cs.RO / 1 / 2607.09701
EgoSteer: A Full-Stack System Towards Steerable Dexterous Manipulation from Egocentric Videos
EgoSteer：一个面向从自我中心视频进行可操控灵巧操作的全栈系统
Abstract
Steerability is a defining capability of generalist robot policies, yet remains largely absent in dexterous-hand systems for lack of large-scale, language-aligned, and action-accurate demonstration data. To address this bottleneck, we present a full-stack system that scales dexterous VLA pre-training from egocentric human videos and enables data-efficient real-robot post-training. It integrates EgoSmith, a data pipeline that curates in-the-wild egocentric videos into 9.6K hours of high-quality pre-training data with 9x higher throughput and better accuracy than prior SOTA; a unified robot stack for teleoperation and human-in-the-loop correction; and EgoSteer, a world-model-enhanced VLA trained on optimized infrastructure. Human-data pre-training equips EgoSteer with language-guided manipulation priors, which are grounded through robot post-training and improved by DAgger refinement. Empirically, EgoSteer robustly executes free-form instructions across 40+ diverse tasks, demonstrating failure recovery, dexterity, and generalization. The pre-trained model also few-shot adapts to complex long-horizon tasks, including box folding, on two embodiments with 75+% success. We open-source the system, data, and model at https://egosteer.github.io/.
Chinese Translation
可操控性是通用机器人策略的一个重要特性，但由于缺乏大规模、与语言对齐且动作准确的示范数据，这一特性在灵巧手系统中仍然大多缺失。为了解决这一瓶颈，我们提出了一个全栈系统，该系统通过自我中心的人类视频扩展灵巧 VLA（Visual Language Alignment）预训练，并实现数据高效的真实机器人后训练。该系统集成了 EgoSmith，一个数据管道，将野外的自我中心视频整理成 9.6K 小时的高质量预训练数据，其吞吐量比之前的最先进技术高出 9 倍，准确性更佳；一个统一的机器人堆栈用于远程操作和人机协作校正；以及 EgoSteer，一个在优化基础设施上训练的世界模型增强的 VLA。人类数据预训练为 EgoSteer 提供了语言引导的操作先验，这些先验通过机器人后训练进行扎根，并通过 DAgger 精炼进行改进。实证结果表明，EgoSteer 在 40 多个多样化任务中稳健地执行自由形式的指令，展示了故障恢复、灵巧性和泛化能力。预训练模型还能够在两个实现上对复杂的长时间任务（包括盒子折叠）进行少量样本适应，成功率超过 75%。我们将在 https://egosteer.github.io/ 上开源该系统、数据和模型。
cs.RO / 2 / 2607.09716
RoboNav-Arm: Agentic AI-Driven Navigation and Obstacle Avoidance for Robotic Manipulator in Cluttered Environments
RoboNav-Arm：基于代理智能的导航与障碍物避让框架，用于杂乱环境中的机器人操控器
Abstract
Robotic manipulators operating in unstructured environments face significant challenges in safely executing goal-directed tasks due to dynamic and unforeseen obstacles, while traditional methods rely on prior knowledge or fixed perception pipelines, limiting adaptability. We propose a framework for safe task execution with effective obstacle avoidance. The environment module performs real-time obstacle detection, 3D localization, and ground surface geometry estimation. It then generates a structured semantic report that includes obstacle positions, object geometry and shape, and whether obstacles lie inside, outside, or within critical interaction zones. A central coordination module manages the overall system by handling tool invocation (e.g., memory and MoveIt collision scene updates), facilitating communication between modules, and continuously monitoring task progress until completion. Furthermore, a planning module selects an appropriate motion planning algorithm, such as RRTConnect, RRT*, or BiTRRT, based on the current environment configuration and goal requirements. The trajectory generated by the planner is further analyzed and refined to ensure safe and collision-free task execution. The proposed approach is evaluated in Gazebo Classic , demonstrating robustness in dynamic scenarios.
Chinese Translation
在非结构化环境中操作的机器人操控器面临着由于动态和不可预见的障碍物而安全执行目标导向任务的重大挑战，而传统方法依赖于先前知识或固定的感知管道，限制了适应性。我们提出了一种安全任务执行的框架，具有有效的障碍物避让能力。环境模块执行实时障碍物检测、三维定位和地面表面几何估计。随后，它生成一个结构化的语义报告，其中包括障碍物位置、物体几何形状以及障碍物是否位于关键交互区域的内部、外部或其中。中央协调模块通过处理工具调用（例如，内存和 MoveIt 碰撞场更新）来管理整体系统，促进模块之间的通信，并持续监控任务进展直至完成。此外，规划模块根据当前环境配置和目标要求选择合适的运动规划算法，如 RRTConnect、RRT* 或 BiTRRT。规划器生成的轨迹进一步分析和优化，以确保安全和无碰撞的任务执行。所提出的方法在 Gazebo Classic 中进行了评估，展示了在动态场景中的鲁棒性。
cs.RO / 3 / 2607.09725
Learning High-Level Decision Making with an Interaction-Aware Attention-Based Network in Autonomous Driving
在自动驾驶中基于交互感知的注意力网络学习高层决策制定
Abstract
Reliable learning-based high-level decision making for lane changes and speed control in automated driving must accommodate dynamically sized inputs due to varying scene traffic flow. DeepSet and its variants represent the state of the art among shared-encoder approaches; however, they neglect explicit traffic interaction modeling, limiting performance in negotiation-intensive scenarios such as intersections. Attention-based methods capture interactions among static and dynamic agents, but incur quadratic memory and computational complexity and provide limited control over representation granularity. Inspired by Perceiver IO, an attention-based architecture, DecisionPerceiver, is proposed to project dynamic agent features into a fixed-size latent space, where feature granularity is regulated by the number of latent queries, improving scalability for larger networks. A finer discretization of the action set is further proposed to increase the performance gain due to interaction awareness. Extensive evaluations across three driving scenarios that require different levels of interaction awareness demonstrate consistent performance gains and generalization across various navigation objectives. In addition, the proposed architecture is assessed in scenarios with an increasing number of vehicles to demonstrate scalability.
Chinese Translation
可靠的基于学习的高层决策制定在自动驾驶中的车道变换和速度控制必须适应由于场景交通流变化而动态变化的输入。DeepSet及其变体代表了共享编码器方法的最新进展；然而，它们忽视了显式的交通交互建模，限制了在如交叉口等需要协商的场景中的性能。基于注意力的方法能够捕捉静态和动态代理之间的交互，但会导致二次记忆和计算复杂度，并且对表示粒度的控制有限。受Perceiver IO启发，提出了一种基于注意力的架构DecisionPerceiver，旨在将动态代理特征投影到固定大小的潜在空间中，其中特征粒度由潜在查询的数量调节，从而提高了大型网络的可扩展性。进一步提出了对动作集进行更细粒度离散化的方案，以提高由于交互感知带来的性能提升。在需要不同程度交互感知的三种驾驶场景中进行了广泛评估，结果表明在各种导航目标下均表现出一致的性能提升和泛化能力。此外，所提架构在车辆数量逐渐增加的场景中进行了评估，以证明其可扩展性。
cs.RO / 4 / 2607.09736
Saturation-Aware Robust Trajectory Optimization for Reusable Launch Vehicles via Differentiable Physics
基于可微物理的可重复使用发射器饱和感知鲁棒轨迹优化
Abstract
The high-angle-of-attack flip maneuver of reusable launch vehicles presents significant challenges for robust trajectory optimization due to the combined effects of highly nonlinear dynamics, aerodynamic uncertainties, and actuator saturation. This paper presents a differentiable physics framework for saturation-aware robust trajectory optimization. At its core, a Differentiable Particle Tube Control (DPTC) scheme is developed to optimize uncertainty evolution through an ensemble-based distribution shaping strategy. State uncertainty is represented by a Lagrangian particle ensemble, while hard actuator projection operators are embedded directly into the computational graph, enabling the joint optimization of the nominal feedforward trajectory and a time-varying feedback policy via end-to-end backpropagation. The proposed framework is evaluated against an automatic differentiation-based Successive Convexification (AD-SCvx) baseline combined with a conventional covariance steering feedback strategy. Six-degree-of-freedom Monte Carlo simulations demonstrate that, although the baseline achieves nominal fuel-optimal solutions, its unconstrained feedback formulation becomes susceptible to actuator saturation under aerodynamic disturbances, leading to degraded closed-loop robustness. In contrast, the proposed DPTC framework proactively performs a constraint-aware performance trade-off by relaxing spatial tracking to preserve critical control authority. These results demonstrate that integrating differentiable physics with ensemble-based optimization provides an effective and practical framework for robust guidance in highly constrained aerospace flight systems.
Chinese Translation
可重复使用发射器的高攻角翻转机动由于高度非线性动力学、气动不确定性和执行器饱和的综合影响，对鲁棒轨迹优化提出了重大挑战。本文提出了一种基于可微物理的饱和感知鲁棒轨迹优化框架。其核心是开发了一种可微粒子管控（Differentiable Particle Tube Control, DPTC）方案，通过基于集成的分布塑形策略来优化不确定性演化。状态不确定性通过拉格朗日粒子集表示，而硬执行器投影算子直接嵌入计算图中，从而使得名义前馈轨迹和时变反馈策略的联合优化能够通过端到端反向传播实现。所提出的框架与基于自动微分的连续凸化（Automatic Differentiation-based Successive Convexification, AD-SCvx）基线相比较，并结合传统的协方差引导反馈策略进行评估。六自由度的蒙特卡洛仿真表明，尽管基线能够实现名义燃料最优解，但其无约束反馈形式在气动扰动下容易受到执行器饱和的影响，导致闭环鲁棒性下降。相比之下，所提出的DPTC框架通过放宽空间跟踪以保持关键控制权，主动进行约束感知的性能权衡。这些结果表明，将可微物理与基于集成的优化相结合，为高度受限的航空航天飞行系统提供了有效且实用的鲁棒引导框架。
cs.RO / 5 / 2607.09741
SWIFT: A Small-World Interaction Framework for Flow-Aware Trajectory Prediction in Autonomous Driving
SWIFT：一种用于自主驾驶流量感知轨迹预测的小世界交互框架
Abstract
Accurate trajectory prediction in autonomous driving hinges on modeling dynamic and context-dependent interactions among traffic agents. However, most existing approaches are purely data-driven and lack structural priors, which limits their generalization under distribution shifts. In this work, interaction modeling is revisited through the structure and dynamics of traffic networks, and SWIFT (Small-World Interaction Framework for Trajectory prediction) is proposed as a unified framework that integrates small-world networks with traffic flow theory. SWIFT introduces structural inductive biases via a Small-World Interaction Network that captures both local and global dependencies, and a Flow Regime Encoder that adapts the interaction structure to scene-level traffic states. Interaction reasoning is further enhanced through a multi-relational graph module that explicitly encodes direct and higher-order agent relationships. Extensive experiments on three real-world datasets, nuScenes, MoCAD, and NGSIM, show that SWIFT consistently outperforms strong baselines in prediction accuracy across diverse traffic regimes. Beyond accuracy, SWIFT exhibits improved generalization to unseen locations and regimes, robustness under noisy observations, and strong performance with limited training data, supporting the effectiveness of its structure-aware design.
Chinese Translation
自主驾驶中的准确轨迹预测依赖于对交通参与者之间动态和上下文依赖交互的建模。然而，大多数现有方法纯粹基于数据，缺乏结构性先验，这限制了它们在分布变化下的泛化能力。在本研究中，通过交通网络的结构和动态重新审视交互建模，提出了SWIFT（Small-World Interaction Framework for Trajectory prediction）作为一个统一框架，将小世界网络与交通流理论相结合。SWIFT通过小世界交互网络引入结构性归纳偏差，捕捉局部和全局依赖关系，并通过流量状态编码器将交互结构适应于场景级交通状态。通过一个多关系图模块进一步增强交互推理，该模块明确编码直接和高阶代理关系。在三个真实世界数据集nuScenes、MoCAD和NGSIM上的大量实验表明，SWIFT在不同交通状态下的预测准确性上始终优于强基线。除了准确性，SWIFT在未见位置和状态下展现出更好的泛化能力，在噪声观察下表现出强健性，并在有限训练数据下表现出强劲性能，支持其结构感知设计的有效性。
cs.RO / 6 / 2607.09756
LLM-Centric Agentic AI for UAV Swarms: Architecture, Enabling Technologies, and Open Problems
面向无人机群的以大型语言模型为中心的自主人工智能：架构、支持技术与开放问题
Abstract
Uncrewed Aerial Vehicle (UAV) swarms have significant potential for applications such as Search and Rescue (SAR) and environmental monitoring, but their real-world deployment is limited by a lack of situational awareness, intermittent connectivity, and significant cybersecurity risks. Agentic Artificial Intelligence (AI) represents a shift from standalone Large Language Model (LLM) toward closed-loop cognitive architectures that integrate perception, memory, reasoning/planning, and action to enable adaptive, goal-directed swarm behavior. Within this framework, Agentic AI provides a unifying structure for autonomous and adaptive swarm operations while expanding the system attack surface compared to conventional AI systems. This paper proposes LLM-Centric Agentic AI for UAV Swarms (LAUS) and reviews key enabling technologies such as onboard and edge computing, 5G/6G connectivity, multimodal intelligence, and cybersecurity mechanisms, and analyzes threats such as Priority Manipulation Attacks (PMA) that can distort decision-making and degrade network performance. Finally, it identifies open research challenges, including hallucination-resistant reasoning, onboard LLM deployment under SWaP constraints, and standardized security benchmarks for perception-reasoning attacks in agentic UAV systems.
Chinese Translation
无人机（UAV）群在搜索与救援（SAR）和环境监测等应用中具有重要潜力，但其在现实世界中的部署受到缺乏情境感知、间歇性连接和显著网络安全风险的限制。自主人工智能（Agentic AI）代表了一种从独立的大型语言模型（LLM）向闭环认知架构的转变，这种架构集成了感知、记忆、推理/规划和行动，以实现自适应的、目标导向的群体行为。在这一框架内，自主人工智能为自主和自适应的群体操作提供了统一结构，同时与传统人工智能系统相比，扩大了系统的攻击面。本文提出了面向无人机群的以大型语言模型为中心的自主人工智能（LAUS），并回顾了关键的支持技术，如机载和边缘计算、5G/6G连接、多模态智能和网络安全机制，同时分析了可能扭曲决策和降低网络性能的威胁，例如优先级操控攻击（PMA）。最后，本文识别了开放的研究挑战，包括抗幻觉推理、在SWaP约束下的机载LLM部署，以及针对自主无人机系统中感知-推理攻击的标准化安全基准。
cs.RO / 7 / 2607.09764
OmniSCS: Omni Safety-Critical Scenario Synthesis for Autonomous Driving via a Fully Editable Driving World
OmniSCS：通过完全可编辑的驾驶世界合成自主驾驶的全安全关键场景
Abstract
The synthesis of safety-critical scenarios (SCS) and their evaluation through closed-loop simulations are crucial for developing robust autonomous driving systems. A key aspect of this process involves editing agent states in both appearance and trajectory levels within existing scenes. However, current methods struggle to preserve data fidelity after scene editing and fail to efficiently generate high-quality SCS through such modifications. To overcome these limitations, we propose OmniSCS, an innovative system that generates photorealistic SCS with high physical fidelity while enabling closed-loop testing in synthetic environments. OmniSCS comprises two key modules: 1) A Fully Editable Driving World Construction module that maintains high-fidelity agent appearance and background during scene editing via dual-strategy agent reconstruction and depth-refinement background reconstruction methods. 2) A SCS Synthesis module that facilitates object insertion and agent trajectory editing to synthesize diverse SCS while preserving data fidelity. Experiments on nuScenes, Waymo, and KITTI datasets show that OmniSCS outperforms state-of-the-art methods in edited scene fidelity. We further validate its ability to enhance autonomous driving algorithms and support real-time (13Hz) closed-loop testing. Overall, OmniSCS provides a safer, more effective, and cost-efficient solution for SCS optimization and testing in autonomous driving.
Chinese Translation
安全关键场景（SCS）的合成及其通过闭环仿真进行评估对于开发稳健的自主驾驶系统至关重要。该过程的一个关键方面是编辑现有场景中代理的状态，包括外观和轨迹层面。然而，当前的方法在场景编辑后难以保持数据的真实性，并且无法通过此类修改有效生成高质量的SCS。为了解决这些局限性，我们提出了OmniSCS，这是一种创新系统，能够生成具有高物理真实性的照片级SCS，同时支持在合成环境中进行闭环测试。OmniSCS包含两个关键模块：1）一个完全可编辑的驾驶世界构建模块，通过双策略代理重建和深度细化背景重建方法，在场景编辑过程中保持高保真度的代理外观和背景。2）一个SCS合成模块，便于对象插入和代理轨迹编辑，以合成多样化的SCS，同时保持数据的真实性。在nuScenes、Waymo和KITTI数据集上的实验表明，OmniSCS在编辑场景的真实性方面优于最先进的方法。我们进一步验证了其增强自主驾驶算法的能力，并支持实时（13Hz）闭环测试。总体而言，OmniSCS为自主驾驶中的SCS优化和测试提供了更安全、更有效和更具成本效益的解决方案。
cs.RO / 8 / 2607.09772
A Risk-Field Enhanced Closed-Loop Digital Twin Framework for Autonomous Driving Safety Validation
一种风险场增强的闭环数字双胞胎框架用于自动驾驶安全验证
Abstract
Autonomous driving systems require reliable safety validation before real-world deployment. However, large-scale road testing is costly, difffcult to reproduce, and inefffcient for exposing rare safety-critical scenarios. Conventional simulation improves repeatability, but an offfine simulator alone cannot continuously connect physical trafffc states, virtual reconstruction, algorithm evaluation, and scenario evolution. This paper proposes a risk-ffeld enhanced closed-loop digital twin framework for autonomous driving safety validation. The framework integrates physical data acquisition, data synchronization, virtual twin reconstruction, risk-aware scenario generation, autonomous driving algorithm evaluation, and safety analysis. A driving risk ffeld is introduced as a uniffed intermediate representation to describe obstacle, lane-departure, road-boundary, time-to-collision, and comfort-related risks around the ego vehicle. The risk ffeld ranks high-risk scenarios in the digital twin scenario library and provides dense safety guidance for reinforcement learning-based driving policies. A simulation-style evaluation protocol is designed to compare conventional reinforcement learning baselines, risk-penalty baselines, and the proposed risk-ffeld guided method. The study indicates that embedding explicit risk structure into digital twins can make autonomous driving validation more targeted, interpretable, and reusable, while its practical effectiveness remains bounded by model ffdelity, risk calibration, and sim-to-real transfer.
Chinese Translation
自动驾驶系统在实际部署之前需要可靠的安全验证。然而，大规模的道路测试成本高昂，难以重现，并且对于暴露罕见的安全关键场景效率低下。传统的仿真提高了可重复性，但单靠仿真器无法持续连接物理交通状态、虚拟重建、算法评估和场景演变。本文提出了一种风险场增强的闭环数字双胞胎框架，用于自动驾驶安全验证。该框架集成了物理数据采集、数据同步、虚拟双胞胎重建、风险感知场景生成、自动驾驶算法评估和安全分析。引入了驾驶风险场作为统一的中介表示，以描述自我车辆周围的障碍物、车道偏离、道路边界、碰撞时间和舒适度相关的风险。风险场在数字双胞胎场景库中对高风险场景进行排序，并为基于强化学习的驾驶策略提供密集的安全指导。设计了一种仿真风格的评估协议，以比较传统的强化学习基线、风险惩罚基线和所提出的风险场引导方法。研究表明，将明确的风险结构嵌入数字双胞胎可以使自动驾驶验证更加有针对性、可解释和可重用，同时其实际有效性仍受限于模型保真度、风险校准和仿真到现实的转移。
cs.RO / 9 / 2607.09776
Maximizing Human Efficiency in Large-Scale Robot Post-Training via VLAC-Cut Guided Pipeline
通过 VLAC-Cut 引导管道最大化大规模机器人后训练中的人类效率
Abstract
When adapting Vision Language Action (VLA) models to downstream tasks, multiple rounds of post training are required because a single round of data cannot resolve all issues, making continuous iterations necessary to progressively address the weaknesses exposed in previous rounds. In this report, we aim to maximize human efficiency during post-training, defined as the policy improvement and task throughput achieved per unit of human labor and time. We propose a human-efficient post-training pipeline that enables a small number of human operators to supervise multiple robots. The pipeline is built around a specialized division of labor: a trained Teleoperator focuses on high-value remote interventions and recovery demonstrations, while a Floor Operator monitors multiple robots, triggers takeovers, and performs physical resets. This role specialization reduces task switching, lowers operator training costs, and allows limited human labor to supervise more robot interaction across a larger fleet. To improve data utilization efficiency, we introduce VLAC-CUT as an automatic rollout curation tool. It segments autonomous robot trajectories into progress-making, idle, failure-inducing, and recovery portions, preserving useful segments while filtering harmful or uninformative ones. The curated rollout data are combined with Human-in-the-Loop data for the next post-training round. We validate the proposed pipeline on four real-world manipulation tasks. Across iterative post-training rounds, the final policies achieve 80\%--95\% success rates and improve task throughput by 1.7$\times$--4.2$\times$ over the base model. Under the same human-intervention budget, VLAC-CUT guided rollout reuse outperforms HITL-only training in both success rate and throughput.
Chinese Translation
在将视觉语言行动（VLA）模型适应于下游任务时，需要多轮后训练，因为单轮数据无法解决所有问题，因此需要持续迭代以逐步解决在之前轮次中暴露的弱点。在本报告中，我们旨在最大化后训练过程中的人类效率，定义为每单位人力和时间所实现的策略改进和任务吞吐量。我们提出了一种人类高效的后训练管道，使少量人类操作员能够监督多个机器人。该管道围绕专业化的劳动分工构建：经过培训的远程操作员专注于高价值的远程干预和恢复演示，而地面操作员则监控多个机器人，触发接管并执行物理重置。这种角色专业化减少了任务切换，降低了操作员培训成本，并允许有限的人力监督更大车队中的更多机器人交互。为了提高数据利用效率，我们引入了 VLAC-CUT 作为自动化回放策划工具。它将自主机器人轨迹分割为进展、闲置、导致失败和恢复的部分，保留有用的片段，同时过滤有害或无信息的部分。策划后的回放数据与人机协作（Human-in-the-Loop）数据结合用于下一轮后训练。我们在四个真实世界的操作任务上验证了所提出的管道。在迭代的后训练轮次中，最终策略的成功率达到 80 ext{%}--95 ext{%}，任务吞吐量比基础模型提高了 1.7$ imes$--4.2$ imes$。在相同的人类干预预算下，VLAC-CUT 引导的回放重用在成功率和吞吐量上均优于仅依赖人机协作的训练。
cs.RO / 10 / 2607.09792
A Comprehensive Survey and Systematic Real-World Evaluation of Embodied Vision-and-Language Navigation
对具身视觉与语言导航的全面调查及系统性现实世界评估
Abstract
Navigation is a fundamental capability of autonomous systems, yet most existing approaches rely on highly structured models and strong prior assumptions, limiting their robustness in open and uncertain real-world environments. Vision-and-Language Navigation (VLN) offers a promising direction by enabling robots to integrate natural language understanding with visual perception in a data-driven manner. Although VLN has attracted increasing research attention, systematic methodological taxonomy and real-world validation remain limited. This survey presents a comprehensive review of VLN research. Specifically, state-of-the-art methods are organized along two orthogonal dimensions: action paradigms, including hierarchical and monolithic frameworks, and model paradigms, including discriminative and generative approaches. A critical analysis of their respective strengths and limitations is provided. Additionally, we conduct a systematic real-world evaluation of representative VLN system configurations on a physical robotic platform. Experiments across ten diverse real-world scenes show a substantial performance gap between simulation and real-world deployment under the tested configurations: a representative monolithic RGB-only method achieves 61% success in simulation but drops to 22% in real-world deployment, while a hierarchical framework achieves a higher real-world success rate of 51%, suggesting stronger robustness in our evaluation setting. Finally, we highlight key challenges in perception, decision-making, and control that must be addressed in future research.
Chinese Translation
导航是自主系统的一项基本能力，但现有大多数方法依赖于高度结构化的模型和强假设，这限制了它们在开放和不确定的现实世界环境中的鲁棒性。视觉与语言导航（Vision-and-Language Navigation, VLN）通过使机器人能够以数据驱动的方式将自然语言理解与视觉感知相结合，提供了一种有前景的方向。尽管VLN吸引了越来越多的研究关注，但系统的方法分类和现实世界的验证仍然有限。本调查对VLN研究进行了全面回顾。具体而言，最先进的方法沿着两个正交维度进行组织：行动范式，包括分层和单体框架，以及模型范式，包括判别和生成方法。提供了对各自优缺点的批判性分析。此外，我们在一个物理机器人平台上对代表性的VLN系统配置进行了系统的现实世界评估。在十个不同的现实世界场景中的实验显示，在测试的配置下，仿真与现实世界部署之间存在显著的性能差距：一个代表性的单体RGB-only方法在仿真中的成功率为61%，但在现实世界部署中降至22%；而一个分层框架在现实世界中的成功率更高，达到了51%，这表明在我们的评估环境中具有更强的鲁棒性。最后，我们强调了在感知、决策和控制方面的关键挑战，这些挑战必须在未来的研究中得到解决。
cs.RO / 11 / 2607.09807
A Biomimetic Myoelectric Tentacle Prosthesis with Sensorless Object Detection and Vibrotactile Feedback
一种生物仿生肌电触手假肢，具备无传感器物体检测和振动触觉反馈
Abstract
This paper presents the design and evaluation of a myoelectric tentacle-shaped prosthesis integrating electromyographic (EMG) control, sensorless object detection, and vibrotactile feedback. The objective was to develop a responsive and intuitive assistive device that adapts to various object shapes while providing sensory feedback to the user. The system relies on EMG signals to control the motion of a flexible, biomimetic structure whose curling geometry follows a logarithmic spiral, enabling it to coil around objects. To ensure stable control, the EMG signal is normalized and filtered, and a threshold-based method identifies user intention. Object contact is detected through a slope-based analysis of motor current, eliminating the need for external sensors, and a haptic feedback strategy based on cumulative vibrotactile stimulation conveys spatial information about the tentacle's configuration. The system was evaluated through quantitative and qualitative tests. The results demonstrate a low response time (77 ms on average), enabling smooth real-time interaction; an object-detection success rate above 90%, confirming robustness despite EMG variability; and an effective haptic feedback strategy that allowed users to reliably identify the folding zone of the tentacle. The proposed biomimetic design promotes further investigation of expressive artificial limbs by prioritizing expressive functionality over adherence to a predefined, anthropomorphic form factor.
Chinese Translation
本文提出了一种集成肌电控制、无传感器物体检测和振动触觉反馈的肌电触手形状假肢的设计与评估。其目标是开发一种响应迅速且直观的辅助设备，能够适应各种物体形状，同时为用户提供感官反馈。该系统依赖肌电信号控制一种灵活的生物仿生结构的运动，其卷曲几何形状遵循对数螺旋，使其能够缠绕物体。为了确保稳定控制，肌电信号经过归一化和滤波处理，并通过基于阈值的方法识别用户意图。物体接触通过对电机电流的斜率分析进行检测，消除了对外部传感器的需求，而基于累积振动刺激的触觉反馈策略则传达了触手配置的空间信息。该系统通过定量和定性测试进行了评估。结果表明，响应时间低（平均77毫秒），实现了流畅的实时交互；物体检测成功率超过90%，确认了在肌电信号变异情况下的鲁棒性；以及有效的触觉反馈策略使用户能够可靠地识别触手的折叠区域。所提出的生物仿生设计促进了对表现性人工肢体的进一步研究，优先考虑表现功能而非遵循预定义的人形外形因素。
cs.RO / 12 / 2607.09815
RASR: Range-Aware Scale Recovery for Metric UAV Navigation
RASR：基于距离感知的度量无人机导航尺度恢复
Abstract
Under Global Navigation Satellite System (GNSS) denial, a UAV controller still needs a distance and heading command it can execute, making accurate metric last-meter navigation essential. Dense pair-geometry foundation models transfer relative structure well, yet the distance scale of their raw metric outputs remains poorly calibrated. Under the relative error metric of PairUAV, correcting only the average scale can still leave costly, distance-dependent residuals near the goal. To address this scale mismatch, Range-Aware Scale Recovery (RASR) separates a transferable scale-recovery core from a protocol-specific calibration module in a per-pair system fixed at inference. The core compresses frozen Matching And Stereo 3D Reconstruction (MASt3R)-style geometry into a compact descriptor and uses global calibration to recover the dominant metric signal. Range-bucket residual correction and command-grid alignment stay inside the calibration module, so they match the command format and evaluation protocol of PairUAV. On the UAVs in Multimedia 2026 PairUAV online evaluation, RASR reaches a total score of 0.003189. Under the PairUAV protocol, frozen pair geometry thus yields stable per-pair distance and heading estimates, while every protocol-specific adjustment stays confined to a calibration module fixed before inference. Code and materials are available at https://github.com/lht-research/rasr-pairuav.
