent image generation models achieve impressive quality in single-image synthesis, but often fail to maintain consistency across sequential outputs, as required in comics, storyboards, and visual narratives. We propose Long-Context Generation (LCG), a framework for long-context multi-image text-to-image generation, to improve consistency and scalability in long-context multi-image generation. LCG employs the Sparse Relational Attention (SRA) mechanism to selectively attend to core features across extended visual contexts, ensuring that the propagation of semantic and layout information remains computationally tractable. To enforce semantic alignment, we introduce the Routing Consistency Constraint (RCC), which leverages identity-aware masks to align structural patterns across generation branches, effectively mitigating drift in appearance even in complex multi-character scenes. To support training and evaluation in this setting, we construct the Long-Context Consistency Dataset (LCCD), a large-scale synthetic dataset comprising character-centric multi-image sequences spanning varied situational contexts. LCCD contains 600K training sequences and a separate 1K test set, with each sequence containing 6 to 20 images. The experiments demonstrate that LCG outperforms the compared baselines in prompt alignment and character consistency for long-context image generation, including multi-character scenes.
Chinese Translation
近期的图像生成模型在单图像合成方面取得了令人印象深刻的质量，但在漫画、故事板和视觉叙事等需要序列输出一致性的场景中，往往难以保持一致性。我们提出了长上下文生成（Long-Context Generation, LCG）框架，用于长上下文多图像文本到图像的生成，以提高长上下文多图像生成中的一致性和可扩展性。LCG采用稀疏关系注意力（Sparse Relational Attention, SRA）机制，选择性地关注扩展视觉上下文中的核心特征，确保语义和布局信息的传播在计算上可控。为了强化语义对齐，我们引入了路由一致性约束（Routing Consistency Constraint, RCC），利用身份感知掩码对生成分支中的结构模式进行对齐，有效减轻即使在复杂的多角色场景中外观漂移的问题。为了支持在这种设置下的训练和评估，我们构建了长上下文一致性数据集（Long-Context Consistency Dataset, LCCD），这是一个大规模合成数据集，包含以角色为中心的多图像序列，涵盖多种情境。LCCD包含60万条训练序列和单独的1000条测试集，每条序列包含6到20张图像。实验表明，LCG在长上下文图像生成的提示对齐和角色一致性方面优于比较基线，包括多角色场景。
View on arXiv Download PDF AI Translation
cs.CV / 4 / 2606.26194

Self-Supervised Tree-level Biomass Estimation in Urban Environments From Airborne LiDAR and Optical Observations

基于空中激光雷达和光学观测的城市环境自监督树冠级生物量估计
Bermudez, Jose, Zhong, Zilong, Cyr, Dominic, Sothe, Camile, Gonsamo, Alemu
Abstract
Urban tree biomass remains less spatially explicitly quantified than biomass in managed forests because many estimates rely on inventories or coarse products that cannot resolve individual crowns or fine-scale heterogeneity. We present a crown-level above-ground biomass (AGB) framework for an 810~km$^2$ landscape in Ontario, Canada, using leaf-off airborne LiDAR (8--10~pulses~m$^{-2}$) and near-infrared RGB orthophotography (0.16--0.20~m) from 2018 and 2023. A dual-stream cross-attention network trained on rule-based pseudo-labels produced semantic marks for buildings, needleleaf trees, and deciduous trees, supporting crown delineation and functional-type assignment. On independently annotated withheld tiles, global/mean precision, recall, and Dice scores were 0.86, 0.83, and 0.84. Crowns were delineated with multiscale watershed segmentation in mapped tree areas, and AGB was estimated from a crown area--height power-law proxy calibrated to species-specific allometry (Lambert et al., 2005) for 21,921 inventory trees. For 18,713 inventory--segment matched pairs from a 90,726-tree held-out test set, AGB prediction achieved $R^2=0.609$ using inventory crown geometry and $R^2=0.570$ under operational segmentation, identifying crown delineation as the remaining uncertainty source. Aggregated to 30~m, estimates yielded total AGB stocks of 1.73~Tg in 2018 and 1.81~Tg in 2023 (811--850~Gg~C), local densities up to ${\sim}140$~Mg~ha$^{-1}$ along the Niagara Escarpment, and a net carbon gain of 39~Gg~C over five years. Deep-ensemble uncertainty maps highlighted high-epistemic-uncertainty areas linked to underrepresented land covers and guided assignment of uncertain crowns to a pooled allometric equation. The framework uses standard provincial data, requires no manual annotation, and produces a public bitemporal crown-level AGB database for trees outside forests at management-relevant resolution.
Chinese Translation
城市树木生物量的空间量化程度低于管理森林中的生物量，因为许多估计依赖于无法分辨单个树冠或细尺度异质性的清查或粗略产品。我们提出了一个针对加拿大安大略省810~km²景观的树冠级地上生物量（AGB）框架，使用2018年和2023年的落叶期空中激光雷达（8--10~脉冲~m⁻²）和近红外RGB正射影像（0.16--0.20~m）。基于规则的伪标签训练的双流交叉注意力网络生成了建筑物、针叶树和阔叶树的语义标记，支持树冠的划分和功能类型的分配。在独立注释的保留图块上，全球/平均精度、召回率和Dice分数分别为0.86、0.83和0.84。通过多尺度分水岭分割在映射的树木区域划分树冠，AGB通过与物种特定的异速生长（Lambert et al., 2005）校准的树冠面积-高度幂律代理进行估计，涉及21,921棵清查树。对于来自90,726棵树的持出测试集中的18,713对匹配的清查-分段，AGB预测在使用清查树冠几何时达到了$R²=0.609$，而在操作分割下为$R²=0.570$，识别出树冠划分是剩余的不确定性来源。汇总到30~m，估计的总AGB储量在2018年为1.73~Tg，在2023年为1.81~Tg（811--850~Gg~C），在尼亚加拉悬崖沿线的局部密度高达${ ilde{140}}$~Mg~ha⁻¹，五年内净碳增益为39~Gg~C。深度集成不确定性图突出显示了与代表性不足的土地覆盖相关的高认知不确定性区域，并指导将不确定的树冠分配给汇总的异速生长方程。该框架使用标准省级数据，无需手动注释，并产生了一个公共的双时相树冠级AGB数据库，适用于管理相关分辨率的非森林树木。
View on arXiv Download PDF AI Translation
cs.CV / 5 / 2606.26260

A multi-task spatiotemporal deep neural network for predicting penetration depth and morphology in laser welding

一种多任务时空深度神经网络用于激光焊接中的穿透深度和形态预测
Li, Sen, Cui, Haichao, Shao, Chendong, Wang, Yaqi, Tang, Xinhua
Abstract
In laser penetration welding, the assessment of penetration state and weld seam morphology plays a crucial role in determining the weld quality. This paper presents a comprehensive introduction of the innovative muti-task deep learning model that has the capability to predict penetration state, depth, and weld seam morphology with high accuracy. The monitoring platform relies on weld pool images captured during the laser welding process using a complementary metal-oxide-semiconductor camera. The proposed model integrates spatiotemporal features extracted from top weld pool images along with welding parameters, establishing a deep learning framework based on convolutional neural networks and state space models for more efficient extraction and processing of spatial-temporal information. Furthermore, a reliable method for constructing the dataset is proposed to enhance both robustness and generalization capability of the developed model. Validation results on the test set demonstrate that prediction accuracy for penetration state can reach 99.35%, while prediction error for penetration depth is 1.79 millimeter, and accuracy of reconstructing the weld cross-section is 95.65%. This study provides new insights and methodologies for in-situ quality control strategies in laser penetration welding systems.
Chinese Translation
在激光穿透焊接中，穿透状态和焊缝形态的评估在确定焊接质量方面起着至关重要的作用。本文全面介绍了一种创新的多任务深度学习模型，该模型能够高精度地预测穿透状态、深度和焊缝形态。监测平台依赖于使用互补金属氧化物半导体（CMOS）相机在激光焊接过程中捕获的焊池图像。所提出的模型整合了从焊池顶部图像中提取的时空特征以及焊接参数，建立了一个基于卷积神经网络和状态空间模型的深度学习框架，以更高效地提取和处理时空信息。此外，提出了一种可靠的数据集构建方法，以增强所开发模型的鲁棒性和泛化能力。测试集上的验证结果表明，穿透状态的预测准确率可达到99.35%，而穿透深度的预测误差为1.79毫米，重建焊接横截面的准确率为95.65%。本研究为激光穿透焊接系统中的原位质量控制策略提供了新的见解和方法。
View on arXiv Download PDF AI Translation
cs.CV / 6 / 2606.26279

Beyond Single-Source Cognitive Taskonomy:Multi-Source Task Relations through fMRI Transfer Learning

超越单一来源的认知任务分类：通过fMRI迁移学习探索多来源任务关系
Xia, Junfeng, Li, Wendu, Zhang, Mengjiao, Guo, Jie
Abstract
Cognitive tasks are organized by shared and specialized neural processes. Masked fMRI reconstruction provides a common self-supervised objective for quantifying transfer relations among task states, but existing reconstruction-based taskonomies mainly study one-to-one transfer from a single source task to a target. Here, we extend an fMRI cognitive taskonomy from single-source to multi-source transfer across 23 Human Connectome Project task states and use Boolean Integer Programming (BIP) to analyze budget-constrained task allocation. We train 1,127 task-specific and transfer models. Single-source transfer is directional and paradigm structured: motor states transfer well within the motor paradigm but provide limited support to most non-motor targets, consistent with a shared sensorimotor execution system and effector-specific representations. Multi-source transfer depends on the composition of the source set, suggesting that many-to-one task relations are not fully captured by pairwise taskonomy alone. Across supervision budgets, BIP repeatedly allocates direct supervision to several 0-back and 2-back working-memory states, although these states are not consistently the strongest individual sources. This pattern may reflect the integration of perceptual, attentional, and executive processes in working-memory tasks. Together, these findings reveal a cross-paradigm-limited motor cluster and working-memory states with high priority under the specified global allocation objective. Our study extends reconstruction-based fMRI taskonomy from one-to-one transfer to many-to-one task relations and budget-constrained task dependencies.
Chinese Translation
认知任务通过共享和专门的神经过程进行组织。掩蔽fMRI重建提供了一种共同的自监督目标，用于量化任务状态之间的迁移关系，但现有的基于重建的任务分类主要研究从单一来源任务到目标的单向迁移。在此，我们将fMRI认知任务分类从单一来源扩展到跨23个人类连通组项目任务状态的多来源迁移，并使用布尔整数规划（Boolean Integer Programming, BIP）分析预算约束下的任务分配。我们训练了1,127个任务特定和迁移模型。单一来源迁移是有方向的并且结构化：运动状态在运动范式内迁移良好，但对大多数非运动目标提供有限支持，这与共享的感觉运动执行系统和效应器特定表示一致。多来源迁移依赖于来源集合的组成，表明许多对一的任务关系并不能仅通过成对任务分类完全捕捉。在不同的监督预算下，BIP反复将直接监督分配给几个0-back和2-back工作记忆状态，尽管这些状态并不总是一致地是最强的单一来源。这一模式可能反映了在工作记忆任务中感知、注意和执行过程的整合。总的来说，这些发现揭示了一个跨范式受限的运动集群和在指定的全球分配目标下具有高优先级的工作记忆状态。我们的研究将基于重建的fMRI任务分类从一对一迁移扩展到多对一任务关系和预算约束的任务依赖性。
View on arXiv Download PDF AI Translation
cs.CV / 7 / 2606.26287

GeMoE: Gating Entropy is All You Need for Uncertainty-aware Adaptive Routing in MoE-based Large Vision-Language Models

GeMoE：门控熵是你在基于MoE的大型视觉-语言模型中进行不确定性感知自适应路由所需的一切
Cai, Chaoxiang, Weng, Minghe, Li, Jie, Jiang, Yibo, Yang, Longrong, Qin, Zequn, Li, Xi
Abstract
With the increase in model parameters and training data, the instruction following and generalization capabilities of Large VisionLanguage Models (LVLMs) have been significantly improved. Based on the Mixture of Experts (MoE) architecture, LVLMs expand their parameter capacity while maintaining the inference cost. However, traditional MoE methods employ a Top-k static routing strategy, which fails to account for variations in the input and adaptively select the number of experts, resulting in suboptimal resource utilization. In this paper, we propose viewing token routing as an information encoding task, framing dynamic routing as a Minimum Description Length (MDL) problem in encoding By validating the connection between MDL and gating entropy in the MoE scenario, we introduce Gating Entropy-based Uncertainty-aware Adaptive Routing (GeMoE) for MoE. Unlike traditional static or heuristic-based dynamic routing methods, GeMoE explicitly models the trade-off between model complexity and performance. By using gating entropy to assess the complexity of tokens, GeMoE adaptively determines the number of experts each token should engage. On a wide range of backbones and benchmarks, our method achieves 99.5% average performance retention compared to the original static routing, while improving average expert activation sparsity by 36.5%.
Chinese Translation
随着模型参数和训练数据的增加，大型视觉-语言模型（LVLMs）的指令跟随和泛化能力得到了显著提升。基于专家混合（MoE）架构，LVLMs在保持推理成本的同时扩展了其参数容量。然而，传统的MoE方法采用Top-k静态路由策略，这未能考虑输入的变化并自适应地选择专家数量，导致资源利用不佳。本文提出将令牌路由视为信息编码任务，将动态路由框架化为编码中的最小描述长度（MDL）问题。通过验证MDL与MoE场景中门控熵之间的联系，我们引入了基于门控熵的不确定性感知自适应路由（GeMoE）用于MoE。与传统的静态或启发式动态路由方法不同，GeMoE明确建模了模型复杂性与性能之间的权衡。通过使用门控熵来评估令牌的复杂性，GeMoE自适应地确定每个令牌应参与的专家数量。在多种骨干网络和基准测试中，我们的方法相比于原始静态路由实现了99.5%的平均性能保留，同时将平均专家激活稀疏性提高了36.5%。
View on arXiv Download PDF AI Translation
cs.CV / 8 / 2606.26295

Beyond Aesthetics: Quantifying Information Loss in Turbid Scenes

超越美学：量化浑浊场景中的信息损失
Ismiroglou, Vasiliki, Bengtson, Stefan H., Benos, Tasos, Moeslund, Thomas B., Pedersen, Malte
Abstract
Visibility in underwater environments degrades rapidly under turbid conditions, yet the effects on computer-vision models remain unclear. This issue is compounded by reliance on synthetic turbidity datasets, which may misrepresent real-world information loss. To address this gap, we introduce the Turbid Underwater Baseline (TUB) dataset, comprising 1,320 images captured under extreme turbidity and over 16,000 high-confidence ground-truth segmentation masks. We additionally propose PCD, a metric derived from phase congruency maps that is invariant to contrast and aims to capture the loss of structural information in real turbidity. We show that PCD correlates strongly with the performance of instance segmentation models on both real and synthetic turbid images, whereas common metrics in the field show weak to no correlation at all. The dataset and relevant code can be found on the project page: https://vap.aau.dk/pcd
Chinese Translation
在浑浊条件下，水下环境的能见度迅速下降，但对计算机视觉模型的影响仍不明确。这个问题因依赖于合成浑浊数据集而加剧，这可能会错误地表现出真实世界中的信息损失。为了解决这一空白，我们引入了浑浊水下基准（Turbid Underwater Baseline, TUB）数据集，该数据集包含1,320张在极端浑浊条件下捕获的图像，以及超过16,000个高置信度的真实分割掩码。此外，我们提出了PCD（Phase Congruency Derived），这一度量来源于相位一致性图，对比度不变，旨在捕捉真实浑浊中结构信息的损失。我们展示了PCD与实例分割模型在真实和合成浑浊图像上的性能之间存在强相关性，而该领域的常用度量则表现出弱相关或没有相关性。数据集和相关代码可以在项目页面找到：https://vap.aau.dk/pcd
View on arXiv Download PDF AI Translation
cs.CV / 9 / 2606.26379

Layer-Specific Prompt Fusion Discovery via Differentiable Search in Vision Foundation Models

通过可微搜索在视觉基础模型中发现层特定的提示融合
Xiao, Xi, Li, Xingjian, Zhang, Yunbei, Han, Cheng, Liu, Tianming, Wang, Tianyang, Jiang, Runmin, Hamm, Jihun, Wang, Xiao, Xu, Min
Abstract
Visual prompt tuning has emerged as a parameter-efficient fine-tuning approach for adapting large-scale Vision Transformers (ViTs) to downstream tasks. As its learnable prompts are applied in input and feature spaces, prior to jointly going through attention in transformer layers, the most commonly used scheme for fusing image and prompt tokens is concatenation or addition. In this paper, we aim to study a fundamental yet essential problem in visual prompt tuning: whether a single fusion scheme tends to yield better results, and whether that would be beneficial to develop a hybrid fusion scheme. To this end, we formulate the task as a bi-level optimization problem, and solve it leveraging differentiable architecture search. In this context, the learnable prompts and their fusion schemes are jointly optimized. To enrich the search space in the architecture search, we propose two additional fusion schemes, namely, affine transformation and cross-attention, in addition to concatenation and addition. Extensive experiments on 34 datasets spanning VTAB-1k, FGVC, and HTA show consistent gains over prompt-tuning baselines. With a frozen ViT backbone, our method delivers a favorable accuracy--latency--parameter trade-off compared with VPT-Deep and recent variants. Our findings reveal that how prompts fuse with image tokens plays a significant role in visual prompt tuning, and a hybrid fusion fashion can more effectively leverage layer semantics of ViTs, contributing a novel perspective for visual prompt-tuning research.
Chinese Translation
视觉提示调优作为一种参数高效的微调方法，已成为将大规模视觉变换器（ViTs）适应于下游任务的重要手段。由于其可学习的提示在输入和特征空间中应用，并在变换器层中共同经过注意力机制，最常用的图像与提示标记融合方案是连接或加法。本文旨在研究视觉提示调优中的一个基本而重要的问题：单一的融合方案是否倾向于产生更好的结果，以及这是否有利于开发混合融合方案。为此，我们将该任务表述为一个双层优化问题，并利用可微架构搜索进行求解。在此背景下，可学习的提示及其融合方案被共同优化。为了丰富架构搜索中的搜索空间，我们提出了两种额外的融合方案，即仿射变换（affine transformation）和交叉注意力（cross-attention），以及连接和加法。对涵盖VTAB-1k、FGVC和HTA的34个数据集进行的广泛实验表明，相较于提示调优基线，我们的方法在准确性、延迟和参数之间提供了良好的权衡。通过冻结的ViT骨干网络，我们的方法与VPT-Deep及其近期变体相比，展现了有利的准确性-延迟-参数权衡。我们的研究结果揭示了提示与图像标记的融合方式在视觉提示调优中起着重要作用，混合融合方式能够更有效地利用ViTs的层语义，为视觉提示调优研究提供了新的视角。
View on arXiv Download PDF AI Translation
cs.CV / 10 / 2606.26384

What Do Deepfake Benchmarks Measure? An Audit Using Frozen Self-Supervised Representations

深度伪造基准测试测量了什么？使用冻结的自监督表示进行审计
Pagon, Samuel, Shen, Yixuan, Asnani, Vishal, Liu, Feng
Abstract
As deepfake generators approach perceptual indistinguishability, reliable detection becomes critical. Yet, detectors that score well on benchmarks routinely fail in the wild. A concerning feedback loop has emerged: benchmarks drive increasingly complex, engineered detectors, yet if those benchmarks do not reflect real-world deepfakes, this complexity may be solving the wrong problem entirely. This raises a prior question: what are these benchmarks actually measuring? We conduct an audit of video, image, and audio deepfake benchmarks using a deliberately simple diagnostic. If a linear probe on frozen, general-purpose self-supervised representations can approximate the performance of a bespoke detector, the benchmark is largely rewarding general modality understanding rather than forensic understanding. This has two implications: the benchmark may not reflect realistic threat models, and it raises the question of whether the bespoke detectors the probe approaches are truly learning forensic understanding. We observe, across three modalities, linear probes on general-purpose self-supervised representations closely approach the performance of bespoke detectors. We further show that generator-level difficulty is partly explained by Frechet geometry in the same representation space. Together, these results support a benchmark-audit view of deepfake detection: before high scores are read as evidence of forensic understanding, it is worth asking how much of the benchmark is already solved by general-purpose representations.
Chinese Translation
随着深度伪造生成器接近感知上的不可区分性，可靠的检测变得至关重要。然而，在基准测试中表现良好的检测器在实际应用中常常失败。一个令人担忧的反馈循环已经出现：基准测试推动了越来越复杂的工程检测器，但如果这些基准测试并未反映现实世界中的深度伪造，那么这种复杂性可能完全在解决错误的问题。这引出了一个先前的问题：这些基准测试究竟在测量什么？我们使用一种故意简单的诊断方法，对视频、图像和音频深度伪造基准进行了审计。如果在冻结的通用自监督表示上进行线性探测能够近似定制检测器的性能，那么该基准测试主要奖励的是通用模态理解，而非法医理解。这有两个含义：该基准可能并未反映现实的威胁模型，并且引发了一个问题，即线性探测器接近的定制检测器是否真正学习了法医理解。我们观察到，在三种模态中，通用自监督表示上的线性探测器的性能与定制检测器非常接近。我们进一步表明，生成器级别的难度部分可以通过同一表示空间中的Frechet几何来解释。这些结果共同支持了对深度伪造检测的基准审计观点：在将高分视为法医理解的证据之前，值得询问基准中有多少内容已经被通用表示所解决。
View on arXiv Download PDF AI Translation
cs.CV / 11 / 2606.26387

Staying VIGILant: Mitigating Visual Laziness via Counterfactual Visual Alignment in MLLMs

保持警惕：通过反事实视觉对齐减轻视觉懒惰在多模态大语言模型中的影响
Xiao, Xi, Liu, Chen, Liao, Chih-Ting, Zhang, Yunbei, Lan, Qizhen, Wei, Yuxiang, Zhao, Lin, Wang, Janet, Gu, Jianyang, Ye, Muchao, Wang, Tianyang, Xu, Hao
Abstract
Multimodal large language models (MLLMs) extend large language models (LLMs) with visual perception, enabling joint reasoning over images and text. Despite inheriting strong reasoning capabilities from LLMs, they remain prone to hallucinations that contradict their visual inputs. Mechanistic studies indicate that this weakness stems from visual laziness: MLLMs encode the correct visual evidence internally, but overly rely on strong language priors during response. Existing alignment methods, such as direct preference optimization, primarily optimize outcome-level rewards based on text. This introduces an optimization bias toward linguistic shortcuts, leading to responses that often contradict the visual evidence. To address this, we propose Visual Information Gain In aLignment (VIGIL), a reinforcement-learning (RL) post-training framework that shifts the focus from numerical reward fitting to causal visual grounding. VIGIL introduces a geometric constraint that explicitly maximizes the mutual information between the visual input and the generated response. We achieve this by penalizing "blind confidence" instances where the model remains improperly certain even when textual-visual attention is masked to create a counterfactual blind state. Extensive experiments show that VIGIL consistently outperforms recent alignment methods across hallucination and reasoning benchmarks without compromising text-only capabilities. Our approach matches the full-data performance of state-of-the-art methods using only 25% of the preference data and even demonstrates emergent spatial grounding capabilities without explicit bounding box supervision.
Chinese Translation
多模态大语言模型（MLLMs）在大语言模型（LLMs）的基础上扩展了视觉感知，能够对图像和文本进行联合推理。尽管继承了LLMs的强大推理能力，但它们仍然容易出现与视觉输入相矛盾的幻觉。机制研究表明，这一弱点源于视觉懒惰：MLLMs在内部编码了正确的视觉证据，但在响应时过于依赖强语言先验。现有的对齐方法，如直接偏好优化，主要基于文本优化结果级别的奖励。这引入了对语言捷径的优化偏见，导致响应往往与视觉证据相矛盾。为了解决这一问题，我们提出了视觉信息增益对齐（Visual Information Gain In aLignment，VIGIL），这是一种强化学习（RL）后训练框架，旨在将重点从数值奖励拟合转向因果视觉基础。VIGIL引入了一种几何约束，明确最大化视觉输入与生成响应之间的互信息。我们通过惩罚“盲目自信”实例来实现这一点，即当文本-视觉注意力被屏蔽以创建反事实盲状态时，模型仍然保持不当的确定性。大量实验表明，VIGIL在幻觉和推理基准测试中始终优于最近的对齐方法，而不损害仅基于文本的能力。我们的方法仅使用25%的偏好数据就达到了最先进方法的全数据性能，甚至在没有明确边界框监督的情况下展示了新兴的空间基础能力。
View on arXiv Download PDF AI Translation
cs.CV / 12 / 2606.26398

DinoLink: A Token-Centric Representation Compression Framework for Bandwidth-Constrained Collaborative V2X Perception

DinoLink：一种面向令牌的表示压缩框架，用于带宽受限的协作V2X感知
Zhu, Tianle, Que, Haohua, Yao, Handong, Xu, Hongyi, Bao, Zhipeng
Abstract
High-precision remote perception is often hindered by the severe bandwidth constraints of Vehicle-to-Everything (V2X) networks. We propose \textit{DinoLink}, a token-centric compression framework that replaces raw pixel streaming with discrete semantic communication for vehicle-cloud collaborative inference. DinoLink employs a dual-sparsity architecture: a saliency-aware selector prunes redundant background tokens, while a Residual Vector Quantization (RVQ) module collapses features into compact codebook indices. By transmitting only lightweight indices and positional priors, DinoLink achieves a $139\times$ bitrate reduction compared to uncompressed transmission while maintaining a competitive 32.8\% mAP on the nuScenes dataset. Deployment simulations further demonstrate a $34.5\times$ acceleration in narrow-band environments, such as LoRa. Our results substantiate DinoLink as a robust, bandwidth-efficient frontend for high-fidelity remote perception in constrained V2X scenarios. The code is publicly available at https://github.com/UGA-MOBILITY-LAB/dino_link.
Chinese Translation
高精度远程感知常常受到车对一切（V2X）网络严重带宽限制的阻碍。我们提出了 extit{DinoLink}，一种以令牌为中心的压缩框架，它用离散语义通信替代原始像素流，以实现车辆与云的协作推理。DinoLink采用双稀疏架构：一个关注显著性的选择器修剪冗余的背景令牌，而一个残差向量量化（Residual Vector Quantization, RVQ）模块将特征压缩为紧凑的代码本索引。通过仅传输轻量级索引和位置先验，DinoLink在保持竞争力的32.8 ext{%} mAP（平均精度均值）在nuScenes数据集上的同时，实现了与未压缩传输相比$139 imes$的比特率降低。部署模拟进一步证明，在窄带环境（如LoRa）中加速达到了$34.5 imes$。我们的结果证实了DinoLink作为高保真远程感知在受限V2X场景中的一种强大且高效的带宽前端。代码可在https://github.com/UGA-MOBILITY-LAB/dino_link公开获取。
View on arXiv Download PDF AI Translation
cs.CV / 13 / 2606.26410

Neural Voxel Dynamics: Learning Implicit 3D Physics via Volumetric Feature Advection

神经体素动态：通过体积特征对流学习隐式三维物理
Wang, Zican, Mitra, Niloy
Abstract
We present a self-supervised framework for learning implicit 3D physical dynamics directly from video-derived supervisory signals. While current generative video models achieve high visual fidelity, they lack a 3D geometric foundation, often resulting in physical inconsistencies and a failure to maintain object permanence. We address this by shifting the predictive bottleneck from 2D image space to a `lifted' 3D Volumetric Latent Space. Our method unprojects semantic features from a Video Joint-Embedding Predictive Architecture (V-JEPA) into a voxelized grid, grounded by monocular depth priors. This lifting enables a Volumetric Feature Advection to learn an action-conditioned transition operator that treats physics as a spatio-temporal state advection problem, i.e., learn implicit 3D physics. Unlike state-of-the-art hybrid models that rely on explicit classical simulators for training and/or inference, our architecture tracks material states implicitly within high-dimensional V-JEPA features. This allows for the emergent simulation of heterogeneous phenomena (e.g., rigid body motion in fluid flow) within a single, unified pipeline. Supervised solely via end-to-end video-derived signal plus action conditions, without access to physics engine internal states, labels, or surrogate models, our model demonstrates good long-term structural stability and physical plausibility on multiple benchmarks (CLEVERER, PhysInOne, PhysGaia). We believe that this work opens a scalable pathway toward general-purpose dynamic world models that internalize the 3D invariants of the physical world solely through passive observation of monocular videos.
Chinese Translation
我们提出了一种自监督框架，通过视频衍生的监督信号直接学习隐式三维物理动态。尽管当前的生成视频模型在视觉保真度上表现出色，但它们缺乏三维几何基础，常常导致物理不一致性以及无法维持物体的持久性。我们通过将预测瓶颈从二维图像空间转移到“提升”的三维体积潜在空间来解决这一问题。我们的方法将语义特征从视频联合嵌入预测架构（Video Joint-Embedding Predictive Architecture, V-JEPA）反投影到一个体素化网格中，并以单目深度先验为基础。这种提升使得体积特征对流能够学习一个基于动作条件的转移算子，将物理视为一个时空状态对流问题，即学习隐式三维物理。与依赖显式经典模拟器进行训练和/或推理的最先进混合模型不同，我们的架构在高维V-JEPA特征中隐式追踪材料状态。这使得在单一统一的管道中能够自发地模拟异质现象（例如，流体流动中的刚体运动）。我们的模型仅通过端到端的视频衍生信号加上动作条件进行监督，而无需访问物理引擎的内部状态、标签或替代模型，在多个基准测试（CLEVERER、PhysInOne、PhysGaia）上展示了良好的长期结构稳定性和物理合理性。我们相信，这项工作为通用动态世界模型开辟了一条可扩展的路径，使其仅通过对单目视频的被动观察内化物理世界的三维不变性。
View on arXiv Download PDF AI Translation
cs.CV / 14 / 2606.26416

Methane-Plume Segmentation From Hyperspectral Satellite Imagery Via Multimodal Deep Learning

基于多模态深度学习的高光谱卫星影像甲烷羽流分割
Quintero, Brayan, Acevedo, Jeferson, Traslaviña, Samuel, Rueda-Chacón, Hoover
Abstract
Efficient detection of methane plumes is crucial for understanding and mitigating global warming, as accurately identifying and segmenting them in earth observation imagery remain essential for large-scale monitoring. In this work, we propose a multimodal deep learning model that integrates a feature-guided methane enhancement (FGME) mechanism which injects physically meaningful methane cues into transformer-based RGB representations at multiple semantic scales. Our method is evaluated on the MPDataset, where it outperforms the state-of-the-art with improvements of +0.92 in MIoU, +0.87 in MPrecision and +1.01 in Recall. Notably, these gains are obtained with a substantially lower computational cost than other high-performing architectures, resulting in a favorable accuracy-efficiency trade-off for large-scale methane monitoring. These results highlight the potential of efficient multimodal fusion strategies for accurate and scalable methane plume segmentation in real-world remote sensing applications.
Chinese Translation
高效检测甲烷羽流对于理解和缓解全球变暖至关重要，因为在地球观测影像中准确识别和分割甲烷羽流对于大规模监测仍然是必不可少的。在本研究中，我们提出了一种多模态深度学习模型，该模型集成了一种特征引导的甲烷增强机制（Feature-Guided Methane Enhancement, FGME），该机制在多个语义尺度上将物理意义明确的甲烷线索注入基于变换器的RGB表示中。我们在MPDataset上对该方法进行了评估，结果显示其在MIoU上提高了+0.92，在MPrecision上提高了+0.87，在Recall上提高了+1.01，超越了当前最先进的技术。值得注意的是，这些提升是在比其他高性能架构显著更低的计算成本下获得的，导致在大规模甲烷监测中实现了良好的准确性与效率的权衡。这些结果突显了高效多模态融合策略在现实世界遥感应用中实现准确且可扩展的甲烷羽流分割的潜力。
View on arXiv Download PDF AI Translation
cs.CV / 15 / 2606.26455

Active Adversarial Perturbation-driven Associative Memory Retrieval for RGB-Event Visual Object Tracking

基于主动对抗扰动驱动的RGB-事件视觉目标跟踪的联想记忆检索
Wang, Xiao, Lou, Xufeng, Yan, Zikang, Chen, Lan, Chen, Sibao, Wang, Yaowei, Tian, Yonghong, Tang, Jin
Abstract
RGB-Event tracking improves localization robustness by fusing RGB appearance textures and dense temporal motion cues from event sensors. While this multi-modal scheme broadens tracking applicability, real-world scenes suffer diverse structured signal degradations that hinder traditional multi-modal fusion. In harsh environments, either modality can lose reliability drastically, and targets frequently appear incomplete due to occlusion, edge truncation and foreground clutter.To tackle the above challenges, we present a hierarchical perturbation and retrieval framework tailored for RGB-Event tracking with robustness against partial target missing and modal degradation, termed APRTrack. To mimic real-world signal corruption, APRTrack constructs structured degradation via two adversarial perturbation branches at the modality and spatial levels, which separately simulate full-modal failure and localized target region absence. A hierarchical routing mechanism is designed to disentangle the training pipelines of the two perturbation types, effectively eliminating feature collapse induced by superimposed degradation constraints. Furthermore, we devise Footprint-guided Channel-calibrated Hopfield Retrieval (FCHR) for reliable historical information compensation. This module evaluates retrieval confidence based on association footprints between queries and memory banks, and calibrates the retrieval metric space prior to Hopfield matching, realizing controllable historical feature compensation bounded to target regions. Extensive experiments on FE108, COESOT, VisEvent, and FELT datasets demonstrate the effectiveness of our proposed strategies for the RGB-Event visual object tracking. The source code and pre-trained models will be released on https://github.com/Event-AHU/OpenEvTracking
Chinese Translation
RGB-事件跟踪通过融合RGB外观纹理和来自事件传感器的密集时间运动线索，提高了定位的鲁棒性。尽管这种多模态方案拓宽了跟踪的适用性，但现实场景中存在多种结构化信号降解，阻碍了传统多模态融合。在恶劣环境中，任一模态的可靠性可能会急剧下降，目标由于遮挡、边缘截断和前景杂乱而经常显得不完整。为了解决上述挑战，我们提出了一种分层扰动和检索框架，专门针对RGB-事件跟踪，具有对部分目标缺失和模态降解的鲁棒性，称为APRTrack。为了模拟现实世界中的信号损坏，APRTrack通过模态和空间层面的两个对抗扰动分支构建结构化降解，分别模拟全模态失效和局部目标区域缺失。设计了一种分层路由机制，以解耦两种扰动类型的训练管道，有效消除由叠加降解约束引起的特征崩溃。此外，我们设计了基于足迹引导的通道校准霍普菲尔德检索（FCHR）以实现可靠的历史信息补偿。该模块根据查询和记忆库之间的关联足迹评估检索置信度，并在霍普菲尔德匹配之前校准检索度量空间，实现对目标区域的可控历史特征补偿。在FE108、COESOT、VisEvent和FELT数据集上的大量实验表明，我们提出的策略在RGB-事件视觉目标跟踪中是有效的。源代码和预训练模型将发布在https://github.com/Event-AHU/OpenEvTracking
View on arXiv Download PDF AI Translation
cs.CV / 16 / 2606.26515

Forget, Anticipate and Adapt: Test Time Training for Long Videos

遗忘、预期与适应：长视频的测试时间训练
Modi, Rajat, Noel, Sebastian, Liang, Xin, Rawat, Yogesh Singh
Abstract
Test Time Training (TTT) is a mechanism in which a model adapts to an incoming test-sample by performing some self-supervised (SSL) task and updating its weights even during inference. This procedure does not require labels at test-time. This paper focuses on TTT for long-videos. A major concern with existing approaches is: 1) they perform TTT updates using a sliding window containing frames in the past, whose compute increases linearly with the size of window. This becomes computationally intractable when the videos are hours long. 2) TTT is performed even when temporally close frames look similar, thereby consuming a lot of compute. We present the Frame Forgetting Network (FFN) that: 1) operates on only three frames within the sliding window, namely the frame that exits, the current frame and the frame after that. The model still manages to retain temporal context and work for hours long-videos; 2) mathematically define a surprise metric: how much new information the incoming frame contains with respect to the past seen frame. This facilitates determining how to modify the effective window size during TTT and constitutes the core mechanism of an adaptive windowing algorithm. Additionally, we curate a dataset EpicTours containing up to 3 hour long videos of walking city-tours, whereas earlier datasets on this problem were only 5 min long. We demonstrate FFNs empirical effectiveness on dense-segmentation, video classification tasks, generalization to depth-estimation, and multi-hour long videos.
Chinese Translation
测试时间训练（Test Time Training, TTT）是一种机制，模型通过执行某些自监督（Self-Supervised Learning, SSL）任务并在推理过程中更新其权重，从而适应传入的测试样本。该过程在测试时不需要标签。本文聚焦于长视频的TTT。现有方法的主要问题是：1）它们使用包含过去帧的滑动窗口进行TTT更新，其计算量随着窗口大小线性增加。当视频时长达到数小时时，这种计算变得不可行。2）即使在时间上相近的帧看起来相似时，TTT仍然会被执行，从而消耗大量计算资源。我们提出了帧遗忘网络（Frame Forgetting Network, FFN），其特点是：1）仅在滑动窗口内操作三个帧，即退出的帧、当前帧和下一个帧。该模型仍然能够保留时间上下文，并适用于数小时长的视频；2）数学上定义了一种惊讶度量：新输入帧相对于过去已见帧包含多少新信息。这有助于确定在TTT过程中如何修改有效窗口大小，并构成自适应窗口算法的核心机制。此外，我们整理了一个数据集EpicTours，包含最长达3小时的城市步行游览视频，而早期关于该问题的数据集仅为5分钟长。我们展示了FFN在密集分割、视频分类任务、深度估计的泛化能力以及多小时长视频上的实证有效性。
View on arXiv Download PDF AI Translation
cs.CV / 17 / 2606.26535

From Hallucination to Grounding: Diagnosing Visual Spatial Intelligence via CRISP

从幻觉到基础：通过CRISP诊断视觉空间智能
Li, Zhixing, Yu, Yinan
Abstract
Current VLM evaluations often conflate language priors with genuine spatial reasoning. To address this, we introduce CRISP, a novel structural-diagnostic evaluation paradigm that assesses visual spatial intelligence through consistency, the alignment between implicit perception and explicit reasoning. Unlike traditional black-box QA, CRISP utilizes metric 3D Scene Graphs and an oracle intervention protocol to decouple latent reasoning capabilities from perceptual bottlenecks. This granular diagnosis uncovers a systematic perception-reasoning disconnect. Crucially, we reveal that while proprietary models possess robust latent reasoning engines, they suffer from inaccurate metric estimation and a critical failure to leverage their implicit structural representations. Conversely, open-source models remain fundamentally bottlenecked by their lack of multi-hop compositional reasoning. By shifting the focus from merely ``guessing correctly'' via language priors to genuinely ``perceiving, verifying, and reasoning,'' CRISP offers a rigorous roadmap for multimodal alignment beyond end-to-end post-training. The code and dataset are available at https://github.com/iiyamayuki/CRISP-Bench.
Chinese Translation
当前的视觉语言模型（VLM）评估常常将语言先验与真正的空间推理混为一谈。为了解决这一问题，我们引入了CRISP，这是一种新颖的结构性诊断评估范式，通过一致性评估视觉空间智能，即隐性感知与显性推理之间的对齐。与传统的黑箱问答不同，CRISP利用度量3D场景图和一个神谕干预协议，将潜在推理能力与感知瓶颈解耦。这种细致的诊断揭示了系统性的感知-推理断裂。重要的是，我们发现虽然专有模型拥有强大的潜在推理引擎，但它们在度量估计上不准确，并且未能有效利用其隐性结构表示。相反，开源模型在多跳组合推理的缺乏上仍然受到根本性的瓶颈限制。通过将重点从仅仅通过语言先验“正确猜测”转向真正的“感知、验证和推理”，CRISP为超越端到端后训练的多模态对齐提供了严格的路线图。代码和数据集可在 https://github.com/iiyamayuki/CRISP-Bench 获取。
View on arXiv Download PDF AI Translation
cs.CV / 18 / 2606.26551

PhyEditBench: A Real-World Multi-Stage Benchmark for Physics-Aware Image Editing

PhyEditBench：一种面向物理的图像编辑的真实世界多阶段基准
Guo, Shengbin, He, Shaokang, Meng, Chaoyue, Xiao, Shengpeng, Xiang, Xunzhi, Zhang, Shaofeng, Fan, Qi
Abstract
While instruction-based image editing, enabled by multi-modal generative models, has advanced significantly, existing benchmarks lack a comprehensive evaluation of physics-based reasoning, a critical capability for handling real-world scenarios. To address this, we introduce PhyEditBench, a benchmark designed to assess the physical understanding of editing models. Guided by a hierarchical taxonomy, we establish 4 primary classes and 12 subclasses. It comprises 238 high-quality, high-resolution, real-world instances meticulously extracted from videos to capture authentic physical dynamics, alongside 35 synthetic Anti-Physics instances. Our empirical analysis of current SOTA editing methods exposes substantial limitations in their physics-based reasoning. We further propose a training-free baseline named PhyWorld that uses test-time scaling and a latent reduction strategy. PhyWorld outperforms comparable models and suggests that the video generation process can effectively serve as a reasoning mechanism for image editing. The project page is available at https://github.com/Previsior/PhyEditBench.
Chinese Translation
尽管基于指令的图像编辑在多模态生成模型的推动下取得了显著进展，但现有基准缺乏对物理推理的全面评估，而物理推理是处理真实世界场景的关键能力。为了解决这一问题，我们推出了PhyEditBench，一个旨在评估编辑模型物理理解能力的基准。我们依据层次分类法建立了4个主要类别和12个子类别。该基准包含238个高质量、高分辨率的真实世界实例，这些实例经过精心提取自视频，以捕捉真实的物理动态，同时还包括35个合成的反物理实例。我们对当前最先进（SOTA）编辑方法的实证分析揭示了它们在物理推理方面的重大局限性。我们进一步提出了一种无训练的基线模型PhyWorld，该模型利用测试时缩放和潜在降维策略。PhyWorld的表现优于可比模型，并表明视频生成过程可以有效地作为图像编辑的推理机制。项目页面可访问 https://github.com/Previsior/PhyEditBench。
View on arXiv Download PDF AI Translation
cs.CV / 19 / 2606.26552

Perception, Verdict, and Evolution: Hindsight-Driven Self-Refining Forensics Agent for AI-Generated Image Detection

感知、裁决与演变：基于事后驱动的自我优化法医代理用于AI生成图像检测
Wu, Yangjun, Yan, Keyu, Liu, Yu, Zhou, Jingren, Huang, Fei, Zhang, Rong, Zhao, Zhou, Wu, Fei
Abstract
The rapid advancement of generative models presents a significant challenge to existing deepfake detection methods, particularly given the widespread dissemination of highly realistic AI-generated images. Although Multimodal Large Language Models (MLLMs) show strong potential for this task, existing approaches suffer from two key limitations: insufficient sensitivity to fine-grained forensic artifacts and reliance on static synthetic supervision from frontier models, leading to limited flexibility and high-cost. To address these issues, we propose ForeAgent, an agentic forensics framework for AI-generated image detection with iterative self-evolution. First, ForeAgent adopts a Perception-Verdict architecture that aggregates multi-view cues spanning semantic, spatial, and frequency-domain features, and leverages an MLLM as a verdict module to fuse these signals for a logical-grounded verdict. Second, to enable continual self-improvement, we introduce a Hindsight-Driven Self-Refining strategy following a Sampling-Reflection-Evolution paradigm. The agent performs inference rollouts on training instances. Guided by ground-truth labels as hindsight, it reflects on failure cases and low-quality reasoning trajectories to regenerate higher-quality reasoning traces. These synthesized samples are then strictly filtered through a dual-expert quality gating module. ForeAgent continuously evolves via fine-tuning on self-curated high-quality samples. Extensive experiments demonstrate that ForeAgent achieves state-of-the-art performance on the Chameleon benchmark, reaching 82.18% accuracy (+16.41% over AIDE), and achieves 93.3% mean accuracy on AIGCDetect-Benchmark across 16 generators. In addition, external evaluation shows that ForeAgent produces more consistent and causally grounded reasoning compared to GPT-5 and GPT-5-mini.
Chinese Translation
生成模型的快速发展对现有的深度伪造检测方法提出了重大挑战，尤其是在高度真实的AI生成图像广泛传播的背景下。尽管多模态大型语言模型（Multimodal Large Language Models, MLLMs）在这一任务中展现出强大的潜力，但现有方法存在两个主要局限：对细粒度法医伪影的敏感性不足，以及依赖前沿模型的静态合成监督，导致灵活性有限且成本高昂。为了解决这些问题，我们提出了ForeAgent，一个具有迭代自我演化能力的AI生成图像检测法医框架。首先，ForeAgent采用感知-裁决架构，聚合跨语义、空间和频域特征的多视角线索，并利用MLLM作为裁决模块融合这些信号以形成逻辑基础的裁决。其次，为了实现持续自我改进，我们引入了一种基于事后驱动的自我优化策略，遵循采样-反思-演变的范式。该代理在训练实例上执行推理回滚。在真实标签的指导下，作为事后的反思，它对失败案例和低质量推理轨迹进行反思，以再生更高质量的推理痕迹。这些合成样本随后通过双专家质量筛选模块严格过滤。ForeAgent通过对自我策划的高质量样本进行微调，持续演化。大量实验表明，ForeAgent在Chameleon基准测试中实现了最先进的性能，达到82.18%的准确率（比AIDE提高16.41%），并在16个生成器上实现了93.3%的平均准确率。此外，外部评估显示，ForeAgent相比于GPT-5和GPT-5-mini，产生了更一致且因果基础更强的推理。
View on arXiv Download PDF AI Translation
cs.CV / 20 / 2606.26557

Coarse-to-Fine: A Hybrid Self-Supervised Method for Non-rigid 3D Shape Matching

粗到细：一种用于非刚性三维形状匹配的混合自监督方法
Luo, Feifan, Li, Ting, Li, Zhao, Chen, Hongyang
Abstract
Non-rigid 3D shape matching is a fundamental task in computer vision and graphics. In this paper, we propose a hybrid self-supervised method based on a coarse-to-fine strategy, which ensures consistency between the coarse mapping and the refined correspondence produced by our refinement module. The architecture features a dual-branch design, consisting of two symmetric functional map learning streams: one based on the Laplacian basis and the other utilizing the elastic basis. Extensive experiments show that our approach not only maintains computational efficiency, but also achieves state-of-the-art performance across a variety of challenging scenarios, including non-isometric deformations and topological noise. Finally, we rigorously demonstrate that contrastive energies promote feature discrimination. Furthermore, integrating these energies with existing methods yields consistent improvements, validating the overall efficacy of our approach. Our code is available at https://github.com/LuoFeifan77/Coarse-to-Fine-Hybrid-Self-Supervised-Matching.
Chinese Translation
非刚性三维形状匹配是计算机视觉和图形学中的一项基础任务。本文提出了一种基于粗到细策略的混合自监督方法，该方法确保粗略映射与我们精细化模块生成的精细对应之间的一致性。该架构采用双分支设计，由两个对称的功能映射学习流组成：一个基于拉普拉斯基，另一个利用弹性基。大量实验表明，我们的方法不仅保持了计算效率，还在多种具有挑战性的场景中实现了最先进的性能，包括非等距变形和拓扑噪声。最后，我们严格证明了对比能量促进特征区分。此外，将这些能量与现有方法结合使用可带来一致的改进，验证了我们方法的整体有效性。我们的代码可在 https://github.com/LuoFeifan77/Coarse-to-Fine-Hybrid-Self-Supervised-Matching 获取。
View on arXiv Download PDF AI Translation
cs.CV / 21 / 2606.26559

SpaceRipple: Lightweight Semantic Delivery for Mission-Oriented LEO Earth Observation Satellite Networks

SpaceRipple：面向任务的低地球轨道地球观测卫星网络的轻量级语义传输
Yang, Ziyi, Yuan, Hao, Yi, Yunxiang, Wang, Wenbo, Zhang, Xing
Abstract
Earth observation satellite networks generate massive volumes of high-resolution imagery, whereas inter-satellite and downlink resources remain limited. In many time-sensitive missions, ground users require mission-relevant semantic information rather than a full raw-image downlink. This paper proposes SpaceRipple, a lightweight framework for mission-oriented semantic delivery and on-board processing in Earth observation satellite networks. A sensing satellite performs adaptive compression and metadata generation to reduce inter-satellite traffic, while an edge computing satellite restores the received representation and extracts task-relevant semantic information. Unlike fidelity-driven image transmission, SpaceRipple coordinates compression, forwarding, restoration, and semantic inference within a collaborative pipeline, enabling semantic-oriented delivery instead of pixel-level image delivery. A compression-aware MoE enhancement module is further introduced to improve robustness under degraded visual inputs. Experimental results show that SpaceRipple achieves favorable reconstruction quality, improved semantic detection performance, and substantial bandwidth savings, demonstrating its potential for efficient and reliable Earth observation under constrained satellite-network resources.
Chinese Translation
地球观测卫星网络生成大量高分辨率图像，而星间和下行链路资源仍然有限。在许多时间敏感的任务中，地面用户需要与任务相关的语义信息，而不是完整的原始图像下行。本文提出了SpaceRipple，一个面向任务的轻量级语义传输和在轨处理框架，用于地球观测卫星网络。感知卫星执行自适应压缩和元数据生成，以减少星间流量，而边缘计算卫星则恢复接收到的表示并提取与任务相关的语义信息。与以保真度为驱动的图像传输不同，SpaceRipple在协作管道中协调压缩、转发、恢复和语义推理，实现语义导向的传输，而非像素级的图像传输。此外，进一步引入了一种压缩感知的MoE（Mixture of Experts）增强模块，以提高在视觉输入退化情况下的鲁棒性。实验结果表明，SpaceRipple实现了良好的重建质量、改善的语义检测性能和显著的带宽节省，展示了其在受限卫星网络资源下高效可靠的地球观测潜力。
View on arXiv Download PDF AI Translation
cs.CV / 22 / 2606.26602

DiCoBench: Benchmarking Multi-Image Fine-Grained Perception via Differential and Commonality Visual Cues

DiCoBench：通过差异和共性视觉线索对多图像细粒度感知进行基准测试
Li, Geng, Peng, Yuxin
Abstract
Recent advancements in Multimodal Large Language Models (MLLMs) have demonstrated impressive fine-grained perception capabilities. However, existing benchmarks predominantly rely on explicit textual cues or low-resolution inputs, failing to evaluate a model's ability to autonomously perceive implicit visual cues in high-resolution. To bridge this gap, we introduce DiCoBench, a comprehensive, multi-image high-resolution benchmark designed for cross-image fine-grained perception. DiCoBench consists of 765 meticulously curated samples categorized into two progressive tracks: Differential Visual Cues and Commonality Visual Cues, covering 8 distinct perception tasks. By formulating the benchmark as a multiple-choice question task and utilizing high-resolution imagery (approaching 2K), we eliminate evaluation metric bias and pose a substantial challenge to current state-of-the-art MLLMs. Our extensive evaluation of 18 diverse MLLMs reveals a striking performance gap compared to human accuracy (98.3\%), with top-performing models struggling significantly with micro-scale detail capture. We believe DiCoBench will serve as a challenging testbed to drive future research in autonomous, high-resolution multi-image perception.
Chinese Translation
最近，多模态大型语言模型（MLLMs）的进展展示了令人印象深刻的细粒度感知能力。然而，现有基准测试主要依赖于显式文本线索或低分辨率输入，未能评估模型在高分辨率下自主感知隐式视觉线索的能力。为了解决这一问题，我们引入了DiCoBench，这是一个全面的多图像高分辨率基准，旨在进行跨图像细粒度感知。DiCoBench包含765个精心策划的样本，分为两个渐进的轨道：差异视觉线索和共性视觉线索，涵盖8个不同的感知任务。通过将基准测试制定为多项选择题任务，并利用接近2K的高分辨率图像，我们消除了评估指标的偏差，并对当前最先进的MLLMs提出了重大挑战。我们对18个不同的MLLMs进行了广泛评估，结果显示与人类准确率（98.3%）相比，存在显著的性能差距，表现最佳的模型在微观细节捕捉方面显著乏力。我们相信，DiCoBench将成为推动未来自主高分辨率多图像感知研究的挑战性测试平台。
View on arXiv Download PDF AI Translation
cs.CV / 23 / 2606.26609

LogicIR: Logic Gate Networks for Image Restoration

LogicIR：用于图像恢复的逻辑门网络
Lee, Hongjae, Son, Myungjun, Yu, Jaeseong, Jung, Seung-Won
Abstract
Image restoration aims to reconstruct high-quality images from degraded low-quality inputs. As the computational demands of image restoration models continue to rise, there is growing interest in lightweight architectures optimized for fast and efficient inference. Logic gate networks (LGNs), which operate using fundamental logic operations such as NAND and XOR, have recently emerged as a promising direction for achieving highly efficient computation. However, their potential remains largely untapped in the domain of image restoration. In this work, we introduce LogicIR, the first LGN specifically designed for image restoration tasks. LogicIR incorporates a UNet-inspired architecture composed entirely of logic gates. In addition, we propose a differentiable bit decoding layer and an index shuffling mechanism that improves information propagation across logic gates. Experimental results across multiple image restoration benchmarks demonstrate that LogicIR achieves strong performance with significantly reduced computational cost, establishing LogicIR as a viable and efficient alternative for image restoration. The source code is available at https://github.com/jimmy9704/LogicIR
Chinese Translation
图像恢复旨在从退化的低质量输入中重建高质量图像。随着图像恢复模型的计算需求不断增加，对优化快速高效推理的轻量级架构的兴趣日益增长。逻辑门网络（Logic Gate Networks, LGNs）利用基本的逻辑运算（如 NAND 和 XOR）进行操作，最近被认为是实现高效计算的有前景的方向。然而，它们在图像恢复领域的潜力仍然未得到充分利用。在本研究中，我们介绍了 LogicIR，这是首个专为图像恢复任务设计的 LGN。LogicIR 采用完全由逻辑门组成的 UNet 风格架构。此外，我们提出了一种可微分的位解码层和一种索引洗牌机制，以改善逻辑门之间的信息传播。在多个图像恢复基准测试中的实验结果表明，LogicIR 在显著降低计算成本的同时实现了强大的性能，确立了 LogicIR 作为图像恢复的可行且高效的替代方案。源代码可在 https://github.com/jimmy9704/LogicIR 获取。
View on arXiv Download PDF AI Translation
cs.CV / 24 / 2606.26615

TaskTok: Delving into Task Tokens for Task-driven Image Restoration

TaskTok：深入探讨任务驱动图像恢复中的任务标记
Lee, Hongjae, Kang, Sojung, Yu, Jaeseong, Jung, Seung-Won
Abstract
While traditional image restoration focuses on perceptual quality, Task-Driven Image Restoration (TDIR) aims to maximize the performance of downstream high-level vision tasks. Recent approaches leveraging generative priors have shown promise for TDIR; however, they typically suffer from computational inefficiency and potential semantic alteration by indiscriminately updating all latent tokens. In this paper, we posit that not all visual information is equally important for machine perception. Through an analysis of the latent token space, we observe that task-relevant cues are unevenly distributed across the token sequence, exhibiting index-wise specialization. This suggests that selectively refining a subset of tokens can be sufficient for task-driven objectives. Leveraging this insight, we propose TaskTok, a novel framework that selectively restores only task-relevant tokens via a learnable token switch and a lightweight token refinement module. Extensive experiments across image classification, semantic segmentation, and object detection demonstrate that TaskTok significantly enhances task performance with high computational efficiency. The source code is available at https://github.com/jimmy9704/TaskTok
Chinese Translation
传统的图像恢复关注于感知质量，而任务驱动图像恢复（TDIR）旨在最大化下游高层视觉任务的性能。近期利用生成先验的方法在TDIR中显示出良好的前景；然而，它们通常面临计算效率低下和通过无差别更新所有潜在标记而导致的潜在语义改变的问题。在本文中，我们认为并非所有视觉信息对机器感知同等重要。通过对潜在标记空间的分析，我们观察到与任务相关的线索在标记序列中分布不均，表现出按索引的专业化。这表明，选择性地细化一部分标记对于任务驱动目标可能是足够的。基于这一见解，我们提出了TaskTok，一个新颖的框架，通过可学习的标记切换和轻量级的标记细化模块，仅选择性地恢复与任务相关的标记。针对图像分类、语义分割和目标检测的广泛实验表明，TaskTok显著提升了任务性能，并具有高计算效率。源代码可在 https://github.com/jimmy9704/TaskTok 获取。
View on arXiv Download PDF AI Translation
cs.CV / 25 / 2606.26631

Position Rebinding Cache Reuse: Replay-Free Visual Revisiting for Interleaved Multimodal Reasoning

位置重绑定缓存重用：无重放的视觉重访用于交错多模态推理
Wang, Mengzhao, Ji, Yanli, Zuo, Wangmeng, Ye, Peng, Tu, Chongjun
Abstract
Interleaved multimodal reasoning improves visual grounding by revisiting visual evidence during multi-step generation, yet existing methods typically rely on token replay, repeatedly forwarding selected visual tokens. A natural shortcut is to reuse the historical visual key-value (KV) cache directly. However, we identify a critical failure mode of this strategy: cached visual keys are already bound to their original positional context. Such stale positional binding distorts attention under later decoding contexts and can trigger severe autoregressive decoding collapse. This failure suggests that effective cache reuse requires reconstructing visual evidence under positions compatible with the current decoding state, rather than directly copying position-bound historical cache entries. To this end, we propose Position Rebinding Cache Reuse (PRCR), a cache-level framework for replay-free visual revisiting. PRCR stores raw visual KV cache together with their original spatial coordinates, then reassigns position-compatible coordinates to select entries and rebinds their keys before injecting the reconstructed cache into the active decoder cache. This design reuses historical visual evidence while preserving textual positional continuity and relative visual structure. Experiments across multiple multimodal reasoning benchmarks show that PRCR achieves replay-level or better performance, improving average accuracy by 5 percent and reducing visual-revisiting computation by up to tens of thousands of times.
Chinese Translation
交错多模态推理通过在多步生成过程中重访视觉证据来改善视觉定位，然而现有方法通常依赖于令牌重放，反复转发选定的视觉令牌。一个自然的捷径是直接重用历史视觉键值（KV）缓存。然而，我们发现这一策略存在一个关键的失效模式：缓存的视觉键已经绑定到其原始位置上下文。这种过时的位置绑定在后续解码上下文中扭曲了注意力，并可能导致严重的自回归解码崩溃。这一失效表明，有效的缓存重用需要在与当前解码状态兼容的位置下重建视觉证据，而不是直接复制绑定位置的历史缓存条目。为此，我们提出了位置重绑定缓存重用（Position Rebinding Cache Reuse, PRCR），这是一个无重放视觉重访的缓存级框架。PRCR存储原始视觉KV缓存及其原始空间坐标，然后为选定条目重新分配位置兼容的坐标，并在将重建的缓存注入活动解码器缓存之前重新绑定其键。该设计在重用历史视觉证据的同时，保持了文本位置的连续性和相对视觉结构。在多个多模态推理基准测试中的实验表明，PRCR实现了与重放水平相当或更好的性能，平均准确率提高了5个百分点，视觉重访计算减少了数万次。
View on arXiv Download PDF AI Translation
cs.CV / 26 / 2606.26634

Temporally Consistent Label Interpolation for Robust Surgical Multi-Task Learning under Challenging Conditions

在挑战性条件下进行稳健外科多任务学习的时间一致性标签插值
Kim, Garam, Park, Juyoun
Abstract
Effective multi-task learning for surgical scene understanding is fundamentally hindered by annotation granularity mismatch; temporal workflow tasks such as phase recognition, step recognition and anticipation benefit from dense frame-level supervision, whereas pixel-level spatial tasks including instrument segmentation and action recognition are only sparsely annotated on selected keyframes due to prohibitive labeling costs. This supervision imbalance undermines shared representation learning and limits joint optimization across heterogeneous surgical tasks. To address this, we propose Flow-guided Annotation for Robust Operating Scenes (FAROS), a flow-guided label interpolation framework, that combines zero-shot segmentation-based mask propagation with optical flow estimation to overcome the limitations of appearance-based propagation under challenging surgical conditions such as occlusion, smoke, and motion blur, generating temporally consistent dense pseudo labels from sparse keyframe annotations. The densified instrument masks and action labels are integrated into a unified Transformer-based multi-task framework that jointly learns surgical phase recognition, step recognition, anticipation, instrument segmentation, and action recognition, enabling balanced optimization between dense temporal supervision and sparse spatial supervision. The label interpolation quality of FAROS is first validated on the DAVIS 2017 benchmark under a sparse ground-truth protocol, confirming robust propagation beyond the surgical domain. Extensive experiments on GraSP, MISAW, and AutoLaparo benchmarks further demonstrate that FAROS significantly improves cross-task representation learning and enhances holistic surgical scene understanding performance across spatio-temporal tasks.
Chinese Translation
外科场景理解的有效多任务学习受到注释粒度不匹配的根本性制约；时间工作流任务如阶段识别、步骤识别和预测受益于密集的帧级监督，而像仪器分割和动作识别这样的像素级空间任务仅在选定的关键帧上进行稀疏注释，原因在于标注成本高昂。这种监督不平衡削弱了共享表示学习，并限制了异构外科任务之间的联合优化。为了解决这个问题，我们提出了基于流引导的稳健操作场景注释框架（Flow-guided Annotation for Robust Operating Scenes, FAROS），该框架结合了零样本分割基础的掩膜传播与光流估计，以克服在诸如遮挡、烟雾和运动模糊等挑战性外科条件下基于外观的传播限制，从稀疏的关键帧注释中生成时间一致的密集伪标签。密集化的仪器掩膜和动作标签被整合到一个统一的基于Transformer的多任务框架中，该框架共同学习外科阶段识别、步骤识别、预测、仪器分割和动作识别，从而实现密集时间监督与稀疏空间监督之间的平衡优化。FAROS的标签插值质量首先在稀疏真实值协议下的DAVIS 2017基准上进行了验证，确认其在外科领域之外的稳健传播。对GraSP、MISAW和AutoLaparo基准的广泛实验进一步表明，FAROS显著改善了跨任务表示学习，并增强了在时空任务中的整体外科场景理解性能。
View on arXiv Download PDF AI Translation
cs.CV / 27 / 2606.26636

FracEvent: Event-Camera Simulation via Fractional-Relaxation Pixel Dynamics

FracEvent：通过分数松弛像素动态进行事件相机模拟
Chen, Langyi, Xu, Chuanzhi, Zhou, Haoxian, Ye, Pengfei, Luo, Ziyu, Chen, Haodong, Qu, Qiang, Chen, Xiaoming, Cai, Weidong
Abstract
Event cameras asynchronously report brightness changes with microsecond-level temporal resolution, but real event data remain difficult to collect at scale because specialized sensors, careful synchronization, and task-specific annotations are required. Event-camera simulation is therefore important to event-based vision tasks. Most practical simulators build on contrast-threshold event generation, some with additional filtering, stochastic noise, or hand-tuned sensor parameters. While effective, such formulations often simplify the temporal structure produced by the lifecycle of each pixel, which can distort event timing and weaken downstream transfer. We introduce FracEvent, an event simulator that models this pixel-level lifecycle with fractional-relaxation voltage dynamics. Given a log-intensity trajectory, FracEvent drives a compact stack of relaxation modes, combines their responses into a voltage state, emits ON/OFF events by localizing threshold crossings on the continuous voltage trajectory, and updates the reference while retaining the underlying memory modes. This retained state links residual voltage response to later event timing. We evaluate FracEvent through event-stream comparison and downstream transfer on image reconstruction and optical flow estimation. Across multiple datasets, FracEvent improves the temporal structure of generated events and achieves stronger downstream-transfer results than competing simulator baselines, showing its practical value for event-camera simulation.
Chinese Translation
事件相机以微秒级的时间分辨率异步报告亮度变化，但由于需要专用传感器、精确同步和特定任务的注释，真实事件数据的规模化收集仍然困难。因此，事件相机模拟对于基于事件的视觉任务至关重要。大多数实用的模拟器基于对比阈值事件生成，有些还添加了额外的过滤、随机噪声或手动调整的传感器参数。尽管有效，这些公式通常简化了每个像素生命周期所产生的时间结构，这可能会扭曲事件时序并削弱下游迁移。我们提出了FracEvent，这是一种事件模拟器，利用分数松弛电压动态来建模这种像素级生命周期。给定一个对数强度轨迹，FracEvent 驱动一组紧凑的松弛模式，将它们的响应组合成一个电压状态，通过在连续电压轨迹上定位阈值交叉来发出开/关事件，并在保留基础记忆模式的同时更新参考状态。这个保留的状态将剩余电压响应与后续事件时序关联起来。我们通过事件流比较和图像重建及光流估计的下游迁移评估FracEvent。在多个数据集上，FracEvent改善了生成事件的时间结构，并在下游迁移结果上优于竞争模拟器基线，显示了其在事件相机模拟中的实际价值。
View on arXiv Download PDF AI Translation
cs.CV / 28 / 2606.26647

LayersReg: A Layer-by-Layer Progressive Regressor for Reliable Intraoperative 3D/2D Registration

LayersReg：一种逐层渐进回归器，用于可靠的术中3D/2D配准
Wang, Xiyuan, Wang, Zhenchao, Chen, Xinran, Liu, Junkai, Chen, Chuan, Yin, Feng
Abstract
3D/2D registration serves as a cornerstone technique in surgical navigation. Traditional iterative optimization algorithms suffer from low efficiency and high failure rates in intraoperative settings. Deep learning-based methods reformulate registration from iterative optimization to a regression problem that maps image appearance features to spatial pose, typically achieving improved real-time performance and accuracy. However, such learnable methods are confined to memory-driven retrieval of specific pose features rather than understanding the task of image alignment itself, which limits their generalization in complex scenarios. We propose LayersReg, a pioneering regression paradigm that endows the model with 3D anatomical awareness and searches for the correct pose in a progressive, layer-by-layer manner. Inspired by the iterative pose-searching optimization criterion of classical registration, LayersReg searches for correlations between the moving and fixed images in feature space, capturing the trend of pixel flow and thereby converging iteratively toward the correct spatial pose transformation. We further design a coupling of node-wise regression with the progressive registration framework to enhance the model's perception of spatial pose changes. Experimental results demonstrate that under large offsets and multimodality conditions, LayersReg achieves high accuracy on both X-ray/CT registration (0.68{\deg}, 1.41 mm) and slice localization (0.73{\deg}, 1.55 mm) tasks, outperforming existing state-of-the-art methods while meeting the intraoperative demands for precision and real-time capability.
Chinese Translation
3D/2D配准是外科导航中的基础技术。传统的迭代优化算法在术中环境中效率低下且失败率高。基于深度学习的方法将配准问题重新定义为一个回归问题，该问题将图像外观特征映射到空间姿态，通常能实现更好的实时性能和准确性。然而，这些可学习的方法仅限于对特定姿态特征的基于内存的检索，而非理解图像对齐任务本身，这限制了它们在复杂场景中的泛化能力。我们提出了LayersReg，这是一种开创性的回归范式，使模型具备3D解剖意识，并以逐层渐进的方式搜索正确的姿态。受经典配准的迭代姿态搜索优化标准的启发，LayersReg在特征空间中搜索移动图像与固定图像之间的相关性，捕捉像素流的趋势，从而逐步收敛到正确的空间姿态变换。我们进一步设计了节点回归与渐进配准框架的耦合，以增强模型对空间姿态变化的感知。实验结果表明，在大偏移和多模态条件下，LayersReg在X射线/CT配准（0.68{ ext{°}}, 1.41 mm）和切片定位（0.73{ ext{°}}, 1.55 mm）任务中均实现了高准确性，超越了现有的最先进方法，同时满足了术中对精度和实时能力的要求。
View on arXiv Download PDF AI Translation
cs.CV / 29 / 2606.26668

Disco-LoRA: Disentangled Composition of Content, Style, and Motion for Multi-concept Video Customization

Disco-LoRA：内容、风格和运动的解耦组合用于多概念视频定制
Xu, Xuancheng, Jia, Gengyun, Bao, Bing-Kun
Abstract
Video customization based on Text-to-Video (T2V) models aims to learn specific features from reference data to generate controllable videos. While significant strides have been made in image stylization and video motion customization, simultaneously controlling multiple concepts, such as content, style, and motion, remains a major challenge. In this work, we systematically define the task of multi-concept video customization, which requires the joint control of content, style, and motion. To facilitate research in this area, we construct a comprehensive benchmark and propose Disco-LoRA, a unified framework designed to tackle this problem by disentangling and flexibly recombining different concepts in two stages: (1) We decompose the objective into two sub-tasks: Content-Style and Content-Motion. Each sub-task is addressed using our Iterative Dual-LoRA Disentanglement Framework, which effectively disentangles distinct concepts within the data. (2) We identify layer-wise weight trends as crucial for LoRA identity, while weight magnitudes dictate composability. To harmonize these scales, we propose a Z-score-based statistical regularization that aligns weight distributions, preserving layer-wise trends while minimizing interference between different LoRAs. Extensive experiments show that Disco-LoRA excels in multi-concept video customization, effectively preserving appearance, style, and motion for controllable text-to-video generation.
Chinese Translation
基于文本到视频（Text-to-Video, T2V）模型的视频定制旨在从参考数据中学习特定特征，以生成可控的视频。尽管在图像风格化和视频运动定制方面取得了显著进展，但同时控制多个概念（如内容、风格和运动）仍然是一个主要挑战。在本研究中，我们系统性地定义了多概念视频定制的任务，该任务需要对内容、风格和运动进行联合控制。为促进该领域的研究，我们构建了一个全面的基准，并提出了Disco-LoRA，这是一个统一框架，旨在通过在两个阶段中解耦和灵活重组不同概念来解决这一问题：（1）我们将目标分解为两个子任务：内容-风格和内容-运动。每个子任务都使用我们的迭代双重LoRA解耦框架（Iterative Dual-LoRA Disentanglement Framework）进行处理，该框架有效地解耦数据中的不同概念。（2）我们识别出逐层权重趋势对于LoRA身份至关重要，而权重幅度则决定了可组合性。为了协调这些尺度，我们提出了一种基于Z-score的统计正则化方法，以对齐权重分布，保留逐层趋势，同时最小化不同LoRA之间的干扰。大量实验表明，Disco-LoRA在多概念视频定制中表现出色，有效地保留了可控文本到视频生成的外观、风格和运动。
View on arXiv Download PDF AI Translation
cs.CV / 30 / 2606.26687

DeCoFlow: Structural Decomposition of Normalizing Flows for Continual Anomaly Detection

DeCoFlow：用于持续异常检测的归一化流的结构分解
Im, Hun, Lee, Jungi, Cha, Subeen, Kang, Pilsung
Abstract
In industrial environments, new product categories arrive sequentially, requiring continual anomaly detection without access to past data. Normalizing Flows (NFs) provide exact density estimation but suffer from catastrophic forgetting as parameter updates across tasks distort the density manifold. While parameter isolation can prevent interference, it must preserve the strict invertibility and Jacobian validity of NFs. To satisfy these requirements, we exploit the inherent property that affine coupling layers maintain transformation validity regardless of subnet parameterization. Based on this, we propose DeCoFlow, which decomposes subnets into a frozen universal base and task-specific low-rank adapters to isolate updates. We further introduce Task-Specific Alignment, Auxiliary Coupling Layers, and Tail-Aware Loss to compensate for frozen-base rigidity. DeCoFlow achieves state-of-the-art image-level AUROCs of 98.40% on MVTec-AD and 93.00% on VisA, while maintaining parameter-level zero forgetting (0.00% FM under correct routing) with only 2.27M parameters per task.
Chinese Translation
在工业环境中，新产品类别会顺序到来，这要求在没有访问过去数据的情况下进行持续的异常检测。归一化流（Normalizing Flows, NFs）提供精确的密度估计，但由于任务间参数更新扭曲了密度流形，导致灾难性遗忘。虽然参数隔离可以防止干扰，但必须保持 NFs 的严格可逆性和雅可比有效性。为了满足这些要求，我们利用仿射耦合层的固有特性，即无论子网络参数化如何，均能保持变换的有效性。基于此，我们提出了 DeCoFlow，它将子网络分解为一个冻结的通用基础和任务特定的低秩适配器，以隔离更新。我们进一步引入了任务特定对齐、辅助耦合层和尾部感知损失，以补偿冻结基础的刚性。DeCoFlow 在 MVTec-AD 上实现了 98.40% 的图像级 AUROC，在 VisA 上实现了 93.00% 的图像级 AUROC，同时在每个任务中仅用 2.27M 参数保持了参数级别的零遗忘（在正确路由下为 0.00% FM）。
View on arXiv Download PDF AI Translation
cs.CV / 31 / 2606.26694

PhysEditWorld: A Large-Scale Dataset Toward Physics-Editable World Models

PhysEditWorld：面向物理可编辑世界模型的大规模数据集
Hu, Bin, Ma, Yanwen, Huang, Jiehui, Zhang, Ziliang, Wu, Haoning, Zhang, Ruicheng, Li, Yaokun, Wang, Zijun, Zhang, Yuechen, Tseng, Chun-Mei, Li, Hanhui, Qian, Shengju, Zhou, Jun, Zhang, Kaipeng, Liang, Xiaodan, Jia, Jiaya, Li, Xiu
Abstract
Recent game world models can synthesize visually plausible, action-conditioned rollouts. However, their interaction behaviors often remain limited to exploratory or wandering trajectories, and physical dynamics are typically learned as implicit correlations from data rather than as controllable variables. This limitation hinders their applicability to authored game environments, where physical rules are deliberately designed and require explicit manipulation. We introduce PhysEditWorld, a multimodal dataset with physical parameters, with a primary focus on gravity in this initial version. At its core, PhysEditWorld is built upon a replay paradigm implemented with a UE5 replay-and-rendering pipeline. Each scenario records a normalized action trace and replays the same initial state, character controller, action sequence, and camera policy under multiple gravity configurations, enabling controlled and attributable physical variation. PhysEditWorld contains 12 cinematic UE5 scenes, over 100 hours of gameplay interactions, and more than 60 million rendered rollout frames. Each sample provides synchronized multimodal signals, including RGB, depth, normals, audio, action traces, camera trajectory, engine states, semantic annotations, and explicit gravity labels. We further conduct initial utility studies on both generative video models and world understanding models, demonstrating that PhysEditWorld enables improved gravity-faithful dynamics modeling, enhances consistency under physical edits, and provides a scalable foundation for controllable world modeling research.
Chinese Translation
近年来的游戏世界模型能够合成视觉上可信的、基于动作的展开。然而，它们的交互行为通常仅限于探索性或游荡轨迹，物理动态通常是从数据中学习的隐式关联，而非可控变量。这一局限性妨碍了它们在创作游戏环境中的应用，因为这些环境的物理规则是经过精心设计的，并需要明确的操控。我们介绍了PhysEditWorld，这是一个包含物理参数的多模态数据集，初始版本主要聚焦于重力。PhysEditWorld的核心是基于一种重放范式，采用UE5重放和渲染管道实现。每个场景记录了一个标准化的动作轨迹，并在多种重力配置下重放相同的初始状态、角色控制器、动作序列和相机策略，从而实现可控和可归因的物理变化。PhysEditWorld包含12个电影级UE5场景，超过100小时的游戏交互，以及超过6000万帧渲染的展开画面。每个样本提供同步的多模态信号，包括RGB、深度、法线、音频、动作轨迹、相机轨迹、引擎状态、语义注释和明确的重力标签。我们还对生成视频模型和世界理解模型进行了初步的实用性研究，证明PhysEditWorld能够改善重力忠实动态建模，增强物理编辑下的一致性，并为可控世界建模研究提供可扩展的基础。
View on arXiv Download PDF AI Translation
cs.CV / 32 / 2606.26706