Chinese Translation
在全球导航卫星系统（GNSS）失效的情况下，无人机控制器仍然需要可以执行的距离和航向指令，因此准确的度量最后一米导航至关重要。密集的配对几何基础模型能够很好地传递相对结构，但其原始度量输出的距离尺度仍然校准不佳。在PairUAV的相对误差度量下，仅纠正平均尺度仍可能导致在目标附近产生昂贵的、依赖于距离的残差。为了解决这一尺度不匹配问题，基于距离感知的尺度恢复（RASR）将可转移的尺度恢复核心与特定协议的校准模块分离，在推理时固定为每对系统。该核心将冻结的匹配与立体三维重建（MASt3R）风格几何压缩为紧凑描述符，并利用全局校准恢复主导的度量信号。范围桶残差校正和指令网格对齐保持在校准模块内，以便与PairUAV的指令格式和评估协议相匹配。在2026年多媒体PairUAV在线评估中的无人机上，RASR的总得分为0.003189。在PairUAV协议下，冻结的配对几何因此能够产生稳定的每对距离和航向估计，而每个特定协议的调整则限制在推理前固定的校准模块内。代码和材料可在 https://github.com/lht-research/rasr-pairuav 获取。
cs.RO / 13 / 2607.09818
TS-Mask VLA: 2D Temporal-Spatial Masking for Vision-Language-Action Model with Effective Bridging
TS-Mask VLA：具有有效桥接的视觉-语言-动作模型的二维时空掩蔽
Abstract
Vision-language-action (VLA) models aim to understand natural-language instructions and visual observations, and to generate and execute corresponding actions as embodied agents. Recently, autoregressive token-based action generation has driven the development of many representative VLA models. However, this paradigm often reduces action generation to next-token prediction, thereby lacking explicit modeling of the spatiotemporal structure of action sequences and the disentanglement between vision-language representations and actions, which can limit performance in long-horizon and complex scenarios. In this paper, we propose TS-Mask VLA, a vision-language-action framework for robot manipulation. TS-Mask VLA is built upon two key designs: (1) a Discrete Diffusion Action Expert equipped with a Bridge Attention conditioning bridge, which enables multi-layer conditioning from the VLM and facilitates more accurate and stable action generation; and (2) a temporal-spatial 2D masking strategy for discrete action tokens that strengthens the model's understanding of cross-time dependencies and inter-dimensional coupling, leading to more structurally consistent action sequences. We conduct extensive experiments on simulation benchmarks and real-world tasks. On LIBERO, TS-Mask VLA achieves a 95.7 percent average success rate with only 0.5B parameters, outperforming significantly larger models. On CALVIN, it attains the best average sequence length of 4.19 and strong long-horizon performance. Comprehensive analyses and ablations further validate the effectiveness of our design.
Chinese Translation
视觉-语言-动作（VLA）模型旨在理解自然语言指令和视觉观察，并生成和执行相应的动作，作为具身代理。最近，自回归基于令牌的动作生成推动了许多代表性 VLA 模型的发展。然而，这一范式往往将动作生成简化为下一个令牌的预测，从而缺乏对动作序列时空结构的明确建模，以及视觉-语言表示与动作之间的解耦，这可能限制了在长时间跨度和复杂场景中的性能。在本文中，我们提出了 TS-Mask VLA，这是一种用于机器人操作的视觉-语言-动作框架。TS-Mask VLA 基于两个关键设计构建：（1）配备桥接注意力条件桥的离散扩散动作专家，能够实现来自 VLM 的多层条件化，并促进更准确和稳定的动作生成；（2）针对离散动作令牌的二维时空掩蔽策略，增强了模型对跨时间依赖性和维度间耦合的理解，从而导致更结构一致的动作序列。我们在仿真基准和现实任务上进行了广泛的实验。在 LIBERO 上，TS-Mask VLA 以仅 0.5B 参数实现了 95.7% 的平均成功率，显著优于更大规模的模型。在 CALVIN 上，它达到了最佳的平均序列长度 4.19，并展现出强大的长时间跨度性能。全面的分析和消融实验进一步验证了我们设计的有效性。
cs.RO / 14 / 2607.09825
More Structure, Not More Capacity: Object-Centric Representations for Visuomotor Imitation Learning
更多结构，而非更多容量：面向物体的视觉运动模仿学习表示
Abstract
Robotic manipulation policies rely on pre-trained vision models that give either a global scene embedding or a dense patch grid. Both mix task-relevant and task-irrelevant features. Object-centric slot representations are a structured alternative: they group features into a few per-object slots. We test what this structure buys on ManiSkill3 PickCube-v1, with a frozen encoder and a held-out-seed evaluation. Holding the policy, goal token, rendering, and calibration fixed and changing only the encoder, a frozen object-centric SPOT representation (DINO ViT-B/16 + Slot Attention) reaches 55.0$\pm$2.9% success, 22.4% above a dense DINO global-feature baseline (32.6 $\pm$ 1.5%), with the same trainable policy and no encoder fine-tuning. More tokens alone do not help: a dense patch grid with 16x the tokens performs no better than the global feature. Adding an explicit 2D spatial goal and native-resolution rendering raises the full system to 68.7$\pm$4.2%, just below a privileged 3D-oracle upper bound (71.7$\pm$4.1%). An automated kinematic failure taxonomy then separates spatial-precision (Near-Miss) failures from object-tracking (No-Grasp) failures: spatial grounding reduces Near-Miss while leaving No- Grasp unchanged. The same taxonomy transfers to the harder StackCube-v1 and points to occlusion as the main bottleneck.
Chinese Translation
机器人操作策略依赖于预训练的视觉模型，这些模型提供全局场景嵌入或密集的补丁网格。这两者都混合了与任务相关和无关的特征。面向物体的槽表示是一种结构化的替代方案：它将特征分组为少数每个物体的槽。我们在 ManiSkill3 PickCube-v1 上测试这种结构的效果，使用固定的编码器和保留的种子评估。保持策略、目标标记、渲染和校准不变，仅更改编码器，一个冻结的面向物体的 SPOT 表示（DINO ViT-B/16 + 槽注意力）达到了 55.0$ ext{±}$2.9% 的成功率，比密集的 DINO 全局特征基线（32.6$ ext{±}$1.5%）高出 22.4%，且使用相同的可训练策略且没有编码器微调。仅增加标记并没有帮助：一个密集的补丁网格，标记数量是原来的 16 倍，表现并不优于全局特征。添加一个明确的 2D 空间目标和原生分辨率渲染将整个系统的成功率提升至 68.7$ ext{±}$4.2%，仅低于特权的 3D 神谕上限（71.7$ ext{±}$4.1%）。随后，自动化的运动学失败分类法将空间精度（近失误）失败与物体跟踪（未抓取）失败分开：空间定位减少了近失误，而未抓取保持不变。同样的分类法可以转移到更难的 StackCube-v1，并指出遮挡是主要瓶颈。
cs.RO / 15 / 2607.09866
Robo-ValueRL: Reliable Value Estimation for Offline-to-Online Reinforcement Learning
Robo-ValueRL：离线到在线强化学习的可靠价值估计
Abstract
Offline-to-online reinforcement learning is promising for generalizable robotic manipulation, yet its full-stack complexity obscures reproduction and diagnosis. Within such systems, value estimation plays a central role in prioritizing heterogeneous data for policy improvement. Despite its importance, the central question remains underexplored: how value-function reliability shapes policy optimization in offline-to-online reinforcement learning. To answer this question, we propose Robo-ValueRL, a unified framework that enables reliable value estimation and systematically traces its downstream effects on policy pretraining and online improvement. Concretely, Robo-ValueRL learns a history-conditioned value estimator and evaluates its reliability through global-progress and local-preference metrics. These resulting value estimates are propagated into quality-conditioned consistency-policy pretraining and a residual adaptation module on online rollouts, providing a unified testbed for analyzing how value reliability shapes downstream policy performance. Across 240 hours of offline demonstrations and over 3,000 online rollout trajectories, our extensive experiments show that downstream performance is strongly associated with value reliability. Reliable value functions provide better action-quality estimates, allowing value-guided offline RL to scale more effectively than quality-agnostic behavior cloning, and stabilize online improvement by prioritizing high-quality rollout data. Integrating reliable value guidance through offline pretraining with online improvement, our system achieves 86% success on millimeter-level precise chip insertion and 84% on generalizable block disassembly. We hope these findings highlight the importance of value-guided data utilization for effective policy improvement from heterogeneous robotic experience.
Chinese Translation
离线到在线强化学习在可推广的机器人操作中展现出良好的前景，但其全栈复杂性使得重现和诊断变得困难。在此类系统中，价值估计在优先考虑异构数据以改善策略方面发挥着核心作用。尽管其重要性不言而喻，但一个核心问题仍未得到充分探讨：价值函数的可靠性如何影响离线到在线强化学习中的策略优化。为了解答这一问题，我们提出了Robo-ValueRL，一个统一框架，能够实现可靠的价值估计，并系统地追踪其对策略预训练和在线改进的下游影响。具体而言，Robo-ValueRL学习一个基于历史的价值估计器，并通过全局进展和局部偏好指标评估其可靠性。这些生成的价值估计被传播到基于质量的连续性策略预训练和在线回放中的残差适应模块，为分析价值可靠性如何影响下游策略性能提供了一个统一的测试平台。在240小时的离线演示和超过3000条在线回放轨迹的广泛实验中，我们的研究表明，下游性能与价值可靠性密切相关。可靠的价值函数提供了更好的行动质量估计，使得基于价值的离线强化学习能够比无质量指导的行为克隆更有效地扩展，并通过优先考虑高质量的回放数据来稳定在线改进。通过将可靠的价值指导与在线改进相结合，我们的系统在毫米级精确芯片插入任务中取得了86%的成功率，在可推广的块拆卸任务中取得了84%的成功率。我们希望这些发现能够强调价值指导的数据利用在异构机器人经验中对有效策略改进的重要性。
cs.RO / 16 / 2607.09911
Diffusion for Long-Horizon Multi-Robot Path Planning in Human-Shared Environments
人类共享环境中长时间跨度多机器人路径规划的扩散方法
Abstract
Multi-robot path planning in human-shared environments requires a delicate balance between robust inter-robot coordination and socially aware behavior. While diffusion models excel at generating predictable, human-like paths, existing generative planners are often restricted to paths of fixed duration and high computational latency, limiting their adaptability to varying goal distances and hindering real-time deployment. We present Multi-Robot Rolling Diffusion (MRRD), a novel framework that enables real-time, long-horizon navigation for large robot teams through dense crowds. MRRD combines a rolling-horizon scheme to accommodate the limited prediction horizon of human motion, parallelized diffusion inference for scalable generation of human-like paths, and a conflict-based-search mechanism for resolving inter-robot collisions. It further incorporates urgency-based temporal conditioning to generate paths with varying speeds and employs differentiated guidance terms to maximize both social awareness around humans and efficient coordination between robots. Experimental results in crowded environments demonstrate that MRRD successfully scales to 15 robots in real-time, significantly outperforming existing baselines in both safety and mission success rates.
Chinese Translation
在人类共享环境中，多机器人路径规划需要在稳健的机器人间协调与社会意识行为之间找到微妙的平衡。尽管扩散模型在生成可预测的类人路径方面表现出色，但现有的生成规划器通常受到固定持续时间和高计算延迟的限制，这限制了它们对不同目标距离的适应性，并妨碍了实时部署。我们提出了多机器人滚动扩散（Multi-Robot Rolling Diffusion, MRRD），这是一个新颖的框架，能够实现大型机器人团队在密集人群中实时、长时间跨度的导航。MRRD结合了滚动时间框架，以适应人类运动的有限预测范围，并通过并行化的扩散推理实现可扩展的类人路径生成，同时采用基于冲突的搜索机制来解决机器人间的碰撞问题。它进一步结合了基于紧迫性的时间条件，以生成具有不同速度的路径，并采用差异化的引导项，以最大化对人类的社会意识和机器人之间的高效协调。在拥挤环境中的实验结果表明，MRRD能够成功扩展至15个机器人实时运行，在安全性和任务成功率方面显著优于现有基准。
cs.RO / 17 / 2607.09947
A Numerically-Robust ROS 2 Port of iG-LIO: Diagnosing and Fixing Toolchain-Induced Failures in Incremental GICP LiDAR-Inertial Odometry
iG-LIO的数值稳健ROS 2移植：诊断和修复增量GICP激光雷达-惯性里程计中的工具链引发的故障
Abstract
iG-LIO is a tightly-coupled LiDAR-inertial odometry system fusing generalized-ICP and point-to-plane constraints in an iterated error-state Kalman filter over an incremental voxel map. We report an open-source ROS 2 Jazzy port of the original ROS 1 implementation and, more importantly, the diagnosis of environment-induced numerical failures that appear only after the port: a mechanically faithful migration -- estimation mathematics left unchanged -- compiled and ran, yet diverged with NaN internal values. Both causes trace to the modern ROS 2 toolchain, not the algorithm: a Quality-of-Service (QoS) mismatch that silently drops and reorders IMU samples, and an uninitialized parallel-reduce accumulator arising from the oneTBB + Eigen combination shipped with current distributions. We further correct Ouster point-field parsing to ensure correct point cloud undistortion with newer Ouster revisions, add Velodyne Velarray M1600 support, provide both a compile-time-gated Livox CustomMsg path and a driver-free path for Livox sensors publishing standard PointCloud2 (e.g. Mid-360), and expose the runtime via YAML. The result has been validated in an Ouster OS0 Rev7, an Ouster OS1 Rev 7, and a Livox MID-360. This report is a citable reference for the port itself, not a claim on the underlying algorithm [1]. The ROS 2 port of iG-LIO described in this document can be found at https://github.com/Forestry-Robotics-UC/ig_lio/tree/ros2-jazzy.
Chinese Translation
iG-LIO是一个紧耦合的激光雷达-惯性里程计系统，通过在增量体素地图上使用迭代误差状态卡尔曼滤波器融合广义ICP和点到平面约束。我们报告了原始ROS 1实现的开源ROS 2 Jazzy移植，更重要的是，诊断了仅在移植后出现的环境引发的数值故障：一个机械上忠实的迁移——估计数学保持不变——编译并运行，但内部值却出现了NaN的发散。这两个原因都追溯到现代ROS 2工具链，而非算法：一个质量服务（QoS）不匹配，默默地丢弃和重排序IMU样本，以及一个未初始化的并行归约累加器，源于当前发行版中包含的oneTBB + Eigen组合。我们进一步修正了Ouster点场解析，以确保与更新的Ouster版本正确的点云去畸变，增加了对Velodyne Velarray M1600的支持，提供了一个编译时受限的Livox CustomMsg路径和一个无驱动路径，以支持发布标准PointCloud2（例如Mid-360）的Livox传感器，并通过YAML公开运行时。该结果已在Ouster OS0 Rev7、Ouster OS1 Rev 7和Livox MID-360上进行了验证。本报告是对该移植的可引用参考，而不是对基础算法的声明[1]。本文档中描述的iG-LIO的ROS 2移植可以在https://github.com/Forestry-Robotics-UC/ig_lio/tree/ros2-jazzy找到。
cs.RO / 18 / 2607.09959
SEAMLiS: Visibility-Aware Safety for Perception-Limited Multi-Robot Exploration
SEAMLiS：面向感知受限多机器人探索的可见性意识安全性
Abstract
Autonomous exploration in unknown environments is typically driven by informative frontiers, viewpoints, or trajectories, while local safety controllers avoid obstacles represented in the current map. Under finite sensing range and limited field of view, this separation can be unsafe: an exploration stack may plan optimistically through unobserved space and steer the sensor toward information gain rather than along the direction of motion, causing hidden obstacles to be detected too late for bounded-actuation avoidance. This paper presents SEAMLiS (Safe Exploration for Autonomous Multi-Robot Systems Under Limited Sensing), a modular execution-layer safety framework for decentralized multi-robot exploration. SEAMLiS preserves the upstream exploration stack, including the goal allocator and local planner, and enforces safety at the execution layer through perception-aware attitude and positional filters. A gatekeeper-based attitude filter switches between a visibility-promoting yaw policy and a velocity-tracking backup policy to preserve visibility of the critical known-free/unknown boundary with sufficient braking margin. A Control Barrier Function (CBF)-based positional filter then avoids known obstacles, newly detected obstacles, and other robots. We provide sufficient collision-avoidance conditions and validate the framework in randomized simulation, Isaac Sim, and Crazyflie hardware experiments. Results show collision-free exploration across tested single- and multi-robot settings while retaining much of the efficiency of visibility-promoting yaw control.
Chinese Translation
在未知环境中的自主探索通常由信息前沿、视点或轨迹驱动，而局部安全控制器则避免当前地图中表示的障碍物。在有限的感知范围和有限的视野下，这种分离可能是不安全的：探索堆栈可能在未观察到的空间中乐观规划，并将传感器引导至信息获取方向，而不是沿着运动方向，导致隐藏障碍物的检测过晚，无法进行有限动作规避。本文提出了SEAMLiS（在有限感知下的自主多机器人系统安全探索），这是一个用于去中心化多机器人探索的模块化执行层安全框架。SEAMLiS保留了上游探索堆栈，包括目标分配器和局部规划器，并通过感知意识的姿态和位置滤波器在执行层强制执行安全性。基于门控的姿态滤波器在促进可见性的偏航策略和跟踪速度的备份策略之间切换，以保持关键已知-自由/未知边界的可见性，并留有足够的制动余量。基于控制障碍函数（Control Barrier Function, CBF）的位置信息滤波器则避免已知障碍物、新检测到的障碍物和其他机器人。我们提供了充分的碰撞避免条件，并在随机仿真、Isaac Sim和Crazyflie硬件实验中验证了该框架。结果显示，在测试的单机器人和多机器人设置中实现了无碰撞探索，同时保留了促进可见性的偏航控制的高效性。
cs.RO / 19 / 2607.09962
Task Planning for Mobile Manipulation in Retail Stores using Foundation Models with Iterative Re-planning
基于基础模型的零售店移动操作任务规划与迭代重新规划
Abstract
Automation in industries such as retail, warehousing and logistics presents opportunities for greater throughput, cost reduction and mitigation of disruptions from labour shortages. Previously, such efforts have focused on back-room operations involving packing and sorting in relatively structured environments. With advances in robotic mobile manipulation hardware and foundation models, automation can now be applied to more variable and human-centric environments such as retail store shelves. In this work, we present a task-planning approach using Large Language Models (LLMs) and Vision-Language Models (VLMs) to address the restocking problem in retail scenarios such as supermarkets. We demonstrate this system on a custom omnidirectional mobile manipulation platform, with user-driven prompts and a feedback-based iterative re-planning approach for error correction. The end-to-end system is validated in a PyBullet simulation environment for pick-and-place tasks.
Chinese Translation
在零售、仓储和物流等行业的自动化为提高产出、降低成本以及缓解劳动力短缺带来的干扰提供了机会。此前，这些努力主要集中在涉及打包和分类的后场操作，这些环境相对结构化。随着机器人移动操作硬件和基础模型的进步，自动化现在可以应用于更具变数和以人为中心的环境，例如零售店货架。在本研究中，我们提出了一种使用大型语言模型（Large Language Models, LLMs）和视觉-语言模型（Vision-Language Models, VLMs）的任务规划方法，以解决超市等零售场景中的补货问题。我们在一个定制的全向移动操作平台上演示了该系统，采用用户驱动的提示和基于反馈的迭代重新规划方法进行错误修正。该端到端系统在PyBullet仿真环境中进行了验证，以执行抓取和放置任务。
cs.RO / 20 / 2607.09968
Chalito: An Extensible Library for Filtering-Based State Estimation in Quadruped Robots
Chalito：一个可扩展的四足机器人基于滤波的状态估计库
Abstract
State estimation is essential for quadruped robots, enabling robust locomotion, navigation, and control. While many estimators have been proposed in the literature, existing implementations are often tied to specific robots or software stacks, making fair comparisons difficult. This lack of a general-purpose benchmarking framework hinders reproducibility and slows down algorithmic innovation. In this paper, we introduce Chalito, an extensible MATLAB/Python library for benchmarking filter-based state estimation algorithms in quadruped robots. Chalito imports robot models directly from URDF, supports multiple filtering approaches, and is designed to be easily extended with new methods. The framework runs on both simulated and real datasets, enabling systematic evaluation across robots and filters. To the best of our knowledge, this is the first open-source library exclusively dedicated to benchmarking filtering algorithms for quadruped robots.
Chinese Translation
状态估计对于四足机器人至关重要，它使得机器人能够实现稳健的运动、导航和控制。尽管文献中提出了许多估计器，但现有的实现往往与特定的机器人或软件栈紧密相关，这使得公平比较变得困难。这种缺乏通用基准框架的情况阻碍了可重复性，并减缓了算法创新的步伐。本文介绍了Chalito，一个可扩展的MATLAB/Python库，用于基于滤波的四足机器人状态估计算法的基准测试。Chalito能够直接从URDF导入机器人模型，支持多种滤波方法，并设计为易于扩展新方法。该框架可以在模拟和真实数据集上运行，能够在不同的机器人和滤波器之间进行系统评估。根据我们所知，这是第一个专门用于基准测试四足机器人滤波算法的开源库。
cs.RO / 21 / 2607.10000
PinFT: Miniature 5-Axis Force/Torque Sensor Embeddable to Tweezer-like Tool
PinFT：可嵌入镊子类工具的小型五轴力/扭矩传感器
Abstract
We present PinFT, a miniature five-axis capacitive force/torque sensor designed for direct tip-level integration into tweezer-like tools. The sensor employs a compact three-PCB stack with segmented plated through-hole electrodes and a silicone elastomer dielectric, enabling five-degree-of-freedom force and torque sensing ($F_x$, $F_y$, $F_z$, $T_x$, $T_y$) through displacement of a central 2\,mm-diameter stainless steel pin. The fabricated prototype was calibrated using a higher-order polynomial mapping, yielding mean absolute errors of approximately 0.23\,N for forces and 2.5\,mN$\cdot$m for torques, with coefficients of determination ($R^2$) exceeding 0.97 across all axes. To demonstrate practical utility, a 3D-printed tweezer integrating PinFT sensors at both tips was mounted on a parallel-jaw gripper and evaluated across three representative manipulation tasks: grasping a sub-millimeter SMD capacitor, pulling a simulated hair from a silicone substrate, and tearing a compliant silicone specimen. In all cases, per-tip force sensing reliably captured characteristic force signatures that distinguish successful manipulation from failure events -- including slip and object ejection -- using gradient-based features derived from internal grasp force and net interaction force. These results demonstrate that direct, per-tip force sensing enables standard parallel-jaw grippers to monitor and interpret fine manipulation tasks performed through a handheld tweezer.
Chinese Translation
我们提出了PinFT，一种小型五轴电容式力/扭矩传感器，旨在直接集成到镊子类工具的尖端。该传感器采用紧凑的三层PCB堆叠，配备分段的镀孔电极和硅橡胶介电材料，能够通过一个直径为2毫米的不锈钢针的位移实现五自由度的力和扭矩感测（$F_x$, $F_y$, $F_z$, $T_x$, $T_y$）。所制造的原型通过高阶多项式映射进行校准，力的平均绝对误差约为0.23 N，扭矩的平均绝对误差为2.5 mN·m，所有轴的决定系数（$R^2$）均超过0.97。为了展示实际应用，我们将集成PinFT传感器的3D打印镊子安装在平行夹爪上，并在三个代表性的操作任务中进行了评估：抓取亚毫米级的表面贴装电容器，从硅基底上拉动模拟头发，以及撕裂柔性硅样品。在所有情况下，每个尖端的力感测可靠地捕捉到了特征力信号，这些信号能够区分成功操作与失败事件——包括滑动和物体弹出——使用从内部抓取力和净交互力中提取的基于梯度的特征。这些结果表明，直接的每个尖端力感测使标准的平行夹爪能够监测和解释通过手持镊子执行的精细操作任务。
cs.RO / 22 / 2607.10014
Runtime Safety Filtering for Learned Small UAS Separation Policies under GNSS Degradation
在GNSS退化下对学习的小型无人机分离策略进行运行时安全过滤
Abstract
Learning-based separation assurance for small Unmanned Aircraft Systems (sUAS) achieves near-zero collision rates in simulation, but assumes accurate position and velocity information from Global Navigation Satellite Systems (GNSS). This assumption fails in urban environments, where multipath propagation, signal blockage, and intentional interference degrade navigation integrity. This raises a fundamental architectural question for deploying learned separation policies under GNSS degradation: should runtime safety mechanisms filter the policy's actions or its observations? This work evaluates both approaches for multi-agent sUAS separation under adversarial GNSS degradation. Both architectures first estimate a worst-case traffic state consistent with bounded observation uncertainty, then diverge: action filtering constrains policy outputs via discrete-time control barrier functions evaluated at the worst-case state, while observation filtering presents the worst-case state directly to the policy as corrected input. Experimental results show that action filtering provides negligible safety improvement, while observation filtering reduces near mid-air collisions by 90% and remains robust to the barrier function's tradeoff between separation distance and closing rate. These results suggest that, for policies with learned safety behaviors, preserving the policy's decision authority outperforms overriding its actions with hand-designed constraints.
Chinese Translation
基于学习的小型无人机系统（sUAS）的分离保障在模拟中实现了接近零的碰撞率，但假设全球导航卫星系统（GNSS）提供准确的位置和速度信息。这一假设在城市环境中失效，因为多径传播、信号阻塞和故意干扰会降低导航的完整性。这引发了一个关于在GNSS退化下部署学习分离策略的基本架构问题：运行时安全机制应该过滤策略的动作还是其观测？本研究评估了在对抗性GNSS退化下，针对多智能体sUAS分离的这两种方法。两种架构首先估计与有界观测不确定性一致的最坏情况交通状态，然后分歧：动作过滤通过在最坏情况状态下评估的离散时间控制障碍函数来约束策略输出，而观测过滤则将最坏情况状态直接作为修正输入呈现给策略。实验结果表明，动作过滤提供的安全改进微乎其微，而观测过滤将近空中碰撞减少了90%，并且在分离距离与闭合速率之间的障碍函数权衡中保持稳健。这些结果表明，对于具有学习安全行为的策略，保留策略的决策权优于用手工设计的约束来覆盖其动作。
cs.RO / 23 / 2607.10037
Plug-and-Play Reweighting for Resilient Collaborative Decision-Making in Connected Autonomous Driving
即插即用重加权用于连接自主驾驶中的弹性协作决策
Abstract
Collaborative decision-making is a fundamental capability in multi-robot systems, such as connected autonomous vehicles. However, perceptual noise and adversarial attacks in collaborators can severely affect decision reliability. Overall, existing methods typically rely on retraining with attack-specific defenses or on restrictive perturbation assumptions to improve resilience, which limits their practicality. In this paper, we propose a novel Resilient Collaborative Decision-Making (RCDM) framework that consists of an attention-based encoder for extracting individual robot perceptual embeddings and an attention-based decoder for fusing collaborator perceptions and making decisions. To improve resilience to corrupted observations, we design a novel plug-and-play reweighting module that down-weights the influence of corrupted inputs by analyzing the consistency of neighborhood points relative to the local structure and assigning smaller weights to points that deviate strongly from the local median. This module can be seamlessly integrated into attention-based collaborative decision-making without requiring additional training. We evaluate our method in high-fidelity simulations, considering perceptual noise and five types of attacks across diverse accident-prone scenarios. Experimental results demonstrate that our approach consistently outperforms existing methods by up to 26% and achieves state-of-the-art resilient performance.