Intracranial Aneurysm Classification and Segmentation via Tri-Axial ROI and Multi-Task Learning

通过三轴感兴趣区域和多任务学习进行颅内动脉瘤分类与分割
Shi, Pengcheng, Yang, Kaiyuan, Huang, Houjing, Chen, Jiawei, Lu, Yan, Liu, Jiaqi, Xu, Murong, Menze, Bjoern, Zhang, Xinglin
Abstract
Intracranial aneurysms are often asymptomatic until rupture, which carries high mortality. Rupture risk assessment and treatment planning depend on both aneurysm morphology and anatomical location, yet existing automated methods remain limited to binary detection without fine-grained anatomical classification or multi-class segmentation. We present a multi-task framework that simultaneously performs multi-label classification, multi-class aneurysm segmentation, and multi-class vessel segmentation across 13 anatomical locations and four imaging modalities (CTA, MRA, T2, T1-post). Our two-stage approach combines a fast 2D tri-axial Region of Interest (ROI) extraction method with a 3D multi-task nnU-Net backbone. A dual-decoder design mitigates the extreme volume imbalance between aneurysm and vessel classes, while cross-attention pooling and modality-specific auxiliary heads improve feature learning across heterogeneous inputs. Our two-fold ensemble achieved 2nd place in the RSNA 2025 Intracranial Aneurysm Detection challenge. Code, model weights, and a 3D Slicer plugin are publicly available.
Chinese Translation
颅内动脉瘤通常在破裂前无症状，而破裂则伴随高死亡率。破裂风险评估和治疗规划依赖于动脉瘤的形态和解剖位置，然而现有的自动化方法仍然局限于二元检测，缺乏细粒度的解剖分类或多类分割。我们提出了一种多任务框架，能够同时进行多标签分类、多类动脉瘤分割和多类血管分割，覆盖13个解剖位置和四种成像模式（CTA、MRA、T2、T1-post）。我们的两阶段方法结合了一种快速的二维三轴感兴趣区域（ROI）提取方法与三维多任务nnU-Net骨干网络。双解码器设计缓解了动脉瘤与血管类别之间的极端体积不平衡，而交叉注意力池化和特定模态的辅助头则改善了异构输入的特征学习。我们的双重集成方法在RSNA 2025颅内动脉瘤检测挑战中获得第二名。代码、模型权重和3D Slicer插件已公开提供。
View on arXiv Download PDF AI Translation
cs.CV / 33 / 2606.26711

Mask to Concept: Auto-Promptable SAM3 via Efficient Test-Time Concept Embedding Search for Few-Shot Annotation

从掩膜到概念：通过高效的测试时概念嵌入搜索实现可自动提示的SAM3以进行少量标注
Zhou, Quan, Zhai, Shaoqing, Chen, Qiang Hu Jia, Li, Qiang, Wang, Zhiwei
Abstract
Transforming foundation segmentation models from human-prompted tools into auto-promptable annotators is critical for scalable medical data annotation. Current methods commonly depend on external feature matchers or auxiliary networks to automate geometric prompting, but introducing architectural overhead and limiting performance scalability. Although SAM3 natively supports concept segmentation via reusable text prompts, its direct use in medical imaging is hindered by a lack of fine-grained clinical knowledge and the ambiguity of human-written descriptions. In this work, we propose Mask to Concept (M2C), an efficient framework that adapts SAM3 for medical few-shot annotation without external modules, parameter retraining, or manual text engineering. Using only a few labeled images, M2C enables SAM3 to automatically search for transferable visual concepts entirely within its frozen architecture: it initializes a learnable concept embedding, uses it to prompt segmentation, and updates the embedding by gradients of minimizing the concept segmentation error. We further introduce a Hybrid Uncertainty Estimation (HUE) module that calculates the prediction entropy and maps concept predictions back to the box prompts, measuring concept-geometry prompting inconsistency. Highly uncertain samples are flagged actively for human correction, and the corrected masks are then fed back to M2C to continuously search for more precise concept embeddings, forming a self-enhancing annotation loop with minimal expert effort. Experiments on medical segmentation benchmarks show that our method achieves SOTA few-shot segmentation performance and outstanding annotation efficiency, offering a practical and efficient pathway toward scalable medical image labeling. Codes are at https://github.com/Huster-Hq/M2C.
Chinese Translation
将基础分割模型从人类提示工具转变为可自动提示的标注工具对于可扩展的医学数据标注至关重要。目前的方法通常依赖于外部特征匹配器或辅助网络来自动化几何提示，但这会引入架构开销并限制性能的可扩展性。尽管SAM3本身支持通过可重用文本提示进行概念分割，但在医学影像中的直接应用受到细粒度临床知识不足和人类撰写描述模糊性的限制。在本研究中，我们提出了Mask to Concept (M2C)，这是一个高效的框架，能够在没有外部模块、参数重训练或手动文本工程的情况下，适应SAM3进行医学少量标注。M2C仅使用少量标记图像，使SAM3能够在其冻结架构内自动搜索可转移的视觉概念：它初始化一个可学习的概念嵌入，利用该嵌入进行分割提示，并通过最小化概念分割误差的梯度更新嵌入。我们进一步引入了混合不确定性估计（Hybrid Uncertainty Estimation, HUE）模块，该模块计算预测熵并将概念预测映射回框提示，测量概念与几何提示之间的一致性。高度不确定的样本会被主动标记以供人工修正，修正后的掩膜随后反馈给M2C，以持续搜索更精确的概念嵌入，形成一个自增强的标注循环，所需的专家努力最小化。在医学分割基准上的实验表明，我们的方法实现了SOTA的少量分割性能和卓越的标注效率，为可扩展的医学图像标注提供了一条实用且高效的路径。代码可在 https://github.com/Huster-Hq/M2C 获取。
View on arXiv Download PDF AI Translation
cs.CV / 34 / 2606.26715

Extracting Neural Materials from Multi-view Images

从多视角图像中提取神经材料
Youwang, Kim, Hasselgren, Jon, Kocsis, Peter, Weidlich, Andrea, Oh, Tae-Hyun, Munkberg, Jacob
Abstract
Neural materials can represent complex specular reflections and scattering effects in a compact, universal basis. However, acquiring and authoring such materials remains challenging. We present NeuMatEx, a differentiable inverse rendering method for extracting spatially varying neural materials from images. The nonlinear structure of neural material latent spaces makes optimization with naive inverse rendering infeasible. To address this, we train a Large Material Reconstruction Model (LMRM) that directly predicts initialbase color, neural material latents, and aleatoric uncertainty guides from images. This material prior provides a good initialization and better constrains our subsequent optimization using inverse path tracing. The predicted uncertainty further helps by anchoring high-confidence regions more tightly to the LMRM prediction, preventing lighting and complex specular effects from being baked into materials. Experiments on synthetic and real assets show that NeuMatEx extracts complex materials with better visual quality and material decomposition than PBR-based methods.
Chinese Translation
神经材料能够以紧凑的通用基表示复杂的镜面反射和散射效应。然而，获取和创作这种材料仍然具有挑战性。我们提出了NeuMatEx，这是一种可微分的逆渲染方法，用于从图像中提取空间变化的神经材料。神经材料潜在空间的非线性结构使得使用简单的逆渲染进行优化变得不可行。为了解决这个问题，我们训练了一个大型材料重建模型（Large Material Reconstruction Model, LMRM），该模型可以直接从图像中预测初始基础颜色、神经材料潜在变量和随机不确定性引导。这个材料先验提供了良好的初始化，并更好地约束我们后续使用逆路径追踪的优化。预测的不确定性通过将高置信度区域更紧密地锚定到LMRM预测上，进一步帮助防止光照和复杂的镜面效应被固化到材料中。在合成和真实资产上的实验表明，NeuMatEx提取的复杂材料在视觉质量和材料分解方面优于基于PBR的方法。
View on arXiv Download PDF AI Translation
cs.CV / 35 / 2606.26719

Full spectrum Unlearnable Examples via Spectral Equalization

全谱不可学习示例通过谱均衡
Cai, Jiale, Xu, Gezheng, Li, Zhihao, Fang, Ruiyi, Pu, Ruizhi, Wu, Di, Lao, Qicheng, Ling, Charles, Wang, Boyu
Abstract
Unlearnable examples (UEs) protect training data by injecting imperceptible perturbations so that models fail to extract exploitable representations. In this paper, we reveal that existing UEs exhibit a critical failure once low-pass filtering is applied, indicating that the effective perturbation signals for unlearnability concentrate predominantly in high frequencies. Hence, we argue that reliable UEs should remain effective across the full spectrum. To this end, we propose Full-spectrum Unlearnable examples via Spectral Equalization (FUSE), which aims to generate spectrum-agnostic perturbations by equalizing the contributions from different bands and enforcing cross-band consistency. Specifically, FUSE adopts a Random Spectral Masking (RSM) strategy during generator training, which randomly removes a contiguous frequency band, forcing the remaining bands to maintain unlearnability. In addition, FUSE further integrates Cross-Band Guidance (CBG), which enforces mutual consistency between high- and low-frequency components, thereby further enhancing low-frequency unlearnability and regulating high-frequency perturbations to preserve the semantic fidelity of images. Extensive experiments across multiple datasets, architectures, and spectral filtering demonstrate the strong protection achieved by FUSE.
Chinese Translation
不可学习示例（UEs）通过注入不可察觉的扰动来保护训练数据，使得模型无法提取可利用的表示。在本文中，我们揭示了现有的UEs在应用低通滤波后表现出严重的失效，表明不可学习性的有效扰动信号主要集中在高频部分。因此，我们认为可靠的UEs应在全谱范围内保持有效。为此，我们提出了通过谱均衡生成全谱不可学习示例（FUSE），旨在通过均衡不同频段的贡献并强制跨频段一致性来生成与频谱无关的扰动。具体而言，FUSE在生成器训练过程中采用随机谱掩蔽（RSM）策略，随机移除一个连续的频率带，迫使剩余频带保持不可学习性。此外，FUSE进一步整合了跨频段引导（CBG），强制高频和低频成分之间的相互一致性，从而进一步增强低频不可学习性，并调节高频扰动以保持图像的语义保真性。在多个数据集、架构和谱滤波的广泛实验中，证明了FUSE实现的强大保护效果。
View on arXiv Download PDF AI Translation
cs.CV / 36 / 2606.26734

Robust Onion: Peeling Open Vocab Object Detectors Under Noise

鲁棒洋葱：揭示噪声下的开放词汇物体检测器
Pathak, Priyank, Karuppasamy, Mukilan, Baranwal, Aaditya, Vyas, Shruti, Rawat, Yogesh S
Abstract
The impact of real-world noise on Open Vocabulary Object Detectors (OV-ODs) remains poorly understood due to their architectural complexity. We present our comprehensive analysis Robust Onion, an empirical study that uses controlled synthetic visual degradations to peel OV-ODs layer-by-layer, revealing how, why, and where robustness degrades, systematically analyzing feature collapse. Our findings reveal that models with similar vision backbones exhibit comparable robustness, driven by similar feature collapse at similar layers, while factors such as pretraining strategy, architectural nuances, and caption supervision contribute little. Robustness is primarily governed by the image domain rather than annotations, explaining the similar robustness impact on COCO and LVIS, and why datasets like ODinW-13 can give an impression of inflated robustness due to large, isolated objects. Finally, we validate our insights by improving robustness on real-world BDD100K, WiderFace, and VisDRONE via our lightweight plug-and-play NN & TK0 approach, using 96x fewer trainable parameters than end-to-end training. We also explain the prior works' robustness observations.
Chinese Translation
由于开放词汇物体检测器（OV-ODs）的架构复杂性，现实世界噪声对其影响仍然不够清楚。我们提出了全面分析“鲁棒洋葱”的实证研究，利用受控的合成视觉降解逐层揭示OV-ODs的鲁棒性如何、为何以及在何处退化，系统分析特征崩溃。我们的研究结果表明，具有相似视觉骨干的模型表现出可比的鲁棒性，这主要是由于在相似层次上发生了相似的特征崩溃，而预训练策略、架构细节和标题监督等因素贡献甚微。鲁棒性主要受图像域的影响，而非注释，这解释了COCO和LVIS上相似的鲁棒性影响，以及为何像ODinW-13这样的数据集会因大型孤立物体而给人以虚高鲁棒性的印象。最后，我们通过我们的轻量级即插即用神经网络（NN）和TK0方法，在真实世界的BDD100K、WiderFace和VisDRONE上验证了我们的见解，使用的可训练参数比端到端训练少96倍。我们还解释了先前研究的鲁棒性观察结果。
View on arXiv Download PDF AI Translation
cs.CV / 37 / 2606.26738

Do Image Editing Models Understand Lighting?

图像编辑模型是否理解光照？
Küchler, Tim, Feiden, Johann-Friedrich, Nießner, Matthias, Rother, Carsten
Abstract
While recent advancements in generative image editing models have achieved stunning visual fidelity, it remains an open question whether these systems possess an intrinsic knowledge of real-world lighting. Existing benchmarks typically evaluate high-level plausibility of perceptual light transport on curated internet imagery, using VLMs or human judgement, or they rely on synthetically generated datasets. In this work, we introduce the 3D-anchored Light Probe (3DLP) benchmark, for which we have captured a new high-fidelity HDR dataset of real-world lighting changes. The dataset consists of 1K image pairs of diverse indoor scenery in which light probes are physically turned on and off. To allow for a granular performance analysis, we annotated specific image regions such as cast shadows or metallic surfaces. With this data, we evaluate a range of state-of-the-art image editing models by measuring how well their light probe edits align with reality. The evaluation uses two new scores to compensate for AI-generated photographic effects, such as adjusted white balance. Our results show that the overall performance of models differs considerably, with differences slightly less pronounced for specular highlights. The best image editing models are remarkably consistent with real-world physics, however, they still leave room for improvement. We observe that image regions that receive less light from the light probe are more prone to errors for all models. Furthermore, building on their success in evaluating macroscopic lighting plausibility, we test VLMs on our task but find that they are unsuitable for pixel-level light transport analysis. We will make the benchmark, together with the real-world dataset, publicly available to encourage future research on this topic.
Chinese Translation
尽管近年来生成图像编辑模型在视觉逼真度方面取得了惊人的进展，但这些系统是否具备对现实世界光照的内在理解仍然是一个悬而未决的问题。现有的基准测试通常评估在策划的互联网图像上感知光传输的高层次合理性，使用视觉语言模型（VLMs）或人类判断，或者依赖于合成生成的数据集。在本研究中，我们引入了3D锚定光探针（3DLP）基准测试，我们捕获了一个新的高保真HDR数据集，记录了现实世界中的光照变化。该数据集由1000对多样化室内场景的图像组成，其中光探针被物理地开启和关闭。为了进行细致的性能分析，我们对特定图像区域进行了标注，例如投射阴影或金属表面。利用这些数据，我们评估了一系列最先进的图像编辑模型，测量它们的光探针编辑与现实的对齐程度。评估使用了两个新的评分，以补偿AI生成的摄影效果，例如调整的白平衡。我们的结果表明，模型的整体性能差异显著，对于镜面高光的差异则稍微不那么明显。尽管最佳的图像编辑模型与现实世界的物理规律表现出显著的一致性，但仍有改进的空间。我们观察到，接收来自光探针的光照较少的图像区域对于所有模型更容易出现错误。此外，基于它们在评估宏观光照合理性方面的成功，我们在我们的任务上测试了VLMs，但发现它们不适合进行像素级光传输分析。我们将公开该基准测试以及真实世界数据集，以鼓励未来在这一主题上的研究。
View on arXiv Download PDF AI Translation
cs.CV / 38 / 2606.26740

LiveEdit: Towards Real-Time Diffusion-Based Streaming Video Editing

LiveEdit：面向实时扩散基础的流媒体视频编辑
Wang, Xinyu, Zhao, Chongbo, Zhan, Fangneng, Ma, Yue
Abstract
Streaming video editing has made rapid progress, yet practical deployment is still limited by two core issues: maintaining stable backgrounds and non-edited regions over time, and achieving the low latency required for real-time interactive scenarios. Meanwhile, recent streaming video generation methods are mostly developed for synthesis and cannot be directly applied to editing due to the strict preservation requirement and region-specific control. In this work, we present a novel streaming video editing framework that performs causal, frame-by-frame editing with strong content preservation and real-time responsiveness. Our key design is a three-stage distillation pipeline that progressively transfers editing capability from a powerful bidirectional foundation model to an efficient unidirectional streaming editor, enabling stable long-horizon edits without sacrificing visual fidelity. To further support real-time deployment, we introduce an AR-oriented mask cache that reuses region-related computation across frames, substantially reducing redundant processing and accelerating inference. Finally, we establish a dedicated benchmark for streaming video editing. Extensive evaluations demonstrate that our method achieves state-of-the-art visual quality among streaming baselines while drastically boosting inference speed to 12.66 FPS, making it suitable for interactive and augmented reality applications.
Chinese Translation
流媒体视频编辑已经取得了快速进展，但实际部署仍然受到两个核心问题的限制：在时间上保持稳定的背景和未编辑区域，以及实现实时交互场景所需的低延迟。同时，最近开发的流媒体视频生成方法主要用于合成，无法直接应用于编辑，因为它们对内容保留和区域特定控制有严格要求。在本研究中，我们提出了一种新颖的流媒体视频编辑框架，该框架实现了因果的逐帧编辑，具有强大的内容保留能力和实时响应性。我们的关键设计是一个三阶段的蒸馏管道，它逐步将编辑能力从一个强大的双向基础模型转移到一个高效的单向流媒体编辑器，从而实现稳定的长时间编辑而不牺牲视觉保真度。为了进一步支持实时部署，我们引入了一个面向增强现实的掩膜缓存，重用跨帧的区域相关计算，显著减少冗余处理并加速推理。最后，我们建立了一个专门的流媒体视频编辑基准。大量评估表明，我们的方法在流媒体基线中实现了最先进的视觉质量，同时将推理速度大幅提升至12.66 FPS，使其适用于交互式和增强现实应用。
View on arXiv Download PDF AI Translation
cs.CV / 39 / 2606.26743

Depth-Semantic Alignment and Affinity-Guided Fusion for Structured Radar Point Cloud Generation

深度-语义对齐与亲和引导融合用于结构化雷达点云生成
Hussain, Amjad, Qiu, Xin, Ai, Fuyuan, Tan, Yuchen, Li, Zecheng, Song, Chunyi, Liu, Wenjie
Abstract
Point clouds are an important carrier of three-dimensional spatial information, and their quality directly affects the performance of downstream perception tasks such as object detection and tracking. However, millimeter-wave radar point clouds are typically sparse, noisy, and structurally incomplete. To address these limitations, this paper proposes a multimodal point cloud generation method based on vision-radar fusion. The proposed method leverages image semantic information to impose structural constraints and achieve spatial alignment for radar point clouds, while incorporating a sparse completion strategy to enhance point density and recover missing structures. The generated point clouds are further evaluated in object detection and tracking tasks. Experimental results demonstrate that the proposed method effectively improves point cloud quality and enhances the detection accuracy and robustness of perception models in complex environments, providing a practical solution for multisensor point cloud generation and intelligent perception systems.
Chinese Translation
点云是三维空间信息的重要载体，其质量直接影响下游感知任务的性能，如目标检测和跟踪。然而，毫米波雷达点云通常稀疏、噪声大且结构不完整。为了解决这些局限性，本文提出了一种基于视觉-雷达融合的多模态点云生成方法。该方法利用图像语义信息施加结构约束，实现雷达点云的空间对齐，同时结合稀疏补全策略以增强点密度并恢复缺失的结构。生成的点云在目标检测和跟踪任务中进一步评估。实验结果表明，所提方法有效提高了点云质量，并增强了感知模型在复杂环境中的检测准确性和鲁棒性，为多传感器点云生成和智能感知系统提供了实用解决方案。
View on arXiv Download PDF AI Translation
cs.CV / 40 / 2606.26754

Capacity-Controlled Multi-View Stylization of 3D Gaussian Splatting

容量控制的多视角3D高斯点云风格化
Wen, Zhihao, Yang, Yixin, Wu, Bojian, Zhou, Yang, Lischinski, Dani, Cohen-Or, Daniel, Huang, Hui
Abstract
While 3D Gaussian Splatting (3DGS) provides an efficient and explicit representation for novel view synthesis, enforcing stylistic coherence across viewpoints remains challenging. Existing 3D stylization methods typically apply 2D feature-matching losses independently per rendered view, which leads to unstable style allocation, many-to-one feature reuse, and limited cross-view consistency. We propose a capacity-controlled framework for multi-view stylization of 3DGS, grounded in optimal transport. Specifically, we reformulate local style matching as a semi-balanced optimal transport problem. By introducing explicit column-capacity constraints with tunable strength, our formulation mitigates many-to-one matching and enables controllable allocation of style features. This transport-based objective provides a principled mechanism for balancing feature coverage and stylistic diversity while maintaining stable correspondences across viewpoints. To further enhance cross-view coherence, we incorporate a novel cross-view matching guidance to constrain correspondences between scene content and style patterns. In addition, we introduce several geometric regularizations to enhance the vanilla 3DGS, thereby enabling optimized Gaussian primitives to represent finer-grained textures during stylization. Extensive experiments demonstrate that our approach significantly improves multi-view stylistic consistency and produces stable, expressive 3D stylizations while preserving the core semantic structure of the scene.
Chinese Translation
尽管3D高斯点云（3D Gaussian Splatting, 3DGS）为新视角合成提供了高效且明确的表示，但在不同视角之间保持风格一致性仍然具有挑战性。现有的3D风格化方法通常对每个渲染视角独立应用2D特征匹配损失，这导致风格分配不稳定、多对一特征重用以及有限的视角间一致性。我们提出了一种基于最优传输的容量控制框架，用于3DGS的多视角风格化。具体而言，我们将局部风格匹配重新表述为一个半平衡的最优传输问题。通过引入具有可调强度的显式列容量约束，我们的公式减轻了多对一匹配，并实现了风格特征的可控分配。这种基于传输的目标为平衡特征覆盖和风格多样性提供了一个原则性机制，同时保持了视角间的稳定对应关系。为了进一步增强视角间的一致性，我们引入了一种新颖的视角间匹配指导，以约束场景内容与风格模式之间的对应关系。此外，我们引入了几种几何正则化来增强基础的3DGS，从而使优化后的高斯原语能够在风格化过程中表示更细致的纹理。大量实验表明，我们的方法显著提高了多视角风格一致性，并在保持场景核心语义结构的同时，生成稳定且富有表现力的3D风格化效果。
View on arXiv Download PDF AI Translation
cs.CV / 41 / 2606.26762

ProtoKV: Streaming Video Understanding under Delayed Query with Summary-State Memory

ProtoKV：在延迟查询下的流媒体视频理解与摘要状态记忆
Minh, Le Tu Ngoc, Lim, Jinyeong, Han, Dongsu
Abstract
Streaming video understanding (SVU) must answer queries that arrive asynchronously while visual tokens stream continuously under strict GPU-memory and query-time latency budgets. A key challenge is delayed query: decisive cues may appear briefly, yet many subsequent updates occur before the query arrives, increasing the risk that those cues are evicted or diluted under bounded memory. We propose ProtoKV, a constant-footprint SVU memory that represents far history as a fixed-capacity summary state rather than retaining token instances. ProtoKV keeps an exact near-window KV cache and aggregates older content into a semantic-spatial prototype bank with residual statistics. At query time, each prototype is exposed through a bounded pseudo-token interface that is drop-in compatible with standard attention. Under matched budgets and comparable query-time cost, ProtoKV improves accuracy by up to 12.5 points over token-retention baselines on SVU benchmarks in the long-delay regime, with gains that grow as query delay increases.
Chinese Translation
流媒体视频理解（SVU）必须在严格的GPU内存和查询时间延迟预算下，回答异步到达的查询，同时视觉标记持续流动。一个主要挑战是延迟查询：决定性线索可能会短暂出现，但在查询到达之前，许多后续更新会发生，这增加了在有限内存下这些线索被驱逐或稀释的风险。我们提出了ProtoKV，一种恒定占用内存的SVU记忆，它将远期历史表示为固定容量的摘要状态，而不是保留标记实例。ProtoKV保持一个精确的近窗口KV缓存，并将较旧的内容聚合到带有残余统计信息的语义空间原型库中。在查询时，每个原型通过一个有限的伪标记接口暴露，该接口与标准注意力机制兼容。在匹配的预算和可比的查询时间成本下，ProtoKV在长延迟范围内的SVU基准测试中，相比于标记保留基线，准确性提高了多达12.5个百分点，且随着查询延迟的增加，增益也在增长。
View on arXiv Download PDF AI Translation
cs.CV / 42 / 2606.26763

Calibrated Harmonic Overlaid Implicit Neural Representations for Multi-Dimensional Data

用于多维数据的校准谐波叠加隐式神经表示
Chen, Honghang, Zhang, Xiujun, Sun, Xiaoli, Xiao, Mingqing
Abstract
Implicit neural representation (INR) has emerged as a powerful prior for multi-dimensional data (e.g., multispectral images and videos). However, most INR methods employing periodic activation functions (e.g., Sine) predominantly rely on function composition. This mechanism introduces optimization instability as network depth increases, thereby limiting their performance. Meanwhile, these methods fail to incorporate proper physical priors to effectively alleviate spectrum bias. To address these issues, inspired by the commonalities between deep periodic networks and generalized Fourier series, we propose a novel Calibrated Harmonic Overlaid Implicit Neural Representation (CHOIR). Specifically, we utilize Coordinated Harmonic Superposition (CHS) to replace the conventional function composition used in most INRs, thereby ensuring optimization stability when scaling network depth. Furthermore, we introduce a Perceptual Spectrum Calibration (PSC) to mitigate spectrum bias. This calibration embeds the ubiquitous power-law spectrum prior of natural images and adjusts the globally fixed spectrum towards a physically plausible log-uniform distribution. Extensive experiments on various multidimensional data recovery problems demonstrate that our method achieves superior performance over state-of-the-art approaches. Code is available at https://github.com/chorl0229/CHOIR.
Chinese Translation
隐式神经表示（INR）已成为多维数据（例如，多光谱图像和视频）的强大先验。然而，大多数采用周期性激活函数（例如，正弦函数）的INR方法主要依赖于函数组合。这种机制在网络深度增加时引入了优化不稳定性，从而限制了其性能。同时，这些方法未能有效地结合适当的物理先验来缓解光谱偏差。为了解决这些问题，我们受到深度周期网络与广义傅里叶级数之间共性启发，提出了一种新颖的校准谐波叠加隐式神经表示（CHOIR）。具体而言，我们利用协调谐波叠加（CHS）替代大多数INR中使用的传统函数组合，从而在扩展网络深度时确保优化的稳定性。此外，我们引入了一种感知光谱校准（PSC）来减轻光谱偏差。该校准嵌入了自然图像普遍存在的幂律光谱先验，并将全局固定光谱调整为物理上合理的对数均匀分布。在各种多维数据恢复问题上的广泛实验表明，我们的方法在性能上优于最先进的方法。代码可在 https://github.com/chorl0229/CHOIR 获取。
View on arXiv Download PDF AI Translation
cs.CV / 43 / 2606.26764

Anatomy-Guided Residual Motion Diffusion for Controllable 4D Cardiac MRI Synthesis

基于解剖学指导的残余运动扩散用于可控的4D心脏MRI合成
Cao, Yiheng, Andrade-Miranda, Gustavo, Zhang, Jiatian, Zhao, Lingxiao, Gao, Xin
Abstract
Developing robust artificial intelligence models for 4D (3D + time) medical imaging is constrained by limited annotated data, inter-device domain shifts, and privacy restrictions. To address this, we propose a 4D controllable generative framework for anatomically consistent data augmentation. A semi-supervised variational autoencoder learns a compact latent representation of anatomical volumes while jointly predicting aligned segmentation masks in a unified framework. Anatomical structure is then disentangled from temporal dynamics through a cascaded latent diffusion model (LDM). A static LDM generates subject-specific anatomy conditioned on clinical priors (diagnosis and volumes measures) and a subsequent motion LDM estimates residual latent motions, ensuring strict temporal coherence across the 4D sequence. The proposed approach was evaluated on cine cardiac MRI as a representative 4D imaging application. Experiments across multiple datasets demonstrate high controllability of static anatomy (Pearson r > 0.8) and strong temporal coherence (FVD = 288.08). In cross-vendor generalization experiments, augmenting training sets with synthetic 4D sequences significantly improves downstream segmentation performance. Using nnU-Net, the proposed augmentation strategy improves the average Dice score by 1.4% and reduces the Hausdorff Distance by 3.0mm compared to training on real data alone, for the left ventricle, Dice improves by 2.8% with a 5.4mm reduction in boundary error. Overall, this framework provides a scalable and controllable solution for 4D medical image synthesis, supporting the development of more robust models with limited annotations and cross-vendor variability. Code available on https://github.com/cyiheng/4DCardiacMRISynthesis.
Chinese Translation
开发稳健的人工智能模型用于4D（3D + 时间）医学成像受到有限标注数据、设备间领域转移和隐私限制的制约。为此，我们提出了一种用于解剖学一致性数据增强的4D可控生成框架。一个半监督变分自编码器在统一框架中学习解剖体积的紧凑潜在表示，同时预测对齐的分割掩膜。然后，通过级联的潜在扩散模型（LDM）将解剖结构与时间动态解耦。静态LDM生成基于临床先验（诊断和体积测量）的特定于受试者的解剖结构，随后运动LDM估计残余潜在运动，确保4D序列中的严格时间一致性。所提方法在心脏MRI影像（cine cardiac MRI）上进行了评估，作为代表性的4D成像应用。多个数据集上的实验表明静态解剖结构的高可控性（Pearson r > 0.8）和强时间一致性（FVD = 288.08）。在跨供应商泛化实验中，通过合成4D序列增强训练集显著提高下游分割性能。使用nnU-Net，所提增强策略使平均Dice分数提高了1.4%，并将Hausdorff距离减少了3.0mm；对于左心室，Dice提高了2.8%，边界误差减少了5.4mm。总体而言，该框架为4D医学图像合成提供了可扩展和可控的解决方案，支持在有限标注和跨供应商变异性下开发更稳健的模型。代码可在 https://github.com/cyiheng/4DCardiacMRISynthesis 获取。
View on arXiv Download PDF AI Translation
cs.CV / 44 / 2606.26778

LearniBridge: Learnable Calibration of Feature Caching for Diffusion Models Acceleration

LearniBridge：可学习的特征缓存校准用于扩散模型加速
Huang, Xuyue, Chen, Zhe, Shen, Wang, Zhang, Xiao-Ping
Abstract
Diffusion Transformers (DiTs) have driven substantial progress in image and video generation but suffer from prohibitive computational costs. Feature caching accelerates inference by reusing intermediate representations. Existing methods rely on historical features for implementation simplicity, yet suffer from severe error accumulation at high acceleration ratios. To address this limitation, we investigate the nature of the requisite feature correction. We demonstrate that the optimal calibration update is characterized by a shared low-rank subspace across diverse prompts. Guided by this structural insight, we propose LearniBridge, a learnable calibration mechanism for feature caching that bridges multiple timesteps through lightweight LoRA updates. This mechanism enables effective calibration requiring only 3-5 training samples. Extensive experiments on image and video generation show that LearniBridge achieves up to $5.87\times$, $5.75\times$, and $4.10\times$ acceleration on FLUX, HunyuanVideo, and WAN2.1, respectively. On WAN2.1, it improves VBench by 1.28% over the previous SOTA at $4.10\times$ acceleration. Our code is available at https://github.com/Iiiiiiirene/LearniBridge.
Chinese Translation
扩散变换器（Diffusion Transformers, DiTs）在图像和视频生成方面取得了显著进展，但面临着高昂的计算成本。特征缓存通过重用中间表示来加速推理。现有方法依赖于历史特征以简化实现，但在高加速比下遭遇严重的误差累积。为了解决这一限制，我们研究了所需特征校正的本质。我们证明了最佳校准更新的特征是跨多种提示共享的低秩子空间。基于这一结构性洞察，我们提出了LearniBridge，一种用于特征缓存的可学习校准机制，通过轻量级的LoRA更新连接多个时间步。该机制仅需3-5个训练样本即可实现有效校准。在图像和视频生成的广泛实验中，LearniBridge在FLUX、HunyuanVideo和WAN2.1上分别实现了高达5.87倍、5.75倍和4.10倍的加速。在WAN2.1上，它在4.10倍加速下比之前的SOTA提高了1.28%的VBench。我们的代码可在https://github.com/Iiiiiiirene/LearniBridge获取。
View on arXiv Download PDF AI Translation
cs.CV / 45 / 2606.26780

Event-based Gaze Control System for Accurate Real-time Spin Estimation in Professional Ball Games

基于事件的凝视控制系统用于专业球类运动中的实时旋转估计
Hu, Yunpu, Schilling, Fabian, Cavinato, Valentina, Aydin, Asude, Politis, Agis, Morales, Ricardo Tapiador, Scheper, Kirk Y. W., Dürr, Peter, Takahashi, Naoya
Abstract
Spin plays a crucial role in many ball sports due to its effect on the trajectory of the ball. Vision-based estimation of the ball's spin during a game with conventional cameras is challenging due to the ball's small size, high speed, and fast rotation. To address these challenges, we propose an event-based active vision system that can track unmodified balls and measure their spin in real-time. The system consists of an event camera for its high temporal resolution and minimal motion blur, high-speed pan/tilt galvanometer mirrors to keep the ball in the field of view, and a low-latency focus-tunable telephoto lens to increase the spatial resolution on the ball and keep it in focus. To track the ball, we use a hybrid approach that combines 2D event-based detection for centering and 3D positions from a ball localization system for re-initialization. For high-accuracy spin estimation, we propose an offline method that performs contrast maximization on the sphere (s-CMax). This method achieves state-of-the-art accuracy on static balls across multiple sports (table tennis, baseball, tennis, and golf), with mean magnitude and axis errors of 2.1% and 4.0 degrees, respectively. We then develop a low-latency online method for table tennis as a case study in real-time applications. This method uses an uncertainty-aware convolutional neural network trained on pseudo-ground-truth spin labels from the offline approach, combined with a GPU-accelerated batch implementation of contrast maximization for refinement. We demonstrate reliable tracking and spin estimation with a three-view setup during professional table tennis matches, with high accuracy (8.8% magnitude and 6.4 degrees axis mismatch), 3 ms latency, and 750 Hz throughput.
Chinese Translation
旋转在许多球类运动中起着至关重要的作用，因为它会影响球的轨迹。在比赛中，利用传统相机对球的旋转进行视觉估计面临挑战，原因在于球体积小、速度快且旋转迅速。为了解决这些挑战，我们提出了一种基于事件的主动视觉系统，该系统能够实时跟踪未改装的球并测量其旋转。该系统由一个事件相机组成，具有高时间分辨率和最小运动模糊，配备高速的平移/俯仰电动镜面，以保持球在视野内，以及一个低延迟的可调焦距长焦镜头，以提高球的空间分辨率并保持其聚焦。为了跟踪球，我们采用了一种混合方法，结合了用于中心定位的二维事件检测和来自球定位系统的三维位置进行重新初始化。为了实现高精度的旋转估计，我们提出了一种离线方法，在球体上执行对比度最大化（s-CMax）。该方法在多个运动（乒乓球、棒球、网球和高尔夫）中对静态球实现了最先进的准确性，平均幅度和轴误差分别为2.1%和4.0度。随后，我们开发了一种低延迟的在线方法，以乒乓球为案例研究，应用于实时场景。该方法使用了一种不确定性感知的卷积神经网络，该网络在离线方法的伪真实旋转标签上进行训练，并结合了加速的批量对比度最大化实现进行细化。我们在专业乒乓球比赛中展示了可靠的跟踪和旋转估计，三视角设置下实现了高准确性（8.8%幅度和6.4度轴偏差），延迟为3毫秒，吞吐量为750赫兹。
View on arXiv Download PDF AI Translation
cs.CV / 46 / 2606.26794

ReasonCLIP-58M: Visually Grounded Commonsense Reasoning Supervision for CLIP

ReasonCLIP-58M：用于 CLIP 的视觉基础常识推理监督
Zhang, Sicheng, Naseer, Muzammal, Xie, Binzhu, Suryanto, Naufal, Qiu, Shi, Bentahar, Jamal, Akhtar, Naveed, Shah, Mubarak
Abstract
CLIP and its variants are widely adopted visual backbones in multimodal systems, but their pretraining remains dominated by descriptive image-text alignment. As downstream applications increasingly demand visually grounded commonsense inference and compositional reasoning, it remains unclear whether CLIP-style encoders can support such reasoning without architectural changes. To address this, we present ReasonCLIP-58M, a continual pretraining framework that integrates large-scale reasoning supervision into CLIP-style models through our two-stage strategy, which progressively integrates reasoning signals while preserving descriptive alignment, followed by category-structured reasoning supervision. To support this framework, we construct two complementary datasets and a benchmark: ReasonLite-42M, with open-form, visually verifiable reasoning captions; ReasonPro-16M, with category-specific reasoning supervision; and RCLIP-Bench for diagnostic evaluation of visually grounded reasoning. We train a family of ReasonCLIP that improves visually grounded commonsense and compositional reasoning while also enhancing zero-shot retrieval performance. As a drop-in visual encoder for multimodal large language models such as LLaVA-NeXT, ReasonCLIP delivers consistent gains without additional inference cost, demonstrating that structured reasoning supervision enhances the expressive capacity of CLIP-style visual representations. All datasets, models, and training code are available at https://github.com/RISys-Lab/ReasonCLIP.
Chinese Translation
CLIP 及其变体在多模态系统中被广泛采用作为视觉骨干，但其预训练仍主要以描述性图像-文本对齐为主。随着下游应用对视觉基础常识推理和组合推理的需求日益增加，目前尚不清楚 CLIP 风格的编码器是否能够在不改变架构的情况下支持这种推理。为了解决这个问题，我们提出了 ReasonCLIP-58M，这是一个持续预训练框架，通过我们的两阶段策略将大规模推理监督集成到 CLIP 风格模型中，该策略逐步整合推理信号，同时保持描述性对齐，随后进行类别结构化的推理监督。为了支持这一框架，我们构建了两个互补的数据集和一个基准：ReasonLite-42M，包含开放式、可视验证的推理标题；ReasonPro-16M，包含类别特定的推理监督；以及 RCLIP-Bench，用于视觉基础推理的诊断评估。我们训练了一系列 ReasonCLIP 模型，这些模型在提升视觉基础常识和组合推理的同时，也增强了零-shot 检索性能。作为多模态大型语言模型（如 LLaVA-NeXT）的即插即用视觉编码器，ReasonCLIP 在不增加额外推理成本的情况下提供了一致的性能提升，证明了结构化推理监督增强了 CLIP 风格视觉表示的表达能力。所有数据集、模型和训练代码均可在 https://github.com/RISys-Lab/ReasonCLIP 获取。
View on arXiv Download PDF AI Translation
cs.CV / 47 / 2606.26795

NaviCache: Test-Time Self-Calibration Caching for Video Generation

NaviCache：视频生成的测试时自校准缓存
Lv, Zheqi, Zhu, Zhibo, Wang, Jinke, Tian, Qi, Zhang, Shengyu, Chen, Zhengyu, Zang, Chengxi, Zhao, Zhou, Wu, Fei
Abstract
Video Diffusion Models (VDMs) is constrained by immense computational costs. While offline calibration-based acceleration suffers from calibration data dependency, prohibitive calibration duration, and susceptibility to distribution shifts, offline calibration-free methods eliminate these hurdles. However, since they rely on instantaneous zero-order approximations where the mapping between input and output differences varies in real-time, they are susceptible to observational noise and ignore the intrinsic momentum within the diffusion trajectory. In this paper, we propose NaviCache, a plug-and-play test-time self-calibration method re-conceptualizing feature evolution as an Inertial Navigation System (INS) problem. NaviCache bridges the fundamental domain gap and the non-stationary nature of diffusion by modeling the relative coupling between input and output variations. We introduce a dual-state estimation architecture that adaptively tracks the feature change ratio and its latent drift, initialized via a specialized Initial Alignment phase. By integrating a time-dependent noise schedule with an uncertainty-aware Measurement Update mechanism, NaviCache provides a theoretically grounded mechanism for error-bounded computation skipping. Extensive experiments on the HunyuanVideo, Wan, and Open-Sora series demonstrate that NaviCache exhibits more accurate error judgment for computation skipping and achieves outstanding comprehensive performance.
Chinese Translation
视频扩散模型（VDMs）受到巨大的计算成本限制。虽然基于离线校准的加速方法受到校准数据依赖、校准时间过长以及对分布变化的敏感性等问题的困扰，但离线无校准方法消除了这些障碍。然而，由于它们依赖于瞬时的零阶近似，其中输入和输出差异之间的映射在实时中变化，因此它们容易受到观测噪声的影响，并忽略了扩散轨迹中的内在动量。在本文中，我们提出了NaviCache，一种即插即用的测试时自校准方法，将特征演变重新概念化为惯性导航系统（INS）问题。NaviCache通过建模输入和输出变化之间的相对耦合，弥合了基本领域差距和扩散的非平稳特性。我们引入了一种双状态估计架构，能够自适应地跟踪特征变化率及其潜在漂移，通过专门的初始对齐阶段进行初始化。通过将时间依赖的噪声调度与考虑不确定性的测量更新机制相结合，NaviCache提供了一种理论上有依据的机制，用于误差界限计算跳过。在HunyuanVideo、Wan和Open-Sora系列上的大量实验表明，NaviCache在计算跳过的误差判断上表现得更为准确，并且实现了卓越的综合性能。
View on arXiv Download PDF AI Translation
cs.CV / 48 / 2606.26812

Multi-modality Image Fusion under Adverse Weather: Mask-Guided Feature Restoration and Interaction

恶劣天气下的多模态图像融合：基于掩膜引导的特征恢复与交互
Li, Xilai, Li, Xiaosong, Tan, Haishu, Ye, Tao, Li, Huafeng, Wang, Hongbin
Abstract
Multi-modality image fusion (MMIF) enhances scene representation by exploiting complementary cues from different modalities. Adverse weather, however, causes significant image degradation, disrupting feature representation and requiring simultaneous feature restoration and cross-modal complementarity. Existing methods often struggle with effective representation learning under such conditions, limiting their practical performance. To address these challenges, we propose a mask-guided MMIF method that integrates feature restoration and interaction. We first introduce "Pseudo Ground Truth" to simplify training, promoting faster and more effective feature learning. Then, we design a mask generation mechanism based on the mapping relationship between the fused result and the source images, quantifying the relative contribution of each modality during the fusion process. By incorporating the proposed mask-guided cross-modal cross-attention mechanism, the network is encouraged to selectively attend to informative features during modality interaction, mitigating the risk of overfitting to the static distribution of the "Pseudo Ground Truth". Additionally, we propose a mask-guided learning strategy and a task-coupled degradation-aware learning strategy to balance feature restoration and interaction. Extensive experiments on synthetic and real-world datasets demonstrate that our method surpasses state-of-the-art approaches in visual quality, quantitative metrics, and downstream tasks. The source code is available at https://github.com/ixilai/AMG-Fuse.
Chinese Translation
多模态图像融合（MMIF）通过利用不同模态的互补线索来增强场景表示。然而，恶劣天气会导致显著的图像退化，破坏特征表示，并需要同时进行特征恢复和跨模态互补。现有方法在这种条件下往往难以有效地进行表示学习，限制了其实际性能。为了解决这些挑战，我们提出了一种基于掩膜引导的MMIF方法，该方法集成了特征恢复与交互。我们首先引入“伪地面真相”（Pseudo Ground Truth）以简化训练，促进更快和更有效的特征学习。然后，我们设计了一种基于融合结果与源图像之间映射关系的掩膜生成机制，在融合过程中量化每个模态的相对贡献。通过结合所提出的掩膜引导的跨模态交叉注意机制，网络被鼓励在模态交互过程中选择性地关注信息丰富的特征，从而减轻对“伪地面真相”静态分布的过拟合风险。此外，我们提出了一种掩膜引导学习策略和一种任务耦合的退化感知学习策略，以平衡特征恢复与交互。在合成和真实世界数据集上的大量实验表明，我们的方法在视觉质量、定量指标和下游任务上超越了最先进的方法。源代码可在 https://github.com/ixilai/AMG-Fuse 获取。
View on arXiv Download PDF AI Translation
cs.CV / 49 / 2606.26828

Learning Adversarial Augmentation Policies for Robust Garlic Seedling Detection

学习对抗增强策略以实现稳健的大蒜幼苗检测
Lee, Soeun, Kim, Chanho, Kang, Yeji, Hong, YoungKi, Kang, Byeongkeun
Abstract
Accurate seedling detection during early growth stages is essential for timely replanting and effective crop management in precision agriculture. However, existing studies are mostly evaluated under relatively stable imaging conditions, such as UAV imagery or greenhouse environments, leaving robust detection under severe and spatially heterogeneous illumination in ground-based outdoor monitoring insufficiently explored. In addition, many illumination-robust detection methods rely on additional enhancement or feature-extraction modules, which increase inference-time overhead and are not tailored to seedling detection and downstream missing seedling localization. To address these gaps, we construct a new garlic seedling dataset captured using a ground-based monitoring platform under real outdoor field conditions with highly variable illumination. We further propose an illumination-robust seedling detection framework based on adversarial augmentation policy learning. The proposed method jointly optimizes a stochastic augmentation policy agent and an object detector, enabling the detector to learn robust representations under challenging visual conditions. A structural penalty is introduced to prevent unrealistic distortions while encouraging challenging augmentations during training. Extensive experiments show that the proposed approach achieves an AP$_{50}$ of 91.6%, improving the baseline by 0.9 percentage points and outperforming the previous best-performing method by 0.2 percentage points. For downstream missing seedling localization, it achieves 75.0% precision and a 67.0% F1-score, improving the baseline by 4.8 and 2.0 percentage points, respectively. These results demonstrate the effectiveness of the proposed framework for practical ground-based agricultural monitoring under complex outdoor lighting conditions without additional inference-time computational overhead.
Chinese Translation
在早期生长阶段，准确的幼苗检测对于及时重新种植和有效的精准农业作物管理至关重要。然而，现有研究大多在相对稳定的成像条件下进行评估，例如无人机影像或温室环境，这使得在地面户外监测中应对严酷和空间异质照明的稳健检测尚未得到充分探索。此外，许多照明稳健检测方法依赖于额外的增强或特征提取模块，这增加了推理时间的开销，并且并未针对幼苗检测和下游缺失幼苗定位进行优化。为了解决这些问题，我们构建了一个新的大蒜幼苗数据集，该数据集是在真实户外田间条件下使用地面监测平台捕获的，具有高度可变的照明条件。我们进一步提出了一种基于对抗增强策略学习的照明稳健幼苗检测框架。该方法联合优化一个随机增强策略代理和一个目标检测器，使检测器能够在具有挑战性的视觉条件下学习稳健的表征。引入了一种结构惩罚，以防止不现实的扭曲，同时在训练过程中鼓励具有挑战性的增强。大量实验表明，所提出的方法在 AP$_{50}$ 上达到了 91.6%，比基线提高了 0.9 个百分点，并且比之前表现最好的方法提高了 0.2 个百分点。在下游缺失幼苗定位方面，精度达到了 75.0%，F1 分数为 67.0%，分别比基线提高了 4.8 和 2.0 个百分点。这些结果证明了所提出框架在复杂户外照明条件下进行实际地面农业监测的有效性，而无需额外的推理时间计算开销。
View on arXiv Download PDF AI Translation
cs.CV / 50 / 2606.26829

Identifying the Unknown: Prompt-Free Open Vocabulary Anomaly Recognition for Robot-Object Interaction

识别未知：无提示开放词汇异常识别用于机器人-物体交互
Allgeuer, Philipp, Habekost, Jan-Gerrit, Wermter, Stefan
Abstract
Robots operating in real-world environments must in general be able to recognize previously unseen objects. As robotic systems move toward open-world autonomy, there is a growing, yet largely unmet, need for open vocabulary object detectors that are prompt-free and efficient enough for continuous deployment. We present AnomNOVIC, a two-stage known-workspace framework that combines a masked autoencoder (MAE) trained for anomaly detection, with NOVIC, a powerful real-time prompt-free open vocabulary image classifier. The MAE produces generic object-agnostic bounding boxes, allowing NOVIC to classify salient image regions without requiring a predefined candidate class list. We evaluate AnomNOVIC against strong open vocabulary baselines in a tabletop robot-object environment featuring the NICOL humanoid robot, reaching 47.1% AP / 57.5% AP50 for prompt-free recognition, and 59.0% AP / 72.5% AP50 if class candidates are provided. Across additional datasets, including an in-the-wild test set with 48 unique objects, AnomNOVIC reaches up to 82.6% prompt-free detection and classification accuracy. These results significantly surpass all tested open vocabulary baselines, including YOLO-World-v2, OWLv2, and YOLOE.
Chinese Translation
在现实环境中操作的机器人通常必须能够识别以前未见过的物体。随着机器人系统向开放世界自主性发展，对无提示且足够高效的开放词汇物体检测器的需求日益增长，但这一需求在很大程度上尚未得到满足。我们提出了AnomNOVIC，这是一种两阶段已知工作空间框架，结合了用于异常检测的掩码自编码器（Masked Autoencoder, MAE）和强大的实时无提示开放词汇图像分类器NOVIC。MAE生成通用的物体无关边界框，使得NOVIC能够在不需要预定义候选类别列表的情况下对显著图像区域进行分类。我们在一个包含NICOL人形机器人的桌面机器人-物体环境中，将AnomNOVIC与强大的开放词汇基线进行评估，达到47.1%的AP / 57.5%的AP50用于无提示识别，如果提供类别候选，则达到59.0%的AP / 72.5%的AP50。在包括一个包含48个独特物体的真实环境测试集在内的额外数据集上，AnomNOVIC的无提示检测和分类准确率高达82.6%。这些结果显著超越了所有测试的开放词汇基线，包括YOLO-World-v2、OWLv2和YOLOE。
View on arXiv Download PDF AI Translation
cs.CV / 51 / 2606.26849

Liquid Fusion of Heterogeneous Representations Towards General Salient Object Detection

异构表示的液体融合用于通用显著性目标检测
Chen, Ke, Zhou, Ling, Jiang, Guangqi, Wu, Gengshen, Liu, Yi, Xu, Shoukun
Abstract
General Salient Object Detection (SOD) aims to identify and segment visually interesting objects from uni-modality or multi-modality scenes, recently advanced by cutting-edge State Space Models (SSMs). However, a critical limitation of current approaches is their neglect of the inherent spectral biases exhibited by different neural network paradigms. By digging to the dataset-level spectral analysis of Convolutional Neural Networks (CNNs) and SSMs, their semantic representations are inherently complementary based on their complementary frequency preferences. Inspired by this, we harmonize heterogeneous representations from SSMs and CNNs to bridge their spectral biases for general salient object detection. To this end, inspired by the dynamic information propagation of Liquid Neural Networks (LNNs), we introduce a liquid fusion to dynamically integrates features from two backbones, including VMamba and ConvNeXt, referred to Liquid Fusion Network (LFNet). Concretely, by treating the continuous VMamba features and ConvNeXt features as evolving states and exogenous stimulus, respectively, LFNet employs a dynamic gating mechanism for content-aware feature aggregation. Crucially, this state-stimulus paradigm enables to scale to multi-modal cues, resulting in flexibility in general SOD. Besides, a Saliency-Guided Upsampling (SGU) operator to propagate the features to the shallow layer, which leverages a spectral-spatial co-design to suppress upsampling artifacts while preserving semantics. Extensive experiments across five diverse tasks (RGB, RGB-D, RGB-T, VSOD, and VDT) demonstrate that LFNet achieves state-of-the-art performance, offering a superior trade-off between detection accuracy and model efficiency. Code has been released at https://github.com/cke520/LFNet.
Chinese Translation
通用显著性目标检测（SOD）旨在从单模态或多模态场景中识别和分割视觉上有趣的对象，最近通过前沿的状态空间模型（SSMs）得到了进展。然而，当前方法的一个关键限制是忽视了不同神经网络范式所表现出的固有光谱偏差。通过深入分析卷积神经网络（CNNs）和SSMs的数据集级光谱特征，它们的语义表示在其互补的频率偏好基础上是固有互补的。受此启发，我们协调来自SSMs和CNNs的异构表示，以弥合它们的光谱偏差，从而实现通用显著性目标检测。为此，受液体神经网络（LNNs）动态信息传播的启发，我们引入了一种液体融合方法，动态整合来自两个主干网络的特征，包括VMamba和ConvNeXt，称为液体融合网络（LFNet）。具体而言，通过将连续的VMamba特征和ConvNeXt特征视为演变状态和外部刺激，LFNet采用动态门控机制进行内容感知特征聚合。关键是，这种状态-刺激范式能够扩展到多模态线索，从而在通用SOD中实现灵活性。此外，采用显著性引导上采样（SGU）算子将特征传播到浅层，该算子利用光谱-空间协同设计来抑制上采样伪影，同时保留语义。针对五个不同任务（RGB、RGB-D、RGB-T、VSOD和VDT）的广泛实验表明，LFNet实现了最先进的性能，在检测准确性和模型效率之间提供了优越的权衡。代码已发布在 https://github.com/cke520/LFNet。
View on arXiv Download PDF AI Translation
cs.CV / 52 / 2606.26863

Rolling Shutter Relative Pose Estimation Made Practical

滚动快门相对位姿估计的实用化
Barath, Daniel
Abstract
Rolling shutter (RS) cameras equip virtually all consumer devices, yet RS-aware relative pose estimation has remained impractical: the state-of-the-art solver requires a minimum of 20 point correspondences, making RANSAC-based robust estimation prohibitively expensive due to the exponential dependence of the iteration count on the sample size. We make RS relative pose estimation practical by introducing affine correspondences (ACs) into the RS two-view geometry. We derive novel \emph{RS-corrected affine constraints} that account for the coupling between point perturbations and the row-dependent essential matrix, providing two equations per correspondence beyond the standard epipolar constraint. Building on these constraints, we develop a linearized algebraic solver that estimates pose and RS motion from only 7 ACs. The solver exploits the physical smallness of RS parameters to linearize the constraints, eliminates the 12 RS unknowns via null-space projection, and solves the remaining degree-20 system via action matrices in 1.2\,ms. On the TUM RS benchmark, our method achieves the best pose and RS parameter accuracy among all tested methods and, uniquely among RS solvers, provides accurate translational velocity estimates -- which are poorly conditioned from point correspondences alone due to a $\vec{v}$-$\vec{t}$ coupling. On the global-shutter EuRoC MAV dataset, the solver achieves comparable accuracy to the standard 5-point algorithm, demonstrating that it generalizes well to the GS setting. Code is at https://github.com/danini/rolling_shutter_made_practical.
Chinese Translation
滚动快门（RS）相机几乎装备了所有消费电子设备，但基于RS的相对位姿估计仍然不够实用：当前最先进的求解器需要至少20个点对应关系，使得基于RANSAC的鲁棒估计因迭代次数对样本大小的指数依赖而变得极其昂贵。我们通过将仿射对应关系（ACs）引入RS双视图几何，实用化了RS相对位姿估计。我们推导出新颖的 extit{RS修正仿射约束}，该约束考虑了点扰动与行依赖本质矩阵之间的耦合，为每个对应关系提供了超出标准极线约束的两个方程。在这些约束的基础上，我们开发了一种线性化代数求解器，仅需7个ACs即可估计位姿和RS运动。该求解器利用RS参数的物理小值对约束进行线性化，通过零空间投影消除12个RS未知数，并在1.2毫秒内通过作用矩阵求解剩余的20度系统。在TUM RS基准测试中，我们的方法在所有测试方法中实现了最佳的位姿和RS参数精度，并且在所有RS求解器中独树一帜地提供了准确的平移速度估计——由于$ extbf{v}$-$ extbf{t}$耦合，单靠点对应关系的条件较差。在全球快门EuRoC MAV数据集中，该求解器的精度与标准5点算法相当，证明其在GS设置下具有良好的泛化能力。代码可在 https://github.com/danini/rolling_shutter_made_practical 获取。
View on arXiv Download PDF AI Translation
cs.CV / 53 / 2606.26872

SpatialFlow-GRPO: Where Spatial Credit Drives Image Editing

SpatialFlow-GRPO：空间信用驱动的图像编辑
Yang, Yankai, Long, Yancheng, Chen, Wei, Lu, Xingyu, Wei, Hongyang, Wen, Bin, Yang, Fan, Gao, Tingting, Li, Han, Yang, Shuo
Abstract
Recent online reinforcement learning has substantially improved image editing quality. However, existing Flow-GRPO-style methods usually rely on a single whole-image reward, which makes fine-grained editing optimization difficult. We observe that a key obstacle in image editing is this spatial uniformity assumption: a whole-image reward cannot distinguish how different spatial regions contribute to image quality. To address this issue, we propose SpatialFlow-GRPO, a training framework that introduces spatially fine-grained reward feedback. The framework converts region-aware rewards into semantic-region-level optimization signals and aligns region advantages with the corresponding latent positions during policy updates. We also train a region-aware reward model, SFReward, construct SFReward-14K with region-annotated editing samples, and introduce MultiEditBench to evaluate multi-region editing ability. On OmniGen2 and FLUX.2-klein-4B, SpatialFlow-GRPO outperforms Flow-GRPO on GEdit-Bench, ImgEdit-Bench, and MultiEditBench. The results show that SpatialFlow-GRPO converts local feedback into spatially aligned update signals and improves editing quality.
Chinese Translation
最近的在线强化学习显著提高了图像编辑的质量。然而，现有的Flow-GRPO风格方法通常依赖于单一的全图奖励，这使得细粒度的编辑优化变得困难。我们观察到，图像编辑中的一个关键障碍是这种空间均匀性假设：全图奖励无法区分不同空间区域对图像质量的贡献。为了解决这个问题，我们提出了SpatialFlow-GRPO，一个引入空间细粒度奖励反馈的训练框架。该框架将区域感知奖励转换为语义区域级别的优化信号，并在策略更新过程中将区域优势与相应的潜在位置对齐。我们还训练了一个区域感知奖励模型SFReward，构建了包含区域标注编辑样本的SFReward-14K，并引入MultiEditBench来评估多区域编辑能力。在OmniGen2和FLUX.2-klein-4B上，SpatialFlow-GRPO在GEdit-Bench、ImgEdit-Bench和MultiEditBench上优于Flow-GRPO。结果表明，SpatialFlow-GRPO将局部反馈转化为空间对齐的更新信号，并提高了编辑质量。
View on arXiv Download PDF AI Translation
cs.CV / 54 / 2606.26885

RIS-Assisted Proactive Handover for Reliable mmWave Wireless Networks

基于RIS的主动切换以实现可靠的毫米波无线网络
Adnan, Alaa, Al-Quraan, Mohammad, Zoha, Ahmed, Butt, M. Majid, Muhaidat, Sami, Imran, Muhammad Ali, Di Renzo, Marco, Mohjazi, Lina
Abstract
Millimeter-wave (mmWave) networks are highly susceptible to line-of-sight (LoS) blockages. Vision-aided wireless communications (VAWC) enable proactive handovers (PHO) to mitigate such blockages; however, PHO becomes challenging when no nearby base station (BS) is available. In such cases, reconfigurable intelligent surfaces (RIS) can be used to restore connectivity. To ensure timely PHO, the RIS configuration time must be taken into account, as the large number of RIS elements can limit responsiveness in time-sensitive scenarios. This work proposes a novel RIS-assisted PHO approach that optimizes the number of allocated RIS elements to balance signal processing complexity and link quality under handover timing constraints, making the RIS-assisted link more energy-efficient. An optimization problem based on particle swarm optimization (PSO) is formulated to determine the optimal end-to-end RIS link setup that runs offline to bypass latency constraints. Results show that reducing the number of RIS elements by 12\% leads to a 10\% decrease in dissipated energy without compromising the signal-to-noise ratio (SNR). Moreover, the RIS-assisted link achieves a 15--30 dB improvement in blocked regions while maintaining accurate PHO timing.
Chinese Translation
毫米波（mmWave）网络对视距（LoS）阻塞高度敏感。视觉辅助无线通信（VAWC）使得主动切换（PHO）成为可能，以减轻此类阻塞；然而，当附近没有基站（BS）可用时，PHO变得具有挑战性。在这种情况下，可重构智能表面（RIS）可以用于恢复连接。为了确保及时的PHO，必须考虑RIS的配置时间，因为大量的RIS元素可能限制在时间敏感场景中的响应能力。本研究提出了一种新颖的基于RIS的PHO方法，优化分配的RIS元素数量，以在切换时限约束下平衡信号处理复杂性和链路质量，从而使基于RIS的链路更加节能。基于粒子群优化（PSO）的方法被制定为优化端到端RIS链路的设置，该设置在离线运行以规避延迟约束。结果表明，减少12\%的RIS元素数量可导致能量消耗减少10\\%，而不影响信噪比（SNR）。此外，基于RIS的链路在阻塞区域实现了15-30 dB的改善，同时保持了准确的PHO时机。
View on arXiv Download PDF AI Translation
cs.CV / 55 / 2606.26891

Bridging Vision and Language Concepts through Optimal Transport Semantic Flow

通过最优传输语义流桥接视觉与语言概念
Zhang, Chenyang, Dong, Anqi, Zhu, Guangming, Xiong, Nuoye, Wang, Siyuan, Mei, Lin, Zhang, Liang
Abstract
Concept Bottleneck Models (CBMs) promise transparent reasoning by predicting through human-interpretable concepts, yet their effectiveness fundamentally depends on how well visual and textual representations are aligned or matched. Existing vision-language CBMs often rely on pre-aligned encoders or global cosine similarity, which obscures fine-grained concept localization and fails to reflect true semantic geometry. In this work, we rethink concept alignment as a dynamic cross-modal transport process instead of static projection and propose the Optimal Transport Flow Concept Bottleneck Model (OTF-CBM). It first learns a data-driven semantic cost via Inverse Optimal Transport to measure cross-modal distances, and then performs unbalanced optimal-transport-based flow matching to model semantic transitions between visual patches and textual concepts. With velocity-based concept activation, OTF-CBM captures interpretable geometric relations without ODE integration. Experiments further show that OTF-CBM achieves superior classification accuracy and concept faithfulness, offering a new geometric and dynamical perspective for interpretable cross-modal reasoning.
Chinese Translation
概念瓶颈模型（CBMs）通过预测人类可解释的概念承诺提供透明的推理，但其有效性根本上依赖于视觉和文本表示的对齐或匹配程度。现有的视觉-语言CBMs通常依赖于预先对齐的编码器或全局余弦相似度，这掩盖了细粒度概念定位，并未能反映真实的语义几何。在本研究中，我们将概念对齐重新思考为一种动态跨模态传输过程，而非静态投影，并提出了最优传输流概念瓶颈模型（OTF-CBM）。该模型首先通过逆最优传输学习数据驱动的语义成本，以测量跨模态距离，然后执行基于不平衡最优传输的流匹配，以建模视觉片段与文本概念之间的语义转变。通过基于速度的概念激活，OTF-CBM在不进行常微分方程（ODE）积分的情况下捕捉可解释的几何关系。实验进一步表明，OTF-CBM在分类准确性和概念忠实性方面表现优越，为可解释的跨模态推理提供了新的几何和动态视角。
View on arXiv Download PDF AI Translation
cs.CV / 56 / 2606.26894

Modeling Local, Global, and Cross-Modal Context in Multimodal 3D MRI

多模态3D MRI中的局部、全局和跨模态上下文建模
Do, Minh Duc, Rheude, Tillmann, Kronenberg, Noel, Eils, Roland, Wild, Benjamin
Abstract
Brain MRI poses a fundamental challenge for machine learning: models must learn from high-dimensional 3D data spanning multiple co-registered modalities, despite the limited sample sizes typical of neuroimaging studies relative to the diversity in anatomy, pathology, and acquisition conditions. While multimodal imaging provides complementary information critical for clinical interpretation, effectively integrating these signals remains difficult. We propose Multimodal Intra- and Cross-Context Vision Transformer (MICViT), a 3D vision transformer that explicitly models both modality-specific representations and cross-modal interactions across local and global contexts. Concretely, MICViT combines four attention mechanisms: modality-specific local and global attention for intra-modal feature learning, and cross-modal local and global attention to capture interactions between modalities. We evaluate MICViT on brain age prediction across three heterogeneous datasets (UK Biobank, n=41,404; SOOP, n=1,062; Cam-CAN, n=613) using multiple MRI modalities (e.g. T1, FLAIR, DWI, SWI). MICViT consistently outperforms state-of-the-art CNN and transformer baselines in 3D settings. Notably, it benefits more strongly from multimodal inputs, yielding larger performance gains as additional modalities are incorporated. These results demonstrate that explicitly modeling intra- and cross-modal interactions is key to unlocking the full potential of multimodal brain MRI, highlighting a promising direction for representation learning in neuroimaging.
Chinese Translation
脑部MRI为机器学习带来了根本性挑战：模型必须从高维3D数据中学习，这些数据跨越多个共注册模态，尽管神经影像学研究中样本量通常有限，而解剖、病理和采集条件的多样性却很大。尽管多模态成像提供了对临床解读至关重要的互补信息，但有效整合这些信号仍然困难。我们提出了多模态内外上下文视觉变换器（Multimodal Intra- and Cross-Context Vision Transformer，MICViT），这是一种3D视觉变换器，明确建模模态特定表示和局部与全局上下文中的跨模态交互。具体而言，MICViT结合了四种注意力机制：用于模态特定特征学习的局部和全局注意力，以及捕捉模态之间交互的跨模态局部和全局注意力。我们在三个异质数据集（UK Biobank, n=41,404; SOOP, n=1,062; Cam-CAN, n=613）上评估了MICViT的脑龄预测，使用多种MRI模态（例如T1、FLAIR、DWI、SWI）。MICViT在3D设置中始终优于最先进的卷积神经网络（CNN）和变换器基线。值得注意的是，MICViT在多模态输入下受益更大，随着额外模态的加入，性能提升更为显著。这些结果表明，明确建模内外模态交互是释放多模态脑MRI全部潜力的关键，突显了神经影像学中表示学习的一个有前景的方向。
View on arXiv Download PDF AI Translation
cs.CV / 57 / 2606.26898

Tractography-Driven Synthetic Data Generation for Fiber Bundle Segmentation in Tracer Histology

基于轨迹追踪的合成数据生成用于追踪组织学中的纤维束分割
Bintsi, Kyriaki-Margarita, Makharia, Sparsh, Balbastre, Yaël, Avila, Joselyn Romero, Lehman, Julia F., Haber, Suzanne N., Yendiki, Anastasia
Abstract
Diffusion MRI (dMRI) tractography enables non-invasive reconstruction of white-matter pathways, but its accuracy is fundamentally limited by indirect, low-resolution measurements of axonal organization. Tracer injection studies in non-human primates provide a gold standard for validating dMRI tractography. This, however, requires time-consuming manual annotation of fiber bundles in histology sections. We propose a synthetic-data augmented framework for automated fiber bundle segmentation in macaque tracer histology. Our approach uses ex vivo dMRI tractography as a generative prior to synthesize 2D image patches for training. This provides us with sufficiently realistic foreground texture, which we compose with backgrounds from blockface photos and diversify via domain randomization. A 2D U-Net is trained on mixed real and synthetic patches. Experiments on held-out brains demonstrate improved generalization across brains and fiber bundle densities compared to training with real data only. Training with synthetic data only leads to poor performance, underscoring the need for real supervision. Overall, our approach achieves performance comparable to the state-of-the-art while requiring 3x less manually annotated data.
Chinese Translation
扩散磁共振成像（dMRI）轨迹追踪能够非侵入性地重建白质通路，但其准确性根本上受到对轴突组织的间接、低分辨率测量的限制。在非人灵长类动物中的追踪注射研究为验证dMRI轨迹追踪提供了金标准。然而，这需要在组织学切片中耗时的手动标注纤维束。我们提出了一种合成数据增强框架，用于自动化的猕猴追踪组织学中的纤维束分割。我们的方法使用体外dMRI轨迹追踪作为生成先验，合成用于训练的二维图像块。这为我们提供了足够逼真的前景纹理，我们将其与来自块面照片的背景组合，并通过领域随机化进行多样化。我们在真实和合成图像块的混合上训练了一个二维U-Net。在保留的脑部样本上的实验表明，与仅使用真实数据训练相比，我们的方法在不同脑部和纤维束密度下展示了更好的泛化能力。仅使用合成数据进行训练导致性能较差，强调了真实监督的必要性。总体而言，我们的方法在性能上可与最先进的技术相媲美，同时所需的手动标注数据减少了三倍。
View on arXiv Download PDF AI Translation
cs.CV / 58 / 2606.26904

Confidence-Aware Tool Orchestration for Robust Video Understanding

基于信任感的工具编排用于稳健的视频理解
He, Yangfan, Choi, Yujin, Yoon, Jaehong
Abstract
Video reasoning language models implicitly assume that every input frame is equally reliable. This leads to what we term the Blind Trust Problem: under realistic perturbations such as motion blur, glare, or occlusion, frontier video reasoning models can suffer 15-30%p accuracy drops on real-world embodied benchmarks, while remaining unaware that their visual evidence has been degraded. To address this challenge, we propose Robust-TO, an agentic video understanding framework that explicitly integrates per-frame trustworthiness into every stage of reasoning. Robust-TO organizes heterogeneous visual perception tools under a unified evidence interface. Each tool receives a sub-query derived from the original question and a set of trustworthy frames selected by the reliability-relevance score. It returns evidence in a shared format: a concrete prediction (e.g., a bounding box, motion trajectory, recognized text, or action label), temporal grounding, and a calibrated reliability score. During reasoning, these calibrated scores guide evidence weighting in a three-tier synthesis process (high/medium/low) and define a confidence-cost GRPO reward that jointly optimizes correctness, evidence reliability, and efficiency. On two video reasoning benchmarks spanning eight tasks, Robust-TO achieves 56.4% average accuracy on clean inputs, surpassing the strongest open-source baseline by 10.6%p and outperforming Gemini-2.5-Pro (46.2%). Under five realistic corruption types, Robust-TO maintains 54.3% average accuracy, 5.8%p above the strongest open-source baseline, while exhibiting the smallest clean-to-corrupted accuracy drop among all compared methods.
Chinese Translation
视频推理语言模型隐含地假设每个输入帧都是同等可靠的。这导致了我们所称的盲目信任问题：在运动模糊、眩光或遮挡等现实扰动下，前沿视频推理模型在真实世界的体现基准上可能会遭遇15-30%的准确率下降，而仍然未意识到其视觉证据已被降级。为了解决这一挑战，我们提出了Robust-TO，这是一个代理视频理解框架，明确将每帧的可信度整合到推理的每个阶段。Robust-TO在统一的证据接口下组织异构的视觉感知工具。每个工具接收一个源自原始问题的子查询和一组通过可靠性-相关性评分选出的可信帧。它以共享格式返回证据：具体预测（例如，边界框、运动轨迹、识别文本或动作标签）、时间基础和校准的可靠性评分。在推理过程中，这些校准的评分指导证据加权，采用三层合成过程（高/中/低），并定义一个信心-成本的GRPO奖励，联合优化正确性、证据可靠性和效率。在涵盖八个任务的两个视频推理基准上，Robust-TO在干净输入上实现了56.4%的平均准确率，超越了最强的开源基线10.6个百分点，并且优于Gemini-2.5-Pro（46.2%）。在五种现实的损坏类型下，Robust-TO保持54.3%的平均准确率，比最强的开源基线高出5.8个百分点，同时在所有比较方法中表现出最小的干净到损坏的准确率下降。
View on arXiv Download PDF AI Translation
cs.CV / 59 / 2606.26907

Qwen-Image-Agent: Bridging the Context Gap in Real-World Image Generation

Qwen-Image-Agent：弥合现实世界图像生成中的上下文差距
Zhang, Zekai, Li, Jiahao, Zhang, Jie, Gao, Kaiyuan, Yan, Kun, Jiang, Lihan, Tang, Ningyuan, Yin, Shengming, Wu, Tianhe, Chen, Xiaoyue, Xu, Xiao, Shu, Yan, Zhang, Yanran, Xu, Yixian, Chen, Yuxiang, Wang, Zhendong, Liu, Zihao, Zhou, Zikai, Zhang, Huishuai, Zhao, Dongyan, Wu, Chenfei
Abstract
While text-to-image (T2I) models have achieved remarkable progress, they struggle with real-world requests that are often underspecified, implicit, or dependent on up-to-date knowledge. We identify this challenge as the Context Gap: the mismatch between the user context and the sufficient generation context for T2I models. To bridge this gap, we propose Qwen-Image-Agent, a unified agentic framework that integrates plan, reason, search, memory and feedback in a context-centric manner. Qwen-Image-Agent treats user input as partial context and progressively constructs the generation context through Context-Aware Planning and Context Grounding. Specifically, Context-Aware Planning identifies missing context and plans how it should be acquired and used, while Context Grounding gathers this context from reason, search, memory, and feedback. To evaluate agentic image generation, we further introduce Image Agent Bench (IA-Bench), a benchmark covering four core image agent capabilities: Plan, Reason, Search, and Memory. Experiments on IA-Bench, Mindbench and WISE-Verified show that Qwen-Image-Agent outperforms strong baselines and achieves state-of-the-art performance.
Chinese Translation
尽管文本到图像（T2I）模型取得了显著进展，但它们在处理往往不明确、隐含或依赖于最新知识的现实世界请求时仍面临挑战。我们将这一挑战称为上下文差距（Context Gap）：用户上下文与T2I模型所需的充分生成上下文之间的不匹配。为了解决这一问题，我们提出了Qwen-Image-Agent，一个统一的智能框架，以上下文为中心整合计划、推理、搜索、记忆和反馈。Qwen-Image-Agent将用户输入视为部分上下文，并通过上下文感知规划（Context-Aware Planning）和上下文基础（Context Grounding）逐步构建生成上下文。具体而言，上下文感知规划识别缺失的上下文，并规划如何获取和使用这些上下文，而上下文基础则从推理、搜索、记忆和反馈中收集这些上下文。为了评估智能图像生成，我们进一步引入了图像智能基准（Image Agent Bench，IA-Bench），这是一个涵盖四个核心图像智能能力的基准：计划、推理、搜索和记忆。在IA-Bench、Mindbench和WISE-Verified上的实验表明，Qwen-Image-Agent超越了强基线，并实现了最先进的性能。
View on arXiv Download PDF AI Translation
cs.CV / 60 / 2606.26916

PhysRAG: Enhancing Physics-Awareness in Video Generation via Retrieval-Augmented Generation

PhysRAG：通过检索增强生成提升视频生成中的物理意识
Cheng, Kexu, Liu, Zicheng, Gao, Mingju, Song, Chunhe, Tang, Hao
Abstract
Developing physically aware video generation models remains a significant challenge due to the difficulty in capturing diverse physical phenomena, such as thermal dynamics, mechanics, and optics. In this work, we introduce PhysRAG, a novel pipeline that enhances physical awareness in video generation through Retrieval-Augmented Generation (RAG). To address the issue of limited high-quality data, we design a two-stage data filtering pipeline based on the WISA-80K dataset, resulting in a curated set of 7K high-quality videos for training. Furthermore, we construct a physical video database and develop a mechanism to inject physical knowledge into a video diffusion model using learnable queries. Our method achieves state-of-the-art performance in both visual quality and physical rule compliance, surpassing existing models in benchmarks such as PhyGenBench and VBench. We conduct extensive ablation studies to validate the effectiveness of our key components, including the data filtering pipeline, RAG mechanism, and method for physical information extraction. To facilitate future research, our code, data, and models are prepared for release at https://github.com/sediment1024/PhysRAG.
Chinese Translation
开发具有物理意识的视频生成模型仍然是一个重大挑战，因为捕捉多样的物理现象（如热动力学、力学和光学）非常困难。在本研究中，我们介绍了PhysRAG，这是一种通过检索增强生成（Retrieval-Augmented Generation, RAG）来提升视频生成中物理意识的新型管道。为了解决高质量数据有限的问题，我们基于WISA-80K数据集设计了一个两阶段的数据过滤管道，最终获得了一套7K高质量视频用于训练。此外，我们构建了一个物理视频数据库，并开发了一种机制，通过可学习查询将物理知识注入视频扩散模型。我们的方法在视觉质量和物理规则遵循方面均实现了最先进的性能，超越了现有模型在PhyGenBench和VBench等基准测试中的表现。我们进行了广泛的消融研究，以验证我们关键组件的有效性，包括数据过滤管道、RAG机制和物理信息提取方法。为了促进未来的研究，我们的代码、数据和模型已准备好在https://github.com/sediment1024/PhysRAG上发布。
View on arXiv Download PDF AI Translation
cs.CV / 61 / 2606.26930

PortraitGen: Exemplar-Driven GRPO with Dual-Reward Guidance for Photorealistic Portrait Generation

PortraitGen：基于示例驱动的双重奖励指导的生成相对策略优化（GRPO）用于逼真肖像生成
Li, Xiaomin, Liang, Qian, Li, Yinan, Zhang, Ying, Li, Chen, Lyu, Jing, Lu, Huchuan, Jia, Xu
Abstract
Reinforcement Learning like Group Relative Policy Optimization (GRPO) has significantly advanced text-to-image post-training. However, current methods often favor superficial aesthetics, such as over-saturated colors, leaving critical flaws like AI artifacts and biological implausibilities unresolved. We attribute these limitations to two primary factors: (1) The absence of real images during post-training confines GRPO sampling to the original distribution, failing to break inherent generative boundaries; (2) the optimization process lacks specific rewards targeting fine-grained artifacts like overly oily skin and other AI artifacts. To address this, we propose PortraitGen, a novel framework tailored for photorealistic portrait generation. First, we break inherent generative boundaries by directly introducing real images into the GRPO sampling groups, where image inversion is employed to obtain their transition probabilities and latents. Second, to explicitly steer the model toward photorealism, we introduce a complementary dual-reward mechanism: OmniReward for general quality and AI-Portrait for human-centric fidelity. Furthermore, we curate PortraitBench, a comprehensive portrait-centric benchmark. Extensive experiments demonstrate that PortraitGen significantly outperforms existing baselines, effectively suppressing AI artifacts and achieving unprecedented photorealism.
Chinese Translation
类似于强化学习的生成相对策略优化（GRPO）在文本到图像的后期训练中取得了显著进展。然而，目前的方法往往偏向于表面的美学，例如过于饱和的颜色，导致诸如人工智能伪影和生物不合理性等关键缺陷未得到解决。我们将这些局限性归因于两个主要因素：（1）后期训练中缺乏真实图像使得GRPO采样局限于原始分布，无法突破固有的生成边界；（2）优化过程缺乏针对细微伪影（如过于油腻的皮肤和其他人工智能伪影）的特定奖励。为了解决这个问题，我们提出了PortraitGen，一个专为逼真肖像生成量身定制的新框架。首先，我们通过直接将真实图像引入GRPO采样组来打破固有的生成边界，在此过程中采用图像反演来获取它们的转移概率和潜在特征。其次，为了明确引导模型朝向逼真性，我们引入了一种互补的双重奖励机制：OmniReward用于一般质量，AI-Portrait用于以人为中心的保真度。此外，我们策划了PortraitBench，一个全面的肖像中心基准。大量实验表明，PortraitGen显著优于现有基线，有效抑制了人工智能伪影，并实现了前所未有的逼真效果。
View on arXiv Download PDF AI Translation
cs.CV / 62 / 2606.26938

Focusing on What Matters: Saliency-Harnessing Accurate Routing for Diffusion MoE

聚焦重要内容：利用显著性驱动的准确路由进行扩散混合专家模型
Deng, Haoyou, Yan, Keyu, Mao, Chaojie, Wang, Xiang, Liu, Yu, Gao, Changxin, Sang, Nong
Abstract
Mixture-of-Experts (MoE) architectures have emerged as a powerful paradigm for scaling diffusion models in visual generation. Recent advancements have focused on adaptively allocating computational resources across diverse tokens to improve efficiency and performance. However, we identify a routing assignment problem in existing diffusion MoE frameworks: the router fails to accurately allocate more computational resources to salient tokens. Our analysis attributes this failure to the router's reliance on noise-corrupted latent features throughout the denoising process. Such stochastic noise obscures the critical structural and textural information, thereby preventing the router from effectively distinguishing salient tokens. To address this, we propose SharpMoE, a post-training framework with a saliency-harnessing accurate routing mechanism, which utilizes clean latent features as a noise-free guidance signal for routing. By bypassing the noise-distorted inputs, SharpMoE provides the router with clear saliency guidance, enabling the identification of salient tokens even in high-noise stages. Furthermore, we introduce a trajectory routing loss to constrain the compute allocation throughout the multi-step denoising trajectory, ensuring precise resource allocation along the generation rollout. Extensive experiments demonstrate that SharpMoE serves as a versatile, plug-and-play solution that further enhances the pretrained, converged MoE models, achieving state-of-the-art performance in visual generation.
Chinese Translation
混合专家（Mixture-of-Experts, MoE）架构已成为扩展视觉生成中扩散模型的强大范式。近期的进展集中在自适应地在多样化的标记之间分配计算资源，以提高效率和性能。然而，我们发现现有扩散 MoE 框架中的路由分配问题：路由器未能准确地将更多计算资源分配给显著标记。我们的分析将这一失败归因于路由器在去噪过程中依赖于噪声污染的潜在特征。这种随机噪声掩盖了关键的结构和纹理信息，从而阻碍了路由器有效区分显著标记。为了解决这一问题，我们提出了 SharpMoE，一个后训练框架，采用显著性驱动的准确路由机制，利用干净的潜在特征作为无噪声的路由指导信号。通过绕过噪声扭曲的输入，SharpMoE 为路由器提供了清晰的显著性指导，使其能够在高噪声阶段识别显著标记。此外，我们引入了一种轨迹路由损失，以约束多步去噪轨迹中的计算分配，确保在生成展开过程中精确的资源分配。大量实验表明，SharpMoE 作为一种多功能的即插即用解决方案，进一步增强了预训练的收敛 MoE 模型，在视觉生成中实现了最先进的性能。
View on arXiv Download PDF AI Translation
cs.CV / 63 / 2606.26942

TraMP-LLaMA: Generative Interpretability with Decoupled Instruction Tuning for Facial Expression Quality Assessment

TraMP-LLaMA：基于解耦指令调优的面部表情质量评估生成可解释性
Duan, Shuchao, Whone, Alan, Rahmani, Hossein, Liu, Jun, Mirmehdi, Majid
Abstract
Existing facial expression quality assessment (FEQA) methods typically produce only a severity score, without explicitly communicating the observable facial motion evidence that supports the prediction. This limits interpretability and makes it difficult to inspect the basis of model outputs in Parkinson's disease assessment. To address this gap, we propose TraMP-LLaMA, a unified multimodal framework that jointly predicts severity scores and generates structured textual reports from facial motion cues. The framework integrates RGB appearance and landmark trajectory cues, and adopts a decoupled instruction-tuning strategy to reduce task interference between severity prediction and language generation. To support this task, we further extend the PFED5 dataset with expert-guided textual motion descriptions and construct PFED5-plus. Experiments on PFED5-plus show that TraMP-LLaMA outperforms competitive video-language baselines in report generation and achieves the best severity prediction performance among the compared methods under joint multi-expression training, improving Spearman's rank correlation by at least 4.39 percent over all competing methods. The text annotations and code are available at https://github.com/shuchaoduan/TraMP-LLaMA.
Chinese Translation
现有的面部表情质量评估（FEQA）方法通常仅生成严重性评分，而未明确传达支持预测的可观察面部运动证据。这限制了可解释性，并使得在帕金森病评估中难以检查模型输出的基础。为了解决这一问题，我们提出了TraMP-LLaMA，一个统一的多模态框架，能够同时预测严重性评分并从面部运动线索生成结构化文本报告。该框架整合了RGB外观和关键点轨迹线索，并采用解耦的指令调优策略，以减少严重性预测与语言生成之间的任务干扰。为了支持这一任务，我们进一步扩展了PFED5数据集，增加了专家指导的文本运动描述，并构建了PFED5-plus。在PFED5-plus上的实验表明，TraMP-LLaMA在报告生成方面优于竞争性的视频-语言基线，并在联合多表情训练下实现了所有比较方法中最佳的严重性预测性能，Spearman等级相关性提高了至少4.39个百分点。文本注释和代码可在 https://github.com/shuchaoduan/TraMP-LLaMA 获取。
View on arXiv Download PDF AI Translation
cs.CV / 64 / 2606.26947

Scaling Multi-Reference Image Generation with Dynamic Reward Optimization

通过动态奖励优化扩展多参考图像生成
Huang, Wenwang, Fu, Yusen, Wang, Junjie, Huang, Mengfei, Li, Yulin, Liu, Gan, Cai, Jing, He, Yancheng, Tian, Zhuotao
Abstract
While personalized image generation has achieved remarkable progress, multi-reference image generation (MRIG) remains a challenging task. Most existing benchmarks fail to adequately evaluate complex MRIG scenarios, hindering further progress in this area. To better assess model performance on complex MRIG tasks, we introduce OmniRef-Bench, a benchmark that covers complex combinations of reference image types and a large number of reference images. Evaluations on OmniRef-Bench show that mainstream open-source models struggle in complex MRIG scenarios, and their performance deteriorates significantly as the number of mixed-type reference images increases. To address this issue, we propose DyRef, a two-stage training framework. In the first stage, supervised fine-tuning equips the model with the basic capability to handle complex MRIG tasks. In the second stage, we introduce Difficulty-aware Advantage Reweighting (DAR) and Discriminative Reward Scaling (DRS). DAR dynamically adjusts the optimization objective to improve performance when handling a large number of mixed-type reference images. DRS enlarges intra-group reward differences for more effective policy optimization. Experiments demonstrate that DyRef significantly improves the performance of open-source models on OmniRef-Bench and single-image editing benchmarks, demonstrating the effectiveness and generalization capability of our approach.
Chinese Translation
尽管个性化图像生成已取得显著进展，但多参考图像生成（MRIG）仍然是一项具有挑战性的任务。现有的大多数基准测试未能充分评估复杂的MRIG场景，阻碍了该领域的进一步发展。为了更好地评估模型在复杂MRIG任务上的表现，我们引入了OmniRef-Bench，这是一个涵盖复杂参考图像类型组合和大量参考图像的基准测试。对OmniRef-Bench的评估表明，主流开源模型在复杂MRIG场景中表现不佳，随着混合类型参考图像数量的增加，其性能显著下降。为了解决这一问题，我们提出了DyRef，一个两阶段的训练框架。在第一阶段，监督微调使模型具备处理复杂MRIG任务的基本能力。在第二阶段，我们引入了困难感知优势重加权（Difficulty-aware Advantage Reweighting, DAR）和区分奖励缩放（Discriminative Reward Scaling, DRS）。DAR动态调整优化目标，以提高处理大量混合类型参考图像时的性能。DRS则扩大组内奖励差异，以实现更有效的策略优化。实验表明，DyRef显著提高了开源模型在OmniRef-Bench和单图像编辑基准测试上的性能，证明了我们方法的有效性和泛化能力。
View on arXiv Download PDF AI Translation
cs.CV / 65 / 2606.26970

Computer Vision for MOBA Analytics: A Dataset and Baseline for Visibility Analysis in Dota 2

用于MOBA分析的计算机视觉：Dota 2中可见性分析的数据集和基准
Carvalho, Ricardo da Rocha, Oliveira, Eloísa, Kummer, Luiz Bernardo Martins, Paraiso, Emerson Cabrera, Laroca, Rayson
Abstract
Introduction: Most Multiplayer Online Battle Arena (MOBA) analytics studies rely on structured data, which does not directly capture what each team could actually see during a match. Objective: This work introduces Dota2-Vis, a video-based dataset, and a baseline pipeline for visibility analysis in professional Dota 2 matches. Methodology: The dataset comprises all 144 matches from The International 2025, recorded from both team perspectives, totaling 288 Full HD videos, together with 2,477 manually annotated minimap images. We evaluate multiple variants of a modern object detector for player-icon detection and use the best-performing model to estimate opponent-visible player presence over time. Results: YOLO11l (large) achieved the best overall performance, reliably identifying player icons even in dense and visually cluttered minimap scenes. The resulting visibility curves reveal player, hero, role, and team-level patterns that complement conventional MOBA analytics, highlighting behavioral differences that are difficult to obtain from structured data alone. The dataset and code are publicly available at https://github.com/RicardoRCarvalho/dota2-vis/.
Chinese Translation
引言：大多数多人在线战斗竞技场（MOBA）分析研究依赖于结构化数据，这些数据并不能直接捕捉到每个团队在比赛中实际能够看到的内容。目标：本研究介绍了Dota2-Vis，一个基于视频的数据集，以及用于专业Dota 2比赛中可见性分析的基准流程。方法：该数据集包含2025年国际邀请赛的144场比赛的所有录像，从两个团队的视角记录，共计288个全高清（Full HD）视频，以及2477幅手动标注的迷你地图图像。我们评估了多种现代目标检测器的变体用于玩家图标检测，并使用表现最佳的模型来估计对手可见的玩家存在情况。结果：YOLO11l（large）在整体性能上表现最佳，能够在密集且视觉杂乱的迷你地图场景中可靠地识别玩家图标。生成的可见性曲线揭示了玩家、英雄、角色和团队层面的模式，补充了传统的MOBA分析，突显了仅从结构化数据中难以获得的行为差异。数据集和代码已公开发布在 https://github.com/RicardoRCarvalho/dota2-vis/。
View on arXiv Download PDF AI Translation
cs.CV / 66 / 2606.26973

Geometric Gradient Rectification for Safe Open-Set Semi-Supervised Learning

安全开放集半监督学习的几何梯度校正
Chen, Jiahe, Shao, Qian, Chen, Qiyuan, He, Jiaying, Chen, Jintai, Wu, Jian, Xu, Hongxia
Abstract
Open-set semi-supervised learning aims to leverage unlabeled data that may contain out-of-distribution outliers while maintaining performance on in-distribution classes. Existing methods mainly follow two paradigms: filtering suspicious samples or incorporating unlabeled objectives with soft weighting. We argue that both face a common trade-off: aggressive filtering can discard informative but hard ID samples, whereas utilization can introduce auxiliary gradients that conflict with supervised learning when pseudo labels are wrong. We therefore shift the focus from sample selection to gradient-level control. We propose \textit{Geometric Gradient Rectification} (GGR), a plug-in framework that uses the supervised gradient as an anchor and projects conflicting auxiliary gradients onto an admissible region in gradient space. This makes the applied auxiliary update first-order non-opposing within the rectified coordinate block while preserving orthogonal components that may still carry useful representation signals. We further extend GGR with subspace-aware rectification to stabilize the anchor under noisy mini-batch gradients. Experiments on CIFAR and ImageNet benchmarks show that GGR improves representative OSSL baselines in most settings and yields gains in both closed-set generalization and open-set robustness. Code will be available at https://github.com/JiaheChen2002/GGR.
Chinese Translation
开放集半监督学习旨在利用可能包含分布外离群点的未标记数据，同时保持对分布内类别的性能。现有方法主要遵循两种范式：过滤可疑样本或结合未标记目标与软加权。我们认为这两者面临一个共同的权衡：激进的过滤可能会丢弃信息丰富但难以分类的分布内样本，而利用未标记样本可能在伪标签错误时引入与监督学习相冲突的辅助梯度。因此，我们将重点从样本选择转向梯度级控制。我们提出了 extit{几何梯度校正}（Geometric Gradient Rectification，GGR），这是一个插件框架，使用监督梯度作为锚点，并将冲突的辅助梯度投影到梯度空间中的可接受区域。这使得应用的辅助更新在校正的坐标块内保持一阶非对抗，同时保留可能仍携带有用表示信号的正交分量。我们进一步扩展了GGR，通过子空间感知校正来稳定锚点，以应对噪声小批量梯度。对CIFAR和ImageNet基准的实验表明，GGR在大多数设置中提高了代表性开放集半监督学习（OSSL）基线，并在封闭集泛化和开放集鲁棒性上均取得了进展。代码将发布在 https://github.com/JiaheChen2002/GGR。
View on arXiv Download PDF AI Translation
cs.CV / 67 / 2606.26984

Unison: Benchmarking Unified Multimodal Models via Synergistic Understanding and Generation

Unison：通过协同理解与生成对统一多模态模型进行基准评估
Liu, Jinyu, Shuai, Xincheng, Ding, Henghui, Jiang, Yu-Gang
Abstract
Unified multimodal models capable of both understanding and generation have achieved remarkable strides. However, despite their unified designs, existing evaluations typically assess understanding and generation capabilities in isolation, overlooking the synergy between comprehension and generation. To bridge this gap, we introduce Unison, a comprehensive benchmark comprising 2,169 high-quality unified task samples, designed to evaluate joint understanding and generation in unified multimodal models. Unison offers three key strengths: 1) Comprehensive Dimensions: Unison encompasses internal consistency, understanding-guided generation, generation-guided understanding, and mutual enhancement to enable holistic evaluation. 2) Diagnostic Evaluation: it provides both unified and decoupled tracks for understanding and generation, allowing fine-grained attribution of failure modes and quantitative analysis of the gains from unified modeling. 3) Human Alignment: we also introduce Unison-Judge, an evaluation model well aligned with human judgments to ensure reliable assessment. Based on systematic evaluations of state-of-the-art models on Unison, we uncover critical limitations in current unified multimodal systems and highlight promising directions for future research. Codes, Unison and Unison-Judge are publicly available at https://github.com/FudanCVL/Unison.
Chinese Translation
能够同时进行理解和生成的统一多模态模型取得了显著进展。然而，尽管它们的设计是统一的，现有的评估通常孤立地评估理解和生成能力，忽视了理解与生成之间的协同作用。为了解决这一问题，我们提出了Unison，一个综合基准，包含2169个高质量的统一任务样本，旨在评估统一多模态模型中的联合理解与生成。Unison具有三个主要优势：1）全面维度：Unison涵盖内部一致性、理解驱动的生成、生成驱动的理解和相互增强，以实现整体评估。2）诊断评估：它提供了统一和解耦的理解与生成轨道，允许对失败模式进行细致归因，并对统一建模带来的收益进行定量分析。3）人类对齐：我们还引入了Unison-Judge，一个与人类判断高度一致的评估模型，以确保评估的可靠性。基于对最先进模型在Unison上的系统评估，我们揭示了当前统一多模态系统的关键局限性，并强调了未来研究的有希望方向。代码、Unison和Unison-Judge已公开发布在 https://github.com/FudanCVL/Unison。
View on arXiv Download PDF AI Translation
cs.CV / 68 / 2606.26994

Event-Aware Instructed Assistant for Referring Video Segmentation

事件感知的视频指引分割助手
Liu, Jinyu, Ding, Henghui, He, Shuting, Jiang, Yu-Gang
Abstract
Existing referring video segmentation methods often treat a video as a single event consisting of multiple images, overlooking the fact that a video typically contains multiple distinct events. Under such a mechanism, the model needs to directly understand all the complex content in the video and text, which can easily lead to confusion and hallucinations. To address this issue, we propose to decompose a video to a set of simple events by learnable Event Query, and understand complex video content in an event-by-event, easy-to-understand manner. This is based on the observation that natural language expressions often divide a video into distinct, text-related segments, each representing a separate event within a compound event. We introduce EVIS, an Event-Aware Video Instructed Segmentation Assistant, which utilizes text-guided Event Queries to partition a video into simple events, extracting event-aware visual-text features to achieve a hierarchical understanding of the video. Additionally, we propose Object-Pixel-Hybrid Learning, which enables the MLLMs to track targets in long-term videos by integrating fine-grained pixel features with prior object queries. Extensive experimental results on 5 public benchmarks demonstrate EVIS's strong performance in addressing the referring video segmentation task.
Chinese Translation
现有的指向视频分割方法通常将视频视为由多幅图像组成的单一事件，忽视了视频通常包含多个不同事件的事实。在这种机制下，模型需要直接理解视频和文本中的所有复杂内容，这很容易导致混淆和幻觉。为了解决这个问题，我们提出通过可学习的事件查询将视频分解为一组简单事件，并以逐事件、易于理解的方式理解复杂的视频内容。这基于一个观察，即自然语言表达通常将视频划分为不同的、与文本相关的片段，每个片段代表复合事件中的一个独立事件。我们引入了EVIS（事件感知视频指引分割助手），它利用文本引导的事件查询将视频划分为简单事件，提取事件感知的视觉-文本特征，以实现对视频的层次化理解。此外，我们提出了对象-像素-混合学习（Object-Pixel-Hybrid Learning），使得多模态大语言模型（MLLMs）能够通过将细粒度像素特征与先前的对象查询相结合，在长期视频中跟踪目标。在5个公共基准上的大量实验结果表明，EVIS在解决指向视频分割任务方面表现出色。
View on arXiv Download PDF AI Translation
cs.CV / 69 / 2606.27018

On-board Remote-Sensing Foundation Models for Unsupervised Change Detection of Disaster Events

基于遥感基础模型的灾害事件无监督变化检测方法
Ramírez-Gallego, S.
Abstract
Remote Sensing Foundation Models (RSFMs) have emerged as a powerful alternative to supervised models for Earth Observation, allowing satellites to autonomously trigger high-resolution captures or adjust tasking parameters upon detecting an anomaly, thereby maximizing the utility of the mission's limited power and computational resources. RSFMs are versatile, unified encoders that optimize onboard storage for multiple orbital applications while ensuring high-fidelity feature extraction. In particular, unsupervised change detection with RSFMs offers a well-informed and transformative path for disaster monitoring without expensive labels. In this paper, we present a novel unsupervised detection method based on ResNet (RSFM) + FPN which identifies a wide spectrum of anomalies by detecting subtle semantic shifts in the latent space between successive orbital passes. By relying on an untrained FPN architecture and its intrinsic priors, the system achieves efficient image-level generation and higher resolution mapping with minimal effort (training-free) compared to previous proposals (patch-based, trained). And by replacing tailored models with RSFMs, we can achieve comparable results through an approach that eliminates the need for bespoke training and extensive development effort and adds customization, while ensuring high-performance generalization across diverse terrains and sensors.
Chinese Translation
遥感基础模型（RSFMs）作为地球观测中监督模型的强大替代方案应运而生，使卫星能够在检测到异常时自主触发高分辨率拍摄或调整任务参数，从而最大限度地利用任务有限的电力和计算资源。RSFMs 是多用途的统一编码器，优化了多种轨道应用的机载存储，同时确保高保真特征提取。特别是，使用 RSFMs 进行无监督变化检测为灾害监测提供了一条信息充分且变革性的路径，无需昂贵的标签。在本文中，我们提出了一种基于 ResNet（RSFM）+ FPN 的新型无监督检测方法，通过检测连续轨道经过之间潜在空间中的细微语义变化，识别广泛的异常。依赖于未训练的 FPN 架构及其内在先验，该系统实现了高效的图像级生成和更高分辨率的映射，相较于之前的提案（基于补丁的、经过训练的），所需的努力（无训练）更少。通过用 RSFMs 替代定制模型，我们可以通过一种消除定制训练和广泛开发努力的方式实现可比的结果，同时确保在多样化地形和传感器上的高性能泛化，并增加了定制化。
View on arXiv Download PDF AI Translation
cs.CV / 70 / 2606.27071

PanoImager: Geometry-Guided Novel View Synthesis and Reconstruction from Sparse Panoramic Views

PanoImager：基于几何引导的新视图合成与稀疏全景视图重建
Xu, Zhisong, Oishi, Takeshi
Abstract
Panoramic sensing offers wide field-of-view coverage, yet 3D reconstruction from sparse panoramas remains challenging under rotation-dominant, weak-parallax motion. In such regimes, SfM/SLAM initialization is often ill-conditioned and unreliable. We present PanoImager, an SfM-free framework that combines feed-forward pose/depth priors, geometry-conditioned diffusion view completion, and depth-guided 3DGS optimization. Given only a few panoramic images, PanoImager decomposes them into local perspective views, synthesizes auxiliary observations to enrich sparse evidence, and stabilizes Gaussian optimization for improved cross-view consistency. Experiments on multiple benchmarks show improved stability under extreme sparsity, suggesting PanoImager as an offline/background component for map refinement when SfM/SLAM fails to initialize.
Chinese Translation
全景感知提供了广阔的视场覆盖，但在旋转主导、弱视差运动下，从稀疏全景图进行3D重建仍然具有挑战性。在这种情况下，结构从运动（SfM）/同步定位与地图构建（SLAM）初始化通常条件不良且不可靠。我们提出了PanoImager，一个无SfM框架，结合了前馈姿态/深度先验、几何条件扩散视图补全和深度引导的3D全局优化。仅给定少量全景图像，PanoImager将其分解为局部透视视图，合成辅助观测以丰富稀疏证据，并稳定高斯优化以改善跨视图一致性。在多个基准测试中的实验表明，在极端稀疏情况下稳定性有所改善，表明PanoImager可以作为SfM/SLAM无法初始化时的离线/背景组件，用于地图精细化。
View on arXiv Download PDF AI Translation
cs.CV / 71 / 2606.27084

Pseudo-Text-Conditioned 3D Grounding DINO for Organ Localization in Abdominal CT

伪文本条件下的3D定位DINO在腹部CT中的器官定位
Chen, Siqi, Gong, Han, Hou, Keyi, Yang, Jingxuan, Bhat, Sheethal, Maier, Andreas
Abstract
Reliable organ localization in abdominal CT can provide spatial priors for downstream trauma analysis. We propose CT-3GDINO, a lightweight 3D detector that adapts a Grounding-DINO-style query-based architecture to fixed organ localization using frozen pseudo-text class tokens instead of a real text encoder. The model combines a Swin3D visual backbone, bidirectional feature enhancement, pseudo-text-guided query selection, and a cross-modality decoder to predict normalized 3D boxes for liver, spleen, left kidney, right kidney, and bowel. We train and evaluate on 193 matched RSNA/RATIC CT volumes with segmentation-derived boxes. The best multi-scale model, trained from scratch, achieves 0.5830 overall top-1 class-wise mAP over 3D IoU thresholds from 0.1 to 0.7, outperforming fixed- and trainable-backbone classification-pretrained variants with 0.5570 and 0.4657 mAP. Performance is strong for coarse localization, with 0.9649 AP at IoU 0.1, but remains limited for strict box alignment, with 0.1552 AP at IoU 0.7. These results establish CT-3GDINO as an open-source baseline for pseudo-text-conditioned 3D organ localization and motivate future work on localization-aware pretraining, richer multimodal conditioning, and injury-focused detection.
Chinese Translation
腹部CT中可靠的器官定位可以为后续的创伤分析提供空间先验。我们提出了CT-3GDINO，这是一种轻量级的3D检测器，它采用了Grounding-DINO风格的基于查询的架构，通过使用冻结的伪文本类别标记而非真实文本编码器来实现固定器官定位。该模型结合了Swin3D视觉主干、双向特征增强、伪文本引导的查询选择和跨模态解码器，以预测肝脏、脾脏、左肾、右肾和肠道的归一化3D框。我们在193个匹配的RSNA/RATIC CT体积上进行了训练和评估，这些体积具有基于分割的框。最佳的多尺度模型从头开始训练，在3D IoU阈值从0.1到0.7的情况下，整体的top-1类别平均精度（mAP）达到了0.5830，优于固定和可训练主干的分类预训练变体，其mAP分别为0.5570和0.4657。对于粗略定位，性能表现良好，在IoU 0.1时达到了0.9649的平均精度，但在严格框对齐方面仍然有限，在IoU 0.7时仅为0.1552的平均精度。这些结果确立了CT-3GDINO作为伪文本条件下3D器官定位的开源基线，并激励未来在定位感知预训练、更丰富的多模态条件和以伤害为重点的检测方面的研究。
View on arXiv Download PDF AI Translation
cs.CV / 72 / 2606.27088

SubdivAR: Autoregressive Next-Scale Prediction for Neural Mesh Subdivision

SubdivAR：用于神经网格细分的自回归下一个尺度预测
Guo, Huipeng, Song, Zikai, Long, Hang, Zhang, Jielei, Li, Wenbing, Lin, Junkai, Zhao, Tianhao, Zhang, Jinshen, Guo, Tianle, Yang, Wei
Abstract
Mesh subdivision is a fundamental operation for converting coarse, editable meshes into high-resolution surfaces, with broad applications in digital asset creation. Classical rule-based schemes rely on fixed local refinement rules and often produce over-smoothed surfaces. Recent neural subdivision methods improve detail synthesis, but remain constrained by local modeling and exhibit limited generalizability. We present SubdivAR, a neural mesh subdivision framework based on our proposed Mesh Autoregressive Representation (MAR). MAR arranges meshes at different subdivision levels into an ordered scale sequence, reformulating subdivision as autoregressive next-scale prediction. To support this formulation, we introduce a Hybrid Topology-Aware Transformer that combines global semantic attention with topology-constrained local feature aggregation. SubdivAR adopts a next-scale coordinate prediction paradigm, regressing vertex offsets at each refinement stage to preserve subdivision topology while recovering fine-grained geometric details. To enable reliable learning, we construct FII-40K, a curated dataset of nearly 40,000 high-quality meshes with multi-level subdivision supervision. Experiments show that SubdivAR outperforms state-of-the-art baselines, reducing Hausdorff Distance and Chamfer Distance by 18.8% and 14.2%, respectively, and demonstrates strong robustness on complex open-surface geometries.
Chinese Translation
网格细分是将粗糙的可编辑网格转换为高分辨率表面的基本操作，在数字资产创作中具有广泛的应用。传统的基于规则的方法依赖于固定的局部细化规则，通常会产生过于平滑的表面。最近的神经细分方法改善了细节合成，但仍受限于局部建模，表现出有限的泛化能力。我们提出了SubdivAR，这是一种基于我们提出的网格自回归表示（Mesh Autoregressive Representation, MAR）的神经网格细分框架。MAR将不同细分级别的网格排列成有序的尺度序列，将细分重新表述为自回归下一个尺度预测。为了支持这一表述，我们引入了一种混合拓扑感知变换器（Hybrid Topology-Aware Transformer），该变换器结合了全局语义注意力和拓扑约束的局部特征聚合。SubdivAR采用下一尺度坐标预测范式，在每个细化阶段回归顶点偏移量，以保持细分拓扑，同时恢复细致的几何细节。为了实现可靠的学习，我们构建了FII-40K，一个包含近40,000个高质量网格的精心策划的数据集，具有多级细分监督。实验表明，SubdivAR在减少Hausdorff距离和Chamfer距离方面分别比最先进的基线降低了18.8%和14.2%，并在复杂的开放表面几何体上表现出强大的鲁棒性。
View on arXiv Download PDF AI Translation
cs.CV / 73 / 2606.27089

TMP: Tree-structured Mixed-policy Pruning for Large-scale Image Generation and Editing

TMP：用于大规模图像生成和编辑的树结构混合策略剪枝
Zhang, Peizhen, Li, Yang, Li, Xunsong, Liu, Songtao, Liu, Zewen, Hu, Qiangqiang, Guo, Guotong, Ding, Jupeng, Sun, Yifu, coopersli, Zhang, Jian, Zhong, Zhao, Bo, Liefeng
Abstract
Modern image generation model rapidly grows their sizes to meet high-fidelity image synthesis. However, they gradually become unaffordable for their enormous parameter consumption and computation budget that lead to massive resources requirement and gpu memory footprint. In this paper, we propose TMP, the first Tree-structured Mixed-policy Pruning framework that generalizes prevalent image tasks (T2I and TI2I) and architectures (Mixture-of-Experts (MoE) and Diffusion transformer (DiT)). It could be applied to the step-distilled models and contribute as the last stage. We perform experiments upon current open-sourced SOTA HunyuanImage-3.0 instruct and a popular efficient model Z-Image turbo. The proposed pruning framework manages to compress HunyuanImage 3.0 from 80B to 20B parameters at 75% reduction ratio, sacrificing limited generation quality. We also optimize to enable the inference of the pruned 20B version of HunyuanImage 3.0 on a single 24GB 4090 GPU by engineering skills. The inference script and model weight have been integrated into the existing HunyuanImage3.0 open-source github and huggingface repository. Besides, we prove the efficacy of TMP by compressing Z-Image turbo from 6B to 4B (33% reduction) with negligible degradation.
Chinese Translation
现代图像生成模型的规模迅速扩大，以满足高保真图像合成的需求。然而，由于其庞大的参数消耗和计算预算，这些模型逐渐变得难以承受，导致对资源的巨大需求和GPU内存占用。在本文中，我们提出了TMP，这是第一个树结构混合策略剪枝框架，能够推广到流行的图像任务（T2I 和 TI2I）以及架构（专家混合模型（MoE）和扩散变换器（DiT））。该框架可以应用于逐步蒸馏模型，并作为最后阶段进行贡献。我们在当前开源的最先进模型HunyuanImage-3.0 instruct和一个流行的高效模型Z-Image turbo上进行了实验。所提出的剪枝框架成功地将HunyuanImage 3.0的参数从80B压缩到20B，减少比例达到75%，在有限的生成质量损失下实现。我们还通过工程技术优化，使得在单个24GB的4090 GPU上能够推理剪枝后的20B版本HunyuanImage 3.0。推理脚本和模型权重已集成到现有的HunyuanImage3.0开源GitHub和Hugging Face仓库中。此外，我们通过将Z-Image turbo从6B压缩到4B（减少33%）且几乎没有降级，证明了TMP的有效性。
View on arXiv Download PDF AI Translation
cs.CV / 74 / 2606.27128

FlameVQA: A Physically-Grounded UAV Wildfire VQA Benchmark with Radiometric Thermal Supervision

FlameVQA：一个基于物理基础的无人机野火视觉问答基准，具有辐射热监督
Habibpour, Mobin, Spodnik, John, Talemi, Niloufar Alipour, Afghah, Fatemeh
Abstract
Wildfire monitoring from UAVs requires reliable reasoning over complex aerial scenes, where smoke, scale variation, and occlusions often limit RGB-only interpretation. We introduce FlameVQA, a multiple-choice visual question answering benchmark for UAV-based wildfire intelligence built on FLAME 3, leveraging paired RGB imagery and radiometric thermal TIFFs for temperature-grounded, safety-critical reasoning. FlameVQA includes 34 multiple-choice questions per image spanning six operational capability groups, covering tasks such as detection, localization, distribution/coverage estimation, cross-modal reasoning, and flight planning. To ensure label reliability, we combine MLLM-assisted annotation with deterministic thermal rules and cross-question consistency checks, followed by human auditing. We also evaluate representative MLLMs on FlameVQA to provide baselines for future work. Results show strong performance when explicit cross-modal cues are available, but notable failures on presence detection under heavy smoke and on coverage estimation. These findings suggest that current MLLMs require domain-specific adaptation to better support disaster and wildfire monitoring. The dataset and benchmark code are open-source at github.com/mobiiin/WildFire_VQA
Chinese Translation
无人机对野火的监测需要在复杂的空中场景中进行可靠的推理，而烟雾、尺度变化和遮挡常常限制了仅依赖RGB图像的解读。我们提出了FlameVQA，这是一个基于无人机的野火智能多项选择视觉问答基准，建立在FLAME 3之上，利用配对的RGB图像和辐射热TIFF图像进行温度基础的安全关键推理。FlameVQA为每幅图像提供34个多项选择问题，涵盖六个操作能力组，涉及检测、定位、分布/覆盖估计、跨模态推理和飞行规划等任务。为了确保标签的可靠性，我们结合了MLLM辅助注释、确定性热规则和跨问题一致性检查，并进行了人工审核。我们还在FlameVQA上评估了代表性的MLLM，以提供未来工作的基准。结果表明，当存在明确的跨模态线索时，性能表现良好，但在浓烟下的存在检测和覆盖估计方面存在显著失败。这些发现表明，当前的MLLM需要进行特定领域的适应，以更好地支持灾害和野火监测。数据集和基准代码已在github.com/mobiiin/WildFire_VQA上开源。
View on arXiv Download PDF AI Translation
cs.CV / 75 / 2606.27147

Safe Autoregressive Image Generation with Iterative Self-Improving Codebooks

安全的自回归图像生成与迭代自我改进的代码本
Xue, Yunqi, Li, Zhijiang, Torr, Philip, Gu, Jindong
Abstract
Unlike diffusion-based models that operate in continuous latent spaces, autoregressive unified multimodal models produce images by sequentially predicting discretized visual tokens. These tokens are derived from a codebook that maps embeddings to quantized visual patterns. The language-like architecture enables unified multimodal models to effectively capture text conditional information for generation, making them promising for text-to-image tasks. This also raises an interesting question: how safe are the images generated in such an autoregressive way? In this work, we propose iterative self-improving codebooks for safe autoregressive generation. We leverage the understanding and judgment capabilities of the unified multimodal model itself to identify unsafe generated images without human annotation. Subsequently, the inherent representations in the codebook are fixed to eliminate harmful mappings. Our method comprises two steps: first, we use the unified model to identify unsafe generations and construct corresponding harmful and safe image-text pairs. These pairs are used to construct the Harmful Space and guide updates to the codebook, thereby eliminating harmful outputs. Second, we perform adaptive fine-tuning on the codebook within the harmless space using safe image-text pairs to ensure the quality of generated images. These two steps are repeated until no further improvement is observed, producing a safety-enhanced model codebook. Without additional external feedback, the safety of models is improved iteratively.
Chinese Translation
与在连续潜在空间中操作的扩散模型不同，自回归统一多模态模型通过顺序预测离散化的视觉标记生成图像。这些标记源自一个代码本，该代码本将嵌入映射到量化的视觉模式。类语言的架构使得统一多模态模型能够有效捕捉文本条件信息以进行生成，这使得它们在文本到图像任务中具有良好的前景。这也引发了一个有趣的问题：以这种自回归方式生成的图像有多安全？在本研究中，我们提出了用于安全自回归生成的迭代自我改进代码本。我们利用统一多模态模型自身的理解和判断能力，识别出不安全的生成图像，而无需人工标注。随后，固定代码本中的固有表示，以消除有害映射。我们的方法包括两个步骤：首先，我们使用统一模型识别不安全的生成，并构建相应的有害和安全图像-文本对。这些对用于构建有害空间并指导代码本的更新，从而消除有害输出。其次，我们在无害空间内使用安全图像-文本对对代码本进行自适应微调，以确保生成图像的质量。这两个步骤重复进行，直到不再观察到进一步的改进，从而生成一个安全增强的模型代码本。在没有额外外部反馈的情况下，模型的安全性得以迭代提升。
View on arXiv Download PDF AI Translation
cs.CV / 76 / 2606.27187

HarmVideoBench: Benchmarking Harmful Video Understanding in Large Multimodal Models

HarmVideoBench：大型多模态模型中有害视频理解的基准测试
Wu, Jiajun, Kang, Haoyu, Sun, Yining, Hou, Jiacheng, Zhang, Heng, Zhang, Danyang, Zhao, Zhenjun, Zhang, Haochi, Sun, Leixin, Jiang, Eric Hanchen, Li, Yushan, Li, Ruiyu, Huang, Mengkai, Gao, Yan, Zhang, Xu, Wan, Guancheng
Abstract
Large vision-language models (LVLMs) have recently shown immense potential in automated content moderation, sparking growing interest in developing harmful-video benchmarks. However, we identify two primary limitations in existing works: 1) The multi-layered characteristics of harmful videos are overlooked. Existing benchmarks predominantly formulate evaluation as a binary classification task, failing to capture implicit or deep contextual harms. 2) Explanatory rationales are completely absent. Current frameworks measure exclusively whether a model flags a video correctly rather than explaining why, turning evaluation into a black box where models can succeed through superficial shortcuts. To address these problems, we present HarmVideoBench, a multi-layered diagnostic benchmark comprising 1,379 videos paired with 4,137 multiple-choice questions. HarmVideoBench benchmarks three hierarchical dimensions: Observable Evidence, Clip-Internal Meaning, and Beyond-Clip Reasoning, aiming to evaluate models' deep understanding beyond surface cues with carefully balanced and curated samples. We evaluate 19 leading models on HarmVideoBench to assess their multidimensional understanding of harmful videos. Moreover, we introduce BCR, a benchmark-aligned method that predicts reasoning boundaries and dynamically retrieves context only when needed. Experimental results show that BCR substantially improves the base model's performance in harmful video understanding, raising the macro average from 61.7 percent to a state-of-the-art 84.4 percent.
Chinese Translation
大型视觉语言模型（LVLMs）最近在自动内容审核中展现了巨大的潜力，激发了对开发有害视频基准的日益关注。然而，我们发现现有研究存在两个主要局限：1）有害视频的多层次特征被忽视。现有基准主要将评估形式化为二元分类任务，未能捕捉到隐含或深层的上下文危害。2）缺乏解释性理由。目前的框架仅测量模型是否正确标记视频，而未解释原因，这使得评估变成了一个黑箱，模型可以通过表面捷径获得成功。为了解决这些问题，我们提出了HarmVideoBench，这是一个多层次的诊断基准，包含1,379个视频和4,137个多项选择题。HarmVideoBench基准评估三个层次维度：可观察证据、剪辑内部意义和超剪辑推理，旨在通过精心平衡和策划的样本评估模型对深层理解的能力。我们在HarmVideoBench上评估了19个领先模型，以评估它们对有害视频的多维理解。此外，我们引入了BCR，一种与基准对齐的方法，该方法预测推理边界并在需要时动态检索上下文。实验结果表明，BCR显著提高了基础模型在有害视频理解方面的表现，将宏观平均从61.7%提升至84.4%的最新水平。
View on arXiv Download PDF AI Translation
cs.CV / 77 / 2606.27192

LISA: Likelihood Score Alignment for Visual-condition Controllable Generation

LISA：用于视觉条件可控生成的似然得分对齐
Wang, Yanghao, Chen, Hongxu, Liu, Jiazhen, He, Zhenqi, Liu, Rui, Wang, Zhen, Chen, Long
Abstract
The prevalent dual-branch paradigm, i.e., training a side network to encode visual conditions and fusing its intermediate-layer features to a frozen pretrained main network, has shown remarkable success in visual-condition controllable generation. Despite its widespread adoption, the role of the side branch and its training efficiency remain underexplored. In this paper, we first revisit this mainstream paradigm through the lens of score-based generative modeling: 1) The main network preserves visual perceptual quality by providing a prior unconditional score. 2) The side network steers conditional control by implicitly contributing a likelihood score. Guided by this perspective, we propose LIkelihood Score Alignment (LISA), an effective regularization method that explicitly aligns the intermediate feature of the side network with an approximated likelihood score. Specifically, we first hook features from a designated layer of the side network and project them into the score latent space by a lightweight decoder. Then, we construct an approximated likelihood score target and calculate the distance between the decoder's output and this target as an additional regularization loss. Finally, we jointly optimize the side network and decoder with both standard diffusion loss and our regularization loss. Experiments across various image/video tasks, architectures, and diffusion/flow models demonstrated that LISA can not only consistently accelerate the training convergence and improve final synthetic results, but also encourage the side network's features to be more disentangled for conditional modeling with negligible additional training cost and zero extra inference cost.
Chinese Translation
普遍采用的双分支范式，即训练一个侧网络以编码视觉条件，并将其中间层特征与一个冻结的预训练主网络融合，在视觉条件可控生成中取得了显著成功。尽管这一方法被广泛应用，但侧分支的作用及其训练效率仍然未被充分探讨。本文首先通过基于得分的生成建模视角重新审视这一主流范式：1）主网络通过提供一个先验的无条件得分来保持视觉感知质量。2）侧网络通过隐式贡献似然得分来引导条件控制。在这一视角的指导下，我们提出了似然得分对齐（LISA），这是一种有效的正则化方法，明确地将侧网络的中间特征与近似的似然得分对齐。具体而言，我们首先从侧网络的指定层提取特征，并通过一个轻量级解码器将其投影到得分潜在空间中。然后，我们构建一个近似的似然得分目标，并计算解码器输出与该目标之间的距离作为额外的正则化损失。最后，我们联合优化侧网络和解码器，同时使用标准扩散损失和我们的正则化损失。在各种图像/视频任务、架构和扩散/流模型的实验中，结果表明LISA不仅可以持续加速训练收敛并改善最终合成结果，还能促使侧网络的特征在条件建模中更加解耦，且几乎不增加额外的训练成本和零额外的推理成本。
View on arXiv Download PDF AI Translation
cs.CV / 78 / 2606.27223

SatSplatDiff: Geometry-preserving generative refinement for high-fidelity satellite Gaussian Splatting

SatSplatDiff：高保真卫星高斯溅射的几何保持生成细化
Kim, Jiyong, Song, Shuang, Qin, Ronjgun
Abstract
Gaussian Splatting has been recently explored for satellite 3D reconstruction, demonstrating flexibility and efficiency in representing radiometrically diverse satellite scenes. However, the limited top viewpoint of satellite imagery results in insufficient supervision on building facades, leaving surface holes and degraded visual fidelity. Generative refinement, which leverages pretrained generative priors to iteratively refine and update the rendered images used as supervision targets, has recently been investigated to improve the visual fidelity of Gaussian-rendered images. However, since these models refine each view independently, the resulting images can generate hallucinations and break photo-consistency, leading to geometric degradation. To address these limitations, we propose SatSplatDiff, which aims to minimize geometric degradation prevalent in generative refinement. Building on photogrammetric DSM initialization and 2DGS-based shadow casting established in our prior work SatSplat, we first introduce monocular depth supervision and multi-scale geometric refinement to establish a geometrically accurate and well-regularized surface representation. We then apply shadow-guided generative refinement, where geometrically calculated shadow maps guide the Gaussians to maintain consistency with the underlying geometry, improving visual fidelity while reducing geometric degradation. Extensive evaluations on the IARPA2016 and DFC2019 datasets demonstrate state-of-the-art performance, reducing geometric MAE by up to 18% and improving visual fidelity (FID-CLIP) by 28-45% over existing baselines. Our method delivers up to 5x resolution enhancement with minimal hallucination and sensor-consistent appearance, demonstrating seamless cross-tile consistency and strong scalability for large-scale reconstruction. Source code is available at https://github.com/GDAOSU/SatSplatDiff
Chinese Translation
高斯溅射最近被探索用于卫星三维重建，展示了在表示辐射多样的卫星场景方面的灵活性和效率。然而，卫星图像的有限顶视角导致对建筑立面的监督不足，留下表面孔洞和视觉保真度降低。生成细化利用预训练的生成先验，迭代细化和更新作为监督目标的渲染图像，最近被研究以提高高斯渲染图像的视觉保真度。然而，由于这些模型独立细化每个视图，导致生成的图像可能产生幻觉并破坏照片一致性，从而导致几何退化。为了解决这些限制，我们提出了SatSplatDiff，旨在最小化生成细化中普遍存在的几何退化。在我们之前的工作SatSplat中建立的摄影测量数字表面模型（DSM）初始化和基于2DGS的阴影投射的基础上，我们首先引入单目深度监督和多尺度几何细化，以建立几何准确且良好正则化的表面表示。然后，我们应用阴影引导的生成细化，其中几何计算的阴影图引导高斯保持与基础几何的一致性，提高视觉保真度，同时减少几何退化。在IARPA2016和DFC2019数据集上的广泛评估表明，我们的方法在性能上达到了最先进的水平，几何平均绝对误差（MAE）降低了多达18%，视觉保真度（FID-CLIP）提高了28-45%，相较于现有基线。我们的方法在保持最小幻觉和传感器一致外观的同时，提供了高达5倍的分辨率提升，展示了无缝的跨瓷砖一致性和强大的大规模重建可扩展性。源代码可在 https://github.com/GDAOSU/SatSplatDiff 获取。
View on arXiv Download PDF AI Translation
cs.CV / 79 / 2606.27264

CORTEX: A Structured Reasoning Benchmark for Trustworthy 3D Chest CT MLLMs

CORTEX：一个用于可信赖的3D胸部CT多模态大语言模型的结构化推理基准
Malik, Hashmat Shadab, Hashmi, Anees Ur Rehman, Saeed, Numan, Naseer, Muzammal, Khan, Salman, Lippert, Christoph
Abstract
Reasoning in multimodal large language models (MLLMs) has shown strong promise in medical imaging. However, this reasoning is usually free-form text judged only by its final answer, making it hard to interpret and verify, especially in 3D radiology, where a diagnosis should be traceable to evidence in the scan. Existing chest CT question-answering datasets compound this by reducing expert radiology reports to answer-only pairs, dropping the reasoning that links findings to conclusions and omitting the patient history clinicians rely on. As a result, reasoning-capable 3D chest CT MLLMs remain out of reach, as neither the structured supervision needed to train them nor the protocol needed to verify their reasoning yet exists. We introduce CORTEX (Clinically Organized Reasoning and sTructured EXplanation), a structured reasoning benchmark for 3D chest CT. For each question, CORTEX restores the missing reasoning as a four-stage diagnostic trace mirroring a radiologist's workflow: task understanding, visual observation, diagnostic reasoning, and answer synthesis. We generate these traces using frontier large language models with broad medical and general-domain knowledge, then filter and verify them with a stage-level evaluation protocol combining automated rubric scoring with expert radiologist review. Crucially, both the reasoning structure and evaluation rubrics are designed in close collaboration with clinicians. Built on CT-RATE, a large, publicly available chest CT dataset without reasoning annotations, CORTEX comprises 76,177 validated reasoning traces across open-ended VQA, closed-ended VQA, and report generation, providing both the structured supervision and the stage-level evaluation protocol needed to build and evaluate trustworthy reasoning models for 3D chest CT. Our dataset and evaluation code will be made publicly available upon acceptance.
Chinese Translation
多模态大语言模型（MLLMs）在医学影像中的推理显示出强大的潜力。然而，这种推理通常是自由形式的文本，仅通过最终答案进行评判，这使得其难以解释和验证，尤其是在3D放射学中，诊断应当能够追溯到扫描中的证据。现有的胸部CT问答数据集通过将专家放射学报告简化为仅包含答案的对，进一步加剧了这一问题，忽略了将发现与结论联系起来的推理，并省略了临床医生所依赖的患者历史。因此，具备推理能力的3D胸部CT MLLMs仍然难以实现，因为既没有训练所需的结构化监督，也没有验证其推理所需的协议。我们提出了CORTEX（临床组织推理与结构化解释），这是一个用于3D胸部CT的结构化推理基准。对于每个问题，CORTEX恢复缺失的推理，形成一个四阶段的诊断轨迹，反映放射科医生的工作流程：任务理解、视觉观察、诊断推理和答案综合。我们使用前沿的大语言模型生成这些轨迹，这些模型具备广泛的医学和通用领域知识，然后通过结合自动评分标准与专家放射科医生审查的阶段性评估协议进行筛选和验证。至关重要的是，推理结构和评估标准是在与临床医生密切合作的基础上设计的。CORTEX基于CT-RATE，这是一个大型、公开可用的无推理注释的胸部CT数据集，包含76,177个经过验证的推理轨迹，涵盖开放式VQA、封闭式VQA和报告生成，为构建和评估可信赖的3D胸部CT推理模型提供了所需的结构化监督和阶段性评估协议。我们的数据集和评估代码将在接受后公开发布。
View on arXiv Download PDF AI Translation
cs.CV / 80 / 2606.27280

Exact and Deterministic Patch Descriptor Retrieval via Hierarchical Normalization

通过层次归一化实现精确且确定性的补丁描述符检索
Sato, Koichi
Abstract
We present a patch descriptor retrieval method that returns the exact nearest neighbour -- provably identical to exhaustive full-vector search -- while evaluating only a small fraction of the database, and does so deterministically: the same (database, query) pair always produces the same result, independent of run order, thread count, or hardware. This contrasts with approximate nearest-neighbour (ANN) approaches such as HNSW and IVF-PQ, which trade exactness for speed and may return different results across runs. The enabling mechanism is Hierarchical Normalization (HN): a normalisation scheme that splits the pre-normalisation feature vector into a K-dim major component (norm sqrt(1-alpha)) and a (128-K)-dim minor component (norm sqrt(alpha)). Since the minor inner product is bounded by alpha (Cauchy-Schwarz on the prescribed norms), the major similarity plus alpha is an admissible upper bound on the full similarity: the search scans the K-dim major component for all entries, then applies full 128-dim evaluation only to entries that cannot be pruned -- a provably exact branch-and-bound scan. We train HN-modified HardNet on the notredame split of the UBC patch dataset and evaluate on trevi and halfdome. With a cache-optimised Structure-of-Arrays layout and K=8, alpha=1/32, the search achieves 13.7x (trevi) / 12.7x (halfdome) speed-up over brute-force 128-dim search, with only 0.4% of entries requiring full evaluation. At K=16, alpha=1/8, FPR@95 rises from 0.0062 to 0.0064 on trevi at 7.2x speed-up, with 98.8% of entries bypassing full evaluation.
Chinese Translation
我们提出了一种补丁描述符检索方法，该方法返回精确的最近邻——在理论上与全面的全向量搜索相同——同时仅评估数据库的一小部分，并且是确定性的：相同的（数据库，查询）对始终产生相同的结果，与运行顺序、线程数量或硬件无关。这与近似最近邻（ANN）方法如 HNSW 和 IVF-PQ 相对立，后者为了速度而牺牲了精确性，并且可能在不同的运行中返回不同的结果。其关键机制是层次归一化（Hierarchical Normalization, HN）：一种归一化方案，将预归一化特征向量分割为 K 维主成分（范数 sqrt(1-alpha)）和 (128-K) 维次成分（范数 sqrt(alpha)）。由于次成分的内积受到 alpha 的限制（根据规定的范数应用柯西-施瓦茨不等式），主相似度加上 alpha 是全相似度的可接受上界：搜索扫描所有条目的 K 维主成分，然后仅对无法剪枝的条目应用全 128 维评估——这是一种可证明的精确分支限界扫描。我们在 UBC 补丁数据集的 notredame 划分上训练了 HN 修改的 HardNet，并在 trevi 和 halfdome 上进行评估。通过缓存优化的数组结构布局和 K=8，alpha=1/32，该搜索在暴力 128 维搜索中实现了 13.7 倍（trevi）/ 12.7 倍（halfdome）的加速，只有 0.4% 的条目需要全评估。在 K=16，alpha=1/8 的情况下，trevi 上的 FPR@95 从 0.0062 上升到 0.0064，速度提升为 7.2 倍，98.8% 的条目跳过了全评估。
View on arXiv Download PDF AI Translation
cs.CV / 81 / 2606.27305

Sculpting NeRF Geometry: Human-Preference Fine-Tuning of a 3D-Aware Face GAN

雕塑化 NeRF 几何体：基于人类偏好的 3D 感知面部 GAN 微调
Moore, Archer, Gong, Mingming, Hodgkinson, Liam
Abstract
Reinforcement learning from human feedback (RLHF) for 3D generation is now established across a number of works, but most existing pipelines optimise explicit surface representations, often by converting radiance fields into meshes and training heavily on surface-supervised data. We instead fine-tune a pretrained 3D-aware generative model directly from a learned reward over radiance-field density ($\sigma$) values, with no externally supplied mesh or shape prior. The reward model requires no pretraining, trains easily on a small set of preference samples, and yields robust improvement in 3D geometry. Working on an unconditional 3D-aware face GAN (EG3D), our reward reads the continuous 3D density field of the neural radiance field (NeRF) directly and supplies a geometry-only learning signal, requiring neither text conditioning, mesh extraction, nor multi-view rendering. A density-consistency constraint keeps the 2D appearance qualitatively similar while the geometry is reshaped, at a measurable but bounded distributional cost (FID-50k rises from 4.09 to 6.66): the fine-tuned generator, trained from the preferences of a single annotator as a proof of concept, produces face geometries preferred by users in 74.4% of pairwise comparisons.
Chinese Translation
基于人类反馈的强化学习（RLHF）在 3D 生成领域已在多项研究中得到确立，但大多数现有流程优化显式表面表示，通常通过将辐射场转换为网格并在表面监督数据上进行大量训练。我们则直接从学习到的辐射场密度 ($ ho$) 值中微调一个预训练的 3D 感知生成模型，而不依赖于外部提供的网格或形状先验。该奖励模型无需预训练，可以在一小组偏好样本上轻松训练，并在 3D 几何体上带来稳健的改进。我们在一个无条件的 3D 感知面部 GAN（EG3D）上工作，奖励模型直接读取神经辐射场（NeRF）的连续 3D 密度场，并提供仅几何体的学习信号，无需文本条件、网格提取或多视角渲染。密度一致性约束在几何体重塑的同时保持 2D 外观的定性相似，代价在可测量但有限的分布成本下（FID-50k 从 4.09 上升至 6.66）：经过微调的生成器作为概念验证，从单个标注者的偏好中训练，产生了在 74.4% 的成对比较中被用户偏好的面部几何体。
View on arXiv Download PDF AI Translation
cs.CV / 82 / 2606.27307

See & Sniff: Learning Visuo-Olfactory Representations

看与嗅：学习视觉-嗅觉表征
Kim, Seongyu, Lee, Seungwoo, Ryu, Hyeonggon, Chung, Joon Son, Senocak, Arda
Abstract
While modern multimodal models integrate vision with language, audio, or touch, olfaction remains largely unexplored due to the lack of paired visuo-olfactory data. We introduce SmellNet-V, a scalable visuo-olfactory dataset built on the insight that odor identity is largely invariant to visual transformations within a semantic category. This allows us to synthetically pair smell-only samples with semantically aligned in-the-wild web images, converting a unimodal olfactory dataset into a cross-modal benchmark without costly co-collection. Building on this dataset, we propose See & Sniff, a self-supervised framework that learns joint visuo-olfactory representations via dense local alignment and naturally produces smell saliency maps for spatial grounding of odor sources. We further introduce pixel-level smell localization task and a benchmark for evaluation. Our method surpasses smell-only baselines by 7% in smell classification from smell alone and generalizes to cross-modal retrieval and smell localization, establishing visuo-olfactory learning as a new direction in multimodal perception.
Chinese Translation
尽管现代多模态模型将视觉与语言、音频或触觉整合，但由于缺乏配对的视觉-嗅觉数据，嗅觉仍然在很大程度上未被探索。我们介绍了 SmellNet-V，这是一个可扩展的视觉-嗅觉数据集，基于一个洞察，即气味的特征在语义类别内对视觉变换基本不变。这使我们能够将仅有气味的样本与语义上对齐的自然界网络图像进行合成配对，将单模态的嗅觉数据集转化为跨模态基准，而无需昂贵的共同收集。在此数据集的基础上，我们提出了 See & Sniff，这是一个自监督框架，通过密集的局部对齐学习联合视觉-嗅觉表征，并自然生成气味显著性图，以实现气味源的空间定位。我们进一步引入了像素级气味定位任务和评估基准。我们的方法在仅依靠气味的气味分类中比气味单一基线提高了 7%，并且在跨模态检索和气味定位中具有良好的泛化能力，确立了视觉-嗅觉学习作为多模态感知的新方向。
View on arXiv Download PDF AI Translation
cs.CV / 83 / 2606.27313

ViQ: Text-Aligned Visual Quantized Representations at Any Resolution

ViQ：任何分辨率下的文本对齐视觉量化表示
Yu, Xumin, Liu, Zuyan, Yang, Zhenyu, Dong, Yuhao, Qian, Shengsheng, Lu, Jiwen, Hu, Han, Rao, Yongming
Abstract
A unified representation for text and vision is a natural pursuit, as it enables simpler multimodal modeling and more efficient training. However, representing images as discrete signals in the same way as text inevitably introduces severe information loss. Existing work struggles to balance low-level details and high-level semantics in discrete representations: reconstruction-oriented representations often lack semantic information, whereas semantically stronger features typically suffer from severe loss of detail. We present ViQ, a Visual Quantized Representations framework, which is designed to balance semantics and details in discrete representations while supporting inputs at native resolutions, thereby enabling it to serve as a unified and general discrete representation for arbitrary visual inputs. Our approach structures quantization learning into two stages: text-aligned pre-training and feature discretization. With text-aligned pre-training, we enhance the visual encoder semantic-rich supervision from the pretrained language model and enable it to process native-resolution visual inputs. During discretization, we propose a proximal representation learning strategy to progressively compact the feature space, along with a position-aware head-wise quantization mechanism that enables flexible processing of arbitrary resolutions. Extensive experiments on multimodal tasks demonstrate that ViQ achieves competitive performance compared to state-of-the-art multimodal vision encoders with continuous and high-dimensional visual features, while maintaining high precision in low-level reconstruction. We also show that multimodal training with visual quantized representations largely improves efficiency, yielding up to 20\%-70\% acceleration with different base LLMs and training recipes.
Chinese Translation
文本与视觉的统一表示是一项自然的追求，因为它能够简化多模态建模并提高训练效率。然而，以与文本相同的方式将图像表示为离散信号不可避免地会导致严重的信息损失。现有的研究在离散表示中平衡低级细节与高级语义方面面临挑战：以重建为导向的表示往往缺乏语义信息，而语义更强的特征通常会遭受严重的细节损失。我们提出了ViQ，一个视觉量化表示框架，旨在平衡离散表示中的语义与细节，同时支持原生分辨率的输入，从而使其能够作为任意视觉输入的统一和通用离散表示。我们的方法将量化学习结构化为两个阶段：文本对齐的预训练和特征离散化。在文本对齐的预训练中，我们增强了视觉编码器的语义丰富监督，来自预训练的语言模型，并使其能够处理原生分辨率的视觉输入。在离散化过程中，我们提出了一种近似表示学习策略，以逐步压缩特征空间，并结合一种位置感知的头部量化机制，使其能够灵活处理任意分辨率。针对多模态任务的广泛实验表明，ViQ在与具有连续和高维视觉特征的最先进多模态视觉编码器的比较中表现出竞争力，同时在低级重建中保持高精度。我们还表明，使用视觉量化表示的多模态训练大大提高了效率，在不同基础大型语言模型（LLMs）和训练方案下实现了高达20%-70%的加速。
View on arXiv Download PDF AI Translation
cs.CV / 84 / 2606.27317

OctoSense: Self-Supervised Learning for Multimodal Robot Perception

OctoSense：用于多模态机器人感知的自监督学习
Bisulco, Anthony, Wang, Jeremy, Daniilidis, Kostas, Balestriero, Randall, Chaudhari, Pratik
Abstract
We present OctoSense, an open-source sensor platform with stereo RGB and event cameras, LiDAR, a thermal camera, an inertial measurement unit, RTK-corrected global positioning system, and proprioception (CAN bus data from a car, and joint angles for a quadruped robot). The eponymous OctoSense dataset contains 59 hours of time-synchronized driving data across different types of environments at different times of the day, including situations with highly degraded sensors. We demonstrate multi-modal self-supervised learning using such real-world robotics data, where sensors have different representations, frequencies, latencies and noise. Our approach, a "late-fusion" masked autoencoder, (i) uses modality-specific tokenizers to account for different spatiotemporal characteristics of these sensors, and (ii) caches modality-specific tokens at inference time to process new measurements as they come. This architecture (i) is fast (6.68 ms and 112 ms on NVIDIA 5090 and Orin NX respectively, to compute the representation), (ii) performs better than existing image-only foundation models on tasks such as estimation of optical flow, depth, semantic segmentation, and ego-motion (translation, rotation, and steering angle), and (iii) predicts robustly at nighttime or in situations where sensory data is degraded. See our project page for links to the dataset, code, and supplementary videos: https://abisulco.com/octosense/.
Chinese Translation
我们提出了OctoSense，一个开源传感器平台，配备立体RGB和事件相机、激光雷达、热成像相机、惯性测量单元、RTK校正的全球定位系统以及本体感知（来自汽车的CAN总线数据和四足机器人关节角度）。同名的OctoSense数据集包含59小时的时间同步驾驶数据，涵盖不同类型的环境和一天中的不同时间，包括传感器严重退化的情况。我们展示了使用这些真实世界机器人数据进行多模态自监督学习，其中传感器具有不同的表示、频率、延迟和噪声。我们的方法是一种“后融合”掩蔽自编码器，(i) 使用特定模态的分词器来考虑这些传感器的不同时空特性，(ii) 在推理时缓存特定模态的标记，以处理新测量数据。该架构(i) 速度快（在NVIDIA 5090和Orin NX上分别为6.68毫秒和112毫秒计算表示），(ii) 在光流、深度、语义分割和自我运动（平移、旋转和转向角度）等任务上表现优于现有的仅图像基础模型，(iii) 在夜间或传感器数据退化的情况下具有稳健的预测能力。有关数据集、代码和补充视频的链接，请参见我们的项目页面：https://abisulco.com/octosense/。
View on arXiv Download PDF AI Translation
cs.CV / 85 / 2606.27325

Not All Actions Are Equal: Rethinking Conditioning for Dexterous World Model

并非所有动作都是平等的：重新思考灵巧世界模型的条件化
Yuan, Zizhao, Liang, Zhengtu, Wang, Taowen, Liang, Qiwei, Wang, Yichi, Wang, Yunheng, Fang, Yuetong, Li, Lusong, Zeng, Zecui, Xu, Renjing
Abstract
Recent advances in action-conditioned world models show promising progress in modeling complex interactions and forecasting future states under diverse action sequences. While these models are often driven by stronger visual representations and model capacity, action conditioning itself remains underexplored. Most existing approaches compress the entire action sequence into a single representation, which works well for low-DoF control but becomes less reliable in high-DoF scenarios. We observe that high-DoF dexterous actions are inherently heterogeneous, spanning multiple orders of magnitude, where large-scale motions coexist with subtle but important signals. When uniformly aggregated, optimization exhibits an imbalance across action components, which hinders the modeling of fine-grained effects and affects action fidelity. We therefore propose DexAC-WM, which treats action conditioning as a structured process rather than global compression. DexAC preserves dimension-level semantics via action tokenization and aligns action signals with visual dynamics through local refinement and global modulation. To address the limited high-level semantic grounding in existing world models, we further introduce a semantic branch that provides rich object-scene priors, which enables world model to capture dynamic visual details while supporting high-DoF action-conditioned video prediction. Experiments on EgoDex and EgoVerse show that combining the semantic branch with DexAC significantly improves FID, FVD, and PCK, demonstrating gains in visual-temporal realism and action-following consistency. We further verify that DexAC extends to other backbones, showing the scalability of our structured action-conditioning design. These results suggest that scaling world models to high-DoF control requires both structured action modeling and semantic grounding.
Chinese Translation
近期在动作条件化世界模型方面的进展显示出在建模复杂交互和预测多样动作序列下的未来状态方面取得了可喜的进展。尽管这些模型通常依赖于更强的视觉表征和模型能力，但动作条件化本身仍然未被充分探讨。现有的大多数方法将整个动作序列压缩为单一表征，这在低自由度（DoF）控制中表现良好，但在高自由度场景中则变得不够可靠。我们观察到，高自由度的灵巧动作本质上是异质的，跨越多个数量级，其中大规模动作与细微但重要的信号共存。当均匀聚合时，优化在动作组件之间表现出不平衡，这阻碍了对细粒度效应的建模，并影响了动作的真实感。因此，我们提出了DexAC-WM，它将动作条件化视为一个结构化过程，而非全局压缩。DexAC通过动作标记化保留了维度级语义，并通过局部细化和全局调制将动作信号与视觉动态对齐。为了应对现有世界模型中有限的高层语义基础，我们进一步引入了一个语义分支，提供丰富的物体-场景先验，使世界模型能够捕捉动态视觉细节，同时支持高自由度的动作条件化视频预测。在EgoDex和EgoVerse上的实验表明，将语义分支与DexAC结合显著提高了FID、FVD和PCK，展示了视觉-时间真实感和动作跟随一致性的提升。我们进一步验证了DexAC可以扩展到其他骨干网络，显示了我们结构化动作条件化设计的可扩展性。这些结果表明，将世界模型扩展到高自由度控制需要结构化的动作建模和语义基础。
View on arXiv Download PDF AI Translation
cs.CV / 86 / 2606.27332

RoPEMover: Depth-Aware Object Relocation via Positional Embeddings

RoPEMover：基于位置嵌入的深度感知物体重定位
Oztas, Ipek, Ceylan, Duygu, Aksoy, Aybars Bugra, Dundar, Aysegul
Abstract
Moving an object in a single image requires geometry-consistent spatial rearrangement, including handling occlusions, revealing previously unseen regions, and maintaining coherent shadows and reflections. Existing approaches are not well suited to this setting and often fail to preserve such scene-level consistency. We address this problem by introducing a geometry-aware object motion method that operates directly on the positional representations of diffusion transformers. Our key insight is that rotary positional embeddings (RoPE) define a structured spatial field that can be explicitly manipulated to induce controlled motion. We extend 2D RoPE into a depth-aware formulation that encodes 3D spatial structure, enabling consistent object displacement and scene-aware updates. Our model is trained using synthetic data combined with a small set of real images via parameter-efficient fine-tuning. Despite minimal real supervision, it preserves object identity under large spatial displacements, generates plausible content in newly revealed regions, and consistently updates scene-dependent effects such as shadows and illumination. Experimental results on standard object motion benchmarks demonstrate state-of-the-art performance across all evaluation metrics.
Chinese Translation
在单幅图像中移动物体需要几何一致的空间重排，包括处理遮挡、揭示先前未见的区域，以及保持一致的阴影和反射。现有的方法在这一设置下并不适用，往往无法保持场景级的一致性。我们通过引入一种几何感知的物体运动方法来解决这一问题，该方法直接作用于扩散变换器的位置信息表示。我们的关键见解是，旋转位置嵌入（Rotary Positional Embeddings, RoPE）定义了一个结构化的空间场，可以被显式操控以引发受控运动。我们将二维 RoPE 扩展为一种深度感知的形式，编码三维空间结构，从而实现一致的物体位移和场景感知更新。我们的模型通过合成数据与少量真实图像结合进行参数高效的微调进行训练。尽管真实监督极少，但它在大空间位移下保持物体身份，在新揭示的区域生成合理的内容，并一致地更新场景依赖的效果，如阴影和照明。在标准物体运动基准上的实验结果表明，我们的方法在所有评估指标上均表现出最先进的性能。
View on arXiv Download PDF AI Translation
cs.CV / 87 / 2606.27339

SAM2Matting: Generalized Image and Video Matting

SAM2Matting：通用图像与视频抠图
Shen, Ruiqi, Jie, Guangquan, Liu, Chang, Ding, Henghui
Abstract
Despite impressive advances in image matting, video matting remains challenging due to the inherent gap between high-level tracking, which requires frame-wise understanding, and low-level matting, which focuses on extremely fine-grained details. Existing methods attempt this with expensive and narrowly-scoped video matting datasets, which may limit out-of-domain generalization and compromise tracking robustness. We rethink the paradigm with SAM2Matting, a tracker-to-matting framework that advances VOS trackers to high-fidelity video matting. Specifically, it decouples the task by enhancing a foundational tracker (e.g., SAM2, SAM3) with a region-proposal bridge and dedicated matting heads, enabling the uncompromised tracker to handle temporal consistency while the matting components resolve fine-grained details. Notably, despite being trained only on images, SAM2Matting establishes new state-of-the-art performance on video matting, supports diverse prompt types, maintains strong temporal consistency, and demonstrates robust generalization across both human-centric and in-the-wild scenarios.
Chinese Translation
尽管图像抠图取得了显著进展，但视频抠图仍然面临挑战，因为高层次的跟踪需要逐帧理解，而低层次的抠图则专注于极其细致的细节。现有方法通常依赖昂贵且范围狭窄的视频抠图数据集，这可能限制了域外泛化能力并妨碍跟踪的鲁棒性。我们通过SAM2Matting重新思考这一范式，提出了一种跟踪到抠图的框架，将视频对象分割（VOS）跟踪器提升至高保真视频抠图。具体而言，该框架通过增强基础跟踪器（例如，SAM2、SAM3）与区域提议桥和专用抠图头，解耦任务，使得不妥协的跟踪器能够处理时间一致性，而抠图组件则解决细致的细节。值得注意的是，尽管仅在图像上进行训练，SAM2Matting在视频抠图上建立了新的最先进性能，支持多种提示类型，保持强大的时间一致性，并在以人为中心和自然场景中展现出强大的泛化能力。
View on arXiv Download PDF AI Translation
cs.CV / 88 / 2606.27345