Chinese Translation
协作决策是多机器人系统（如连接自主车辆）中的一项基本能力。然而，感知噪声和协作者的对抗攻击可能严重影响决策的可靠性。总体而言，现有方法通常依赖于针对特定攻击的防御进行重新训练，或依赖于限制性扰动假设来提高弹性，这限制了它们的实用性。本文提出了一种新颖的弹性协作决策（Resilient Collaborative Decision-Making, RCDM）框架，该框架包括一个基于注意力的编码器，用于提取单个机器人感知嵌入，以及一个基于注意力的解码器，用于融合协作者的感知并做出决策。为了提高对损坏观测的弹性，我们设计了一种新颖的即插即用重加权模块，通过分析邻域点相对于局部结构的一致性来降低损坏输入的影响，并对与局部中位数偏离较大的点分配较小的权重。该模块可以无缝集成到基于注意力的协作决策中，而无需额外的训练。我们在高保真模拟中评估了我们的方法，考虑了感知噪声和五种类型的攻击，涵盖了多种事故易发场景。实验结果表明，我们的方法在性能上始终优于现有方法，提升幅度可达26%，并实现了最先进的弹性表现。
cs.RO / 24 / 2607.10099
Direct Rotor Thrust Sensing and Feedback Control for Disturbance Rejection of Multirotors Using Load-cells
基于负载传感器的多旋翼飞行器直接转子推力感知与反馈控制以抑制干扰
Abstract
Gust disturbances, dynamic vertical inflow and ground effect are key adverse aerodynamic phenomena that induce variations in the forces acting on a multirotor and complicate its flight control. Miniature rotorcraft typically rely on simplified modelling of such effects to compute adjustments in thrust to counteract these forces. In the most basic case, disturbance force estimations are derived from the aircraft's motion and the generated thrust is assumed to exactly match that requested by the controller. However, such systems rely on the aircraft's trajectory to be affected before disturbances can be sensed and compensated. Numerous approaches presented over the last 15-20 years aim to reject external disturbances more quickly, but challenges remain. This paper presents a new approach in this category by measuring the instantaneous force of the rotors directly at the point of generation using load-cells and implementing high-speed control to accurately track the desired thrust. Measurements from load-cells were previously considered too noisy to provide meaningful input, but the experiments presented in the paper using purpose-built hardware from low-cost commodity components in single- and dual rotor see-saw models and a flying aircraft demonstrate both the feasibility and the effectiveness of the approach in the presence of complex aerodynamic phenomena.
Chinese Translation
气流干扰、动态垂直流入和地面效应是导致多旋翼飞行器所受力变化的关键不利气动现象，这些现象使飞行控制变得复杂。微型旋翼机通常依赖于对这些效应的简化建模，以计算推力的调整以抵消这些力。在最基本的情况下，干扰力的估计是基于飞行器的运动，而生成的推力被假设与控制器请求的推力完全匹配。然而，这种系统依赖于飞行器的轨迹在干扰被感知和补偿之前受到影响。在过去15-20年中，许多方法旨在更快地抑制外部干扰，但仍然存在挑战。本文提出了一种新方法，通过使用负载传感器直接在转子产生点测量瞬时力，并实施高速控制以准确跟踪所需推力。负载传感器的测量之前被认为噪声过大，无法提供有意义的输入，但本文中使用低成本商品组件构建的专用硬件在单旋翼和双旋翼跷跷板模型以及飞行器上的实验展示了该方法在复杂气动现象下的可行性和有效性。
cs.RO / 25 / 2607.10132
TAC-LOCO: Unified Whole-Body Control for Quadrupedal TACtile-Informed LOCO-Manipulation
TAC-LOCO：四足机器人触觉信息驱动的统一全身控制
Abstract
Dynamic loco-manipulation requires legged robots to coordinate whole-body motion while maintaining stable physical interaction with grasped objects under uncertain external forces. While tactile sensing has been widely studied for robotic manipulation, its role in dynamic whole-body control remains largely unexplored. Existing works without tactile feedback commonly grasp firmly rather than regulate the grasp according to the interaction. We propose TAC-LOCO, a tactile-augmented unified reinforcement learning framework that encodes tactile array observations from compliant grippers into a compact latent representation and joins it with proprioception for unified control of the legs, arm, and gripper. With effective grasp stability reward design, the policy learns to simultaneously track body velocity and end-effector trajectories, moderate grasp force, and prevent object slip under both gradual load changes and sudden release events. We deploy the policy zero-shot on a Unitree Go2 with an Interbotix WidowX 250 arm and tactile gripper, demonstrating dynamic tactile-informed loco-manipulation under varying external interactions, achieving a 47% reduction in grasping force and an object drop rate of less than 1%.
Chinese Translation
动态的运动操控要求腿式机器人在与被抓取物体保持稳定物理交互的同时，协调全身运动，以应对不确定的外部力量。尽管触觉传感在机器人操控中得到了广泛研究，但其在动态全身控制中的作用仍然未得到充分探索。现有的没有触觉反馈的工作通常是牢固抓握，而不是根据交互调节抓握。我们提出了TAC-LOCO，这是一种增强触觉的统一强化学习框架，它将来自柔性夹持器的触觉阵列观察编码为紧凑的潜在表示，并与本体感知结合，实现对腿部、手臂和夹持器的统一控制。通过有效的抓握稳定性奖励设计，策略学习同时跟踪身体速度和末端执行器轨迹，调节抓握力量，并防止物体在逐渐负载变化和突然释放事件下滑落。我们在Unitree Go2上零样本部署该策略，配备Interbotix WidowX 250手臂和触觉夹持器，展示了在不同外部交互下的动态触觉信息驱动的运动操控，实现了抓握力量减少47%和物体掉落率低于1%。
cs.RO / 26 / 2607.10161
Millimeter Wave Radar: From Synthetic Aperture to Probabilistic Mapping
毫米波雷达：从合成孔径到概率映射
Abstract
Robust probabilistic mapping is essential for autonomous robotic systems operating in challenging environments. While traditional sensors fail in adverse conditions such as smoke and fog, millimeter wave (mmWave) radar sensors offer reliable sensing in such conditions. However, creating accurate probabilistic maps from radar data presents significant challenges due to the inherently sparse and noisy characteristics of radio wave measurements and signal processing steps. In an attempt to address these issues, we establish a complete pipeline from raw radar signals to probabilistic occupancy maps, incorporating Synthetic Aperture Radar processing followed by a probabilistic modeling step. We conduct extensive validation across indoor environments, comparing our approach against different signal processing and probabilistic modeling approaches. We also evaluate mapping quality through downstream path planning performance analysis. Furthermore, we investigate the impact of key parameters and antenna array configuration on mapping performance. The experimental results demonstrate both the effectiveness and limitations of SAR-based probabilistic mapping for real-world robotic deployment. To facilitate future research and broader adoption, we contribute an open-source cascaded mmWave radar dataset with an accompanying GPU-accelerated signal processing pipeline available at https://github.com/rpl-cmu/rpm.
Chinese Translation
稳健的概率映射对于在复杂环境中运行的自主机器人系统至关重要。传统传感器在烟雾和雾霾等恶劣条件下表现不佳，而毫米波（mmWave）雷达传感器在这些条件下提供了可靠的感知。然而，由于无线电波测量和信号处理步骤固有的稀疏性和噪声特性，从雷达数据创建准确的概率地图面临重大挑战。为了解决这些问题，我们建立了一个从原始雷达信号到概率占用地图的完整流程，结合了合成孔径雷达（Synthetic Aperture Radar）处理和概率建模步骤。我们在室内环境中进行了广泛的验证，将我们的方法与不同的信号处理和概率建模方法进行了比较。我们还通过下游路径规划性能分析评估了映射质量。此外，我们研究了关键参数和天线阵列配置对映射性能的影响。实验结果展示了基于SAR的概率映射在实际机器人部署中的有效性和局限性。为了促进未来的研究和更广泛的应用，我们贡献了一个开源的级联mmWave雷达数据集，并提供了一个伴随的GPU加速信号处理流程，地址为 https://github.com/rpl-cmu/rpm 。
cs.RO / 27 / 2607.10170
From Non-Rigid to Rigid: Safe Acquisition of Rigid Communication Graphs under Limited Sensing
从非刚性到刚性：在有限感知下安全获取刚性通信图
Abstract
Communication graph rigidity is a fundamental requirement in many multi robot formation control approaches. However, ensuring and maintaining a rigid communication topology becomes challenging in practice due to limited sensing ranges and dynamic operating conditions. This paper provides a method for achieving an inter robot collision free, rigid time varying communication graph, where communication links are established or broken according to limited sensing ranges, without assuming an initial rigid graph. In addition, the proposed approach guarantees the realization of a rigid graph for heterogeneous nonlinear multi robot systems. A computationally lean, distributed quadratic optimization-based controller is developed for a leader follower architecture, acquiring rigidity based on hierarchical second-order consensus among robots. Follower agents do not require global absolute positions of any agent, including their own. The proposed method is validated through both simulations and hardware experiments in a motion-capture environment, demonstrating reliable performance under the limited sensing capabilities of individual robots.
Chinese Translation
通信图的刚性是许多多机器人编队控制方法中的基本要求。然而，由于感知范围有限和动态操作条件，确保和维持刚性通信拓扑在实践中变得具有挑战性。本文提供了一种方法，以实现机器人间无碰撞的刚性时变通信图，其中通信链接根据有限的感知范围建立或断开，而不假设初始刚性图。此外，所提出的方法保证了异构非线性多机器人系统中刚性图的实现。为领导-跟随架构开发了一种计算上高效的分布式基于二次优化的控制器，通过机器人之间的层次二阶共识来获取刚性。跟随代理不需要任何代理的全局绝对位置，包括它们自己的位置。所提出的方法通过在运动捕捉环境中的仿真和硬件实验得到了验证，展示了在个体机器人有限感知能力下的可靠性能。
cs.RO / 28 / 2607.10172
On the Efficiency of LoRA Fine-Tuning for Vision-Language-Action Models in Industrial Robotic Manipulation
工业机器人操作中视觉-语言-动作模型的LoRA微调效率研究
Abstract
Deploying billion-parameter Vision-Language-Action (VLA) models on industrial hardware requires fine-tuning to bridge the embodiment gap. Full Fine-Tuning (FFT) provides maximal plasticity but requires data centre-grade GPUs. We present a systematic study of Low-Rank Adaptation (LoRA) for $\pi_0$, a flow-matching VLA, evaluated on four precision assembly tasks with a UR5e robotic manipulator. Across a sweep of LoRA ranks (r=8 to 256), allocation strategies, and component-freezing ablations, we find no statistically significant advantage of FFT over certain LoRA configurations. Performance saturates at r=32, and uniform allocation across the Vision-Language-Model (VLM) backbone and action expert proves sufficient. Freezing the VLM or restricting the vision encoder to LoRA significantly degrades performance, indicating that embodiment adaptation requires both semantic and visual plasticity. These results suggest that LoRA at r=32 with full vision encoder fine-tuning is a practical approach, reducing static peak VRAM from 36.2 to 10.8 GiB (parameters and optimizer states, activation memory excluded) without detectable performance loss.
Chinese Translation
在工业硬件上部署十亿参数的视觉-语言-动作（VLA）模型需要进行微调以弥补体现差距。完全微调（FFT）提供了最大的可塑性，但需要数据中心级别的GPU。我们对低秩适应（LoRA）进行了系统研究，针对$ ext{π}_0$这一流匹配的VLA模型，在四个精密装配任务中使用UR5e机器人操纵器进行评估。在对LoRA秩（r=8至256）、分配策略和组件冻结消融的全面测试中，我们发现某些LoRA配置相比于FFT没有统计学上的显著优势。性能在r=32时达到饱和，且在视觉-语言模型（VLM）主干和动作专家之间均匀分配证明是足够的。冻结VLM或将视觉编码器限制为LoRA显著降低了性能，表明体现适应需要语义和视觉的可塑性。这些结果表明，使用r=32的LoRA结合完全视觉编码器微调是一种实用的方法，将静态峰值VRAM从36.2 GiB降低到10.8 GiB（不包括参数和优化器状态、激活内存），且没有可检测的性能损失。
cs.RO / 29 / 2607.10180
ActiveFly-Bench: Aligning Embodied Question Answering with Vision-Language-Action for Aerial Embodied Perception
ActiveFly-Bench：将具身问答与视觉-语言-行动对齐以实现空中具身感知
Abstract
We introduce ActiveFly-Bench, the first benchmark to bridge cyberspace reasoning and physical-world interaction for UAV embodied perception. The benchmark decomposes active perception into three hierarchical tasks: Aerial Embodied Question Answering (Air-EQA), Observation Behavior Planning (OBP), and Fine-grained Language-guided UAV Control (FLUC), explicitly connecting high-level task understanding, behavior planning, and low-level control. The datasets are collected from both real-world and simulated outdoor environments for training and evaluation. We further develop ActiveFly, a closed-loop UAV agent that integrates visual-language reasoning with fine-grained control, and deploy it on a physical UAV platform. Experiments with representative VLMs and VLA models show that current UAV agents still struggle with behavior planning, viewpoint adjustment, and robust task completion in active perception. These results establish ActiveFly-Bench as a new testbed for embodied aerial intelligence.
Chinese Translation
我们介绍了ActiveFly-Bench，这是第一个旨在连接网络空间推理与物理世界交互的无人机具身感知基准。该基准将主动感知分解为三个层次的任务：空中具身问答（Air-EQA）、观察行为规划（OBP）和细粒度语言引导的无人机控制（FLUC），明确连接了高层次任务理解、行为规划和低层次控制。这些数据集来自真实世界和模拟户外环境，用于训练和评估。我们进一步开发了ActiveFly，一个将视觉-语言推理与细粒度控制相结合的闭环无人机代理，并将其部署在物理无人机平台上。与代表性的视觉-语言模型（VLMs）和视觉-语言-行动模型（VLA）进行的实验表明，目前的无人机代理在主动感知中的行为规划、视角调整和任务完成的鲁棒性方面仍然存在困难。这些结果确立了ActiveFly-Bench作为具身空中智能的新测试平台。
cs.RO / 30 / 2607.10206
Source-Lifted Flow Matching for Intervenable Multimodal Imitation
可干预的多模态模仿学习中的源提升流匹配
Abstract
Flow-matching policies are promising for imitation learning because they model complex multimodal action distributions. However, their stochasticity is largely passive: repeated sampling may yield diverse behaviors, but users cannot directly choose among valid continuations from the same state. We propose Source-Lifted Flow Matching (SL-FM), a source-intervenable flow-matching policy that exposes such a handle while keeping the velocity field shared and latent-free. The handle selects only the source endpoint of the conditional flow, not a mode-specific field, preserving the standard formulation while avoiding decomposition into separate mode-conditioned dynamics. The core mechanism is \textbf{Orthogonal Source Lifting}, designed to prevent path-crossing ambiguity. Instead of partitioning target actions by mode, SL-FM lifts handle-specific sources into auxiliary orthogonal coordinates and keeps targets in the original action subspace. This preserves the demonstrated action distribution while allowing one shared field to carry different branches without merging at crossings. To keep handles usable across states, we learn a state-dependent source mixture end to end and use a responsibility floor, giving each handle weak supervision and mitigating dead modes. Experiments on crossing-flow diagnostics and robot-control benchmarks show that SL-FM converts passive source randomness into an actionable intervention variable. It removes crossing-induced composite trajectories, changes future routes in 91.1\% of matched-prefix interventions, and achieves strong free-deployment performance, with improvements in several benchmark settings. Overall, source geometry provides actionable multimodal control without conditioning the velocity field on the selected mode.
Chinese Translation
流匹配策略在模仿学习中具有良好的前景，因为它们能够建模复杂的多模态动作分布。然而，它们的随机性在很大程度上是被动的：重复采样可能产生多样的行为，但用户无法直接从同一状态中选择有效的延续。我们提出了源提升流匹配（Source-Lifted Flow Matching, SL-FM），这是一种可干预的流匹配策略，能够提供这样的控制，同时保持共享的速度场和潜在的无干扰性。该控制仅选择条件流的源端点，而不是特定模式的场，保留了标准公式，同时避免分解为独立的模式条件动态。核心机制是正交源提升（Orthogonal Source Lifting），旨在防止路径交叉模糊。SL-FM不通过模式对目标动作进行划分，而是将特定控制的源提升到辅助正交坐标中，并将目标保持在原始动作子空间。这保留了已展示的动作分布，同时允许一个共享场在交叉处承载不同的分支而不合并。为了保持控制在不同状态下的可用性，我们端到端地学习了状态依赖的源混合，并使用责任底线，给予每个控制弱监督并减轻死模式的影响。在交叉流诊断和机器人控制基准测试中的实验表明，SL-FM将被动的源随机性转化为可操作的干预变量。它消除了由交叉引起的复合轨迹，在91.1%的匹配前缀干预中改变了未来的路径，并在多个基准设置中实现了强大的自由部署性能。总体而言，源几何提供了可操作的多模态控制，而无需将速度场条件化于所选模式。
cs.RO / 31 / 2607.10243
Diffusion-Residual Model Predictive Steering Control for Vehicle Stabilization at the Limit of Handling under Model Uncertainty
基于扩散-残差模型的预测控制转向系统在处理极限下的车辆稳定性控制
Abstract
At the limit of handling, a stabilizing MPC depends on the yaw-rate reference it tracks and the stable-handling envelope it enforces, both operating-point-dependent and unknown a priori, so fixed or worst-case settings are either too conservative or unsafe. We learn this uncertainty with a conditional diffusion residual model and apply it to the controller's reference and constraints rather than its control law. Conditioned on the steering command, the model returns the residual's mean and a predictive spread: the mean re-sizes the tracked yaw reference, while the spread, propagated over the prediction horizon, tightens the stable-handling envelope through a one-sided chance back-off. Together these form the proposed diffusion-residual MPC (D-res), so caution is anticipated ahead of the tracking error rather than corrected after it by a high-gain loop. Because only two moments per command are needed, the generator is tabulated offline and the online controller adds a single table lookup to the baseline MPC, with no in-loop diffusion; it runs within the 100 Hz budget on an NVIDIA Jetson AGX Xavier (worst-case 4.08 ms per step). Across a 7-DOF model and high-fidelity CarMaker co-simulation spanning vehicle, tire, road, and maneuver diversity, D-res reduces peak side-slip where the fixed bicycle model is least accurate and restores directional stability on low-friction maneuvers, where the fixed reference over-commands the available grip.
Chinese Translation
在处理极限下，稳定的模型预测控制（MPC）依赖于其跟踪的偏航率参考和强制执行的稳定处理包络，这两者都是依赖于操作点且事先未知的，因此固定或最坏情况的设置要么过于保守，要么不安全。我们利用条件扩散残差模型来学习这种不确定性，并将其应用于控制器的参考和约束，而不是控制律。基于转向指令，该模型返回残差的均值和预测范围：均值重新调整跟踪的偏航参考，而范围在预测时间范围内传播，通过单侧机会回退来收紧稳定处理包络。这两者共同构成了所提出的扩散-残差模型预测控制（D-res），因此在跟踪误差之前预期谨慎，而不是通过高增益回路在之后进行修正。由于每个指令只需两个时刻，生成器在离线时进行表格化，在线控制器在基线MPC中增加了一个单一的表查找，而没有在回路中进行扩散；它在NVIDIA Jetson AGX Xavier上以100 Hz的预算运行（最坏情况下每步4.08毫秒）。在一个7自由度模型和高保真CarMaker联合仿真中，涵盖了车辆、轮胎、道路和机动多样性，D-res在固定自行车模型最不准确的地方减少了峰值侧滑，并在低摩擦机动中恢复了方向稳定性，在这些情况下，固定参考过度指挥了可用的抓地力。
cs.RO / 32 / 2607.10288
PIER-Flow: Physics-Informed Efficient Rectified Flow for Real-Time Mobile Robot Navigation
PIER-Flow：物理信息高效整流流用于实时移动机器人导航
Abstract
Autonomous navigation in dense and highly dynamic environments requires both physically feasible control and low-latency replanning. Optimization-based methods such as Model Predictive Control (MPC) explicitly handle robot kinematics and safety constraints, but repeated nonlinear optimization can limit real-time responsiveness. Deterministic behavior-cloning policies enable efficient inference but may fail to represent multimodal avoidance behaviors, whereas diffusion policies capture multimodality at the cost of time-consuming iterative denoising. We propose PIER-Flow (Physics-Informed Efficient Rectified Flow), a lightweight navigation policy for mobile robots. By distilling an MPC expert into a continuous-time Ordinary Differential Equation (ODE), PIER-Flow achieves single-step action generation through parallel latent sampling and lightweight feasibility selection. We introduce a physics-informed training objective to enforce kinematic consistency, paired with an asynchronous action chunking architecture for robust sim-to-real deployment. Extensive simulations demonstrate that PIER-Flow achieves a 98.85\% success rate and zero collisions, with an average inference of $\sim$1.29 ms, which accelerates planning by 37.2$\times$ compared to MPC and over 800$\times$ against standard diffusion models. Crucially, real-world deployment on a resource-constrained edge computer further achieves an approximately stable inference latency of $\sim$5.3 ms, avoiding the latency spikes and freezing events observed with planning baselines.
Chinese Translation
在密集且高度动态的环境中，自主导航需要既符合物理可行性的控制，又具备低延迟的重新规划能力。基于优化的方法，如模型预测控制（Model Predictive Control, MPC），明确处理机器人运动学和安全约束，但重复的非线性优化可能限制实时响应能力。确定性行为克隆策略能够实现高效推理，但可能无法有效表示多模态避障行为，而扩散策略则以耗时的迭代去噪为代价来捕捉多模态性。我们提出了PIER-Flow（物理信息高效整流流），一种轻量级的移动机器人导航策略。通过将MPC专家提炼为连续时间的常微分方程（Ordinary Differential Equation, ODE），PIER-Flow通过并行潜在采样和轻量级可行性选择实现单步动作生成。我们引入了一种物理信息训练目标，以强制执行运动学一致性，并配合异步动作分块架构以实现稳健的仿真到现实部署。大量仿真表明，PIER-Flow实现了98.85%的成功率和零碰撞，平均推理时间约为1.29毫秒，相比MPC加速规划37.2倍，相较于标准扩散模型加速超过800倍。关键的是，在资源受限的边缘计算机上的实际部署进一步实现了约5.3毫秒的稳定推理延迟，避免了在规划基线中观察到的延迟峰值和冻结事件。
cs.RO / 33 / 2607.10294
Soft Eversion Robots for Colonoscopy: Challenges, Open Problems, and Emerging Solutions
用于结肠镜检查的软翻转机器人：挑战、开放问题与新兴解决方案
Abstract
Conventional colonoscopy remains limited by patient discomfort and procedural risks, motivating research into compliant robotic alternatives. Eversion robots, which advance via pressure-driven tip growth, eliminate sliding friction against the colon wall and offer a less intrusive approach to traversal. However, no existing design simultaneously satisfies all clinical requirements. This paper benchmarks four recent eversion robot architectures against the key anatomical and clinical constraints of colonoscopy, including colon length, minimum luminal diameter, bending angle, and working-channel needs. We identify the central trade-offs each design reveals, particularly the difficulty of integrating steering and payload delivery without sacrificing the soft, low-friction behaviour that makes eversion robots attractive. We then provide design guidance across material selection, steering strategy, and payload delivery, and highlight open problems for clinical translation.
Chinese Translation
传统的结肠镜检查受到患者不适和程序风险的限制，这促使了对顺应性机器人替代方案的研究。翻转机器人通过压力驱动的尖端生长前进，消除了与结肠壁的滑动摩擦，提供了一种更不具侵入性的穿越方式。然而，现有设计尚未能同时满足所有临床要求。本文对四种近期的翻转机器人架构进行了基准测试，评估其在结肠镜检查的关键解剖和临床约束下的表现，包括结肠长度、最小腔径、弯曲角度和工作通道需求。我们识别出每种设计所揭示的核心权衡，特别是在不牺牲使翻转机器人具有吸引力的软性、低摩擦特性情况下，整合转向和负载传递的难度。随后，我们提供了在材料选择、转向策略和负载传递方面的设计指导，并强调了临床转化中的开放问题。
cs.RO / 34 / 2607.10336
PrismAD: Decoupled Planning via Semantic Mixture-of-Planners for End-to-End Autonomous Driving
PrismAD：基于语义混合规划者的解耦规划用于端到端自主驾驶
Abstract
This letter presents PrismAD, a decoupled end-to-end autonomous driving framework based on a Semantic Mixture-of-Planners. Existing planners usually aggregate heterogeneous scene tokens into a coupled representation space, forcing a single planning branch to jointly model agent interaction, road geometry, and driving intention. Such coupling may weaken factor-specific reasoning and obscure the contribution of different planning cues. To address this limitation, PrismAD partitions scene tokens into interaction, geometry, and intent groups, and assigns them to independent planning experts with the same architecture but separate parameters. Each expert learns a specialized motion-planning representation, while a semantics-aware router adaptively aggregates expert predictions with separate routing weights for motion prediction and ego planning. Sparse top-$K$ activation with noisy gating is further introduced to improve routing robustness and reduce unnecessary expert computation. Extensive experiments on the nuScenes open-loop dataset and NeuroNCAP closed-loop benchmark demonstrate that PrismAD exhibits competitive performance. Our code will be released soon.
Chinese Translation
本文提出了PrismAD，一个基于语义混合规划者的解耦端到端自主驾驶框架。现有的规划者通常将异构场景标记聚合到一个耦合的表示空间中，迫使单一规划分支共同建模代理交互、道路几何和驾驶意图。这种耦合可能削弱特定因素的推理能力，并模糊不同规划线索的贡献。为了解决这一局限性，PrismAD将场景标记划分为交互、几何和意图组，并将它们分配给具有相同架构但参数独立的规划专家。每个专家学习专门的运动规划表示，而一个语义感知路由器则自适应地聚合专家预测，为运动预测和自我规划分配独立的路由权重。进一步引入稀疏的 top-$K$ 激活与噪声门控，以提高路由的鲁棒性并减少不必要的专家计算。在nuScenes开放循环数据集和NeuroNCAP闭环基准测试上的大量实验表明，PrismAD表现出具有竞争力的性能。我们的代码将很快发布。
cs.RO / 35 / 2607.10369
VINE: Taming Generative Control Policies for Reinforcement Learning
VINE：驯化生成控制策略以用于强化学习
Abstract
Flow-matching policies have emerged as an effective policy parameterization for robot learning. They iteratively generate actions from noise, enabling highly expressive modeling of complex and multimodal action distributions. However, prior works observed that scaling these policies with value-gradient reinforcement learning (RL) often leads to training instability. Existing methods attribute this instability to iterative generation and therefore avoid end-to-end value-gradient optimization by sacrificing iterative generation, high expressiveness, or value-gradient optimization. Contrary to prior belief, we show the instability does not stem from iterative generation itself, but from the vanilla sampling strategy originally designed for behavior cloning, which becomes brittle under value-gradient RL. Motivated by this insight, we propose VINE, an RL-oriented sampling method that enables stable end-to-end value-gradient optimization for flow-matching policies. Instead of following a single flow trajectory, VINE reconstructs a new interpolation state at every denoising step, creating a stable differentiable path for value-gradient propagation while remaining compatible with the original flow-matching denoising process. As a result, VINE preserves the expressiveness and iterative generation of flow-matching without sacrificing end-to-end value-gradient optimization. Despite performing end-to-end backpropagation through all ten denoising steps, VINE achieves stable policy improvement and consistently outperforms state-of-the-art RL methods on the OGBench offline RL benchmark and real-world robotic manipulation task. Videos are available on our website: https://agibottech.github.io/vine.