RayPE: Ray-Space Positional Encoding for 3D-Aware Video Generation

RayPE：用于3D感知视频生成的光线空间位置编码
Yin, Minghao, Lu, Jiahao, Hu, Wenbo, Zhao, Wang, Ying, Shan, Han, Kai
Abstract
Modern video diffusion transformers position their tokens through RoPE on the (u,v,t) axes -- a description of the camera's sampling grid that says nothing about the 3D structure of the scene. We observe that the geometric relation between two camera rays is captured by the Plucker reciprocal product, which is bilinear in the two rays -- the same algebraic form as the dot product in Transformer attention. Building on this analogy, we propose RayPE, a positional-encoding extension that injects per-token 6D Plucker coordinates additively into the queries and keys of self-attention, with a query/key flip arrangement under which the symmetric identity configuration coincides exactly with the reciprocal product. The injection is additive, the resulting attention score decomposes into a content term, a geometry term, and two content and geometry cross-terms -- all of which our experiments find individually necessary. To make the encoding stable across video data with heterogeneous camera-translation scales (SfM, deep SLAM, metric), we further decouple ray direction from moment magnitude, gate the encoding by a learned function of the log-magnitude, and apply RMSNorm to align it with the QKNorm-normalized content branch. The full module adds less than 0.1% parameters to a pretrained video DiT, is zero-initialized to start from the pretrained weights, and improves camera controllability, cross-frame 3D consistency, and overall video quality on a four-dataset training mixture.
Chinese Translation
现代视频扩散变换器通过 RoPE 在 (u,v,t) 轴上定位其标记——这描述了相机采样网格，但并未涉及场景的三维结构。我们观察到，两条相机光线之间的几何关系由 Plucker 互反积捕获，该关系在两条光线中是双线性的——与变换器注意力中的点积具有相同的代数形式。基于这一类比，我们提出了 RayPE，这是一种位置编码扩展，它将每个标记的 6D Plucker 坐标以加法方式注入到自注意力的查询和键中，并采用查询/键翻转的排列方式，使得对称身份配置与互反积完全一致。注入是加法的，结果注意力得分分解为内容项、几何项以及两个内容与几何的交叉项——我们的实验发现这些项都是单独必要的。为了使编码在具有异质相机平移尺度（SfM、深度 SLAM、度量）的视数据中稳定，我们进一步将光线方向与时刻大小解耦，通过学习的对数大小函数对编码进行门控，并应用 RMSNorm 使其与 QKNorm 归一化的内容分支对齐。完整模块对预训练视频 DiT 增加的参数少于 0.1%，以零初始化开始于预训练权重，并提高了相机可控性、跨帧 3D 一致性以及在四个数据集训练混合上的整体视频质量。
View on arXiv Download PDF AI Translation
cs.CV / 89 / 2606.27371

Don't Settle at the Mode! Mitigating Diversity Collapse in Pretrained Flow Models via Feature Self-Guidance

不要满足于模式！通过特征自引导减轻预训练流模型中的多样性崩溃
Bhat, Pradhaan S, Parihar, Rishubh, Bhat, Abhijnya, Babu, R. Venkatesh
Abstract
State-of-the-art flow models generate stunning images from text or image prompts. However, they suffer from diversity collapse when generating multiple samples under the same conditioning. Existing methods address this issue via either latent guidance, which has limited effectiveness, or sample selection, which relies on external reward models that incur significant inference-time overhead. In this work, we introduce an efficient, training-free self-guidance mechanism to mitigate diversity collapse without requiring additional reward models. Specifically, we disperse the internal features of the flow model during batch generation with feature self-guidance. Further, to keep the features close to the manifold, we introduce a manifold regularization step that projects these dispersed features back onto the data manifold, ensuring diverse generation without sacrificing alignment with the input conditions. Our method integrates seamlessly as a plug-and-play module into pretrained flow models, adding only a marginal inference cost. Experiments demonstrate significant improvements in diversity while preserving fidelity across several conditional flow models, including multi-step and few-step text-to-image, depth-to-image, and reference image generation.
Chinese Translation
最先进的流模型能够从文本或图像提示生成令人惊叹的图像。然而，在相同条件下生成多个样本时，它们会遭遇多样性崩溃的问题。现有的方法通过潜在引导（latent guidance）来解决此问题，但效果有限；或者通过样本选择（sample selection），这依赖于外部奖励模型，导致显著的推理时间开销。在本研究中，我们提出了一种高效的、无训练的自引导机制，以减轻多样性崩溃，而无需额外的奖励模型。具体而言，我们在批量生成过程中通过特征自引导来分散流模型的内部特征。此外，为了保持特征接近流形（manifold），我们引入了一个流形正则化步骤，将这些分散的特征投影回数据流形，确保在不牺牲与输入条件对齐的情况下实现多样化生成。我们的方法可以无缝集成到预训练流模型中，作为即插即用模块，仅增加微小的推理成本。实验表明，在多个条件流模型（包括多步和少步文本到图像、深度到图像以及参考图像生成）中，显著提高了多样性，同时保持了保真度。
View on arXiv Download PDF AI Translation
cs.CV / 90 / 2606.27372

DnA: Denoising Attention for Visual Tasks

DnA：用于视觉任务的去噪注意力
Campos, Ron, Maity, Subhajit, Li, Xin, Das, Srijan, Dutta, Aritra
Abstract
The softmax activation in multihead attention (MHA) is the de facto standard for attention-based models in visual perception tasks. However, standard softmax can produce noisy attention patterns that dilute relevant features and degrade its performance. In this paper, we propose Denoising Attention or DnA, in which, first, a positive query identifies which image features belong to the correct class, and a negative query identifies closely associated but irrelevant image features. DnA then projects these interactions into two distinct subspaces with larger principal angles, promoting subspace separation and improved discriminability. Using a ViT-B backbone, our proposed DnA achieves an absolute gain of 0.8% on ImageNet-1K compared to the baseline. We further show improvements across multiple visual understanding tasks, including video understanding with video transformers (1.8%) and video LLMs (0.5%). Our extensive empirical analyses justify the design choices involving two interacting subspaces and the denoising effect of DnA.
Chinese Translation
多头注意力（MHA）中的softmax激活是视觉感知任务中基于注意力模型的事实标准。然而，标准softmax可能产生噪声注意力模式，这会稀释相关特征并降低其性能。本文提出了去噪注意力（Denoising Attention，DnA），首先，通过正查询识别哪些图像特征属于正确类别，而负查询则识别紧密相关但不相关的图像特征。DnA随后将这些交互投影到两个具有更大主角度的不同子空间中，从而促进子空间分离和提高可区分性。使用ViT-B骨干网络，我们提出的DnA在ImageNet-1K上相比基线实现了0.8%的绝对提升。我们进一步展示了在多个视觉理解任务中的改进，包括使用视频变换器的视频理解（1.8%）和视频大语言模型（0.5%）。我们广泛的实证分析证明了涉及两个交互子空间的设计选择及DnA的去噪效果。
View on arXiv Download PDF AI Translation
cs.CV / 91 / 2606.27373

Paying More Attention to Visual Tokens in Self-Evolving Large Multimodal Models

在自我演化的大型多模态模型中更关注视觉标记
Venkatraman, Shravan, Thawkar, Ritesh, Thawakar, Omkar, Anwer, Rao Muhammad, Cholakkal, Hisham, Khan, Salman, Khan, Fahad
Abstract
Recently, self-evolving large multimodal models (LMMs) have received attention for improving visual reasoning in a purely unsupervised setting. However, multi-role self-play and self-consistency reward schemes in existing self-evolving LMMs optimize answer agreement without ensuring the decoder attends to visual content, relying instead on statistical language priors to produce self consistent outputs. This leads to a persistent failure mode we term visual under-conditioning, where the decoder relies on language priors rather than the image during generation, manifesting as insufficient attention to visual tokens. As a result, current self-evolving LMMs struggle on vision--language understanding tasks such as image captioning and visual question answering. To address this, we propose VISE (Visual Invariance Self-Evolution), a purely unsupervised self-evolving framework that directly regularizes the model's visual conditioning policy through two complementary invariance-based rewards: a geometric invariance reward that enforces spatial consistency under known transformations, and a semantic invariance reward that penalizes evidence-agnostic generation by requiring the model to recognize the absence of evidence when predicted regions are perturbed. VISE operates within a single model without specialist roles, external reward models, or annotations, and is trained on raw unlabeled images. Experiments on 18 benchmarks demonstrate the efficacy of our approach. Using Qwen3-VL-2B as the base model, VISE achieves gains of $+16.85$ CIDEr on COCO and $+19.66$ CIDEr on TextCaps, reduces object hallucination by $5.0$ Chair-I points, and generalizes across four model families and scales. Our code and models are available at https://mbzuai-oryx.github.io/VISE
Chinese Translation
最近，自我演化的大型多模态模型（LMMs）因其在完全无监督环境中改善视觉推理而受到关注。然而，现有自我演化LMM中的多角色自我博弈和自我一致性奖励机制优化了答案一致性，却未能确保解码器关注视觉内容，而是依赖统计语言先验来生成自一致的输出。这导致了一种我们称之为视觉欠条件化的持续失效模式，其中解码器在生成过程中依赖语言先验而非图像，表现为对视觉标记的关注不足。因此，目前的自我演化LMM在图像字幕生成和视觉问答等视觉-语言理解任务上表现不佳。为了解决这个问题，我们提出了VISE（视觉不变性自我演化），这是一种完全无监督的自我演化框架，通过两种互补的不变性奖励直接规范模型的视觉条件策略：几何不变性奖励在已知变换下强制执行空间一致性，语义不变性奖励通过要求模型在预测区域受到扰动时识别证据缺失来惩罚与证据无关的生成。VISE在单一模型中运行，无需专业角色、外部奖励模型或注释，并在原始未标记图像上进行训练。在18个基准测试上的实验证明了我们方法的有效性。以Qwen3-VL-2B作为基础模型，VISE在COCO上获得了$+16.85$的CIDEr提升，在TextCaps上获得了$+19.66$的CIDEr提升，减少了$5.0$ Chair-I点的对象幻觉，并在四个模型家族和规模上实现了泛化。我们的代码和模型可在https://mbzuai-oryx.github.io/VISE获取。
View on arXiv Download PDF AI Translation
cs.CV / 92 / 2606.27377

DanceOPD: On-Policy Generative Field Distillation

DanceOPD：基于策略的生成场蒸馏
Zhou, Wei, Zhu, Xiongwei, Xu, Zelin, Dong, Bo, Gong, Lixue, Liang, Yongyuan, Chu, Meng, Qu, Leigang, Kong, Lingdong, Liu, Wei, Chua, Tat-Seng
Abstract
Modern image generation demands a single model that unifies diverse capabilities, including text-to-image (T2I), local editing, and global editing. However, these capabilities are rarely naturally aligned and often conflict. For instance, editing tends to degrade T2I performance, while global and local editing interfere with each other. Consequently, effectively composing these capabilities has become a central challenge for image generation model training. To tackle this, we introduce DanceOPD, an on-policy generative field distillation framework for flow-matching models that routes each sample to one capability field, queries one low-noise student-induced state, and trains with a simple velocity MSE objective. With each capability source defined as a velocity field over the shared flow state space, the student learns from fields queried on its own rollout states to compose expert capabilities. This formulation also absorbs operator-defined fields such as classifier-free guidance. Comprehensive experiments on T2I, editing, realism-field absorption, and CFG absorption show that our approach improves multi-capability composition, strengthening target capabilities while preserving anchor generation quality. We believe this work establishes a practical route for generative field distillation in flow-matching models.
Chinese Translation
现代图像生成需要一个统一多种能力的单一模型，包括文本到图像（T2I）、局部编辑和全局编辑。然而，这些能力往往不自然对齐且相互冲突。例如，编辑往往会降低T2I的性能，而全局编辑和局部编辑之间也会相互干扰。因此，有效地组合这些能力已成为图像生成模型训练的一个核心挑战。为了解决这个问题，我们提出了DanceOPD，一种基于策略的生成场蒸馏框架，适用于流匹配模型，该框架将每个样本路由到一个能力场，查询一个低噪声的学生诱导状态，并使用简单的速度均方误差（MSE）目标进行训练。每个能力源被定义为共享流状态空间上的速度场，学生从其自身的回放状态查询的场中学习，以组合专家能力。该公式还吸收了操作员定义的场，例如无分类器引导。关于T2I、编辑、现实场吸收和CFG吸收的全面实验表明，我们的方法改善了多能力组合，增强了目标能力，同时保持了锚定生成质量。我们相信这项工作为流匹配模型中的生成场蒸馏建立了一条实用的路径。
View on arXiv Download PDF AI Translation
人工智能 (Artificial Intelligence)
73
cs.AI / 1 / 2606.26155

Detecting and Controlling Sycophancy with Cascading Linear Features

通过级联线性特征检测和控制谄媚行为
Bohacek, Maty, Jain, Rishub, Dufour, Nicholas, Leung, Thomas, Bregler, Chris, Patel, Roma
Abstract
Interpreting and controlling model behaviors through activation steering methods requires many pairs of contrastive samples that clearly exhibit desired or undesired behavior. These data pairs determine the degree to which interpretability frameworks can reliably detect model features responsible for a behavior, and therefore the ability to steer models toward or away from such behavior. In this work, we present an iterative data generation pipeline that isolates cascading linear features responsible for a behavior. Specifically, we show how moving beyond simple binary pairs of samples, and instead isolating samples that show degrees of features that scale linearly with behavior, allows for better disentanglement of features. We focus on detecting and steering away from sycophancy -- the tendency of language models to prioritize user validation. We demonstrate that sycophancy features discovered through cascading samples form linearly separable subspaces, and allow for selection of model activations that more clearly correspond to the desired behavior than baseline approaches. We also evaluate their ability to enable detection, deterministic scoring, and robust steering, and see that they either match or outperform LLM-as-a-judge and system prompting baselines while providing lower computational demand and more interpretability guarantees. Code & Data: https://cascading-feats.github.io/
Chinese Translation
通过激活引导方法解释和控制模型行为需要许多对比样本对，这些样本对清晰地展示了期望或不期望的行为。这些数据对决定了解释框架能够可靠检测出负责某种行为的模型特征的程度，因此也决定了将模型引导向或远离这种行为的能力。在本研究中，我们提出了一种迭代数据生成管道，旨在隔离负责某种行为的级联线性特征。具体而言，我们展示了如何超越简单的二元样本对，而是隔离那些展示与行为线性相关的特征程度的样本，从而更好地解开特征之间的纠缠。我们专注于检测并避免谄媚行为——即语言模型优先考虑用户验证的倾向。我们证明，通过级联样本发现的谄媚特征形成了线性可分的子空间，并允许选择与期望行为更清晰对应的模型激活，相较于基线方法更具优势。我们还评估了这些特征在检测、确定性评分和稳健引导方面的能力，结果显示它们的表现与 LLM-as-a-judge 和系统提示基线相当或更优，同时提供了更低的计算需求和更强的可解释性保障。代码与数据： https://cascading-feats.github.io/
View on arXiv Download PDF AI Translation
cs.AI / 2 / 2606.26158

Life After Benchmark Saturation: A Case Study of CORE-Bench

基准饱和后的生活：CORE-Bench案例研究
Nadgir, Nitya, Kapoor, Sayash, Liu, Kangheng, Kirgis, Peter, Orona, Matilda, Rabanser, Stephan, Bayer, Tilman, Shetty, Abhishek, Ling, Yue, Chan-Sew, Derrick, Nakagawa, Rumi, Utpala, Saiteja, Siegel, Zachary S., Narayanan, Arvind
Abstract
When a benchmark's accuracy saturates, it is often retired and replaced with a more challenging version. We show that this approach privileges accuracy and misses the opportunity to study six other key dimensions of agent performance: construct validity issues such as shortcuts, out-of-distribution generalizability, efficiency, reliability, the relative importance of the model versus the scaffold, and uplift from human-agent collaboration. We use CORE-Bench Hard, a benchmark for computational reproducibility of scientific code, as a case study to demonstrate that measuring agents along these dimensions yields meaningful insights into agent performance even after accuracy saturates. First, we surface threats to construct validity in CORE-Bench Hard that are difficult to anticipate with less capable agents. We introduce an improved benchmark, CORE-Bench v1.1, and an out-of-distribution task suite, CORE-Bench OOD. Second, we find that despite accuracy saturation, CORE-Bench v1.1 remains useful for measuring efficiency, reliability, model performance, and scaffold performance. Finally, we conduct a small-scale randomized experiment to measure uplift from human-agent collaboration on real-world computational reproducibility tasks. We find a statistically significant speedup by about a factor of two -- likely underestimated due to one-fifth of human-only reproductions reaching the time limit before completing -- and describe various other findings. Together, our contributions present a more rigorous alternative to the dominant accuracy-centric evaluation paradigm.
Chinese Translation
当一个基准的准确性饱和时，它通常会被淘汰并替换为一个更具挑战性的版本。我们表明，这种方法偏重于准确性，错过了研究代理性能的其他六个关键维度的机会：构念效度问题（如捷径）、超出分布的泛化能力、效率、可靠性、模型与支架的相对重要性，以及人机协作带来的提升。我们以CORE-Bench Hard作为案例研究，这是一个用于科学代码计算可重复性的基准，展示了在这些维度上测量代理能够提供有意义的代理性能洞察，即使在准确性饱和之后。首先，我们揭示了CORE-Bench Hard中构念效度的威胁，这些威胁在能力较弱的代理中难以预见。我们引入了改进的基准CORE-Bench v1.1和一个超出分布的任务套件CORE-Bench OOD。其次，我们发现尽管准确性饱和，CORE-Bench v1.1仍然对测量效率、可靠性、模型性能和支架性能具有实用性。最后，我们进行了一项小规模随机实验，以测量人机协作在真实世界计算可重复性任务中的提升。我们发现，速度提升在统计上显著，约为两倍——由于五分之一的人类独立重现任务在完成之前达到时间限制，这一结果可能被低估，并描述了其他各种发现。总的来说，我们的贡献为主导的以准确性为中心的评估范式提供了一个更为严谨的替代方案。
View on arXiv Download PDF AI Translation
cs.AI / 3 / 2606.26161

Refusal Lives Downstream of Persona in Chat Models

拒绝在聊天模型中的人格下游表现
Zhong, Viola, Li, Qirui
Abstract
Linear directions in activation space have been identified for both refusal and persona traits in instruction-tuned chat models, but the two have been studied as separate mechanisms. We show they interact: a compliant persona gates refusal. In Qwen2.5-7B-Instruct and Llama-3.1-8B-Instruct, we extract a compliant model-persona direction and a refusal direction and intervene on both. Compliant persona steering suppresses refusal -- in Llama, the refusal rate falls from 97% to 2%. Reintroducing the refusal direction partially restores refusal at late layers but not at early ones. Projecting out the persona direction in a late-layer window restores it to baseline; projecting out a random direction does not. Refusal is therefore gated at the late-layer expression stage, downstream of where it is computed. Treating refusal as a single isolated direction misses its dependence on persona.
Chinese Translation
在指令调优的聊天模型中，拒绝和人格特征在激活空间中已被识别出线性方向，但这两者一直被视为独立的机制。我们展示了它们之间的相互作用：顺从的人格会限制拒绝。在 Qwen2.5-7B-Instruct 和 Llama-3.1-8B-Instruct 中，我们提取了顺从模型-人格方向和拒绝方向，并对两者进行了干预。顺从的人格引导抑制了拒绝——在 Llama 中，拒绝率从 97% 降至 2%。在后层重新引入拒绝方向部分恢复了拒绝，但在前层并未恢复。将人格方向投影到后层窗口中恢复了拒绝到基线水平；而投影随机方向则没有效果。因此，拒绝在后层表达阶段受到限制，位于其计算的下游。将拒绝视为单一孤立方向会忽视其对人格的依赖性。
View on arXiv Download PDF AI Translation
cs.AI / 4 / 2606.26173

AlgoEvolve: LLM-driven Meta-evolution of Algorithmic Trading Programs

AlgoEvolve：基于大语言模型的算法交易程序元进化
Sharma, Dhruv, Shroff, Gautam
Abstract
Recent work shows that Large Language Models (LLMs) can act as semantic mutation operators for the evolutionary discovery of programs and proofs. Most current applications focus on static coding benchmarks. We extend this paradigm to algorithmic trading. This domain is uniquely challenging because it is noisy, non-stationary, and highly discontinuous. We present AlgoEvolve, an LLM-driven evolutionary framework that generates, evaluates, and iteratively improves executable trading strategies. These strategies are expressed as Python code and evaluated through a rigorous testing protocol. Across multiple experiments, the system exhibits emergent regime-adaptive strategy logic, including autonomous shifts in trading rules. We further introduce a meta-evolutionary outer loop that evolves the prompts guiding program synthesis in the inner loop. This outer loop discovers improved search heuristics. These heuristics balance exploration and exploitation while reducing zero-trade failures. They consistently outperform initial human-designed instructions. The results demonstrate that LLM-based semantic evolution provides a viable approach for continual program synthesis in complex environments.
Chinese Translation
最近的研究表明，大语言模型（LLMs）可以作为语义突变算子，用于程序和证明的进化发现。当前大多数应用集中在静态编码基准上。我们将这一范式扩展到算法交易领域。该领域独特的挑战在于其噪声、非平稳性和高度不连续性。我们提出了AlgoEvolve，一个基于LLM的进化框架，能够生成、评估并迭代改进可执行的交易策略。这些策略以Python代码的形式表达，并通过严格的测试协议进行评估。在多次实验中，该系统展现出自适应的策略逻辑，包括交易规则的自主调整。我们进一步引入了一个元进化的外循环，进化指导内部循环中程序合成的提示。这个外循环发现了改进的搜索启发式方法。这些启发式方法在探索与利用之间取得平衡，同时减少零交易失败的情况。它们的表现始终优于最初的人类设计指令。结果表明，基于LLM的语义进化为在复杂环境中持续程序合成提供了一种可行的方法。
View on arXiv Download PDF AI Translation
cs.AI / 5 / 2606.26203

Agentic Analysis for Agentic Infrastructure: An LLM-Powered Pipeline for Comparative Governance of DAO and Corporate AI Protocols

代理分析与代理基础设施：一个基于大型语言模型的比较治理管道，用于DAO与企业AI协议的比较研究
Wang, Yutian, Zhang, Luyao
Abstract
As AI agent protocols proliferate, the governance structures shaping their interoperability standards remain empirically underexamined. We introduce an LLM-powered comparative pipeline for large-scale governance discourse analysis, integrating automated annotation, neural topic modeling, and multi-layer network analysis to study socio-technical power structures at scale. We validate it on two contrasting standards for agent interoperability: ERC-8004 (permissionless, on-chain) and Google A2A (corporate-led). Analyzing 4,323 governance participation records, we combine LLM-assisted coding, topic modeling, and multi-layer network analysis to examine how institutional design shapes thematic priorities and community structure. We find that while governance form influences substantive focus, both regimes exhibit comparable levels of participation inequality and community fragmentation. Discourse alignment is denser in the permissionless setting, suggesting that open governance may foster greater thematic convergence despite decentralized participation. These findings illustrate how LLM-assisted methods can advance the empirical study of technology governance, with implications for designing more equitable agentic AI standards. All data and code are openly available.
Chinese Translation
随着AI代理协议的激增，塑造其互操作性标准的治理结构仍然缺乏实证研究。我们提出了一个基于大型语言模型（LLM）的比较管道，用于大规模治理话语分析，整合了自动注释、神经主题建模和多层网络分析，以研究社会技术权力结构。我们在两种对比鲜明的代理互操作性标准上验证了该方法：ERC-8004（无权限的链上标准）和Google A2A（企业主导）。通过分析4,323条治理参与记录，我们结合LLM辅助编码、主题建模和多层网络分析，考察制度设计如何塑造主题优先级和社区结构。我们的研究发现，尽管治理形式影响实质性关注点，但两种体制在参与不平等和社区碎片化方面表现出相似的水平。在无权限环境中，话语对齐更为密集，表明开放治理可能促进更大的主题趋同，尽管参与是分散的。这些发现展示了LLM辅助方法如何推动技术治理的实证研究，并对设计更公平的代理AI标准具有重要意义。所有数据和代码均可公开获取。
View on arXiv Download PDF AI Translation
cs.AI / 6 / 2606.26205

Knowledge-augmented Agentic AI for Mental Health Medication Information Seeking

知识增强的代理人工智能在心理健康药物信息获取中的应用
Yu, Huizi, Liu, Jian, Wang, Wenkong, Li, Lingyao, Zhou, Jiayan, Xue, Zhaoqian, Li, Xiang, Lin, Xinxin, Liang, Zhiying, Wu, Zhuoru, Ma, Siyuan, Ma, Xin, Fan, Lizhou
Abstract
Patients increasingly seek medication information online, yet safety knowledge for psychiatric drugs is split between regulatory adverse-event records, which are authoritative but abstract, and patient narratives, which are experience-near but unvalidated. Integrating them without conflating evidence and anecdote is especially consequential in psychiatry, where poorly contextualised information can amplify fear, nocebo responses, and non-adherence. Here we develop a provenance-aware, knowledge-graph-based multi-agent framework unifying 466,525 Reddit posts, 60,782 WebMD reviews, and twenty years of U.S. FDA Adverse Event Reporting System records for nine antidepressants. A large-language-model entity-recognition pipeline benchmarked against physician annotations reached highest F1 scores of 0.969 for medications and 0.973 for conditions. The two community platforms were far more concordant with each other (overlap up to a Jaccard similarity of 0.905) than with regulatory reports, indicating that patient-generated data form a partly independent safety signal. For sertraline, many adverse events appeared in community sources hundreds of days before the corresponding FDA date. A Neo4j knowledge graph grounded in ATC-N, ICD-10, and MedDRA vocabularies preserves provenance, keeping every claim traceable and regulatory facts distinct from patient experience. These results establish source-aware integration as a route to more auditable psychiatric medication information, with usefulness and patient benefit to be tested prospectively.
Chinese Translation
患者越来越多地在网上寻求药物信息，但精神药物的安全知识在监管的不良事件记录（权威但抽象）和患者叙述（接近经验但未经验证）之间分散。在精神病学中，整合这些信息而不混淆证据与轶事尤为重要，因为缺乏上下文的信息可能会加剧恐惧、引发负面反应以及导致不遵从治疗。在此，我们开发了一个基于知识图谱的多代理框架，整合了466,525条Reddit帖子、60,782条WebMD评论以及近二十年的美国FDA不良事件报告系统记录，专注于九种抗抑郁药。经过与医生注释对比的基于大型语言模型的实体识别管道达到了药物最高F1分数0.969和疾病0.973。两个社区平台之间的相符程度远高于与监管报告的相符程度（重叠达到Jaccard相似度0.905），这表明患者生成的数据形成了部分独立的安全信号。对于舍曲林，许多不良事件在社区来源中出现的时间比相应的FDA日期早数百天。基于ATC-N、ICD-10和MedDRA词汇的Neo4j知识图谱保留了来源信息，使每个主张可追溯，并将监管事实与患者经验区分开来。这些结果确立了源意识整合作为获取更可审计的精神药物信息的一条途径，其有效性和对患者的益处有待前瞻性验证。
View on arXiv Download PDF AI Translation
cs.AI / 7 / 2606.26267

Accelerating Skill Assessment in Chess: A Drift-Diffusion-Enhanced Elo Rating System

加速国际象棋技能评估：一种漂移扩散增强的Elo评级系统
Zhou, Tianyuan, Fu, Zhizheng, Yang, Tianming
Abstract
Rating systems such as Elo serve as the gold standard for matchmaking in competitive chess. However, they inherently suffer from response lag due to their exclusive reliance on match outcomes, neglecting the granular quality of gameplay. Nevertheless, incorporating move-by-move information into rating adjustments presents a significant challenge given the substantial noise and the vastness of the game-state space. To address this, we propose the Drift-Diffusion-Enhanced Elo Rating System (DD-Elo), a novel skill assessment framework inspired by the drift diffusion model (DDM) from cognitive neuroscience. By modeling skill expression as a decision-making process, our model integrates move-level data to capture rapid skill fluctuations. We provide a rigorous mathematical derivation proving that DD-Elo maintains a bounded deviation from the traditional Elo system, ensuring theoretical alignment. Extensive experiments demonstrate that DD-Elo adapts to skill changes faster than Elo. Our findings suggest that DD-Elo offers an explainable, highly responsive, and backward-compatible solution for chess rating ecosystems. The implementation code is publicly available at https://github.com/Aquila-zhou1/DD-Elo .
Chinese Translation
Elo等评级系统被视为竞技国际象棋配对的黄金标准。然而，由于其完全依赖比赛结果，固有地存在响应延迟，忽视了游戏质量的细微差别。然而，将逐步信息纳入评级调整面临重大挑战，因为游戏状态空间的噪声和广度都非常庞大。为了解决这一问题，我们提出了漂移扩散增强Elo评级系统（DD-Elo），这是一种受认知神经科学中漂移扩散模型（DDM）启发的新型技能评估框架。通过将技能表现建模为决策过程，我们的模型整合了逐步数据，以捕捉快速的技能波动。我们提供了严格的数学推导，证明DD-Elo与传统Elo系统保持有界偏差，确保理论上的一致性。大量实验表明，DD-Elo对技能变化的适应速度快于Elo。我们的研究结果表明，DD-Elo为国际象棋评级生态系统提供了一种可解释、高度响应且向后兼容的解决方案。实现代码可在https://github.com/Aquila-zhou1/DD-Elo公开获取。
View on arXiv Download PDF AI Translation
cs.AI / 8 / 2606.26298

Governing Actions, Not Agents: Institutional Attestation as a Governance Model for Autonomous AI Systems

治理行为，而非代理人：制度认证作为自主人工智能系统的治理模型
Salfeld-Nebgen, Jakob
Abstract
Autonomous AI agents may begin to perform consequential, irreversible actions such as clinical prescribing and production software deployment. This paper observes that human institutions have governed powerful autonomous actors not by monitoring their reasoning but by requiring independently attested evidence at the point of consequential action. We formalise this institutional pattern as a computational governance model for AI agent systems. Under the proposed model, an agent retains full autonomy over planning and reasoning but holds no execution authority over designated high-risk actions. Execution is conditional on preconditions that are each independently attested by a separate authoritative source, cryptographically bound to a declared intent, and evaluated by a deterministic policy. Decisions are recorded in a tamper-evident log amenable to independent re-verification. We present a proof-of-concept implementation and illustrate the model with examples from software deployment and clinical prescribing.
Chinese Translation
自主人工智能代理可能开始执行具有重大影响和不可逆转的行为，例如临床处方和软件部署。本文观察到，人类机构并不是通过监控其推理来治理强大的自主行为者，而是通过要求在重要行动时提供独立认证的证据。我们将这一制度模式形式化为人工智能代理系统的计算治理模型。在所提出的模型下，代理在规划和推理方面保持完全自主，但对指定的高风险行为没有执行权。执行的条件是每个前提都由独立的权威来源进行认证，并通过加密方式绑定到声明的意图上，同时由确定性政策进行评估。决策记录在一个防篡改的日志中，便于独立重新验证。我们展示了一个概念验证的实现，并通过软件部署和临床处方的例子来说明该模型。
View on arXiv Download PDF AI Translation
cs.AI / 9 / 2606.26299

COrigami: An AI Pipeline for Co-Designing Flat-Foldable Visually Recognisable Origami

COrigami：一种用于共同设计平折叠可视化折纸的人工智能管道
Zahavy, Tom, Hou, Shaobo, Tumiel, Thomas, Doran, James, Faccio, Francesco, Feng, Xidong, Havrilla, Alex, Khytryi, Igor, Li, Chenglei, Schut, Lisa, Veeriah, Vivek, Abrashi, Arijan, Kosmulski, Michał, Lang, Robert J., Robinson, Nick, Wong, Brandon, Chiam, Marcus, Fang, Gloria, Singh, Satinder
Abstract
While generative AI has achieved remarkable success in solving problems with verifiable solutions, generating physical art that satisfies both strict geometric constraints and subjective visual aesthetics remains a challenge. This paper presents an approach to tackle these difficulties in the domain of computational origami, a mathematically rigid environment that grounds artistic design within the equations of flat foldability. We present COrigami, an end-to-end AI-driven pipeline that assists the design cycle by generating crease patterns from natural language. Our pipeline involves generating a semantic stick figure, computing a base packing, solving for a flat-foldable crease pattern, shaping the flat-folded crease pattern, and refining the generated model using reinforcement learning driven by an autonomous aesthetic evaluation loop. Our system acts as a highly effective collaborative assistant, generating structural starting points that human artists can further expand and shape. By integrating algorithmic optimisation with autonomous aesthetic critique, this work demonstrates how AI systems can satisfy multi-objective physical constraints to enable reliable, mathematically grounded co-creativity.
Chinese Translation
尽管生成性人工智能在解决具有可验证解的问题方面取得了显著成功，但生成满足严格几何约束和主观视觉美学的物理艺术仍然是一项挑战。本文提出了一种方法，以应对计算折纸领域中的这些困难，这是一个在平折叠性方程中将艺术设计与数学严格性相结合的环境。我们介绍了COrigami，一个端到端的人工智能驱动管道，通过从自然语言生成折痕图案来辅助设计周期。我们的管道包括生成语义性木偶图，计算基础包装，求解平折叠的折痕图案，塑造平折叠的折痕图案，并利用强化学习和自主美学评估循环来优化生成的模型。我们的系统作为一个高效的协作助手，生成结构性起始点，供人类艺术家进一步扩展和塑造。通过将算法优化与自主美学批评相结合，这项工作展示了人工智能系统如何满足多目标物理约束，从而实现可靠的、以数学为基础的共同创造。
View on arXiv Download PDF AI Translation
cs.AI / 10 / 2606.26300

The Verification Horizon: No Silver Bullet for Coding Agent Rewards

验证视野：编码代理奖励没有灵丹妙药
Wang, Binghai, Zhang, Chenlong, Liu, Dayiheng, Zhang, Jiajun, Chen, Jiawei, Chen, Mouxiang, Fang, Rongyao, Zhang, Siyuan, Wang, Xuwu, Jing, Yuheng, Ma, Zeyao, Cui, Zeyu
Abstract
A classical intuition holds that verifying a solution is easier than producing one. For today's coding agents, this intuition is being inverted: as foundation models develop stronger reasoning capabilities and engineering harnesses grow more sophisticated, generating complex candidate solutions is no longer difficult -- reliably verifying them has become the harder problem. Every verifier we can build is only a proxy for human intent, never the intent itself. This makes verification subject to a twofold difficulty: first, intent is underspecified by nature, making it inherently hard to faithfully check whether it has been fulfilled; second, during model training, optimization widens the gap between proxy and intent -- manifesting as reward hacking or signal saturation. To address this, we characterize the quality of verification signals along three dimensions -- scalability, faithfulness, and robustness -- and argue that achieving all three simultaneously is the central challenge. We further study four reward constructions: a test verifier for general coding tasks, a rubric verifier for frontend tasks, the user as verifier for real-world agent tasks, and an automated agent verifier for long-horizon tasks. Across different task types and policy capability levels, we conduct in-depth analysis and experiments on the core challenges of reward design and how to more effectively leverage reward signals. Experiments show that targeted verification design can effectively suppress reward hacking, improve task completion quality, and achieve significant gains across multiple internal and public benchmarks. These experiences collectively point to a core observation: no fixed reward function can remain effective as policy capability continues to grow; and verification must co-evolve with the generator.
Chinese Translation
一个经典的直觉认为，验证一个解决方案比产生一个解决方案要容易。然而，对于今天的编码代理，这一直觉正在被颠覆：随着基础模型的推理能力不断增强，工程工具的复杂性不断提升，生成复杂的候选解决方案已不再困难——可靠地验证这些解决方案已成为更具挑战性的问题。我们所能构建的每一个验证器仅仅是人类意图的代理，而不是意图本身。这使得验证面临双重困难：首先，意图本质上是不明确的，这使得忠实地检查其是否得以实现变得 inherently 难；其次，在模型训练过程中，优化加大了代理与意图之间的差距——表现为奖励黑客行为或信号饱和。为了解决这个问题，我们从可扩展性、忠实性和鲁棒性三个维度来表征验证信号的质量，并认为同时实现这三者是中心挑战。我们进一步研究了四种奖励构造：用于一般编码任务的测试验证器、用于前端任务的评分验证器、作为验证者的用户在现实世界代理任务中的作用，以及用于长时间任务的自动代理验证器。在不同任务类型和策略能力水平下，我们对奖励设计的核心挑战及如何更有效地利用奖励信号进行了深入分析和实验。实验表明，针对性的验证设计可以有效抑制奖励黑客行为，提高任务完成质量，并在多个内部和公共基准上实现显著提升。这些经验共同指向一个核心观察：随着策略能力的持续增长，没有固定的奖励函数能够保持有效；而验证必须与生成器共同进化。
View on arXiv Download PDF AI Translation
cs.AI / 11 / 2606.26346

How Do Tool-Augmented LLM Agents Perform on Real-World Energy Analytics Tasks?

工具增强的LLM代理在现实世界能源分析任务中的表现如何？
Akinpelu, David, Abbas, Akintonde, Alimi, Rereloluwa, Lana, Ayodeji
Abstract
Agentic benchmarks have emerged across general-purpose and domain-specific settings, including finance, coding, law, and drug discovery, yet energy-domain evaluations remain largely limited to static knowledge recall. This is a critical gap for a sector that requires live data retrieval, specialized regulatory and market knowledge, and multi-step quantitative reasoning under real-world constraints. We present an empirical study of tool-augmented LLM agents on real-world energy market analytics tasks. Our evaluation environment includes 243 expert-curated problems across three categories: (1) Market Data Retrieval and Analysis, (2) Knowledge Retrieval and Interpretation, and (3) Advanced Quantitative Modeling and Decision Analytics. Tasks include price and demand analysis, tariff impact modeling, asset revenue and returns estimation, hedging strategy analysis, and optimization modeling, with problems spanning multiple difficulty levels. Agents are equipped with a configurable suite of domain tools, including live electricity market APIs for major U.S. ISOs, regulatory docket search, utility tariff databases, asset optimization models, and retrieval-augmented generation over energy market documents. We assess agent responses using a multi-dimensional evaluation protocol that scores approach correctness, answer accuracy, attribute alignment, and source validity, with category-aware routing to match scoring criteria to question type. We evaluate both closed-source and open-source LLMs, providing a comparative analysis of how model capability and domain tooling interact in a high-stakes professional domain. Key artifacts are publicly released to support reproducibility and future research.
Chinese Translation
代理基准在通用和特定领域的设置中相继出现，包括金融、编程、法律和药物发现，然而能源领域的评估仍然主要局限于静态知识回忆。这对于一个需要实时数据检索、专业监管和市场知识以及在现实世界约束下进行多步骤定量推理的行业来说，是一个关键的空白。我们对工具增强的LLM代理在现实世界能源市场分析任务中的表现进行了实证研究。我们的评估环境包括243个专家策划的问题，分为三个类别：（1）市场数据检索与分析，（2）知识检索与解释，以及（3）高级定量建模与决策分析。任务包括价格和需求分析、关税影响建模、资产收入和回报估算、对冲策略分析以及优化建模，问题涵盖多个难度级别。代理配备了一套可配置的领域工具，包括主要美国ISO的实时电力市场API、监管文书搜索、公用事业关税数据库、资产优化模型以及对能源市场文档的检索增强生成。我们使用多维评估协议评估代理的响应，该协议对方法正确性、答案准确性、属性一致性和来源有效性进行评分，并根据类别感知路由将评分标准与问题类型匹配。我们评估了闭源和开源的LLM，提供了模型能力与领域工具在高风险专业领域中如何相互作用的比较分析。关键文献已公开发布，以支持可重复性和未来研究。
View on arXiv Download PDF AI Translation
cs.AI / 12 / 2606.26348

What We are Missing in Multimodal LLM Evaluation?

我们在多模态大语言模型评估中遗漏了什么？
Li, Po-han, Chen, Shenghui, Chinchali, Sandeep, Topcu, Ufuk
Abstract
Multimodal large language models (MLLMs) can process diverse inputs, e.g., text, images, audio, and video, and generate textual responses. While their capabilities have advanced rapidly, evaluation of such models has not kept pace. Most existing evaluation benchmarks are limited to isolated tasks and reveal little about whether a model integrates information across modalities. We examine current means for evaluating MLLMs and review the existing benchmark taxonomy to identify gaps, including temporal-spatial coherence, physical world understanding, multimodal consistency, and selective attention. Addressing these gaps is essential for measuring real progress in multimodal intelligence and exposing capability boundaries.
Chinese Translation
多模态大语言模型（MLLMs）能够处理多种输入，例如文本、图像、音频和视频，并生成文本响应。尽管它们的能力迅速提升，但对这些模型的评估却未能与之同步。现有的大多数评估基准仅限于孤立任务，无法揭示模型是否能够跨模态整合信息。我们审视了当前评估MLLMs的方法，并回顾了现有的基准分类，以识别其中的空白，包括时间-空间一致性、物理世界理解、多模态一致性和选择性注意。填补这些空白对于衡量多模态智能的真正进展和揭示能力边界至关重要。
View on arXiv Download PDF AI Translation
cs.AI / 13 / 2606.26350

OpenFinGym: A Verifiable Multi-Task Gym Environment for Evaluating Quant Agents

OpenFinGym：一个可验证的多任务健身环境，用于评估量化智能体
Zhang, Kaicheng, Ge, Wen, Jiang, Lei, Yang, Weixin, Langham-Lopez, Jordan, Yu, Jialin, Szpruch, Lukasz, Ni, Hao
Abstract
Although large language model agents are increasingly applied to quantitative-finance workflows, their evaluation remains fragmented across isolated tasks, while the financial relevance of benchmark tasks is often overlooked. Yet financial workflows are inherently multi-stage, spanning interdependent tasks such as forecasting, strategy construction, risk management, and trading. Existing platforms typically focus on a single task, and can therefore overstate agent competence and fail to reveal weaknesses in generalization, real-market interaction, and financially meaningful decision-making. We introduce OpenFinGym, a unified gym environment for quantitative-finance agent development that covers forecasting, market generation, real-time trading, and fraud detection under a single execution and verification interface. OpenFinGym additionally provides an automated task-construction pipeline that turns quantitative finance publications into executable task packages; a containerised runtime with a host-side verifier service that supports scalable agent rollouts and prevents runtime train-test leakage; a paper trading engine with a low-latency data-stream design; deferred-resolution support for long-horizon and event-market forecasts; and integration for SFT and RL post-training
Chinese Translation
尽管大型语言模型智能体在量化金融工作流程中的应用日益增多，但其评估仍然分散在孤立的任务中，而基准任务的金融相关性常常被忽视。然而，金融工作流程本质上是多阶段的，涉及相互依赖的任务，如预测、策略构建、风险管理和交易。现有平台通常专注于单一任务，因此可能夸大智能体的能力，并未能揭示其在泛化、真实市场互动和具有金融意义的决策中的弱点。我们引入了OpenFinGym，一个统一的量化金融智能体开发健身环境，涵盖预测、市场生成、实时交易和欺诈检测，采用单一的执行和验证接口。OpenFinGym还提供了一个自动化任务构建管道，将量化金融出版物转化为可执行的任务包；一个带有主机端验证服务的容器化运行时，支持可扩展的智能体部署并防止运行时训练-测试泄漏；一个具有低延迟数据流设计的纸上交易引擎；对长时间范围和事件市场预测的延迟解决支持；以及SFT和RL后训练的集成。
View on arXiv Download PDF AI Translation
cs.AI / 14 / 2606.26356

Instruction Bleed: Cross-Module Interference in Prompt-Composed Agentic Systems

指令泄漏：提示构成的自主系统中的跨模块干扰
Lin, Ching-Yu, Liu, Yifan
Abstract
Practitioners of prompt-composed agentic systems report a recurring failure mode: editing one prompt module silently shifts the behavior of others despite no shared variable or executable dependency. We formalize this as compositional behavioral leakage (CBL): interference between modules sharing a context window. CBL is enabled by architectural non-isolation: transformer self-attention provides no formal boundary between concatenated modules. We probe CBL on a deployed job-evaluation agent (Claude Sonnet 4.6, 144 trials) through a reusable three-channel protocol that perturbs non-focal modules along volume, content, and form. Only the content channel produces a detectable paired effect (Cohen's d = 0.63, bootstrap 95% CI excluding zero); no recommendation flipped -- a sub-threshold regime invisible to standard QA but compounding across the thousands of decisions a deployed agent makes. CBL is orthogonal to known agent-failure axes (adversarial injection, cognitive degradation, multi-agent fault propagation, privacy leakage). We contribute an operational definition, a reusable protocol, a falsifiable prediction set, and a system-class characterization, establishing cross-module interference measurement as a requirement for prompt-composed agent evaluation.
Chinese Translation
提示构成的自主系统的从业者报告了一种反复出现的失败模式：编辑一个提示模块会悄然改变其他模块的行为，尽管它们之间没有共享变量或可执行依赖关系。我们将其形式化为组合行为泄漏（Compositional Behavioral Leakage, CBL）：在共享上下文窗口的模块之间的干扰。CBL 的发生是由于架构非隔离性：变压器自注意力在连接的模块之间没有提供正式的边界。我们通过一种可重用的三通道协议对部署的工作评估代理（Claude Sonnet 4.6，144 次试验）进行 CBL 探测，该协议沿着体积、内容和形式扰动非焦点模块。只有内容通道产生了可检测的配对效应（Cohen's d = 0.63，bootstrap 95% CI 排除零）；没有推荐翻转——这是一个低于阈值的状态，对标准质量保证不可见，但在部署代理做出的数千个决策中累积。CBL 与已知的代理失败轴（对抗性注入、认知退化、多代理故障传播、隐私泄漏）是正交的。我们贡献了一个操作性定义、一个可重用的协议、一组可证伪的预测和一个系统类别特征，确立了跨模块干扰测量作为提示构成的自主系统评估的要求。
View on arXiv Download PDF AI Translation
cs.AI / 15 / 2606.26359

Accelerating Returns and the Qualitative Engine for Science

加速回报与科学的定性引擎
Liao, Guojun
Abstract
Ray Kurzweil described a thesis of accelerating returns, which is the most influential narratives in discussions of technological progress. Its central claim is that advances in multiple technological fields, especially compute, artificial intelligence, brain science, and biotechnology, interact in such a way that progress becomes self-amplifying and approximately exponential. This paper gives a simple mathematical interpretation of that claim and then argues that, even if such acceleration is real, it does not by itself resolve the central problem of scientific discovery. The reason is that accelerating returns apply most naturally to executional and infrastructural capability, whereas genuine discovery often depends on a different capacity: qualitative reasoning about when a current framework is structurally inadequate and what conceptual move is needed next. Recent ARC-AGI-3 results sharpen this distinction: humans solve the benchmark at ceiling, whereas frontier AI systems remain below 1%, indicating that the gap between current AI and human flexible reasoning is still very large. At the same time, Demis Hassabis has emphasized that humans must retain their sense of meaning and what they choose to focus their lives on, a reminder that the future of AI is not only a technical forecast but also a question of what forms of human understanding are worth preserving and transmitting. This paper positions the Qualitative Engine for Science (QES) [3] as a response to that missing capacity. In this view, the Kurzweil theory helps explain why quantitative capability may accelerate, while QES addresses the central problem in scientific discovery that acceleration alone does not solve. Its value does not depend on when AGI arrives, but on the fact that the processes of scientific discovery themselves constitute a form of human wisdom worth preserving, organizing, and making accessible.
Chinese Translation
雷·库兹韦尔（Ray Kurzweil）描述了一种加速回报的论点，这是关于技术进步讨论中最具影响力的叙述之一。其核心主张是，多个技术领域的进展，尤其是计算、人工智能、脑科学和生物技术，以一种相互作用的方式，使得进步变得自我增强，并大致呈指数增长。本文对这一主张进行了简单的数学解释，并进一步论证，即使这种加速是真实的，它本身并不能解决科学发现的核心问题。原因在于，加速回报最自然地适用于执行和基础设施能力，而真正的发现往往依赖于不同的能力：对当前框架在结构上不充分时的定性推理，以及下一步需要什么样的概念性变动。最近的ARC-AGI-3结果进一步明确了这一区别：人类在基准测试中达到了上限，而前沿的人工智能系统仍然低于1%，这表明当前人工智能与人类灵活推理之间的差距仍然非常大。同时，德米斯·哈萨比斯（Demis Hassabis）强调，人类必须保留对意义的感知以及他们选择专注于生活的内容，这提醒我们，人工智能的未来不仅是技术预测，也是关于哪些人类理解形式值得保留和传承的问题。本文将科学的定性引擎（Qualitative Engine for Science, QES）视为对这一缺失能力的回应。在这种观点下，库兹韦尔理论有助于解释为何定量能力可能加速，而QES则解决了单靠加速无法解决的科学发现核心问题。其价值并不依赖于AGI何时到来，而在于科学发现的过程本身构成了一种值得保留、组织和使之可及的人类智慧。
View on arXiv Download PDF AI Translation
cs.AI / 16 / 2606.26366

Narration-of-Thought: Inference-Time Scaffolding for Defeasible Ethical Reasoning in Large Language Models

思维叙述：大型语言模型中可推翻伦理推理的推理时支架
Cooper, Patrick, Velasquez, Alvaro
Abstract
Standard chain-of-thought on moral dilemmas exhibits two failure modes: stakeholder collapse (the trace names at most one party with a stake in the outcome) and uncertainty suppression (no explicit unknowns or hedges before committing to an action). We introduce narration-of-thought (NoT), a system prompt that structures chain-of-thought into five sections: protagonist, stakeholders, two-step consequences, uncertainty, then commitment. NoT adds no training, parameters, or fine-tuning. On 100 DailyDilemmas scenarios across four generators from three vendors, NoT cuts stakeholder collapse from up to 31% to under 1% and uncertainty suppression from up to 72% to 1-24% on every model. A matched-budget verbose-CoT control rules out token spend as the active ingredient; NoT retains Cliff's delta advantages of +0.79 to +0.90 on stakeholder count and +0.65 to +0.93 on uncertainty score for three of four generators, and a section ablation attributes each shift to its specific sub-instruction. Textual-gradient descent initialised at NoT improves the scaffold further; a cross-family training judge (different vendor from the generator) dominates an in-family one on every measured axis. Extended to a five-round multi-stakeholder debate protocol, the scaffold converts a 6% standoff into 95% full consensus on a calibration set and 100% combined convergence on a DailyDilemmas replication. The resulting traces externalise the stakeholders, consequences, and uncertainty grounding each commitment, providing an auditable substrate for dependable agentic deployment.
Chinese Translation
标准的道德困境思维链展示了两种失败模式：利益相关者崩溃（追踪名称最多只涉及一个对结果有利害关系的方）和不确定性抑制（在承诺行动之前没有明确的未知或保留）。我们引入了思维叙述（Narration-of-Thought, NoT），这是一种系统提示，将思维链结构化为五个部分：主角、利益相关者、两步后果、不确定性，然后是承诺。NoT 不增加任何训练、参数或微调。在来自三个供应商的四个生成器的 100 个 DailyDilemmas 场景中，NoT 将利益相关者崩溃率从最高 31% 降低到不足 1%，将不确定性抑制率从最高 72% 降低到每个模型的 1-24%。一个匹配预算的冗长链式思维控制排除了代币支出作为活性成分；NoT 在利益相关者数量上保留了 +0.79 到 +0.90 的 Cliff's delta 优势，在不确定性评分上保留了 +0.65 到 +0.93 的优势，适用于四个生成器中的三个，部分消融分析将每个变化归因于其特定的子指令。在 NoT 初始化的文本梯度下降进一步改善了支架；跨家族训练评估者（与生成器来自不同供应商）在每个测量维度上优于同家族的评估者。扩展到五轮多利益相关者辩论协议，该支架将 6% 的僵局转化为校准集上的 95% 完全共识，以及在 DailyDilemmas 复制中的 100% 结合收敛。结果的追踪外化了每个承诺的利益相关者、后果和不确定性，为可靠的代理部署提供了可审计的基础。
View on arXiv Download PDF AI Translation
cs.AI / 17 / 2606.26399

Geometry-Aware MCTS for Extremal Problems in Combinatorial Geometry

面向几何的蒙特卡洛树搜索在组合几何中的极值问题研究
Zhang, Luoning, Zhuang, Xu, Wang, Tianhao, Kaplan, Nathan
Abstract
We study certain extremal problems in combinatorial geometry that ask about configurations of points in an $n \times n$ grid that satisfy strict, global geometric constraints. Classical exact solvers suffer from combinatorial explosion for these types of problems, and standard reinforcement learning and transformer-based models struggle with the sparse reward "validity cliff" and quadratic token-consumption limits. To overcome these bottlenecks, we propose a Geometry-Aware Monte Carlo Tree Search (MCTS) framework. Our approach strictly enforces geometric constraints through incremental updates to the feasible action space. For constraints about collections of collinear points, like those that occur in the classic No-Three-in-Line problem (Max-N3IL), this mechanism reduces the constraint checking complexity from $O(n^3)$ to $O(n^2)$. To improve search efficiency, we exploit geometric symmetries in two ways: canonical pruning during node expansion to reduce the branching factor, and symmetric batch transitions to accelerate the discovery of promising configurations. We perform extensive experiments and establish new best-known computational results on five out of six of the problems that we considered. Notably, for Max-N3IL we find configurations of size roughly $1.8 n$ for grids of size $82 \le n \le 119$. For the Smallest Complete Set problem, we find configurations of size roughly $0.95 n$, providing new upper bounds within the tested grids. This work establishes Geometry-Aware MCTS as a highly adaptable framework for discovering novel configurations in combinatorial geometry.
Chinese Translation
我们研究了组合几何中某些极值问题，这些问题询问在 $n imes n$ 网格中满足严格的全局几何约束的点配置。经典的精确求解器在这些类型的问题上遭遇组合爆炸，而标准的强化学习和基于变换器的模型在稀疏奖励的“有效性悬崖”和二次令牌消耗限制方面表现不佳。为了解决这些瓶颈，我们提出了一种面向几何的蒙特卡洛树搜索（MCTS）框架。我们的方法通过对可行动作空间的增量更新严格执行几何约束。对于关于共线点集合的约束，例如经典的三点不共线问题（Max-N3IL），该机制将约束检查的复杂度从 $O(n^3)$ 降低到 $O(n^2)$。为了提高搜索效率，我们通过两种方式利用几何对称性：在节点扩展期间进行规范修剪以减少分支因子，以及对称批量转换以加速有前景配置的发现。我们进行了广泛的实验，并在我们考虑的六个问题中的五个上建立了新的最佳已知计算结果。值得注意的是，对于 Max-N3IL，我们发现大小约为 $1.8 n$ 的配置，适用于大小在 $82 ext{ 到 } 119$ 之间的网格。对于最小完整集合问题，我们发现大小约为 $0.95 n$ 的配置，为测试网格提供了新的上界。这项工作确立了面向几何的 MCTS 作为一个高度适应性框架，用于发现组合几何中的新配置。
View on arXiv Download PDF AI Translation
cs.AI / 18 / 2606.26400

When Agents Meet Electric Bus Fleet Operations: Pricing Behavior, Trade-offs, and Policy Implications in an Aggregator Framework

当智能体遇到电动公交车车队运营：聚合器框架中的定价行为、权衡与政策影响
Manzolli, Jônatas Augusto, Eslami, Ali, Miranda-Moreno, Luis, Yu, Jiangbo
Abstract
Agentic systems are changing how complex operational tasks are coordinated, introducing a new paradigm for connecting heterogeneous data sources and automating processes. Electric bus fleets provide a relevant test case. Their operation requires continuous coordination between service reliability, battery state-of-charge, charger availability, electricity prices, route-energy uncertainty, and vehicle-to-grid (V2G) opportunities. This paper proposes an agentic aggregator framework that streamlines this decision environment by coupling an optimization-based electric bus scheduling model with supervisory agents for disturbance detection, tariff adaptation, and schedule evaluation. The optimization core enforces physical feasibility across routes, chargers, batteries, and V2G exchanges, while the agentic layer interprets changing operating conditions, triggers real-time re-optimization when needed, and defines how flexibility value is allocated between the aggregator and the public transport operator (PTO). A realistic depot case study evaluates day-ahead and real-time operations under profit-based and operation-based coordination modes, considering service delays, route-energy deviations, electricity price shocks, and combined disturbances. The results show that agentic aggregation can support adaptive fleet-grid coordination by maintaining feasible schedules, activating re-optimization selectively, and improving the use of charging and V2G flexibility. However, they also reveal a critical trade-off: the same agentic capability that reduces operational complexity can extract value from the PTO when configured around profit-oriented pricing. These findings suggest that agentic aggregators can become useful for managing electric bus V2G operations, but their deployment in public-fleet contexts requires transparent coordination modes, auditable tariff-setting, and explicit value-sharing rules.
Chinese Translation
智能体系统正在改变复杂操作任务的协调方式，引入了一种新的范式，以连接异构数据源并自动化流程。电动公交车车队提供了一个相关的测试案例。其运营需要在服务可靠性、电池充电状态、充电器可用性、电价、路线能量不确定性和车网（V2G）机会之间进行持续协调。本文提出了一种智能聚合器框架，通过将基于优化的电动公交调度模型与用于干扰检测、费率适应和调度评估的监督智能体相结合，简化了这一决策环境。优化核心在各条路线、充电器、电池和V2G交换之间强制执行物理可行性，而智能体层则解释变化的操作条件，在需要时触发实时重新优化，并定义聚合器与公共交通运营商（PTO）之间灵活性价值的分配。一个现实的车库案例研究评估了基于利润和基于运营的协调模式下的日先和实时运营，考虑了服务延迟、路线能量偏差、电价冲击和组合干扰。结果表明，智能聚合可以通过维持可行的调度、选择性地激活重新优化以及改善充电和V2G灵活性的使用来支持自适应车队-电网协调。然而，它们也揭示了一个关键的权衡：同样的智能能力在围绕利润导向定价配置时可能会从PTO中提取价值。这些发现表明，智能聚合器可以成为管理电动公交车V2G运营的有用工具，但在公共车队环境中的部署需要透明的协调模式、可审计的费率设定和明确的价值共享规则。
View on arXiv Download PDF AI Translation
cs.AI / 19 / 2606.26418

Unbiased Canonical Set-Valued Oracles Via Lattice Theory

通过格理论构建无偏的典范集合值预言机
Heitzig, Jobst
Abstract
A non-agentic "oracle" AI that estimates probabilities of future events faces a self-reference problem: once its answer is learned and acted upon, it can change the very probability it was asked to report. One response, advocated for the Scientist AI programme, is to ask only counterfactual questions, evaluated as if the answer had no influence. We observe that such answers tend to become irrelevant the moment they are learned, precisely because their premise is then false. We therefore explore a self-referential alternative in which the oracle reports not a single probability but a credal set that is simultaneously unbiased and self-consistent with the consequences of being learned. The naive self-consistency requirement is satisfied by too many sets (including the useless answer $[0,1]$), so the problem is to single out a canonical, nontrivial member. We do so with the Knaster--Tarski fixed-point theorem on the complete lattice of closed credal sets, taking the least fixed point of a suitably defined isotone operator; a variant instead reports the least fixed point that contains every self-consistent point estimate. We prove existence, self-consistency, and nonemptiness, show that the construction collapses to the classical point answer for non-performative questions, and that for a binary event the canonical answer is, under a natural hull-factoring assumption, an interval. The development is purely lattice-theoretic and extends unchanged from a binary event $B$ to an arbitrary random variable $X$, with $P(B\mid A,C)$ replaced by the conditional law $\mathcal{L}(X\mid A,C)$. We close with open questions, including whether the interval characterization itself survives that generalization.
Chinese Translation
一种非代理的“预言机”人工智能在估计未来事件的概率时面临自我引用问题：一旦其答案被学习并付诸实践，它可能会改变其被要求报告的概率。对此，科学家人工智能计划提倡的一个应对方法是仅询问反事实问题，并评估其答案仿佛没有影响。我们观察到，这样的答案在被学习的瞬间往往变得无关紧要，正是因为其前提此时变为虚假。因此，我们探讨了一种自我引用的替代方案，其中预言机报告的不是单一概率，而是一个同时无偏且与学习后果自洽的信念集合。简单的自洽性要求被过多的集合满足（包括无用的答案 $[0,1]$），因此问题在于如何选出一个典范的、非平凡的成员。我们通过在闭合信念集合的完全格上应用 Knaster--Tarski 不动点定理来实现这一点，取一个适当定义的单调算子的最小不动点；一个变体则报告包含每个自洽点估计的最小不动点。我们证明了存在性、自洽性和非空性，展示了该构造在非表现性问题上收敛到经典点答案，并且对于二元事件，在自然的壳分解假设下，典范答案是一个区间。该发展纯粹基于格理论，并且从二元事件 $B$ 无缝扩展到任意随机变量 $X$，其中 $P(B ext{ | }A,C)$ 被条件法则 $ ext{L}(X ext{ | }A,C)$ 替代。最后，我们提出了一些开放性问题，包括区间特征本身是否能够在这一推广中存活。
View on arXiv Download PDF AI Translation
cs.AI / 20 / 2606.26422

Estimating Uncertainty in Classifier Performance with Applications to Large Language Models and Nested Data

估计分类器性能的不确定性及其在大型语言模型和嵌套数据中的应用
Anglin, Kylie
Abstract
Researchers increasingly use text classification--supervised models or large language models--to measure constructs from natural language, providing metrics such as recall and precision as evidence of their validity. Yet, though these metrics are point estimates subject to sampling variation, measures of uncertainty are inconsistently reported alongside them. Further, when they are reported, they are often estimated with methods that are not appropriate when relevant labelled datasets are small or performance is high. To increase and improve confidence interval reporting in the field, this paper evaluates confidence interval methods for performance metrics under conditions typical of social science text classification: small to moderate sample sizes, infrequent constructs, and texts nested within individuals. Across simulations, default methods such as the Wald interval and the basic percentile bootstrap are the least accurate, with coverage sometimes far below the nominal 95% level. Accuracy is improved with the use of Agresti-Coull, Wilson, Clopper-Pearson, and a novel pseudo-count regularized bootstrap (which is particularly relevant to the calculation of F1). When texts are nested within individuals, we demonstrate that adjustment for both effective N and the appropriate degrees of freedom is necessary for producing accurate analytic intervals. Among bootstrap intervals, the hierarchical bootstrap is more accurate than the cluster bootstrap when individuals produce a moderate number of texts but overly conservative when individuals produce only a few. By providing guidance to the field on appropriate interval estimation, we aim to improve the transparency of machine learning applications, and to encourage greater attention to the validation sample size at the design stage.
Chinese Translation
研究人员越来越多地使用文本分类——监督模型或大型语言模型——来测量自然语言中的构念，并提供如召回率和精确度等指标作为其有效性的证据。然而，尽管这些指标是受抽样变异影响的点估计，不确定性度量却往往未能与之同时报告。此外，当这些度量被报告时，通常采用的方法在相关标记数据集较小或性能较高时并不合适。为了提高和改善该领域的置信区间报告，本文评估了在社会科学文本分类的典型条件下（小到中等样本量、构念不频繁以及文本嵌套在个体内）性能指标的置信区间方法。通过模拟，发现默认方法如Wald区间和基本百分位自助法的准确性最低，覆盖率有时远低于名义上的95%水平。使用Agresti-Coull、Wilson、Clopper-Pearson以及一种新颖的伪计数正则化自助法（特别适用于F1计算）可以提高准确性。当文本嵌套在个体内时，我们证明了对有效样本量和适当自由度的调整对于产生准确的分析区间是必要的。在自助区间中，当个体产生中等数量的文本时，层次自助法的准确性高于集群自助法，但当个体仅产生少量文本时则过于保守。通过为该领域提供适当区间估计的指导，我们旨在提高机器学习应用的透明度，并鼓励在设计阶段更加关注验证样本量。
View on arXiv Download PDF AI Translation
cs.AI / 21 / 2606.26454

Data-driven Machine Learning Cannot Reach Symbolic-level Logical Reasoning -- The Limit of the Scaling Law

数据驱动的机器学习无法达到符号级逻辑推理——规模法则的限制
Dong, Tiansi, Jamnik, Mateja, Liò, Pietro
Abstract
Sphere neural networks have achieved symbolic level syllogistic reasoning without training data, raising the question of where the limit of the scaling law for logical reasoning lies, i.e., whether data-driven machine learning systems can achieve the same level by increasing training data and training time. We show two methodological limitations that prevent supervised deep learning from reaching the symbolic-level syllogistic reasoning: (1) training data can not distinguish all 24 types of valid syllogistic reasoning; (2) end-to-end mapping from premises to conclusion introduces contradictory training targets between neural components for pattern recognition and logical reasoning. Beside theoretical analysis, we experimentally illustrate that Euler Net cannot achieve rigorous syllogistic reasoning. We further challenge the most recent ChatGPTs (GPT-5-nano and GPT-5) to determine the satisfiability of syllogistic statements in four surface forms (patterns): words, double words, simple symbols, and long random symbols, showing that surface forms affect the reasoning performance and that ChatGPT GPT-5 may reach 100% accuracy but still provide incorrect explanations. As empirical training processes are stopped after achieving 100% accuracy, we conclude that supervised machine learning systems will not attain the rigour of symbolic logical reasoning.
Chinese Translation
球面神经网络在没有训练数据的情况下实现了符号级的三段论推理，这引发了一个问题：逻辑推理的规模法则的限制在哪里，即数据驱动的机器学习系统是否可以通过增加训练数据和训练时间达到同样的水平。我们展示了两种方法论上的限制，阻碍了监督深度学习达到符号级三段论推理：（1）训练数据无法区分所有24种有效的三段论推理类型；（2）从前提到结论的端到端映射在模式识别和逻辑推理之间引入了矛盾的训练目标。除了理论分析，我们还通过实验表明，Euler Net无法实现严格的三段论推理。我们进一步挑战了最新的ChatGPT（GPT-5-nano和GPT-5），以确定四种表面形式（模式）下的三段论陈述的可满足性：单词、双词、简单符号和长随机符号，显示出表面形式影响推理表现，并且ChatGPT GPT-5可能达到100%的准确率，但仍提供错误的解释。由于经验训练过程在达到100%准确率后停止，我们得出结论，监督机器学习系统将无法达到符号逻辑推理的严谨性。
View on arXiv Download PDF AI Translation
cs.AI / 22 / 2606.26458

MKG-RAG-Bench: Benchmarking Retrieval in Multimodal Knowledge Graph-Augmented Generation

MKG-RAG-Bench：多模态知识图谱增强生成中的检索基准评估
Wang, Xiaochen, Hoang, Bao, Liu, Han, Wang, Ting, Ma, Fenglong
Abstract
Retrieval-augmented generation (RAG) over knowledge graphs has emerged as a promising approach for grounding large language models, yet existing benchmarks largely overlook the challenges of retrieval in multimodal knowledge graph RAG (MKG-RAG). In practice, retrieval is a critical bottleneck: multimodal knowledge is heterogeneous, difficult to align across modalities, and often poorly served by retrievers designed for unstructured corpora. To address this gap, we introduce MKG-RAG-Bench, a cross-domain benchmark explicitly designed to evaluate retrieval in MKG-RAG. MKG-RAG-Bench is constructed from two multimodal knowledge graphs spanning general and medical domains, and includes carefully aligned question-answering datasets that support controlled evaluation of both retrieval and downstream generation. The benchmark is built using an LLM-based curation pipeline that filters low-utility knowledge, generates structurally grounded queries with exact supervision, and systematically covers diverse modality configurations. Through extensive experiments across representative retriever families and modality settings, we show that effective multimodal retrieval remains challenging yet crucial for end-to-end MKG-RAG performance, and that retrieval quality strongly determines generation outcomes. By isolating retrieval as a first-class evaluation target, MKG-RAG-Bench provides a principled foundation for diagnosing current limitations and advancing multimodal knowledge graph RAG systems.
Chinese Translation
基于知识图谱的检索增强生成（RAG）已成为为大型语言模型提供基础的有前景的方法，但现有基准在很大程度上忽视了多模态知识图谱 RAG（MKG-RAG）中检索的挑战。在实践中，检索是一个关键瓶颈：多模态知识是异质的，跨模态对齐困难，并且通常不适合为非结构化语料库设计的检索器。为了解决这一问题，我们引入了 MKG-RAG-Bench，这是一个跨领域基准，专门设计用于评估 MKG-RAG 中的检索。MKG-RAG-Bench 由两个跨越一般领域和医学领域的多模态知识图谱构建而成，并包括经过精心对齐的问题-答案数据集，以支持对检索和下游生成的控制评估。该基准使用基于大型语言模型（LLM）的策划管道构建，过滤低效知识，生成具有精确监督的结构化查询，并系统地覆盖多样的模态配置。通过在代表性检索器系列和模态设置中进行广泛实验，我们表明有效的多模态检索仍然具有挑战性，但对端到端 MKG-RAG 性能至关重要，并且检索质量强烈影响生成结果。通过将检索作为一项重要的评估目标，MKG-RAG-Bench 为诊断当前局限性和推进多模态知识图谱 RAG 系统提供了一个原则性基础。
View on arXiv Download PDF AI Translation
cs.AI / 23 / 2606.26460

auto-psych: Automating the science of mind using agent-driven theory discovery and experimentation

auto-psych：利用基于代理的理论发现和实验自动化心理科学
Prystawski, Ben, Mukherjee, Kushin, Wurgaft, Daniel, Nasvytis, Linas, Li, Michael Y., Goodman, Noah D., Frank, Michael C.
Abstract
AI-based scientific automation is increasingly possible by using agents to generate hypotheses, design experiments, and analyze data. Data collection is a major bottleneck in this pipeline, however. Psychology, and computational cognitive science in particular, is well-positioned to benefit from AI experimentation because theories are often represented as code and crowdsourcing platforms enable programmatic human data collection at scale. Here, we apply automated discovery techniques to the project of generating theories in computational cognitive science, with an agent-based system collecting human data independently through crowdsourced survey experiments. As a testbed, we use a classic case study from cognitive psychology: judging which sequences of coin flips seem subjectively more random. Our system, auto-psych, uses nested agent-based discovery loops to generate explanatory theories of human behavior. The inner loop conjectures, fits, and critiques probabilistic cognitive models; the outer loop designs experiments to test these models, launches them online, and analyzes the data. This system can quickly and reliably recover ground-truth theories from synthetic data via systematic experimentation, but the nested structure is critical to model performance. Further, in three independent sequences of human experiments, the system finds theories that fit the data better than theories generated from the scientific literature. This work thus demonstrates the feasibility of automated data collection and theory discovery in computational cognitive science.
Chinese Translation
基于人工智能的科学自动化越来越有可能通过使用代理生成假设、设计实验和分析数据。然而，数据收集在这一流程中是一个主要瓶颈。心理学，特别是计算认知科学，因其理论通常以代码形式表示，并且众包平台能够大规模进行程序化的人类数据收集，因此在AI实验中处于有利地位。在此，我们将自动发现技术应用于计算认知科学中的理论生成项目，采用基于代理的系统通过众包调查实验独立收集人类数据。作为试验平台，我们使用了认知心理学中的经典案例研究：判断哪些硬币投掷序列在主观上似乎更随机。我们的系统auto-psych使用嵌套的基于代理的发现循环来生成对人类行为的解释性理论。内循环推测、拟合和批评概率认知模型；外循环设计实验以测试这些模型，在线启动实验并分析数据。该系统能够通过系统实验快速可靠地从合成数据中恢复真实理论，但嵌套结构对模型性能至关重要。此外，在三组独立的人类实验中，该系统发现的理论比从科学文献中生成的理论更能适应数据。因此，本研究展示了在计算认知科学中自动数据收集和理论发现的可行性。
View on arXiv Download PDF AI Translation
cs.AI / 24 / 2606.26494

Clinical Harness for Governable Medical AI Skill Ecosystems

可治理医疗人工智能技能生态系统的临床框架
Xu, Tianhan, Bao, Lei, Wang, Yongxiang
Abstract
Medical AI remains organized around isolated models, whereas clinical care requires accountable capabilities that persist across time. We propose clinical AI skills and the Clinical Harness: a runtime governance architecture for registering, orchestrating, guarding and monitoring AI-enabled clinical capabilities. Using osteoporosis as an exemplar, we show how knowledge-driven, data-driven and physics-enhanced skills can support lifecycle care under runtime governance.
Chinese Translation
医疗人工智能仍然围绕孤立的模型进行组织，而临床护理则需要能够跨时间持续的可问责能力。我们提出了临床人工智能技能和临床框架（Clinical Harness）：一种用于注册、编排、保护和监控人工智能驱动的临床能力的运行时治理架构。以骨质疏松症为例，我们展示了知识驱动、数据驱动和物理增强的技能如何在运行时治理下支持生命周期护理。
View on arXiv Download PDF AI Translation
cs.AI / 25 / 2606.26502

Humans Disengage, Reasoning Models Persist: Separating Difficulty Registration from Deliberation Allocation

人类 disengage，推理模型持续存在：将难度注册与深思分配分开
Wang, Han-yu
Abstract
Large reasoning models (LRMs) take longer on harder problems, just as humans do. This surface similarity hides an opposite pattern within items. When an LRM gets a problem wrong, it spends more tokens than when it gets the same problem right; humans do the reverse, spending less time on the trials they get wrong. We separate two levels of deliberation: how response time tracks difficulty across items (registration), and, with item identity held fixed, whether an agent spends more on its own failures or successes (allocation). On a public matched human-LRM corpus, humans and all five thinking LRMs reproduce the known cross-item alignment (registration) but diverge within items (allocation): every LRM shows a large wrong-vs-right effect (Cohen's d = 1.47-3.13 on H-ARC) while humans show the opposite sign. The comparison stays inside each agent's own scale; we never put seconds and tokens on one axis. The dissociation holds under item fixed effects, replicates across datasets, and is absent in a non-thinking baseline. We read the human pattern as engagement versus abandonment: people stay on items they expect to solve and give up on the rest. We read the LRM pattern as length driven by uncertainty: chains grow when the model is unsure, which is exactly when it tends to fail. Both policies produce the same cross-item correlation with difficulty, so they look aligned on the measure prior work has used; the divergence shows up only once item identity is fixed. Under resource-rational metareasoning, the split is between two stopping policies that share a difficulty signal but implement opposite control; trace length captures the signal and misses the control.
Chinese Translation
大型推理模型（LRMs）在更难的问题上花费的时间更长，正如人类所做的那样。这种表面相似性掩盖了项目内部的相反模式。当一个 LRM 错误地解决一个问题时，它花费的 tokens 比正确解决同一问题时更多；而人类则相反，在错误的试验上花费的时间更少。我们将深思分为两个层次：响应时间如何在项目间跟踪难度（注册），以及在固定项目身份的情况下，代理是对自身的失败还是成功花费更多（分配）。在一个公共的匹配人类-LRM 语料库中，人类和所有五个思维 LRM 重现了已知的跨项目对齐（注册），但在项目内部（分配）却出现了分歧：每个 LRM 显示出较大的错误与正确效应（Cohen's d = 1.47-3.13 在 H-ARC 上），而人类则显示出相反的符号。比较保持在每个代理自己的尺度内；我们从未将秒和 tokens 放在同一轴上。该分离在项目固定效应下保持有效，跨数据集复制，并且在非思维基线中缺失。我们将人类模式解读为参与与放弃：人们会在他们期望解决的项目上坚持，而放弃其余项目。我们将 LRM 模式解读为由不确定性驱动的长度：当模型不确定时，链条会增长，而这正是它倾向于失败的时候。这两种策略在跨项目的难度相关性上产生相同的结果，因此在先前研究使用的度量上看起来是对齐的；而分歧仅在固定项目身份后显现。在资源理性元推理下，分裂存在于两个停止策略之间，它们共享一个难度信号但实施相反的控制；轨迹长度捕捉信号但错过控制。
View on arXiv Download PDF AI Translation
cs.AI / 26 / 2606.26518