Chinese Translation
流匹配策略已成为机器人学习中一种有效的策略参数化方法。它们通过噪声迭代生成动作，从而能够高度表达复杂和多模态的动作分布。然而，先前的研究观察到，使用价值梯度强化学习（RL）来扩展这些策略通常会导致训练不稳定。现有方法将这种不稳定归因于迭代生成，因此通过牺牲迭代生成、高表达能力或价值梯度优化来避免端到端的价值梯度优化。与之前的观点相反，我们表明这种不稳定并非源于迭代生成本身，而是源于最初为行为克隆设计的普通采样策略，在价值梯度RL下变得脆弱。基于这一见解，我们提出了VINE，一种面向强化学习的采样方法，能够实现流匹配策略的稳定端到端价值梯度优化。VINE在每个去噪步骤中重构一个新的插值状态，而不是遵循单一的流轨迹，从而为价值梯度传播创造了一个稳定的可微路径，同时与原始流匹配去噪过程兼容。因此，VINE在不牺牲端到端价值梯度优化的情况下，保留了流匹配的表达能力和迭代生成。尽管在所有十个去噪步骤中执行端到端反向传播，VINE仍实现了稳定的策略改进，并在OGBench离线RL基准和真实世界的机器人操控任务中始终优于最先进的RL方法。视频可在我们的网站上查看：https://agibottech.github.io/vine。
cs.RO / 36 / 2607.10372
Robotic Contextual Awareness for Human-Robot Collaboration and Environmental Understanding
机器人情境意识在人与机器人协作与环境理解中的应用
Abstract
The transition of autonomous mobile robots from controlled industrial settings to dynamic, human-centric environments, such as manufacturing, logistics, and healthcare, has made their safe and autonomous operation a critical area of research. These sophisticated machines must be capable of perceiving, understanding, and interacting with their surroundings to navigate freely and perform complex tasks. A significant obstacle to achieving this is the lack of comprehensive contextual awareness, which requires a robot to recognize its spatial environment and identify the objects and actors within it. Without this perceptual knowledge, robots struggle to plan adaptive behaviors or engage in meaningful interaction with humans. This thesis presents novel solutions to this challenge by exploring two distinct but complementary research directions. The first direction involves human re-identification and tracking to improve Human-Robot Collaboration. Our developed approach enables a mobile robot to recognize a specific person, facilitating targeted collaboration while ignoring other individuals. The second direction focuses on enhancing the robot's overall perceptual capabilities to understand its environment geometrically and semantically. Geometric information is vital for motion planning and collision avoidance, while semantic knowledge provides the robot with a richer understanding for more advanced interaction. Both solutions are driven by the improvement of the semantical understanding of robots that enhance their knowledge of their surroundings, allowing a smoother and more natural interaction between robots, humans, and the environment. The contributions of this work in human re-identification and environmental understanding represent a significant step toward a future where robots are more contextually aware, enabling safer coexistence and more effective collaboration.
Chinese Translation
自主移动机器人从受控的工业环境过渡到动态的人本环境，如制造、物流和医疗保健，使其安全和自主操作成为一个关键的研究领域。这些复杂的机器必须能够感知、理解并与周围环境互动，以便自由导航和执行复杂任务。实现这一目标的一个重大障碍是缺乏全面的情境意识，这要求机器人能够识别其空间环境并识别其中的物体和参与者。如果没有这种感知知识，机器人在规划适应性行为或与人类进行有意义的互动时会面临困难。本论文通过探索两个不同但互补的研究方向，提出了应对这一挑战的新解决方案。第一个方向涉及人类的重新识别和跟踪，以改善人与机器人协作。我们开发的方法使移动机器人能够识别特定的人，从而促进针对性的协作，同时忽略其他个体。第二个方向则侧重于增强机器人的整体感知能力，以几何和语义的方式理解其环境。几何信息对于运动规划和避碰至关重要，而语义知识则为机器人提供了更丰富的理解，以实现更高级的互动。这两种解决方案都旨在提高机器人的语义理解能力，增强其对周围环境的认知，从而实现机器人、人类和环境之间更加顺畅和自然的互动。本研究在人类重新识别和环境理解方面的贡献，代表了朝着一个机器人更具情境意识的未来迈出了重要一步，从而实现更安全的共存和更有效的协作。
cs.RO / 37 / 2607.10374
Navigating the Crowd: Non-linear MPC with Social Forces Dynamics for Human-Aware Robot Navigation
在人群中导航：基于社会力动态的非线性模型预测控制用于人类感知机器人导航
Abstract
Safe and socially compliant navigation remains a fundamental challenge for autonomous robots operating in human-populated environments. Beyond collision avoidance, robots must anticipate human motion and respect personal space to ensure human comfort. Model Predictive Control (MPC) offers a robust alternative to classical and data-driven methods, although its effectiveness strongly depends on accurate human motion prediction and efficient computation. This paper introduces SFM-NMPC, a Social Force Model-based Non-linear Model Predictive Control framework that embeds human motion prediction directly within the optimization loop. By incorporating the Social Force Model into the dynamic model of surrounding agents, the controller jointly predicts the trajectories of humans and robots over the prediction horizon, thereby enabling socially-aware planning. A tailored set of social cost functions guides the optimization toward human-compliant behaviors. Despite the increased model complexity, the proposed formulation runs in real time at 20 Hz. Extensive simulated testing in crowded environments demonstrates that SFM-NMPC outperforms state-of-the-art baselines in social compliance metrics while maintaining efficient and smooth navigation. Visual trajectory analysis and an ablation study further highlight the contribution of the embedded SFM dynamics and social cost terms, confirming the effectiveness of the proposed approach for real-world social navigation.
Chinese Translation
安全且符合社会规范的导航仍然是自主机器人在人员密集环境中面临的基本挑战。除了避免碰撞之外，机器人还必须预测人类的运动并尊重个人空间，以确保人类的舒适感。模型预测控制（Model Predictive Control, MPC）提供了一种比传统方法和数据驱动方法更为稳健的替代方案，尽管其有效性在很大程度上依赖于准确的人类运动预测和高效的计算。本文提出了SFM-NMPC，一种基于社会力模型的非线性模型预测控制框架，该框架将人类运动预测直接嵌入优化循环中。通过将社会力模型纳入周围代理的动态模型，控制器在预测范围内共同预测人类和机器人的轨迹，从而实现社会感知的规划。一组量身定制的社会成本函数引导优化朝向符合人类行为的方向。尽管模型复杂性增加，但所提出的公式以20 Hz的实时速度运行。在拥挤环境中的广泛模拟测试表明，SFM-NMPC在社会合规性指标上优于最先进的基线，同时保持高效和平滑的导航。视觉轨迹分析和消融研究进一步突显了嵌入的社会力动态和社会成本项的贡献，确认了所提出方法在现实世界社会导航中的有效性。
cs.RO / 38 / 2607.10437
Interleaved POMDP Planning for Multi-Object Search in Unknown Multi-Room Household Environments
未知多房间家庭环境中的多目标搜索交错POMDP规划
Abstract
Multi-object search in unknown household environments requires planning under extensive uncertainty - from unknown object locations to cluttered spaces with unobserved obstacles. POMDPs offer a principled framework for such problems but remain intractable in large domains. We propose Inter-POMDP, a novel interleaved POMDP planning algorithm that decomposes this challenge into two interacting levels: a high-level POUCT planner reasons over object distributions using LLM-informed histogram beliefs, while a low-level motion planner models navigation uncertainty with obstacle-aware particle beliefs as domain knowledge to guide high-level POUCT. This interleaved design balances planning quality and efficiency despite the large search space across unknown multi-room environments. Both simulation and real-world experiments show that our Inter-POMDP algorithm reduces collision counts by up to 63%, navigation steps by up to 35%, and detection counts by up to 32% compared with baseline methods. Full videos are https://sites.google.com/view/inter-pomdp
Chinese Translation
在未知家庭环境中进行多目标搜索需要在广泛的不确定性下进行规划——从未知的物体位置到充满未观察障碍物的杂乱空间。部分可观测马尔可夫决策过程（POMDP）为此类问题提供了一个原则性的框架，但在大领域中仍然难以处理。我们提出了Inter-POMDP，这是一种新颖的交错POMDP规划算法，它将这一挑战分解为两个相互作用的层次：高层的部分可观测不确定性树（POUCT）规划器使用基于大语言模型（LLM）信息的直方图信念对物体分布进行推理，而低层的运动规划器则通过障碍物感知的粒子信念建模导航不确定性，将其作为领域知识来指导高层的POUCT。这种交错设计在未知多房间环境中尽管搜索空间庞大，仍然平衡了规划质量和效率。模拟和现实世界实验表明，我们的Inter-POMDP算法与基线方法相比，减少了高达63%的碰撞次数、35%的导航步骤和32%的检测次数。完整视频请访问 https://sites.google.com/view/inter-pomdp
cs.RO / 39 / 2607.10438
Large Language Model Enhanced Differentiable Trajectory Planning for IoT-Enabled Autonomous Driving
增强大语言模型的可微分轨迹规划用于物联网驱动的自动驾驶
Abstract
Autonomous driving planning is a key component of IoT-enabled intelligent transportation systems, requiring vehicles to generate safe, efficient, and executable trajectories in complex urban environments from multi-source contextual information. While imitation learning (IL) has shown promise on large-scale datasets, IL-based planners still suffer from limited coverage of complex long-tail interactions, weak consistency with downstream constrained refinement, and insufficient use of high level scene semantics under real time constraints. To address these issues, this paper proposes a large language model (LLM) enhanced differentiable trajectory planning framework for IoT-enabled autonomous driving. Specifically, we introduce a surrounding agent centric data augmentation strategy to reorganize sur rounding agent trajectories as additional planning supervision, thereby improving the training distribution without collecting additional raw data. We further design a complexity-aware asyn chronous LLM-based semantic enhancement module to extract scene-related high-level semantic features with controlled online overhead. In addition, a differentiable optimization module is incorporated to refine generated trajectories with explicit residual penalties while backpropagating optimization gradients to the upstream planner. Experiments show that the proposed method achieves the best overall scores of 83.63 and 78.29 on the nuPlan closed-loop nonreactive and reactive Hard20 benchmarks, respectively, and CARLA-ROS tests further verify its online deployment and real time closed-loop execution capability.
Chinese Translation
自动驾驶规划是物联网智能交通系统的关键组成部分，要求车辆在复杂城市环境中从多源上下文信息生成安全、高效且可执行的轨迹。尽管模仿学习（IL）在大规模数据集上显示出前景，但基于IL的规划器仍然存在复杂长尾交互覆盖有限、与下游约束优化一致性较弱以及在实时约束下对高层场景语义利用不足等问题。为了解决这些问题，本文提出了一种增强大语言模型（LLM）的可微分轨迹规划框架，用于物联网驱动的自动驾驶。具体而言，我们引入了一种围绕代理中心的数据增强策略，以重新组织周围代理轨迹作为额外的规划监督，从而在不收集额外原始数据的情况下改善训练分布。我们进一步设计了一种复杂度感知的基于LLM的语义增强模块，以提取与场景相关的高层语义特征，并控制在线开销。此外，结合了一个可微分优化模块，以显式残差惩罚精炼生成的轨迹，同时将优化梯度反向传播到上游规划器。实验表明，所提出的方法在nuPlan闭环非反应性和反应性Hard20基准测试中分别取得了最佳总体得分83.63和78.29，CARLA-ROS测试进一步验证了其在线部署和实时闭环执行能力。
cs.RO / 40 / 2607.10504
SUREFlow: State-space Uncertainty-aware REsidual Flow Matching for Robust Robot Manipulation
SUREFlow：状态空间不确定性感知残差流匹配用于鲁棒机器人操作
Abstract
Generative vision-language-action policies have advanced robot manipulation, but they often exhibit instability under noise, partial observability, and stochastic initial conditions. During extended rollouts, small velocity errors accumulate, degrading execution reliability. Existing diffusion and flow-based policies typically assume homoscedastic residuals and lack explicit uncertainty modeling within action generation, limiting robustness during iterative rollout. We propose SUREFlow, a state-space uncertainty-aware residual flow matching framework built on a Mamba backbone. The method jointly predicts action velocities and input-dependent residual uncertainty, enabling selective refinement of unreliable action dimensions without environment feedback while preserving computational efficiency. On LIBERO, SUREFlow achieves 92.5% average success rate (SR), outperforming the Mamba-based MaIL by 34.2%. On LIBERO-PRO, it attains around 49% SR using only 179M parameters, achieving performance comparable to large VLAs with 3-7B parameters. SUREFlow source code is available on: https://github.com/tanvirnwu/SUREFlow
Chinese Translation
生成性视觉-语言-动作策略推动了机器人操作的发展，但在噪声、部分可观测性和随机初始条件下，它们往往表现出不稳定性。在长时间的执行过程中，小的速度误差会积累，从而降低执行的可靠性。现有的扩散和基于流的方法通常假设同方差残差，并且在动作生成中缺乏明确的不确定性建模，限制了在迭代执行过程中的鲁棒性。我们提出了SUREFlow，一种基于Mamba骨干网的状态空间不确定性感知残差流匹配框架。该方法共同预测动作速度和输入依赖的残差不确定性，使得在没有环境反馈的情况下能够选择性地优化不可靠的动作维度，同时保持计算效率。在LIBERO上，SUREFlow实现了92.5%的平均成功率（SR），比基于Mamba的MaIL提高了34.2%。在LIBERO-PRO上，仅使用179M参数便达到了约49%的成功率，表现与大型视觉-语言-动作模型（VLA）相当，这些模型的参数量在30亿到70亿之间。SUREFlow的源代码可在以下链接获取：https://github.com/tanvirnwu/SUREFlow
cs.RO / 41 / 2607.10553
SLIDER: Sparse History-Guided Aerial Robot Target Search using Sliding Local Maps
SLIDER：基于滑动局部地图的稀疏历史引导空中机器人目标搜索
Abstract
Efficient exploration and target search in large-scale unknown environments remain challenging for aerial robots due to the demands of broad spatial coverage, fine-grained perception, and real-time decision-making. This paper presents SLIDER, a lightweight and memory-efficient framework that avoids reliance on globally dense maps by combining a local sliding map with sparse global history information. A novel observation quality evaluation method is proposed, leveraging historical poses and sensor models to assess point cloud data in real-time, enabling efficient frontier detection. To support scalable and responsive planning, an incremental viewpoint clustering strategy dynamically adapts to local updates, significantly reducing the number of candidate targets and decreasing computational load. A sparse global topological map is incrementally maintained to assist global planning and cost evaluation. Extensive simulations and real-world experiments demonstrate that the proposed system outperforms state-of-the-art methods in memory usage, decision latency, and search efficiency.
Chinese Translation
在大规模未知环境中，高效的探索和目标搜索对空中机器人仍然是一个挑战，因为这需要广泛的空间覆盖、细致的感知和实时决策。本文提出了SLIDER，一个轻量级且内存高效的框架，通过将局部滑动地图与稀疏的全球历史信息相结合，避免对全球密集地图的依赖。我们提出了一种新颖的观测质量评估方法，利用历史姿态和传感器模型实时评估点云数据，从而实现高效的边界检测。为了支持可扩展和响应迅速的规划，我们提出了一种增量视点聚类策略，能够动态适应局部更新，显著减少候选目标数量并降低计算负担。一个稀疏的全球拓扑地图被增量维护，以辅助全球规划和成本评估。大量的仿真和实际实验表明，所提出的系统在内存使用、决策延迟和搜索效率方面优于现有的最先进方法。
cs.RO / 42 / 2607.10565
BucketKD: A Safety-Aware Bucket-Based Knowledge Distillation Framework for End-to-End Motion Planning
BucketKD：一种安全意识的基于桶的知识蒸馏框架用于端到端运动规划
Abstract
End-to-end motion planning has emerged as a promising paradigm in autonomous driving, directly mapping raw sensor data to control commands via deep neural networks. Despite its advantages, its large model size hinders deployment in resource-constrained platforms. In this paper, we present BucketKD, a bucket-based knowledge distillation framework that yields compact and safety-aware end-to-end planners. Compared to the state-of-the-art approach, which relies on simplified planning state representations, BucketKD discretizes critical environmental variables into adaptive buckets that capture richer scene semantics while preserving efficiency. In addition, we design a safety-aware waypoint attention mechanism that evaluates each waypoint's risk level by accounting for both obstacle proximity and relative motion through a time-to-collision (TTC) formulation widely used in transportation research. This enables the student model to better retain safety-critical behaviors during distillation. Extensive experiments in CARLA using the Bench2Drive dataset show that BucketKD significantly outperforms the state-of-the-art in both planning accuracy and safety while maintaining strong compression ratios.
Chinese Translation
端到端运动规划已成为自动驾驶领域一种有前景的范式，通过深度神经网络直接将原始传感器数据映射到控制命令。尽管其具有优势，但其庞大的模型尺寸限制了在资源受限平台上的部署。本文提出了BucketKD，一种基于桶的知识蒸馏框架，能够生成紧凑且具有安全意识的端到端规划器。与依赖简化规划状态表示的最先进方法相比，BucketKD将关键环境变量离散化为自适应桶，从而捕捉更丰富的场景语义，同时保持效率。此外，我们设计了一种安全意识的航点注意机制，通过考虑障碍物接近度和相对运动，利用广泛应用于交通研究的碰撞时间（TTC）公式来评估每个航点的风险水平。这使得学生模型在蒸馏过程中能够更好地保留安全关键行为。在CARLA中使用Bench2Drive数据集的广泛实验表明，BucketKD在规划准确性和安全性方面显著优于最先进的方法，同时保持强大的压缩比。
cs.RO / 43 / 2607.10597
Underwater Dead Reckoning with Deployable Situation-Triggered Covariance Scheduling
可部署情境触发协方差调度的水下航位推算
Abstract
Underwater dead reckoning estimates vehicle position when vision is unavailable and external positioning cannot be assumed. A single set of filter parameters can work well in many situations, but fixed tuning may be poorly matched during turns, motion transitions, or periods when sensor measurements are less reliable. This paper presents the Situation-Triggered Calibrated Adaptive Robust Extended Kalman Filter for a BlueROV2. An onboard probabilistic trigger identifies the current motion situation while one error-state filter runs continuously. When the trigger is confident, the filter changes only to the corresponding pre-calibrated process- and measurement-noise matrices; the state estimate, covariance history, dynamics, and measurement models are not reset or replaced. The trigger, noise profiles, and a one-time Doppler velocity log yaw-alignment correction are calibrated offline using sparse AprilTag-supervised pool runs. A separate validation set selects the scheduling policy, which is then fixed before held-out testing. Across four held-out pool runs, the method reduces label-weighted mean per-run translation root-mean-square error from 0.488 m to 0.471 m relative to the same filter backbone with one global noise profile, and every held-out run favors the scheduled method. A paired bootstrap over 10-second segments gives a candidate-minus-baseline difference of -0.017 m with a 95% confidence interval of [-0.024, -0.008] m, while orientation error remains essentially unchanged. These results indicate that situation-aware covariance scheduling provides a modest but consistent vision-free dead-reckoning improvement without switching estimators or resetting the filter.
Chinese Translation
水下航位推算在视觉不可用且无法假设外部定位时估计车辆位置。虽然在许多情况下，单一的滤波器参数集可以很好地工作，但在转弯、运动过渡或传感器测量不可靠的时期，固定调优可能不匹配。本文提出了一种针对BlueROV2的情境触发校准自适应鲁棒扩展卡尔曼滤波器（Situation-Triggered Calibrated Adaptive Robust Extended Kalman Filter）。一种基于概率的触发器在一个误差状态滤波器持续运行的同时识别当前的运动情境。当触发器有信心时，滤波器仅更改为相应的预校准过程和测量噪声矩阵；状态估计、协方差历史、动态和测量模型不会被重置或替换。触发器、噪声特征以及一次性多普勒速度日志偏航对齐校正通过稀疏的AprilTag监督池运行进行离线校准。一个单独的验证集选择调度策略，然后在保留测试之前固定。通过四次保留池运行，该方法将标签加权的每次运行的平移均方根误差从0.488米降低到0.471米，相对于具有一个全局噪声特征的相同滤波器骨架，并且每次保留运行都偏向于调度方法。对10秒段进行配对自助法分析，候选减去基线的差异为-0.017米，95%置信区间为[-0.024, -0.008]米，而方向误差基本保持不变。这些结果表明，情境感知的协方差调度提供了适度但一致的无视觉航位推算改进，而无需切换估计器或重置滤波器。
cs.RO / 44 / 2607.10625
Dual-Process Atomic Skill Learning: Decoupling Semantic Reasoning and Real-Time Control
双过程原子技能学习：解耦语义推理与实时控制
Abstract
Language-conditioned Imitation Learning (IL) is essential for enabling robots to perform complex tasks following natural language instructions. However, generalizing to multi-step compositional tasks remains a significant challenge. While hierarchical approaches attempt to address this by decomposing tasks into atomic skills, existing methods often suffer from training instability and codebook collapse due to the tight coupling between high-level skill reasoning and low-level action generation in joint training paradigms. Inspired by the Dual-Process Theory of cognition, we propose Dual-Process Atomic Skill Learning (DASL), a novel asynchronous hierarchical imitation learning framework that decouples slow semantic reasoning from fast, real-time motion control. DASL comprises a Slow-Frequency Policy that predicts interpretable, discrete skills via Vector Quantization, and a High-Frequency Policy that leverages a latent diffusion model and a Decision Transformer to generate precise actions conditioned on these latent skills. By asynchronously coordinating these modules and utilizing diffusion to structure the latent space, our framework mitigates the skill codebook interference problem common in joint training paradigms. Evaluations across simulation benchmarks and experiment demonstrate that DASL significantly outperforms state-of-the-art baselines, excelling in skill acquisition and compositional generalization to unseen instructions. GitHub page: https://github.com/Hatakekaka/DASL
Chinese Translation
语言条件下的模仿学习（IL）对于使机器人能够根据自然语言指令执行复杂任务至关重要。然而，向多步骤组合任务的推广仍然是一个重大挑战。尽管层次化方法试图通过将任务分解为原子技能来解决这一问题，但现有方法通常由于高层技能推理与低层动作生成在联合训练范式中的紧密耦合而遭遇训练不稳定和代码本崩溃的问题。受到双过程认知理论的启发，我们提出了双过程原子技能学习（DASL），这是一种新颖的异步层次化模仿学习框架，能够将缓慢的语义推理与快速的实时运动控制解耦。DASL包含一个慢频率策略，通过向量量化预测可解释的离散技能，以及一个高频率策略，利用潜在扩散模型和决策变换器生成基于这些潜在技能的精确动作。通过异步协调这些模块并利用扩散结构化潜在空间，我们的框架减轻了联合训练范式中常见的技能代码本干扰问题。在仿真基准和实验中的评估表明，DASL显著优于最先进的基线，在技能获取和对未见指令的组合泛化方面表现出色。GitHub页面：https://github.com/Hatakekaka/DASL
cs.RO / 45 / 2607.10630
World Models as Adversaries: Multi-Agent Self-Play Fine-Tuning for Robust Motion Planning
作为对手的世界模型：多智能体自我博弈微调用于鲁棒运动规划
Abstract
Robust motion planning in dense traffic requires autonomous vehicles to interact in rare and safety-critical scenarios that are underrepresented in naturalistic driving data. Although adversarial training offers a feasible solution, existing methods often rely on external scenario generators, heuristic perturbations, or simulator-heavy rollouts, which makes them difficult to integrate with modern autoregressive planners. Here, we cast adversarially robust planner learning as a constrained min-max game and propose Adversarial World Modeling (AWM), a theoretically grounded multi-agent self-play fine-tuning framework. Since solving the exact game is intractable, AWM introduces a principled decoupled solver. In the inner minimization, the planner's predictive world model is converted into a role-conditioned adversary that learns sparse, scene-adaptive attack coalitions via counterfactual credit assignment. In the outer maximization, the ego planner optimizes a regret-aware robust best response against the frozen AWM, utilizing tail-risk weighting and reference-anchored trust regions to improve hard-case recovery while preserving nominal driving behavior. Experiments on the nuPlan and InterPlan benchmarks demonstrate that our method generates transferable adversarial interactions and yields a robust planner that achieves competitive closed-loop performance in both nominal and highly interactive long-tail scenarios. Theoretical analysis justifies the decoupled solver and the main optimization components.
Chinese Translation
在密集交通中，鲁棒的运动规划要求自主车辆在自然驾驶数据中稀缺且安全关键的场景中进行交互。尽管对抗训练提供了一种可行的解决方案，但现有方法通常依赖于外部场景生成器、启发式扰动或重度依赖模拟器的回放，这使得它们难以与现代自回归规划器集成。在此，我们将对抗鲁棒规划器学习视为一个约束的最小-最大博弈，并提出对抗世界建模（Adversarial World Modeling, AWM），这是一个理论基础扎实的多智能体自我博弈微调框架。由于精确求解该博弈是不可行的，AWM 引入了一个原则性的解耦求解器。在内部最小化中，规划器的预测世界模型被转换为一个角色条件的对手，通过反事实信用分配学习稀疏的、场景自适应的攻击联盟。在外部最大化中，自我规划器针对冻结的 AWM 优化一个考虑遗憾的鲁棒最佳响应，利用尾部风险加权和参考锚定的信任区域来改善困难案例的恢复，同时保持名义驾驶行为。在 nuPlan 和 InterPlan 基准上的实验表明，我们的方法生成可转移的对抗交互，并产生一个鲁棒的规划器，在名义和高度互动的长尾场景中实现竞争性的闭环性能。理论分析证明了解耦求解器和主要优化组件的合理性。
cs.RO / 46 / 2607.10649
Coverage Path Planning: Classical Foundations, Recent Advances, and Future Directions
覆盖路径规划：经典基础、近期进展与未来方向
Abstract
Coverage path planning (CPP) is a fundamental problem in robot motion planning, whose aim is to produce robot trajectories that provide complete coverage of target workspaces while minimizing task-specific objectives such as path length, overlap, number of turns, and energy consumption. CPP has widespread applications in cleaning, inspection, mapping, agriculture, manufacturing, surveillance, demining, and environmental monitoring. Although classical CPP has been extensively studied, recent advances have extended CPP beyond single-robot settings to multi-robot systems, complex 3D environments, constrained platforms, learning-based coverage planning, and visual coverage tasks. This paper presents a comprehensive survey of 125 representative works published primarily between 2015 and 2026, while presenting the evolution of recent developments in light of the classical CPP methods published before 2015. The CPP methods are organized into six main categories: single-robot CPP, multi-robot CPP, 3D CPP, constrained CPP, learning-based CPP, and visual CPP. For each category, the review summarizes the main planning formulations, representative algorithms, strengths, and limitations. In addition, the review analyzes how environmental knowledge, workspace geometry, robot constraints, sensing objectives, and coordination requirements shape the CPP problem. The survey further discusses open challenges in scalable online planning, multi-robot coordination, 3D and visual coverage, unified platform-constrained and resource-aware coverage, and learning-enhanced coverage. Thus, the survey provides a structured overview of recent CPP developments and future research directions.