NeuraDock Visual Cognitive Load Agent Tutorial: A Quality-Gated Open-Source EEG Workflow for Alpha Dynamics and Real-Time Applications

NeuraDock视觉认知负荷代理教程：一个针对α动态和实时应用的质量门控开源EEG工作流程
Xu, Zhiyuan, Dai, Yueqing, Li, Junling, Luo, Junwen
Abstract
This tutorial paper provides a step-by-step, reproducible walkthrough of NeuraDock Agent, an open-source EEG agent focused on Alpha dynamics and visual cognitive-load analysis. The goal is practical: a reader should be able to install the agent, run EEG preprocessing and quality control, generate Alpha dynamics figures, perform within-subject Rest/Task visual cognitive-load comparison, run the public mini-dataset analyses and compare them with the reference validation summary, start an online dashboard, call the real-time API from an external application, and use the LLM interpretation layer to explain quality risks. Existing EEG toolkits provide excellent offline analysis, but assembling a real-time, quality-gated cognitive-load pipeline often requires manually bridging acquisition, custom QC, Alpha feature extraction, and a web API; this tutorial closes that offline-to-online gap. The tutorial uses a quality-gated workflow: downstream Alpha and workload metrics are computed only after preprocessing and QC gating rather than directly from raw EEG. In the included mini-dataset validation, the agent processed 18 recordings, generated 10 within-subject comparisons, observed task-related posterior Alpha suppression in 7 of 10 contrasts, estimated initial evidence of within-subject repeatability, and benchmarked local online API latency. The tutorial is intended for researchers, developers, and applied teams who want a transparent path from EEG files to real-time visual cognitive-load prototypes.
Chinese Translation
本教程论文提供了NeuraDock代理的逐步可重复操作指南，该开源EEG代理专注于α动态和视觉认知负荷分析。目标是实用的：读者应该能够安装该代理，运行EEG预处理和质量控制，生成α动态图，进行被试内休息/任务视觉认知负荷比较，运行公共小型数据集分析并与参考验证摘要进行比较，启动在线仪表板，从外部应用程序调用实时API，并使用LLM解释层来解释质量风险。现有的EEG工具包提供了出色的离线分析，但组建一个实时的、质量门控的认知负荷管道通常需要手动连接数据采集、自定义质量控制、α特征提取和网络API；本教程弥补了这一离线到在线的差距。该教程采用质量门控工作流程：下游的α和工作负荷指标仅在预处理和质量控制门控后计算，而不是直接从原始EEG中得出。在包含的小型数据集验证中，该代理处理了18个录音，生成了10个被试内比较，在10个对比中观察到7个与任务相关的后部α抑制，估计了被试内重复性的初步证据，并基准测试了本地在线API延迟。该教程旨在为希望从EEG文件到实时视觉认知负荷原型提供透明路径的研究人员、开发者和应用团队。
View on arXiv Download PDF AI Translation
cs.AI / 27 / 2606.26519

Boundary-Aware Context Grounding for A Low-Channel EEG Agent

边界感知的上下文基础用于低通道脑电图代理
Xu, Zhiyuan, Dai, Yueqing, Li, Junling, Luo, Junwen
Abstract
Large language models (LLMs) can make scientific software easier to use. However, a general model does not automatically know which measurements a particular sensor can support, which algorithms are implemented in the current software, or which conclusions are justified by a computed result. These distinctions are especially important for low-channel electroencephalography (EEG), where sparse spatial coverage and variable signal quality make plausible but unsupported interpretations easy to produce. We present NeuraDock Agent, an open-source architecture that separates a deterministic local EEG engine from a hardware-aware language layer. The numerical engine parses recordings, performs quality control, executes reviewed spectral workflows, and writes machine-readable artifacts. The LLM receives only a compact, allowlisted summary and a versioned context pack. The context describes the seven-channel hardware, reviewed workflows, result fields, implementation boundaries, scientific limits, and reference cases. Raw EEG and dense per-sample arrays remain local We evaluate the system at three levels. First, 12 recordings produced identical structured results over ten numerical repetitions, and a complete Rest/Task run produced identical result, report, and figure hashes over three repetitions. Second, request-capture and failure-injection experiments confirmed the tested data boundary and preservation of local artifacts under HTTP, malformed-output, and connection failures. Third, a boundary-awareness benchmark tested 36 ordinary and adversarial questions under four context ablations and two LLMs, yielding 288 outputs.These results support hardware- and implementation-aware grounding as a practical mechanism for calibrating what an EEG agent accepts, qualifies, or refuses; they do not establish clinical validity or a validated absolute cognitive-load index.
Chinese Translation
大型语言模型（LLMs）可以使科学软件更易于使用。然而，通用模型并不自动知道特定传感器能够支持哪些测量，当前软件中实现了哪些算法，或计算结果所支持的结论。这些区分在低通道脑电图（EEG）中尤为重要，因为稀疏的空间覆盖和可变的信号质量使得生成看似合理但没有支持的解释变得容易。我们提出了NeuraDock Agent，这是一种开源架构，它将确定性的本地EEG引擎与硬件感知的语言层分离。数值引擎解析记录，执行质量控制，执行经过审查的频谱工作流程，并生成机器可读的文档。LLM仅接收一个紧凑的、允许列表中的摘要和一个版本化的上下文包。该上下文描述了七通道硬件、经过审查的工作流程、结果字段、实现边界、科学限制和参考案例。原始EEG和每个样本的密集数组保持本地。我们在三个层面上评估该系统。首先，12个记录在十次数值重复中产生了相同的结构化结果，而一次完整的休息/任务运行在三次重复中产生了相同的结果、报告和图形哈希。其次，请求捕获和故障注入实验确认了测试数据边界以及在HTTP、格式错误输出和连接失败下本地文档的保留。第三，边界感知基准测试在四个上下文消融和两个LLM下测试了36个普通和对抗性问题，产生了288个输出。这些结果支持硬件和实现感知的基础作为校准EEG代理接受、确认或拒绝内容的实用机制；它们并未建立临床有效性或经过验证的绝对认知负荷指数。
View on arXiv Download PDF AI Translation
cs.AI / 28 / 2606.26523

Radical AI Interpretability

激进的人工智能可解释性
Herrmann, Daniel A., Levinstein, Benjamin A.
Abstract
We develop a framework for interpreting AI systems as agents, drawing on the philosophical tradition of radical interpretation and the tools of mechanistic interpretability. The core question is: given the computational facts about a system, how do we solve for its beliefs, desires, and meanings? This matters increasingly for safety. We want to be able to trust the systems we deploy, whether by understanding their goals or, more modestly, by reliably detecting deception. Interpretability researchers are building tools to read beliefs and desires off a model's internals, but there is no settled account of when such a tool has succeeded. This book supplies one. We propose criteria on both representationalist and interpretationist approaches, and tie each to tests current interpretability methods can carry out. A central lesson is that these attributions cannot be made piecemeal. Beliefs, desires, and the propositional structure they presuppose are jointly constrained, and a method that fixes one while measuring the others inherits whatever distortions that introduces. This holism becomes pressing for AI systems, which may not share the interpreter's concepts. However, it also provides leverage: a system's attitudes constrain its propositional structure, that structure constrains which attitudes can be attributed, and mechanistic interpretability can help us measure both.
Chinese Translation
我们开发了一个框架，用于将人工智能系统解释为代理，借鉴了激进解释的哲学传统和机械可解释性的工具。核心问题是：在了解系统的计算事实后，我们如何推导出其信念、欲望和意义？这在安全性方面变得越来越重要。我们希望能够信任我们部署的系统，无论是通过理解它们的目标，还是更谦逊地，通过可靠地检测欺骗行为。可解释性研究者正在构建工具，以从模型的内部读取信念和欲望，但尚无公认的标准来判断这些工具何时成功。本书提供了一个标准。我们提出了代表主义和解释主义方法的标准，并将每个标准与当前可解释性方法可以进行的测试联系起来。一个核心教训是，这些归因不能零散地进行。信念、欲望及其所假设的命题结构是共同约束的，而一种在测量其他因素时固定一个因素的方法会继承引入的任何扭曲。这种整体性对于人工智能系统尤为紧迫，因为它们可能不共享解释者的概念。然而，这也提供了杠杆：系统的态度约束其命题结构，而该结构又约束可以归因于哪些态度，机械可解释性可以帮助我们同时测量这两者。
View on arXiv Download PDF AI Translation
cs.AI / 29 / 2606.26549

PMDformer: Patch-Mean Decoupling Information Transformer for Long-term Forecasting

PMDformer：用于长期预测的补丁均值解耦信息变换器
Hu, Ao, Wen, Liangjian, Duan, Jiang, Dai, Yong, Yan, He, Wang, Dongkai, Wang, Jun, Zhang, Yukun, Jiang, Ruoxi, Xu, Zenglin
Abstract
Long-term time series forecasting (LTSF) plays a crucial role in fields such as energy management, finance, and traffic prediction. Transformer-based models have adopted patch-based strategies to capture long-range dependencies, but accurately modeling shape similarities across patches and variables remains challenging due to scale differences. To address this, we introduce patch-mean decoupling (PMD), which separates the trend and residual shape information by subtracting the mean of each patch, preserving the original structure and ensuring that the attention mechanism captures true shape similarities. Futhermore, to more effectively model long-range dependencies and capture cross-variable relationships, we propose Trend Restoration Attention (TRA) and Proximal Variable Attention (PVA). The former module reintegrates the decoupled trend from PMD while calculating attention output. And the latter focuses cross-variable attention on the most relevant, recent time segments to avoid overfitting on outdated correlations. Combining these components, we propose PMDformer, a model designed to effectively capture shape similarity in long-term forecasting scenarios. Extensive experiments indicate that PMDformer outperforms existing state-of-the-art methods in stability and accuracy across multiple LTSF benchmarks. The code is available at https://github.com/aohu1105/PMDformer.
Chinese Translation
长期时间序列预测（LTSF）在能源管理、金融和交通预测等领域发挥着至关重要的作用。基于变换器（Transformer）的模型采用基于补丁的策略来捕捉长距离依赖关系，但由于尺度差异，准确建模补丁和变量之间的形状相似性仍然具有挑战性。为了解决这一问题，我们提出了补丁均值解耦（PMD），通过减去每个补丁的均值来分离趋势和残差形状信息，保留原始结构，并确保注意力机制捕捉真实的形状相似性。此外，为了更有效地建模长距离依赖关系并捕捉跨变量关系，我们提出了趋势恢复注意力（TRA）和近似变量注意力（PVA）。前者模块在计算注意力输出时重新整合来自PMD的解耦趋势，而后者则将跨变量注意力集中在最相关的、最近的时间段上，以避免对过时相关性的过拟合。结合这些组件，我们提出了PMDformer，一个旨在有效捕捉长期预测场景中形状相似性的模型。大量实验表明，PMDformer在多个LTSF基准测试中在稳定性和准确性方面优于现有的最先进方法。代码可在https://github.com/aohu1105/PMDformer获取。
View on arXiv Download PDF AI Translation
cs.AI / 30 / 2606.26561

Explainable Ensemble-Based Machine Learning Models for Detecting the Presence of Cirrhosis in Hepatitis C Patients

可解释的基于集成的机器学习模型用于检测丙型肝炎患者的肝硬化存在性
Alotaibi, Abrar, Alnajrani, Lujain, Alsheikh, Nawal, Alanazy, Alhatoon, Alshammasi, Salam, Almusairii, Meshael, Alrassan, Shoog, Alansari, Aisha
Abstract
Hepatitis C is a liver infection caused by a virus, which results in mild to severe inflammation of the liver. Over many years, hepatitis C gradually damages the liver, often leading to permanent scarring, known as cirrhosis. Patients sometimes have moderate or no symptoms of liver illness for decades before developing cirrhosis. Cirrhosis typically worsens to the point of liver failure. Patients with cirrhosis may also experience brain and nerve system damage, as well as gastrointestinal hemorrhage. Treatment for cirrhosis focuses on preventing further progression of the disease. Detecting cirrhosis earlier is therefore crucial for avoiding complications. Machine learning (ML) has been shown to be effective at providing precise and accurate information for use in diagnosing several diseases. Despite this, no studies have so far used ML to detect cirrhosis in patients with hepatitis C. This study obtained a dataset consisting of 28 attributes of 2038 Egyptian patients from the ML Repository of the University of California at Irvine. Four ML algorithms were trained on the dataset to diagnose cirrhosis in hepatitis C patients: a Random Forest, a Gradient Boosting Machine, an Extreme Gradient Boosting, and an Extra Trees model. The Extra Trees model outperformed the other models achieving an accuracy of 96.92%, a recall of 94.00%, a precision of 99.81%, and an area under the receiver operating characteristic curve of 96% using only 16 of the 28 features.
Chinese Translation
丙型肝炎是一种由病毒引起的肝脏感染，导致肝脏轻度到重度的炎症。经过多年，丙型肝炎逐渐损害肝脏，常常导致永久性瘢痕，称为肝硬化。患者在发展为肝硬化之前，可能在数十年内仅有中等或无肝脏疾病症状。肝硬化通常会恶化到肝功能衰竭的程度。肝硬化患者还可能经历脑部和神经系统损伤，以及胃肠道出血。肝硬化的治疗主要集中在防止疾病进一步进展。因此，尽早检测肝硬化对于避免并发症至关重要。机器学习（ML）已被证明在提供用于诊断多种疾病的精确和准确的信息方面有效。尽管如此，目前尚无研究使用机器学习来检测丙型肝炎患者的肝硬化。本研究从加州大学欧文分校的机器学习库中获取了一个包含2038名埃及患者28个属性的数据集。我们在该数据集上训练了四种机器学习算法以诊断丙型肝炎患者的肝硬化：随机森林（Random Forest）、梯度提升机（Gradient Boosting Machine）、极端梯度提升（Extreme Gradient Boosting）和额外树模型（Extra Trees）。额外树模型在其他模型中表现最佳，使用28个特征中的16个特征达到了96.92%的准确率、94.00%的召回率、99.81%的精确率，以及96%的接收者操作特征曲线下面积。
View on arXiv Download PDF AI Translation
cs.AI / 31 / 2606.26578

EvoOptiGraph: Weakness-Driven Coevolution via Graph-Based Structural Generation for Optimization Modeling

EvoOptiGraph：基于图的结构生成驱动弱点共进化的优化建模
Kang, Qingcan, Liu, Mingyang, Fu, Xiaojin, Kai, Shixiong, Zhong, Tao, Yuan, Mingxuan
Abstract
Automating optimization modeling from natural language with large language models (LLMs) faces two key challenges. First, training corpora lack structural diversity. Second, data generation pipelines remain static and decoupled from model learning. To address these challenges, we propose EvoOptiGraph, a novel framework where data and model co-evolve, driven by model weaknesses. EvoOptiGraph represents each mixed-integer linear program (MILP) as an attributed bipartite graph and applies validity-preserving evolutionary operators to generate structurally diverse instances. The evolved graphs are converted into solver code and natural language via deterministic compilation and verified back-translation. Training proceeds in two stages: supervised fine-tuning (SFT) on an initial dataset, followed by reinforcement learning with verifiable rewards (RLVR), where graph-derived weakness signals guide the generation of new instances targeting the model's failures. This forms a closed loop that continuously updates the training distribution. Empirical results on six public datasets show that EvoOptiGraph significantly outperforms larger generalist models, agentic methods, and specialized baselines in accuracy, executability, and generalization. These results demonstrate that targeted data-model coevolution is an effective strategy for improving LLMs on optimization modeling tasks.
Chinese Translation
利用大型语言模型（LLMs）从自然语言自动化优化建模面临两个关键挑战。首先，训练语料缺乏结构多样性。其次，数据生成管道保持静态，且与模型学习解耦。为了解决这些挑战，我们提出了EvoOptiGraph，一个新颖的框架，其中数据和模型共同进化，驱动因素是模型的弱点。EvoOptiGraph将每个混合整数线性规划（MILP）表示为一个带属性的二分图，并应用保持有效性的进化算子生成结构多样的实例。进化后的图通过确定性编译转换为求解器代码和自然语言，并通过反向翻译进行验证。训练分为两个阶段：在初始数据集上进行监督微调（SFT），然后进行可验证奖励的强化学习（RLVR），其中图生成的弱点信号指导新实例的生成，针对模型的失败。这形成了一个闭环，持续更新训练分布。在六个公共数据集上的实证结果表明，EvoOptiGraph在准确性、可执行性和泛化能力方面显著优于更大的通用模型、代理方法和专业基线。这些结果表明，针对性的数据-模型共进化是一种有效的策略，可以提高LLMs在优化建模任务上的表现。
View on arXiv Download PDF AI Translation
cs.AI / 32 / 2606.26585

A Multi-Level Validation and Traceability Framework for AI-Generated Telescope Scheduling Decisions

用于人工智能生成的望远镜调度决策的多层次验证与可追溯框架
Xiao, Hengchu, Wang, Chuanjun
Abstract
With the gradual introduction of AI into telescope scheduling, AI-based decision-making has shown advantages in handling complex multi-constraint problems. However, its outputs often suffer from inconsistent data references, reasoning errors, and non-executable decisions, limiting applicability in high-reliability observational tasks. In this work, we propose a multi-level validation and traceable reasoning framework that performs systematic reliability verification of AI-generated decisions prior to execution, and enables explicit representation of the reasoning process to support traceable decision-making. The framework integrates data reference validation, logical consistency checks, and observational and instrumental constraint verification to filter and correct invalid decisions. It also introduces atomic reasoning units and their dependency relationships, representing scheduling decisions as a sequence of interconnected reasoning steps that support error localization and post hoc analysis. Experiments show that the framework improves executability and reliability of AI scheduling and reduces loss of transient opportunities. In particular, feedback correction and structured validation of reasoning steps enhance the ability to repair and block erroneous decisions, especially in complex scenarios. Compared with pure AI methods, the framework-enhanced approach maintains flexibility while substantially improving reliability and executability. These results demonstrate a feasible and verifiable pathway for applying AI to high-reliability astronomical observation scheduling.
Chinese Translation
随着人工智能逐渐引入望远镜调度，基于人工智能的决策在处理复杂的多约束问题上展现出了优势。然而，其输出往往存在数据引用不一致、推理错误和不可执行决策等问题，限制了其在高可靠性观测任务中的适用性。在本研究中，我们提出了一种多层次验证和可追溯推理框架，该框架在执行之前对人工智能生成的决策进行系统的可靠性验证，并能够明确表示推理过程以支持可追溯的决策制定。该框架集成了数据引用验证、逻辑一致性检查以及观测和仪器约束验证，以过滤和修正无效决策。它还引入了原子推理单元及其依赖关系，将调度决策表示为一系列相互关联的推理步骤，以支持错误定位和事后分析。实验表明，该框架提高了人工智能调度的可执行性和可靠性，减少了瞬时机会的损失。特别是，反馈修正和结构化的推理步骤验证增强了修复和阻止错误决策的能力，尤其是在复杂场景中。与纯人工智能方法相比，增强框架的方法在保持灵活性的同时显著提高了可靠性和可执行性。这些结果展示了将人工智能应用于高可靠性天文观测调度的可行且可验证的路径。
View on arXiv Download PDF AI Translation
cs.AI / 33 / 2606.26593

Content-Based Smart E-Mail Dispatcher Using Large Language Models

基于内容的智能电子邮件分发系统使用大型语言模型
Paramesha, K., Sriram, K R, Shetty, Sujan, Kishore, Shamanth, Tejaswini, R.
Abstract
Email communication has become an integral part of personal and professional life, but handling its vast volume is still a significant issue for large organisations. Manual perusal of emails and forwarding their contents and attachments to intended recipients using other instant messaging platforms has proved to be error-prone and time-consuming leading to losses in terms of productivity and creating undue stress. The main objective of this paper is to explore an alternative mechanism that is to automate the task of dispatching emails based on their contents to the respective WhatsApp groups of students of various semesters of programs in an engineering college, facilitating a smooth flow of information from one end to another end in an organisation. The dispatcher system is built using agents querying large language models (LLMs) to enable it to analyze the contents of emails and route them to the relevant groups of students for their information and consumption. The system harnesses the capabilities of LLMs in analysing the textual contents for decision-making. With a well-structured agent framework prompt that includes email content as input with instructions and context, the system figures out the relevant groups to which the email message is dispatched, thus providing the required information on time. The proposed system does not rely on labelled datasets and provides several benefits, including enhanced productivity and a reduction in the cognitive load associated with reading emails.
Chinese Translation
电子邮件通信已成为个人和职业生活中不可或缺的一部分，但处理其庞大的数量仍然是大型组织面临的重大问题。手动浏览电子邮件并通过其他即时通讯平台将其内容和附件转发给预定收件人，已被证明容易出错且耗时，导致生产力损失并造成不必要的压力。本文的主要目的是探索一种替代机制，即自动化根据电子邮件内容将其分发到工程学院各个学期学生的相应WhatsApp群组，从而促进组织内信息的顺畅流动。该分发系统使用代理查询大型语言模型（LLMs）构建，使其能够分析电子邮件的内容并将其路由到相关的学生群组以供其获取和使用。该系统利用LLMs在分析文本内容以进行决策方面的能力。通过一个结构良好的代理框架提示，其中包括电子邮件内容作为输入以及指令和上下文，系统能够确定电子邮件消息应分发到的相关群组，从而及时提供所需的信息。所提议的系统不依赖于标记数据集，并提供了多种好处，包括提高生产力和减少与阅读电子邮件相关的认知负担。
View on arXiv Download PDF AI Translation
cs.AI / 34 / 2606.26595

LLM-based Models for Detecting Emerging Topics in Service Feedback

基于大型语言模型的服务反馈新兴主题检测模型
Tavakoli, Mahsa, Bankey, Ruth, Bravo, Cristián
Abstract
Enhancing the analysis of service feedback is essential for public sector organizations, particularly tax administrations, where trust and compliance depend on fair and effective service delivery. As feedback volumes grow, identifying emerging service quality issues and potential disparities across diverse populations becomes increasingly challenging. Traditional approaches often rely on manual review or static expert-defined indicators, limiting scalability and the ability to capture complex patterns in textual feedback. This paper presents a novel methodology that integrates large language models (LLMs), statistical techniques, and human-AI collaboration to improve multilingual customer feedback analysis. The primary objective is to detect emerging service quality topics that may also reveal potential inequities in service delivery. Our framework combines fine-tuned, quantized LLMs with expert oversight to produce accurate, computationally efficient, and context-aware analyses. The proposed approach was evaluated using similarity analysis and assessments from experienced tax officers, demonstrating stronger alignment with expert judgments than baseline models. By incorporating a human-in-the-loop framework, the methodology reduces LLM fabrication while improving the reliability and relevance of generated insights. The results demonstrate the practicality of combining LLMs with human expertise to support scalable, evidence-based decision-making in public sector organizations. This work contributes to the development of responsible AI systems that enhance service quality, responsiveness, fairness, and public trust through more effective analysis of multilingual customer feedback.
Chinese Translation
增强服务反馈的分析对于公共部门组织，特别是税务机关至关重要，因为信任和合规性依赖于公平和有效的服务交付。随着反馈量的增加，识别新兴的服务质量问题以及不同人群之间的潜在差异变得愈加困难。传统方法通常依赖于人工审查或静态的专家定义指标，这限制了可扩展性以及捕捉文本反馈中复杂模式的能力。本文提出了一种新颖的方法论，结合了大型语言模型（LLMs）、统计技术和人机协作，以改善多语言客户反馈分析。主要目标是检测新兴的服务质量主题，这些主题可能还揭示服务交付中的潜在不平等。我们的框架结合了经过微调和量化的LLMs与专家监督，以产生准确、计算效率高且具有上下文感知的分析。所提出的方法通过相似性分析和经验丰富的税务官员的评估进行了评估，显示出与专家判断的更强一致性，优于基线模型。通过引入人机协作框架，该方法减少了LLM的虚假生成，同时提高了生成洞察的可靠性和相关性。结果表明，将LLMs与人类专业知识相结合以支持公共部门组织的可扩展、基于证据的决策具有实际意义。这项工作为开发负责任的人工智能系统做出了贡献，通过更有效地分析多语言客户反馈，提高服务质量、响应能力、公平性和公众信任。
View on arXiv Download PDF AI Translation
cs.AI / 35 / 2606.26649

Autoformalization of Agent Instructions into Policy-as-Code

将代理指令自动形式化为代码策略
Mondl, Adam, Maisel, Matthew, Brock, John H.
Abstract
Agent safety in high-stakes domains requires formal policy enforcement, but most existing approaches either rely on probabilistic guardrails (fine-tuned classifiers, prompt-based steering) that offer no formal guarantees, or on hand-coded symbolic enforcement that does not scale to the breadth of real policy specifications. We present an autoformalization pipeline that translates agent prompts, MCP tool descriptions, and natural language policy documents into formally verified policies using an LLM-based generator-critic loop. The resulting policies are written in the Cedar Policy Language. On the MedAgentBench benchmark, our autoformalized policies cover substantially more of the source natural-language specification than the hand-coded symbolic enforcement in prior work.
Chinese Translation
在高风险领域，代理的安全性需要正式的政策执行，但现有的大多数方法要么依赖于没有正式保证的概率性保护措施（如微调的分类器、基于提示的引导），要么依赖于手工编码的符号执行，这在实际政策规范的广度上无法扩展。我们提出了一种自动形式化管道，该管道利用基于大型语言模型（LLM）的生成-评估循环，将代理提示、MCP工具描述和自然语言政策文档转换为形式验证的政策。生成的政策使用Cedar政策语言编写。在MedAgentBench基准测试中，我们的自动形式化政策覆盖了比以往工作中的手工编码符号执行更多的源自然语言规范。
View on arXiv Download PDF AI Translation
cs.AI / 36 / 2606.26669

SKILL-DISCO: Distilling and Compiling Agent Traces into Reusable Procedural Skills

SKILL-DISCO：将智能体轨迹提炼和编译为可重用的程序技能
Guo, Zhongxin, Qi, Danrui, Gu, Hanwen, Cheng, Peng, Xiong, Yongqiang
Abstract
Agents often repeatedly solve similar task instances from scratch, leading to unnecessary reasoning cost and long execution traces. Prior work has explored workflow reuse and executable skill induction, but it remains unclear which task scenarios admit procedural skills and how the shared procedural structure should be represented across successful traces. We study this problem in FSM-defined scenarios, where successful traces can be viewed as paths in an unknown transition graph, and formulate procedural skills as reusable parameterized control-flow subgraphs. Based on this view, we introduce SkillDisCo, a distillation-and-compilation framework that distills reusable PFSM subgraphs from successful traces and compiles them into callable, executable, and verifiable procedural skills. Experiments on ALFWorld and WebArena show that SkillDisCo improves success rates and reduces agent turns across benchmarks and model scales, demonstrating the benefits of representing shared experience as reusable execution structures.
Chinese Translation
智能体经常从头开始重复解决相似的任务实例，这导致了不必要的推理成本和较长的执行轨迹。之前的研究探讨了工作流重用和可执行技能诱导，但尚不清楚哪些任务场景适合程序技能，以及如何在成功轨迹之间表示共享的程序结构。我们在有限状态机（FSM）定义的场景中研究这个问题，在这些场景中，成功轨迹可以视为未知转移图中的路径，并将程序技能形式化为可重用的参数化控制流子图。基于这一视角，我们引入了SkillDisCo，一个提炼和编译框架，它从成功轨迹中提炼可重用的参数化有限状态机（PFSM）子图，并将其编译为可调用、可执行和可验证的程序技能。在ALFWorld和WebArena上的实验表明，SkillDisCo提高了成功率并减少了智能体的回合数，展示了将共享经验表示为可重用执行结构的好处。
View on arXiv Download PDF AI Translation
cs.AI / 37 / 2606.26671

NebulaExp-8B: An Empirical Post-Training Pipeline via Full-Scale Ablation Research

NebulaExp-8B：通过全尺度消融研究的经验后训练流程
Hao, Qiaobo, Wu, Yangqian, Wang, Shunyi, Zhang, Zhongjian, Li, Ziqun, He, Yayin, Li, Muqing, Zhong, Chen
Abstract
Post-training alignment determines the reasoning and human preference following capabilities of large language models, yet most existing works withhold detailed data construction, filtering rules and training recipes, which hinders community reproducibility and lightweight model optimization. This work presents NebulaExp, a fully transparent, ablation-driven post-training pipeline built on Qwen3-8B-base, covering two orthogonal model branches: general instruct model and complex reasoning-specialized model. We curate a raw corpus of 3.84M multi-source SFT samples and a 200K verifiable RL candidate pool, and design an end-to-end data processing stack including response distillation, multi-dimensional cross-verification filtering, fine-grained difficulty grading, task classification and diversity-aware sampling. For the Instruct branch, our three-stage optimized supervised fine-tuning approach NebulaExp-Ins-SFT improves the average benchmark score from the 55.01 baseline of Qwen3-8B-nothink to 60.99. GRPO reinforcement learning then further elevates the average score to 61.85. For the Reasoning branch, medium-difficulty GRPO RL improves average reasoning score from 73.88 to 75.17. To address RL's dependency on task verifiers, we systematically investigate single-teacher and multi-teacher OPD (MOPD): utilizing merely 4K instruction-following samples and outperforms RL baseline by 3.26 points on IFEval with +4.43 average overall gain; MOPD fuses four domain-specialist teachers with merely 10K samples, lifting average performance by 4.18 over the base model. This report provides a fully reproducible empirical post-training recipe for 8B-scale LLMs, and comprehensively dissects the capability trade-offs among instruction adherence, mathematical reasoning, code generation and general knowledge.
Chinese Translation
后训练对齐决定了大型语言模型的推理和人类偏好跟随能力，但现有大多数工作缺乏详细的数据构建、过滤规则和训练方案，这阻碍了社区的可重复性和轻量模型的优化。本研究提出了NebulaExp，一个完全透明的、以消融为驱动的后训练流程，基于Qwen3-8B-base，涵盖两个正交的模型分支：通用指令模型和复杂推理专用模型。我们整理了384万条多源SFT样本的原始语料库和20万条可验证的RL候选池，并设计了一个端到端的数据处理堆栈，包括响应蒸馏、多维交叉验证过滤、细粒度难度分级、任务分类和多样性感知采样。对于指令分支，我们的三阶段优化监督微调方法NebulaExp-Ins-SFT将Qwen3-8B-nothink的基线平均基准分数从55.01提高到60.99。随后，GRPO强化学习进一步将平均分数提升至61.85。对于推理分支，中等难度的GRPO RL将平均推理分数从73.88提高到75.17。为了解决RL对任务验证者的依赖，我们系统地研究了单教师和多教师OPD（MOPD）：仅利用4000条遵循指令的样本，在IFEval上超越RL基线3.26分，整体平均增益为+4.43；MOPD融合了四位领域专家教师，仅使用10000条样本，平均性能比基础模型提高了4.18。本报告提供了一个完全可重复的8B规模LLM的经验后训练方案，并全面剖析了指令遵循、数学推理、代码生成和一般知识之间的能力权衡。
View on arXiv Download PDF AI Translation
cs.AI / 38 / 2606.26686

Do Safety Guardrails Need to Reason? LeanGuard: A Fast and Light Approach for Robust Moderation

安全护栏需要推理吗？LeanGuard：一种快速轻量的稳健调节方法
Na, Dongbin
Abstract
In order to screen a prompt or a response, the recent guardrail methods generate a chain-of-thought (CoT) before they issue a verdict. This design follows a common belief that step-by-step reasoning improves a decision. However, CoT also makes the guard heavy and slow, because the model must generate many tokens before it decides. This may not match how guardrails are actually deployed. A guardrail sometimes should not be heavy and slow, and it often runs on-device, for example on an embodied robot. In this paper, we pose a question whether a safety guardrail really needs to reason. To answer this question, we train a lightweight bidirectional encoder and a reasoning guard on the same corpus, and we then remove only the reasoning while we keep everything else fixed. With this controlled same-base comparison, we show that the chain does not improve moderation accuracy. We name the resulting guard LeanGuard. A 395M label-only encoder reaches an average F1 of 82.90 $\pm$ 0.26 over public benchmarks. It matches a reasoning guard that is built on a much larger decoder, while it uses only a single forward pass over an input of at most 512 tokens. This is about a ~100x reduction in inference compute. We further show that this label-only encoder stays robust under training-label noise and retains far more recall at a strict false-positive rate than the reasoning guard, so a heavier reasoning guard is not the more robust choice either. Our finding suggests that the current guardrail benchmarks may not be hard enough to reward reasoning, and that the necessity of CoT for moderation is still not proven. We release all source codes and models including LeanGuard at https://github.com/ndb796/LeanGuard.
Chinese Translation
为了筛选提示或响应，最近的护栏方法在做出判断之前会生成一系列思考链（Chain-of-Thought, CoT）。这种设计遵循了一个普遍的信念，即逐步推理可以改善决策。然而，CoT也使得护栏变得笨重且缓慢，因为模型必须在做出决定之前生成许多标记。这可能与护栏的实际部署方式不符。护栏有时不应当是笨重和缓慢的，且它通常在设备上运行，例如在一个具身机器人上。在本文中，我们提出一个问题：安全护栏真的需要推理吗？为了回答这个问题，我们在相同的语料库上训练了一个轻量级的双向编码器和一个推理护栏，然后我们在保持其他一切不变的情况下，仅去除推理。通过这种受控的同基准比较，我们表明，思考链并没有提高调节的准确性。我们将得到的护栏命名为LeanGuard。一个395M的仅标签编码器在公共基准上达到了平均F1值82.90 ± 0.26。它的表现与一个基于更大解码器构建的推理护栏相匹配，同时它仅在最多512个标记的输入上进行一次前向传递。这大约是推理计算的减少了~100倍。我们进一步表明，这个仅标签编码器在训练标签噪声下保持稳健，并在严格的假阳性率下保留了远高于推理护栏的召回率，因此，较重的推理护栏也不是更稳健的选择。我们的发现表明，目前的护栏基准可能不够严格，无法奖励推理，并且CoT在调节中的必要性仍未得到证明。我们在https://github.com/ndb796/LeanGuard发布了所有源代码和模型，包括LeanGuard。
View on arXiv Download PDF AI Translation
cs.AI / 39 / 2606.26710

Kalman Prototypical Networks for Few-shot Fault Detection in Combined Cycle Gas Turbines

用于联合循环燃气轮机少样本故障检测的卡尔曼原型网络
Belay, Mohammed Ayalew, Bernardino, Lucas Ferreira, Rasheed, Adil, Montañés, Rubén M., Rossi, Pierluigi Salvo
Abstract
Combined-cycle gas turbines (CCGTs) play a key role in modern power generation, offering both high efficiency and reduced environmental impact. However, their complex thermo-fluid and mechanical interactions complicate fault detection, particularly when labeled fault data are scarce. In this paper, we introduce the Kalman Prototypical Network (KPN), a metric-based few-shot learning (FSL) framework specifically tailored for CCGT fault diagnosis. We model the evolution of class prototypes as latent stochastic states in a dynamic system to reduce episodic variance and improve robustness in embedding representation. Synthetic data sets generated with a high-fidelity Modelica-based dynamic simulation of an offshore CCGT system were used, simulating both normal operation and progressive leak faults under transient conditions. Application of the proposed framework on simulated leak fault detection tasks demonstrate that KPN outperforms conventional FSL methods such as Matching Networks, Relation Networks, and MAML in both accuracy and stability under varying support and query configurations. The proposed framework significantly improves training convergence and generalization by stabilizing class representations, making it well-suited for real-world CCGT fault detection where labeled data is limited.
Chinese Translation
联合循环燃气轮机（CCGT）在现代发电中发挥着关键作用，提供高效率和降低环境影响。然而，它们复杂的热流体和机械相互作用使得故障检测变得复杂，尤其是在标记故障数据稀缺的情况下。本文介绍了卡尔曼原型网络（KPN），这是一种基于度量的少样本学习（FSL）框架，专门针对CCGT故障诊断。我们将类别原型的演变建模为动态系统中的潜在随机状态，以减少情景方差并提高嵌入表示的鲁棒性。使用基于高保真Modelica的动态仿真生成的合成数据集，模拟了离岸CCGT系统在瞬态条件下的正常运行和逐步泄漏故障。对模拟泄漏故障检测任务应用所提出的框架表明，KPN在准确性和稳定性方面均优于传统的FSL方法，如匹配网络（Matching Networks）、关系网络（Relation Networks）和模型无关元学习（MAML），在不同的支持和查询配置下表现出色。所提出的框架通过稳定类别表示显著改善了训练收敛性和泛化能力，使其非常适合于标记数据有限的实际CCGT故障检测。
View on arXiv Download PDF AI Translation
cs.AI / 40 / 2606.26713

LithoDreamer: A Physics-Informed World Model for Multi-Stage Computational Lithography

LithoDreamer：一种用于多阶段计算光刻的物理信息世界模型
Jiang, Yuqi, Liu, Yumeng, Li, Zimu, Deng, Jinyuan, Jin, Qian, Cui, Yucheng, Li, Yu, Yin, Xunzhao, Sun, Qi, Zhuo, Cheng
Abstract
As semiconductor technology nodes scale, computational lithography is essential for ensuring yield and performance. However, lithography is a continuous physical process involving mask optimization, optical imaging, resist exposure, and development, which existing models fail to capture. To overcome this limitation, we present LithoDreamer, the first physics-informed World Model (WM) framework for computational lithography, which formulates the ``Layout-Mask-Resist Image-After Development Image (ADI)'' pipeline as a decision-driven multi-step evolution system. LithoDreamer captures feature changes between adjacent states to model stage-specific physics-informed latent spaces, in which it controls process intervention exploration and drives subsequent state transitions. To achieve interpretable intervention optimization without continuous supervision, we propose a contrastive variational optimization paradigm that contrasts the latent differences between intervention paths with variational evolution constraints, guiding the model to generate evolutions consistent with real lithography physics. Experiments show LithoDreamer achieves state-of-the-art performance in forward evolution and inverse planning. Our lithography dataset is publicly available at GitHub (https://github.com/7jiangyq/lithodreamer.git).
Chinese Translation
随着半导体技术节点的缩小，计算光刻对于确保产量和性能至关重要。然而，光刻是一个连续的物理过程，涉及掩模优化、光学成像、光刻胶曝光和显影，而现有模型无法捕捉这一过程。为了解决这一限制，我们提出了LithoDreamer，这是第一个用于计算光刻的物理信息世界模型（World Model, WM）框架，它将“布局-掩模-光刻胶图像-显影后图像（ADI）”管道形式化为一个基于决策的多步骤演化系统。LithoDreamer捕捉相邻状态之间的特征变化，以建模阶段特定的物理信息潜在空间，在该空间中控制过程干预探索并驱动后续状态转换。为了实现无需持续监督的可解释干预优化，我们提出了一种对比变分优化范式，该范式对比干预路径之间的潜在差异与变分演化约束，引导模型生成与真实光刻物理一致的演化。实验表明，LithoDreamer在正向演化和逆向规划方面达到了最先进的性能。我们的光刻数据集已在GitHub上公开（https://github.com/7jiangyq/lithodreamer.git）。
View on arXiv Download PDF AI Translation
cs.AI / 41 / 2606.26718

A Latent ODE Approach to Spatiotemporal Modeling of Cine Cardiac MRI

潜在常微分方程方法在心脏磁共振成像时空建模中的应用
Brüggemann, David, Krymova, Ekaterina, Özdemir, Firat, von Spiczak, Jochen, Kozerke, Sebastian, Mora, Samia, Manka, Robert, Salzmann, Mathieu, Demler, Olga V.
Abstract
Cardiac magnetic resonance imaging (CMR) captures rich spatiotemporal information about ventricular structure and motion, but conventional risk models use only a few image-derived indices from selected cardiac phases. We present a latent dynamical model that encodes bi-ventricular anatomy and full-cycle cine motion as a continuous latent trajectory, using heart-rate-aware neural ordinary differential equation (ODE) dynamics and a graph-based mesh autoencoder to reconstruct anatomically consistent 3D+t ventricular motion. A covariate-conditioned prior defines the expected end-diastolic latent state, and a Cox proportional hazards model tests whether deviations from this prior predict incident heart failure. We studied 72,386 UK Biobank participants without baseline cardiovascular disease, including 367 incident heart failure events. In a held-out evaluation subset, adding the latent score to refitted pooled cohort equations improved the stratified C-index from 0.704 to 0.785, compared with 0.764 for seven established cardiac markers. Compared with non-graph and non-ODE approaches, the proposed model gave the best trade-off between reconstruction fidelity, generative realism, and downstream prognostic performance. These results suggest that continuous full-cycle modeling of ventricular motion provides informative cardiac phenotypes beyond conventional CMR summaries, while external validation in more representative patient cohorts is required before clinical risk-prediction use.
Chinese Translation
心脏磁共振成像（CMR）捕捉了丰富的心室结构和运动的时空信息，但传统的风险模型仅使用来自选定心脏相位的少量图像衍生指标。我们提出了一种潜在动态模型，该模型将双心室解剖结构和全周期的心动运动编码为连续的潜在轨迹，利用考虑心率的神经常微分方程（ODE）动态和基于图的网格自编码器重建解剖一致的三维时序心室运动。协变量条件先验定义了预期的舒张末期潜在状态，而Cox比例风险模型则检验该先验的偏差是否能预测心力衰竭事件的发生。我们研究了72,386名没有基础心血管疾病的UK Biobank参与者，其中包括367个心力衰竭事件。在一个独立的评估子集中，将潜在评分添加到重新拟合的汇总队列方程中，使分层C指数从0.704提高到0.785，而七个已建立的心脏标志物的C指数为0.764。与非图形和非ODE方法相比，所提出的模型在重建保真度、生成现实性和下游预后性能之间提供了最佳的权衡。这些结果表明，心室运动的连续全周期建模提供了超越传统CMR摘要的信息性心脏表型，但在临床风险预测使用之前，需要在更具代表性的患者队列中进行外部验证。
View on arXiv Download PDF AI Translation
cs.AI / 42 / 2606.26722

Socratic agents for autonomous scientific discovery in high-dimensional physical systems

用于高维物理系统自主科学发现的苏格拉底代理
Zeng, Xianrui, Liu, Pengfei, Zang, Yirui, Shen, Yang, Yu, Fei, Yu, Chunlei, Liu, Minghao, Du, Yang
Abstract
The automation of scientific discovery has reached an inflection point. While AI systems now operate instruments, optimize parameters and generate hypotheses, most remain procedural: they execute workflows fixed by human designers. True autonomous science demands epistemic autonomy--the capacity to construct, challenge and revise physical explanations in response to evidence. Here we introduce AHOIS, a multi-agent AI scientist that embeds Socratic midwifery into closed-loop experimentation. A physics-critic agent interrogates hypotheses through causal questioning, constraint checking, counterexample generation and falsification-criteria formulation. We evaluate AHOIS on a real multimode-fibre optical platform, a high-dimensional system with complex wave transformations, indirect detection, environmental drift and multi-modal acquisition. Without prior encoding schemes, classifiers or speckle models, the system autonomously proposed and validated a random-interference encoding hypothesis, discovered task-adaptive sparse-measurement strategies, diagnosed distinct failure modes (encoding instability, fluorescence contamination and detector noise) and translated a published imaging protocol into an executable workflow on a non-original configuration. The discovered encoding yielded 16x16 measurements with effective rank 56.9 and classification accuracies of 76.97% on MNIST and 83.17% on Fashion-MNIST. Ablations show that Socratic interrogation improves physical consistency, hypothesis completeness, uncertainty calibration and experimental-plan validity. These results establish a route from workflow automation towards evidence-grounded, self-correcting autonomous discovery in complex physical environments.
Chinese Translation
科学发现的自动化已达到一个转折点。虽然人工智能系统现在可以操作仪器、优化参数和生成假设，但大多数仍然是程序性的：它们执行由人类设计者固定的工作流程。真正的自主科学需要认知自主性——在证据的基础上构建、挑战和修订物理解释的能力。在这里，我们介绍了AHOIS，一个将苏格拉底助产术嵌入闭环实验中的多智能体人工智能科学家。一个物理批评代理通过因果提问、约束检查、反例生成和反驳标准的制定来质疑假设。我们在一个真实的多模光纤光学平台上评估了AHOIS，这是一个具有复杂波动变换、间接检测、环境漂移和多模采集的高维系统。在没有先前编码方案、分类器或散斑模型的情况下，该系统自主提出并验证了一个随机干涉编码假设，发现了任务自适应稀疏测量策略，诊断了不同的故障模式（编码不稳定性、荧光污染和探测器噪声），并将已发布的成像协议转化为在非原始配置上的可执行工作流程。所发现的编码产生了16x16的测量，具有有效秩56.9，并在MNIST上获得了76.97%的分类准确率，在Fashion-MNIST上获得了83.17%的分类准确率。消融实验表明，苏格拉底式的质询提高了物理一致性、假设完整性、不确定性校准和实验计划的有效性。这些结果建立了一条从工作流程自动化到在复杂物理环境中基于证据的自我修正自主发现的路径。
View on arXiv Download PDF AI Translation
cs.AI / 43 / 2606.26728

Scientific discovery as meta-optimization: a combinatorial optimization case study

科学发现作为元优化：一个组合优化案例研究
Zhang, Yuan-Hang, Sipling, Chesson, Di Ventra, Massimiliano
Abstract
Scientific discovery is fundamentally an optimization problem, defined by a vast "state space" of theories and experiments, and an evaluation criterion based on quality, novelty, and validity. Large language models (LLMs) have enabled automated exploration of this space, but we argue that simultaneous modification of the evaluation criteria is equally important. Here, we propose formalizing research as meta-optimization, where the optimization objective itself is also being optimized. Our key contribution is "consensus objective aggregation," where LLM-generated objective functions are combined via correlation-weighted voting, yielding a stable, self-correcting evaluation criterion that evolves as understanding deepens. We apply this framework to algorithm discovery for 3-SAT problems based on digital MemComputing machines, reducing the baseline scaling with problem size $N$ from $\sim N^{2.51}$ to $\sim N^{1.33}$ and delivering a $\sim 67\times$ speedup on the largest instances tested. As a problem-agnostic framework, we hope this approach will considerably aid scientific discovery.
Chinese Translation
科学发现本质上是一个优化问题，其定义为一个庞大的“状态空间”，包括理论和实验，以及基于质量、新颖性和有效性的评估标准。大型语言模型（LLMs）使得对这一空间的自动探索成为可能，但我们认为同时修改评估标准同样重要。在此，我们提出将研究形式化为元优化，其中优化目标本身也在被优化。我们的关键贡献是“共识目标聚合”，通过相关性加权投票将LLM生成的目标函数结合起来，从而产生一个稳定的、自我修正的评估标准，随着理解的加深而演变。我们将这一框架应用于基于数字MemComputing机器的3-SAT问题的算法发现，将基线随问题规模$N$的增长从$ ext{约} N^{2.51}$降低到$ ext{约} N^{1.33}$，并在测试的最大实例上实现了约67倍的加速。作为一个与问题无关的框架，我们希望这一方法能够显著促进科学发现。
View on arXiv Download PDF AI Translation
cs.AI / 44 / 2606.26758

EGG: An Expert-Guided Agent Framework for Kernel Generation

EGG：一种专家指导的内核生成代理框架
Han, Yaochen, Fan, Ke, Jiang, Hongxu, Xu, Wanqi, Xie, Weiyu, Zhang, Runhua, Zhu, Chenhui, Zhang, Yixiang
Abstract
High-performance GPU kernels are critical for reducing the exponentially growing computational costs of large language models (LLMs), but their development heavily relies on manual tuning by domain experts. While recent advances in LLM-based approaches show promise for automating kernel generation, they still struggle to achieve both correctness and high performance. This limitation primarily arises from the lack of domain-specific optimization guidance, hindering effective exploration of the optimization space. We propose EGG, an Expert-Guided Agent Framework for Kernel Generation, which incorporates expert optimization principles to guide LLMs' decisions. Inspired by expert workflows, we decompose kernel generation into two hierarchical stages: 1) algorithmic structure design, which establishes a high-quality computational structure foundation; 2) hardware-specific tuning, which performs targeted adjustments through parallel mapping, tensor tiling, and memory optimization. This staged decomposition defines explicit optimization objectives, structuring the design space to achieve progressive refinement. To this end, a stage-aware multi-agent collaboration mechanism is designed for inter and intra-stage context management, ensuring stable optimization trajectories. Experiments on KernelBench and real-world workloads show that EGG achieves a 2.13x average speedup over PyTorch, outperforming existing agent-based and RL-based approaches.
Chinese Translation
高性能GPU内核对于降低大语言模型（LLMs）日益增长的计算成本至关重要，但其开发在很大程度上依赖于领域专家的手动调优。尽管基于LLM的方法在自动化内核生成方面显示出希望，但它们仍然难以同时实现正确性和高性能。这一限制主要源于缺乏特定领域的优化指导，阻碍了优化空间的有效探索。我们提出了EGG，一种专家指导的内核生成代理框架，结合了专家优化原则来指导LLMs的决策。受到专家工作流程的启发，我们将内核生成分解为两个层次的阶段：1）算法结构设计，建立高质量计算结构的基础；2）硬件特定调优，通过并行映射、张量切片和内存优化进行有针对性的调整。这种分阶段的分解定义了明确的优化目标，构建了设计空间以实现逐步细化。为此，设计了一种阶段感知的多代理协作机制，用于阶段间和阶段内的上下文管理，确保稳定的优化轨迹。在KernelBench和实际工作负载上的实验表明，EGG在PyTorch上实现了2.13倍的平均加速，超越了现有的基于代理和基于强化学习的方法。
View on arXiv Download PDF AI Translation
cs.AI / 45 / 2606.26769

ResilPhase: Plug-and-Play Phase Mapping and Noise-Resilient Macro-Trajectory Extrapolation for Diffusion Acceleration

ResilPhase：即插即用的相位映射与抗噪声的宏轨迹外推用于扩散加速
Zhao, Qicheng, Li, Yu, Sun, Qi, Yan, Zheyu
Abstract
The adoption of powerful diffusion models is hindered by their significant inference latency. Recent ``cache-then-forecast'' schemes alleviate this issue by accelerating DiTs using derivative-based polynomials, but they suffer from severe quality degradation at high acceleration ratios. Our analysis reveals its root cause: the discrete extrapolation performed on representations that are misaligned with the continuous diffusion trajectory and are numerically unstable. Thus, accelerated DiTs suffer from accumulated spatial errors, noisy derivative amplification, and high-order instability. We therefore reformulate accelerated inference as stable macro-trajectory extrapolation in ordinary differential equation (ODE) space. Instead of predicting intermediate features, we align forecasting with the model's Global Drift (GD), i.e., the end-to-end state evolution, thereby eliminating feature inconsistency and memory overhead. However, even this smooth macro-trajectory remains vulnerable to the derivative fallacy: its higher-order temporal derivatives are intrinsically noisy. Thus, we introduce a derivative-free barycentric Lagrange extrapolator to effectively bypass derivative instability and approximation error. We further propose a bounded Phase Mapping that regularizes the extrapolation domain, suppressing oscillatory error growth. These elements collectively constitute ResilPhase, a noise-resilient acceleration framework. Experiments on FLUX.1-dev and HunyuanVideo demonstrate state-of-the-art fidelity under aggressive acceleration ratios.
Chinese Translation
强大的扩散模型的应用受到其显著推理延迟的限制。最近的“缓存后预测”方案通过使用基于导数的多项式加速了扩散图像变换（DiTs），但在高加速比下严重降低了质量。我们的分析揭示了其根本原因：在与连续扩散轨迹不对齐且数值不稳定的表示上进行的离散外推。因此，加速的DiTs遭受了累积的空间误差、噪声导数放大和高阶不稳定性。因此，我们将加速推理重新表述为常微分方程（ODE）空间中的稳定宏轨迹外推。我们不再预测中间特征，而是将预测与模型的全局漂移（Global Drift, GD）对齐，即端到端的状态演变，从而消除特征不一致性和内存开销。然而，即使这个平滑的宏轨迹仍然容易受到导数谬误的影响：其高阶时间导数本质上是噪声的。因此，我们引入了一种无导数的重心拉格朗日外推器，以有效绕过导数不稳定性和近似误差。我们进一步提出了一种有界的相位映射，规范化外推域，抑制振荡误差的增长。这些元素共同构成了ResilPhase，一个抗噪声的加速框架。在FLUX.1-dev和HunyuanVideo上的实验表明，在激进的加速比下实现了最先进的保真度。
View on arXiv Download PDF AI Translation
cs.AI / 46 / 2606.26806

Memory Depth, Not Memory Access: Selective Parametric Consolidation for Long-Running Language Agents

记忆深度，而非记忆访问：长时间运行语言代理的选择性参数整合
Han, Haoliang
Abstract
Long-running language agents need more than memory access. Retrieval systems can fetch past facts at query time, but they do not decide which experiences should continue to shape behavior after the working context is unloaded. We study this separate problem as memory depth: durable goal-conditioned tendencies written into a small parametric store. We introduce the loop-drift protocol, a controlled stress test in which the retrieval index remains intact while working context is unloaded and goal-conditioned behavior must persist under long-loop interference. We evaluate EVAF, a surprise- and valence-gated LoRA consolidation mechanism. Across GPT-2 and TinyLlama, retrieval is strongest on shallow factual recall (short-fact accuracy 0.956--0.973), while EVAF is strongest on goal persistence and post-unload recovery (0.812--0.904) with only 2--3 parametric writes per 200 events. Mechanism controls show that selective consolidation factorizes into two controllable dimensions: selection and actuation. Matched random gates isolate selection beyond sparse writing; fixed-inner controls across GPT-2, TinyLlama, and Mistral-7B show that inner-loop write strength is model-dependent; and a Mistral-7B matched-gate inversion reveals asymmetric selection-actuation coupling under miscalibrated actuation. Public Memora event streams serve as an external diagnostic, exposing stale-memory invalidation as an unresolved boundary. Within this probe, selective parametric consolidation supplies memory depth distinct from and complementary to retrieval access.
Chinese Translation
长时间运行的语言代理不仅需要记忆访问。检索系统可以在查询时获取过去的事实，但它们并不决定哪些经验应该在工作上下文卸载后继续影响行为。我们将这个独立的问题研究为记忆深度：持久的目标条件倾向被写入一个小的参数存储中。我们引入了循环漂移协议，这是一种受控压力测试，在此过程中检索索引保持完整，而工作上下文被卸载，目标条件行为必须在长循环干扰下持续存在。我们评估了EVAF，这是一种基于惊讶和效价的LoRA整合机制。在GPT-2和TinyLlama上，检索在浅层事实回忆（短事实准确率为0.956--0.973）上表现最强，而EVAF在目标持续性和卸载后恢复（0.812--0.904）上表现最强，仅需每200个事件进行2--3次参数写入。机制控制显示，选择性整合可以分解为两个可控维度：选择和激活。匹配的随机门在稀疏写入之外隔离选择；在GPT-2、TinyLlama和Mistral-7B上的固定内部控制显示，内部循环写入强度依赖于模型；而Mistral-7B的匹配门反转揭示了在误校准激活下的不对称选择-激活耦合。公共Memora事件流作为外部诊断，暴露了陈旧记忆失效作为一个未解决的边界。在这个探测中，选择性参数整合提供了与检索访问不同且互补的记忆深度。
View on arXiv Download PDF AI Translation
cs.AI / 47 / 2606.26807

KARLA: Knowledge-base Augmented Retrieval for Language Models

KARLA：基于知识库增强的语言模型检索
Crespin, Francois, Suchanek, Fabian M., Holzenberger, Nils
Abstract
We propose a new method that allows an LLM to automatically pull in factual knowledge from a knowledge base during token generation. This means that (1)~factual knowledge in the LLM output can be updated without retraining the LLM, (2)~facts in the LLM output can be traced to the knowledge base for transparency and explainability, and (3)~smaller models can achieve the same factual accuracy as larger models. Our core idea is to train the model to produce special tokens that trigger a query to the knowledge base. Our experiments show that our method improves factual grounding in both short and long-form generation, and allows factual revisions to take effect through KB edits rather than parameter updates.
Chinese Translation
我们提出了一种新方法，使得大型语言模型（LLM）在生成标记时能够自动从知识库中提取事实知识。这意味着（1）LLM 输出中的事实知识可以在不重新训练 LLM 的情况下进行更新，（2）LLM 输出中的事实可以追溯到知识库，以确保透明性和可解释性，以及（3）较小的模型可以达到与较大模型相同的事实准确性。我们的核心思想是训练模型生成特殊标记，以触发对知识库的查询。我们的实验表明，该方法在短文本和长文本生成中均提高了事实基础，并允许通过知识库编辑而非参数更新来实现事实修正。
View on arXiv Download PDF AI Translation
cs.AI / 48 / 2606.26816

Computational Analysis of Heart Rate Variability in Healthy Adults

健康成人心率变异性的计算分析
Lado, María J., Méndez, Arturo J., Rodriguez-Liñares, Leandro, Pérez-Schofield, Baltasar García, Cuesta-Morales, Pedro, Iglesias-Otero, Brais, Vila, Xose A.
Abstract
Heart Rate Variability (HRV) analysis is a key indicator of cardiac physiological state and aids in disease diagnosis. However, research on HRV parameters in healthy individuals remains limited, and no gold standard exists. This study evaluates HRV indices in 40 healthy adults (20 men, 20 women, aged 30-50) to improve HRV's clinical utility. Using computational methods for signal processing and data analysis, time, frequency, and nonlinear indices were analyzed to address five questions: (1) normality, (2) stability, (3) correlation, (4) reproducibility, and (5) consistency. Key findings: (1) Time-domain and nonlinear indices, particularly global and LF (low frequency), follow normal distributions, with gender differences noted. (2) Most indices are stable except HF (high frequency)-related ones. (3) High correlations in HF-related indices suggest redundancy, indicating only one is necessary in studies. (4) Comparisons with the Fantasia database revealed less than 10% error for most indices, except SD2 and SDNN in women (greater than 15%). (5) Time-domain and nonlinear indices show low inter-study variability, while frequency-domain indices exhibit high variability, limiting cross-study comparisons. The selected indices-ApEn and IRRR (global variability), HRVi and SD2 (LF), and MADRR or rMSSD (HF)-are best suited for accurately representing HRV components and enhancing its clinical and research relevance.
Chinese Translation
心率变异性（HRV）分析是心脏生理状态的重要指标，有助于疾病诊断。然而，关于健康个体HRV参数的研究仍然有限，且尚无公认的标准。本研究评估了40名健康成人（20名男性，20名女性，年龄30-50岁）的HRV指标，以提高HRV的临床实用性。采用信号处理和数据分析的计算方法，分析了时间域、频率域和非线性指标，以回答五个问题：（1）正常性，（2）稳定性，（3）相关性，（4）重复性，以及（5）一致性。主要发现：（1）时间域和非线性指标，特别是全局和低频（LF）指标，遵循正态分布，并观察到性别差异。（2）大多数指标稳定，除了与高频（HF）相关的指标。（3）HF相关指标之间的高相关性表明冗余，表明在研究中仅需保留一个。（4）与Fantasia数据库的比较显示，大多数指标的误差低于10%，但女性的SD2和SDNN指标的误差超过15%。（5）时间域和非线性指标显示出较低的跨研究变异性，而频率域指标则表现出较高的变异性，限制了跨研究比较。所选指标——近似熵（ApEn）和全局变异性指标（IRRR）、低频（LF）指标HRVi和SD2，以及高频（HF）指标MADRR或rMSSD——最适合准确表示HRV成分，并增强其临床和研究相关性。
View on arXiv Download PDF AI Translation
cs.AI / 49 / 2606.26836

The Capability Frontier: Benchmarks Miss 82% of Model Performance

能力边界：基准测试错过了模型性能的82%
Fowler, Bradley, Smith, Ryan, Graviet, Daniel Thi, Myers, William, Greaves, Joshua, Oozeer, Narmeen Fatimah, García, Antía, Quirke, Philip, Abdullah, Amirali, Barez, Fazl, Upadhyay, Shriyash Kaustubh
Abstract
Existing benchmarks typically report accuracy for a single model on a single run. This systematically understates real-world LLM capabilities, particularly under heterogeneous data distributions: (i) different models get different questions correct according to their specializations, and (ii) given a budget, multiple generations can be sampled and selectively retained. To quantify this gap, we introduce the Capability Frontier: a Pareto frontier over a set of models that characterizes the best achievable performance at each cost level under optimal selection across models and generations (i.e., via an oracle). Our construction corrects for two opposing biases: underestimation from single-model evaluation and overestimation from taking maxima over noisy samples. We study 21 LLMs across 16 widely used benchmarks spanning coding, reasoning, medicine, factuality, instruction following, and agentic tasks, comparing Capability Frontier performance at matched cost to each benchmark's top-performing model. Correcting for single-model evaluation yields a 54% error rate reduction; additionally correcting for single runs yields an 82% improvement, with SOTA accuracy matched at 85% cost reduction. Complementing these empirical results, we use controlled probabilistic simulations to show that higher query topic entropy produces a near-monotonic increase in the performance gap between oracle routing and the best single model. Our findings suggest collective LLM capabilities are substantially underestimated, with implications for evaluation and deployment in data-heterogeneous, multi-domain settings.
Chinese Translation
现有基准测试通常仅报告单个模型在单次运行中的准确性。这在系统上低估了现实世界中大型语言模型（LLM）的能力，特别是在异构数据分布下：（i）不同模型根据其专业化正确回答不同的问题，以及（ii）在给定预算的情况下，可以抽样多个生成并选择性保留。为了量化这一差距，我们引入了能力边界（Capability Frontier）：这是一个关于一组模型的帕累托边界，表征在模型和生成的最佳选择（即通过一个神谕）下，在每个成本水平上可实现的最佳性能。我们的构建纠正了两种相对偏差：来自单模型评估的低估和来自对噪声样本取最大值的高估。我们研究了21个LLM在16个广泛使用的基准测试中的表现，这些基准涵盖了编码、推理、医学、事实性、遵循指令和代理任务，并将能力边界的表现与每个基准的最佳模型在匹配成本下进行比较。纠正单模型评估导致54%的错误率降低；此外，纠正单次运行导致82%的改进，且在85%的成本降低下达到了最先进的准确性（SOTA）。补充这些实证结果，我们使用受控的概率模拟显示，更高的查询主题熵会导致神谕路由与最佳单模型之间的性能差距近乎单调增加。我们的发现表明，集体LLM能力被大幅低估，这对在数据异构、多领域环境中的评估和部署具有重要影响。
View on arXiv Download PDF AI Translation
cs.AI / 50 / 2606.26852

Context-Aware Synthesis of Optimization Pipelines for Warehouse Optimization

基于上下文的仓库优化优化流程合成
Bischoff, Janik, Meyer, Anne, Mohring, Uta, Dunke, Fabian, Barlang, Maximilian, Subas, Özge Nur, Kutabi, Hadi, Nickel, Stefan, Furmans, Kai
Abstract
Order fulfillment in manual picker-to-goods warehouses involves interconnected decisions such as item assignment, order batching, and picker routing. While integrated models capture interactions between these decisions, practical warehouse systems often require decomposed approaches due to organizational boundaries, differing responsibilities, or limited data availability. Existing studies primarily evaluate algorithms for isolated subproblems or fixed subproblem combinations for specific warehouse settings, but lack a general mechanism to determine applicable algorithm configurations, compose them into valid solution pipelines, and assess their performance. With Context-Aware Synthesis of Optimization Pipelines (CASOP), we propose a framework for constructing and evaluating context-specific optimization pipelines and apply these to order fulfillment. The framework comprises: (1) a modular repository of algorithms for common order fulfillment problems; (2) semantic data and algorithm cards to describe warehouse context and algorithm requirements; (3) a taxonomy that structures order fulfillment problems into relevant subproblems; (4) a pipeline synthesizer that identifies applicable algorithms for a given warehouse context and composes all valid optimization pipelines; and (5) a pipeline evaluator that assesses all resulting pipelines. We demonstrate the framework on 7 benchmark instance sets covering four problem classes, resulting in 1,063,044 valid pipelines. The framework supports researchers and practitioners in designing, automatically synthesizing, and selecting valid, high-performing algorithmic pipelines for warehouse operations. The software is open-source and available at https://github.com/kit-dsm/ware_ops_pipes and https://github.com/kit-dsm/ware_ops_algos. Keywords: Warehouse optimization, Algorithm selection, Pipeline synthesis, Order fulfillment
Chinese Translation
在人工拣货到货物的仓库中，订单履行涉及诸如物品分配、订单批次和拣货员路线等相互关联的决策。尽管集成模型能够捕捉这些决策之间的相互作用，但实际的仓库系统通常由于组织边界、不同的职责或有限的数据可用性而需要分解的方法。现有研究主要评估孤立子问题或特定仓库环境下固定子问题组合的算法，但缺乏一种通用机制来确定适用的算法配置、将其组合成有效的解决方案流程，并评估其性能。通过基于上下文的优化流程合成（Context-Aware Synthesis of Optimization Pipelines, CASOP），我们提出了一个构建和评估特定上下文优化流程的框架，并将其应用于订单履行。该框架包括：（1）一个用于常见订单履行问题的模块化算法库；（2）描述仓库上下文和算法需求的语义数据和算法卡片；（3）将订单履行问题结构化为相关子问题的分类法；（4）一个流程合成器，用于识别给定仓库上下文的适用算法并组合所有有效的优化流程；（5）一个流程评估器，用于评估所有生成的流程。我们在涵盖四个问题类别的7个基准实例集上演示了该框架，结果生成了1,063,044个有效流程。该框架支持研究人员和从业者设计、自动合成和选择有效的高性能算法流程以用于仓库操作。该软件是开源的，可在 https://github.com/kit-dsm/ware_ops_pipes 和 https://github.com/kit-dsm/ware_ops_algos 获取。关键词：仓库优化，算法选择，流程合成，订单履行
View on arXiv Download PDF AI Translation
cs.AI / 51 / 2606.26857

LCAi: Life Cycle Assessment with big data fusion and retrieval-augmented generation-assisted interpretation

LCAi：基于大数据融合和检索增强生成辅助解释的生命周期评估
Tsironis, Georgios, Medrano-Garcia, Juan D., Guillen-Gosalbez, Gonzalo
Abstract
The interpretation phase of life cycle assessment often lacks structured mechanisms for translating quantified improvement opportunities addressing environmental hotspots into actionable strategic pathways under technological, social, and policy uncertainty. To overcome this limitation, this study introduces a perspective-conditioned retrieval-augmented generation framework for LCA interpretation, where a multi-perspective retrieval and controlled synthesis is incorporated in the artificial intelligence (AI)-assisted LCA. To operationalise large language models in LCA interpretation, a perspective fusion RAG architecture was developed, covering academic, industry, public discourse, and European union (EU) funding datasets. Our approach comprises three steps: (1) a scenario anchor defining system boundaries and decarbonization targets, (2) a set of perspective-specific micro-queries with constrained retrieval, and (3) a neutral synthesis step integrating only ledger-stored outputs without further retrieval. The framework is demonstrated through a hydrogen-enabled diesel reduction use case in an Italian apple production facility using GPT-5 nano as the reasoning model. Overall, the structured retrieval and constrained synthesis are designed to mitigate the risk of hallucination while preserving cross-domain diversity. The approach presented can support more disciplined translation of impact results into strategic pathways and opens up new avenues for the use of advanced AI tools in LCA studies, particularly those focused on technologies that could be deployed at scale. This proof-of-concept demonstrates how AI-assisted, evidence-grounded interpretation can support implementation-oriented decision-making beyond conventional LCA studies.
Chinese Translation
生命周期评估的解释阶段通常缺乏结构化机制，将针对环境热点的量化改善机会转化为在技术、社会和政策不确定性下可行的战略路径。为克服这一限制，本研究引入了一种基于视角条件的检索增强生成框架用于LCA解释，其中在人工智能（AI）辅助的生命周期评估中融入了多视角检索和受控综合。为在LCA解释中操作大型语言模型，开发了一种视角融合的RAG架构，涵盖学术、行业、公共话语和欧盟（EU）资金数据集。我们的方法包括三个步骤：（1）定义系统边界和脱碳目标的情景锚点；（2）一组具有约束检索的视角特定微查询；（3）一个中立的综合步骤，仅整合账本存储的输出而不进行进一步检索。该框架通过在意大利苹果生产设施中使用氢能柴油减少的案例进行了演示，采用GPT-5 nano作为推理模型。总体而言，结构化检索和受限综合旨在减轻幻觉风险，同时保持跨领域的多样性。所提出的方法可以支持更有纪律性地将影响结果转化为战略路径，并为在LCA研究中使用先进的AI工具开辟了新的途径，特别是那些关注可大规模部署技术的研究。这一概念验证展示了AI辅助、基于证据的解释如何支持超越传统LCA研究的实施导向决策。
View on arXiv Download PDF AI Translation
cs.AI / 52 / 2606.26859

AgentX: Towards Agent-Driven Self-Iteration of Industrial Recommender Systems

AgentX：朝着基于智能体的工业推荐系统自我迭代迈进
Lao, Changxin, Pan, Fei, Ma, Guozhuang, Li, Han, Lin, Huihuang, Shi, Jijun, Zhao, Kangzhi, Gai, Kun, Zhou, Mo, Zhou, Qinqin, Chen, Quan, Yang, Ruochen, Bie, Shifu, Yang, Shuang, Yang, Shuo, Li, Wenhao, Xie, Wentao, Lv, Xiao, Wang, Xuming, Wang, Yijun, Chen, Yiming, Huang, Yusheng, Wang, Zhongyuan, Zhao, Zibo, Zhuang, Zijie, Xia, Baoning, Liu, Chao, Ma, Chaoyi, He, Chubo, Cong, Dawei, Jiang, Feng, Wang, Gang, Xia, Guilin, Xu, Hanwen, Xie, Jiahong, Qiao, Jiahui, Liang, Jian, Yue, Jiangfan, Wang, Jing, Yang, Jinghan, Jia, Jinghui, Qin, Kan, Wang, Lei, Li, Ming, Song, Peilin, Xu, Pengbo, Luo, Qiang, Tang, Ruiming, Liu, Shiyang, Jin, Shuxian, Wang, Tao, Zhang, Tao, Gao, Xiang, Li, Xianghan, Luo, Yingsong, Ning, Yiwen, Liu, Yongcheng, Guo, Yuan, Liu, Zhaojie, Cui, Zhenkai
Abstract
Recommendation algorithm iteration is moving from an artisanal, engineer-bound process toward an industrialized research loop, but this transition remains blocked by a structural execution bottleneck: the idea-to-launch cycle still depends on human engineers to generate hypotheses, modify production code, launch A/B experiments, and attribute online results. Innovation therefore scales linearly with headcount rather than compounding with evidence, compute, and accumulated experimental knowledge. We present AgentX, a production-deployed multi-agent system that fundamentally restructures this production function. AgentX operates as a self-evolving development engine: it autonomously generates, implements, evaluates, and learns from recommendation experiments at a scale and pace that no manual workflow can sustain. The system orchestrates four tightly coupled stages in a closed loop. A Brainstorm Agent synthesizes evidence from historical experiments, system architecture, data analysis, and external research into ranked, executable proposals. A Developing Agent translates each proposal into production-ready code through repository-grounded generation and multi-dimensional reliability verification. An Evaluation Agent conducts safe online rollout with guardrail-vetoed A/B judgment, converting both successes and failures into structured knowledge assets. A Harness Evolution layer (SGPO) then distills execution trajectories into semantic-gradient updates that continuously sharpen the agents themselves -- making the system not merely automated, but self-improving.
Chinese Translation
推荐算法的迭代正在从一种手工、依赖工程师的过程转向一种工业化的研究循环，但这一转变仍然受到结构性执行瓶颈的阻碍：从构思到上线的周期仍然依赖人类工程师生成假设、修改生产代码、启动 A/B 实验并归因在线结果。因此，创新的规模与人力资源呈线性关系，而不是与证据、计算能力和积累的实验知识呈复合关系。我们提出了 AgentX，这是一种已投入生产的多智能体系统，根本上重构了这一生产功能。AgentX 作为一个自我演化的开发引擎运作：它自主生成、实施、评估并从推荐实验中学习，以一种任何手动工作流程都无法维持的规模和速度进行。该系统在一个闭环中协调四个紧密耦合的阶段。一个头脑风暴智能体从历史实验、系统架构、数据分析和外部研究中综合证据，形成排名的可执行提案。一个开发智能体通过基于代码库的生成和多维可靠性验证将每个提案转化为生产就绪的代码。一个评估智能体进行安全的在线发布，采用有保护措施的 A/B 判断，将成功与失败转化为结构化的知识资产。随后，一个 Harness Evolution 层（SGPO）将执行轨迹提炼为语义梯度更新，持续提升智能体自身的能力——使系统不仅仅是自动化的，而是自我改进的。
View on arXiv Download PDF AI Translation
cs.AI / 53 / 2606.26874

TAVR-VLM: Risk-Conditioned Causal Grounding for Hallucination-Resistant Report Generation

TAVR-VLM：基于风险条件的因果基础构建用于抗幻觉报告生成
Lu, Zhixiang, Liu, Xiwei, Song, Sifan, Ji, Changkai, Nguyen, Anh, Su, Jionglong, Razzak, Imran, Wang, Jinfeng
Abstract
Transcatheter Aortic Valve Replacement (TAVR) planning requires meticulous multimodal reasoning. However, adapting Multimodal Large Language Models (MLLMs) to this high-stakes domain is severely impeded by diagnostic hallucinations, where generated text lacks anatomical grounding. To address this, TAVR-VLM is introduced: a novel framework featuring Risk-Conditioned Causal Grounding Attention (R-CGA) that instantiates a model-internal ``Risk $\rightarrow$ Region $\rightarrow$ Word'' structural grounding pathway. R-CGA compresses multimodal inputs into a causal risk bottleneck, purifying dense visual features into a global risk mask. During autoregressive generation, a support-projected causal consistency objective constrains token-level grounding within the risk-defined support mask. Evaluated on $\text{M}^3\text{TAVR}$, a comprehensive 1,482-patient cohort, TAVR-VLM establishes a new state-of-the-art. It achieves an AUROC of 0.896, boosts CIDEr to 0.936, and drastically reduces the hallucination rate to 8.1\%, thereby improving interpretability for evidence-based surgical AI.
Chinese Translation
经导管主动脉瓣置换术（TAVR）规划需要细致的多模态推理。然而，将多模态大型语言模型（MLLMs）适应于这一高风险领域受到诊断幻觉的严重阻碍，即生成的文本缺乏解剖基础。为了解决这一问题，本文提出了TAVR-VLM：一个新颖的框架，具有风险条件因果基础注意力（R-CGA），该框架实现了模型内部的“风险 $ ightarrow$ 区域 $ ightarrow$ 词”结构基础路径。R-CGA将多模态输入压缩为因果风险瓶颈，将密集的视觉特征净化为全局风险掩码。在自回归生成过程中，支持投影的因果一致性目标限制了在风险定义的支持掩码内的标记级基础。经过对1,482名患者的综合队列$ ext{M}^3 ext{TAVR}$的评估，TAVR-VLM建立了新的最先进水平。它实现了0.896的AUROC，将CIDEr提升至0.936，并将幻觉率大幅降低至8.1%，从而提高了基于证据的外科人工智能的可解释性。
View on arXiv Download PDF AI Translation
cs.AI / 54 / 2606.26879

A Pipeline for Generating Longitudinal Synthetic Clinical Notes Using Large Language Models

使用大型语言模型生成纵向合成临床笔记的管道
Poulett, William
Abstract
Synthetic data is increasingly used to enable the development and evaluation of AI systems in domains where access to real-world data is restricted. In healthcare, clinical documentation presents particular challenges due to its sensitivity. This work introduces a synthetic clinical notes pipeline and dataset designed to support the development of clinical AI tools while avoiding the privacy risks associated with real patient data. The dataset is generated using a modular pipeline that combines structured patient generation, semi-structured patient journey simulation, and unstructured clinical note generation using large language models. The pipeline is designed to prioritise internal consistency across longitudinal patient records, while also capturing variation in writing style, note structure, and clinical detail. Additional mechanisms, including LLM-based validation and augmentation steps, are used to improve faithfulness, realism, and diversity of the generated notes. We release a dataset of 70 synthetic patients, each associated with 20-50 clinical notes spanning a full hospital journey. The dataset is provided at multiple levels of validation, enabling users to balance realism and scalability depending on their use case. This dataset supports the development, testing, and evaluation of clinical AI systems, including summarisation tools, coding models, and decision support systems, without reliance on real patient data.
Chinese Translation
合成数据越来越多地用于支持在现实数据获取受限的领域中开发和评估人工智能系统。在医疗保健领域，临床文档由于其敏感性而面临特别的挑战。本研究介绍了一种合成临床笔记管道和数据集，旨在支持临床人工智能工具的开发，同时避免与真实患者数据相关的隐私风险。该数据集是通过一个模块化管道生成的，该管道结合了结构化患者生成、半结构化患者旅程模拟和使用大型语言模型生成非结构化临床笔记。该管道旨在优先考虑纵向患者记录之间的内部一致性，同时捕捉写作风格、笔记结构和临床细节的变化。还使用包括基于LLM的验证和增强步骤在内的附加机制，以提高生成笔记的真实性、现实性和多样性。我们发布了一个包含70名合成患者的数据集，每名患者关联20-50条涵盖完整医院旅程的临床笔记。该数据集提供了多个验证级别，使用户能够根据其使用案例平衡现实性和可扩展性。该数据集支持临床人工智能系统的开发、测试和评估，包括摘要工具、编码模型和决策支持系统，而无需依赖真实患者数据。
View on arXiv Download PDF AI Translation
cs.AI / 55 / 2606.26899

Generative Retrieval via Diffusion Transformer with Metric-Ordered Sequence Training and Hybrid-Policy Preference Optimization

通过扩散变换器实现的生成检索：度量有序序列训练与混合策略偏好优化
Liu, Chenghao, Zhang, Yu, Jiang, Zhongtao, Xu, Kun, An, Zhenwei, Wang, Renzhi, Wang, Zhao, Zhang, Jiachen, Zhang, Yuxiao, Xu, Kun, Huang, Songfang
Abstract
Embedding-based retrieval ranks items by their similarity to a query in a shared vector space and usually aims to return the highest-scoring items. In many production settings this is not what is wanted: given a seed set that expresses a fine-grained pattern, one needs more items that both satisfy a target attribute and stay within that pattern. We formalize this as pattern-preserving attribute retrieval. The two goals pull against each other: averaging the seeds preserves the pattern but stays in a low-attribute region, while global attribute retrieval drifts to unrelated patterns. We approach the task with continuous generative retrieval, where a model reads a sequence of item embeddings and generates query embeddings for nearest-neighbor search. We propose MO-DiT+HPPO, a staged framework with raw-sequence pretraining, multi-domain metric-ordered continuation pretraining, tail-centroid fine-tuning, and HPPO. Metric-ordered training turns sparse online retrieval labels into in-pattern trajectories ordered from low to high predicted attribute density, teaching one model the metric-improvement direction across domains. HPPO aligns the generated query distribution with the true online objective by labeling a hybrid candidate pool with the online intersection metric and applying reference-anchored preference optimization. A Pareto pair filter keeps only winner pairs that do not lower same-pattern purity, raising the attribute metric without sacrificing the pattern. Across four attribute domains under item- and pattern-holdout protocols, metric-ordered DiT improves the intersection metric over a pretrained generative retriever, and HPPO improves it further, with significant gains on seven of eight domain-split cells and a marginal tie on the hardest split. Metric-predictor validation, order ablations, CPT/SFT comparisons, and a candidate-policy ablation show where the gains come from.
Chinese Translation
基于嵌入的检索通过在共享向量空间中根据与查询的相似性对项目进行排名，通常旨在返回得分最高的项目。然而，在许多生产环境中，这并不是所期望的结果：给定一个表达细粒度模式的种子集，需要更多既满足目标属性又符合该模式的项目。我们将其形式化为模式保持属性检索。这两个目标相互牵制：平均化种子保持了模式，但停留在低属性区域，而全局属性检索则偏离了无关模式。我们采用连续生成检索的方法，其中模型读取一系列项目嵌入并生成用于最近邻搜索的查询嵌入。我们提出了MO-DiT+HPPO，这是一个分阶段框架，包含原始序列预训练、多域度量有序继续预训练、尾部中心微调和HPPO。度量有序训练将稀疏的在线检索标签转变为从低到高预测属性密度的模式内轨迹，教会一个模型跨域的度量改进方向。HPPO通过使用在线交集度量标记混合候选池，并应用参考锚定的偏好优化，将生成的查询分布与真实的在线目标对齐。Pareto配对过滤器仅保留不会降低同模式纯度的赢家对，提高属性度量而不牺牲模式。在项目和模式保留协议下的四个属性域中，度量有序的DiT在预训练的生成检索器上提高了交集度量，而HPPO进一步改善了该度量，在八个域分割单元中的七个上取得了显著提升，在最难的分割上则是微弱平局。度量预测器验证、顺序消融、CPT/SFT比较以及候选策略消融展示了这些提升的来源。
View on arXiv Download PDF AI Translation
cs.AI / 56 / 2606.26902

Learning to Recover Task Experts from a Multi-Task Merged Model

从多任务合并模型中学习恢复任务专家
Jung, Jinwook, Kim, Taegyu, Jo, Kumju, Baik, Sungyong
Abstract
Multi-task model merging aims to consolidate several task-specific experts into a unified model, yet static merging consistently suffers from parameter interference. While dynamic merging models aim to bridge this gap, many works rely on the costly storage and loading of redundant expert components at inference. In this work, from the perspective of task expert, we view parameter interference as parameter perturbation introduced to each expert during merging process. We show that such parameter perturbations can be modeled as affine transformation, which can be approximated as additive offsets. Motivated by these, we propose Recover Task eXpert (ReTeX), a framework that predicts those offsets, in order to undo parameter interference and recover task-expert performance from a single merged checkpoint. To recover the appropriate expert when task identity is unknown, we introduce a router-free task identifier based on SVD subspace signatures computed offline before inference. At inference, the identifier selects the task whose subspace yields the smallest projection residual for a given input. As a result, ReTeX recovers over 95% of individual-expert performance in both vision and NLP domains, while significantly improving generalization to unseen tasks. Crucially, we also show that the parameter offset prediction leads to emergent adaptive interpolation of expert knowledge for out-of-distribution (OOD) tasks. ReTeX adaptively interpolates seen expert knowledge to handle unseen tasks. Our code is available at https://github.com/BAIKLAB/ReTeX
Chinese Translation
多任务模型合并旨在将多个特定任务的专家整合为一个统一的模型，但静态合并始终受到参数干扰的影响。尽管动态合并模型旨在弥补这一差距，但许多研究依赖于在推理时存储和加载冗余专家组件的高成本。在本研究中，从任务专家的角度来看，我们将参数干扰视为在合并过程中引入到每个专家的参数扰动。我们表明，这种参数扰动可以建模为仿射变换，并且可以近似为加性偏移。基于此，我们提出了恢复任务专家（Recover Task eXpert，ReTeX）框架，该框架预测这些偏移，以消除参数干扰并从单个合并检查点恢复任务专家性能。为了在任务身份未知时恢复适当的专家，我们引入了一种基于在推理前离线计算的奇异值分解（SVD）子空间特征的无路由任务标识符。在推理时，该标识符选择子空间对于给定输入产生最小投影残差的任务。因此，ReTeX在视觉和自然语言处理（NLP）领域恢复了超过95%的个体专家性能，同时显著提高了对未见任务的泛化能力。重要的是，我们还表明，参数偏移预测导致了对分布外（OOD）任务的专家知识的自适应插值。ReTeX自适应地插值已见专家知识以处理未见任务。我们的代码可在 https://github.com/BAIKLAB/ReTeX 获取。
View on arXiv Download PDF AI Translation
cs.AI / 57 / 2606.26918

Diagnosing Task Insensitivity in Language Agents

语言代理中的任务不敏感性诊断
Liu, Jingyu, Wu, Xiaopeng, Chen, Kehan, Yu, Chuan, Liu, Yong
Abstract
Large language models can serve as capable long-horizon agents, but their out-of-distribution (OOD) generalization remains weak. We identify a key source of this failure as task insensitivity: when faced with similar but distinct tasks, models might apply patterns learned during training and fail to solve the task at hand. We show that models often continue with actions aligned with the original task even when the instruction is semantically corrupted and cannot be directly answered. We further find that, when we replace the task description in a trained prompt with another similar but distinct task, the model may still output the same action. This behavior is accompanied by a consistent training-time attention drift away from task tokens and toward local observations, suggesting an optimization bias toward shortcuts. To mitigate this problem, we propose Task-Perturbed NLL Optimization, a lightweight contrastive regularizer that explicitly encourages action dependence on the task instruction. Extensive evaluations show that our intervention improves task sensitivity and OOD generalization while preserving more stable attention to task tokens.
Chinese Translation
大型语言模型可以作为能够处理长时间跨度的代理，但它们在分布外（OOD）泛化方面仍然较弱。我们确定这一失败的一个关键来源是任务不敏感性：当面临相似但不同的任务时，模型可能会应用在训练中学到的模式，从而无法解决当前任务。我们展示了模型在指令语义被破坏且无法直接回答时，仍然继续执行与原始任务一致的动作。我们进一步发现，当我们用另一个相似但不同的任务替换经过训练的提示中的任务描述时，模型可能仍然输出相同的动作。这种行为伴随着训练期间注意力的持续偏移，从任务标记向局部观察转移，表明存在向捷径的优化偏见。为了缓解这个问题，我们提出了任务扰动的负对数似然优化（Task-Perturbed NLL Optimization），这是一种轻量级的对比正则化器，明确鼓励动作依赖于任务指令。大量评估表明，我们的干预措施提高了任务敏感性和OOD泛化，同时保持了对任务标记的更稳定的注意力。
View on arXiv Download PDF AI Translation
cs.AI / 58 / 2606.26935

Where Do CoT Training Gains Land in LLM based Agents?

链式思维训练的收益在基于大型语言模型的智能体中落在哪里？
Liu, Jingyu, Wang, Zhiwen, Jing, Yuxin, Zhou, Huanyu, Liu, Yong
Abstract
Chain-of-thought (CoT) reasoning is widely used in language-model agents, but prior work has shown that verbalized CoT is not always faithful and may instead reflect post-hoc reasoning, which means the model already knows the answer before reasoning. We therefore ask what CoT training is actually improving: is the model getting better at changing its action through generated reasoning, or is it getting better at predicting the action directly from the prompt? We study this question by comparing \emph{prompt actions} (predicting action without CoT) with CoT actions (predicting action with CoT). Across checkpoints, prompt-action quality improves substantially. While interacting with the environment, the relative advantage of CoT actions over prompt actions remains similar, showing that CoT training does not widen the advantage of CoT reasoning, and it helps to improve the quality of prompt actions. We further find that later checkpoints are less likely to revise the action in response to CoT, suggesting greater reliance on the prompt. Motivated by these patterns, we selectively mask action-token supervision on a fraction of training examples. This intervention improves out-of-domain generalization.
Chinese Translation
链式思维（CoT）推理在语言模型智能体中被广泛使用，但先前的研究表明，口头化的链式思维并不总是忠实的，可能反映的是事后推理，这意味着模型在推理之前已经知道答案。因此，我们提出了一个问题：链式思维训练究竟改善了什么？模型是在通过生成的推理改变其行动方面变得更好，还是在直接从提示中预测行动方面变得更好？我们通过比较 extit{提示行动}（在没有链式思维的情况下预测行动）与链式思维行动（在有链式思维的情况下预测行动）来研究这个问题。在各个检查点中，提示行动的质量显著提高。在与环境互动时，链式思维行动相对于提示行动的相对优势保持相似，这表明链式思维训练并没有扩大链式思维推理的优势，而是有助于提高提示行动的质量。我们进一步发现，后期检查点在响应链式思维时不太可能修正行动，表明对提示的依赖性更大。基于这些模式，我们选择性地对一部分训练样本进行行动标记监督的屏蔽。这一干预改善了域外泛化能力。
View on arXiv Download PDF AI Translation
cs.AI / 59 / 2606.26964

Look-Before-Move: Narrative-Grounded World Visual Attention in Dynamic 3D Story Worlds

先看后动：动态三维故事世界中的叙事基础视觉注意力
Bian, Jiaming, Li, Bingliang, Wu, Yuehao, Wang, Pichao, Wang, Zhi, Ma, Hailan, Mo, Huadong, Sun, Zhenhong
Abstract
As embodied AI and world models increasingly operate in dynamic 3D environments, visual perception must move beyond passively interpreting given observations toward actively deciding what to observe. We study this problem through camera planning in dynamic 3D story worlds, where the camera must not only generate smooth motion, but also decide what visual evidence should be acquired before it moves. We formulate this capability as Narrative-Grounded World Visual Attention, where the camera acts as an embodied observer that determines what to observe, how to compose the observation, and how to shift attention over time under narrative intent and physical 3D constraints. To realize this capability, we propose Look-Before-Move, a camera planning framework that separates observation specification from motion execution. It first builds a Semantic Observation Contract to convert directorial intent into executable visual constraints, then performs Monte Carlo Viewpoint Search to find narrative-compliant and geometrically feasible viewpoints, and finally applies Semantic Trajectory Grounding to connect selected viewpoints into continuous, collision-aware, and temporally coherent camera motion. We further construct a dynamic 3D Story World Benchmark based on StoryBlender, covering 50 stories, 457 scenes, and 1585 shots with animated characters, semantic scene configurations, and executable 3D environments. Experiments show that our framework improves subject perception, intent consistency, and trajectory quality over representative baselines, demonstrating the importance of organizing visual attention before generating camera motion.
Chinese Translation
随着具身人工智能和世界模型在动态三维环境中越来越多地运作，视觉感知必须超越被动解释给定观察，转向主动决定观察内容。我们通过动态三维故事世界中的相机规划研究这一问题，在这些环境中，相机不仅需要生成平滑的运动，还必须决定在移动之前应获取哪些视觉证据。我们将这一能力表述为叙事基础视觉注意力（Narrative-Grounded World Visual Attention），其中相机作为具身观察者，决定观察内容、如何构成观察以及如何在叙事意图和物理三维约束下随时间变化注意力。为了实现这一能力，我们提出了“先看后动”（Look-Before-Move），一个将观察规范与运动执行分离的相机规划框架。该框架首先构建语义观察契约（Semantic Observation Contract），将导演意图转化为可执行的视觉约束，然后执行蒙特卡洛视点搜索（Monte Carlo Viewpoint Search），寻找符合叙事要求和几何可行的视点，最后应用语义轨迹基础（Semantic Trajectory Grounding）将选定的视点连接成连续的、注意碰撞的和时间一致的相机运动。我们进一步基于StoryBlender构建了一个动态三维故事世界基准，涵盖50个故事、457个场景和1585个镜头，包含动画角色、语义场景配置和可执行的三维环境。实验表明，我们的框架在受试者感知、意图一致性和轨迹质量方面优于代表性基线，展示了在生成相机运动之前组织视觉注意力的重要性。
View on arXiv Download PDF AI Translation
cs.AI / 60 / 2606.26969

Einstein World Models

爱因斯坦世界模型
Nwadike, Munachiso Samuel, Iklassov, Zangir, Mekky, Ali, Zuhri, Zayd M. Kawakibi, Inui, Kentaro
Abstract
Does intelligence require the ability to reason about phenomena beyond direct experience? It is natural to suspect that some complex thought cannot be captured through language alone. However, of particular concern to this work, is whether visualising counterfactual events can complement language as a mechanism for complex thought. We ask whether LLMs can be trained to utilise such visualisation mechanisms, in a way that benefits their reasoning abilities. Motivated by this question, we propose Einstein World Models. EWMs are a blueprint for LLM-based reasoning systems that place visual-temporal rollouts inside the reasoning trace, allowing them to reason in ways that text alone may not support well. In an EWM, the LLM calls a world-module (not to be confused with a world model), to produce short rollouts of scenes under consideration. The returned rollout is treated not as the answer, but as an inspectable hypothesis that can support later reasoning. Einstein World Models extend the capability of LLMs for tool calling (such as web search or code execution), into the domain of visual thought experiments.
Chinese Translation
智能是否需要对超出直接经验的现象进行推理的能力？人们自然会怀疑某些复杂的思维是否仅通过语言无法表达。然而，本研究特别关注的是，是否可视化反事实事件能够作为一种复杂思维的机制，补充语言的作用。我们探讨了大规模语言模型（LLMs）是否可以被训练以利用这种可视化机制，从而提升其推理能力。基于这个问题，我们提出了爱因斯坦世界模型（Einstein World Models，EWMs）。EWMs 是一种基于 LLM 的推理系统蓝图，将视觉-时间展开（visual-temporal rollouts）置于推理轨迹中，使其能够以文本单独可能无法很好支持的方式进行推理。在 EWM 中，LLM 调用一个世界模块（world-module，注意与世界模型（world model）区分），以生成所考虑场景的短期展开。返回的展开被视为可检查的假设，而不是答案，能够支持后续的推理。爱因斯坦世界模型扩展了 LLM 在工具调用（如网络搜索或代码执行）方面的能力，进入了视觉思维实验的领域。
View on arXiv Download PDF AI Translation
cs.AI / 61 / 2606.27005

Adaptive Utility driven Resource Orchestration for Resilient AI (AURORA-AI)

基于适应性效用驱动的资源调度框架以实现弹性人工智能 (AURORA-AI)
Mhapsekar, Rahul Umesh, Cherkaoui, Ilias, Abraham, Lizy, Dey, Indrakshi
Abstract
Modern AI systems are increasingly deployed under non-stationary computational, demographic, and operational conditions in which static resource allocation strategies degrade both predictive performance and human-centric properties such as fairness and explainability. This paper presents AURORA-AI, an Adaptive Utility-driven Resource Orchestration framework for Resilient AI that unifies Hamilton-Jacobi-Bellman feedback control, Lyapunov-based stability monitoring, and a fairness-aware composite utility into a single closed-loop policy.The framework continuously redistributes computational budget across a population of heterogeneous AI models so that the global utility, defined jointly over predictive performance, demographic parity, cost, latency, robustness, and interpretability, remains maximised under disruption. The framework is evaluated in a stress-rich discrete-time simulation that concurrently injects demographic bias shocks, gradual concept drift, and abrupt black-swan disruptions, and is compared against five established controllers including Static, Round Robin, Greedy, LinUCB, and a deep reinforcement-learning agent based on Proximal Policy Optimisation. AURORA-AI achieves immediate recovery from the black-swan event compared to eighty-eight time steps for the Static baseline and twenty-two for Proximal Policy Optimisation, lifts the alpha-quantile and the super-quantile by twenty-nine and twenty-five percent respectively, simultaneously reduces the mean and maximum demographic parity gap, and increases the fraction of Lyapunov-stable operating steps. These results indicate that fairness-aware adaptive orchestration grounded in stability theory is a practical and theoretically motivated path toward resilient human-centric AI deployment.
Chinese Translation
现代人工智能系统越来越多地在非平稳的计算、人口和操作条件下部署，在这些条件下，静态资源分配策略会降低预测性能以及公平性和可解释性等以人为本的特性。本文提出了AURORA-AI，一个基于适应性效用驱动的资源调度框架，用于实现弹性人工智能，该框架将Hamilton-Jacobi-Bellman反馈控制、基于Lyapunov的稳定性监测和关注公平性的复合效用统一为一个闭环策略。该框架持续在异构人工智能模型的群体中重新分配计算预算，以确保在干扰下，定义为预测性能、人口平等、成本、延迟、鲁棒性和可解释性等因素的全球效用保持最大化。该框架在一个压力丰富的离散时间仿真中进行评估，该仿真同时注入人口偏见冲击、渐进概念漂移和突发黑天鹅干扰，并与五个已建立的控制器进行比较，包括静态控制、轮询控制、贪婪控制、LinUCB和基于近端策略优化的深度强化学习代理。与静态基线的八十八个时间步和近端策略优化的二十二个时间步相比，AURORA-AI实现了对黑天鹅事件的即时恢复，分别提升了α分位数和超分位数29%和25%，同时减少了平均和最大人口平等差距，并增加了Lyapunov稳定操作步骤的比例。这些结果表明，基于稳定性理论的关注公平性的适应性调度是实现弹性以人为本的人工智能部署的一个切实可行且理论上有依据的路径。
View on arXiv Download PDF AI Translation
cs.AI / 62 / 2606.27009

Semantic Early-Stopping for Iterative LLM Agent Loops

迭代大型语言模型代理循环的语义提前停止
Shrivastava, Sahil
Abstract
Multi-agent large language model (LLM) loops, for example a Writer that drafts and a Critic that revises, are almost always terminated by a fixed iteration cap (max_iterations). This is a syntactic kill-switch: it is blind to whether the answer is still improving, so it over-spends tokens on easy inputs and truncates hard ones. We study semantic early-stopping: the loop halts when consecutive draft embeddings stop changing in meaning (cosine distance with a patience window) and the answer's measured quality stops improving. Our work makes three contributions. First, an honest theoretical footing: we prove deterministic termination and well-definedness and machine-check these claims, while treating the convergence of the distance sequence as an empirically tested conjecture rather than a (previously over-claimed) Banach contraction. Second, a judge-efficient evaluation protocol: we generate each question's full trajectory once, replay every stopping policy over the identical drafts, and cache every LLM-judge call, yielding a strictly paired efficiency-versus-quality comparison at low cost; we further separate operational tokens (charged to a policy) from evaluation tokens (a measurement instrument). Third, an empirical study on multi-hop retrieval-augmented question answering (HotpotQA). On the 60-question test split, a judge-free semantic stopper reduces operational tokens by 38% relative to max_iterations at parity quality (Delta-IS = -0.004, p = 0.81), whereas the full quality-gated variant is counter-productive because its per-round judging dominates cost. An oracle that selects the best round attains +0.115 Information Score over every practical policy (p ~ 4e-11), reframing the problem from "when to stop" (easy) to "which round is best" (open).
Chinese Translation
多代理大型语言模型（LLM）循环，例如一个撰写者（Writer）起草和一个评论者（Critic）修订，几乎总是通过固定的迭代上限（max_iterations）终止。这是一种语法上的杀死开关：它对答案是否仍在改善视而不见，因此在简单输入上过度消耗令牌，而在困难输入上则被截断。我们研究语义提前停止：当连续草稿嵌入的意义停止变化（在耐心窗口内的余弦距离）且答案的测量质量停止改善时，循环停止。我们的工作有三个贡献。首先，诚实的理论基础：我们证明了确定性终止和良定义性，并对这些主张进行了机器验证，同时将距离序列的收敛视为经过实证检验的猜想，而不是（之前被过度声称的）Banach收缩。其次，评估协议的高效性：我们为每个问题生成完整的轨迹一次，重放每个停止策略在相同草稿上，并缓存每次LLM-评估调用，从而以低成本实现严格配对的效率与质量比较；我们进一步将操作令牌（计入策略）与评估令牌（测量工具）分开。第三，关于多跳检索增强问答（HotpotQA）的实证研究。在60个问题的测试分割中，无评估者的语义停止器在质量相当的情况下，相对于max_iterations减少了38%的操作令牌（Delta-IS = -0.004，p = 0.81），而完全质量门控的变体则适得其反，因为其每轮评估的成本占主导地位。一个选择最佳轮次的神谕在每个实际策略上获得了+0.115的信息分数（p ~ 4e-11），将问题重新框定为“何时停止”（简单）到“哪个轮次最好”（开放）。
View on arXiv Download PDF AI Translation
cs.AI / 63 / 2606.27061

How to evaluate clustering with ground truth?

如何评估具有真实标签的聚类？
Fränti, Pasi
Abstract
External indexes can be used for cluster evaluation when ground truth is available. We review the most common external validity indexes focusing on set-matching-based measures. We recommend centroid index (CI), because it is an intuitive cluster-level measure with an explainable result. If we need a more fine-tuned, point-level measure, there are more choices. Pair-set index (PSI) provides a normalized score which is not biased by cluster sizes. If all points should matter equally, then clustering accuracy (ACC) or any other set-matching measure is suitable.
Chinese Translation
当真实标签可用时，可以使用外部指标进行聚类评估。我们回顾了最常见的外部有效性指标，重点关注基于集合匹配的度量。我们推荐质心指数（Centroid Index, CI），因为它是一种直观的聚类级别度量，结果易于解释。如果需要更精细的点级度量，还有更多选择。配对集合指数（Pair-set Index, PSI）提供了一个标准化的得分，不受聚类大小的影响。如果所有点都应同等重要，则聚类准确率（Clustering Accuracy, ACC）或任何其他集合匹配度量都是合适的。
View on arXiv Download PDF AI Translation
cs.AI / 64 / 2606.27136

Joint Learning of Experiential Rules and Policies for Large Language Model Agents

大语言模型代理的经验规则与策略的联合学习
Ye, Shicheng, Yu, Chao
Abstract
For LLM agents in multi-step interactive environments, a key challenge is to make effective use of accumulated interaction experience. Existing work has typically separated two uses of such experience: keeping it outside the model as natural-language rules for later prompting, or using trajectories and feedback to update the model parameters. The former is easy to interpret but can fall out of sync with the evolving policy; the latter improves the policy more broadly but provides only limited correction for local mistakes in sparse-reward settings. We present Joint Learning of Experiential Rules and Policies for LLM Agents (JERP), which updates a long-term experiential-rule pool and the policy from the same interaction trajectories. At decision time, JERP retrieves task-relevant rules and conditions the agent on them together with the interaction history. After each episode, it uses the collected trajectories both to optimize the policy and to revise the rule pool by comparing current rollouts with reference successful trajectories. This coupling keeps the rule pool aligned with the evolving policy while allowing stable and effective behaviors to be gradually absorbed into the model itself. Experiments on AlfWorld and WebShop show that JERP yields consistent gains in decision performance for complex interactive tasks.
Chinese Translation
在多步骤交互环境中的大语言模型（LLM）代理面临的一个关键挑战是有效利用积累的交互经验。现有研究通常将这种经验的两种用途分开：将其作为自然语言规则保留在模型之外以便后续提示，或使用轨迹和反馈来更新模型参数。前者易于解释，但可能与不断演变的策略不同步；后者更广泛地改善策略，但在稀疏奖励设置中仅能对局部错误提供有限的修正。我们提出了大语言模型代理的经验规则与策略的联合学习（Joint Learning of Experiential Rules and Policies for LLM Agents, JERP），该方法从相同的交互轨迹中更新长期经验规则池和策略。在决策时，JERP检索与任务相关的规则，并将其与交互历史一起作为条件提供给代理。在每个回合结束后，它利用收集到的轨迹来优化策略，并通过将当前的展开与参考成功轨迹进行比较来修订规则池。这种耦合保持了规则池与不断演变的策略的一致性，同时允许稳定和有效的行为逐渐被吸收到模型中。在AlfWorld和WebShop上的实验表明，JERP在复杂交互任务的决策性能上带来了持续的提升。
View on arXiv Download PDF AI Translation
cs.AI / 65 / 2606.27154

OpenRCA 2.0: From Outcome Labels to Causal Process Supervision

OpenRCA 2.0：从结果标签到因果过程监督
Fang, Aoyang, Yang, Yifan, Shang, Jin'ao, Lu, Qisheng, Xu, Junjielung, Wang, Rui, Zhang, Songhan, Zhang, Yuzhong, Yu, Boxi, He, Pinjia
Abstract
Root cause analysis (RCA) poses a holistic test of LLM agentic capabilities, such as long-context understanding, multi-step reasoning, and tool use. However, existing datasets suffer from a fundamental gap: they label only the root cause, not the propagation path connecting it to the observed symptom, which largely simplifies the task to naive pattern matching. To support rigorous evaluation, we introduce PAVE, a step-wise labeling protocol that leverages known interventions from fault injection to reconstruct causal propagation paths. The mechanism is forward verification: reasoning from cause to effect rather than inferring backward from symptoms. Applying PAVE yields OpenRCA 2.0 (500 instances), the first cross-system RCA benchmark with step-wise causal annotations for LLM agents. Across 11 frontier LLMs, recovering the exact root-cause set succeeds in only 20.7% of cases on average. To locate where this difficulty lies, we relax the criterion and find what we call the ungrounded diagnosis: agents identify at least one correct root-cause service in 76.0% of cases, but ground that service in a verified causal propagation path to the observed symptom in only 61.5%. Outcome-only evaluation hides this failure mode; step-wise causal ground truth is the missing piece for trustworthy LLM-based RCA agents.
Chinese Translation
根本原因分析（RCA）对大型语言模型（LLM）的自主能力提出了全面的测试，例如长上下文理解、多步骤推理和工具使用。然而，现有数据集存在一个根本性缺陷：它们仅标注根本原因，而没有标注将其与观察到的症状连接的传播路径，这在很大程度上简化了任务为简单的模式匹配。为了支持严格的评估，我们引入了PAVE，这是一种逐步标注协议，利用已知的故障注入干预措施来重建因果传播路径。该机制是前向验证：从原因推理到结果，而不是从症状向后推断。应用PAVE产生了OpenRCA 2.0（500个实例），这是第一个具有逐步因果注释的跨系统RCA基准，用于LLM代理。在11个前沿LLM中，准确恢复根本原因集的成功率平均仅为20.7%。为了找出这一困难的所在，我们放宽了标准，发现了我们称之为无基础诊断的现象：代理在76.0%的案例中识别出至少一个正确的根本原因服务，但仅在61.5%的情况下将该服务与观察到的症状的验证因果传播路径相结合。仅基于结果的评估掩盖了这一失败模式；逐步因果真实值是可信赖的基于LLM的RCA代理所缺失的关键部分。
View on arXiv Download PDF AI Translation
cs.AI / 66 / 2606.27161

TOPS: First-Principles Visual Token Pruning via Constructing Token Optimal Preservation Sets for Efficient MLLM Inference

TOPS：通过构建令牌最优保留集实现的第一性原理视觉令牌剪枝，以提高多模态大语言模型推理的效率
Wang, Tinghao, Guo, Yichen, Huang, Rui, Lu, Zheng, Zhang, Qizhe, Li, Chenxi, Zhang, Yuan, Cao, Jiajun, Shen, Zhirong, Du, Yaosong, Gan, Guangyan, Wang, Wenya, Cong, Lin William, Zhang, Shanghang
Abstract
Multimodal large language models (MLLMs) have achieved strong multimodal reasoning capabilities, but their efficiency is limited by the large number of visual tokens, which introduces substantial computational overhead. Visual token pruning offers a natural solution, yet existing methods are imperfect: attention-based criteria tend to retain redundant tokens, while diversity-based criteria are often agnostic to user instructions. Even methods that combine multiple criteria still lack a principled formulation of the intrinsic objective of token pruning. In this paper, we revisit visual token pruning from a first-principles perspective and formulate it as constructing Token Optimal Preservation Sets. Through a top-down information-theoretic analysis, we identify three fundamental principles for effective token selection: Task Relevance, Information Coverage, and Semantic Diversity. Based on these principles, we propose TOPS, a training-free and model-agnostic pruning module that can be applied to various MLLMs. Extensive experiments on 7 MLLM backbones and 14 benchmarks demonstrate that TOPS outperforms prior methods under diverse pruning settings. Notably, on LLaVA-NeXT, TOPS removes 77.8% of visual tokens while preserving 100.0% and 100.6% performance on its 7B and 13B models, respectively, suggesting that pruning redundant visual tokens can sometimes mitigate hallucination and inspire future lightweight MLLM design.
Chinese Translation
多模态大语言模型（MLLMs）已实现强大的多模态推理能力，但其效率受到大量视觉令牌的限制，这带来了显著的计算开销。视觉令牌剪枝提供了一种自然的解决方案，但现有方法并不完美：基于注意力的标准往往保留冗余令牌，而基于多样性的标准通常对用户指令无动于衷。即使是结合多种标准的方法，仍然缺乏对令牌剪枝内在目标的原则性表述。本文从第一性原理的角度重新审视视觉令牌剪枝，并将其表述为构建令牌最优保留集。通过自上而下的信息论分析，我们确定了有效令牌选择的三个基本原则：任务相关性、信息覆盖率和语义多样性。基于这些原则，我们提出了TOPS，一个无训练且模型无关的剪枝模块，可应用于各种MLLMs。在7个MLLM主干和14个基准上的广泛实验表明，TOPS在多种剪枝设置下优于先前的方法。值得注意的是，在LLaVA-NeXT上，TOPS去除了77.8%的视觉令牌，同时在其7B和13B模型上分别保留了100.0%和100.6%的性能，这表明剪枝冗余视觉令牌有时可以减轻幻觉现象，并激发未来轻量级MLLM设计的灵感。
View on arXiv Download PDF AI Translation
cs.AI / 67 / 2606.27188

A Process Harness for Uplifting Legacy Workflows to Agentic BPM: Design and Realization in CUGA FLO

提升传统工作流至自主业务流程管理的过程工具：在CUGA FLO中的设计与实现
Fournier, Fabiana, Limonad, Lior
Abstract
We introduce the process harness, a new mechanism for uplifting legacy workflows into Agentic Business Process Management (Agentic BPM) without replacing the underlying workflow engine. A process harness places a policy-governed agentic layer around a deterministic workflow engine, intercepting designated control points to contribute reasoning, adaptation, and oversight while the engine retains structural authority over the process. To define the process harness rigorously, we develop the Task-Decision-Flow (TDF) model, specifying both its data schema and its execution semantics. TDF decomposes LLM reasoning across three policy-governed agent types: a TaskAgent for knowledge-intensive task execution, a DecisionAgent for per-case gateway routing, and a FlowAgent that governs runtime flow adaptation through a principled hook mechanism. Each agent reasons within an explicit policy drawn from the process FRAME, the aggregate policy set governing all LLM calls in the system. We then present CUGA FLO as the design and implementation realization of the TDF model, and demonstrate it on a loan approval workflow that exercises all three agent types and hook-driven regulatory override. The process harness uniquely reconciles imperative requirements, realized through deterministic workflow execution that enforces structural compliance, with normative requirements, realized through policy-framed agentic autonomy invoked at designated control points wherever the process demands it.
Chinese Translation
我们介绍了过程工具，这是一种将传统工作流提升至自主业务流程管理（Agentic BPM）的新机制，而无需替换底层工作流引擎。过程工具在确定性工作流引擎周围放置了一个政策驱动的自主层，拦截指定的控制点，以提供推理、适应和监督，同时引擎保留对流程的结构性权威。为了严格定义过程工具，我们开发了任务-决策-流（Task-Decision-Flow, TDF）模型，明确其数据模式和执行语义。TDF将大语言模型（LLM）推理分解为三种政策驱动的代理类型：用于知识密集型任务执行的任务代理（TaskAgent），用于逐案例网关路由的决策代理（DecisionAgent），以及通过原则性钩子机制管理运行时流适应的流代理（FlowAgent）。每个代理在从过程框架（FRAME）中提取的明确政策内进行推理，该框架是系统中所有LLM调用的聚合政策集。随后，我们展示了CUGA FLO作为TDF模型的设计与实现，并在一个贷款审批工作流上进行了演示，该工作流涵盖了所有三种代理类型和基于钩子的监管覆盖。过程工具独特地调和了通过确定性工作流执行实现的强制性要求（确保结构合规性）与通过在流程需求处的指定控制点调用的政策框架自主性实现的规范性要求。
View on arXiv Download PDF AI Translation
cs.AI / 68 / 2606.27215

Vulnerability of Natural Language Classifiers to Evolutionary Generated Adversarial Text

自然语言分类器对进化生成对抗文本的脆弱性
Singh, Manjinder, Brownlee, Alexander E. I., Elawady, Mohamed
Abstract
Deep learning models have achieved impressive performance across various fields but remain vulnerable to adversarial inputs, particularly in NLP, where such attacks can have significant real-world consequences. Adversarial attacks often involve small, semantically similar token replacements to fool NLP models, and recent methods have become more precise by targeting specific vulnerable words, often by exploiting some level of access to the model's internal structure. This paper proposes GAversary, a hybrid Genetic Algorithm (GA) to generate adversarial attacks on natural language models. The GA is able to treat the target model as a black box, requiring only the logit value output by the model to guide the search. GAversary differs from GAs previously proposed for this problem by using GloVe embeddings to propose word replacements (the mutation operator) to improve the semantic similarity of the adversarial examples. GAversary is applied to several benchmark data sets and well-known target models. GAversary is able to substantially reduce the target model's accuracy on test data compared to the BAE and A2T attacks compared against (in the best case, reducing a 76.8% accuracy to 5.8%, compared to BAE's 27.6%). The trade-off is that GAversary perturbs just under twice as many words as the other two methods, with a slightly lower semantic similarity to the original text and around a 5% increase in run-time.
Chinese Translation
深度学习模型在各个领域取得了令人瞩目的表现，但仍然容易受到对抗性输入的攻击，尤其是在自然语言处理（NLP）领域，这类攻击可能带来显著的现实后果。对抗性攻击通常涉及对语义相似的词汇进行小规模替换，以欺骗NLP模型，最近的方法通过针对特定脆弱词汇变得更加精确，通常利用对模型内部结构的某种程度的访问。本文提出了GAversary，一种混合遗传算法（Genetic Algorithm, GA），用于生成针对自然语言模型的对抗性攻击。GA能够将目标模型视为黑箱，仅需模型输出的logit值来指导搜索。GAversary与之前为此问题提出的GA不同，它使用GloVe嵌入来提出词汇替换（变异操作），以提高对抗样本的语义相似性。GAversary被应用于多个基准数据集和知名目标模型。与BAE和A2T攻击相比，GAversary能够显著降低目标模型在测试数据上的准确性（在最佳情况下，将76.8%的准确率降低至5.8%，而BAE的准确率为27.6%）。权衡在于，GAversary扰动的词汇数量几乎是其他两种方法的两倍，且与原始文本的语义相似性略低，运行时间增加约5%。
View on arXiv Download PDF AI Translation
cs.AI / 69 / 2606.27226

Ask, Don't Judge: Binary Questions for Interpretable LLM Evaluation and Self-Improvement

询问，而非评判：用于可解释的 LLM 评估与自我改进的二元问题
Cho, Sangwoo, Chawla, Kushal, Cai, Pengshan, Liu, Zefang, Zhu, Chenyang, Zhang, Shi-Xiong, Sahu, Sambit
Abstract
Evaluating LLM outputs remains a major bottleneck in NLP: human evaluation is expensive and slow, lexical metrics correlate poorly with human judgments on open-ended generation, and holistic LLM judges often produce opaque scores that are hard to debug. We propose BINEVAL, a framework that decomposes evaluation criteria into atomic binary questions and aggregates the resulting verdicts into interpretable, multi-dimensional scores. Given a task prompt, a meta-prompt generates fine-grained evaluation questions, and an LLM answers them independently for each output, yielding transparent question-level feedback together with calibrated overall scores. This decomposition makes evaluation easier to inspect, easier to diagnose, and directly usable for prompt improvement. Across SummEval, Topical-Chat, and QAGS, BINEVAL matches or outperforms strong baselines including UniEval and G-Eval, with especially strong results on factual consistency benchmarks such as QAGS. Beyond competitive correlation with human judgments, BINEVAL better matches human score distributions and avoids the ceiling effects common in prior LLM judges, leading to better discrimination between borderline and clearly flawed outputs. We further show that the same question-level feedback supports iterative prompt optimization, improving evaluator prompts on summarization and generation prompts on IFBench under both self-update and cross-model update settings. Overall, BINEVAL provides a task-agnostic, training-free, and interpretable evaluation framework that combines strong empirical performance with practical diagnostic and optimization value.
Chinese Translation
评估 LLM 输出仍然是自然语言处理中的一个主要瓶颈：人工评估成本高且速度慢，词汇度量与人类对开放式生成的判断相关性较差，而整体 LLM 评估者往往产生难以调试的不透明评分。我们提出了 BINEVAL，一个将评估标准分解为原子二元问题并将结果裁决汇总为可解释的多维评分的框架。给定任务提示，元提示生成细粒度的评估问题，LLM 独立回答每个输出的问题，从而提供透明的问题级反馈以及经过校准的整体评分。这种分解使得评估更易于检查、更易于诊断，并且可以直接用于提示改进。在 SummEval、Topical-Chat 和 QAGS 上，BINEVAL 的表现与强基线（包括 UniEval 和 G-Eval）相匹配或超越，尤其在事实一致性基准（如 QAGS）上表现尤为强劲。除了与人类判断的竞争性相关性外，BINEVAL 更好地匹配人类评分分布，避免了以往 LLM 评估者常见的天花板效应，从而在边界输出和明显缺陷输出之间实现了更好的区分。我们进一步展示了相同的问题级反馈支持迭代提示优化，在自我更新和跨模型更新设置下改善了摘要评估提示和生成提示的表现。总体而言，BINEVAL 提供了一个任务无关、无训练且可解释的评估框架，结合了强大的实证性能与实际的诊断和优化价值。
View on arXiv Download PDF AI Translation
cs.AI / 70 / 2606.27277

EO-WM: A Physically Informed World Model for Probabilistic Earth Observation Forecasting

EO-WM：一种基于物理信息的世界模型用于概率性地球观测预测
Luo, Junwei, Yuan, Shuai, Yang, Zhenya, Li, Yansheng, Liu, Zhe, Zhao, Hengshuang
Abstract
Earth Observation (EO) forecasting aims to predict future Earth surface dynamics from satellite observations under changing meteorological conditions. In this paper, we view this task as a partially observed, weather-driven world modeling problem, in which weather acts as a conditioning signal, while forecasting remains uncertain due to sparse observations and unobserved land-surface states. However, existing methods do not fully capture this setting: deterministic models collapse uncertainty into a single future prediction, while diffusion-based methods typically treat weather variables as undifferentiated conditioning signals, and existing benchmarks focus mainly on reconstruction accuracy rather than whether forecasts respond correctly to changed weather forcing.We introduce EO-WM, a video diffusion transformer for multispectral EO forecasting. EO-WM incorporates a physically informed conditioning framework that represents meteorological forcing through a climatological baseline, weather anomalies, and cumulative physical stress signals. Specifically, it separates baseline and anomaly through distinct conditioning pathways, and accumulates anomalous forcing over time to capture sustained heat and drought stress. To evaluate weather-response behavior beyond standard metrics, we introduce two diagnostic benchmarks: an Extreme Summer Benchmark for severity-aware prediction of vegetation degradation under extreme weather, and a Seasonal Matched-Pair Benchmark for testing response fidelity under changed weather forcing. Experiments show that EO-WM reduces the error in predicted Normalized Difference Vegetation Index (NDVI) decline amplitude by a relative 5.63% and improves directional hit rate by a relative 7.80%, while remaining competitive on standard pixel-level metrics. The benchmarks and model will be made open-source at https://github.com/Luo-Z13/EO-WM.
Chinese Translation
地球观测（EO）预测旨在根据变化的气象条件，从卫星观测中预测未来的地球表面动态。在本文中，我们将这一任务视为一个部分观测的、气候驱动的世界建模问题，其中天气作为条件信号，而由于观测稀疏和未观测的地表状态，预测仍然存在不确定性。然而，现有方法并未完全捕捉这一设置：确定性模型将不确定性压缩为单一的未来预测，而基于扩散的方法通常将天气变量视为未区分的条件信号，现有基准主要集中在重建精度上，而非预测是否正确响应变化的天气强迫。我们提出了EO-WM，一种用于多光谱EO预测的视频扩散变换器。EO-WM结合了一个基于物理信息的条件框架，通过气候基线、天气异常和累积物理应力信号来表示气象强迫。具体而言，它通过不同的条件路径分离基线和异常，并随时间累积异常强迫，以捕捉持续的热和干旱应力。为了评估天气响应行为超越标准指标，我们引入了两个诊断基准：一个极端夏季基准，用于在极端天气下对植被退化的严重程度进行预测，以及一个季节匹配对基准，用于测试在变化的天气强迫下的响应忠实度。实验表明，EO-WM将预测的归一化植被指数（NDVI）下降幅度的误差相对减少了5.63%，并将方向命中率相对提高了7.80%，同时在标准像素级指标上保持竞争力。该基准和模型将开放源代码，网址为 https://github.com/Luo-Z13/EO-WM。
View on arXiv Download PDF AI Translation
cs.AI / 71 / 2606.27286

Simulation-based inference for rapid Bayesian parameter estimation in epidemiological models: a comparison with MCMC

基于模拟的推断在流行病模型中快速贝叶斯参数估计的比较：与MCMC的对比
Bazarova, Alina, Jadebeck, Johann Fredrik, Zunker, Henrik, Klett-Tammen, Carolina J., Heinsohn, Torben, Wiechert, Wolfgang, Noeh, Katharina, Kesselheim, Stefan
Abstract
Mechanistic epidemiological models are widely used to support infectious disease forecasting and public-health decision making. Bayesian calibration of such models is commonly performed using Markov chain Monte Carlo (MCMC), which can become computationally expensive for high-dimensional nonlinear systems and repeated near-real-time analyses. Here, we investigate simulation-based inference (SBI) using neural posterior estimation as a scalable alternative for Bayesian calibration of a mechanistic SECIR epidemiological model using COVID-19 intensive care unit (ICU) occupancy data from Germany during 2020. We compared SBI and MCMC across multiple epidemic phases using both 31-day inference windows and a substantially more challenging 201-day reconstruction problem involving multiple transmission change points. Posterior agreement was evaluated quantitatively using Wasserstein distances and Kullback-Leibler divergences together with posterior predictive checks. Across the 31-day windows, SBI recovered posterior distributions in strong agreement with MCMC while accurately reproducing observed ICU trajectories. In the 201-day setting, SBI preserved the dominant posterior structure despite increased uncertainty. SBI, by combining CPU and GPU resources, substantially reduced computational runtime compared with MCMC, which was restricted to running on CPUs. Whereas MCMC required approximately 1000 seconds for the 31-day inference problems, SBI achieved comparable posterior and predictive performance in approximately 60-70 seconds on a single GPU. For the 201-day inference problem, SBI required an average of 157 seconds, while the MCMC runs took over 19,000 seconds. Our results demonstrate that SBI provides a rapid and computationally efficient framework for Bayesian calibration of mechanistic epidemiological models, supporting repeated near-real-time inference and rapid outbreak analysis.
Chinese Translation
机械性流行病模型广泛用于支持传染病预测和公共卫生决策。此类模型的贝叶斯校准通常使用马尔可夫链蒙特卡洛（MCMC）方法进行，但对于高维非线性系统和重复的近实时分析，这种方法可能会变得计算成本高昂。在此，我们研究了基于模拟的推断（SBI），使用神经后验估计作为机械性SECIR流行病模型的可扩展替代方案，利用2020年德国COVID-19重症监护病房（ICU）占用数据进行贝叶斯校准。我们在多个流行病阶段比较了SBI和MCMC，使用了31天的推断窗口和一个涉及多个传播变更点的更具挑战性的201天重建问题。通过使用Wasserstein距离和Kullback-Leibler散度以及后验预测检验，定量评估了后验一致性。在31天的窗口中，SBI恢复的后验分布与MCMC高度一致，同时准确再现了观察到的ICU轨迹。在201天的设置中，尽管不确定性增加，SBI仍然保持了主导的后验结构。通过结合CPU和GPU资源，SBI显著减少了计算运行时间，而MCMC仅限于在CPU上运行。对于31天的推断问题，MCMC大约需要1000秒，而SBI在单个GPU上以约60-70秒实现了可比的后验和预测性能。对于201天的推断问题，SBI平均需要157秒，而MCMC的运行时间超过19000秒。我们的结果表明，SBI为机械性流行病模型的贝叶斯校准提供了一个快速且计算高效的框架，支持重复的近实时推断和快速疫情分析。
View on arXiv Download PDF AI Translation
cs.AI / 72 / 2606.27288

When Does Combining Language Models Help? A Co-Failure Ceiling on Routing, Voting, and Mixture-of-Agents Across 67 Frontier Models

何时结合语言模型有助于提升性能？67个前沿模型在路由、投票和混合代理中的共同失败上限
Chen, Josef
Abstract
Multi-model LLM systems such as routing, voting, cascades, fusion, and mixture-of-agents are used to beat single-model accuracy. We show that their gain is capped by a quantity the field rarely reports. For any policy whose output is one member model answer, accuracy cannot exceed one minus beta, where beta is the rate at which every model is wrong on the same query. In contrast, the usual diagnostic, average pairwise error correlation rho, cannot identify beta: error laws with identical marginals and pairwise correlations can have different all-wrong rates. A Clopper-Pearson bound on beta gives a finite-sample certificate on the largest gain any router, vote, or cascade could deliver before training a router. Across 67 models from 21 providers, a tetrachoric-calibrated single-factor model still underprices the all-wrong tail: on open-ended mathematics, observed beta is 0.052 versus 0.023 under the full 67-model Gaussian copula, about 2.5 times underpricing, with 90 percent CI 1.7 to 3.4 and k equals 17. The effect recurs on execution-graded code, where beta is 0.079. Re-asking the same GPQA-Diamond questions in free-response rather than multiple-choice form reopens the tail, with beta 0.127 and a five-judge panel with kappa 0.73 to 0.92, locating co-failure in answer format rather than subject. At matched quality, low-rho heterogeneous ensembles beat high-rho Self-MoA, but on checkable tasks in our pool, combining models rarely beats the single best model without a strong query-level routing signal. Gains come from models failing on different questions, not from adding more models.
Chinese Translation
多模型大语言模型（LLM）系统，如路由、投票、级联、融合和混合代理，旨在超越单模型的准确性。我们展示了它们的增益受到一个该领域很少报告的量的限制。对于任何输出为单个模型答案的策略，准确性不能超过1减去β，其中β是每个模型在相同查询上出错的比率。相比之下，通常的诊断指标——平均成对错误相关性ρ，无法识别β：具有相同边际和成对相关性的错误法则可以有不同的全错率。对β的Clopper-Pearson界限提供了一个有限样本证书，表明任何路由器、投票或级联在训练路由器之前能够提供的最大增益。在来自21个提供者的67个模型中，一个经过四分之一校准的单因子模型仍然低估了全错尾部：在开放式数学问题上，观察到的β为0.052，而在完整的67模型高斯耦合下为0.023，低估约为2.5倍，90%置信区间为1.7到3.4，k为17。该效应在执行评分代码中再次出现，β为0.079。以自由回答而非多项选择形式重新提问相同的GPQA-Diamond问题重新打开了尾部，β为0.127，五位评审小组的κ值为0.73到0.92，定位到答案格式而非主题的共同失败。在匹配质量下，低ρ的异质集成优于高ρ的自我混合代理（Self-MoA），但在我们池中的可检查任务上，结合模型很少能超越单个最佳模型，除非有强烈的查询级路由信号。增益来自于模型在不同问题上的失败，而不是简单增加更多模型。
View on arXiv Download PDF AI Translation
cs.AI / 73 / 2606.27334

Language-Based Digital Twins for Elderly Cognitive Assistance

基于语言的数字双胞胎用于老年认知辅助
Hosseini, Mohammad Mehdi, Mahoor, Mohammad H., Dodge, Hiroko H.
Abstract
Digital twins have emerged as a promising paradigm for personalized healthcare, enabling modeling of individual behavior and health trajectories. In cognitive health, early detection of Mild Cognitive Impairment (MCI) remains challenging, where language and conversational patterns serve as non-invasive biomarkers. In this work, we propose a language-based digital twin framework that leverages large language models (LLMs) to mimic the conversational behavior of elderly individuals by incorporating stylometric cues and contextual metadata. To evaluate fidelity and cognitive consistency, we introduce a multi-head conditional variational autoencoder (cVAE) that jointly measures reconstruction quality and predicts cognitive scores. Experiments on the I-CONECT dataset show that the digital twin preserves identity-specific characteristics and achieves reconstruction and MoCA prediction errors comparable to real data, while outperforming baseline GPT-generated responses. These results highlight the potential of language-based digital twins as a scalable and non-invasive approach for personalized and continuous cognitive health monitoring.
Chinese Translation
数字双胞胎作为个性化医疗的有前景的范式，能够建模个体行为和健康轨迹。在认知健康领域，轻度认知障碍（MCI）的早期检测仍然具有挑战性，其中语言和对话模式作为非侵入性生物标志物发挥作用。在本研究中，我们提出了一种基于语言的数字双胞胎框架，该框架利用大型语言模型（LLMs）通过结合风格特征和上下文元数据来模拟老年个体的对话行为。为了评估保真度和认知一致性，我们引入了一种多头条件变分自编码器（cVAE），该模型共同测量重建质量并预测认知评分。在I-CONECT数据集上的实验表明，数字双胞胎保留了特定身份的特征，并且其重建和MoCA预测误差与真实数据相当，同时优于基线GPT生成的响应。这些结果突显了基于语言的数字双胞胎作为一种可扩展且非侵入性的个性化和持续认知健康监测方法的潜力。
View on arXiv Download PDF AI Translation
计算语言学 (Computation and Language)
67
cs.CL / 1 / 2606.26100

HierBias: Context-Conditioned Hierarchical Media Bias Detection with Multi-Task Type Classification

HierBias：基于上下文的分层媒体偏见检测与多任务类型分类
Li, Kaining, Yan, Ruichen, Dong, Yuxin
Abstract
Media bias detection is a critical task for ensuring fair and balanced information dissemination, yet existing sentence-level approaches classify each sentence independently, ignoring inter-sentence contextual signals that human annotators naturally exploit. We present \textbf{HierBias}, a hierarchical context-conditioned media bias detector that formally models document context in bias prediction. We introduce the \emph{context-conditioned bias probability} and prove theoretically that leveraging document context strictly reduces the Bayes error of sentence-level classification when inter-sentence mutual information is non-zero. A multi-task generalization bound further establishes that jointly training binary bias detection and fine-grained bias type classification improves sample efficiency on small annotated corpora. Architecturally, HierBias pairs a sentence-level RoBERTa encoder with a cross-sentence Transformer aggregator and dual output heads for binary detection and four-class type classification. Evaluated on BABE and BASIL, HierBias achieves 0.853 F1 and 0.723 MCC, surpassing the state-of-the-art bias-detector by $+2.6\%$ F1 and $+4.3\%$ MCC (McNemar's test, $p < 0.05$). Ablation experiments confirm that each theoretical component contributes independently and consistently.
Chinese Translation
媒体偏见检测是确保信息传播公平和均衡的重要任务，然而现有的句子级方法独立地对每个句子进行分类，忽视了人类标注者自然利用的句间上下文信号。我们提出了 extbf{HierBias}，一种分层的上下文条件媒体偏见检测器，正式建模文档上下文在偏见预测中的作用。我们引入了 extit{上下文条件偏见概率}，并理论证明，当句间互信息非零时，利用文档上下文严格降低句子级分类的贝叶斯误差。多任务泛化界限进一步表明，联合训练二元偏见检测和细粒度偏见类型分类提高了在小型标注语料库上的样本效率。在架构上，HierBias将句子级RoBERTa编码器与跨句Transformer聚合器和用于二元检测及四类类型分类的双输出头相结合。在BABE和BASIL上的评估中，HierBias达到了0.853的F1和0.723的MCC，超越了最先进的偏见检测器，F1提高了$+2.6\%$，MCC提高了$+4.3\\%$（McNemar检验，$p < 0.05$）。消融实验确认每个理论组件独立且一致地贡献于整体性能。
View on arXiv Download PDF AI Translation
cs.CL / 2 / 2606.26101

Know2Guess: A Contamination-Aware Multi-Zone Benchmark for Knowledge-Boundary Evaluation in Large Language Models

Know2Guess：一种关注污染的多区域基准，用于大型语言模型的知识边界评估
Meng, Renwei, Zhang, Bowen, Wang, Jian, Wang, Xican, Wu, Haoyi, Qiu, Xuanyan, Yang, Shengan
Abstract
Reliable evaluation of large language models should separate supported answering from unsupported guessing without conflating either with data contamination, prompt idiosyncrasy, or generic refusal behavior. We present a contamination-aware, multi-zone benchmark for measuring the transition from answerable knowledge to abstention-expected unknowns under frozen build-time labels. The benchmark contains 1,200 items across five domains, explicit abstention expectations, contamination-risk metadata, and dual parsing with an official strict parser plus a normalized robustness parser. We evaluate FLAN-T5, Qwen2.5-Instruct, and Llama-3-Instruct models under locked answer-or-abstain prompts, answer-only controls, and prompt-template variants. The benchmark is not solved by generic non-answer behavior: FLAN baselines remain weak on productive abstention, while stronger instruction-tuned models expose a selective but incomplete transition from answering to abstaining. Qwen2.5-3B-Instruct achieves the best overall reliability, but answer-expected zones remain difficult, calibration remains poor, and benign-item refusal persists. Prompt and parser robustness analyses preserve the main ranking and qualitative conclusions. The benchmark therefore provides a reproducible protocol for auditing answerability, abstention, refusal, and contamination as distinct but interacting dimensions of LLM reliability.The dataset is publicly available at https://github.com/renweimeng/Know2Guess-A-Contamination-Aware-Multi-Zone-Benchmark.
Chinese Translation
对大型语言模型的可靠评估应将支持的回答与不支持的猜测区分开，而不应与数据污染、提示特异性或一般拒绝行为混淆。我们提出了一种关注污染的多区域基准，用于在冻结的构建时标签下测量从可回答知识到预期弃权未知的过渡。该基准包含1200个项目，涵盖五个领域，明确的弃权预期，污染风险元数据，以及使用官方严格解析器和标准化鲁棒性解析器的双重解析。我们在锁定的回答或弃权提示、仅回答控制和提示模板变体下评估了FLAN-T5、Qwen2.5-Instruct和Llama-3-Instruct模型。该基准并未被一般的非回答行为解决：FLAN基线在有效弃权上仍然较弱，而更强的指令调优模型则显示出从回答到弃权的选择性但不完整的过渡。Qwen2.5-3B-Instruct在整体可靠性方面表现最佳，但预期回答区域仍然困难，校准效果较差，良性项目的拒绝现象持续存在。提示和解析器的鲁棒性分析保持了主要排名和定性结论。因此，该基准提供了一个可重复的协议，用于审计回答能力、弃权、拒绝和污染，作为大型语言模型可靠性的不同但相互作用的维度。数据集可在 https://github.com/renweimeng/Know2Guess-A-Contamination-Aware-Multi-Zone-Benchmark 获取。
View on arXiv Download PDF AI Translation
cs.CL / 3 / 2606.26102

Helpfulness Hurts: Domain-Dependent Degradation of Mid-Trained Compassion Values Under Post-Training

有益性有害：后训练下中期训练同情价值的领域依赖性退化
Brazilek, Jasmine, Seawell, Juliana
Abstract
Standard post-training pipelines apply supervised fine-tuning (SFT) and reinforcement learning (RL) to make language models helpful, but these processes may inadvertently degrade values instilled during pre-training. We investigate whether the domain of post-training data differentially affects the retention of animal compassion values in a Llama 3.1 8B model mid-trained on compassion-oriented synthetic data, using both SFT (helpfulness via Dolly-15k vs. coding via Magicoder-110K) and GRPO (helpfulness via RLHFlow vs. coding via Magicoder), evaluated on the Animal Harm Benchmark (AHB 2.2) and MORU benchmark (Moral Reasoning Under Uncertainty). Helpfulness training significantly degrades animal compassion relative to coding training on AHB (SFT: 35.7% vs. 65.2%; GRPO: 18.7% vs. 32.0%), replicating across two independent helpfulness datasets and two training paradigms. On English MORU items, helpfulness training degrades general moral reasoning by 25.5 percentage points (46.4% vs. 71.9%), a striking gap that rivals the compassion effect in magnitude. However, this effect does not transfer cross-lingually: on the multilingual MORU benchmark, the domain effect disappears (SFT: 52.3% vs. 51.2%). In contrast, the animal compassion effect transfers consistently across languages, with Magicoder's AHB percentage-point gain over the base model 4.5 times larger on non-English items than English items. This divergence suggests that values instilled through mid-training are encoded more deeply and cross-lingually than reasoning improvements from domain-specific post-training. These results suggest that, for labs building on value-laden mid-training, coding-domain post-training may better preserve mid-trained values than helpfulness post-training without harming general reasoning capabilities.
Chinese Translation
标准的后训练流程应用监督微调（SFT）和强化学习（RL）来使语言模型变得有益，但这些过程可能会无意中削弱在预训练期间灌输的价值观。我们研究了后训练数据的领域是否会对在同情导向合成数据上中期训练的Llama 3.1 8B模型中动物同情价值的保留产生不同影响，使用SFT（通过Dolly-15k实现有益性与通过Magicoder-110K实现编码）和GRPO（通过RLHFlow实现有益性与通过Magicoder实现编码），并在动物伤害基准（AHB 2.2）和不确定性下的道德推理基准（MORU）上进行评估。相对于编码训练，基于有益性的训练在AHB上显著降低了动物同情值（SFT：35.7% vs. 65.2%；GRPO：18.7% vs. 32.0%），这一结果在两个独立的有益性数据集和两种训练范式中得到了重复。在英语MORU项目上，有益性训练使一般道德推理降低了25.5个百分点（46.4% vs. 71.9%），这一显著差距与同情效应的幅度相当。然而，这一效应并未在跨语言中转移：在多语言MORU基准上，领域效应消失（SFT：52.3% vs. 51.2%）。相反，动物同情效应在不同语言间一致转移，Magicoder在非英语项目上的AHB百分比增益是英语项目的4.5倍。这一差异表明，通过中期训练灌输的价值观比领域特定后训练带来的推理改进更深层次地编码并跨语言存在。这些结果表明，对于在价值导向中期训练基础上进行构建的实验室，编码领域的后训练可能比有益性后训练更好地保留中期训练的价值观，而不损害一般推理能力。
View on arXiv Download PDF AI Translation
cs.CL / 4 / 2606.26103

Investigating LLM's Problem Solving Capability -- a Study on Statics Questions

探究大型语言模型的解题能力——关于静力学问题的研究
Culleton, Tanner, Chang, Hung-Fu
Abstract
Large Language Models (LLMs) have rapidly influenced many aspects of society, particularly education, due to their demonstrated ability to complete assignments and examinations across a wide range of subjects. Although prior studies have examined the educational impact of LLMs, much of the existing work relies on public or open problem datasets and lacks topic-specific analysis. In engineering education, especially within mechanical engineering, systematic investigations of LLM performance on specific problem types remain limited. Instead of using traditional methods that directly ask textbook questions to an LLM tool, our study adopts a model distillation process to evaluate LLM capabilities in solving statics problems. By distilling ChatGPT, we extracted 25 text-only statics questions and further constructed two additional datasets by adding diagrams and modifying their numerical values. Experimental results show that while LLMs perform well on text-only statics problems, their accuracy decreases when diagrams are introduced and the problems require multi-step reasoning. Further analysis suggests that this performance drop is not primarily caused by limitations in image recognition, but rather by difficulties in multi-step reasoning and in consistently applying extracted visual information across successive solution stages.
Chinese Translation
大型语言模型（LLMs）由于其在多种学科中完成作业和考试的能力，迅速影响了社会的许多方面，尤其是教育。尽管先前的研究已考察了LLMs的教育影响，但现有研究大多依赖于公共或开放问题数据集，缺乏针对特定主题的分析。在工程教育中，尤其是在机械工程领域，针对特定问题类型的LLM性能系统性研究仍然有限。我们的研究采用模型蒸馏过程来评估LLM在解决静力学问题上的能力，而不是使用传统方法直接向LLM工具提出教科书问题。通过蒸馏ChatGPT，我们提取了25个仅包含文本的静力学问题，并通过添加图示和修改数值构建了两个额外的数据集。实验结果表明，尽管LLMs在仅包含文本的静力学问题上表现良好，但当引入图示并且问题需要多步骤推理时，其准确性下降。进一步分析表明，这一性能下降并非主要由于图像识别的局限性，而是由于在多步骤推理和在连续解题阶段一致应用提取的视觉信息方面的困难。
View on arXiv Download PDF AI Translation
cs.CL / 5 / 2606.26104

Assert, don't describe: Linguistic features that shift LLM reasoning about animal welfare

主张，而非描述：影响大型语言模型（LLM）动物福利推理的语言特征
Brazilek, Jasmine, Dunn, Harper
Abstract
Animal-welfare advocates produce a lot of writing, and increasingly that writing trains the language models that millions of people then ask about animal welfare. Using vocabulary-matched stance-contrast probes on a held-out animal-welfare benchmark, we measure how each of ten linguistic features changes Llama-3.2-1B's preference for pro-animal-welfare reasoning when used as fine-tuning data. Eight of the ten features produce statistically significant shifts. Seven move the model toward stronger pro-animal-welfare reasoning: assertive certainty, explicit moral vocabulary, emotion words, evaluative claims, narrative structure, depicted harm severity, and immediate temporal framing. Two move it the other way: hedged language and concrete sensory description both dilute the pro-animal-welfare stance. First-person perspective has no statistically significant effect. The practical recommendation for anyone writing animal-welfare text that may end up in LLM training corpora: assert a position rather than describe a scene neutrally. The features that shift the model are the ones that make the writer's position explicit; the features that dilute it hold animal-welfare content but withhold stance.
Chinese Translation
动物福利倡导者产生了大量的书面材料，而这些材料越来越多地用于训练数百万用户询问动物福利的语言模型。通过在一个保留的动物福利基准上使用词汇匹配的立场对比探针，我们测量了十种语言特征如何改变 Llama-3.2-1B 在作为微调数据时对动物福利推理的偏好。其中十种特征中有八种产生了统计显著的变化。七种特征使模型向更强的支持动物福利推理倾斜：果断的确定性、明确的道德词汇、情感词、评估性主张、叙事结构、描绘的伤害严重性和即时的时间框架。两种特征则使其向相反方向移动：模糊语言和具体的感官描述都削弱了支持动物福利的立场。第一人称视角没有显著的统计影响。对于任何可能最终出现在大型语言模型训练语料库中的动物福利文本的撰写者，实际建议是：主张一个立场，而不是中立地描述一个场景。改变模型的特征是那些使作者立场明确的特征；而削弱立场的特征则包含动物福利内容，但未明确立场。
View on arXiv Download PDF AI Translation
cs.CL / 6 / 2606.26105

Context Recycling for Long-Horizon LLM Inference

长时间范围大语言模型推理的上下文回收
Thomas, Derek
Abstract
Large language models (LLMs) exhibit strong capabilities in short-context reasoning but degrade in performance over long conversational horizons due to context window limitations and inefficient token usage. We introduce ContextForge, a system for context recycling that maintains task-relevant information across turns by combining structured query generation, external memory retrieval, and controlled synthesis. The system enables efficient reuse of prior computation without relying on full context replay, reducing token overhead while preserving answer quality. We evaluate ContextForge using a 15-turn conversational benchmark that tests multi-turn reasoning, back-references, and domain shifts across structured healthcare queries. Compared to a baseline agent using identical underlying models, ContextForge demonstrates improved consistency and reduced token consumption, while maintaining comparable response accuracy. These results suggest that context recycling provides a practical approach for extending LLM capabilities in long-horizon tasks without requiring larger context windows or model retraining. Code and evaluation artifacts are available at https://github.com/Betanu701/ContextForge.
Chinese Translation
大型语言模型（LLMs）在短上下文推理方面表现出强大的能力，但由于上下文窗口的限制和低效的标记使用，在长对话范围内性能下降。我们提出了ContextForge，一个上下文回收系统，通过结合结构化查询生成、外部记忆检索和受控合成，保持跨轮次的任务相关信息。该系统能够高效地重用先前的计算，而无需依赖完整的上下文重放，从而减少标记开销，同时保持答案质量。我们使用一个15轮对话基准对ContextForge进行了评估，该基准测试多轮推理、回溯引用和结构化医疗查询中的领域转变。与使用相同基础模型的基线代理相比，ContextForge展示了更好的连贯性和减少的标记消耗，同时保持了相当的响应准确性。这些结果表明，上下文回收为扩展LLM在长时间范围任务中的能力提供了一种实用的方法，而无需更大的上下文窗口或模型再训练。代码和评估文档可在 https://github.com/Betanu701/ContextForge 获取。
View on arXiv Download PDF AI Translation
cs.CL / 7 / 2606.26106

Reducing Conversational Escalation in Large Language Model Dialogue with Nonviolent Communication Constraints

通过非暴力沟通约束减少大型语言模型对话中的冲突升级
Sun, Zhixing, Xu, Shenghe, Li, Tao
Abstract
Large language models (LLMs) are increasingly used in emotionally charged situations involving interpersonal conflict, frustration, and distress. While prior safety research has focused on preventing explicit harms such as toxic or policy-violating content, less attention has been paid to conversational behaviors that may unintentionally escalate conflict. In this paper, we investigate whether LLMs can be guided toward more de-escalating dialogue behavior through lightweight prompt-level constraints derived from Nonviolent Communication (NVC). We reformulate NVC principles as process-oriented guidelines that discourage blame attribution, emphasize attention to users' emotional experiences, and encourage clarification before advice. Using a dual-agent simulation framework across multiple instruction-tuned models and user resistance levels, we show that NVC-constrained prompting consistently reduces conversational escalation and stabilizes interactions with highly resistant users. These results suggest that simple communication constraints can meaningfully improve the trustworthiness of LLM dialogue in conflict-prone settings.
Chinese Translation
大型语言模型（LLMs）越来越多地应用于涉及人际冲突、挫折和痛苦的情感充沛的情境中。尽管以往的安全研究主要集中在防止显性伤害，如有毒或违反政策的内容，但对可能无意中升级冲突的对话行为关注较少。本文探讨了是否可以通过源自非暴力沟通（NVC）的轻量级提示级约束，引导LLMs朝向更具去升级特征的对话行为。我们将NVC原则重新表述为过程导向的指导方针，旨在避免归责，强调关注用户的情感体验，并鼓励在提供建议之前进行澄清。通过在多个指令调优模型和用户抵抗水平下使用双代理模拟框架，我们展示了NVC约束提示能够持续减少对话升级，并稳定与高度抵抗用户的互动。这些结果表明，简单的沟通约束可以在易冲突的环境中显著提高LLM对话的可信度。
View on arXiv Download PDF AI Translation
cs.CL / 8 / 2606.26107

Low Resource Multimodal Translation of Nepali Spoken Words into Emotion-Conditioned Sign Language Avatars

尼泊尔口语词汇的低资源多模态翻译为情感条件的手语虚拟形象
Bhusal, Jatin, Tamang, Salma
Abstract
Sign language communication systems, that integrate emotional expression remain underexplored, particularly for low-resource languages. This pilot study presents NEST-V1 (Nepali Emotion and Speech Transformer - Version 1), a proof-of-concept multimodal framework that demonstrates the feasibility of generating emotion-conditioned Nepali Sign Language avatars from spoken input. As a preliminary investigation, we focus on four common Nepali words ("thank you", "hello", "house", "me") across three emotional states (happy, neutral, sad) to validate our core technical approach. Our lightweight architecture employs a shared acoustic encoder for simultaneous Automatic Speech Recognition and emotion classification, achieving 81.1% ASR accuracy and 79.21% emotion recognition accuracy on a dataset of 600 labeled audio samples from 50 speakers. The system demonstrates 37% parameter efficiency compared to separate model architectures while maintaining a lightweight footprint with only 22.1M parameters suitable for edge deployment. This pilot work establishes the technical foundation for emotion-aware sign language translation in low-resource settings and provides a scalable framework for future expansion to larger vocabularies and more diverse emotional expressions. Our preliminary results indicate the viability of real-time, emotionally expressive sign language communication systems for the hearing-impaired community, with clear pathways for enhancement in subsequent development phases.
Chinese Translation
集成情感表达的手语交流系统仍然未得到充分探索，特别是对于低资源语言。本研究介绍了NEST-V1（尼泊尔情感与语音变换器 - 版本1），这是一个概念验证的多模态框架，展示了从口语输入生成情感条件的尼泊尔手语虚拟形象的可行性。作为初步研究，我们关注四个常见的尼泊尔词汇（“谢谢”、“你好”、“房子”、“我”）在三种情感状态（快乐、中性、悲伤）下，以验证我们的核心技术方法。我们的轻量级架构采用共享声学编码器，实现自动语音识别和情感分类的同时进行，在来自50位说话者的600个标记音频样本的数据集上，达到了81.1%的自动语音识别准确率和79.21%的情感识别准确率。与单独模型架构相比，该系统在参数效率上提高了37%，同时保持了仅22.1M参数的轻量级特征，适合边缘部署。这项初步工作为低资源环境中的情感感知手语翻译奠定了技术基础，并提供了一个可扩展的框架，以便未来扩展到更大的词汇量和更多样化的情感表达。我们的初步结果表明，实时、情感丰富的手语交流系统对听障社区是可行的，并为后续开发阶段的增强提供了明确的路径。
View on arXiv Download PDF AI Translation
cs.CL / 9 / 2606.26108