Chinese Translation
覆盖路径规划（CPP）是机器人运动规划中的一个基本问题，其目标是生成机器人轨迹，以实现对目标工作空间的完全覆盖，同时最小化特定任务的目标，如路径长度、重叠、转弯次数和能耗。CPP在清洁、检查、制图、农业、制造、监视、排雷和环境监测等领域有广泛应用。尽管经典的CPP已经得到了广泛研究，但近期的进展将CPP从单机器人设置扩展到了多机器人系统、复杂的三维环境、受限平台、基于学习的覆盖规划和视觉覆盖任务。本文对2015年至2026年间发表的125篇代表性研究进行了全面的综述，同时展示了在2015年之前发表的经典CPP方法的基础上，近期发展的演变。CPP方法被组织为六个主要类别：单机器人CPP、多机器人CPP、三维CPP、受限CPP、基于学习的CPP和视觉CPP。对于每个类别，综述总结了主要的规划公式、代表性算法、优点和局限性。此外，综述分析了环境知识、工作空间几何、机器人约束、感知目标和协调需求如何影响CPP问题。该综述进一步讨论了可扩展在线规划、多机器人协调、三维和视觉覆盖、统一平台受限和资源感知覆盖以及增强学习覆盖等开放挑战。因此，该综述提供了近期CPP发展的结构化概述和未来研究方向。
cs.RO / 47 / 2607.10655
Artificial Foveated Perception for Mitigating Shortcut Learning in Robotic Foundation Models
人工注视感知以减轻机器人基础模型中的捷径学习
Abstract
Robotic foundation models have recently made substantial progress in multi-task capability, cross-embodiment transfer, and language-conditioned control. Yet robust deployment across diverse real-world settings remains difficult, in part because policies often fail to distinguish causally relevant visual structure from spurious scene-level correlations. We identify this failure mode as shortcut learning: the tendency to exploit predictive but non-causal correlations in the training distribution rather than the task-relevant visual evidence that determines successful action. Although shortcut learning has been extensively studied in computer vision and broader machine learning, its role in robotic foundation models remains comparatively underexplored. We propose Artificial Foveated Perception (AFP), a lightweight, policy-agnostic module that takes the same vision and language inputs as Vision-Language-Action and World Action Model pipelines and predicts task-conditioned masks over relevant objects, the robot, and other action-critical regions. We use these masks primarily as an auxiliary grounding signal during fine-tuning, aligning policy attention with task-relevant regions while leaving the core architecture unchanged. After fine-tuning, the policy executes on the original observation stream without requiring AFP in the control loop. We evaluate AFP across state-of-the-art robotic foundation models and show that foveated perception reduces fine-tuning time, suppresses overfitting, and improves generalization under environmental perturbations. Ablations over mask quality and grounding-loss design further show that these gains arise from directing policy learning toward task-relevant visual evidence. These results suggest that task-conditioned foveated perception is a practical mechanism for making robotic foundation models more robust, data-efficient, and scalable.
Chinese Translation
机器人基础模型最近在多任务能力、跨实体转移和语言条件控制方面取得了显著进展。然而，在多样化的现实环境中实现稳健部署仍然困难，部分原因是策略往往无法区分因果相关的视觉结构与虚假的场景级相关性。我们将这种失效模式称为捷径学习：即倾向于利用训练分布中预测性但非因果的相关性，而非决定成功行动的任务相关视觉证据。尽管捷径学习在计算机视觉和更广泛的机器学习中得到了广泛研究，但其在机器人基础模型中的作用仍然相对未被深入探讨。我们提出了人工注视感知（Artificial Foveated Perception, AFP），这是一种轻量级、与策略无关的模块，它接受与视觉-语言-行动和世界行动模型管道相同的视觉和语言输入，并预测与任务相关的物体、机器人及其他行动关键区域的条件掩码。我们主要将这些掩码作为微调过程中的辅助基础信号，旨在将策略注意力与任务相关区域对齐，同时保持核心架构不变。微调后，策略在原始观察流上执行，而无需在控制循环中使用 AFP。我们在最先进的机器人基础模型上评估 AFP，结果表明，注视感知减少了微调时间，抑制了过拟合，并改善了在环境扰动下的泛化能力。对掩码质量和基础损失设计的消融实验进一步表明，这些收益源于将策略学习引导至任务相关的视觉证据。这些结果表明，任务条件的注视感知是一种实用机制，可以使机器人基础模型更加稳健、数据高效和可扩展。
cs.RO / 48 / 2607.10682
SensorPerch: Sense Wherever and Whenever it Matters
SensorPerch：在关键时刻随时随地感知
Abstract
Existing robotic perception is constrained by sensors that are either robot-mounted or permanently fixed in the environment, locking perception to a limited set of viewpoints. Yet as robots perform increasingly diverse tasks, the most informative viewpoint shifts from one task to the next-often somewhere onboard sensor and static infrastructure can not readily satisfy. To address this gap, we propose SensorPerch, a novel realization of active perception that decouples sensing from both the robot embodiment and the environment by treating sensors as independent physical entities that the robot can autonomously detach and re-attach within the environment. SensorPerch presents one realization of this paradigm: a lightweight, wireless, reconfigurable sensor platform that can perch on diverse surfaces, paired with a viewpoint-selection framework that determines task-optimal sensor placements. Together, these enable robots to construct task-relevant viewpoints on demand, independent of the robot's current position and available fixed infrastructure. We demonstrate the paradigm on two task classes: (i) object-coupled perception, where SensorPerch enables persistent object-state detection beyond the robot's current position, achieving successful event detection even when the robot is not nearby; and (ii) policy-coupled perception, where SensorPerch allows robots to construct diverse, policy-specific viewpoints for various policies, achieving success rates comparable to those obtained using oracle viewpoints.
Chinese Translation
现有的机器人感知受到传感器的限制，这些传感器要么安装在机器人上，要么永久固定在环境中，从而将感知锁定在有限的视角集合中。然而，随着机器人执行越来越多样化的任务，最具信息量的视角会随着任务的不同而变化——通常是在机器人和静态基础设施无法满足的地方。为了解决这一问题，我们提出了SensorPerch，一种新颖的主动感知实现，它通过将传感器视为独立的物理实体，使得感知与机器人本体和环境解耦，机器人可以自主地在环境中拆卸和重新安装传感器。SensorPerch展示了这一范式的一种实现：一个轻量级、无线、可重构的传感器平台，可以停留在各种表面上，并配备一个视角选择框架，以确定任务最优的传感器位置。这些功能使得机器人能够根据需要构建与任务相关的视角，而不受机器人当前的位置和可用固定基础设施的限制。我们在两个任务类别上演示了这一范式：（i）物体耦合感知，其中SensorPerch使得在机器人当前位置信息之外持续检测物体状态成为可能，即使机器人不在附近也能成功检测事件；（ii）策略耦合感知，其中SensorPerch允许机器人为各种策略构建多样的、特定于策略的视角，其成功率与使用oracle视角获得的成功率相当。
cs.RO / 49 / 2607.10706
Action Map Policy: Learning 3D Closed-loop Manipulation via Pixel Classification
动作映射策略：通过像素分类学习三维闭环操作
Abstract
The action space poses a major challenge in robot learning, since it is often high-dimensional, can span long time horizons, and frequently admits multi-modal optimal solutions. A good choice of action representation and loss function can help to address these concerns, but there are often trade offs. We propose Action Map Policy (AMP), which casts 3D closed-loop manipulation policy learning as a classification problem in image space. While classification has been an effective formulation in generative language models, applying it to robot action learning is difficult because naively discretizing high-dimensional continuous actions explodes the token vocabulary. Our key idea is to project 3D actions onto the camera image planes and treat each pixel location as a discrete class, thus controlling dimensionality while retaining multi-modality. This method supports millimeter-level precision for high-dimensional actions without requiring a prohibitively large vocabulary, while preserving fine-grained pixel-wise visual signals. Furthermore, it can predict the entire action chunk in a single forward pass, avoiding complex noise scheduling and iterative denoising while achieving substantially faster inference than diffusion policies. Experiments on various manipulation tasks show that AMP outperforms strong baselines, achieving higher success rates, faster inference, and enhanced spatial reasoning.
Chinese Translation
动作空间在机器人学习中构成了重大挑战，因为它通常是高维的，能够跨越较长的时间范围，并且经常存在多模态的最优解。良好的动作表示和损失函数的选择可以帮助解决这些问题，但往往存在权衡。我们提出了动作映射策略（Action Map Policy, AMP），将三维闭环操作策略学习视为图像空间中的分类问题。尽管分类在生成语言模型中是一种有效的表述，但将其应用于机器人动作学习却很困难，因为简单地离散化高维连续动作会导致词汇量的爆炸。我们的关键思路是将三维动作投影到相机图像平面上，并将每个像素位置视为一个离散类别，从而在保持多模态的同时控制维度。这种方法支持高维动作的毫米级精度，而无需一个过于庞大的词汇量，同时保留细粒度的像素级视觉信号。此外，它可以在单次前向传播中预测整个动作块，避免复杂的噪声调度和迭代去噪，同时实现比扩散策略显著更快的推理速度。在各种操作任务上的实验表明，AMP优于强基线，达到了更高的成功率、更快的推理速度和增强的空间推理能力。
cs.RO / 50 / 2607.10815
Learning Roller-Skating Motions of Humanoid Robots Based on Adversarial Motion Priors
基于对抗运动先验的类人机器人滑轮滑行动作学习
Abstract
Humanoid roller-skating is difficult because the robot must coordinate whole-body balance, rolling contacts, and velocity-dependent posture regulation. This paper presents an adversarial motion prior based reinforcement learning framework for two humanoid roller-skating gaits: Pump Glide skating and Push Glide skating. The two gait datasets are collected independently through motion capture and retargeted to the humanoid robot separately. The retargeted data are then smoothed and resampled into reference motion states for AMP training. The two gaits are learned by independent AMP training pipelines with separate reference datasets, separate policies, and independent reward architectures. Simulation experiments are designed to evaluate gait quality, velocity tracking, turning, and gait-specific reward ablations.
Chinese Translation
类人滑轮滑行是一项困难的任务，因为机器人必须协调全身平衡、滚动接触和速度依赖的姿态调节。本文提出了一种基于对抗运动先验的强化学习框架，用于两种类人滑轮滑行步态：泵滑（Pump Glide）和推滑（Push Glide）。这两种步态数据集通过动作捕捉独立收集，并分别重新定向到类人机器人。重新定向的数据随后被平滑处理并重新采样为AMP训练的参考运动状态。这两种步态通过独立的AMP训练流程进行学习，使用不同的参考数据集、不同的策略和独立的奖励架构。设计了仿真实验以评估步态质量、速度跟踪、转向以及步态特定的奖励消融实验。
cs.RO / 51 / 2607.10842
D-SafeMPC: Diffusion-Driven Safe Model Predictive Control with Discrete-Time Control Barrier Functions
D-SafeMPC：基于扩散驱动的安全模型预测控制与离散时间控制障碍函数
Abstract
A key limitation on the use of diffusion models in robotic planning is their inability to inherently enforce safety or dynamical constraints, which often results in physically infeasible or unsafe outputs. Hybrid approaches that employ model predictive control (MPC) to address this problem can be unstable, as poor trajectory initializations from the diffusion model prevent the MPC from converging to a safe and feasible solution. To overcome these challenges, we propose D-SafeMPC, which enhances the interaction between diffusion and control. Our method guides the reverse diffusion process with control barrier functions (CBFs) and control Lyapunov functions (CLFs) and employs an iterative-projection scheme where an MPC refines the trajectory at each denoising step. This steers sampling toward safe, goal-directed regions and provides reliable MPC warm starts. In simulations on a Franka manipulator across four scenarios (one static-obstacle and three dynamic-obstacle settings) and in a sim-to-real experiment on a physical Franka robot, D-SafeMPC improves safety, task success rates, and planning efficiency over state-of-the-art baselines. To facilitate reproducibility, our source code and experimental configurations are available in a repository at https://github.com/erdiphd/D-SafeMPC
Chinese Translation
扩散模型在机器人规划中的一个关键限制是其无法固有地强制执行安全或动态约束，这常常导致物理上不可行或不安全的输出。采用模型预测控制（MPC）来解决这一问题的混合方法可能不稳定，因为扩散模型的糟糕轨迹初始化会阻止MPC收敛到安全和可行的解决方案。为克服这些挑战，我们提出了D-SafeMPC，它增强了扩散与控制之间的互动。我们的方法通过控制障碍函数（CBFs）和控制李雅普诺夫函数（CLFs）引导反向扩散过程，并采用迭代投影方案，在每个去噪步骤中通过MPC细化轨迹。这将采样引导至安全、目标导向的区域，并提供可靠的MPC热启动。在对Franka机械手进行的四种场景（一个静态障碍和三个动态障碍设置）的仿真以及在物理Franka机器人上的仿真到现实实验中，D-SafeMPC在安全性、任务成功率和规划效率方面优于最先进的基线。为了促进可重复性，我们的源代码和实验配置可在https://github.com/erdiphd/D-SafeMPC的仓库中获取。
cs.RO / 52 / 2607.10879
3D Scene Graph Prediction: Generating Hierarchical Models from Partially Observed Environments
3D场景图预测：从部分观察环境生成层次模型
Abstract
Generating realistic 3D indoor scenes is an area of growing interest in computer vision and robotics. Existing methods, often motivated by applications such as interior design, generally focus on object layout generation within a single room. The generation of high-level scene structure, such as room-level layout and traversability, remains underexplored despite its importance for robotics applications. In this paper, we consider the case where a robot has explored part of an environment and needs to predict the unexplored parts to support downstream tasks such as exploration or object search. We propose a top-down framework for synthesizing hierarchical 3D scene graphs, including a room layer -- describing the floor plan and traversability -- and an object layer modeling object layouts within each room. For the room layer, we propose a novel mixed-domain graph diffusion model jointly predicting room categories, floor boundaries, and traversability between rooms. Via corruption and masking, this model supports partial constraints such as incomplete floor plans, avoiding the need for partially observed training data. For the object layer, we integrate an existing mixed discrete-continuous diffusion model for joint prediction of object categories, locations, sizes, and orientations within each room given the floor plan. We compare our method with state-of-the-art occupancy-based and LLM-based floor plan generation methods on a standard benchmark. Compared with an occupancy-based learning baseline, our method generalizes substantially better to out-of-distribution partial floor plans. We also demonstrate our integrated prediction pipeline on real-world scenes from robot-collected data, enabling prediction beyond explored areas.
Chinese Translation
生成逼真的3D室内场景是计算机视觉和机器人领域日益受到关注的研究方向。现有方法通常受到室内设计等应用的启发，主要集中在单个房间内的物体布局生成上。尽管高层次场景结构（如房间级布局和可通行性）的生成对机器人应用至关重要，但仍然未得到充分探索。本文考虑一种情况，即机器人已探索环境的部分区域，并需要预测未探索部分以支持后续任务，如探索或物体搜索。我们提出了一种自上而下的框架，用于合成层次化的3D场景图，包括一个房间层——描述平面图和可通行性——以及一个物体层，建模每个房间内的物体布局。对于房间层，我们提出了一种新颖的混合域图扩散模型，联合预测房间类别、地面边界和房间间的可通行性。通过破坏和遮蔽，该模型支持部分约束，如不完整的平面图，避免了对部分观察训练数据的需求。对于物体层，我们整合了现有的混合离散-连续扩散模型，以在给定平面图的情况下，联合预测每个房间内的物体类别、位置、大小和方向。我们在标准基准上将我们的方法与最先进的基于占用率和基于大语言模型（LLM）的平面图生成方法进行了比较。与基于占用率的学习基线相比，我们的方法在分布外的部分平面图上具有更好的泛化能力。我们还在机器人收集的数据的真实场景上演示了我们的集成预测管道，使得预测超越已探索区域成为可能。
cs.RO / 53 / 2607.10892
A Single Diffusion-Policy Controller for Multi-Task Block Pushing with Zero-Shot Sim-to-Real Transfer
用于多任务方块推动的单一扩散策略控制器，具备零样本模拟到现实转移能力
Abstract
Diffusion policies have shown promising empirical performance in representing and learning complex maneuvers for robots using behavior cloning (BC). In this paper, we explore training diffusion policies from scratch using reinforcement learning (RL) for multi-task robotic manipulation. Specifically, we aim to train a single diffusion policy for block-pushing tasks with multiple shapes. The proposed framework features a simple policy loss function, which is a reweighted evidence lower bound used in BC-based diffusion policy training and can seamlessly serve as the policy learning module in RL algorithms. To address the exploration challenges arising from the absence of demonstrations, we incorporate reverse curriculum generation and objective-centric representations. Combined with the expressiveness of diffusion policies, our design supports learning of multi-task block-pushing policies in our sparse-reward simulation setting. We further evaluate whether the trained diffusion policy transfers in zero-shot to real-world tasks under varying environmental conditions including goal positions, block shapes, block weights and surface friction, providing evidence that this pipeline can transfer to our real-world block-pushing setup under the tested variations.
Chinese Translation
扩散策略在使用行为克隆（BC）进行复杂动作的表示和学习方面表现出了良好的经验性能。本文探讨了从零开始使用强化学习（RL）训练扩散策略以实现多任务机器人操作。具体而言，我们旨在为具有多种形状的方块推动任务训练一个单一的扩散策略。所提出的框架具有简单的策略损失函数，该函数是用于基于BC的扩散策略训练的重加权证据下界，并可以无缝地作为RL算法中的策略学习模块。为了解决由于缺乏示范而带来的探索挑战，我们结合了反向课程生成和目标中心表示。结合扩散策略的表现力，我们的设计支持在稀疏奖励的仿真环境中学习多任务方块推动策略。我们进一步评估训练好的扩散策略是否能够在不同环境条件下（包括目标位置、方块形状、方块重量和表面摩擦）零样本转移到现实世界任务，提供证据表明该流程能够在测试的变化下转移到我们的现实世界方块推动设置中。
cs.RO / 54 / 2607.10925
Mapping Pamir: Multi-Session Visual-Inertial SLAM and 3D Reconstruction of an Underwater Shipwreck
帕米尔映射：多会话视觉惯性SLAM与水下船舶残骸的三维重建
Abstract
This paper presents a framework for multi-session mapping of underwater environments utilizing an affordable action camera. The Visual-Inertial data are augmented by water depth recordings from a dive computer. SVIn2, an open-source VI-SLAM framework, is utilized to generate a trajectory and a sparse reconstruction for each session. Utilizing the keyframes extracted from SVIn2 and the estimated camera poses, a Structure-from-Motion (SfM) framework, COLMAP, is employed for global optimization and to produce a dense reconstruction of the target environment. The presence of calibration targets at fixed locations, when available, is used to estimate the coordinate transformation between different data collection sessions, thus transforming the different sessions into the same coordinate frame. The proposed pipeline is employed for the mapping of a shipwreck off the coast of Barbados. For the first time, both the exterior and the accessible interior parts of the wreck were mapped in two sessions, while a third session employed two cameras with different fields of view.
Chinese Translation
本文提出了一种利用经济实惠的运动相机进行水下环境多会话映射的框架。视觉-惯性数据通过潜水计算机的水深记录进行增强。采用开源的VI-SLAM框架SVIn2生成每个会话的轨迹和稀疏重建。利用从SVIn2提取的关键帧和估计的相机姿态，采用结构从运动（Structure-from-Motion, SfM）框架COLMAP进行全局优化，并生成目标环境的密集重建。当可用时，固定位置的标定目标的存在用于估计不同数据采集会话之间的坐标变换，从而将不同会话转换到同一坐标系。所提出的流程用于对巴巴多斯海岸附近的一处船舶残骸进行映射。首次在两个会话中同时映射了残骸的外部和可进入的内部部分，而第三个会话则使用了两个具有不同视场的相机。
cs.RO / 55 / 2607.10975
Real-Time Rulebook-Aware Nonlinear MPC for Autonomous Driving with Priority-Biased Tiered Slacks
实时规则感知非线性模型预测控制器用于具有优先级偏向分层松弛的自主驾驶
Abstract
Autonomous-vehicle motion planners must resolve conflicts among safety, regulation, comfort, and efficiency in real time while exposing those decisions for audit. We present W-SQP, a weighted tiered-slack nonlinear model predictive controller (NMPC) that compiles nine driving-rule families into a four-tier shared-slack nonlinear program solved online with CasADi and IPOPT; the name denotes the weighted quadratic slack penalty, not a sequential-quadratic-programming solver. Strongly separated tier penalties bias residual violations toward lower-priority rules while leaving actuation bounds hard. The controller replans from its executed state at $10$\,Hz and records per-rule residuals on every cycle. A $90$\,ms solver-time limit returns an anytime iterate that is projected through the vehicle dynamics before execution; median and maximum observed wall-clock solve times were $28$ and $104$\,ms. We evaluate W-SQP in closed loop on 150 Waymo Open Motion Dataset scenarios in Waymax against reactive and proposal-and-select baselines, and introduce a log-independent protocol that separates safety and regulatory compliance from resemblance to the recorded human trajectory. Under this protocol, W-SQP shows no systematic group-level deficit relative to expert replay on the log-independent safety and regulatory rules, with several localized regressions in the hardest, highest-divergence scenarios. The results characterize W-SQP as an auditable, priority-biased, anytime-capable NMPC prototype rather than a hard-real-time or formally safe controller.
Chinese Translation
自主车辆的运动规划者必须在实时中解决安全、法规、舒适性和效率之间的冲突，同时对这些决策进行审计。我们提出了 W-SQP，一种加权分层松弛非线性模型预测控制器（NMPC），它将九个驾驶规则家族编译成一个四层共享松弛的非线性规划，并通过 CasADi 和 IPOPT 在线求解；该名称表示加权二次松弛惩罚，而不是序列二次规划求解器。强分离的层惩罚将剩余违规偏向于低优先级规则，同时保持执行约束的严格性。控制器以 10 Hz 的频率从其执行状态重新规划，并在每个周期记录每条规则的剩余值。90 毫秒的求解时间限制返回一个随时可用的迭代解，该解在执行前通过车辆动力学进行投影；观察到的中位数和最大墙钟求解时间分别为 28 毫秒和 104 毫秒。我们在 Waymax 上的 150 个 Waymo Open Motion Dataset 场景中对 W-SQP 进行了闭环评估，与反应式和提议-选择基线进行了比较，并引入了一种与日志无关的协议，该协议将安全和法规合规性与记录的人类轨迹的相似性分开。在该协议下，W-SQP 在与专家重放的日志无关的安全和法规规则的比较中，未显示出系统性的群体级缺陷，但在最困难、最高偏差的场景中出现了一些局部回归。结果将 W-SQP 特征化为一个可审计的、优先级偏向的、随时可用的 NMPC 原型，而不是一个硬实时或正式安全的控制器。
cs.RO / 56 / 2607.10991
Think When It Matters: Conditional VLM Reasoning for Social Navigation with RL Policies
在关键时刻思考：基于条件的视觉-语言模型推理用于强化学习策略的社会导航
Abstract
As mobile robots become more integrated into everyday human environments, social robot navigation is becoming essential for ensuring human comfort, safety, and trust. While reinforcement learning (RL) navigation policies provide the fast inference and reactive behavior necessary for real-time deployment, they still lack flexible semantic reasoning capabilities and often fail to generalize to complex social scenarios. Recent approaches have increasingly turned to vision-language models (VLMs) in place of RL policies to improve semantic and social reasoning in robot navigation. Nevertheless, their high computational cost and slow inference remain major barriers to real-time deployment. To overcome these limitations, we introduce HUMA (Hybrid Understanding for Multi-modal social Navigation), a hybrid architecture that dynamically balances the computational efficiency of RL policies with the deep semantic understanding of VLMs. Our approach uses a reactive RL policy to handle low-density, routine navigation tasks, while conditioning it on a post-trained high-level VLM when a human enters sensitive situations, such as the robot's proximity zone. We evaluate HUMA on the Social-MP3D and Social-HM3D benchmarks, where it achieves task success improvements of 20% and 3%, respectively, while significantly reducing personal space violations and human collisions against state-of-the-art baselines. Extensive ablation studies validate each architectural component, and real-world deployment on the Miroka\"i mobile robot further demonstrates the practical viability of our approach.
Chinese Translation
随着移动机器人越来越多地融入日常人类环境，社会机器人导航对于确保人类的舒适、安全和信任变得至关重要。尽管强化学习（RL）导航策略提供了实时部署所需的快速推理和反应行为，但它们仍然缺乏灵活的语义推理能力，并且往往无法在复杂的社会场景中进行泛化。最近的研究越来越多地转向视觉-语言模型（VLM）来替代RL策略，以改善机器人导航中的语义和社会推理。然而，它们的高计算成本和缓慢推理仍然是实时部署的主要障碍。为了解决这些限制，我们提出了HUMA（多模态社会导航的混合理解），这是一种混合架构，动态平衡RL策略的计算效率与VLM的深度语义理解。我们的方法使用反应式RL策略处理低密度的常规导航任务，同时在发生人类进入敏感情况（例如机器人接近区域）时，基于后训练的高层VLM进行条件调整。我们在Social-MP3D和Social-HM3D基准上评估HUMA，分别实现了20%和3%的任务成功率提升，同时显著减少了个人空间侵犯和人类碰撞，相较于最先进的基线。大量的消融研究验证了每个架构组件的有效性，并且在Miroka"i移动机器人上的实际部署进一步证明了我们方法的实用性。
cs.RO / 57 / 2607.10999
Wearing A Coat: Dual-Arm Robot-Assisted Dressing with Differentiable Clothing Simulation
穿上外套：双臂机器人辅助穿衣的可微分服装模拟
Abstract
The development of assistive robots for dressing tasks serves to augment human convenience and improve the quality of life for individuals with physical impairments. However, due to the intricate contact interactions between garments and the human limbs during dressing, most robot-assisted dressing algorithms treat clothing as an assembly of discrete segments, thereby struggling to manage the partial worn garments under contact constraints. To overcome this challenge, we propose a novel robotic dressing control algorithm that integrates realtime differentiable clothing simulation. The simulation algorithm employs explicit iterative scheme with intentionally introduced higher-order perturbations to enhance computational efficiency while maintaining stability under large time-step conditions. Through simulation, we resolve the garment state under contact constraints, which then enables a multi-phase control strategy for successful coat dressing assistance. To further improve real-time performance, we introduce a constrained local model along with its corresponding optimization solver, permitting high-frequency local compensation for the differentiable simulation based global controller. Finally, we experimentally validate our approach through both simulated and physical dressing scenarios, conclusively demonstrating its feasibility and efficacy
Chinese Translation
为穿衣任务开发辅助机器人旨在提升人类便利性，并改善身体残疾人士的生活质量。然而，由于在穿衣过程中服装与人体肢体之间复杂的接触交互，大多数机器人辅助穿衣算法将服装视为离散段的组合，从而在接触约束下难以处理部分穿着的服装。为了解决这一挑战，我们提出了一种新颖的机器人穿衣控制算法，该算法集成了实时可微分服装模拟。该模拟算法采用显式迭代方案，并故意引入高阶扰动，以提高计算效率，同时在大时间步条件下保持稳定性。通过模拟，我们解决了在接触约束下的服装状态，从而实现了成功的外套穿着辅助的多阶段控制策略。为了进一步提高实时性能，我们引入了一个受限局部模型及其相应的优化求解器，允许对可微分模拟基础上的全局控制器进行高频局部补偿。最后，我们通过模拟和实际穿衣场景对我们的方法进行了实验验证，明确证明了其可行性和有效性。
cs.RO / 58 / 2607.11004
Affordance-Based Manipulation Planning with Text Goals and Sim-to-Real Generalisation via Real-to-Sim Image Conversion
基于可供性的操控规划与文本目标的实现及通过真实到仿真图像转换的仿真到现实泛化
Abstract
We present a manipulation planning system based on affordance recognition and action effect prediction. The system reasons through possible futures in visual form, and evaluates candidate plans by agreement of predicted outcomes with text-based goals set at run-time, using a multi-modal goal-matching module. Positions of objects named in the goal text are tracked through predictions even when occluded, making it possible to generate action plans even when objects become occluded, or when their initial descriptors cease to identify them in future states. We further expand the system with an image conversion module for translating real-world state images with objects of varied shapes and visual appearances into a consistent visual appearance, to facilitate manipulation planning in a physical robot setup. We evaluate performance of the system's modules in isolation and demonstrate the integrated system's manipulation planning capabilities on a set of challenging tasks in both simulation and on hardware.