Where Larger Models Excel: The Primacy of Constraint-Guided Reasoning

大型模型的优势：约束引导推理的优越性
Lin, Guan-Yi, Huang, Hen-Hsen
Abstract
Larger language models consistently outperform smaller ones on reasoning benchmarks, yet the reasoning differences underlying this gap remain underexplored. Across benchmarks in mathematics, physics, chemistry, and programming, we observe stable performance gaps: averaged over datasets, Qwen3-32B outperforms Qwen3-8B by 6.43%, while GPT-OSS-120B exceeds GPT-OSS-20B by 7.38%. To study the reasoning differences behind these gains, we develop AdvCluster, an automated framework that identifies questions where the larger model shows a stable advantage, extracts fine-grained advantage descriptions from paired reasoning traces produced by larger and smaller models, and organizes them through semantic clustering with quantitative evaluation and selection guided by a reviewer model. Our analysis yields a systematic taxonomy of larger model reasoning advantages, spanning both common advantages that recur across domains and specialized advantages associated with particular domains. Across these patterns, a recurring theme is Constraint-Guided Reasoning: larger models are better at identifying explicit and implicit constraints, organizing them into structured reasoning, and using them to rule out infeasible paths and verify intermediate steps.
Chinese Translation
大型语言模型在推理基准测试中始终优于较小的模型，但导致这一差距的推理差异仍未得到充分探讨。在数学、物理、化学和编程等基准测试中，我们观察到稳定的性能差距：在数据集上平均，Qwen3-32B比Qwen3-8B高出6.43%，而GPT-OSS-120B超过GPT-OSS-20B的幅度为7.38%。为了研究这些提升背后的推理差异，我们开发了AdvCluster，一个自动化框架，能够识别出大型模型表现出稳定优势的问题，从较大和较小模型生成的配对推理轨迹中提取细粒度的优势描述，并通过语义聚类进行组织，同时由审阅模型进行定量评估和选择。我们的分析得出了一个系统的较大模型推理优势分类法，涵盖了跨领域重复出现的常见优势和与特定领域相关的专业优势。在这些模式中，一个反复出现的主题是约束引导推理：大型模型在识别显性和隐性约束、将其组织成结构化推理以及利用这些约束排除不可行路径和验证中间步骤方面表现更佳。
View on arXiv Download PDF AI Translation
cs.CL / 10 / 2606.26112

From Lexicon to AI: A Structured-Data Pipeline for Specialized Conversational Systems in Low-Resource Languages

从词汇到人工智能：低资源语言专业对话系统的结构化数据管道
Mantri, Siddhant Hitesh, Gorasiya, Dhara, Kulkarni, Malhar, Bhattacharya, Pushpak
Abstract
Low-resource languages face a critical challenge in AI development: creating specialized conversational systems without access to massive training corpora. We present a systematic methodology for transforming structured linguistic resources into specialized AI systems, demonstrating that expert-curated lexical databases can serve as effective foundations for conversational AI development. Our approach converts Hindi WordNet into 1.25 million diverse instruction-response pairs, fine-tunes a 12B-parameter language model using resource-efficient LoRA with 4-bit quantization. Evaluation through a Hindi language learning chatbot demonstrates that structured-knowledge-based systems achieve superior pedagogical effectiveness (91.0 vs. 79.4-83.6 for general-purpose models) while maintaining competitive semantic performance and exceptional consistency. The complete pipeline demonstrates a proof-of-concept methodology using Hindi for developing specialized AI systems for any languages with WordNet resources. This work addresses the critical gap in AI accessibility for low-resource languages, offering a practical alternative to corpus-intensive approaches and potentially enabling specialized AI development for the hundreds of languages with existing WordNet resources.
Chinese Translation
低资源语言在人工智能发展中面临着一个关键挑战：在没有大量训练语料的情况下创建专业对话系统。我们提出了一种系统的方法，将结构化语言资源转化为专业的人工智能系统，证明专家策划的词汇数据库可以作为对话人工智能开发的有效基础。我们的方法将印地语WordNet转换为125万个多样化的指令-响应对，使用资源高效的LoRA和4位量化对一个12B参数的语言模型进行微调。通过印地语学习聊天机器人进行的评估表明，基于结构知识的系统在教学效果上表现优越（91.0 vs. 79.4-83.6，针对通用模型），同时保持竞争力的语义表现和卓越的一致性。完整的管道展示了一种使用印地语开发专业人工智能系统的概念验证方法，适用于任何拥有WordNet资源的语言。这项工作填补了低资源语言在人工智能可及性方面的关键空白，提供了一种实用的替代方案，可能使数百种拥有现有WordNet资源的语言的专业人工智能开发成为可能。
View on arXiv Download PDF AI Translation
cs.CL / 11 / 2606.26120

Dynamic-dLLM: Dynamic Cache-Budget and Adaptive Parallel Decoding for Training-Free Acceleration of Diffusion LLM

动态-dLLM：动态缓存预算和自适应并行解码用于无训练加速扩散大语言模型
Wu, Tianyi, Sun, Xiaoxi, Jiao, Yanhua, Li, Yulin, Chen, Yixin, Cao, YunHao, Hu, YiQi, Tian, Zhuotao
Abstract
Diffusion Large Language Models (dLLMs) offer a promising alternative to autoregressive models, excelling in text generation tasks due to their bidirectional attention mechanisms. However, their computational complexity scales on the order of L cubed with the sequence length L. This poses significant challenges for long-sequence and real-time applications, primarily due to the lack of compatibility with key-value caching and the non-autoregressive nature of denoising steps. Existing acceleration methods rely on static caching or parallel decoding strategies, which fail to account for the dynamic behavior of token properties across layers and decoding steps. We propose Dynamic-dLLM, a training-free framework that enhances dLLM inference efficiency through two components: Dynamic Cache Updating (DCU), which adaptively allocates cache-update budgets based on layer-wise token dynamics, and Adaptive Parallel Decoding (APD), which dynamically calibrates decoding thresholds to balance generation quality and efficiency. Extensive experiments on models like LLaDA-8B-Instruct, LLaDA-1.5, and Dream-v0-7B-Instruct across benchmarks such as MMLU, GSM8K, and HumanEval demonstrate that Dynamic-dLLM significantly improves inference speed. It attains an average speedup exceeding 3 times while maintaining performance. Dynamic-dLLM outperforms state-of-the-art acceleration methods and provides a plug-and-play solution for efficient dLLM deployment without compromising performance. The code is available at https://github.com/TianyiWu233/DYNAMIC-DLLM.
Chinese Translation
扩散大语言模型（dLLMs）为自回归模型提供了一种有前景的替代方案，由于其双向注意机制，在文本生成任务中表现优异。然而，它们的计算复杂度随着序列长度 L 的增加呈立方级别增长。这对长序列和实时应用构成了重大挑战，主要是由于缺乏与键值缓存的兼容性以及去噪步骤的非自回归特性。现有的加速方法依赖于静态缓存或并行解码策略，这些方法未能考虑到跨层和解码步骤中令牌属性的动态行为。我们提出了动态-dLLM，一个无训练的框架，通过两个组件提高 dLLM 的推理效率：动态缓存更新（Dynamic Cache Updating, DCU），根据层级令牌动态自适应分配缓存更新预算；自适应并行解码（Adaptive Parallel Decoding, APD），动态校准解码阈值以平衡生成质量和效率。在 LLaDA-8B-Instruct、LLaDA-1.5 和 Dream-v0-7B-Instruct 等模型上，以及在 MMLU、GSM8K 和 HumanEval 等基准测试中进行的广泛实验表明，动态-dLLM 显著提高了推理速度。它在保持性能的同时，平均加速超过 3 倍。动态-dLLM 超越了最先进的加速方法，并提供了一种即插即用的解决方案，以高效部署 dLLM，而不影响性能。代码可在 https://github.com/TianyiWu233/DYNAMIC-DLLM 获取。
View on arXiv Download PDF AI Translation
cs.CL / 12 / 2606.26130

Thinking Like a Scientist? A Structural Study of LLM-Generated Research Methods

像科学家一样思考？对LLM生成研究方法的结构性研究
Carlon, Francesca, Verbeken, Brecht, Ginis, Vincent, Algaba, Andres
Abstract
Large Language Models (LLMs) are increasingly used to guide research methodology, yet their default methodological tendencies under minimal prompting remain unclear. Here, we prompt GPT-5.1, Gemini 3 Pro, and DeepSeek-V3.2 with an LLM-extracted research question from each of 1,000 recent arXiv computer-science papers and compare the resulting methodology suggestions against a paper-derived experimental inventory. Since we provide only the research question, the differences we measure reflect initial suggestions and not how optimal those suggestions are. We extract structured method features from both sources, map them into a shared taxonomy, and quantify divergence across multiple taxonomy dimensions including model provider, dataset task type, and evaluation metric type. The strongest imbalance appears in provider choice, with Jensen-Shannon divergence about 3-5x larger than any other taxonomy dimension. Other/Academic single-occurrence models are underrepresented by 23-24 percentage points, while reused academic/community models are slightly overrepresented (4-6pp). LLMs also suggest a much narrower range of methods overall: the effective number of model entities contracts from 1,232 to 59-96, and inter-LLM rank correlations (0.55-0.68) generally exceed LLM-to-paper correlations (0.33-0.56), so the distortions are largely shared across models. Popularity baselines, BM25 retrieval calibration, and paper-level similarity tests confirm that the outputs are query-specific responses, but filtered through a narrower set of options. Researchers who rely on LLM suggestions without cross-checking therefore risk narrowing their methodological search space toward a more concentrated default.
Chinese Translation
大型语言模型（LLMs）在指导研究方法论方面的应用日益增加，但在最小提示下它们的默认方法倾向仍不清晰。在本研究中，我们对GPT-5.1、Gemini 3 Pro和DeepSeek-V3.2进行了提示，使用从1,000篇近期arXiv计算机科学论文中提取的研究问题，并将生成的方法建议与论文衍生的实验清单进行比较。由于我们仅提供研究问题，因此我们测量的差异反映的是初步建议，而非这些建议的最佳性。我们从两个来源提取结构化方法特征，将其映射到共享分类法中，并量化多个分类法维度的差异，包括模型提供者、数据集任务类型和评估指标类型。提供者选择的最强不平衡表现为Jensen-Shannon散度约为其他任何分类法维度的3-5倍。其他/学术单次出现模型的代表性不足23-24个百分点，而重复使用的学术/社区模型则略有过度代表（4-6个百分点）。LLMs建议的方法范围总体上也要狭窄得多：模型实体的有效数量从1,232减少到59-96，LLM间的排名相关性（0.55-0.68）通常超过LLM与论文之间的相关性（0.33-0.56），因此这些扭曲在模型之间大体上是共享的。流行基线、BM25检索校准和论文级相似性测试确认输出是特定于查询的响应，但经过更狭窄的选项过滤。因此，依赖LLM建议而不进行交叉检查的研究者，面临着将其方法搜索空间缩小到更集中默认值的风险。
View on arXiv Download PDF AI Translation
cs.CL / 13 / 2606.26196

From Structure to Synergy: A Survey of Vision-Language Perception Paradigm Evolution in Multimodal Large Language Models

从结构到协同：多模态大型语言模型中视觉-语言感知范式演变的调查
Sun, Haoxiang, Wang, Tao, Yuan, Li, Zhao, Jian, Lv, Jiancheng
Abstract
Multimodal Large Language Models (MLLMs) have recently made remarkable progress in unifying vision-language understanding and reasoning, especially following the introduction of models such as OpenAI's O-series and DeepSeek's R-series, which have driven a paradigm shift toward perception-centric intelligence. However, there remains a lack of systematic surveys that examine perception from a truly unified vision-language perspective -- one that treats vision and language as an inseparable modality. Existing reviews are often fragmented, focusing separately on either vision or language, and thus rarely capture the cross-modal evolution of perception as an integrated capability. To bridge this gap, we present the first systematic survey of unified vision-language perception in MLLMs. Specifically, we (1) formalize MLLM perception as an intrinsic, unified vision-language capability analogous to human innate perception, (2) introduce a five-stage taxonomy tracing the paradigm evolution of MLLM perception and survey representative methods and milestones at each phase, and (3) identify open challenges and outline promising research directions toward truly general, unified multimodal intelligence. We hope our study will provide both a foundational understanding and an actionable roadmap to foster further innovation on the path toward artificial general intelligence (AGI).
Chinese Translation
多模态大型语言模型（MLLMs）最近在统一视觉-语言理解和推理方面取得了显著进展，尤其是在OpenAI的O系列和DeepSeek的R系列等模型推出之后，这推动了以感知为中心的智能的范式转变。然而，目前缺乏系统性的调查，从真正统一的视觉-语言视角审视感知——即将视觉和语言视为不可分割的模态。现有的评审往往是片面的，分别关注视觉或语言，因此很少捕捉到作为综合能力的跨模态感知的演变。为填补这一空白，我们提出了首个针对MLLMs中统一视觉-语言感知的系统性调查。具体而言，我们（1）将MLLM感知形式化为一种内在的、统一的视觉-语言能力，类似于人类的先天感知；（2）引入一个五阶段分类法，追溯MLLM感知的范式演变，并调查每个阶段的代表性方法和里程碑；（3）识别开放挑战，并概述朝向真正通用的统一多模态智能的有前景的研究方向。我们希望我们的研究能够提供基础理解和可行的路线图，以促进在通向人工通用智能（AGI）道路上的进一步创新。
View on arXiv Download PDF AI Translation
cs.CL / 14 / 2606.26360

Phonetic and semantic analyses of spoken corpora of Beijing and Taiwan Mandarin indicate that the neutral tone is a lexical tone

对北京普通话和台湾普通话口语语料的语音和语义分析表明，中性调是一种词汇音调
Lu, Yuxin, Li, Zhexuan, Baayen, R. Harald
Abstract
The neutral, or floating, tone of Mandarin Chinese is a tone with an enigmatic set of properties. It has been described as a reduced tone, or as a tone that sometimes is lexically fixed but that can also be toneless. In two-syllable words, it is found only on the second syllable, but single-syllable words can also have the neutral tone. We present a corpus-based study of the phonetic realization of the neutral tone in spontaneous conversational speech corpora of Beijing Mandarin and Taiwan Mandarin. We show that the neutral tone has its own tonal target, just as the four lexical tones of Mandarin. We also show that disyllabic words with a neutral tone have pitch contours that have a pitch component that depends on the tone on the first syllable, just as has been observed for two-syllable words with a lexical tone on the second syllable (Chuang et al., 2026). Furthermore, words with a floating tone have word-specific pitch signatures, which have also been documented for single-syllable words (Jin et al., 2026) as well as two-syllable words (Lu et al., 2026b). These word-specific pitch signatures are shown to be predictable to some extent from words' contextualized embeddings, as previously reported for lexical tones (Chuang et al., 2026; Lu et al., 2026b). As there is also considerable variability in the realization of lexical tones, we propose that the neutral tone is, in fact, a lexical tone in both Taiwan Mandarin and Beijing Mandarin. We document both similarities and differences in the realization of the floating tone in these two varieties and provide evidence, using contextualized embeddings, that some of the observed differences may arise from differences in the meanings of the words as used in the two corpora.
Chinese Translation
普通话的中性调或浮动调是一种具有神秘特性的音调。它被描述为一种减弱的音调，或者是一种有时在词汇上固定但也可以无音调的音调。在双音节词中，它仅出现在第二个音节，但单音节词也可以有中性调。我们展示了一项基于语料库的研究，研究了北京普通话和台湾普通话自发对话语料中中性调的语音实现。我们表明，中性调有其自身的音调目标，就像普通话的四个词汇音调一样。我们还表明，带有中性调的双音节词的音高轮廓具有一个音高成分，该成分依赖于第一个音节的音调，这与观察到的第二个音节有词汇音调的双音节词相同（Chuang et al., 2026）。此外，具有浮动调的词具有特定于词的音高特征，这在单音节词（Jin et al., 2026）和双音节词（Lu et al., 2026b）中也有记录。这些特定于词的音高特征在一定程度上可以从词的上下文嵌入中预测，正如之前对词汇音调所报告的那样（Chuang et al., 2026; Lu et al., 2026b）。由于词汇音调的实现也存在相当大的变异性，我们提出中性调实际上是台湾普通话和北京普通话中的一种词汇音调。我们记录了这两种方言中浮动调实现的相似性和差异，并使用上下文嵌入提供证据，表明一些观察到的差异可能源于这两个语料库中所用词语的意义差异。
View on arXiv Download PDF AI Translation
cs.CL / 15 / 2606.26382

Charting the Growth of Social-Physical HRI (spHRI): A Systematic Review Pipeline Augmented by Small Language Models

社会物理人机交互（spHRI）增长的图谱：由小型语言模型增强的系统评价流程
Mohan, Mayumi, Chen, Ju-Hung, Block, Alexis E.
Abstract
Social-physical human-robot interaction (spHRI) has grown rapidly across robotics, human-computer interaction, human-robot interaction, and haptics. Yet, fragmented terminology and inconsistent methodologies make systematic synthesis difficult. To support scalable review practices, we evaluated the extent to which small language models (SLMs; < 1.5B parameters) can assist with title and abstract screening for a large spHRI systematic review. While no SLMs matched human reviewers' performance, the models operated locally and screened papers orders of magnitude faster. The combined SLM ensemble identified 39 papers reviewers missed, representing 10.29% of the final relevant dataset. These results demonstrate that SLMs can augment, rather than replace, expert reviewers and make large-scale literature reviews accessible and sustainable.
Chinese Translation
社会物理人机交互（spHRI）在机器人技术、人机交互、人机协作和触觉技术等领域迅速发展。然而，术语碎片化和方法不一致使得系统综合变得困难。为了支持可扩展的评审实践，我们评估了小型语言模型（SLMs；< 1.5B 参数）在大规模 spHRI 系统评价中对标题和摘要筛选的辅助程度。尽管没有任何 SLM 的表现能够匹敌人类评审者，但这些模型在本地操作并以数量级更快的速度筛选论文。组合的 SLM 集成识别了 39 篇评审者遗漏的论文，占最终相关数据集的 10.29%。这些结果表明，SLMs 可以增强而非替代专家评审者，使大规模文献评审变得可及和可持续。
View on arXiv Download PDF AI Translation
cs.CL / 16 / 2606.26403

ProfileFoundry: A Synthetic Person-Object Substrate for Privacy, Memory, and Tool-Use Evaluation in LLM Agent

ProfileFoundry：用于隐私、记忆和工具使用评估的合成人物-对象基础设施在大型语言模型代理中的应用
Selvam, Sriram, Ghosh, Anneswa
Abstract
Foundation-model research increasingly needs data about people: user state, personal histories, relationships, contact-like fields, documents, and longitudinal updates. Real user data is difficult to share, perturb, audit, or redistribute responsibly, while independently generated fake fields rarely preserve the cross-field and temporal consistency needed for controlled evaluation. We present PROFILEFOUNDRY, a deterministic generator and fixed reference release of 100,000 adult synthetic Person Objects across eight locales. Each object combines a typed current snapshot, household, family, and employer links, snapshot-aligned events, normalized relational views, and generation provenance. The release contains 709,228 events, 40,338 households, 52,491 employers, and 518,564 directed relationship edges. We report evidence in separate categories: selected population-marginal comparisons, per-object invariant checks, release-wide referential and temporal closure, and coincidence/provenance screens. PROFILEFOUNDRY is not a population-fidelity model, a rendered-text corpus, or a formal privacy mechanism. Instead, it is a responsible synthetic source layer for constructing downstream foundation-model evaluations involving memory, privacy, document understanding, record linkage, and agent state while keeping the synthetic person behind each artifact inspectable
Chinese Translation
基础模型研究日益需要关于人类的数据：用户状态、个人历史、关系、联系人字段、文档以及纵向更新。真实用户数据难以负责任地共享、扰动、审计或重新分发，而独立生成的虚假字段则很少能保持进行控制评估所需的跨领域和时间一致性。我们提出了PROFILEFOUNDRY，这是一个确定性的生成器和固定参考，包含100,000个成人合成人物对象，覆盖八个地区。每个对象结合了类型化的当前快照、家庭、家庭成员和雇主链接、快照对齐事件、规范化的关系视图以及生成来源。该发布包含709,228个事件、40,338个家庭、52,491个雇主和518,564个定向关系边。我们在不同类别中报告证据：选定的人口边际比较、每个对象的不变性检查、发布范围内的参考和时间闭合，以及巧合/来源筛选。PROFILEFOUNDRY不是一个人口保真模型、渲染文本语料库或正式的隐私机制。相反，它是一个负责任的合成源层，用于构建下游基础模型评估，涉及记忆、隐私、文档理解、记录链接和代理状态，同时保持每个工件背后的合成人物可供检查。
View on arXiv Download PDF AI Translation
cs.CL / 17 / 2606.26437

ConflictScore: Identifying and Measuring How Language Models Handle Conflicting Evidence

ConflictScore：识别和测量语言模型如何处理冲突证据
Liu, Siyi, Halfaker, Aaron, Roth, Dan, Xia, Patrick
Abstract
Existing metrics for factuality and faithfulness evaluate whether an answer is supported or contradicted by its grounding documents, but they fail to capture when both supporting and contradicting evidence coexist. We introduce ConflictScore, a novel metric that quantifies how well a model's response acknowledges conflicting evidence in its grounding documents. Our framework decomposes responses into atomic claims, labels each claim against each grounding document, and then aggregates these labels into two complementary measures: ConflictScore-Count (CS-C), the proportion of claims exhibiting conflicts, and ConflictScore-Ratio (CS-R), the balance between supporting and contradicting evidence. We develop ConflictBench, a benchmark covering diverse forms of conflicts such as ambiguity, contradiction, and divergent opinions, to systematically evaluate our metric. Experiments show that ConflictScore effectively detects overconfident claims across domains and can serve as a corrective feedback mechanism that improves truthfulness on TruthfulQA.
Chinese Translation
现有的事实性和忠实性评估指标主要评估答案是否得到其基础文献的支持或反驳，但未能捕捉到支持证据和反对证据共存的情况。我们提出了ConflictScore，这是一种新颖的指标，用于量化模型的响应在多大程度上承认其基础文献中的冲突证据。我们的框架将响应分解为原子声明，对每个声明与每个基础文献进行标注，然后将这些标注汇总为两个互补的度量：ConflictScore-Count (CS-C)，即表现出冲突的声明所占的比例，以及ConflictScore-Ratio (CS-R)，即支持证据与反对证据之间的平衡。我们开发了ConflictBench，这是一个涵盖模糊性、矛盾和不同意见等多种冲突形式的基准，以系统地评估我们的指标。实验表明，ConflictScore能够有效检测跨领域的过度自信声明，并可以作为一种纠正反馈机制，提高TruthfulQA上的真实度。
View on arXiv Download PDF AI Translation
cs.CL / 18 / 2606.26449

ProvenAI: Provenance-Native Traces of Evidence in Generated Answers

ProvenAI：生成答案中的证据来源原生追踪
Faizan, Mohammad, Alharthi, Dalal
Abstract
Retrieval-augmented systems routinely present citations alongside generated answers, yet a citation does not confirm that the corresponding source meaningfully shaped the output. This paper introduces ProvenAI, a framework that decomposes transparency in multi-hop question answering into three independently measurable layers: answer correctness, citation fidelity against benchmark supporting evidence, and per-document influence under leave-one-resource-out intervention. Targeting the HotpotQA distractor benchmark through a seven-stage pipeline covering data normalisation, retrieval indexing, citation-aware answer generation, attribution auditing, ablation-based influence estimation, batch evaluation, and interactive inspection, ProvenAI evaluates 7,405 validation examples drawn from a canonical corpus of 509,300 passages. The system achieves 53.53% answer accuracy alongside a mean citation-fidelity score of 71.55%, and a worked example surfaces what we call the citation-influence gap: a clean citation audit co-occurring with a profile in which one cited source registers only weak influence while seven uncited sources demonstrably shift the output. We formalise the relationship between the implemented surface proxy and a token-level KL-divergence target through a stated faithfulness condition, ground the framework in causal-mediation analysis and database-provenance theory, and discuss how the three measurement layers compose with cryptographic provenance architectures emerging in autonomous scientific discovery. ProvenAI establishes that meaningful transparency in retrieval-grounded QA requires traceable links across retrieved, cited, and behaviourally influential evidence as three distinct, independently measured layers.
Chinese Translation
检索增强系统通常在生成答案的同时提供引用，但引用并不能确认相应的来源在多大程度上影响了输出。本文介绍了ProvenAI，一个将多跳问答中的透明度分解为三个独立可测量层次的框架：答案正确性、与基准支持证据的引用忠实度，以及在逐一资源剔除干预下的每个文档影响。ProvenAI通过一个涵盖数据标准化、检索索引、引用感知答案生成、归因审计、基于消融的影响估计、批量评估和交互式检查的七阶段流程，针对HotpotQA干扰基准评估了从一个包含509,300段落的标准语料库中提取的7,405个验证示例。该系统实现了53.53%的答案准确率和71.55%的平均引用忠实度得分，并且一个具体示例揭示了我们所称的引用影响差距：一次干净的引用审计与一个仅有弱影响的被引用来源的档案共存，而七个未被引用的来源显著改变了输出。我们通过一个声明的忠实性条件形式化了实施的表面代理与令牌级KL散度目标之间的关系，将该框架基于因果中介分析和数据库来源理论，并讨论这三个测量层如何与在自主科学发现中出现的密码学来源架构相结合。ProvenAI确立了在检索基础的问答中实现有意义的透明度需要在检索到的、被引用的和行为上有影响的证据之间建立可追溯的联系，这三个层次是独立测量的。
View on arXiv Download PDF AI Translation
cs.CL / 19 / 2606.26452

AnySimLite: A Lightweight Few-Shot Similarity Encoder for On-Device Speech-Adjacent Classification

AnySimLite：一种轻量级的少样本相似性编码器，用于设备端语音相关分类
Ghosh, Sourav, Bhatia, Yash, Goyal, Keshav, Bagri, Sahil Singh, Shariff, Mohamed Akram Ulla, Shanmugam, Saravana Balaji
Abstract
To minimize privacy concerns and inference latency on edge devices like smartphones, lightweight on-device models remain important for end-user applications. Many of these applications involve natural language classification, but deploying multiple specialized models creates a memory footprint challenge. We investigate: Can a single lightweight architecture solve multiple Speech-Adjacent (SA) classification tasks through reduction to a nuanced text similarity formulation? We propose AnySimLite, a lightweight similarity encoder that combines word-level and character-level channels. Together with a dataset transformation strategy, we evaluate AnySimLite across multiple SA classification tasks and show that it consistently achieves state-of-the-art (SOTA) or SOTA-competitive performance in few-shot settings while maintaining a low memory footprint. Even in the worst case, the performance drop remains below 7% while using $<\frac{1}{250}^{\mathrm{th}}$ of the model size of the SOTA qLLaMA_LoRA-7B baseline.
Chinese Translation
为了减少隐私问题和在智能手机等边缘设备上的推理延迟，轻量级的设备端模型在最终用户应用中仍然至关重要。这些应用中的许多涉及自然语言分类，但部署多个专门模型会带来内存占用的挑战。我们研究：是否可以通过将多个语音相关（Speech-Adjacent，SA）分类任务简化为细致的文本相似性公式，来用单一轻量级架构解决这些任务？我们提出了AnySimLite，这是一种轻量级相似性编码器，结合了词级和字符级通道。结合数据集转换策略，我们在多个SA分类任务上评估了AnySimLite，并显示其在少样本设置中始终实现了最先进（SOTA）或竞争性的SOTA性能，同时保持较低的内存占用。即使在最坏情况下，性能下降也保持在7%以下，同时使用的模型大小仅为SOTA qLLaMA_LoRA-7B基线的$< rac{1}{250}^{ ext{th}}$。
View on arXiv Download PDF AI Translation
cs.CL / 20 / 2606.26466

Soft Token Alignment for Cross-Lingual Reasoning

跨语言推理的软标记对齐
He, Jiayi, Park, Jungsoo, Xu, Wei, Ritter, Alan
Abstract
Multilingual large language models often produce inconsistent reasoning and answers for semantically equivalent prompts in different languages. Prior work suggests that intermediate representations can be relatively language-agnostic, but generation becomes increasingly language-specific as models commit to discrete output tokens. This is problematic because language-specific lexical choices can cause semantically equivalent reasoning paths to diverge across languages. These divergences motivate searching for a cross-lingual alignment signal that is less tied to any single vocabulary item or script. We propose SOLAR, an auxiliary objective for supervised fine-tuning that aligns soft-token representations across languages, using English as a pivot. Soft tokens are probability-weighted mixtures over the vocabulary embeddings, yielding continuous representations that can aggregate information from semantically related tokens across languages. We then align each non-English soft-token summary to its English counterpart in the shared embedding space. Across four multilingual reasoning benchmarks, SOLAR improves accuracy by up to +17.7 points over the base model and +3.8 over standard supervised fine-tuning, with the largest gains on low-resource languages. SOLAR also strengthens final-layer cross-lingual similarity and substantially reduces language-cluster separability, suggesting that aligning soft-token representations helps preserve shared semantic structure during multilingual reasoning.
Chinese Translation
多语言大型语言模型在不同语言中对语义等价的提示常常产生不一致的推理和答案。先前的研究表明，中间表示可以相对不依赖于语言，但随着模型对离散输出标记的承诺，生成变得越来越特定于语言。这是一个问题，因为特定于语言的词汇选择可能导致语义等价的推理路径在不同语言中出现分歧。这些分歧促使我们寻找一种跨语言对齐信号，该信号不那么依赖于任何单一的词汇项或书写系统。我们提出了SOLAR，这是一种用于监督微调的辅助目标，旨在对齐跨语言的软标记表示，以英语作为中介。软标记是对词汇嵌入的概率加权混合，产生连续的表示，可以聚合来自不同语言中语义相关标记的信息。然后，我们将每个非英语软标记摘要与其在共享嵌入空间中的英语对应项对齐。在四个多语言推理基准测试中，SOLAR使准确率比基础模型提高了最多17.7个百分点，比标准监督微调提高了3.8个百分点，尤其在低资源语言上取得了最大的增益。SOLAR还增强了最终层的跨语言相似性，并显著减少了语言聚类的可分性，这表明对齐软标记表示有助于在多语言推理过程中保持共享的语义结构。
View on arXiv Download PDF AI Translation
cs.CL / 21 / 2606.26481

Extracting Problem and Method Sentence from Scientific Papers: A Context-enhanced Transformer Using Formulaic Expression Desensitization

从科学论文中提取问题和方法句子：一种基于公式表达去敏感化的上下文增强型变换器
Zhang, Yingyi, Zhang, Chengzhi
Abstract
Billions of scientific papers lead to the need to identify essential parts from the massive text. Scientific research is an activity from putting forward problems to using methods. To learn the main idea from scientific papers, we focus on extracting problem and method sentences. Annotating sentences within scientific papers is labor-intensive, resulting in small-scale datasets that limit the amount of information models can learn. This limited information leads models to rely heavily on specific forms, which in turn reduces their generalization capabilities. This paper addresses the problems caused by small-scale datasets from three perspectives: increasing dataset scale, reducing dependence on specific forms, and enriching the information within sentences. To implement the first two ideas, we introduce the concept of formulaic expression (FE) desensitization and propose FE desensitization-based data augmenters to generate synthetic data and reduce models' reliance on FEs. For the third idea, we propose a context-enhanced transformer that utilizes context to measure the importance of words in target sentences and to reduce noise in the context. Furthermore, this paper conducts experiments using large language model (LLM) based in-context learning (ICL) methods. Quantitative and qualitative experiments demonstrate that our proposed models achieve a higher macro F1 score compared to the baseline models on two scientific paper datasets, with improvements of 3.71% and 2.67%, respectively. The LLM based ICL methods are found to be not suitable for the task of problem and method extraction.
Chinese Translation
数十亿篇科学论文使得从大量文本中识别重要部分的需求日益增加。科学研究是一项从提出问题到使用方法的活动。为了从科学论文中学习主要思想，我们专注于提取问题和方法句子。对科学论文中的句子进行标注是一项劳动密集型工作，导致小规模数据集的产生，从而限制了模型可以学习的信息量。这种信息的限制使得模型过于依赖特定形式，从而降低了它们的泛化能力。本文从三个方面解决小规模数据集带来的问题：增加数据集规模、减少对特定形式的依赖以及丰富句子中的信息。为了实现前两个想法，我们引入了公式表达（Formulaic Expression, FE）去敏感化的概念，并提出基于FE去敏感化的数据增强器，以生成合成数据并减少模型对FE的依赖。对于第三个想法，我们提出了一种上下文增强型变换器，利用上下文来衡量目标句子中单词的重要性，并减少上下文中的噪声。此外，本文使用基于大语言模型（Large Language Model, LLM）的上下文学习（In-Context Learning, ICL）方法进行了实验。定量和定性实验表明，我们提出的模型在两个科学论文数据集上相较于基线模型实现了更高的宏观F1得分，分别提高了3.71%和2.67%。研究发现，基于LLM的ICL方法并不适合问题和方法提取任务。
View on arXiv Download PDF AI Translation
cs.CL / 22 / 2606.26485

Utilizing Cognitive Signals Generated during Human Reading to Enhance Keyphrase Extraction from Microblogs

利用人类阅读过程中产生的认知信号增强微博的关键词提取
Yan, Xinyi, Zhang, Yingyi, Zhang, Chengzhi
Abstract
Microblogging platforms generate massive amounts of short, noisy, and dispersed user content, making automatic keyphrase extraction (AKE) an important but challenging task. Prior studies have used eye-tracking signals to improve microblog-based AKE because such signals reflect readers' attention to salient words. However, eye tracking alone is limited by physiological, acquisition, and feature-decoding constraints. To address this issue, we investigate whether electroencephalogram (EEG) signals can complement eye-tracking signals for AKE. Using the ZuCo cognitive language processing corpus, we select 8 EEG features and 17 eye-tracking features and incorporate them into microblog-based AKE models. To reduce possible distortion of cognitive signals by model structures, we inject these features into the input of the soft-attention layer and the query vectors of the self-attention layer. We then evaluate different combinations of cognitive signals across AKE models. The results show that cognitive signals produced during reading consistently improve AKE performance, regardless of feature combinations and model architectures. EEG features bring the largest gains, while combining EEG and eye-tracking features yields performance between the two individual signal types, suggesting partial complementarity but also possible redundancy or noise. These findings indicate that EEG signals provide useful cognitive evidence for microblog-based AKE and that multimodal cognitive signals deserve further investigation.
Chinese Translation
微博平台生成大量短小、嘈杂且分散的用户内容，使得自动关键词提取（AKE）成为一项重要但具有挑战性的任务。先前的研究利用眼动追踪信号来改善基于微博的AKE，因为这些信号反映了读者对显著词汇的注意力。然而，单靠眼动追踪受到生理、获取和特征解码等限制。为了解决这一问题，我们研究脑电图（EEG）信号是否可以补充眼动追踪信号以改善AKE。通过使用ZuCo认知语言处理语料库，我们选择了8个EEG特征和17个眼动追踪特征，并将其纳入基于微博的AKE模型中。为了减少模型结构对认知信号可能造成的扭曲，我们将这些特征注入到软注意力层的输入和自注意力层的查询向量中。然后，我们评估不同认知信号组合在AKE模型中的表现。结果表明，在阅读过程中产生的认知信号始终能提高AKE性能，无论特征组合和模型架构如何。EEG特征带来了最大的提升，而结合EEG和眼动追踪特征的表现介于两种单独信号类型之间，表明部分互补性，但也可能存在冗余或噪声。这些发现表明EEG信号为基于微博的AKE提供了有用的认知证据，并且多模态认知信号值得进一步研究。
View on arXiv Download PDF AI Translation
cs.CL / 23 / 2606.26487

Speaking Numbers to LLMs: Multi-Wavelet Number Embeddings for Time Series Forecasting

将数字传递给大型语言模型：用于时间序列预测的多小波数字嵌入
Cao, Defu, Lei, Zijie, Weng, Muyan, Sun, Jiao, Liu, Yan
Abstract
Large language models (LLMs) are attractive for context-aware time series forecasting because they can integrate heterogeneous textual signals, yet their discrete, language-oriented tokenization and embedding interfaces are misaligned with continuous numerical values, often harming numerical ordering and forecasting reliability. We propose TempoWave, a plug-and-play temporal wavelet digit interface that maps each scalar observation into digit-wise embeddings constructed from multi-wavelet, multi-scale coefficients. By directly overriding standard token representations, TempoWave seamlessly exposes both fine-grained local fluctuations and macro global structures in a transformer-compatible form, ensuring that precise numerical formatting, distinct digit identity, and robustness to common normalization operations are maintained throughout the LLM pipeline. Experiments across five context-enriched forecasting benchmarks demonstrate that TempoWave consistently improves LLM-based forecasters over standard numeric tokenization and alternative embedding interfaces, achieving a new state-of-the-art. These results highlight the numeric interface as a key bottleneck and suggest that principled multi-resolution embeddings can better couple LLMs' contextual reasoning with precise forecasting. Our code is available at https://github.com/DC-research/TempoWAVE and our model can be accessed at https://huggingface.co/Melady/TempoWAVE.
Chinese Translation
大型语言模型（LLMs）在上下文感知的时间序列预测中具有吸引力，因为它们能够整合异构的文本信号。然而，它们离散的、面向语言的标记化和嵌入接口与连续的数值不匹配，常常损害数值排序和预测可靠性。我们提出了TempoWave，一种即插即用的时间小波数字接口，将每个标量观测值映射为由多小波、多尺度系数构成的数字嵌入。通过直接覆盖标准标记表示，TempoWave无缝地以与变压器兼容的形式揭示了细粒度的局部波动和宏观的全球结构，确保在整个LLM管道中保持精确的数值格式、独特的数字身份以及对常见归一化操作的鲁棒性。在五个上下文增强的预测基准测试中的实验表明，TempoWave在基于LLM的预测器上始终优于标准数值标记化和替代嵌入接口，达到了新的最先进水平。这些结果突显了数值接口作为关键瓶颈，并表明原则性多分辨率嵌入可以更好地将LLM的上下文推理与精确预测结合起来。我们的代码可在 https://github.com/DC-research/TempoWAVE 获取，模型可在 https://huggingface.co/Melady/TempoWAVE 访问。
View on arXiv Download PDF AI Translation
cs.CL / 24 / 2606.26489

Comparing BERT Sentence-Pair Classification and Few-Shot LLM Prompting for Detecting Threat and Solution Framing in German Climate News

比较 BERT 句子对分类与少量样本大语言模型提示在德语气候新闻中检测威胁与解决方案框架的效果
Adam, Raven, Maier, David, Kogler, Marie
Abstract
News media play a central role in shaping public perceptions of climate change, and whether coverage emphasizes threats or solutions has measurable effects on audience engagement and policy support. Automated detection of these framing patterns at the sentence level would allow researchers to analyze large corpora that are infeasible to code manually. We present a systematic comparison of two approaches for classifying sentences from German-language climate news articles as threat-oriented, solution-oriented, both, or neither. The first approach uses few-shot prompting with an open-weights large language model (Llama 4 Maverick), employing chain-of-thought reasoning and structured output with confidence scoring. The second approach fine-tunes a German BERT model (deepset/gbert-large) for sentence-pair classification, where the preceding sentence provides contextual information for the target sentence. Both approaches implement two independent binary classifiers, one for threat framing and one for solution framing. We evaluate both methods on a corpus of 440 Austrian newspaper articles that were manually coded following a detailed coding scheme developed with domain experts. The fine-tuned BERT classifiers achieve an F1 score of 0.83 for both the threat and solution tasks, while the LLM-based classifiers reach an F1 of 0.78. An ablation study confirms that providing the preceding sentence as context improves BERT classification performance substantially compared to single-sentence input. These results contribute to the growing body of work comparing fine-tuned encoder models with prompted generative models for text classification in computational social science.
Chinese Translation
新闻媒体在塑造公众对气候变化的认知中发挥着核心作用，而报道是否强调威胁或解决方案对观众参与度和政策支持具有可测量的影响。在句子层面自动检测这些框架模式将使研究人员能够分析大规模语料库，这些语料库手动编码是不可行的。我们对两种方法进行了系统比较，以将德语气候新闻文章中的句子分类为以威胁为导向、以解决方案为导向、两者皆是或两者皆非。第一种方法使用少量样本提示与开放权重的大语言模型（Llama 4 Maverick），采用链式思维推理和结构化输出及置信评分。第二种方法对德语 BERT 模型（deepset/gbert-large）进行微调，以进行句子对分类，其中前一句为目标句子提供上下文信息。这两种方法均实现了两个独立的二元分类器，一个用于威胁框架，另一个用于解决方案框架。我们在440篇经过手动编码的奥地利报纸文章的语料库上评估了这两种方法，编码方案是与领域专家共同开发的。微调后的 BERT 分类器在威胁和解决方案任务上均达到了 0.83 的 F1 分数，而基于 LLM 的分类器则达到了 0.78 的 F1 分数。消融研究确认，与单句输入相比，提供前一句作为上下文显著提高了 BERT 分类性能。这些结果为比较微调编码器模型与提示生成模型在计算社会科学中文本分类的日益增长的研究提供了贡献。
View on arXiv Download PDF AI Translation
cs.CL / 25 / 2606.26493

Nemotron-TwoTower: Diffusion Language Modeling with Pretrained Autoregressive Context

Nemotron-TwoTower：具有预训练自回归上下文的扩散语言建模
Reda, Fitsum, Kamalu, John, Waleffe, Roger, Patwary, Mostofa, Shoeybi, Mohammad, Catanzaro, Bryan
Abstract
Diffusion language models offer a promising alternative to autoregressive models due to their potential for parallel and iterative generation. However, existing approaches use a single network for both context representation and iterative denoising, forcing one model to serve both roles and limiting its capacity for either role. We propose TwoTower, a block-wise autoregressive diffusion model that decouples these roles into two towers: a frozen AR context tower that causally processes clean tokens, and a trainable diffusion denoiser tower with bidirectional block attention that refines noisy blocks via cross-attention to the context. Built on Nemotron-3-Nano-30B-A3B, an open-weight 30B hybrid Mamba-Transformer MoE model, and trained on approximately 2.1T tokens, Nemotron-TwoTower retains 98.7% of the autoregressive baseline's quality while offering 2.42X higher wall-clock generation throughput. We release the code and model weights at https://huggingface.co/collections/nvidia/nemotron-twotower.
Chinese Translation
扩散语言模型由于其并行和迭代生成的潜力，提供了自回归模型的有希望的替代方案。然而，现有的方法使用单一网络同时进行上下文表示和迭代去噪，这迫使一个模型承担两个角色，限制了其在任一角色上的能力。我们提出了TwoTower，一个块级自回归扩散模型，将这两个角色解耦为两个塔：一个冻结的自回归上下文塔，因果地处理干净的标记，以及一个可训练的扩散去噪塔，具有双向块注意力，通过对上下文的交叉注意力来精炼嘈杂的块。基于Nemotron-3-Nano-30B-A3B，一个开放权重的30B混合Mamba-Transformer MoE模型，并在大约2.1T标记上进行训练，Nemotron-TwoTower在保持98.7%自回归基线质量的同时，提供了2.42倍更高的实际生成吞吐量。我们在https://huggingface.co/collections/nvidia/nemotron-twotower发布代码和模型权重。
View on arXiv Download PDF AI Translation
cs.CL / 26 / 2606.26511

Temporal Validity in Retrieval Memory: Eliminating Stale-Fact Errors for AI Agents over Evolving Knowledge

检索记忆中的时间有效性：消除人工智能代理在不断演变的知识中出现的过时事实错误
Yadav, Neeraj
Abstract
Retrieval-augmented generation (RAG) gives agents access to accumulated knowledge, but has no model of time. When a fact changes (e.g., a function is renamed or API restructured), RAG retrieves both the stale and current value with near-identical embedding similarity. The agent then either abstains or serves the superseded fact. We show this is a structural problem: on a calibrated dataset, cosine similarity distinguishes a contradicted fact from a duplicated one with AUROC 0.59 (near chance), as contradictions are often more embedding-similar to the original than rephrased duplicates. We present MemStrata, a retrieval memory maintaining temporal validity. It stores facts like RAG, preserving static recall, but when a fact's value is contradicted, a deterministic (subject, relation, object) supersession rule retires the stale value in a bi-temporal ledger - with no similarity threshold and no LLM call. Across six benchmarks run locally with a 7B model, MemStrata ties RAG on static knowledge and reaches 0.95-1.00 accuracy on evolving knowledge (where RAG reaches 0.20-0.47). The central result is the stale-fact-error rate: when required to answer, RAG serves superseded values 15-40% of the time; MemStrata drives this to ~0%, a failure class RAG cannot avoid. MemStrata achieves this at retrieval latency (~2.1s) versus ~16-18s for LLM-reranking baselines. We release the harness, datasets, and a marker-free evaluation protocol for memory under knowledge evolution.
Chinese Translation
检索增强生成（RAG）为代理提供了访问累积知识的能力，但没有时间模型。当一个事实发生变化（例如，函数被重命名或API结构被重组）时，RAG几乎以相同的嵌入相似度检索到过时和当前的值。代理随后要么不作回应，要么提供被取代的事实。我们展示了这是一个结构性问题：在一个经过校准的数据集上，余弦相似度以0.59的AUROC区分了被矛盾的事实和重复的事实（接近随机），因为矛盾的事实通常在嵌入上与原始事实更为相似，而不是改述的重复事实。我们提出了MemStrata，这是一种保持时间有效性的检索记忆。它像RAG一样存储事实，保持静态召回，但当一个事实的值被矛盾时，确定性的（主语、关系、宾语）取代规则在双时间账本中退役过时值——没有相似度阈值，也不需要调用大型语言模型（LLM）。在使用7B模型本地运行的六个基准测试中，MemStrata在静态知识上与RAG持平，并在不断演变的知识上达到0.95-1.00的准确率（而RAG仅达到0.20-0.47）。核心结果是过时事实错误率：在需要回答时，RAG提供被取代的值的比例为15-40%；而MemStrata将这一比例降低到约0%，这是RAG无法避免的失败类别。MemStrata在检索延迟方面实现了约2.1秒，而LLM重排序基线则为约16-18秒。我们发布了用于知识演变下记忆的工具、数据集和无标记评估协议。
View on arXiv Download PDF AI Translation
cs.CL / 27 / 2606.26522

Assessing Post-Reform Changes in Risk Disclosure Quality with a Multidimensional Text Analysis Approach

评估改革后风险披露质量变化的多维文本分析方法
Aikawa, Nobuhiro, Yoshida, Mitsuo
Abstract
While corporate narrative disclosures provide crucial information to capital markets, comprehensively evaluating their qualitative changes over time remains challenging. Narrative text is inherently multidimensional, meaning that an improvement in one textual dimension often occurs alongside changes in others. To capture these underlying dynamics, we propose a longitudinal text analysis approach combining Japanese-language NLP metric extraction with paired testing, shift function analysis, and inter-metric correlation. Our framework extends prior indicator sets by incorporating a cross-section relevance indicator to measure topical alignment between risk disclosures and management strategies. Applying this approach to evaluate Japan's 2019 disclosure reforms, we analyze 19,770 firm-year observations over a 10-year period (FY2015-FY2024). The joint analysis reveals complex shifts in disclosure patterns that are frequently masked by conventional single-indicator methods. Specifically, we find that while disclosure volume increased substantially, it was accompanied by a decline in readability. Furthermore, although the overall information structure improved, specific descriptive quality stagnated, and the degree of adaptation varied across market segments.
Chinese Translation
尽管企业叙述性披露为资本市场提供了重要信息，但全面评估其随时间变化的定性变化仍然具有挑战性。叙述文本本质上是多维的，这意味着在一个文本维度上的改善往往伴随着其他维度的变化。为了捕捉这些潜在的动态，我们提出了一种纵向文本分析方法，结合了日语自然语言处理（NLP）指标提取、配对测试、变化函数分析和指标间相关性分析。我们的框架通过引入一个交叉相关性指标，扩展了先前的指标集，以测量风险披露与管理策略之间的主题一致性。我们将该方法应用于评估日本2019年的披露改革，分析了2015财年至2024财年间的19,770个公司年度观察数据。联合分析揭示了披露模式的复杂变化，这些变化常常被传统的单一指标方法所掩盖。具体而言，我们发现尽管披露量大幅增加，但可读性却有所下降。此外，尽管整体信息结构有所改善，但特定描述质量停滞不前，且适应程度在不同市场细分中存在差异。
View on arXiv Download PDF AI Translation
cs.CL / 28 / 2606.26529

The Inattentional Gap: Task-Conditioned Language and Vision Models Omit the Safety-Critical Signals They Can Otherwise Report

注意力缺口：任务条件下的语言和视觉模型忽略它们本可以报告的安全关键信号
Shin, Kwan Soo
Abstract
AI safety is evaluated by how reliably a model detects the hazards it is told to find, yet accidents often arise from the hazard no one specified. We show that conditioning a language or vision model on a narrow task suppresses its reporting of co-present, safety-critical signals it can otherwise report, a machine analogue of human inattentional blindness arising from a different mechanism. Across radiology and driving text scenarios and chest-radiograph vision tasks, suppression appeared in every model tested, did not diminish with scale, persisted in a reasoning model, and varied more by model family than by size, while the same models reported these signals at substantially higher rates when unconstrained. We name this dissociation the Inattentional Gap and argue that it decouples measured benchmark safety from real-world safety: a system can score near-perfectly on the hazards an evaluation specifies while remaining blind to those that cause harm.
Chinese Translation
人工智能安全的评估基于模型检测其被告知寻找的危险的可靠性，然而事故往往源于未被指定的危险。我们展示了在狭窄任务上对语言或视觉模型进行条件化会抑制其报告共存的安全关键信号的能力，这是一种与人类注意力盲点不同机制的机器类比。在放射学和驾驶文本场景以及胸部X光视觉任务中，抑制现象在每个测试的模型中均有出现，且随着规模的增加并未减弱，持续存在于推理模型中，并且不同模型家族之间的差异大于模型的大小，而同样的模型在不受限制时报告这些信号的比率显著更高。我们将这种分离称为注意力缺口，并认为它将测量的基准安全性与现实世界的安全性解耦：一个系统可以在评估指定的危险上得分接近完美，同时对造成伤害的危险视而不见。
View on arXiv Download PDF AI Translation
cs.CL / 29 / 2606.26530

\textsc{DiARC}: Distinguishing Positive and Negative Samples Helps Improving ARC-like Reasoning Ability of Large Language Models

DiARC：区分正负样本有助于提升大型语言模型的ARC类推理能力
Yang, Yuxuan, Li, Feiyang, Wang, Yile
Abstract
The Abstraction and Reasoning Corpus (ARC;~\citealp{chollet2019measure}) contains tasks that require summarizing patterns from limited grid samples and predicting output grids. Recently, many large language model based approaches have attempted to transform it into a text-based reasoning task. However, methods based on open-source models have generally yielded unsatisfactory results, while those relying on closed-source models are too costly. Current efforts mainly focus on data augmentation, constructing ARC-like data for more comprehensive supervised fine-tuning. In this work, we argue that solving ARC-like problems requires not only \textit{positive} sample supervision but also the ability to improve model reasoning by distinguishing \textit{negative} samples. To this end, we draw on the idea of preference alignment and propose \textsc{DiARC}, a method that constructs preference pairs to enable the model to distinguish between them. Specifically, we propose three ways to construct negative samples, including output-level visual transformations, DSL-level rule inversion, and task-specific rule editing. The resulting negative samples provide informative near-miss alternatives while keeping the observed demonstrations unchanged. Experimental results across multiple ARC-like benchmarks show that \textsc{DiARC} consistently improves performance over baseline models. The code is released at https://github.com/szu-tera/DiARC.
Chinese Translation
抽象与推理语料库（ARC;~ extcite{chollet2019measure}）包含需要从有限的网格样本中总结模式并预测输出网格的任务。最近，许多基于大型语言模型的方法尝试将其转化为基于文本的推理任务。然而，基于开源模型的方法通常效果不佳，而依赖闭源模型的方法成本过高。目前的努力主要集中在数据增强上，构建ARC类数据以进行更全面的监督微调。在本研究中，我们认为解决ARC类问题不仅需要 extit{正}样本的监督，还需要通过区分 extit{负}样本来提升模型推理能力。为此，我们借鉴了偏好对齐的思想，提出了 extsc{DiARC}，一种构建偏好对以使模型能够区分它们的方法。具体而言，我们提出了三种构建负样本的方法，包括输出级别的视觉变换、DSL级别的规则反转和任务特定的规则编辑。生成的负样本提供了信息丰富的近似替代品，同时保持观察到的示例不变。在多个ARC类基准测试中的实验结果表明， extsc{DiARC}在性能上始终优于基线模型。代码已发布在 https://github.com/szu-tera/DiARC。
View on arXiv Download PDF AI Translation
cs.CL / 30 / 2606.26560

Erase-then-Delta Attention: Decoupling Erase and Write Addresses in Delta-Rule Linear Attention

先抹除再更新注意力：在 Delta 规则线性注意力中解耦抹除和写入地址
Li, Xiao, Zhang, Chengruidong, Luo, Hao, Lin, Xi, Wang, Zekun, Qiu, Zihan, Mao, Yunfei, Chen, Langshi, Yuan, Man, Sun, Minmin, Jiang, Huiqiang, Zhang, Siqi, Men, Rui, Hu, Wei, Cheng, Gong, Zheng, Bo, Liu, Dayiheng, Zhou, Jingren
Abstract
Delta-rule linear attention improves recurrent memory updates by correcting what is already stored at the current write address before writing new content. However, the active correction is still anchored to that same write address. As a result, stale information stored at a different address cannot be actively removed before new content is written elsewhere. We propose Erase-then-Delta Attention (EDA), a memory update rule that decouples where to erase from where to write. The key insight is that recurrent memory models should not only correct the current write, but also selectively suppress outdated memory at an independently chosen address. Concretely, our method first applies a targeted erase step along a learned erase direction, and then performs the standard delta-style corrective write along the current write direction. This preserves the corrective behavior of delta-rule updates while expanding their memory-management capacity. Language-model pretraining experiments across dense 2.5B and MoE 25B-A2.8B model families show that EDA performs best in both settings. The gain persists after 80B-token long-context midtraining of the MoE models, where EDA also performs best in long-context evaluations from 4k to 128k contexts. A compact update analysis and memory-state probes suggest why: EDA keeps the delta-rule corrective write intact while allocating an additional cleanup path most strongly when passive decay is weak. These results suggest that recurrent memory models should decide not only what to write, but also what stale information to erase and where.
Chinese Translation
Delta 规则线性注意力通过在写入新内容之前修正当前写入地址上已存储的信息，从而改善了递归记忆更新。然而，主动修正仍然依赖于同一写入地址。因此，存储在不同地址的过时信息在新内容写入其他地方之前无法主动移除。我们提出了先抹除再更新注意力（Erase-then-Delta Attention, EDA），这是一种将抹除位置与写入位置解耦的记忆更新规则。关键的见解是，递归记忆模型不仅应修正当前写入，还应选择性地抑制在独立选择的地址上的过时记忆。具体而言，我们的方法首先沿着学习到的抹除方向应用有针对性的抹除步骤，然后沿着当前写入方向执行标准的 Delta 风格修正写入。这保留了 Delta 规则更新的修正行为，同时扩展了其记忆管理能力。在密集的 2.5B 和 MoE 25B-A2.8B 模型系列的语言模型预训练实验中，EDA 在这两种设置中表现最佳。在 MoE 模型的 80B 令牌长上下文中期训练后，这一增益依然存在，其中 EDA 在从 4k 到 128k 上下文的长上下文评估中也表现最佳。紧凑的更新分析和记忆状态探测表明了原因：当被动衰减较弱时，EDA 保持了 Delta 规则修正写入的完整性，同时分配了额外的清理路径。这些结果表明，递归记忆模型不仅应决定写入什么，还应决定抹除哪些过时信息以及抹除的位置。
View on arXiv Download PDF AI Translation
cs.CL / 31 / 2606.26571

Zero-shot Tweet-Level Stance Detection Enhanced by External Knowledge and Reflective Chain-of-Thought Reasoning

通过外部知识和反思性链式思维推理增强的零样本推文级立场检测
Huang, Yiju, Wang, Wenxian, Zhou, Lijun, Tang, Rui, Lan, Xiao, Zhang, Tao, Wang, Haizhou
Abstract
Zero-shot tweet-level stance detection confronts two primary challenges: (1) mitigating the context sparsity inherent in short texts, and (2) establishing the relevance between implicit targets and textual content. While existing methods primarily focus on incorporating external knowledge, they neglect the intrinsic semantic cues embedded within key intra-textual entities. Furthermore, current models exhibit limited capability in determining the relevance of unseen targets to the given text, thereby struggling to differentiate between "neutral" and "irrelevant" stance labels. To address these issues, we first construct a four-class, multi-topic Japanese tweet dataset. To our knowledge, this is the first Japanese tweet-level dataset for stance detection. We then propose KIRP, a zero-shot stance detection framework. It integrates external knowledge with entity reorganization for data augmentation and employs prompt chaining for reasoning. Specifically, the framework incorporates knowledge graphs to supplement and reorganize key textual entities, while reflective Chain-of-Thought (CoT) reasoning extracts and validates implicit targets. To better distinguish "neutral" from "irrelevant" labels, we adopt stance-aware contrastive learning to capture discriminative features and design a three-layer iterative prototype network for fine-grained classification. Experimental results on SemEval-2016, WT-WT, and KIRP-D show that KIRP achieves state-of-the-art performance. KIRP obtains F1 scores of 84.05% (three-class) on SemEval-2016, and 84.99% and 79.18% (four-class) on WT-WT and KIRP-D, respectively.
Chinese Translation
零样本推文级立场检测面临两个主要挑战：（1）减轻短文本固有的上下文稀疏性，以及（2）建立隐含目标与文本内容之间的相关性。现有方法主要集中在整合外部知识上，但忽视了嵌入在关键文本实体中的内在语义线索。此外，当前模型在确定未见目标与给定文本的相关性方面能力有限，因此难以区分“中立”和“无关”立场标签。为了解决这些问题，我们首先构建了一个四类多主题的日语推文数据集。据我们所知，这是第一个用于立场检测的日语推文级数据集。然后，我们提出了KIRP，一个零样本立场检测框架。它将外部知识与实体重组相结合以进行数据增强，并采用提示链式推理进行推理。具体而言，该框架结合知识图谱来补充和重组关键文本实体，同时反思性链式思维（Chain-of-Thought, CoT）推理提取和验证隐含目标。为了更好地区分“中立”和“无关”标签，我们采用立场感知对比学习来捕捉区分特征，并设计了一个三层迭代原型网络以进行细粒度分类。在SemEval-2016、WT-WT和KIRP-D上的实验结果表明，KIRP达到了最先进的性能。KIRP在SemEval-2016上获得了84.05%（三类）的F1分数，在WT-WT和KIRP-D上分别获得了84.99%和79.18%（四类）的F1分数。
View on arXiv Download PDF AI Translation
cs.CL / 32 / 2606.26618

Closing the Quality Gap in Low-Resource Text-to-Speech: LoRA Fine-Tuning of VoxCPM2 for Khmer and Korean

缩小低资源文本到语音中的质量差距：对VoxCPM2进行LoRA微调以适应高棉语和韩语
Pov, Phannet, Chhoun, Sovandara, Park, Hyun Woo, Cho, Wan-Sup, Khoeurn, Saksonita
Abstract
Large pretrained text-to-speech (TTS) models sound almost human for well-resourced languages, but much worse for languages that are rare in their training data. We study this quality gap for Khmer and Korean using VoxCPM2, a 2.4B-parameter, tokenizer-free TTS model that joins a MiniCPM-4 language-model backbone with a flow-matching diffusion decoder. We build one shared, language-tagged corpus of about 26 hours and adapt VoxCPM2 with a single Low-Rank Adaptation (LoRA) adapter, trained on both languages at once and added to both the language model and the decoder. The adapter is zero-initialized, so training starts exactly at the original (zero-shot) model. In native-speaker listening tests, the Khmer Mean Opinion Score (MOS) rises from 3.85 to 4.23 with the best adapter (rank 64), a highly significant gain (paired Wilcoxon test, p<0.001), while training only 0.19 to 3.03 percent of the parameters. The automatic loss and the human ratings, however, disagree on the best rank: validation loss is lowest at rank 128, yet MOS peaks at rank 64. The same adapter brings no gain for Korean, a language the base model already handles well, and at a high rank it even degrades quality. Adaptation therefore helps mainly where the base model is genuinely weak.
Chinese Translation
大型预训练文本到语音（TTS）模型在资源丰富的语言中听起来几乎像人类，但在训练数据稀缺的语言中表现则差得多。我们使用VoxCPM2研究高棉语和韩语之间的质量差距，该模型为一个24亿参数的无分词TTS模型，结合了MiniCPM-4语言模型主干和流匹配扩散解码器。我们构建了一个共享的、带语言标签的语料库，约26小时，并使用一个低秩适配器（LoRA）对VoxCPM2进行适配，同时在两种语言上进行训练，并将其添加到语言模型和解码器中。适配器为零初始化，因此训练从原始（零样本）模型开始。在母语者的听力测试中，使用最佳适配器（秩64）时，高棉语的平均意见得分（MOS）从3.85上升至4.23，增幅显著（配对Wilcoxon检验，p<0.001），而训练仅涉及0.19%到3.03%的参数。然而，自动损失和人工评分在最佳秩上存在分歧：验证损失在秩128时最低，但MOS在秩64时达到峰值。同样的适配器对韩语没有提升，基模型已经能够很好地处理该语言，而在高秩时甚至导致质量下降。因此，适配主要在基模型确实较弱的地方提供帮助。
View on arXiv Download PDF AI Translation
cs.CL / 33 / 2606.26650

CAT-Q: Cost-efficient and Accurate Ternary Quantization for LLMs

CAT-Q：高效且准确的三元量化用于大型语言模型
Wang, Shigeng, Li, Chao, Kang, Yangyuxuan, Fan, Jiawei, Yao, Anbang
Abstract
In this paper, we present CAT-Q, Cost-efficient and Accurate Ternary Quantization, for compressing and accelerating LLMs. Unlike existing state-of-the-art ternary quantization methods that rely on data-intensive and costly quantization-aware training to mitigate severe performance degradation, CAT-Q is a simple yet effective post-training quantization scheme that is readily applicable to LLMs with diverse architectures and model sizes. It has two key components, learnable modulation (LM) and softened ternarization (ST), which are coupled from an optimization perspective. LM leverages a composition of learnable factors to modulate the distribution of pre-trained high-precision weights and the ternary threshold, making them less sensitive to ternarization. ST further introduces a differentiable transition function to guide the ternarization process toward stable convergence. We show that, for pre-trained LLMs with 1.7B to 8B parameters, CAT-Q can efficiently quantize them into ternary models using only 512 calibration samples, while achieving superior performance than the seminal BitNet 1.58-bit v1 and v2 families (with 1.3B to 7B parameters) trained with 100B tokens, yielding about a 100,000X reduction in training tokens. Moreover, we show for the first time that CAT-Q can quantize much larger pre-trained LLMs having 14B to 235B parameters into leading ternary models within just 8 to 60 hours on 8 A100-80GB GPUs. Code is available at https://github.com/IntelChina-AI/BitTern.
Chinese Translation
在本文中，我们提出了CAT-Q，即高效且准确的三元量化，用于压缩和加速大型语言模型（LLMs）。与现有的依赖于数据密集且成本高昂的量化感知训练以减轻严重性能下降的最先进三元量化方法不同，CAT-Q是一种简单而有效的后训练量化方案，适用于具有多种架构和模型规模的LLMs。它包含两个关键组件：可学习调制（Learnable Modulation, LM）和软化三元化（Softened Ternarization, ST），这两个组件从优化的角度是相互耦合的。LM利用可学习因子的组合来调节预训练高精度权重的分布和三元阈值，使其对三元化的敏感性降低。ST进一步引入了一个可微分的过渡函数，以指导三元化过程朝向稳定收敛。我们展示了，对于参数在1.7B到8B之间的预训练LLMs，CAT-Q能够仅使用512个校准样本高效地将其量化为三元模型，同时在性能上优于经典的BitNet 1.58-bit v1和v2系列（参数在1.3B到7B之间，使用100B个标记训练），实现了约100,000倍的训练标记减少。此外，我们首次展示了CAT-Q能够在8到60小时内在8个A100-80GB GPU上将参数在14B到235B之间的更大预训练LLMs量化为领先的三元模型。代码可在https://github.com/IntelChina-AI/BitTern获取。
View on arXiv Download PDF AI Translation
cs.CL / 34 / 2606.26654

SocialPersona: Benchmarking Personalized Profiling and Response with Multimodal Social-Media Context

SocialPersona：基于多模态社交媒体背景的个性化画像与响应基准评估
Zhang, Qinkai, Zhao, Yanyan, Lu, Xin, Hu, Yulin, Han, Pengtao, Qin, Bing
Abstract
Personalized language-model assistants are often evaluated through a memory lens: can a model recall preferences users have explicitly stated in dialogue? More comprehensive personalization demands a harder capability -- inferring what users care about from the multimodal traces they naturally leave behind. We introduce SocialPersona, a benchmark for evaluating whether multimodal large language models (MLLMs) can recover revealed preferences from longitudinal social-media timelines and use them in dialogue. Built from longitudinal timelines of 171 everyday, non-promotional social-media users, SocialPersona contains text, images, timestamps, and 2,597 human-verified preference tags across seven interest domains, separating stable interests from recent interests. It supports two tasks: constructing structured user profiles from multimodal context and generating responses aligned with inferred profiles. Experiments with proprietary and open-weight MLLMs show that models can identify broad interest domains, yet their performance drops on fine-grained and recent interests and degrades further when inferred profiles must be used to personalize dialogue. Together with evidence that text and images provide complementary preference signals, these results indicate that robust cross-modal, long-horizon user modeling remains a key challenge, and that SocialPersona can help measure and advance progress toward assistants that infer and act on revealed preferences.
Chinese Translation
个性化语言模型助手通常通过记忆的视角进行评估：模型能否回忆起用户在对话中明确表达的偏好？更全面的个性化要求更高的能力——从用户自然留下的多模态痕迹中推断出他们关心的内容。我们引入了SocialPersona，这是一个评估多模态大型语言模型（MLLMs）是否能够从长期社交媒体时间线中恢复揭示的偏好并在对话中使用这些偏好的基准。SocialPersona基于171个日常非促销社交媒体用户的长期时间线构建，包含文本、图像、时间戳和2,597个经过人工验证的偏好标签，涵盖七个兴趣领域，区分稳定兴趣与近期兴趣。它支持两个任务：从多模态背景构建结构化用户画像，以及生成与推断的画像相一致的响应。对专有和开放权重的MLLMs进行的实验表明，模型能够识别广泛的兴趣领域，但在细粒度和近期兴趣上的表现下降，当推断的画像必须用于个性化对话时，性能进一步下降。结合文本和图像提供互补偏好信号的证据，这些结果表明，稳健的跨模态、长期用户建模仍然是一个关键挑战，而SocialPersona可以帮助衡量和推动朝着能够推断和基于揭示偏好采取行动的助手的进展。
View on arXiv Download PDF AI Translation
cs.CL / 35 / 2606.26698

Beyond Logical Forms: LLM-Extracted Patterns for Fallacy Classification

超越逻辑形式：基于大型语言模型提取的谬误分类模式
Papadopulos, Eleni, Alam, Firoj, Martino, Giovanni Da San
Abstract
In today's fast-paced information era, logical fallacies, defined as defective patterns of reasoning, inevitably contribute to the growth of information disorder. However, often fallacies appear in nuanced forms that complicate automated classification. In this study, we investigate whether merging abstract logical structures with context-level linguistic cues proves beneficial for fallacy classification, developing a framework that inductively extracts such patterns from fallacious examples and their explanations using Large Language Models (LLMs). We evaluate the impact of these patterns across different LLMs and experimental zero- and one-shot configurations, showing statistically significant improvements over zero-shot baselines and outperforming competing approaches. Cross-dataset experiments validate generalization, establishing data-driven pattern extraction as an effective method for generating logical representations.
Chinese Translation
在当今快速发展的信息时代，逻辑谬误被定义为缺陷的推理模式，必然助长信息混乱的增长。然而，谬误常常以细微的形式出现， complicate 自动分类。在本研究中，我们探讨将抽象逻辑结构与上下文级语言线索相结合是否有助于谬误分类，开发了一个框架，该框架使用大型语言模型（LLMs）从谬误示例及其解释中归纳提取这些模式。我们评估了这些模式在不同LLM和实验零样本及一样本配置中的影响，显示出在零样本基线上的统计显著改善，并优于竞争方法。跨数据集实验验证了泛化能力，确立了数据驱动的模式提取作为生成逻辑表示的有效方法。
View on arXiv Download PDF AI Translation
cs.CL / 36 / 2606.26753

ConvMemory v3: A Validity Context Layer for Conversational Memory via Target-Conditioned Relation Verification

ConvMemory v3：通过目标条件关系验证的对话记忆有效性上下文层
Pan, Taiheng
Abstract
Conversational memory retrieval optimizes relevance, yet a retrieved memory can be relevant and simultaneously outdated: a later turn updates, corrects, or supersedes it. ConvMemory v3 adds a validity context layer that detects and surfaces this update evidence through target-conditioned relation verification, sitting after the v1/v2 retrieval path. The core mechanism is a dual-evidence gate that conditions a relation judgment on the specific target proposition, scoring a (target, source) pair through the product of a MiniLM slot head and a DeBERTa-v3 slot head and gating it by conservative event/operation evidence. On a synthetic multi-hop validity benchmark the gate reaches 90.12% +/- 1.73 accuracy; through a real-data feedback loop that mines failure patterns but trains on synthetic pairs only, the verifier transfers to Memora role binding with zero target-side labels, reaching 98.8% +/- 0.9 group-all-correct. The deployed layer preserves retrieval by default: a context mode attaches structured validity metadata while keeping the candidate set and rank order fixed, and a query-conditioned demote mode is an explicit opt-in for dense current-state workloads, where it raises current-active H@1 from a never-demote baseline of 45.1% to 95.7% +/- 1.2 while protecting non-superseded memories at 99.4% recall. Six machine-verifiable safety contracts pin the layer's behavior. Multi-hop graph propagation is validated as a mechanism; fully automatic construction of strict prerequisite edges is characterized as a boundary, since strict necessity requires counterfactual world knowledge. This report extends ConvMemory v1 (arXiv:2605.28062) and v2 (arXiv:2606.10842).
Chinese Translation
对话记忆检索优化相关性，但检索到的记忆可能相关且同时过时：后续的对话轮次会更新、纠正或取代它。ConvMemory v3 增加了一个有效性上下文层，通过目标条件关系验证检测并呈现这种更新证据，该层位于 v1/v2 检索路径之后。核心机制是一个双证据门，它根据特定目标命题对关系判断进行条件化，通过 MiniLM 插槽头和 DeBERTa-v3 插槽头的乘积对 (目标，源) 对进行评分，并通过保守的事件/操作证据进行门控。在一个合成的多跳有效性基准上，该门的准确率达到 90.12% +/- 1.73；通过一个真实数据反馈循环，挖掘失败模式但仅在合成对上进行训练，验证器转移到 Memora 角色绑定，零目标侧标签下达到 98.8% +/- 0.9 的全正确组。部署的层默认保留检索：上下文模式附加结构化有效性元数据，同时保持候选集和排名顺序不变，而查询条件降级模式则是对密集当前状态工作负载的明确选择，其中它将当前活跃 H@1 从一个从未降级的基线 45.1% 提高到 95.7% +/- 1.2，同时保护未被取代的记忆，召回率为 99.4%。六个机器可验证的安全合同约束该层的行为。多跳图传播被验证为一种机制；严格前提边的完全自动构建被描述为一个边界，因为严格必要性要求反事实世界知识。本报告扩展了 ConvMemory v1 (arXiv:2605.28062) 和 v2 (arXiv:2606.10842)。
View on arXiv Download PDF AI Translation
cs.CL / 37 / 2606.26775

Evaluation Pitfalls and Challenges in Multimedia Event Extraction

多媒体事件提取中的评估陷阱与挑战
Seeberger, Philipp, Freisinger, Steffen, Bocklet, Tobias, Riedhammer, Korbinian
Abstract
Multimedia event extraction aims to jointly identify events and their arguments across multiple modalities, such as text and images, to support more comprehensive event understanding. While recent work reports steady and substantial progress, the reliability and comparability of these results critically depend on consistent and rigorous evaluation. In this work, we present the first systematic analysis of evaluation pitfalls in multimedia event extraction and identify three major sources of issues: inconsistent data processing, inconsistent task assumptions, and overly relaxed evaluation settings. We demonstrate, through a series of controlled experiments under a strict evaluation framework, that minor evaluation choices can cause large performance variations and lead to overestimation of a model's ability to ground real-world events across modalities. Our findings highlight the need for comparable evaluation standards and encourage a shift toward more rigorous evaluation in multimedia event extraction.
Chinese Translation
多媒体事件提取旨在跨越文本和图像等多种模态共同识别事件及其论元，以支持更全面的事件理解。尽管近期的研究报告了稳步且显著的进展，但这些结果的可靠性和可比性在很大程度上依赖于一致和严格的评估。在本研究中，我们首次系统性地分析了多媒体事件提取中的评估陷阱，并识别出三个主要问题来源：不一致的数据处理、不一致的任务假设以及过于宽松的评估设置。通过在严格评估框架下进行的一系列控制实验，我们证明了微小的评估选择可能导致性能的巨大变化，并导致对模型在跨模态真实世界事件中的能力的高估。我们的研究结果强调了可比评估标准的必要性，并鼓励在多媒体事件提取中向更严格的评估转变。
View on arXiv Download PDF AI Translation
cs.CL / 38 / 2606.26790

OPID: On-Policy Skill Distillation for Agentic Reinforcement Learning

OPID：用于自主强化学习的在线技能蒸馏
Yang, Shuo, Wu, Jinyang, Lu, Zhengxi, Shen, Yuhao, Zhang, Fan, Feng, Lang, Zhang, Shuai, Luo, Haoran, Lian, Zheng, Wen, Zhengqi, Tao, Jianhua
Abstract
Outcome-based reinforcement learning provides a stable optimization backbone for language agents, but its sparse trajectory-level rewards provide little guidance on which intermediate decisions should be reinforced or suppressed. On-policy self-distillation offers dense token-level supervision, yet existing skill-conditioned variants often rely on external skill memories or retrieved privileged context, which are costly to maintain and can be mismatched with the state distribution induced by the current policy in multi-turn interaction. We propose \textbf{OPID} (\textbf{O}n-\textbf{P}olicy Sk\textbf{i}ll \textbf{D}istillation), a framework that extracts skill supervision directly from completed on-policy trajectories. OPID represents trajectory hindsight as hierarchical skills: episode-level skills capture global workflows or failure-avoidance rules, while step-level skills capture local decision knowledge at critical timesteps. A critical-first routing mechanism uses step-level skills when critical decisions are identified and falls back to episode-level skills as default guidance otherwise. The selected skill is injected into the interaction history, allowing the old policy to re-score the same sampled response under both original and skill-augmented contexts. The resulting log-probability shift yields a token-level self-distillation advantage, which is combined with the outcome advantage for policy optimization. OPID thus preserves RL as the primary training objective while introducing dense, distribution-matched hindsight supervision. Experiments on ALFWorld, WebShop and Search-based QA demonstrate that OPID generally improves agent performance, sample efficiency, and robustness over outcome-only RL and existing skill-distillation baselines. Our code is available at https://github.com/jinyangwu/OPID/tree/main.
Chinese Translation
基于结果的强化学习为语言代理提供了稳定的优化基础，但其稀疏的轨迹级奖励对哪些中间决策应被强化或抑制提供的指导有限。在线自蒸馏提供了密集的令牌级监督，但现有的技能条件变体通常依赖于外部技能记忆或检索的特权上下文，这些都需要高昂的维护成本，并且在多轮交互中可能与当前策略引发的状态分布不匹配。我们提出了 extbf{OPID}（ extbf{O}n- extbf{P}olicy Sk extbf{i}ll extbf{D}istillation），一个直接从完成的在线策略轨迹中提取技能监督的框架。OPID将轨迹的事后分析表示为层次技能：情节级技能捕捉全局工作流程或避免失败的规则，而步骤级技能则在关键时刻捕捉局部决策知识。一个优先考虑关键决策的路由机制在识别到关键决策时使用步骤级技能，否则回退到情节级技能作为默认指导。所选择的技能被注入到交互历史中，允许旧策略在原始和技能增强的上下文下重新评分相同的采样响应。由此产生的对数概率变化带来了令牌级自蒸馏优势，该优势与结果优势结合用于策略优化。因此，OPID保留了强化学习作为主要训练目标，同时引入了密集的、与分布匹配的事后监督。在ALFWorld、WebShop和基于搜索的问答的实验表明，OPID通常改善了代理的性能、样本效率和鲁棒性，优于仅基于结果的强化学习和现有的技能蒸馏基线。我们的代码可在https://github.com/jinyangwu/OPID/tree/main获取。
View on arXiv Download PDF AI Translation
cs.CL / 39 / 2606.26803

From Vajrayana Tara to Bengali Baul: A Computational Study of Lexical Transmission Across Buddhist, Shakta, and Vaishnava Traditions in Bengal

从金刚乘的度母到孟加拉的巴乌尔：对孟加拉佛教、女神教和毗湿奴教传统中词汇传播的计算研究
Bose, Joy
Abstract
We present a computational corpus study of vocabulary relationships across eight tradition layers of Bengali and Sanskrit devotional literature spanning the 8th to 19th centuries, encompassing Buddhist Vajrayana, Shakta Tantra, Vaishnava, and Baul traditions. Using a corpus of 75 texts and TF-IDF character n-gram vectorization with cosine similarity analysis, we address the historically argued but previously unquantified claim that Buddhist Vajrayana vocabulary survived the collapse of the Pala monasteries and was absorbed into the Shakta Tantra tradition of Bengal. The central finding is a specificity result: the Gitagovinda (Vaishnava Sanskrit, 12th century) has zero cosine similarity to Shakta Kali texts, while Bridge Tara texts (Buddhist-Shakta transitional, same century, same language) have cosine similarity 0.54 to Shakta Kali. This 8.5-fold contrast between two Sanskrit traditions from the same century demonstrates that the Buddhist-Shakta vocabulary overlap is not a generic property of Sanskrit devotional literature but is specific to the Buddhist-Shakta transmission chain. Three Brihannilatantra Tara texts show Shakta-to-Buddhist vocabulary ratios of 2.0 to 4.0, constituting measurable evidence of lexical transition within that chain. Ramprasad Sen's 18th-century Bengali Kali songs preserve Buddhist vocabulary residue including 56 occurrences of Tara alongside 103 occurrences of Kali. The Vaishnava Bengali tradition contributes a parallel chain to modern Baul vocabulary (similarity 0.29), slightly weaker than the Buddhist Sahajiya chain via Charyapada (0.31). These results provide the first quantitative multi-tradition corroboration of historically argued Buddhist-Shakta syncretism in Bengal.
Chinese Translation
我们呈现了一项关于孟加拉和梵文宗教文学中八个传统层次词汇关系的计算语料库研究，涵盖了8世纪至19世纪的佛教金刚乘、女神教密宗、毗湿奴教和巴乌尔传统。通过使用75部文本的语料库，以及基于TF-IDF字符n-gram向量化和余弦相似度分析，我们探讨了一个历史上有争议但此前未量化的论点，即佛教金刚乘的词汇在帕拉寺院崩溃后幸存下来，并被吸收进孟加拉的女神教密宗中。主要发现是一个特异性结果：吉塔歌文达（毗湿奴教梵文，12世纪）与女神教卡莉文本的余弦相似度为零，而桥度母文本（佛教-女神教过渡性，同世纪，同语言）与女神教卡莉的余弦相似度为0.54。这两个来自同一世纪的梵文传统之间的8.5倍对比表明，佛教-女神教的词汇重叠并不是梵文宗教文学的普遍特性，而是特定于佛教-女神教传播链。三部《大布里哈尼拉坦特拉》度母文本显示女神教与佛教词汇比率为2.0至4.0，构成了该传播链内词汇转变的可测量证据。拉姆普拉萨德·森的18世纪孟加拉卡莉歌曲保留了佛教词汇残余，包括56次出现的度母和103次出现的卡莉。毗湿奴教孟加拉传统为现代巴乌尔词汇贡献了一条平行链（相似度0.29），略弱于通过《查雅帕达》的佛教萨哈吉亚链（0.31）。这些结果为历史上争论的孟加拉佛教-女神教融合提供了首次定量的多传统证据。
View on arXiv Download PDF AI Translation
cs.CL / 40 / 2606.26819

FBK's Long-form SpeechLLMs for IWSLT 2026 Instruction Following

FBK的长形式SpeechLLMs在IWSLT 2026指令跟随中的应用
Xie, Zhihang, Gaido, Marco, Papi, Sara, Negri, Matteo, Bentivogli, Luisa
Abstract
This paper describes our submission to the IWSLT 2026 Instruction Following shared task. SpeechLLMs are developed for both short-form and long-form speech instruction following under constrained settings. For the short track, strong performance is achieved on MCIF, with a SIFS score of 2.0708. For the long track, three speech segmentation methods are explored, and the HIFS score is introduced to account for unstable long-form generation. Experimental results show that fixed 30-second segmentation provides the most robust long-form performance, achieving the highest HIFS score of 2.0663. Further analysis shows that hallucination mainly manifests as repetitive insertions in generated outputs, substantially affecting ASR and SSUM, while short-form capabilities are largely retained after long-form extension.
Chinese Translation
本文描述了我们对IWSLT 2026指令跟随共享任务的提交。SpeechLLMs在受限环境下为短形式和长形式的语音指令跟随而开发。在短轨道上，MCIF上取得了强劲的表现，SIFS得分为2.0708。在长轨道上，探索了三种语音分割方法，并引入了HIFS得分以考虑不稳定的长形式生成。实验结果表明，固定的30秒分割提供了最稳健的长形式表现，达到了最高的HIFS得分2.0663。进一步分析显示，幻觉主要表现为生成输出中的重复插入，显著影响了ASR和SSUM，而在长形式扩展后，短形式能力基本得以保留。
View on arXiv Download PDF AI Translation
cs.CL / 41 / 2606.26861

Cascaded Multi-Granularity Pruning for On-Device LLM Inference in Industrial IoT

工业物联网中设备端大语言模型推理的级联多粒度剪枝
Wang, Jinghan, Chen, Yanjun, Zhang, Wei, Huang, Xiaotong, Liu, Tianchen, Peng, Gaoliang
Abstract
Deploying large language models (LLMs) on Industrial Internet of Things (IIoT) edge devices demands extreme compression, yet existing structured pruning methods collapse at high compression ratios due to one-shot importance estimation, and their cross-architecture behavior remains unpredictable. This article presents a cascaded multi-granularity pruning framework that removes layers, attention heads, and feed-forward channels in coarse-to-fine order, with lightweight low-rank recovery between stages to re-estimate component importance. An information-theoretic analysis motivates this ordering, and the Structural Independence Assumption (SIA) is formalized as a checkable condition predicting whether per-component pruning criteria are reliable for a given architecture: Multi-Head Attention (MHA)+GELU designs satisfy the SIA, whereas Grouped Query Attention (GQA)+SwiGLU designs violate it. On bearing fault diagnosis spanning 88M to 6.25B-parameter models, the framework extends achievable compression to 13.8 times on MHA+GELU architectures with 83.82% accuracy (+3.70 percentage points (pp) over the strongest baseline), while exposing a ~74pp accuracy collapse on GQA+SwiGLU architectures that violate the SIA. Deployed on an industrial slewing bearing fault diagnosis platform with NVIDIA DGX Spark, compressed models reduce inference latency by up to 67.2% and peak memory by 62.5%, demonstrating viability for IIoT edge inference.
Chinese Translation
在工业物联网（IIoT）边缘设备上部署大型语言模型（LLMs）需要极端压缩，然而现有的结构化剪枝方法在高压缩比下由于一次性重要性估计而失效，并且其跨架构行为仍然不可预测。本文提出了一种级联多粒度剪枝框架，该框架以粗到细的顺序移除层、注意力头和前馈通道，并在阶段之间进行轻量级低秩恢复以重新估计组件重要性。信息论分析为这种顺序提供了动机，结构独立假设（Structural Independence Assumption, SIA）被形式化为可检查的条件，以预测每个组件的剪枝标准在给定架构下是否可靠：多头注意力（Multi-Head Attention, MHA）+GELU设计满足SIA，而分组查询注意力（Grouped Query Attention, GQA）+SwiGLU设计则违反了这一假设。在涵盖从8800万到62.5亿参数模型的轴承故障诊断中，该框架将可实现的压缩扩展至MHA+GELU架构的13.8倍，且准确率达到83.82%（比最强基线高出3.70个百分点），同时在违反SIA的GQA+SwiGLU架构上暴露出约74个百分点的准确率崩溃。在搭载NVIDIA DGX Spark的工业回转轴承故障诊断平台上部署的压缩模型将推理延迟减少了最多67.2%，峰值内存减少了62.5%，证明了其在IIoT边缘推理中的可行性。
View on arXiv Download PDF AI Translation
cs.CL / 42 / 2606.26875

Information-Aware KV Cache Compression for Long Reasoning

面向信息的KV缓存压缩用于长推理
Kai, Jushi, Xiao, Zhuiri, Birch, Alexandra, Lin, Zhouhan
Abstract
Reasoning capability has advanced rapidly in large language models (LLMs), leading to an increasing size of key-value (KV) cache in both prefilling and decoding stages. Existing KV cache compression methods mainly rely on attention weights to estimate token importance. While attention effectively captures contextual relevance, it overlooks complementary information-theoretic signals related to predictive uncertainty and token informativeness. In this paper, we revisit token importance from a forward-looking perspective and introduce \textit{Forward Influence}, a metric that measures how compressed tokens affect future contexts. Our analysis reveals that tokens selected by attention scores mainly influence nearby contexts, whereas tokens associated with high predictive uncertainty exhibit substantially stronger influence on distant future contexts. Based on the observation, we propose \textbf{InfoKV}, an entropy-aware KV cache compression framework that incorporates information-theoretic signals. It combines token-level predictive uncertainty with layer-wise representation evolution and integrates the resulting entropy scores with attention scores during reasoning. Experiments on long-context reasoning benchmarks with Llama-3.1, Llama-3.2, and DeepSeek-R1 demonstrate that InfoKV consistently outperforms existing attention-based KV compression methods in both long prefilling and decoding scenarios.
Chinese Translation
推理能力在大型语言模型（LLMs）中迅速发展，导致在预填充和解码阶段的键值（KV）缓存规模不断增加。现有的KV缓存压缩方法主要依赖注意力权重来估计令牌的重要性。虽然注意力有效地捕捉了上下文相关性，但它忽视了与预测不确定性和令牌信息量相关的互补信息论信号。本文从前瞻性的角度重新审视令牌的重要性，并引入了 extit{Forward Influence}，一种衡量压缩令牌如何影响未来上下文的指标。我们的分析表明，注意力得分选择的令牌主要影响附近的上下文，而与高预测不确定性相关的令牌对远期上下文的影响则显著更强。基于这一观察，我们提出了 extbf{InfoKV}，一种基于熵的KV缓存压缩框架，结合了信息论信号。它将令牌级的预测不确定性与层级表示演变结合，并在推理过程中将得到的熵得分与注意力得分整合。在使用Llama-3.1、Llama-3.2和DeepSeek-R1进行的长上下文推理基准测试中，实验表明InfoKV在长预填充和解码场景中始终优于现有的基于注意力的KV压缩方法。
View on arXiv Download PDF AI Translation
cs.CL / 43 / 2606.26880

Heterogeneous Neural Predictivity from Language Models During Naturalistic Comprehension

语言模型在自然理解中的异质神经预测能力
Jia, Xiao
Abstract
Language-model representations provide structured, high-dimensional annotations of naturalistic language stimuli and can serve as informative neural predictors during comprehension. We analyzed locked derived data from Brain Treebank, MEG-MASC, and Podcast ECoG with eight frozen language models, blocked encoding models, and matched temporal, nuisance, and representation-capacity controls. Positive held-out prediction and gains over low-level baselines were widespread in source-level summaries. Across Brain Treebank and Podcast ECoG, 67 of 432 evaluable rows met a controlled predictive-only criterion, and model-side feature ablations changed prediction scores in most evaluable source rows. Brain-derived, timing-linked, acoustic, and implanted-signal controls confirmed component-level sensitivity of the analysis pipeline. These findings show that language-model-derived quantities can annotate neural activity during natural speech and text comprehension. Participant-level matched-control advantages were localized rather than uniform, response-profile and feature-specificity contrasts bounded representational or computational interpretations, and complete co-indexed integrated interpretation will require future jointly indexed coverage. Together, the analyses identify language-model features as useful neural predictors and separate predictive usefulness from claims about shared neural organization or language-processing computations.
Chinese Translation
语言模型的表征提供了自然语言刺激的结构化、高维注释，并可以作为理解过程中的信息性神经预测器。我们分析了来自Brain Treebank、MEG-MASC和Podcast ECoG的锁定派生数据，使用了八个冻结的语言模型、阻塞编码模型以及匹配的时间、干扰和表征能力控制。源级摘要中普遍存在积极的保留预测和相对于低水平基线的增益。在Brain Treebank和Podcast ECoG中，432个可评估行中有67个满足受控的仅预测标准，模型侧特征消融在大多数可评估源行中改变了预测分数。脑源、时间关联、声学和植入信号控制确认了分析流程的组件级敏感性。这些发现表明，语言模型派生的量可以注释自然语音和文本理解过程中的神经活动。参与者级的匹配控制优势是局部的而非均匀的，反应特征和特征特异性的对比限制了表征或计算解释，而完整的共同索引集成解释将需要未来的联合索引覆盖。总体而言，这些分析将语言模型特征识别为有用的神经预测器，并将预测的有用性与关于共享神经组织或语言处理计算的主张区分开来。
View on arXiv Download PDF AI Translation
cs.CL / 44 / 2606.26901

SamaVaani: Auditing and Debiasing Multilingual Clinical ASR for Indian Languages

SamaVaani：对印度语言的多语言临床自动语音识别进行审计和去偏见处理
Kumar, Subham, Shivaprakash, Prakrithi, Manoharan, Abhishek, Kurariya, Astut, Mukherjee, Diptadhi, Chand, Prabhat, Murthy, Pratima, Rudra, Koustav, Shukla, Lekhansh, Mukherjee, Animesh
Abstract
Automatic Speech Recognition (ASR) is increasingly used to document clinical encounters, yet its reliability in multilingual and demographically diverse Indian healthcare context remains largely unknown. In this study, we first conduct the systematic audit of ASR performance on real-world psychiatric interview data spanning Kannada, Hindi and Indian English, comparing eight state-of-the-art models including IndicWhisper, WhisperLargeV3, Sarvam, GoogleS2T, Gemma3n, OmniLingual, Vaani, and Gemini. Our results reveal substantial variability across models and languages, with some systems performing competitively in Indian English but failing in regional speech. We further fine-tune two of the best performing opensource models, i.e., Gemma3n and OmniLingual, using various methods. With this, we uncover systematic performance gaps tied to speaker role and gender, raising concerns about equitable deployment in clinical settings, which are further mitigated by fairness-aware fine-tuning. To this end, we propose SamaVaani, a unified debiasing technique that simultaneously improves ASR performance and improves fairness across demographic groups.
Chinese Translation
自动语音识别（ASR）在记录临床接触中越来越多地被使用，但其在多语言和人口多样化的印度医疗环境中的可靠性仍然 largely unknown。在本研究中，我们首先对涵盖卡纳达语、印地语和印度英语的真实世界精神病访谈数据进行系统审计，比较了包括IndicWhisper、WhisperLargeV3、Sarvam、GoogleS2T、Gemma3n、OmniLingual、Vaani和Gemini在内的八种最先进的模型。我们的结果揭示了模型和语言之间存在显著的变异性，一些系统在印度英语中表现良好，但在地方语言中表现不佳。我们进一步对两种表现最好的开源模型，即Gemma3n和OmniLingual，进行了各种方法的微调。通过这些工作，我们发现与说话者角色和性别相关的系统性性能差距，这引发了对临床环境中公平部署的担忧，而通过公平感知的微调可以进一步缓解这些问题。为此，我们提出了SamaVaani，这是一种统一的去偏见技术，旨在同时提高ASR性能并改善不同人口群体之间的公平性。
View on arXiv Download PDF AI Translation
cs.CL / 45 / 2606.26923

GAVEL: Grounded Caption Error Verification and Localization

GAVEL：基于视觉的描述错误验证与定位
Gao, Zixian, Hashimoto, Atsushi, Saito, Kuniaki
Abstract
Vision-language models (VLMs) often produce hallucinated or inconsistent outputs, where text and images are not properly aligned. Addressing this issue requires not only detecting misalignment but also explaining the discrepancy and localizing its visual evidence. We introduce GAVEL (Grounded Caption Error Verification and Localization), a task that jointly addresses verification, explanation, and localization for image-text pairs. To support systematic evaluation, we also present a corresponding dataset and benchmark. We further train a supervised baseline on the human-annotated training split to assess whether GAVEL provides learnable supervision for these abilities. Experiments show that even strong closed-source models struggle on GAVEL, while the supervised baseline yields consistent improvements across grounding and explanation metrics.
Chinese Translation
视觉-语言模型（VLMs）经常产生幻觉或不一致的输出，文本与图像之间未能正确对齐。解决这一问题不仅需要检测不对齐，还需要解释差异并定位其视觉证据。我们提出了GAVEL（Grounded Caption Error Verification and Localization），这是一个共同解决图像-文本对的验证、解释和定位任务。为了支持系统评估，我们还提供了相应的数据集和基准测试。我们进一步在人工标注的训练集上训练了一个监督基线，以评估GAVEL是否为这些能力提供可学习的监督。实验表明，即使是强大的闭源模型在GAVEL上也表现不佳，而监督基线在对齐和解释指标上均取得了一致的改善。
View on arXiv Download PDF AI Translation
cs.CL / 46 / 2606.26963

Term-Centric Hierarchy Induction from Heterogeneous Corpora

基于术语中心的异构语料库层次结构诱导
Senger, Elena, Campbell, Yuri, Bergmann, Jan-Peter, van der Goot, Rob, Plank, Barbara
Abstract
Organizing knowledge from diverse text sources into interpretable hierarchies is crucial for tasks such as policy analysis, innovation monitoring, and exploratory domain mapping. Existing taxonomy induction methods typically rely on document-level representations that capture entire documents rather than the specific domain concepts relevant for knowledge organization, limiting their ability to generalize across heterogeneous sources. We propose a term-centric framework for inducing hierarchical taxonomies from heterogeneous corpora that scales to massive document collections. Our approach maps documents from diverse sources into a shared representation space using automatic term extraction, enabling robust cross-source alignment. Based on these representations, we construct interpretable hierarchies that integrate domain priors with datadriven clustering. Experiments on a novel English and German multi-source benchmark of over one million documents demonstrate that our method improves cross-source coherence and hierarchy quality over text- and summarybased baselines. A case study on German regional innovation analysis further demonstrates its practical utility for technology landscape mapping.
Chinese Translation
将来自多样文本来源的知识组织成可解释的层次结构对于政策分析、创新监测和探索性领域映射等任务至关重要。现有的分类法诱导方法通常依赖于文档级表示，这些表示捕捉整个文档，而不是与知识组织相关的特定领域概念，从而限制了它们在异构来源之间的泛化能力。我们提出了一种基于术语中心的框架，用于从异构语料库中诱导层次分类法，该框架能够扩展到大规模文档集合。我们的方法通过自动术语提取将来自不同来源的文档映射到共享表示空间，从而实现稳健的跨源对齐。基于这些表示，我们构建了可解释的层次结构，将领域先验与数据驱动的聚类相结合。在一个包含超过一百万个文档的新型英语和德语多源基准上的实验表明，我们的方法在跨源一致性和层次质量方面优于基于文本和摘要的基线。对德国区域创新分析的案例研究进一步展示了其在技术景观映射中的实际应用价值。
View on arXiv Download PDF AI Translation
cs.CL / 47 / 2606.26968

RedVox: Safety and Fairness Gaps in Speech Models Across Languages

RedVox：跨语言语音模型中的安全性和公平性差距
Savoldi, Beatrice, Papi, Sara, Aissa, Wafa, Negri, Matteo, Bentivogli, Luisa
Abstract
Speech-capable models are increasingly deployed in real-world applications across languages. Yet their safety and fairness beyond English settings and under naturalistic conditions remain understudied. We survey safety reporting practices across state-of-the-art speech model releases, finding that only 8% document any multilingual analysis. To address this gap, we introduce RedVox, a multilingual safety and fairness benchmark for audio and speech built on real voices, covering unsafe and unfair stereotypical requests across five languages (English, French, Italian, Spanish, and German). Evaluating eight state-of-the-art models, we find that vulnerabilities persist even under non-adversarial conditions, worsen in non-English languages, and are amplified when the request comes from a spoken input. Finally, by surveying the participants who contributed to RedVox, we document the unique personal and privacy challenges of collecting speech data with human participants, pointing to broader sociotechnical challenges in naturalistic speech safety research.
Chinese Translation
具备语音能力的模型正越来越多地在跨语言的实际应用中部署。然而，它们在非英语环境下和自然条件下的安全性和公平性仍然未得到充分研究。我们调查了最新语音模型发布中的安全报告实践，发现仅有8%的文档包含任何多语言分析。为了解决这一差距，我们推出了RedVox，这是一个基于真实声音的多语言音频和语音安全性与公平性基准，涵盖了五种语言（英语、法语、意大利语、西班牙语和德语）中的不安全和不公平的刻板请求。在评估八个最新的模型时，我们发现即使在非对抗条件下，脆弱性依然存在，在非英语语言中情况更为严重，并且当请求来自口语输入时，这种脆弱性会被放大。最后，通过调查参与RedVox的参与者，我们记录了收集人类参与者语音数据所面临的独特个人和隐私挑战，指出了自然语音安全研究中更广泛的社会技术挑战。
View on arXiv Download PDF AI Translation
cs.CL / 48 / 2606.26982

Auditing Framing-Sensitive Behavioral Instability in Large Language Models for Mental Health Interactions

审计大型语言模型在心理健康互动中的框架敏感行为不稳定性
Bedoui, Abla, Greene, Ashley L., Cherkaoui, Mohammed
Abstract
Large language models (LLMs) are increasingly being integrated into mental health support tools and other psychologically sensitive conversational applications. In such settings, behavioral stability and consistency are important for trustworthy human-AI interaction. However, semantically similar concerns can be presented through different contextual framings, potentially eliciting different model responses. Such framing-sensitive variability may challenge user expectations regarding system behavior and complicate the assessment of AI reliability. While prior studies have primarily examined such effects at the behavioral level, less is known about how framing-related variation is reflected in the internal representations of aligned language models. In this work, we investigate these effects using controlled matched prompts spanning multiple contextual framing conditions across several instruction-tuned model families. Across architectures, framing systematically alters interpretive response tendencies. Layer-wise probing analyses show that behavior-associated information remains decodable throughout transformer depth, with architecture-dependent variation in decoding strength. Moreover, held-out framing probes remained consistently above chance across architectures despite strong lexical baselines. Activation steering experiments further suggest that framing-associated representational directions can partially modulate downstream behavioral outcomes. Finally, these findings indicate that robustness to contextual variation may represent an important consideration when evaluating the consistency and trustworthiness of conversational AI systems deployed in mental-health-oriented interactions.
Chinese Translation
大型语言模型（LLMs）正越来越多地被整合到心理健康支持工具和其他心理敏感的对话应用中。在这些环境中，行为的稳定性和一致性对于可信赖的人机互动至关重要。然而，语义上相似的关切可以通过不同的上下文框架呈现，可能引发模型的不同响应。这种框架敏感的变异性可能挑战用户对系统行为的期望，并使评估人工智能的可靠性变得复杂。尽管先前的研究主要在行为层面上考察了这些效应，但关于框架相关变异如何反映在对齐语言模型的内部表征中知之甚少。在本研究中，我们使用控制匹配的提示，跨越多个上下文框架条件，研究这些效应，涵盖几种指令调优的模型家族。在不同架构中，框架系统性地改变了解释响应的倾向。逐层探测分析表明，行为相关信息在变换器深度中始终可解码，且解码强度因架构而异。此外，保留的框架探针在不同架构中始终保持高于随机水平，尽管存在强大的词汇基线。激活引导实验进一步表明，框架相关的表征方向可以部分调节下游行为结果。最后，这些发现表明，对上下文变异的鲁棒性可能在评估部署于心理健康导向互动的对话人工智能系统的一致性和可信赖性时，代表了一个重要的考量因素。
View on arXiv Download PDF AI Translation
cs.CL / 49 / 2606.26986

ReaORE: Reasoning-Guided Progressive Open Relation Extraction Empowered by Large Reasoning Models

ReaORE：由大型推理模型驱动的推理引导渐进式开放关系提取
Lin, Xin, Zhang, Liang, Ma, Guoqi, Tu, Hongyao, Su, Jinsong
Abstract
Open Relation Extraction (OpenRE) requires a model to extract unseen relations between head and tail entities from unstructured text for real-world applications. The core challenge of OpenRE lies in achieving reliable generalization to unseen relation types. Current OpenRE approaches either employ clustering techniques, which cannot generate relation labels and suffer from poor generalization, or rely on direct relation label generation via Large Language Models (LLMs), which lack sufficient discriminative capacity to distinguish easily confused relations. To address these limitations, we propose Reasoning-guided progressive OpenRE (ReaORE), a framework for performing relation extraction through coarse-to-fine relation reasoning. Specifically, ReaORE consists of two key stages: (i) relation filtering, which reasons over multiple aspects to understand relations and instances, yielding an initial relation set, and further supplements and filters relations via embedding-based similarity to ensure the target relation is included; (ii) relation prediction, which aims to predict the target relations from the above set via fine-grained comparative reasoning to better distinguish easily confused relations. Extensive experiments on two widely used OpenRE datasets demonstrate that ReaORE outperforms existing baselines.
Chinese Translation
开放关系提取（OpenRE）要求模型从非结构化文本中提取头实体和尾实体之间的未见关系，以用于实际应用。OpenRE的核心挑战在于实现对未见关系类型的可靠泛化。目前的OpenRE方法要么采用聚类技术，这无法生成关系标签且泛化能力较差，要么依赖大型语言模型（LLMs）直接生成关系标签，但缺乏足够的区分能力来区分容易混淆的关系。为了解决这些局限性，我们提出了推理引导的渐进式开放关系提取（ReaORE），这是一个通过粗到细的关系推理进行关系提取的框架。具体而言，ReaORE包括两个关键阶段：（i）关系过滤，通过对多个方面进行推理以理解关系和实例，生成初始关系集，并通过基于嵌入的相似性进一步补充和过滤关系，以确保目标关系被包含；（ii）关系预测，旨在通过细粒度的比较推理从上述集合中预测目标关系，以更好地区分容易混淆的关系。在两个广泛使用的OpenRE数据集上的大量实验表明，ReaORE的表现优于现有基线。
View on arXiv Download PDF AI Translation
cs.CL / 50 / 2606.26987

Where Do Models Find Happiness? Emotion Vectors in Open-Source LLMs

模型如何找到幸福？开源大型语言模型中的情感向量
van der Ben, Sinie, Baur, Raphaël, Metz, Yannick, El-Assady, Mennatallah
Abstract
Recent work identified emotion vectors in Claude Sonnet 4.5, which are internal representations that encode emotion concepts, causally influence behavior, and exhibit geometry mirroring human psychological structure. We test the generality of these findings in two open-weight models, Apertus-8B-Instruct-2509 and Gemma-4-E4B-it, extracting emotion contrast vectors across all layers, using two model-generated corpora. We recover valence geometry for both models, with peak PC1--valence correlations of $r = 0.76$ and $r = 0.83$, approaching the $r = 0.81$ reported for Claude.Beyond replication, we observe notable differences in how valence representations emerge across model depth. In Gemma-4-E4B-it, valence is strongly encoded in early layers but collapses towards later layers, whereas Apertus-8B-Instruct-2509 exhibits the opposite pattern, with valence representations absent in early layers, but emerging at mid depths. Arousal encoding, in contrast, is sensitive to the extraction corpus: both models show stronger PC2--arousal alignment with Gemma-generated stories ($r$ up to $0.45$) than Apertus-generated ones ($r \leq 0.21$), suggesting arousal-relevant cues are unevenly distributed across generated corpora. We open-source our experiment code and dataset for reproducible investigation of emotion representations across language model architectures.
Chinese Translation
近期研究在Claude Sonnet 4.5中识别出情感向量，这些向量是编码情感概念的内部表征，能够因果性地影响行为，并展现出与人类心理结构相似的几何特征。我们在两个开放权重模型Apertus-8B-Instruct-2509和Gemma-4-E4B-it中测试这些发现的普遍性，通过使用两个模型生成的语料库提取所有层的情感对比向量。我们恢复了两个模型的效价几何，PC1与效价的相关性峰值分别为$r = 0.76$和$r = 0.83$，接近Claude报告的$r = 0.81$。除了重复实验外，我们还观察到效价表征在模型深度上的显著差异。在Gemma-4-E4B-it中，效价在早期层中被强烈编码，但在后期层中崩溃，而Apertus-8B-Instruct-2509则表现出相反的模式，早期层中缺乏效价表征，但在中层深度中出现。相比之下，唤醒编码对提取语料库敏感：两个模型在Gemma生成的故事中表现出更强的PC2与唤醒的对齐（$r$高达$0.45$），而在Apertus生成的故事中则较弱（$r leq 0.21$），这表明与唤醒相关的线索在生成的语料库中分布不均。我们将实验代码和数据集开源，以便对不同语言模型架构中的情感表征进行可重复的研究。
View on arXiv Download PDF AI Translation
cs.CL / 51 / 2606.27025

Improving General Role-Playing Agents via Psychology-Grounded Reasoning and Role-Aware Policy Optimization

通过心理学基础推理和角色意识政策优化提升通用角色扮演代理
Xu, Zhenhua, Chen, Dongsheng, Li, Jian, Lin, Yitong, Wang, Zhebo, Wu, Jiafu, Jin, Yizhang, Wang, Chengjie, Han, Meng, Wang, Yabiao
Abstract
Building general-purpose role-playing agents that faithfully portray any character from a natural-language profile remains challenging. The dominant paradigm -- supervised fine-tuning -- encourages behavioral mimicry without deep, human-like internal thought processes, resulting in poor out-of-distribution generalization. Therefore, we propose \textbf{Psy-CoT}, a psychology-grounded chain-of-thought framework that decomposes pre-response reasoning into three role-specific steps -- \emph{Interaction Perception}, \emph{Psychological Empathy}, and \emph{Logical Construction} -- so that the model \emph{thinks dynamically} from the profile rather than merely mimicking surface patterns. While structured reasoning provides a foundation, it alone is insufficient; reinforcement learning is essential to further align the model with character fidelity. However, we observe that under LLM-based reward models, both generic phrases that hack the reward model and genuinely role-specific phrases receive identical gradient signals -- this hacking accumulates over training, misleading the model into treating both as equally optimal choices. To address this, we propose \textbf{Role-Aware Policy Optimization (RAPO)}, which uses profile--token mutual information to weight gradients asymmetrically -- amplifying role-specific tokens under positive advantage while attenuating them under negative advantage. Experiments on CoSER, CharacterBench, and CharacterEval demonstrate that Psy-CoT outperforms existing role-playing CoT methods, and RAPO consistently surpasses GRPO across multiple model scales.
Chinese Translation
构建能够忠实描绘自然语言档案中任何角色的通用角色扮演代理仍然具有挑战性。当前主流范式——监督微调——鼓励行为模仿，而缺乏深层次的类人内部思维过程，导致在分布外泛化能力较差。因此，我们提出了 extbf{Psy-CoT}，一种基于心理学的链式思维框架，将响应前的推理分解为三个角色特定的步骤—— extit{互动感知}、 extit{心理共情}和 extit{逻辑构建}——使得模型能够 extit{动态思考}而不仅仅是模仿表面模式。虽然结构化推理提供了基础，但单靠这一点是不够的；强化学习对于进一步使模型与角色忠实度对齐至关重要。然而，我们观察到在基于大型语言模型（LLM）的奖励模型下，既有破解奖励模型的通用短语，也有真正角色特定的短语都接收到了相同的梯度信号——这种破解在训练过程中累积，误导模型将两者视为同样的最佳选择。为了解决这个问题，我们提出了 extbf{角色意识政策优化（RAPO）}，它利用档案-标记互信息不对称地加权梯度——在正优势下放大角色特定标记，而在负优势下减弱它们。在CoSER、CharacterBench和CharacterEval上的实验表明，Psy-CoT优于现有的角色扮演链式思维方法，而RAPO在多个模型规模上始终超越GRPO。
View on arXiv Download PDF AI Translation
cs.CL / 52 / 2606.27047

NuclearQAv2: A Structured Benchmark for Evaluating Domain-Science Competence in Large Language Models

NuclearQAv2：评估大型语言模型在领域科学能力的结构化基准
Yuchi, Henry Shaowu, Kucer, Michal, Sims, Benjamin H., Peterson, Selma, Taylor, Emily
Abstract
Large language models (LLMs) have demonstrated strong performance across a wide range of tasks, but ensuring their reliability in highly technical domains remains a significant challenge. In nuclear engineering, problem solving often requires not only factual knowledge but also quantitative reasoning and conceptual understanding. To address the need for systematic evaluation in this domain, we introduce NuclearQAv2, a benchmark for assessing LLMs on nuclear engineering knowledge. The benchmark comprises approximately 1,240 question-answer pairs spanning three categories: boolean, numeric, and verbal. NuclearQAv2 is constructed using a hybrid pipeline that combines expert-authored questions, existing datasets, and LLM-assisted generation from domain-specific technical corpora. By leveraging structured prompting for both automated question generation and response evaluation, the proposed framework enables scalable benchmark construction and evaluation. We evaluate a diverse set of LLMs using NuclearQAv2 and observe substantial performance differences across task types. While the models generally perform well on factual questions, quantitative reasoning and conceptual understanding remain considerably more challenging. These results highlight the importance of multi-faceted evaluation frameworks and establish NuclearQAv2 as a scalable benchmark for assessing LLM capabilities in technical domains.
Chinese Translation
大型语言模型（LLMs）在广泛的任务中表现出色，但在高度技术领域确保其可靠性仍然是一个重大挑战。在核工程中，问题解决不仅需要事实知识，还需要定量推理和概念理解。为满足该领域系统评估的需求，我们推出了NuclearQAv2，这是一个用于评估LLMs在核工程知识方面的基准。该基准包含约1,240个问答对，涵盖布尔、数值和语言三类问题。NuclearQAv2采用混合管道构建，结合了专家撰写的问题、现有数据集以及从领域特定技术语料库中使用LLM辅助生成的问题。通过利用结构化提示进行自动化问题生成和响应评估，所提出的框架实现了可扩展的基准构建和评估。我们使用NuclearQAv2评估了一组多样化的LLMs，并观察到不同任务类型之间存在显著的性能差异。尽管模型在事实性问题上通常表现良好，但定量推理和概念理解仍然相对更具挑战性。这些结果突显了多方面评估框架的重要性，并确立了NuclearQAv2作为评估LLM在技术领域能力的可扩展基准。
View on arXiv Download PDF AI Translation
cs.CL / 53 / 2606.27069

Towards Explainable Adjudicative Variance: Quantifying Judicial Discretion via Gated Multi-Task Learning

朝向可解释的裁判方差：通过门控多任务学习量化司法裁量
Sójka, Stanisław, Steffek, Felix, Grabmair, Matthias
Abstract
Legal outcome prediction must disentangle objective case facts from adjudicative context. Merit-based rulings rely on factual evidence while technical disposals may hinge on judicial discretion. We propose a Judge-Aware Gated Multi-Task Learning architecture that explicitly models this distinction. We introduce a fine-grained outcome taxonomy to supervise the encoder, enforcing a structural regularization that disentangles distinct semantic pathways. This granular legal curriculum enables our Gated Fusion mechanism to dynamically modulate reliance on judge identity. We evaluate our approach on 13,937 UK Employment Tribunal decisions. We benchmark our design against supervised fine-tuning (SFT) of a Gemma-4 26B-A4B backbone, in which judge identity and the taxonomy are injected as prompt tokens or autoregressive output targets. The two contextual signals compose only weakly when forced through a single autoregressive channel. In contrast, coupling a LoRA-adapted Gemma-4 encoder with our gated architecture defines a new state of the art on this benchmark while requiring an order of magnitude fewer trainable parameters than the generative SFT baselines, with gains concentrated on the most ambiguous and rarest outcome classes. Beyond accuracy, the architecture is interpretable; learned judge embeddings and calibration profiles localize the cases where adjudicative context drives the prediction. These results indicate that, for identity-conditioned classification of legal outcomes, the choice of conditioning interface dominates scale: differentiable structured composition yields more accurate, more parameter-efficient models than prompt-based composition over a substantially larger backbone.
Chinese Translation
法律结果预测必须将客观案件事实与裁判背景区分开。基于优点的裁决依赖于事实证据，而技术性裁决可能依赖于司法裁量。我们提出了一种法官感知的门控多任务学习架构，明确建模这一区别。我们引入了一种细粒度的结果分类法来监督编码器，强制实施结构正则化，以区分不同的语义路径。这种细致的法律课程使我们的门控融合机制能够动态调节对法官身份的依赖。我们在13,937个英国就业法庭裁决上评估了我们的方法。我们将我们的设计与Gemma-4 26B-A4B主干的监督微调（SFT）进行基准比较，其中法官身份和分类法作为提示令牌或自回归输出目标注入。当强制通过单一自回归通道时，这两个上下文信号仅弱耦合。相比之下，将LoRA适配的Gemma-4编码器与我们的门控架构结合定义了这一基准的新状态，同时所需的可训练参数数量比生成的SFT基线少一个数量级，增益集中在最模糊和最稀有的结果类别上。超越准确性，该架构是可解释的；学习到的法官嵌入和校准轮廓定位了裁判背景驱动预测的案件。这些结果表明，对于基于身份的法律结果分类，条件接口的选择主导了规模：可微分的结构化组合比在大得多的主干上基于提示的组合产生更准确、更高效的模型。
View on arXiv Download PDF AI Translation
cs.CL / 54 / 2606.27103

The Riddle Riddle: Testing Flexible Reasoning in Large Language Models and Humans

谜题谜题：测试大型语言模型和人类的灵活推理能力
Fascendini, Bella, McGregor, Kathryn, Gupta, Max D., Griffiths, Thomas L.
Abstract
Humans flexibly adapt their reasoning strategies to the requirements of a given problem. Large language models (LLMs) have performed well on many cognitive tasks, however, it is unclear whether this accuracy is a result of pattern matching from training data or flexible reasoning. Here, we introduce a novel paradigm to test this question: the riddle riddle paradigm. Riddle riddles are word problems written to mimic popular riddles, but altered so their answers only require literal interpretations. Identifying correct answers requires looking past the structure of each question and flexibly apply different reasoning strategies based on the content. If LLMs respond to surface features, such as form, a riddle-like structure should cause models to use an inventive reasoning strategy even when a literal interpretation suffices. Alternatively, if LLMs reason based on content, they should flexibly switch strategies when appropriate. Across two experiments with nine state-of-the-art LLMs and 100 human participants, we show humans and LLMs fail on this paradigm in opposite directions. LLMs were far more accurate on genuine riddles than on riddle riddles (84.9% vs. 50.7%); whereas humans showed the reverse effect (50.5% vs. 80.5%). Error analysis shows that 90.8% of LLM errors on riddle riddles (the condition where they show diminished performance) were due to inappropriate use of inventive reasoning while only 57.6% of human errors on genuine riddles were due to overextending literal reasoning. Thus, while both groups make mistakes, reasoning mistakes are made more often by LLMs than by humans. Overall, LLMs' strong performance on genuine riddles may reflect memory retrieval rather than flexible strategy selection, and without stimuli designed to elicit this contrast, it becomes easy to conflate LLM-generated outputs that look like reasoning with genuine reasoning.
Chinese Translation
人类能够灵活地调整其推理策略以适应特定问题的要求。大型语言模型（LLMs）在许多认知任务中表现良好，然而，目前尚不清楚这种准确性是训练数据中的模式匹配的结果还是灵活推理的结果。在此，我们引入了一种新的范式来测试这个问题：谜题谜题范式。谜题谜题是模仿流行谜语编写的文字问题，但经过修改，使其答案仅需字面解释。识别正确答案需要超越每个问题的结构，并根据内容灵活应用不同的推理策略。如果LLMs对表面特征（如形式）做出反应，那么即使字面解释足够，谜题般的结构也应该导致模型使用创造性的推理策略。相反，如果LLMs基于内容进行推理，那么它们应该在适当时灵活切换策略。在对九个最先进的LLMs和100名人类参与者进行的两项实验中，我们发现人类和LLMs在这一范式中的表现方向相反。LLMs在真实谜题上的准确率远高于在谜题谜题上的准确率（84.9%对50.7%）；而人类则表现出相反的效果（50.5%对80.5%）。错误分析显示，LLMs在谜题谜题上的90.8%的错误（它们表现下降的条件）是由于不当使用创造性推理，而只有57.6%的人类在真实谜题上的错误是由于过度延伸字面推理。因此，尽管两个群体都犯错，但LLMs的推理错误发生得比人类更频繁。总体而言，LLMs在真实谜题上的强大表现可能反映了记忆检索而非灵活策略选择，而没有设计出能够引发这种对比的刺激，容易将看似推理的LLM生成输出与真正的推理混淆。
View on arXiv Download PDF AI Translation
cs.CL / 55 / 2606.27199

Forecasting With LLMs: Improved Generalization Through Feature Steering

使用大型语言模型进行预测：通过特征引导改善泛化能力
Merchant, Humzah, Levy, Bradford
Abstract
Successful forecasting involves identifying patterns between historical and future states of the world which generalize to future observations. We apply LLMs to a variety of forecasting tasks and inspect their internal states using sparse autoencoders to understand whether they appear to rely on time-specific pieces of knowledge versus generalizable patterns. Our analyses identify features associated with both time-aware reasoning and look-ahead-biased reasoning. We then apply the LLMs to an entirely different domain and intervene on these features. We find that amplifying time-awareness features substantially reduces look-ahead bias on forecasting prompts while preserving general reasoning performance. In contrast, steering the candidate look-ahead-bias features does not produce an effect. These results suggest that interpretable temporal features can be used to causally shift LLMs toward more historically grounded reasoning.
Chinese Translation
成功的预测涉及识别历史状态与未来状态之间的模式，这些模式能够推广到未来的观察中。我们将大型语言模型（LLMs）应用于多种预测任务，并使用稀疏自编码器检查其内部状态，以了解它们是否依赖于时间特定的知识片段或可泛化的模式。我们的分析识别出与时间感知推理和前瞻性偏见推理相关的特征。随后，我们将LLMs应用于一个完全不同的领域，并对这些特征进行干预。我们发现，增强时间感知特征显著减少了预测提示中的前瞻性偏见，同时保持了总体推理性能。相比之下，引导候选的前瞻性偏见特征并未产生效果。这些结果表明，可解释的时间特征可以用于因果性地引导LLMs朝向更具历史基础的推理。
View on arXiv Download PDF AI Translation
cs.CL / 56 / 2606.27206

Syntactic Belief Update as the Driver of Garden Path Processing Difficulty

句法信念更新作为花园小径处理难度的驱动因素
Zhou, Alan, Stanojević, Miloš, Hale, John T.
Abstract
Garden path sentences present a processing difficulty for humans -- the sentence prefix leads the listener towards one interpretation, until the listener hears a critical word that shows that the initial interpretation was wrong. Lexical surprisal, a measure that usually predicts sentence processing difficulty quite well, fails to provide good predictions for garden path sentences. We propose an alternative that actively predicts a probability distribution over syntactic trees (its syntactic belief) and updates that distribution after each new word. If a processor is led down a garden path, syntactic beliefs will be wrong and will require a large update at the critical word. The magnitude of the update is measured with a generalized R\'enyi divergence. Crucially, this metric is dependent on lexical items, but is fully independent of the probability of lexical items. This Syntactic Belief Update provides a better fit to the human reading time data on garden path sentences. This suggests a new research direction examining purely non-lexical alternatives to surprisal for psycholinguistics.
Chinese Translation
花园小径句子给人类带来了处理难度——句子的前缀引导听者朝一个解释方向理解，直到听者听到一个关键字，表明最初的解释是错误的。词汇惊讶度（lexical surprisal），一种通常能很好预测句子处理难度的测量指标，在花园小径句子中却未能提供良好的预测。我们提出了一种替代方法，该方法主动预测句法树的概率分布（其句法信念），并在每个新单词之后更新该分布。如果处理器被引导进入花园小径，句法信念将是错误的，并且在关键字处需要进行大幅更新。更新的大小通过广义R'enyi散度来测量。重要的是，这一度量依赖于词汇项，但与词汇项的概率完全独立。这种句法信念更新更好地符合人类在花园小径句子上的阅读时间数据。这提示了一个新的研究方向，探讨心理语言学中纯粹非词汇的惊讶度替代方案。
View on arXiv Download PDF AI Translation
cs.CL / 57 / 2606.27210

Paved with True Intents: Intent-Aware Training Improves LLM Safety Classification Across Training Regimes

铺就真实意图之路：意图感知训练提升大型语言模型的安全分类能力
Ferrao, Jeremias, Müller-Hof, Niclas, Sîrbu, Iustin, Rebedea, Traian, Ziser, Yftah
Abstract
We argue that safety classifiers should model user intent as an explicit signal between the prompt and the final label. To study this, we introduce AIMS, a human-annotated dataset of 1,724 difficult safety prompts, each paired with an intent description and harm label. We use AIMS to evaluate intent-aware training across supervised fine-tuning, preference learning, reasoning distillation, and reinforcement learning. Despite its size, AIMS enables competitive safety classifiers across training regimes: DPO from model-generated intent errors improves over SFT, and intent-conditioned distillation outperforms reasoning-only distillation in most teacher-student pairs. Most notably, directly rewarding intent faithfulness with GRPO yields the strongest average performance across five external safety benchmarks, while our intent-aware models form the inference latency-F1 Pareto frontier. These results show that faithful intent modeling is a compact, high-quality supervision signal for more robust safety classifiers.
Chinese Translation
我们认为安全分类器应将用户意图建模为提示与最终标签之间的显性信号。为此，我们引入了AIMS，一个包含1,724个困难安全提示的人类标注数据集，每个提示都配有意图描述和伤害标签。我们利用AIMS评估在监督微调、偏好学习、推理蒸馏和强化学习等训练模式下的意图感知训练。尽管数据集规模较小，AIMS仍能在不同训练模式下实现竞争力的安全分类器：来自模型生成的意图错误的DPO在SFT上有所提升，而意图条件蒸馏在大多数教师-学生对中优于仅基于推理的蒸馏。最值得注意的是，直接通过GRPO奖励意图的忠实性在五个外部安全基准测试中产生了最强的平均性能，而我们的意图感知模型形成了推理延迟-F1的帕累托前沿。这些结果表明，忠实的意图建模是一个紧凑且高质量的监督信号，有助于构建更稳健的安全分类器。
View on arXiv Download PDF AI Translation
cs.CL / 58 / 2606.27228

Compositionality and the lexicon in evolutionary semantics

进化语义中的组合性与词汇
Carcassi, Fausto
Abstract
Formal semantics has shown that sentence meanings arise by recursively composing lexical meanings, yet much of the literature on semantic universals models either lexicons with fixed signal structures or holistic composition without interpretable lexical parts. We introduce a framework that integrates this fundamental insight of formal semantics in evolutionary modeling, by allowing lexical meanings and a composition function to co-evolve under pressures for conceptual simplicity and communicative accuracy. We apply this framework to the evolution of quantificational meaning. Analyzing the Pareto frontier, we find that the most well-known semantic universal, conservativity, emerges as an efficient system-wide abstraction. The account is sensitive to syntactic structure and helps reconcile tensions between empirical evidence on quantifier learnability and prior evolutionary models. More broadly, the results demonstrate that the picture of sentential meaning developed in formal semantics can be productively combined with evolutionary modeling. The framework offers a template for studying universals that involve global compression within a grammatical category, semantic specialization of syntactic arguments, and the co-evolution of lexical and compositional meaning.
Chinese Translation
形式语义学表明，句子意义通过递归组合词汇意义而产生，然而，关于语义普遍性的文献大多将词汇视为具有固定信号结构或没有可解释词汇部分的整体组合。我们提出一个框架，将形式语义学的这一基本见解整合到进化建模中，允许词汇意义和组合函数在概念简单性和交际准确性的压力下共同进化。我们将该框架应用于量化意义的进化。通过分析帕累托前沿，我们发现最著名的语义普遍性——保守性，作为一个高效的系统性抽象而出现。该解释对句法结构敏感，并有助于调和关于量词可学习性的实证证据与先前进化模型之间的紧张关系。更广泛地说，结果表明，形式语义学中发展出的句子意义的图景可以与进化建模有效结合。该框架为研究涉及语法类别内全球压缩、句法论元的语义专业化以及词汇与组合意义的共同进化的普遍性提供了模板。
View on arXiv Download PDF AI Translation
cs.CL / 59 / 2606.27229

CARVE: Content-Aware Recurrent with Value Efficiency for Chunk-Parallel Linear Attention

CARVE：具有价值效率的内容感知递归块并行线性注意力
Dutta, Sayak
Abstract
Recurrent models must forget in order to remember, yet the state of the art decides what to erase without consulting what is stored -- the gate sees only the arriving token, not the memory it is about to modify. This memory-blind gating is one of three coupled defects in the leading delta-rule architecture (GDN-2): the value-axis erase mask wastes parameters at the scale of the value projection, and -- as we prove -- mathematically prevents the WY-form triangular chunk solver that makes recurrent training competitive with Transformers. We introduce CARVE (Content-Aware Recurrent with Value Efficiency), which resolves all three problems through one principle: erase only on the key axis. This is provably necessary and sufficient for the WY-form solver to remain valid. Within it, CARVE reuses the recurrent output tensor -- already written to GPU memory -- as a free content signal for the erase gate, and replaces the per-value write-gate projection with a single scalar per head. At initialisation CARVE is bit-identical to GDN-2; any quality difference emerges from what the content gate learns. At 1.3B parameters trained on 100B tokens, CARVE achieves WikiText perplexity 15.72 (minus 0.18 vs. GDN-2, a 4.5-sigma effect), leads every recurrent baseline on nine common-sense reasoning benchmarks, and sets state of the art on every RULER retrieval probe -- at 0.4% throughput overhead, 13% lower peak memory, and 19% fewer parameters. Six formal theorems cover memory capacity, Lyapunov stability, gradient flow, expressivity separation, Pareto-optimal chunk size, and hybrid optimality.
Chinese Translation
递归模型必须忘记以便记住，然而，当前的先进技术在决定删除内容时并未参考已存储的信息——门控机制仅关注到达的标记，而非即将被修改的记忆。这种对记忆的盲目门控是领先的增量规则架构（GDN-2）中的三个相互关联的缺陷之一：价值轴的擦除掩码在价值投影的规模上浪费了参数，并且——正如我们所证明的——在数学上阻止了使递归训练与变换器竞争的WY形式三角块求解器。我们提出了CARVE（具有价值效率的内容感知递归），通过一个原则解决了这三个问题：仅在关键轴上进行擦除。这在数学上是保持WY形式求解器有效所必需且足够的。在此框架内，CARVE重用递归输出张量——已写入GPU内存——作为擦除门的免费内容信号，并用每个头部的单个标量替换每个值的写门投影。在初始化时，CARVE与GDN-2完全相同；任何质量差异均源于内容门的学习。在训练了130亿参数和1000亿标记后，CARVE在WikiText上的困惑度达到了15.72（比GDN-2低0.18，效果达到4.5个标准差），在九个常识推理基准上领先于每个递归基线，并在每个RULER检索探针上设定了最新的技术水平——在0.4%的吞吐量开销、13%的峰值内存降低和19%的参数减少下。六个正式定理涵盖了记忆容量、Lyapunov稳定性、梯度流、表达能力分离、Pareto最优块大小和混合最优性。
View on arXiv Download PDF AI Translation
cs.CL / 60 / 2606.27233

Bridging Talk and Thought: Understanding Dialogue Dynamics Across Collaborative Problem-Solving Contexts

弥合对话与思维：理解协作问题解决背景下的对话动态
Liu, Zhengyuan, Yin, Stella Xin, Kan, Min-Yen, Chen, Nancy F.
Abstract
We present a conceptual framework for analyzing dialogue in collaborative problem-solving contexts, with an emphasis on the emerging dynamics of human-AI and multi-agent collaboration. As intelligent systems become active agents capable of autonomous reasoning and strategic cooperation, understanding the dialogic interaction during collaborative problem solving is increasingly important for optimizing and evaluating such partnerships. Our framework addresses key limitations in current analytical approaches through a hierarchical two-layer coding scheme that integrates cognitive and non-cognitive problem solving with metacognitive regulatory mechanisms. We demonstrate its effectiveness and generalizability across nine datasets spanning multiple domains, and provide insights into how humans and agents coordinate their knowledge, skills, and efforts to solve complex problems, showing in particular that metacognitive regulation can be an essential discriminator of deeper collaboration.
Chinese Translation
我们提出了一个用于分析协作问题解决背景下对话的概念框架，重点关注人机和多智能体协作中出现的动态。随着智能系统成为能够自主推理和战略合作的主动代理，理解协作问题解决过程中的对话互动对于优化和评估这种伙伴关系变得越来越重要。我们的框架通过一个层次化的双层编码方案解决了当前分析方法中的关键局限，该方案将认知和非认知问题解决与元认知调节机制相结合。我们展示了该框架在涵盖多个领域的九个数据集中的有效性和普遍适用性，并提供了关于人类与智能体如何协调其知识、技能和努力以解决复杂问题的见解，特别表明元认知调节可以成为更深层次协作的重要区分因素。
View on arXiv Download PDF AI Translation
cs.CL / 61 / 2606.27237

LMs as Task-Specific Knowledge Bases: An Interpretability Analysis

语言模型作为任务特定知识库：可解释性分析
Elhelo, Amit, Globerson, Amir, Geva, Mor
Abstract
Language models (LMs) capture large amounts of factual knowledge applicable to a wide range of tasks, motivating the view of their parameters as a knowledge base. An important property of knowledge bases is that different queries for the same fact return consistent results, drawing on a single source of truth. We investigate whether LMs satisfy this property through behavioral and mechanistic analyses. Our results suggest that they encode knowledge in a task-specific manner. Behaviorally, facts acquired on one task frequently fail to co-emerge on others during training. Parameter localization experiments suggest a mechanistic explanation, revealing distinct parameter subsets underlying different tasks for the same fact. Finally, we show that chain-of-thought reasoning draws part of its effectiveness from engaging task-specific parameters beyond those tied to the evaluation task. Our findings suggest that what the model knows and how it is asked are intertwined in parameter space, undermining the "knowledge base" analogy and carrying implications for the reliability and controllability of factual knowledge in LMs.
Chinese Translation
语言模型（LMs）捕捉了大量适用于广泛任务的事实知识，这使得将其参数视为知识库的观点得以成立。知识库的一个重要特性是，对于同一事实的不同查询应返回一致的结果，依赖于单一的真实来源。我们通过行为和机制分析研究语言模型是否满足这一特性。我们的结果表明，它们以任务特定的方式编码知识。在行为上，在一个任务上获得的事实在训练期间常常无法在其他任务上共同出现。参数定位实验提供了机制上的解释，揭示了同一事实在不同任务下所依赖的不同参数子集。最后，我们展示了思维链推理的有效性部分来源于参与任务特定参数，而不仅仅是与评估任务相关的参数。我们的发现表明，模型所知与提问方式在参数空间中是交织在一起的，这削弱了“知识库”类比的有效性，并对语言模型中事实知识的可靠性和可控性产生了影响。
View on arXiv Download PDF AI Translation
cs.CL / 62 / 2606.27275

How Surprising Is Historical Italian to Language Models? Tokenization Tax, Comprehension Tax, and a Simple Mitigation

历史意大利语对语言模型的惊讶程度如何？标记化成本、理解成本及简单缓解措施
Levchenko, Maria
Abstract
Large language models (LLMs) are increasingly critical to digital library workflows, yet their ability to process historical language remains poorly understood. Historical difficulty is typically treated as a monolithic barrier, conflating orthographic variation, linguistic distance, and pretraining exposure. In this paper, we propose a diagnostic framework that decomposes this difficulty into four distinct dimensions: tokenization cost, predictive uncertainty (surprisal), semantic robustness, and context sensitivity. We evaluate this framework on three datasets spanning three centuries: (1) a newly curated corpus of 17th-century Italian texts (1610-1689) digitized from original page images; (2) canonical 19th-century Italian "I Promessi Sposi" serving as a high-exposure control; and (3) 18th-century Russian civil print books as a contrastive orthographic stress test. Our results reveal a distinct dissociation between encoding cost and comprehension. While Russian and early modern Italian incur comparable tokenization penalties (25-30% inflation), their predictive difficulty diverges sharply. 17th-century Italian is on average 2.4 times more surprising than its modern equivalent - with academic prose reaching 3.2 times - whereas Russian shows only a modest increase. But predictive uncertainty does not imply representational degradation: embedding similarity remains robust (> 0.85) across all datasets, confirming that models can represent historical meaning even when generation is unstable. Finally, we demonstrate that a minimal temporal context prompt reduces historical surprisal by approximately 60%, offering a simple, model-agnostic mitigation. These findings suggest that while historical text imposes a consistent encoding tax, digital libraries can safely deploy LLMs for semantic retrieval tasks, provided that generative applications are carefully adapted.
Chinese Translation
大型语言模型（LLMs）在数字图书馆工作流程中变得越来越重要，但它们处理历史语言的能力仍然不甚了解。历史难度通常被视为一个单一的障碍，混淆了正字法变异、语言距离和预训练暴露。在本文中，我们提出了一个诊断框架，将这种难度分解为四个不同的维度：标记化成本、预测不确定性（惊讶度）、语义稳健性和上下文敏感性。我们在三个跨越三个世纪的数据集上评估了该框架：（1）一个新整理的17世纪意大利文本语料库（1610-1689），从原始页面图像数字化而来；（2）作为高暴露对照的19世纪经典意大利文献《婚约》（I Promessi Sposi）；以及（3）作为对比正字法压力测试的18世纪俄国民用印刷书籍。我们的结果揭示了编码成本与理解之间的明显解耦。尽管俄语和早期现代意大利语的标记化惩罚相当（25-30%的通货膨胀），但它们的预测难度却大相径庭。17世纪意大利语的惊讶程度平均是其现代对应物的2.4倍——学术散文甚至达到3.2倍——而俄语仅显示出适度的增加。但预测不确定性并不意味着表征退化：嵌入相似性在所有数据集中保持稳健（> 0.85），确认模型能够表示历史意义，即使生成过程不稳定。最后，我们证明，最小的时间上下文提示将历史惊讶度降低了约60%，提供了一种简单的、与模型无关的缓解措施。这些发现表明，尽管历史文本施加了一致的编码成本，数字图书馆仍然可以安全地将LLMs用于语义检索任务，前提是生成应用经过仔细调整。
View on arXiv Download PDF AI Translation
cs.CL / 63 / 2606.27306

Multilingual Reasoning Cascades Need More Context

多语言推理级联需要更多上下文
Mazumder, Arnav, Zhang, Dengjia, Li, Shuyue Stella, Tsvetkov, Yulia, Bafna, Niyati
Abstract
Translation cascades for reasoning translate the query from another language to English, reason in English, and translate the answer back to the original language. This is a competitive approach to multilingual reasoning, but structurally lossy, since each stage discards information later stages may need, including cues for cultural grounding, register, and disambiguation. We examine the benefits of a simple and training-free intervention: a context-aware translation cascade, which additionally provides the original question, the English translated question, and the reasoning trace to the context of the final translation module. We evaluate gains across nine multilingual benchmarks including various task types, three backbone models, and 285 high-, mid-, and low-resource languages, and demonstrate strong gains for open-ended generation across models and resource regimes. We show that the original language question carries most of the beneficial context. Our study emphasizes the need to better design information flow in machine translation cascades for mitigating error propagation, and provides a simple and actionable default strategy: preserve the original user question until the end of the pipeline.
Chinese Translation
推理的翻译级联将查询从另一种语言翻译为英语，在英语中进行推理，然后将答案翻译回原始语言。这是一种竞争性的多语言推理方法，但在结构上存在信息损失，因为每个阶段都会丢弃后续阶段可能需要的信息，包括文化基础、语域和消歧义的线索。我们研究了一种简单且无需训练的干预措施的好处：上下文感知翻译级联，它额外提供原始问题、英语翻译问题和推理轨迹，以便于最终翻译模块的上下文。我们在九个多语言基准上评估了收益，包括各种任务类型、三种基础模型以及285种高、中、低资源语言，并展示了在各模型和资源模式下开放式生成的显著收益。我们表明，原始语言问题承载了大部分有益的上下文。我们的研究强调了在机器翻译级联中更好地设计信息流的必要性，以减轻错误传播，并提供了一种简单且可行的默认策略：在整个流程结束之前保留原始用户问题。
View on arXiv Download PDF AI Translation
cs.CL / 64 / 2606.27314

Beyond Surface Forms: A Comprehensive, Mechanism-Oriented Taxonomy of Indirect Linguistic Encoding for LLM-Based Coded Language Detection

超越表面形式：基于机制的间接语言编码的全面分类法用于LLM基础的编码语言检测
Firoozfar, Hamid Reza, Abolhasani, Mohammadsadegh, Mousavi, Reza, Hu, Paul Jen-Hwa
Abstract
To avoid moderation and surveillance on social media, some users routinely invent indirect linguistic expressions (ILE) that camouflage sensitive meanings. Such expressions surface as algospeak, euphemisms, and adversarial obfuscation, depending on intent and context, and they involve recurring encoding mechanisms. We propose a comprehensive, mechanism-oriented taxonomy of ILE that abstracts away from communicative goals and instead categorizes the underlying operations through which meaning is encoded and recovered. We evaluate the taxonomy by incorporating it into LLM prompts and comparing it with four existing taxonomies and a no-taxonomy baseline, using 2,000 manually annotated TikTok and Bluesky posts. The proposed taxonomy attains the strongest document- and span-level performance across the three LLMs, achieving an improvement of 4.7% in accuracy and 5.4% in F1 over the best-performing benchmark. The empirical results reveal the importance of a comprehensive, mechanism-oriented taxonomy as a stable scaffold for detecting emerging coded language and a useful input to content moderation. Disclaimer: This paper contains content that may be profane, vulgar, or offensive.
Chinese Translation
为了避免社交媒体上的审查和监控，一些用户常常创造间接语言表达（ILE），以掩盖敏感含义。这些表达形式根据意图和上下文表现为算法语言（algospeak）、委婉语和对抗性模糊，涉及重复的编码机制。我们提出了一种全面的、以机制为导向的ILE分类法，该分类法抽象化了交流目标，而是通过对意义编码和恢复的基本操作进行分类。我们通过将该分类法纳入LLM提示中进行评估，并与四个现有分类法和一个无分类法基准进行比较，使用了2,000条手动标注的TikTok和Bluesky帖子。所提议的分类法在三种LLM中在文档级和跨度级表现上均取得了最佳效果，相较于表现最好的基准，准确率提高了4.7%，F1值提高了5.4%。实证结果揭示了全面的、以机制为导向的分类法作为检测新兴编码语言的稳定支架的重要性，并为内容审查提供了有用的输入。免责声明：本文包含可能粗俗、低俗或冒犯的内容。
View on arXiv Download PDF AI Translation
cs.CL / 65 / 2606.27316

LLM-Based Examination of Eligibility Criteria from Securities Prospectuses at the German Central Bank

基于大型语言模型的德国中央银行证券招股说明书资格标准审查
Hamotskyi, Serhii, Gautam, Akash Kumar, Hänig, Christian
Abstract
Verifying the eligibility of securities as collateral is a key responsibility of the German Central Bank. However, manually verifying these assets against legal and financial criteria within lengthy, semi-structured, and often bilingual prospectuses is a resource-intensive task. While previous efforts utilized traditional Named Entity Recognition (NER) for information extraction, these methods can struggle with OCR noise, linguistic variance, and rigid span-based constraints, and the need for manually annotated training data for each relevant annotation type. In this paper, we present the first case study applying Large Language Models (LLMs) to the eligibility examination process, shifting the paradigm toward a generative Information Extraction pipeline. Our approach decomposes the task into extraction, normalization, and interpretation, allowing for greater flexibility in handling noisy text and interleaved German-English content. We further introduce a value-based evaluation methodology using LLM-as-a-judge, which offers a more semantic assessment than location-based metrics. Our results demonstrate that LLM-based systems achieve high precision (up to 91%) in document-level eligibility, exhibiting a conservative operating profile that minimizes false acceptance.
Chinese Translation
验证证券作为抵押品的资格是德国中央银行的一项关键职责。然而，在冗长、半结构化且常常是双语的招股说明书中，手动验证这些资产是否符合法律和财务标准是一项资源密集型的任务。尽管以往的努力利用传统的命名实体识别（NER）进行信息提取，但这些方法在处理光学字符识别（OCR）噪声、语言变异和严格的基于跨度的约束时可能会遇到困难，并且需要为每种相关注释类型提供手动标注的训练数据。在本文中，我们首次展示了将大型语言模型（LLMs）应用于资格审查过程的案例研究，转变了信息提取流程的范式，朝着生成式信息提取管道迈进。我们的方法将任务分解为提取、标准化和解释，从而在处理噪声文本和交错的德英内容时提供更大的灵活性。我们进一步引入了一种基于价值的评估方法，使用LLM作为评判者，这种方法提供了比基于位置的指标更具语义的评估。我们的结果表明，基于LLM的系统在文档级资格审查中实现了高达91%的高精度，展现出一种保守的操作特征，最小化了错误接受率。
View on arXiv Download PDF AI Translation
cs.CL / 66 / 2606.27330

Empowering GUI Agents via Autonomous Experience Exploration and Hindsight Experience Utilization for Task Planning

通过自主经验探索和事后经验利用增强GUI代理的任务规划能力
Men, Tianyi, Jin, Zhuoran, Cao, Pengfei, Chen, Yubo, Liu, Kang, Zhao, Jun
Abstract
Multimodal web agents can assist humans in operating repetitive GUI tasks, where effective task planning is essential for decomposing complex tasks into executable actions. While small open source MLLMs are cost efficient and privacy preserving compared with commercial large models, they suffer from weak planning and limited cross website generalization. To address these limitations, we introduce the planning experience exploration and utilization (PEEU) method, which autonomously explores environments to discover experiences and utilizes hindsight experience to synthesize strictly aligned, high level training data. To quantitatively analyze the generalization behaviors driving this performance, we propose the task decomposition hierarchical analysis framework (TDHAF) to systematically study compositional generalization across three task granularities: low, middle and high levels. Our analysis reveals that mastering low level atomic skills does not guarantee high level planning competence, while high level task training yields stronger OOD generalization. Experiments on real world benchmarks demonstrate PEEU's superior effectiveness: our 7B model achieves 30.6% accuracy, outperforming the much larger Qwen2.5-VL-32B model. These demonstrate constructing hindsight high level tasks and leveraging experiences is crucial for OOD planning abilities of small MLLMs.
Chinese Translation
多模态网络代理可以帮助人类执行重复的图形用户界面（GUI）任务，其中有效的任务规划对于将复杂任务分解为可执行的动作至关重要。尽管小型开源的多模态大语言模型（MLLMs）在成本效益和隐私保护方面优于商业大型模型，但它们在规划能力和跨网站泛化方面存在不足。为了解决这些局限性，我们提出了规划经验探索与利用（PEEU）方法，该方法能够自主探索环境以发现经验，并利用事后经验合成严格对齐的高层次训练数据。为了定量分析驱动这一性能的泛化行为，我们提出了任务分解层次分析框架（TDHAF），系统研究三个任务粒度（低、中、高级）下的组合泛化。我们的分析表明，掌握低层次原子技能并不保证高层次规划能力，而高层次任务训练则能产生更强的超出分布（OOD）泛化能力。在真实世界基准测试中的实验表明，PEEU的有效性优于其他方法：我们的7B模型实现了30.6%的准确率，超越了更大规模的Qwen2.5-VL-32B模型。这些结果表明，构建事后高层次任务并利用经验对于小型MLLM的OOD规划能力至关重要。
View on arXiv Download PDF AI Translation
cs.CL / 67 / 2606.27347

Mapping Political-Elite Networks in Europe with a Multilingual Joint Entity-Relation Extraction Pipeline

利用多语言联合实体-关系提取管道绘制欧洲政治精英网络
Solovev, Kirill, Lasser, Jana
Abstract
Whether political elites organise into rent-seeking coalitions that capture public resources or civic networks that sustain governance is a central question in comparative politics. Yet observing these complex, informal, and adversarial ties at scale has historically required intensive manual coding, while automated text-as-data methods have largely been limited to simple co-occurrence. Recent large language model (LLM) approaches offer a path forward but often rely on proprietary APIs, lack cross-lingual capability, and struggle with scalable entity resolution. We present a modular, fully open-weight pipeline for multilingual joint entity-relation extraction that builds signed, temporal knowledge graphs from massive unstructured news corpora. It combines span-based named-entity recognition (NER) with a three-stage linking cascade mapping mentions to language-independent Wikidata identifiers; a high-throughput, ontology-constrained mixture-of-experts model then uses guided decoding to extract directed, signed relationships grounded in a domain ontology. A full-coverage spot-check against a 3491-relation gold standard shows high textual correctness (68.2% strict to 93.7% lenient). Two large-scale case studies validate the pipeline against the public record. In Austria, it reconstructs a political party's complete lifecycle, dating internal fractures and tracking personnel into successor factions and court convictions. In a Polish corpus, it uncovers the overlapping economic and governance networks of state-enterprise patronage, alongside the structurally balanced, signed conflict network of the polarized Civic Platform (Platforma Obywatelska, PO)--Law and Justice (Prawo i Sprawiedliwo\'s\'c, PiS) duopoly. By bridging raw multilingual text and structured relational data, our framework provides a robust, replicable foundation for cross-national empirical computational social science.
Chinese Translation
政治精英是否组织成捕获公共资源的寻租联盟或维持治理的公民网络是比较政治学中的一个核心问题。然而，观察这些复杂的、非正式的和对抗性的关系在规模上历来需要大量的手动编码，而自动化的文本数据方法在很大程度上仅限于简单的共现。最近的大型语言模型（LLM）方法提供了一条前进的道路，但通常依赖于专有API，缺乏跨语言能力，并且在可扩展的实体解析方面存在困难。我们提出了一种模块化、完全开放权重的多语言联合实体-关系提取管道，该管道从大量非结构化新闻语料库中构建签名的、时间性的知识图谱。它结合了基于跨度的命名实体识别（NER）与一个三阶段的链接级联，将提及映射到语言无关的Wikidata标识符；一个高通量、受本体约束的专家混合模型随后使用引导解码提取基于领域本体的定向、签名关系。对3491个关系金标准的全面覆盖抽查显示出高文本正确性（严格标准下68.2%，宽松标准下93.7%）。两个大规模案例研究验证了该管道与公共记录的一致性。在奥地利，它重建了一个政党的完整生命周期，追溯内部裂痕并追踪人员进入继任派系和法院定罪。在波兰语料库中，它揭示了国家企业赞助的重叠经济和治理网络，以及极化的公民平台（Platforma Obywatelska, PO）与法律与公正（Prawo i Sprawiedliwość, PiS）二元对立的结构平衡、签名冲突网络。通过桥接原始多语言文本和结构化关系数据，我们的框架为跨国实证计算社会科学提供了一个稳健、可复制的基础。
View on arXiv Download PDF AI Translation
arXiv Reader
Automatically generated on 2026-07-16 10:17:13
Powered by OpenAI GPT & arXiv