Chinese Translation
我们提出了一种基于可供性识别和动作效果预测的操控规划系统。该系统通过视觉形式推理可能的未来，并通过多模态目标匹配模块评估候选计划与运行时设定的基于文本的目标之间的预测结果一致性。即使在被遮挡的情况下，目标文本中提到的物体位置也能通过预测进行跟踪，从而使得即使在物体被遮挡或其初始描述符在未来状态中无法识别的情况下，也能生成动作计划。我们进一步扩展了该系统，增加了一个图像转换模块，用于将具有不同形状和视觉外观的真实世界状态图像转换为一致的视觉外观，以促进物理机器人设置中的操控规划。我们在隔离条件下评估了系统各模块的性能，并展示了集成系统在仿真和硬件上对一系列挑战性任务的操控规划能力。
cs.RO / 59 / 2607.11018
Whole-Body Semantic-to-Actuation Grounding of Elephant-Inspired Soft-Trunk Motion via Lightweight Flow Matching
基于轻量级流匹配的象 inspired 软树干运动的全身语义到驱动的基础
Abstract
For close-contact human-robot interaction (HRI), trunk-like continuum manipulators provide a physical channel for diverse whole-body expression, but grounding open-vocabulary responses into such robots is difficult: end-effector motion underspecifies body shape, whereas direct whole-body commands are high-dimensional and hard to keep feasible. We propose a whole-body semantic-to-actuation grounding framework for elephant-inspired soft-trunk HRI based on lightweight flow matching. The framework converts responses from a multimodal large language model into bounded, morphology-aligned intent-intensity tuples, parameterizes tendon-actuation trajectories with compact Catmull-Rom spline controls, and uses a rectified-flow generator to sample feasible whole-body trunk motions. Experiments show that the proposed framework improves held-out grounding correctness from 25.0% to 77.2% over a raw-response dense-regression baseline. Compared with a denoising-diffusion baseline, it improves correctness from 71.9% to 77.2% and reduces inference time from 7.86 ms to 4.87 ms while preserving motion diversity. A 100-participant physical HRI study further shows that adding the generated soft-trunk motion channel increases the positive overall-satisfaction rating from 46% to 82% over the audiovisual-only baseline.
Chinese Translation
对于近距离人机交互（HRI），树干状连续操纵器提供了多样化全身表达的物理通道，但将开放词汇响应嵌入此类机器人中是困难的：末端执行器的运动不足以指定身体形状，而直接的全身指令则是高维的，难以保持可行性。我们提出了一种基于轻量级流匹配的象 inspired 软树干 HRI 的全身语义到驱动的基础框架。该框架将来自多模态大型语言模型的响应转换为有界的、形态对齐的意图-强度元组，用紧凑的 Catmull-Rom 样条控制参数化肌腱驱动轨迹，并使用整流流生成器来采样可行的全身树干运动。实验表明，所提出的框架将保留的基础正确性从 25.0% 提高到 77.2%，相较于原始响应密集回归基线。与去噪扩散基线相比，其正确性从 71.9% 提高到 77.2%，推理时间从 7.86 毫秒减少到 4.87 毫秒，同时保持运动多样性。一项包含 100 名参与者的物理 HRI 研究进一步表明，增加生成的软树干运动通道使整体满意度评分从 46% 提高到 82%，相较于仅有视听的基线。
cs.RO / 60 / 2607.11027
SegDiff: Segmented Trajectory Diffusion for Consistent and Adaptive Robot Manipulation
SegDiff：用于一致性和自适应机器人操控的分段轨迹扩散
Abstract
Imitation learning enables robots to acquire manipulation skills from demonstrations by mapping observations to actions. Existing approaches predict either short-horizon continuous action sequences or discrete keyposes. However, continuous prediction methods suffer from compounding errors due to short prediction horizons and struggle with multi-modal action distributions, whereas keypose-based methods necessitate an external planner, constraining real-time applicability. To address these challenges, we introduce SegDiff, a closed-loop visuomotor policy that integrates the strengths of both paradigms. SegDiff decomposes demonstrations into motion segments between keyposes and learns to predict the continuous trajectory from the current state to the next keypose, enabling long-horizon prediction with real-time refinement. Furthermore, we leverage the capability of diffusion models and DDIM inversion to propose a Dynamic Temporal Ensembling mechanism, which allows the policy to efficiently respond to dynamic environments and mitigate discontinuities caused by inconsistent multi-modal sampling. SegDiff demonstrates significant performance gains over existing approaches across various simulated and real-world scenarios, indicating its strong ability to reason over extended temporal dependencies while maintaining real-time adaptability and control stability.
Chinese Translation
模仿学习使机器人能够通过将观察映射到动作来从示范中获取操控技能。现有方法要么预测短时间范围内的连续动作序列，要么预测离散的关键姿态。然而，连续预测方法由于短预测范围而遭受累积误差，并且在多模态动作分布上表现不佳，而基于关键姿态的方法则需要外部规划器，限制了其实时应用性。为了解决这些挑战，我们提出了SegDiff，一种闭环视觉运动策略，整合了两种范式的优点。SegDiff将示范分解为关键姿态之间的运动段，并学习从当前状态到下一个关键姿态的连续轨迹预测，从而实现长时间范围的预测和实时优化。此外，我们利用扩散模型和DDIM反演的能力，提出了一种动态时间集成机制，使得策略能够有效应对动态环境，并减轻因不一致的多模态采样造成的间断性。SegDiff在各种模拟和真实场景中表现出显著的性能提升，表明其在推理扩展时间依赖性方面的强大能力，同时保持实时适应性和控制稳定性。
cs.RO / 61 / 2607.11029
Learning to Navigate Efficiently with Only 0.58M Trainable Parameters
仅用 0.58M 可训练参数高效导航的学习
Abstract
Recent progress in visual navigation has largely been driven by scale: end-to-end policies with hundreds of millions of parameters trained on billions of frames or large-scale simulated data. We ask how much of this scale a single task family actually requires, and what structure can substitute for it. We propose a decomposed navigation model in which operations with known closed-form structure, such as projective geometry, occupancy, and coordinate transforms, are computed analytically and serve as interfaces between three small learned modules: an egress predictor that grounds the episode goal as a local subgoal in the current view, a navigation predictor that estimates a goal-conditioned posterior over where trajectories travel, and an endpoint-pinned residual diffusion generator that samples trajectory shapes from this posterior. The system trains only 0.58M out of a total of 22.7M parameters, on 44k frames in under one GPU-hour, yet approaches the performance of state-of-the-art models on navigation tasks across 6060 point-goal episodes and 60 environments, while having 233x fewer trainable parameters, the lowest collision rate among all evaluated methods, and 50 Hz inference speed. The decomposition further transfers to no-goal exploration by retraining only the 123k-parameter egress head, and its failure modes under sensor corruption are transparent and analytically correctable.
Chinese Translation
近年来，视觉导航的进展在很大程度上是由规模驱动的：端到端策略拥有数亿个参数，并在数十亿帧或大规模模拟数据上进行训练。我们探讨单一任务系列实际上需要多少规模，以及什么结构可以替代它。我们提出了一种分解导航模型，其中已知封闭形式结构的操作，如投影几何、占用和坐标变换，都是通过解析计算的，并作为三个小型学习模块之间的接口：一个出口预测器将情节目标作为当前视图中的局部子目标进行定位，一个导航预测器估计目标条件下的轨迹旅行后验，以及一个端点固定的残差扩散生成器从该后验中采样轨迹形状。该系统仅训练了 0.58M 个参数（总共 22.7M 个参数），在不到一个 GPU 小时内处理了 44k 帧，然而在 6060 个点目标情节和 60 个环境的导航任务中，其性能接近最先进模型，同时可训练参数少了 233 倍，且在所有评估方法中具有最低的碰撞率和 50 Hz 的推理速度。该分解进一步通过仅重新训练 123k 参数的出口头部转移到无目标探索，其在传感器损坏下的失效模式是透明且可解析的。
cs.RO / 62 / 2607.11031
GraspGraphNet: Graph-Structured Multi-Embodiment Dexterous Grasp Generation
GraspGraphNet：图结构多体现灵巧抓取生成
Abstract
Dexterous grasp generation across robot hands is challenging because hands differ in kinematic topology, actuation dimensions, and native command spaces. We introduce GraspGraphNet, a topology-aware grasp generation framework that represents each hand as a URDF-derived kinematic graph and directly generates executable palm poses and joint configurations. GraspGraphNet combines hierarchical object surface encoding, differentiable forward kinematics, and dynamic world-edge message passing to model evolving robot-object interactions. It applies conditional flow matching directly in executable palm-pose and joint-state space, avoiding post-processing optimization, inverse kinematics, and retargeting. Using a shared model trained on Barrett Hand, Allegro Hand, and Shadow Hand, GraspGraphNet achieves an average success rate of 83.48% with 40ms inference time per grasp on a 40-object benchmark. Without retraining, the same model achieves 72.70% success on controlled finger-removal variants, demonstrating robustness to hand-topology variations. These results suggest that graph-structured hand representations can effectively support dexterous grasp generation across robot hands with different kinematic structures. Project: https://lysees.github.io/graspgraphnet-page
Chinese Translation
在机器人手之间生成灵巧抓取是具有挑战性的，因为手的运动拓扑、驱动维度和固有指令空间各不相同。我们提出了GraspGraphNet，一种拓扑感知的抓取生成框架，它将每只手表示为一个基于URDF的运动图，并直接生成可执行的掌姿态和关节配置。GraspGraphNet结合了分层物体表面编码、可微分前向运动学和动态世界边缘消息传递，以建模不断变化的机器人-物体交互。它在可执行的掌姿态和关节状态空间中直接应用条件流匹配，避免了后处理优化、逆运动学和重新目标化。使用在Barrett手、Allegro手和Shadow手上训练的共享模型，GraspGraphNet在一个包含40个物体的基准测试中实现了83.48%的平均成功率，抓取推理时间为40毫秒。在不重新训练的情况下，同一模型在控制指移除变体上取得了72.70%的成功率，展示了对手部拓扑变化的鲁棒性。这些结果表明，图结构的手部表示可以有效支持具有不同运动结构的机器人手之间的灵巧抓取生成。项目网址：https://lysees.github.io/graspgraphnet-page
cs.RO / 63 / 2607.11041
PAKE: Learning Whole-Body Loco-Manipulation with Partial Kinematic Embeddings
PAKE：通过部分运动学嵌入学习全身运动操控
Abstract
Loco-manipulation has recently shown promising capabilities; however, achieving high-precision control, managing the high-dimensional action space induced by many degrees of freedom (DoFs), and fully exploiting the inherent redundancy of whole-body systems remain challenging. In this paper, we propose a novel whole-body control framework that effectively addresses these challenges by decomposing the complex loco-manipulation problem into partial reference motion generation and low-level imitation control. We introduce a new Kinematic Normalizing Flow (KNF) model, trained on a large-scale kinematic dataset, that generates diverse yet feasible partial reference motions. A high-level controller is then trained to navigate the KNF's latent space to exploit redundant solutions, while a low-level controller ensures physically feasible and accurate motion execution. We validate our approach on the quadrupedal robot equipped with a six-DoF robotic arm. In simulation, experimental results show that our approach significantly outperforms state-of-the-art methods in terms of tracking accuracy and feasible workspace coverage. For hardware deployment, we evaluate the system over 24 episodes across 8 different mobile loco-manipulation tasks. The system achieves end-effector pose-tracking errors of 4.5 cm and 0.14 rad, while maintaining accurate locomotion tracking with linear and angular velocity errors of 0.1 m/s and 0.01 rad/s, respectively, outperforming competitive baselines. Our method represents a practical and powerful solution for accurate and generalized whole-body loco-manipulation in high-DoF robotic systems, with promising potential for diverse downstream robotic tasks.
Chinese Translation
运动操控最近展现了令人期待的能力；然而，实现高精度控制、管理由多个自由度（DoFs）引起的高维动作空间，以及充分利用全身系统固有的冗余性仍然具有挑战性。本文提出了一种新颖的全身控制框架，有效地通过将复杂的运动操控问题分解为部分参考运动生成和低级模仿控制来应对这些挑战。我们引入了一种新的运动学归一化流（Kinematic Normalizing Flow, KNF）模型，该模型在大规模运动学数据集上训练，能够生成多样且可行的部分参考运动。随后，训练一个高级控制器以在KNF的潜在空间中导航，从而利用冗余解，而一个低级控制器则确保物理上可行且准确的运动执行。我们在配备六自由度机械臂的四足机器人上验证了我们的方法。在仿真中，实验结果表明，我们的方法在跟踪精度和可行工作空间覆盖方面显著优于最先进的方法。在硬件部署方面，我们在8个不同的移动运动操控任务中评估了系统的24个实验回合。该系统实现了末端执行器位姿跟踪误差为4.5厘米和0.14弧度，同时保持准确的运动跟踪，线性和角速度误差分别为0.1米/秒和0.01弧度/秒，超越了竞争基线。我们的方法代表了一种实用且强大的解决方案，适用于高自由度机器人系统中的准确和广泛的全身运动操控，并在多样的下游机器人任务中展现了良好的潜力。
cs.RO / 64 / 2607.11099
Desc++: Efficient Descriptor Enhancement for Data Association in Existing Visual SLAM Systems
Desc++：现有视觉SLAM系统中数据关联的高效描述符增强
Abstract
Reliable visual data association is fundamental to visual SLAM (V-SLAM), as it directly determines the quality of the camera pose estimation and map consistency. However, the handcrafted descriptors used by most mature real-time systems degrade under illumination and viewpoint changes, while learning-based front-ends that address this weakness typically require replacing the extraction-and-matching pipeline and introduce substantial computational overhead. Descriptor enhancement offers a compromise by refining existing descriptors within their original format, yet current methods rely on simplified attention mechanisms whose limited contextual modeling constrains the achievable matching quality. To resolve this trade-off between contextual expressiveness and efficiency, we propose Desc++, a lightweight enhancement module that jointly encodes descriptor representations and keypoint geometry and aggregates spatial context through a hybrid architecture that combines order-agnostic global attention with geometry-aware sequential modeling in linear time. The enhanced descriptors retain their original dimensionality and matching interface, enabling integration into deployed V-SLAM systems without modifying the pipeline. Experiments across descriptor matching, correspondence analysis, and system-level benchmarks with four different V-SLAM systems demonstrate that Desc++ improves matching accuracy over the state-of-the-art enhancement method, translates these gains into more accurate and stable trajectory estimation, and achieves a favorable balance between accuracy and efficiency for practical integration into existing real-time V-SLAM pipelines.
Chinese Translation
可靠的视觉数据关联是视觉SLAM（V-SLAM）的基础，因为它直接决定了相机位姿估计和地图一致性的质量。然而，大多数成熟的实时系统使用的手工描述符在光照和视角变化下表现不佳，而针对这一弱点的基于学习的前端通常需要替换提取和匹配流程，并引入大量的计算开销。描述符增强提供了一种折中方案，通过在其原始格式内精炼现有描述符来实现，但当前方法依赖于简化的注意力机制，其有限的上下文建模限制了可实现的匹配质量。为了解决上下文表达能力与效率之间的权衡，我们提出了Desc++，一个轻量级增强模块，它联合编码描述符表示和关键点几何，并通过一种混合架构聚合空间上下文，该架构结合了无序全局注意力与几何感知的序列建模，且在线性时间内完成。增强后的描述符保留了其原始维度和匹配接口，使其能够在不修改流程的情况下集成到已部署的V-SLAM系统中。通过对四种不同V-SLAM系统进行描述符匹配、对应分析和系统级基准测试的实验表明，Desc++在匹配准确性上优于最先进的增强方法，将这些增益转化为更准确和稳定的轨迹估计，并在准确性与效率之间实现了良好的平衡，以便于实际集成到现有的实时V-SLAM流程中。
cs.RO / 65 / 2607.11119
VIA: Visual Interface Agent for Robot Control
VIA：用于机器人控制的视觉接口代理
Abstract
Robot manipulation is a complex task that requires visual understanding, physical reasoning, planning, and closed-loop control. General-purpose foundation models (FMs) have grown remarkably capable of some of these, especially vision and reasoning. To leverage this for generalist robot policies, current methods typically involve converting existing FMs into vision-language-action (VLA) models by fine-tuning on robot data to output low-level actions. However, VLAs are often orders of magnitude smaller than frontier FMs given the limited data and compute available for fine-tuning, which in turn limits their general capability. Inspired by the growing ability of FMs to operate software through visual interfaces, we ask whether that same competence suffices to control a robot. We present VIA (Visual Interface Agent for robot control), a framework that recasts robot control as an agentic task: an off-the-shelf FM-powered agent drives a manipulator through a browser-based 3D interface by taking screenshots, issuing intuitive commands, observing the outcome, and adjusting. The agent receives no robot-specific fine-tuning and no access to privileged state information: it perceives visual input and acts through a small set of general tools. VIA inherits the agent's general reasoning, closed-loop error recovery, and ability to plan and re-plan from what it observes. It solves a diverse suite of tabletop manipulation tasks zero-shot with both Claude Code and Codex. With the strongest model (Fable 5) it achieves 96.7% success on three LIBERO-Goal tasks and 100% on a long-horizon rainbow assembly task. Performance improves with the scale and strength of the underlying model. These results suggest that frontier agents already possess skills that transfer directly to robot control given the right interface: your coding or computer-use agent is, in a sense, secretly a robot-control agent.
Chinese Translation
机器人操作是一项复杂的任务，涉及视觉理解、物理推理、规划和闭环控制。通用基础模型（FMs）在这些任务中，尤其是在视觉和推理方面，表现出显著的能力。为了将这一能力应用于通用机器人策略，目前的方法通常涉及将现有的FMs转换为视觉-语言-动作（VLA）模型，通过在机器人数据上进行微调以输出低级动作。然而，由于微调所需的数据和计算资源有限，VLAs的规模通常比前沿FMs小几个数量级，这反过来限制了它们的通用能力。受到FMs通过视觉接口操作软件能力日益增强的启发，我们提出了一个问题：这种能力是否足以控制机器人。我们提出了VIA（用于机器人控制的视觉接口代理），这是一个将机器人控制重新定义为代理任务的框架：一个基于现成FM的代理通过浏览器的3D接口驱动操纵器，方法是截取屏幕截图、发出直观命令、观察结果并进行调整。该代理没有进行任何特定于机器人的微调，也没有访问特权状态信息：它通过一小组通用工具感知视觉输入并采取行动。VIA继承了代理的通用推理、闭环错误恢复能力以及从观察中进行规划和重新规划的能力。它在零样本情况下解决了一系列多样的桌面操作任务，使用了Claude Code和Codex。使用最强的模型（Fable 5），它在三个LIBERO-Goal任务上达到了96.7%的成功率，在一个长时间跨度的彩虹组装任务上达到了100%的成功率。随着基础模型的规模和强度的增加，性能也有所提升。这些结果表明，前沿代理已经具备了可以直接转移到机器人控制的技能，只需合适的接口：从某种意义上说，你的编码或计算机使用代理，实际上是一个机器人控制代理。
cs.RO / 66 / 2607.11128
Comparison-Based Ordinal Learning for Proactive Driving Risk Assessment
基于比较的序数学习用于主动驾驶风险评估
Abstract
Real-time driving risk assessment provides an essential basis for proactive safety by identifying and quantifying the danger of ongoing road interactions before adverse outcomes occur. However, due to the scarcity of collision data and frame-level risk labels, existing driving risk assessment methods often rely on surrogate objectives, which may imperfectly align with true collision risk and not faithfully reflect the relative danger of driving interaction. This paper proposes a comparison-based ordinal risk learning framework that learns collision-relevant risk scores from pairwise supervision in driving data, directly modeling relative risk ordering without requiring numerical frame-level risk labels. We derive pairwise comparisons from three sources of event-structured driving data for such ordinal risk learning: temporal progression within safety-critical sequences, event-level contrast between dangerous and normal interactions, and physics-based counterfactual perturbations. On this basis, instantiations with three risk-scoring function parameterizations are implemented, including directly learning risk scores from comparison data, and aligning existing single or multiple surrogate-based risk models. The proposed framework is evaluated on the 100-Car and SHRP2 naturalistic driving datasets using a proactive collision warning task. Results show that the proposed framework improves high-recall risk discrimination, warning precision, and warning lead time over representative surrogate-based baselines across both in-distribution and out-of-distribution evaluations. These results suggest that the proposed framework can contribute to proactive safety research by providing more reliable risk assessment for automated driving systems and safety-critical driving interactions.
Chinese Translation
实时驾驶风险评估为主动安全提供了重要基础，通过在不良结果发生之前识别和量化正在进行的道路交互的危险性。然而，由于碰撞数据和帧级风险标签的稀缺，现有的驾驶风险评估方法往往依赖于替代目标，这可能与真实的碰撞风险不完全一致，并且未能真实反映驾驶交互的相对危险性。本文提出了一种基于比较的序数风险学习框架，该框架通过对驾驶数据的成对监督学习与碰撞相关的风险评分，直接建模相对风险排序，而无需数值帧级风险标签。我们从三种事件结构化的驾驶数据源中推导出成对比较，以进行这种序数风险学习：安全关键序列中的时间进展、危险与正常交互之间的事件级对比，以及基于物理的反事实扰动。在此基础上，实施了三种风险评分函数参数化的实例，包括直接从比较数据学习风险评分，以及对现有单一或多重基于替代的风险模型进行对齐。所提出的框架在100-Car和SHRP2自然驾驶数据集上进行了主动碰撞警告任务的评估。结果表明，所提出的框架在高召回风险区分、警告精度和警告提前时间方面优于代表性的基于替代的基线，无论是在分布内还是分布外的评估中。这些结果表明，所提出的框架可以通过为自动驾驶系统和安全关键驾驶交互提供更可靠的风险评估，促进主动安全研究。
cs.RO / 67 / 2607.11167
Pix2Act: Image-Space Manipulation Policies with Equivariant Augmentation
Pix2Act：具有等变增强的图像空间操作策略
Abstract
Representing manipulation actions as 2D trajectories in the camera plane provides a compact and interpretable basis for learning complex 3D manipulation policies. However, it also creates challenges from out-of-frame trajectories and limited precision. We propose Pix2Act, an imitation learning method that addresses these challenges by generating continuous image-space keypoint trajectories in each camera plane and losslessly recovering end-effector poses via triangulation. This reformulates high-dimensional 3D control as a simpler, more learnable 2D prediction problem. Crucially, it aligns observations and actions in the same coordinate space, enabling equivariant transformations to jointly rotate individual camera images together with their image-space actions. We analyze the symmetry properties of this augmentation and design a network architecture that can fuse multiple camera views while respecting their per-view rotations. As a result, Pix2Act implicitly enlarges the support of the data distribution and learns invariant action structures across transformations, yielding improved generalization and overall performance. Across diverse simulated and real-world manipulation tasks, Pix2Act outperforms state-of-the-art baselines and remains robust under camera perturbations.
Chinese Translation
将操作动作表示为相机平面中的二维轨迹为学习复杂的三维操作策略提供了一个紧凑且易于解释的基础。然而，这也带来了来自帧外轨迹和精度有限的挑战。我们提出了Pix2Act，一种模仿学习方法，通过在每个相机平面中生成连续的图像空间关键点轨迹，并通过三角测量无损恢复末端执行器姿态，从而解决这些挑战。这将高维三维控制重新构造成一个更简单、更易学习的二维预测问题。关键在于，它将观察和动作对齐在同一坐标空间中，使得等变变换能够共同旋转各个相机图像及其图像空间动作。我们分析了这种增强的对称性特性，并设计了一种网络架构，可以融合多个相机视图，同时尊重它们各自的视图旋转。因此，Pix2Act隐式地扩大了数据分布的支持，并学习了跨变换的不变动作结构，从而提高了泛化能力和整体性能。在各种模拟和真实世界的操作任务中，Pix2Act的表现优于最先进的基线，并在相机扰动下保持稳健。
cs.RO / 68 / 2607.11184
GeoGS-SLAM: Online Monocular Reconstruction Using Gaussian Splatting with Geometric Priors
GeoGS-SLAM：基于几何先验的高斯点云在线单目重建
Abstract
SLAM methods based on 3D Gaussian Splatting (3DGS) have demonstrated impressive tracking and mapping performance, but typically require additional geometric information from external depth sensors. Meanwhile, recent SLAM systems that leverage geometric priors from pre-trained feed-forward models enable real-time dense reconstruction, yet often discard original RGB information during optimization, thus degrading overall reconstruction quality. We present GeoGS-SLAM, an online monocular dense reconstruction system that combines the 3DGS-based map representation with learned geometric priors. Given uncalibrated RGB input, we first employ a feed-forward visual geometry model to predict camera and scene priors. The Gaussian scene map is then expanded by directly sampling Gaussian primitives from both RGB input and geometric priors. Camera poses and the scene map are jointly optimized through a coarse-to-fine strategy that minimizes both photometric and geometric losses. To ensure global consistency, we further incorporate online loop closure detection and pose graph optimization. Extensive experiments across indoor and outdoor benchmarks demonstrate that GeoGS-SLAM achieves superior rendering quality and tracking accuracy compared to state-of-the-art methods while maintaining online real-time performance. Project page: https://rlgao.github.io/geogs_slam.
Chinese Translation
基于3D高斯点云（3DGS）的SLAM方法展示了令人印象深刻的跟踪和地图构建性能，但通常需要来自外部深度传感器的额外几何信息。同时，最近的SLAM系统利用来自预训练前馈模型的几何先验实现实时稠密重建，但在优化过程中往往会丢弃原始RGB信息，从而降低整体重建质量。我们提出了GeoGS-SLAM，这是一种在线单目稠密重建系统，结合了基于3DGS的地图表示和学习的几何先验。给定未校准的RGB输入，我们首先采用前馈视觉几何模型来预测相机和场景先验。然后，通过直接从RGB输入和几何先验中采样高斯原语来扩展高斯场景地图。相机位姿和场景地图通过一种粗到细的策略共同优化，以最小化光度损失和几何损失。为了确保全局一致性，我们进一步结合了在线回环闭合检测和位姿图优化。在室内和室外基准测试中的大量实验表明，GeoGS-SLAM在渲染质量和跟踪精度上优于最先进的方法，同时保持在线实时性能。项目页面：https://rlgao.github.io/geogs_slam。
cs.RO / 69 / 2607.11204
Stop to Decide: Latency-Aware Proprioceptive Navigation Primitives for Mapping-Free Quadruped Inspection
停止决策：基于延迟感知的无地图四足机器人检查导航原语
Abstract
Compute-constrained quadrupeds often run their navigation loop far below the controller's design rate: sharing the onboard Jetson Orin with the vision pipeline slows our stair loop to about 15 Hz. This latency breaks a standard proprioceptive pattern: declaring stair-summit arrival from the body-pitch signal while still climbing. On a stepped platform whose 50 cm top is shorter than the robot (Unitree Go2, about 75 cm), in-motion detection overshoots the top edge with probability rising with the per-period advance v/f (the slowest about 15 Hz cell partly diluted by a separate non-arrival mode), whereas a climb-settle cadence holds overshoot near zero at every loop rate (pooled 22/45 vs 1/45 over about 30/20/15 Hz; Fisher p about 2.4e-7; 7/15 vs 0/15 at the deployed about 15 Hz). A logistic dose-response model in v/f captures the failure; a pre-specified 40 Hz out-of-sample test favours the protocol-clean fit (33% observed vs 43%/22% predicted), giving a deployment rule (critical loop rate about 19 Hz at 0.30 m/s). The detector sits in a fully onboard, mapping-free and learning-free stack: built-in inertial measurement unit, four foot-force channels, three 1-D ranges, one line camera, chaining line-following, a three-segment maneuver for 90-degree corners in a 55 cm corridor (20/20 contact-free vs 14/20 with 12 wall contacts for in-place yaw; exit-heading error 1.56 degrees vs 5.64 degrees), and stair traversal, completing the inspection course in 18/20 trials (90%). Results are from a single course geometry, platform, and operator.
Chinese Translation
计算资源受限的四足机器人通常以远低于控制器设计速率的频率运行其导航循环：与视觉处理管道共享板载的 Jetson Orin 导致我们的楼梯循环频率降至约 15 Hz。这种延迟打破了标准的本体感知模式：在仍在攀爬时通过身体俯仰信号声明到达楼梯顶端。在一个高度为 50 厘米、低于机器人（Unitree Go2，约 75 厘米）的阶梯平台上，运动检测以概率超出顶边，且该概率随着每周期的前进速度 v/f 增加（最慢约 15 Hz 的单元部分被单独的未到达模式稀释），而爬升-稳定节奏在每个循环速率下将超出保持在接近零（在约 30/20/15 Hz 下的合并数据为 22/45 对比 1/45；Fisher p 值约为 2.4e-7；在部署的约 15 Hz 下为 7/15 对比 0/15）。v/f 中的逻辑剂量反应模型捕捉了失败；一个预先指定的 40 Hz 的外样本测试支持协议清晰的拟合（观察到的 33% 对比预测的 43%/22%），给出了一个部署规则（临界循环速率约为 19 Hz，速度为 0.30 m/s）。检测器位于一个完全板载、无地图且无学习的堆栈中：内置的惯性测量单元、四个足部力量通道、三个一维测距、一个线性相机，链式跟踪线，三个段的机动用于 55 厘米走廊中的 90 度转角（20/20 无接触对比 14/20，带有 12 次墙接触用于原地偏航；出口航向误差为 1.56 度对比 5.64 度），以及楼梯穿越，在 18/20 次试验中完成检查课程（90%）。结果来自单一的课程几何、平台和操作员。
cs.RO / 70 / 2607.11270
Towards Predictive, Aligned, and Scalable Robot Learning
朝向预测性、对齐性和可扩展的机器人学习
Abstract
Learning, at its core, extends beyond memorization to the ability to reason and solve novel problems by navigating a space of possibilities. We introduce Lumo-2, a latent world-action model that generates actions by reasoning over world dynamics in latent space. The learned latent world dynamics capture physically grounded visual transitions, naturally encoding future possibilities and providing a unified substrate for cross-modal alignment. This formulation enables predictive reasoning akin to world modelling while remaining lightweight and focused on physical dynamics relevant to control. Central to our approach is the hypothesis that action generation quality is governed by the geometry of the latent space. We observe that standard reconstruction-based action tokenization objectives induce representations biased toward low-level signal fidelity, leading to misalignment between reconstruction quality and downstream control performance. To address this limitation, we propose a multi-stage modality pre-alignment strategy in which action representations are progressively aligned with latent world dynamics, vision, and language. This process enforces cross-modal consistency, promotes abstraction, and induces a structured latent space for predictive reasoning. We provide a systematic empirical study of latent world modelling and modality alignment, analyzing their roles in scaling laws and out-of-distribution generalization. Results show that Lumo-2 consistently outperforms strong vision-language-action (VLA) and world-action model (WAM) baselines, with gains on challenging real-world tasks requiring temporal reasoning, physical understanding, or high control complexity, including long-horizon and dexterous manipulation. These findings suggest that structured multimodal alignment and predictive reasoning are fundamental principles for advancing embodied intelligence.
Chinese Translation
学习的核心超越了单纯的记忆，具备了通过在可能性空间中导航来推理和解决新问题的能力。我们介绍了Lumo-2，一个潜在世界-动作模型，通过在潜在空间中推理世界动态来生成动作。学习到的潜在世界动态捕捉了物理基础的视觉转换，自然地编码了未来的可能性，并为跨模态对齐提供了统一的基础。这种表述使得预测性推理类似于世界建模，同时保持轻量化，专注于与控制相关的物理动态。我们的方法核心假设是，动作生成的质量受潜在空间几何的支配。我们观察到，基于标准重建的动作标记目标会导致偏向低级信号保真度的表征，从而导致重建质量与下游控制性能之间的不对齐。为了解决这一局限性，我们提出了一种多阶段模态预对齐策略，在该策略中，动作表征逐步与潜在世界动态、视觉和语言对齐。这个过程强制执行跨模态一致性，促进抽象，并诱导出一个结构化的潜在空间以进行预测性推理。我们提供了潜在世界建模和模态对齐的系统实证研究，分析它们在规模法则和分布外泛化中的作用。结果表明，Lumo-2在强大的视觉-语言-动作（VLA）和世界-动作模型（WAM）基准上始终表现优越，在需要时间推理、物理理解或高控制复杂度的挑战性现实任务中，包括长时间跨度和灵巧操作，均取得了显著的提升。这些发现表明，结构化的多模态对齐和预测性推理是推动具身智能发展的基本原则。
cs.RO / 71 / 2607.11340
CR-Solver: GPU-Accelerated Kinematics Solver for Tendon-driven Continuum Robots
CR-Solver：用于腱驱动连续机器人 的 GPU 加速运动学求解器
Abstract
Continuum robots provide intrinsic compliance, high dexterity, and safe physical interaction, enabling navigation and manipulation in confined and unstructured environments. Despite recent advances in sensing and control, heightening the need for precise motion generation, most widely used planning libraries are grounded in rigid-body assumptions, creating a critical gap for fast and practical tools for continuum robots. To address this, we present CR-Solver, a two-stage, optimization-based solver for the motion generation of tendon-driven continuum robots. Our method unifies inverse kinematics, path following, and trajectory planning within a single constrained nonlinear optimization framework. Leveraging GPU-accelerated parallel optimization, CR-Solver delivers fast, accurate, and constraint-aware solutions. We validate our approach on three tasks, demonstrating significant speedups over traditional CPU-based solvers while achieving a consistently high success rate above 95% and millimeter-level accuracy. The solver is implemented in pure Python, reducing the barrier to adoption and offering a practical, extensible foundation for continuum robots' high-performance motion planning.
Chinese Translation
连续机器人具有内在的柔顺性、高灵活性和安全的物理交互能力，使其能够在狭窄和非结构化环境中进行导航和操作。尽管在传感和控制方面取得了近期进展，提升了对精确运动生成的需求，但大多数广泛使用的规划库仍基于刚体假设，这为连续机器人提供快速实用工具造成了关键性缺口。为了解决这个问题，我们提出了 CR-Solver，这是一种基于优化的两阶段求解器，用于腱驱动连续机器人的运动生成。我们的方法在一个单一的约束非线性优化框架内统一了逆运动学、路径跟踪和轨迹规划。通过利用 GPU 加速的并行优化，CR-Solver 提供了快速、准确且考虑约束的解决方案。我们在三个任务上验证了我们的方法，展示了相较于传统的基于 CPU 的求解器显著的速度提升，同时成功率始终保持在 95% 以上，精度达到毫米级。该求解器采用纯 Python 实现，降低了采用门槛，并为连续机器人的高性能运动规划提供了一个实用且可扩展的基础。
cs.RO / 72 / 2607.11377
A Glimpse into Long-term Physical Coexistence with Intelligent Robots
与智能机器人长期物理共存的初步探索
Abstract
Long-term physical coexistence with intelligent robots requires more than capable robot policies. A persistent robotic assistant must support diverse user-facing interfaces, maintain long-horizon memory of people and preferences, coordinate across robot embodiments, and translate human intent into safe physical execution. We introduce PHILIA, a multi-robot agent built around a robot gateway abstraction. PHILIA retains the rich interaction and tool ecosystem of OpenClaw while exposing robot-local runtimes, onboard perception, navigation, speaker, and robot policies through a unified capability interface. This design decouples low-frequency, high-semantic agent reasoning from high-frequency, low-level robot execution, enabling plug-and-play integration of user interfaces, robot embodiments, and policy backends. As a result, the user experience becomes compositional: advances in user interfaces, robot embodiments, robot policies, navigation, or interaction algorithms can improve the overall experience without redesigning the system. We validate the architecture on Astribot S1 robots while designing the robot gateway contract to support future heterogeneous robot platforms through a shared capability interface for observation, task execution, navigation, speech playback, status monitoring, and task cancellation. We present representative use cases in which agent memory and scene understanding are grounded in robot actions. These span interactive household scenarios, ranging from simple organization to challenging long-horizon and dexterous service tasks, such as packing a backpack and lifting a garbage bag. We highlight the human-robot interaction flow, where contextual understanding of user intent and preferences, together with human-in-the-loop confirmation or adjustment during execution, is essential for effective assistance.
Chinese Translation
与智能机器人长期物理共存不仅需要强大的机器人策略。一个持久的机器人助手必须支持多样化的用户接口，保持对人及其偏好的长期记忆，协调不同机器人形态，并将人类意图转化为安全的物理执行。我们介绍了PHILIA，一个围绕机器人网关抽象构建的多机器人代理。PHILIA保留了OpenClaw丰富的交互和工具生态系统，同时通过统一的能力接口暴露机器人本地运行时、机载感知、导航、扬声器和机器人策略。该设计将低频、高语义的代理推理与高频、低级的机器人执行解耦，从而实现用户接口、机器人形态和策略后端的即插即用集成。因此，用户体验变得可组合：用户接口、机器人形态、机器人策略、导航或交互算法的进步可以在不重新设计系统的情况下改善整体体验。我们在Astribot S1机器人上验证了该架构，同时设计机器人网关合同，以通过共享能力接口支持未来异构机器人平台的观察、任务执行、导航、语音播放、状态监控和任务取消。我们展示了代表性的用例，其中代理记忆和场景理解与机器人动作相结合。这些用例涵盖了互动家庭场景，从简单的整理到具有挑战性的长期和灵巧服务任务，例如打包背包和提起垃圾袋。我们强调了人机交互流程，其中对用户意图和偏好的上下文理解，以及在执行过程中的人机确认或调整，对于有效的协助至关重要。
cs.RO / 73 / 2607.11386
From Sketch Prior to Trajectories: A Mission-Oriented Coordinated Navigation Framework for Indoor UAV Swarm
从草图先验到轨迹：面向任务的室内无人机群协调导航框架
Abstract
UAV swarm for applications, such as indoor inspection, security patrol, and logistics delivery, are often mission-oriented rather than exploration-oriented. In these tasks, UAVs are required to visit task-relevant regions in a prescribed sequence, and such region-level mission information can often be obtained from pre-deployment sketch-map priors, such as floor plans, CAD layouts, or evacuation diagrams. Although these tasks are executed in three-dimensional space, UAVs usually fly within a specific altitude layer or a nearly fixed altitude range on each floor, making mission-level region transitions mainly governed by planar connectivity. Based on these observations, this paper proposes a mission-oriented coordinated navigation framework that exploits sketch-map priors for multi-UAV indoor operations. Onboard observations are used to perform topological alignment, and the aligned prior is fused with online observations to construct a mission-oriented traversability representation. A layered 2D--3D coordinated navigation framework is further developed, where 2D guided path planning generates mission-oriented guide paths and guide-driven 3D trajectory optimization produces dynamically feasible and collision-free trajectories. Simulation and real-world experiments validate the effectiveness of the proposed framework in structured multi-room indoor environments and further demonstrate its coordinated navigation capability under both communication-available and communication-loss conditions. Multi-floor simulation results show the scalability of the system to layered indoor structures.
Chinese Translation
无人机群在室内检查、安全巡逻和物流配送等应用中，通常是面向任务而非探索。在这些任务中，无人机需要按照规定的顺序访问与任务相关的区域，而这种区域级的任务信息通常可以通过预部署的草图地图先验获得，例如平面图、CAD布局或疏散图。尽管这些任务是在三维空间中执行的，但无人机通常在每层楼内飞行于特定的高度层或几乎固定的高度范围，使得任务级区域的转换主要受平面连通性的影响。基于这些观察，本文提出了一种面向任务的协调导航框架，利用草图地图先验进行多无人机的室内操作。通过机载观测进行拓扑对齐，并将对齐后的先验与在线观测融合，以构建面向任务的可通行性表示。进一步开发了一个分层的2D-3D协调导航框架，其中2D引导路径规划生成面向任务的引导路径，而引导驱动的3D轨迹优化则产生动态可行且无碰撞的轨迹。仿真和实际实验验证了所提框架在结构化多房间室内环境中的有效性，并进一步展示了其在通信可用和通信丢失条件下的协调导航能力。多层仿真结果显示了该系统在分层室内结构中的可扩展性。
cs.RO / 74 / 2607.11397
WALA Learning Executable Latent Actions from Action-Labeled Demonstrations and Action-Free Videos
WALA：从动作标记示范和无动作视频中学习可执行的潜在动作
Abstract
Generalizable robot policies typically rely on action-labeled robot demonstrations, which are expensive to collect and difficult to scale. In contrast, large-scale human and robot videos contain rich physical interactions but often lack executable robot action labels. We present WALA, a framework for learning executable latent actions from both action-labeled demonstrations and action-free videos. WALA first pretrains a semantic-geometric latent action model from videos by modeling the evolution between current observations and sparsely sampled future observations. Instead of reconstructing raw pixels, WALA predicts future deltas in the DINOv3 feature space and dense depth space, preserving task-relevant semantic and geometric structure while reducing sensitivity to appearance details. During policy training, the pretrained encoder provides stable latent action targets, and the decoder serves as a trainable latent world model. The latent actions generated by the vision-language backbone are jointly supervised by robot action prediction, latent action target matching, and future dynamics prediction. This enables action-labeled demonstrations to provide executable control supervision, while action-free videos contribute dynamics supervision without requiring robot action annotations. Experiments show that WALA achieves strong performance on RoboTwin, sets a new state-of-the-art result on RoboCasa with 75.2% average success, and improves both policy performance and generalization in real-world manipulation tasks.
Chinese Translation
通用机器人策略通常依赖于动作标记的机器人示范，这些示范收集成本高且难以扩展。相比之下，大规模的人类和机器人视频包含丰富的物理交互，但往往缺乏可执行的机器人动作标签。我们提出了WALA，一个从动作标记示范和无动作视频中学习可执行潜在动作的框架。WALA首先通过建模当前观察与稀疏采样未来观察之间的演变，从视频中预训练一个语义-几何潜在动作模型。WALA并不重建原始像素，而是预测DINOv3特征空间和密集深度空间中的未来增量，保留与任务相关的语义和几何结构，同时降低对外观细节的敏感性。在策略训练期间，预训练的编码器提供稳定的潜在动作目标，解码器则作为可训练的潜在世界模型。由视觉-语言主干生成的潜在动作通过机器人动作预测、潜在动作目标匹配和未来动态预测进行联合监督。这使得动作标记示范能够提供可执行的控制监督，而无动作视频则在不需要机器人动作注释的情况下贡献动态监督。实验表明，WALA在RoboTwin上表现出色，在RoboCasa上创造了75.2%的平均成功率的新最优结果，并改善了现实世界操作任务中的策略性能和泛化能力。
cs.RO / 75 / 2607.11427
EDAR: Learning Environment-Dependent Action Representations for Robotic Manipulation
EDAR：学习环境依赖的动作表示用于机器人操作
Abstract
Learning effective action representations is critical for robotic manipulation, where raw control trajectories are often noisy, redundant, and difficult to model directly. Existing methods mainly encode the structure of the action stream itself, treating the role of actions in the environment as implicit. Yet manipulation is about changing the world: the same action segment can induce different outcomes under different scene contexts, making action semantics inherently environment-dependent. We propose EDAR, an Environment-Dependent Action Representation that grounds action tokens in both executable control structure and expected visual consequences. By coupling motor commands with their environment-conditioned effects, EDAR encourages the learned action space to capture interaction semantics rather than merely command-level patterns. Experiments on simulated and real-robot manipulation benchmarks demonstrate that EDAR improves downstream policy learning, especially in long-horizon manipulation. These results highlight the importance of grounding action representations in executable control structure and environment-conditioned visual change.
Chinese Translation
学习有效的动作表示对于机器人操作至关重要，因为原始控制轨迹通常是嘈杂的、多余的，并且难以直接建模。现有方法主要编码动作流本身的结构，将动作在环境中的作用视为隐含。然而，操作的本质在于改变世界：相同的动作片段在不同场景上下文中可能会引发不同的结果，使得动作语义本质上依赖于环境。我们提出了EDAR（Environment-Dependent Action Representation），一种将动作标记与可执行控制结构和预期视觉后果相结合的环境依赖动作表示。通过将运动指令与其环境条件下的效果相结合，EDAR鼓励学习到的动作空间捕捉交互语义，而不仅仅是命令级别的模式。在模拟和真实机器人操作基准上的实验表明，EDAR提高了下游策略学习的效果，特别是在长时间操作中。这些结果强调了将动作表示与可执行控制结构和环境条件下的视觉变化相结合的重要性。
cs.RO / 76 / 2607.11481
Towards Human-level Dexterous Teleoperation
迈向人类水平的灵巧远程操作
Abstract
Humans routinely wield tools, swap grasps, and reposition objects within a single hand, seamlessly orchestrating contact transitions that span translation, reorientation, and finger gaiting. Endowing robot dexterous hands with this level of in-hand dexterity through teleoperation requires precise control of object motion via dynamic hand-object contact, yet current teleoperation systems remain far from this capability. To bridge this gap, we take a major step towards human-level dexterous teleoperation by introducing TeleDexter, a hand-object co-tracking controller that maps operator intent into learned, low-level contact execution. The controller is trained on consecutive co-tracking subgoals derived from human reference motions, utilizing a hybrid reward that couples sparse subgoal objectives with dense tracking rewards to enable learning across diverse interaction modalities rather than frame-wise trajectory imitation. The entire pipeline requires only single-stage RL and, with random action masking and domain randomization, transfers zero-shot to the real robot. We evaluate TeleDexter on seven challenging dexterous teleoperation tasks spanning object reorientation and long-horizon tool use across two dexterous hands, achieving a 75% average success rate where all baselines consistently fail. Furthermore, the collected demonstrations successfully train autonomous policies via behavioral cloning, marking a concrete step towards human-level dexterous teleoperation.
Chinese Translation
人类通常能够熟练地使用工具、交换抓握方式，并在单只手内重新定位物体，毫不费力地协调涉及平移、重新定向和手指步态的接触转换。通过远程操作赋予机器人灵巧手这种级别的手内灵巧性，需要通过动态的手-物体接触精确控制物体运动，但目前的远程操作系统仍远未达到这一能力。为弥补这一差距，我们通过引入TeleDexter，一个手-物体协同跟踪控制器，迈出了实现人类水平灵巧远程操作的重要一步，该控制器将操作员意图映射为学习到的低级接触执行。控制器在基于人类参考动作的连续协同跟踪子目标上进行训练，利用一种混合奖励机制，将稀疏的子目标目标与密集的跟踪奖励结合，以便在多样的交互模式中进行学习，而不是逐帧轨迹模仿。整个流程仅需单阶段强化学习，并通过随机动作屏蔽和领域随机化实现零样本迁移到真实机器人。我们在七个具有挑战性的灵巧远程操作任务上评估了TeleDexter，这些任务涵盖了物体重新定向和跨越两个灵巧手的长时间工具使用，取得了75%的平均成功率，而所有基线方法均持续失败。此外，收集到的演示成功地通过行为克隆训练了自主策略，标志着向人类水平灵巧远程操作迈出了实质性的一步。
cs.RO / 77 / 2607.11498
See like a Robot: Robot-Centric Pointmaps for Vision-Language-Action Models
像机器人一样观察：面向视觉-语言-动作模型的机器人中心点图
Abstract
Vision-language-action (VLA) models predict robot actions from visual observations and language instructions. These actions are defined in the robot's own 3D coordinate frame, yet most VLAs observe the scene in the camera frame, creating a frame mismatch between where the scene is observed and where actions are defined. The mismatch is benign under a fixed viewpoint, where the policy can memorize a single observation-to-action mapping, but grows harder as large-scale datasets aggregate demonstrations across diverse camera setups and the policy must generalize this mapping across viewpoints. We address this mismatch with robot-centric pointmaps, images whose pixels store the 3D coordinates of scene points in the robot frame. Pointmaps provide robot-frame 3D geometry while preserving the dense H x W grid expected by pretrained 2D VLAs, so they integrate into existing VLAs with minimal architectural change. On RoboCasa, pointmaps improve both pi0.5 and SmolVLA and outperform representative camera-viewpoint and 3D-aware baselines. In real-robot experiments, their advantage over an RGB-only policy widens when the camera is moved to a placement unseen during training.
Chinese Translation
视觉-语言-动作（VLA）模型根据视觉观察和语言指令预测机器人动作。这些动作是在机器人的自身3D坐标系中定义的，但大多数VLA是在相机坐标系中观察场景，从而造成了观察场景的位置与定义动作的位置之间的框架不匹配。在固定视点下，这种不匹配是良性的，因为策略可以记忆单一的观察到动作的映射，但随着大规模数据集汇聚来自不同相机设置的演示，这种映射的泛化变得更加困难。我们通过机器人中心点图来解决这一不匹配问题，点图的像素存储了场景点在机器人坐标系中的3D坐标。点图提供了机器人坐标系下的3D几何信息，同时保留了预训练的2D VLA所期望的密集H x W网格，因此它们可以以最小的架构变化集成到现有的VLA中。在RoboCasa上，点图提升了pi0.5和SmolVLA的性能，并超越了代表性的相机视点和3D感知基线。在真实机器人实验中，当相机移动到训练期间未见过的位置时，它们相较于仅使用RGB的策略的优势进一步扩大。
cs.RO / 78 / 2607.11570
ERR@HRI 3.0 Challenge: Multimodal Detection of Errors and Anticipation in Human-Robot Interactions
ERR@HRI 3.0 挑战：人机交互中的多模态错误检测与预判
Abstract
As robots become increasingly integrated into human environments, their ability to detect and respond to errors remains critical for maintaining user trust and interaction quality. While recent advances in machine learning have improved error detection capabilities, most approaches are limited to specific contexts, controlled settings, or pre-extracted features, limiting their generalizability and applicability to real-world conditions. To address this challenge, the third edition of the ERR@HRI Challenge (ERR@HRI 3.0) provided researchers with two complementary datasets that enable end-to-end innovation in methods for both detecting and preventing errors in human-robot interaction. The challenge offered raw, non-anonymized video data from naturalistic settings: (1) the Bystander Affect Detection (BAD) dataset, containing webcam recordings of 45 participants' spontaneous reactions to robot and human failure scenarios; and (2) the Bad Idea dataset, featuring 29 participants' anticipatory facial responses while predicting action outcomes before failures occur. Both datasets were collected via crowdsourcing, capturing the inherent variability of real-world conditions. This naturalistic variability, while challenging, provides an authentic testbed for developing robust error detection systems. Participants developed multimodal machine learning models for bystander reaction detection (Track 1) and anticipatory outcome prediction (Track 2), with an optional cross-dataset generalization track (Track 3). Three teams submitted valid models, all of which surpassed our convolutional neural network baselines. This paper describes the datasets, tasks, baselines, and results of ERR@HRI 3.0, and discusses implications for building generalizable, context-aware, and anticipatory error detection systems for human-robot interaction.
Chinese Translation
随着机器人越来越多地融入人类环境，其检测和响应错误的能力对于维持用户信任和交互质量至关重要。尽管最近在机器学习方面的进展提高了错误检测能力，但大多数方法仍然局限于特定的上下文、受控环境或预提取特征，限制了它们在现实世界条件下的普适性和适用性。为了解决这一挑战，ERR@HRI 挑战的第三届（ERR@HRI 3.0）为研究人员提供了两个互补的数据集，以支持在检测和预防人机交互中的错误方面的端到端创新。该挑战提供了来自自然环境的原始、非匿名视频数据：(1) 旁观者情感检测（Bystander Affect Detection, BAD）数据集，包含45名参与者对机器人和人类失败场景的自发反应的网络摄像头录制；(2) 坏主意（Bad Idea）数据集，展示了29名参与者在预测行动结果之前的预判面部反应。这两个数据集通过众包方式收集，捕捉了现实世界条件固有的变异性。这种自然变异性虽然具有挑战性，但为开发强健的错误检测系统提供了真实的测试平台。参与者开发了多模态机器学习模型用于旁观者反应检测（Track 1）和预判结果预测（Track 2），并提供了一个可选的跨数据集泛化轨道（Track 3）。三支团队提交了有效模型，所有模型均超过了我们的卷积神经网络基线。本文描述了 ERR@HRI 3.0 的数据集、任务、基线和结果，并讨论了构建可泛化、上下文感知和预判错误检测系统在人机交互中的意义。
cs.RO / 79 / 2607.11603
WarpMPC: Large-Batch MPC on GPU via ADMM with Unrolled $LDL^\top$ Factorization
WarpMPC：基于ADMM的GPU大批量MPC通过展开$LDL^ op$分解
Abstract
This paper introduces numerical optimizations for maximizing throughput on GPU when solving large batches (10,000 to over 100,000) of sequential quadratic programming (SQP) iterations, where all problems have the same structure. The optimizations are implemented in a toolbox WarpMPC for model-predictive control (MPC) in JAX and Warp. Based on the insight that all MPC problem instances in a batch share the same sparsity in time, cost, and constraints, we propose unrolling sparse linear factorizations and solves, which dominate alternating direction method of multipliers (ADMM) solver runtime. We avoid memory access bottlenecks and wasting computations via optimized memory layout, padding-reducing segmentation of the unrolled factorization, and dependency level scheduled backsolves, additionally accelerating sensitivity computation. We achieve throughputs of 8,000 to 250,000 SQP iterations per second on nonlinear cartpole, quadrotor, and humanoid robot benchmarks, outperforming baselines by 3$\times$ to 25$\times$. We illustrate practical usefulness by synthesizing a dataset and training a neural network approximation of an MPC in under 4 minutes that stabilizes a nano quadrotor in hardware experiments.
Chinese Translation
本文介绍了在GPU上解决大批量（从10,000到超过100,000个）顺序二次规划（SQP）迭代时最大化吞吐量的数值优化方法，其中所有问题具有相同的结构。这些优化方法被实现于一个名为WarpMPC的工具箱中，该工具箱用于在JAX和Warp中进行模型预测控制（MPC）。基于所有批次中的MPC问题实例在时间、成本和约束方面共享相同稀疏性的洞察，我们提出了展开稀疏线性分解和求解的方法，这些方法主导了交替方向乘子法（ADMM）求解器的运行时间。我们通过优化内存布局、减少填充的展开分解分段和依赖级别调度的回代求解，避免了内存访问瓶颈和计算浪费，同时加速了灵敏度计算。在非线性倒立摆、四旋翼和类人机器人基准测试中，我们实现了每秒8,000到250,000个SQP迭代的吞吐量，超越了基线3倍至25倍。我们通过合成数据集并在不到4分钟内训练一个神经网络近似的MPC，展示了其实用性，该MPC在硬件实验中稳定了一个纳米四旋翼。
cs.RO / 80 / 2607.11624
SKooP: Symmetric Koopman Predictions for Faster and More Generalizable Legged Robot Locomotion with Reinforcement Learning
SKooP：对称Koopman预测在强化学习下加速和增强四足机器人行走的泛化能力
Abstract
Reinforcement learning (RL) algorithms classically suffer from poor sample efficiency. In robotics, a recent line of work has emerged addressing this problem by encoding physics priors in the learning process. However, most of these approaches are validated on well-defined, low-dimensional benchmark systems rather than high-dimensional robots with complex nonlinear dynamics. In this paper, we introduce \textit{SKooP (Symmetric Koopman Predictions)}, an approach combining the advantages of morphological symmetries with those of a Koopman model learned via autoencoder to enhance policy learning. SKooP learns a Koopman model of the system dynamics alongside the policy. The resulting Koopman predictions are used as privileged observations for the critic, allowing the agent to learn based on smoother, more informative features. We also incorporate group symmetries into the actor, critic, encoder and decoder networks to produce a highly equivariant policy. The SKooP approach is validated via in-depth analysis of the learned Koopman models and symmetric policies to showcase how each of these influences the agent's performance. We also show that the learned policies are transferable to different simulation environments. Our results show that SKooP consistently reduces convergence time and increases the learned reward for multiple challenging bipedal locomotion tasks on a quadruped robot. Project page: https://evelyd.github.io/SymmetricKoopmanPredictions/
Chinese Translation
强化学习（RL）算法通常存在样本效率低下的问题。在机器人领域，最近出现了一系列研究，旨在通过在学习过程中编码物理先验来解决这一问题。然而，这些方法大多是在定义明确、维度较低的基准系统上进行验证，而非在具有复杂非线性动力学的高维机器人上进行验证。本文介绍了 extit{SKooP（对称Koopman预测）}，一种结合形态对称性优势与通过自编码器学习的Koopman模型的策略学习方法。SKooP在学习策略的同时学习系统动力学的Koopman模型。生成的Koopman预测作为特权观测用于评论者，使得代理能够基于更平滑、更具信息量的特征进行学习。我们还将群体对称性纳入演员、评论者、自编码器和解码器网络，以产生高度等变的策略。通过对学习到的Koopman模型和对称策略的深入分析，验证了SKooP方法，展示了这些因素如何影响代理的性能。我们还表明，学习到的策略可以迁移到不同的仿真环境中。我们的结果表明，SKooP在多个具有挑战性的双足行走任务中，始终减少了收敛时间并增加了学习到的奖励，适用于四足机器人。项目页面：https://evelyd.github.io/SymmetricKoopmanPredictions/
cs.RO / 81 / 2607.11633
Breaking the 15% Barrier: A Real-World Data-Driven System for Proactive Social Robot Triggered by User Nonverbal Cues
突破15%的障碍：基于真实世界数据的主动社交机器人系统，通过用户非语言线索触发
Abstract
Service robots in retail stores increasingly rely on cascaded speech pipelines (STT-LLM-TTS), yet many customer-robot interactions are initiated or guided by nonverbal behaviors such as approaching, waving, pointing, or showing items. This paper studies such cues in a real-world store deployment with a teleoperated humanoid robot and shows that a non-negligible portion of robot turns are triggered by nonverbal behaviors rather than spoken input, revealing a limitation of audio-only dialogue systems. In a 6-day in-the-wild deployment, 15.3\% of robot utterances were initiated by users' nonverbal behaviors rather than spoken input. Based on an analysis of observed customer behaviors, we define a set of frequent, service-relevant nonverbal cues and develop a real-time multi-person, multi-label recognizer that runs online from video. We then propose a dialogue framework that conditions LLM-based utterance generation on recognized nonverbal cue tokens, and optionally leverages a vision-language model when items are shown, enabling proactive robot responses without hand-crafted rules. We evaluate the approach offline on nonverbal-triggered turns and demonstrate an online prototype that reacts to users' nonverbal cues in real time.
Chinese Translation
零售店中的服务机器人越来越依赖级联语音管道（STT-LLM-TTS），然而许多客户与机器人的互动是由非语言行为如接近、挥手、指向或展示物品等引发或引导的。本文研究了在真实商店部署中使用遥控类人机器人时的这些线索，并显示出相当一部分机器人的发言是由非语言行为而非口头输入触发的，这揭示了仅依赖音频的对话系统的局限性。在为期6天的实际部署中，15.3%的机器人发言是由用户的非语言行为而非口头输入发起的。基于对观察到的客户行为的分析，我们定义了一组频繁且与服务相关的非语言线索，并开发了一个实时的多用户、多标签识别器，该识别器能够在线处理视频数据。随后，我们提出了一个对话框架，该框架基于识别到的非语言线索标记来调节基于大语言模型（LLM）的发言生成，并在展示物品时可选地利用视觉-语言模型，从而实现主动的机器人响应，而无需手工编写规则。我们在非语言触发的发言上对该方法进行了离线评估，并展示了一个在线原型，能够实时响应用户的非语言线索。
cs.RO / 82 / 2607.11638
DA-Nav: Direction-Aware City-Scale Vision-Language Navigation
DA-Nav：方向感知的城市规模视觉-语言导航
Abstract
City-scale outdoor navigation is currently hindered by the heavy reliance on dense maps or costly navigation supervision. In this work, we introduce a novel paradigm for leveraging directional instructions from commercial navigation tools (e.g., Google Maps). To bridge the gap between commercial instructions and executable navigation actions, while mitigating long-horizon error accumulation through robust trajectory recovery, we propose DA-Nav, a Direction-Aware vision-language Navigation framework that reformulates navigation as a discrete spatial grounding problem on the egocentric 2D image plane. To achieve trajectory recovery, DA-Nav employs a Chain-of-Thought (CoT) reasoning process encompassing deviation assessment, action prediction, and target grid selection. We further introduce ReDA, a dataset that provides direction-aware instructions and recovery trajectories to enhance spatial grounding and support CoT recovery reasoning. Extensive experiments in CARLA demonstrate that DA-Nav achieves a high success rate of 56.16% in unseen urban environments, outperforming existing State-of-The-Art (SoTA) methods while maintaining a substantially stronger recovery capability. Furthermore, without fine-tuning, DA-Nav seamlessly adapts to both quadruped and humanoid robots, enabling stable kilometer-scale closed-loop outdoor navigation in complex real world environments.
Chinese Translation
城市规模的户外导航目前受到对密集地图或昂贵导航监督的严重依赖的制约。在本研究中，我们引入了一种新颖的范式，利用来自商业导航工具（例如，Google Maps）的方向指令。为了弥合商业指令与可执行导航动作之间的差距，同时通过稳健的轨迹恢复减轻长时间范围内的误差积累，我们提出了DA-Nav，一个方向感知的视觉-语言导航框架，将导航重新定义为在自我中心的二维图像平面上的离散空间定位问题。为了实现轨迹恢复，DA-Nav采用了一种链式思维（Chain-of-Thought, CoT）推理过程，包括偏差评估、动作预测和目标网格选择。我们进一步引入了ReDA，一个提供方向感知指令和恢复轨迹的数据集，以增强空间定位并支持CoT恢复推理。在CARLA中的广泛实验表明，DA-Nav在未见过的城市环境中实现了56.16%的高成功率，超越了现有的最先进（State-of-The-Art, SoTA）方法，同时保持了显著更强的恢复能力。此外，在没有微调的情况下，DA-Nav能够无缝适应四足机器人和类人机器人，使其能够在复杂的现实环境中实现稳定的公里级闭环户外导航。
cs.RO / 83 / 2607.11643
Xiaomi-Robotics-U0: Unified Embodied Synthesis with World Foundation Model
小米机器人U0：具有世界基础模型的统一具身合成
Li, Xinghang, Guo, Jun, Li, Qiwei, Qian, Long, Lai, Hang, Wang, Yueze, Yan, Hongyu, Cao, Jiahang, Chen, Xi, Qu, Jingen, Song, Jiaxi, Sun, Nan, Zhao, Hanye, Liu, Futeng, Peng, Wanli, Wang, Heyun, Wang, Yunhong, Xia, Caoyu, Zhao, Jack, Xiang, Diyun, Ye, Hangjun, Qu, Heng, Liu, Huaping, Li, Jason
Abstract
Recent foundation image and video generation models offer strong generalization and controllability, but their direct application to embodied scenarios is limited by requirements for multi-view consistency, geometric coherence, and robot embodiment constraints. Existing methods typically adapt foundation models with limited robot data, often sacrificing visual knowledge acquired during large-scale pre-training. We present Xiaomi-Robotics-U0, a 38-billion-parameter multimodal autoregressive model for unified embodied synthesis. It treats embodied generation as an extension of foundation image and video generation and jointly optimizes text-to-image generation, image editing, embodied scene generation, embodied transfer, and embodied video generation. This unified framework preserves the generalization of the pre-trained world foundation model while adapting it to embodied settings. Xiaomi-Robotics-U0 is the first model to support high-quality multi-view scene generation across multiple robot embodiments and to introduce structured, controllable embodied transfer for fine-grained editing while preserving multi-view consistency and interaction dynamics. It achieves state-of-the-art results on single-step and sequential generation tasks, outperforming GPT-Image-2.0 in human evaluations of embodied scene generation and transfer, ranking first on World Arena for embodied video generation, and improving the out-of-distribution success rate of pi_0.5 from 36.9% to 63.2% on challenging real-world manipulation tasks. These results show that foundation world models can serve both as embodied world models and scalable data engines for embodied intelligence. Code and checkpoints are available at https://robotics.xiaomi.com/xiaomi-robotics-u0.html.
Chinese Translation
近期的基础图像和视频生成模型展现了强大的泛化能力和可控性，但其在具身场景中的直接应用受到多视角一致性、几何一致性和机器人具身约束的限制。现有方法通常使用有限的机器人数据来调整基础模型，往往牺牲了在大规模预训练中获得的视觉知识。我们提出了小米机器人U0，这是一个拥有380亿参数的多模态自回归模型，用于统一具身合成。它将具身生成视为基础图像和视频生成的扩展，并联合优化文本到图像生成、图像编辑、具身场景生成、具身迁移和具身视频生成。该统一框架在将预训练的世界基础模型适应于具身环境的同时，保留了其泛化能力。小米机器人U0是第一个支持跨多个机器人具身的高质量多视角场景生成的模型，并引入了结构化、可控的具身迁移，以便进行细粒度编辑，同时保持多视角一致性和交互动态。它在单步和序列生成任务上实现了最先进的结果，在具身场景生成和迁移的人类评估中超越了GPT-Image-2.0，并在具身视频生成的World Arena中排名第一，同时将pi_0.5在具有挑战性的现实世界操作任务中的分布外成功率从36.9%提高到63.2%。这些结果表明，基础世界模型可以同时作为具身世界模型和具身智能的可扩展数据引擎。代码和检查点可在 https://robotics.xiaomi.com/xiaomi-robotics-u0.html 获取。
cs.RO / 84 / 2607.11651
Coordinated Incremental Trajectory Tracking of a Tailsitter Drone
协调增量轨迹跟踪的尾坐式无人机
Abstract
This paper derives an analytical differential flatness transform for a tailsitter Unmanned Aerial Vehicle (UAV) under coordinated flight conditions using a simplified aerodynamic model. The proposed framework is formulated exclusively using rotation matrices, avoiding the ambiguities inherent to Euler angle representations. The method extends the applicability of an existing state-of-the-art differential flatness-based controller to flight regimes involving a significant vertical velocity component, where the previous approach becomes inapplicable. The proposed framework is validated experimentally with trajectories that highlight its advantages in these regimes.
Chinese Translation
本文在协调飞行条件下，基于简化的气动模型推导了尾坐式无人机（UAV）的解析微分平坦变换。所提出的框架完全使用旋转矩阵进行构建，避免了欧拉角表示法固有的模糊性。该方法扩展了现有的基于微分平坦性的最先进控制器的适用性，适用于涉及显著垂直速度分量的飞行状态，而之前的方法在这种情况下不再适用。通过实验验证了所提出框架的有效性，所选轨迹突显了其在这些状态下的优势。
cs.RO / 85 / 2607.11654
A Model for Mediating Multi-Modal Human Intent into Safe Maneuvers for UAVs
一种将多模态人类意图转化为无人机安全机动的中介模型
Abstract
Direct human interaction with autonomous UAV systems can be enabled through modalities such as speech, gestures, and graphical interfaces. However, interpreting such inputs as directly executable commands introduces safety risks in dynamic environments. Operator requests may conflict with terrain constraints, inter-UAV separation requirements, or flight-envelope limitations. In this paper, we present a requirements-governed maneuver-response model that mediates multi-modal human intent into safe UAV maneuvers by treating operator inputs as bounded maneuver requests rather than direct commands. Requested maneuvers are mapped to constrained motion primitives and processed through a structured request-evaluate-execute pipeline. Each request is interpreted with associated confidence, validated against terrain, separation, workspace, and flight-envelope constraints, and either constrained, rejected, or executed under continuous runtime monitoring. We further formalize the approach as a requirements-based specification model in which maneuver primitives are associated with explicit preconditions, invariants, guard conditions, and postconditions governing admissibility, execution safety, and emergency handling. These requirements support runtime verification and future reactive synthesis approaches. We present an initial lab-based validation demonstrating that voice and GUI-based inputs can be reliably interpreted and safely executed as constrained maneuver requests.
Chinese Translation
通过语音、手势和图形界面等多种方式，可以实现人类与自主无人机系统的直接互动。然而，将这些输入解释为可直接执行的命令在动态环境中引入了安全风险。操作员的请求可能与地形限制、无人机间隔要求或飞行包络限制相冲突。本文提出了一种基于需求的机动响应模型，该模型通过将操作员输入视为有界的机动请求而非直接命令，从而将多模态人类意图转化为安全的无人机机动。请求的机动被映射到受限的运动原语，并通过结构化的请求-评估-执行管道进行处理。每个请求都带有相关的置信度进行解释，并根据地形、间隔、工作空间和飞行包络限制进行验证，最终决定是约束、拒绝还是在持续的运行时监控下执行。我们进一步将该方法形式化为一种基于需求的规范模型，其中机动原语与明确的前提条件、不变性、保护条件和后置条件相关联，以规范可接受性、执行安全性和紧急处理。这些需求支持运行时验证和未来的反应合成方法。我们展示了初步的实验室验证，证明了基于语音和图形用户界面输入的请求可以被可靠地解释并安全地作为受限机动请求执行。
cs.RO / 86 / 2607.11657
Trajectory Planning and Certification for 3-DOF Robot Manipulators Using Real Quantifier Elimination Based on Comprehensive Gr\"obner Systems
基于综合格罗布纳系统的3自由度机器人手臂轨迹规划与认证的实数量词消除方法
Abstract
We propose an algorithm and its implementation for trajectory planning and certification for 3-DOF robot manipulators. The method uses Real Quantifier Elimination (QE) based on Comprehensive Gr\"obner Systems (CGS), also known as the CGS-QE method. The main advantage of the proposed method is its efficiency in trajectory planning and solution certification. This efficiency comes from the effective use of the CGS. First, for trajectory planning, we solve the inverse kinematics problem at each point along the trajectory via Gr\"obner basis computation. This usually requires recalculating the Gr\"obner basis at every point, which is time-consuming. We avoid this by computing the CGS for a parametric system. Here, the end-effector coordinates are parameters. This approach streamlines the algorithm. Second, for solution certification, the CGS-QE method certifies that an inverse kinematics solution exists at any point along the end-effector's trajectory. Our method also certifies solutions for trajectories composed of line segments and cubic natural splines. The algorithm is implemented within the computer algebra system Risa/Asir.
Chinese Translation
我们提出了一种用于3自由度机器人手臂的轨迹规划与认证的算法及其实现。该方法基于综合格罗布纳系统（Comprehensive Gr"obner Systems, CGS）的实数量词消除（Real Quantifier Elimination, QE），也称为CGS-QE方法。所提方法的主要优点在于其在轨迹规划和解的认证方面的高效性。这种高效性得益于CGS的有效利用。首先，在轨迹规划中，我们通过计算格罗布纳基解决轨迹上每一点的逆运动学问题。这通常需要在每一点重新计算格罗布纳基，耗时较长。我们通过为参数系统计算CGS来避免这一点。在这里，末端执行器的坐标作为参数。该方法简化了算法。其次，对于解的认证，CGS-QE方法可以证明在末端执行器轨迹的任意一点存在逆运动学解。我们的方法还能够认证由线段和三次自然样条构成的轨迹的解。该算法在计算机代数系统Risa/Asir中实现。
cs.RO / 87 / 2607.11686
Self-Healing Visual Recovery for Autonomous Ground Vehicles Using Camera-Only Visual Odometry
基于仅摄像头视觉里程计的自主地面车辆自愈视觉恢复
Abstract
Low-cost unmanned ground vehicles are often used in indoor places like warehouses, inspection corridors, and farm rows, where painted floor lines guide the robot. Line following is useful because it only needs one camera and little computing power, but it can fail when the line is blocked or turns sharply and goes out of view. Sensor-rich platforms tolerate this through hardware redundancy (LiDAR, GPS, multiple cameras), but camera-only systems must recover at runtime with no additional infrastructure. This paper presents a lightweight, two-stage recovery approach that restores guideline tracking without LiDAR, GPS, or a GPU. When the line is lost, the robot first turns in place while slowly relaxing its color checks and waiting for confirmation across multiple frames (Stage 1). If the line is still not found, monocular visual odometry moves the robot back to saved breadcrumb positions before it tries again (Stage 2). The system uses a depth-gated HSV line tracker, a YOLOv8n obstacle detector, and a visual odometry breadcrumb mapper, and it runs at 20 Hz on CPU-only hardware. The controller embeds a complete MAPE-K loop within a single 50 ms control tick, with no external adaptation manager required. The approach is evaluated across 119 fault-injected episodes on three Webots simulation courses. The method was successful in 86.6% of cases, with a median recovery time of 3.26 seconds. These results demonstrate that reliable visual recovery is feasible on camera-only UGVs within practical cost and computational limits.
Chinese Translation
低成本无人地面车辆通常用于仓库、检查走廊和农田等室内场所，在这些地方，涂漆的地面线条为机器人提供导航。线条跟随技术的优势在于只需一台摄像头和较少的计算能力，但当线条被遮挡或急剧转弯而失去视野时，系统可能会失败。配备多种传感器的平台通过硬件冗余（如激光雷达、全球定位系统、多摄像头）来容忍这种情况，而仅依赖摄像头的系统必须在运行时恢复，而无需额外的基础设施。本文提出了一种轻量级的两阶段恢复方法，能够在没有激光雷达、全球定位系统或图形处理单元的情况下恢复指导线跟踪。当失去线条时，机器人首先在原地旋转，同时逐渐放宽颜色检查，并等待多个帧的确认（阶段1）。如果仍然找不到线条，单目视觉里程计会将机器人移动回保存的面包屑位置，然后再尝试（阶段2）。该系统使用深度门控的HSV线条跟踪器、YOLOv8n障碍物检测器和视觉里程计面包屑映射器，并在仅使用CPU的硬件上以20 Hz的频率运行。控制器在单个50毫秒的控制周期内嵌入了完整的MAPE-K循环，无需外部适应管理器。该方法在三个Webots仿真课程上进行了119个故障注入实验的评估，成功率为86.6%，中位恢复时间为3.26秒。这些结果表明，在实际成本和计算限制内，仅依赖摄像头的无人地面车辆上实现可靠的视觉恢复是可行的。
cs.RO / 88 / 2607.11688
Automated Synthesis of Facial Mechanisms for Conversational Animatronic Robots
面部机制的自动合成用于对话式仿生机器人
Abstract
Animatronic faces are a central component of socially interactive robots, enabling rich nonverbal communication through facial articulation. However, state-of-the-art animatronic faces are typically tailored systems: each new facial geometry requires extensive manual mechanical redesign, making large-scale personalization prohibitively slow and costly. In this work, we pursue automated and scalable mechanical face synthesis, aiming to rapidly generate a physically realizable facial mechanism for a wide range of facial geometries. We introduce a parametric, linkage-driven mechanical face template whose topology and actuator layout are explicitly parameterized to support systematic scaling and retargeting across diverse facial morphologies. Building on this template, we propose a hierarchical automatic design algorithm that takes a single 2D portrait as input, reconstructs a target 3D face, and synthesizes a collision-free, manufacturable internal mechanism. The algorithm combines anatomy-guided feasible motion volumes, Action Unit (AU)-derived trajectory-based expressiveness objectives, and a collision-driven outer-loop refinement strategy. Beyond hardware synthesis, we argue that future mechanical faces deployed at scale must engage in bidirectional, multi-turn conversation rather than functioning solely as speaking or listening heads. To this end, we develop a dual-identity conversational facial motion synthesis framework that jointly models speaking and listening behaviors from audio, producing temporally coherent 3D facial motion suitable for physical execution. We validate our system through extensive experiments, including (i) quantitative evaluation of automatic mechanism synthesis across diverse facial geometries, (ii) comparisons against manual mechanical design, (iii) benchmarks on conversational facial motion synthesis and real-time deployment, and (iv) perceptual user studies.
Chinese Translation
仿生面部是社会互动机器人中的核心组成部分，通过面部动作实现丰富的非语言交流。然而，最先进的仿生面部通常是定制系统：每种新的面部几何形状都需要进行大量的手动机械重新设计，这使得大规模个性化变得极为缓慢和昂贵。在本研究中，我们追求自动化和可扩展的机械面部合成，旨在快速生成适用于广泛面部几何形状的物理可实现面部机制。我们引入了一种参数化的、基于连杆驱动的机械面部模板，其拓扑结构和驱动器布局被明确参数化，以支持在多样的面部形态中进行系统性的缩放和重定向。在此模板的基础上，我们提出了一种分层自动设计算法，该算法以单个二维肖像作为输入，重建目标三维面部，并合成一个无碰撞、可制造的内部机制。该算法结合了解剖学指导的可行运动体积、基于动作单元（Action Unit, AU）导出的轨迹表达目标，以及基于碰撞的外循环优化策略。除了硬件合成，我们认为未来大规模部署的机械面部必须参与双向、多轮对话，而不仅仅是作为说话或倾听的头部。为此，我们开发了一种双身份对话面部运动合成框架，该框架从音频中联合建模说话和倾听行为，生成适合物理执行的时间一致的三维面部运动。我们通过广泛的实验验证了我们的系统，包括（i）对多样面部几何形状的自动机制合成的定量评估，（ii）与手动机械设计的比较，（iii）对话面部运动合成和实时部署的基准测试，以及（iv）感知用户研究。
cs.RO / 89 / 2607.11689
From World Action Models to Embodied Brains: A Roadmap for Open-World Physical Intelligence
从世界行动模型到具身大脑：开放世界物理智能的路线图
Abstract
Artificial general intelligence ultimately requires agents that can reason and act in the physical world. Action models, vision-language-action policies, and world models have advanced this goal, while World Action Models (WAMs) are particularly promising because they connect candidate interventions with predicted consequences. However, progress remains fragmented: models use incompatible action spaces and prediction targets, datasets and tasks follow different conventions, and runtime systems expose limited interfaces for reuse and evaluation. We review the evolution toward WAMs and organize these limitations into three coupled gaps: model roles and representations, objectives and standardization, and system composition. Building on this analysis, we propose a co-evolution roadmap for physical intelligence centered on the \emph{embodied brain}, a long-term model target for integrating multimodal context, comparing candidate interventions, and issuing state-transition or capability requests rather than direct actuator commands. WAMs provide promising prototypes for its predictive functions, while a physical harness grounds model outputs through tools, controllers, verification, and trace logging. Shared contracts align heterogeneous models, data, tasks, and embodiments, and closed-loop post-training converts verified interaction into reusable experience. Together, these components define a modular physical-intelligence stack for adaptive and self-improving embodied agents.
Chinese Translation
人工通用智能最终需要能够在物理世界中推理和行动的智能体。行动模型、视觉-语言-行动策略和世界模型推动了这一目标的发展，而世界行动模型（World Action Models, WAMs）特别有前景，因为它们将候选干预与预测后果连接起来。然而，进展仍然是碎片化的：模型使用不兼容的行动空间和预测目标，数据集和任务遵循不同的规范，运行时系统提供有限的接口用于重用和评估。我们回顾了WAMs的发展，并将这些局限性组织为三个相互关联的缺口：模型角色与表示、目标与标准化，以及系统组成。在此分析的基础上，我们提出了一个以 extit{具身大脑}为中心的物理智能共演化路线图，这是一个长期的模型目标，旨在整合多模态上下文、比较候选干预，并发出状态转换或能力请求，而不是直接发出执行器命令。WAMs为其预测功能提供了有前景的原型，而物理支架通过工具、控制器、验证和跟踪日志将模型输出与实际应用相结合。共享契约使异构模型、数据、任务和具身体对齐，而闭环后训练将验证的交互转化为可重用的经验。这些组件共同定义了一个模块化的物理智能堆栈，适用于自适应和自我改进的具身智能体。
cs.RO / 90 / 2607.11690
Requirement-Driven Design of Whole-Body Social Tactile Sensing via Virtual Human-Robot Interaction
基于需求驱动的全身社交触觉感知设计：虚拟人机交互的应用
Abstract
Tactile sensing for social-physical human-robot interaction (spHRI) is designed in a hardware-driven manner, where predefined sensor configurations constrain coverage, spatial resolution, and the range of recognizable gestures. We propose a requirement-driven framework that derives sensing requirements, specifically spatial resolution and placement, directly from interaction data. Using a VR-based platform with haptic feedback, we collected high-resolution whole-body contact distributions across multiple social scenarios, from which we identified nine recurring social touch gestures. Eight gestures were selected for controlled data collection with 18 participants, yielding an open-source dataset of 5,520 trials. Analysis of contact distributions and simulated tactile encodings provides quantitative baselines for skin coverage and sensor density on a humanoid robot platform. While demonstrated on a single robot platform, the methodology is designed to be transferable to other robot morphologies, potentially enabling morphology-specific sensing requirements to be derived prior to hardware fabrication.
Chinese Translation
社交物理人机交互（spHRI）的触觉感知采用硬件驱动的方式设计，其中预定义的传感器配置限制了覆盖范围、空间分辨率和可识别手势的范围。我们提出了一种基于需求驱动的框架，该框架直接从交互数据中推导出感知需求，特别是空间分辨率和传感器布置。通过使用具有触觉反馈的虚拟现实（VR）平台，我们收集了多个社交场景下的高分辨率全身接触分布数据，从中识别出九种重复出现的社交触摸手势。选取了八种手势进行控制数据收集，共有18名参与者，生成了一个包含5,520次试验的开源数据集。对接触分布和模拟触觉编码的分析为类人机器人平台上的皮肤覆盖和传感器密度提供了定量基准。尽管在单一机器人平台上进行了演示，但该方法论旨在可转移到其他机器人形态，潜在地使得在硬件制造之前能够推导出特定形态的感知需求。
cs.RO / 91 / 2607.11724
High-level spatial Dubins airplane-based reference smoothing with low-level geometric tracking for quadrotor control
基于高层空间Dubins飞机模型的四旋翼控制参考平滑与低层几何跟踪
Abstract
A method for the control of quadrotors is presented. It is composed of a high-level reference smoothing step and a low-level reference tracking step. The high-level step leverages the Dubins airplane model for dimensionality reduction and reduced computational complexity, and exploits its structure for decoupling, spatial modeling and the formulation of a small linear program. The low-level step leverages a geometric tracking controller, which is based on the full quadrotor model. The method is designed for the tracking of references subject to lateral constraints along the path. An example is the tracking of references along obstacle contours. It is differentiated between two different setups. Either the high-level planning step is conducted once and offline, or, alternatively, the high-level planning step is conducted recedingly online in closed-loop over a limited spatial prediction horizon.
Chinese Translation
本文提出了一种四旋翼控制方法。该方法由高层参考平滑步骤和低层参考跟踪步骤组成。高层步骤利用Dubins飞机模型进行维度降低和计算复杂度的减少，并利用其结构实现解耦、空间建模以及小型线性规划的制定。低层步骤则基于完整的四旋翼模型，采用几何跟踪控制器。该方法旨在跟踪沿路径的受侧向约束的参考点。例如，跟踪沿障碍物轮廓的参考点。本文区分了两种不同的设置：一种是高层规划步骤一次性离线进行，另一种是高层规划步骤在有限空间预测范围内以闭环方式逐步在线进行。
cs.RO / 92 / 2607.11734
NeuralActuator: Neural Actuation Modeling for Robot Dynamics and External Force Perception
NeuralActuator：用于机器人动力学和外部力感知的神经驱动建模
Abstract
Differentiable simulators have advanced policy learning and model-based control, yet actuator dynamics remain an important source of sim-to-real error. This is particularly acute on low-cost platforms, where the linear current-to-torque relation $\tau = K_tI$ becomes unreliable during commanded-target tracking because of friction, hysteresis, backlash, and thermal effects. We present NeuralActuator, a neural actuator model that jointly predicts (i) a simulator-equivalent generalized-effort surrogate for trajectory propagation on low-cost servo platforms, (ii) external force with a contact-probability gate for sensorless force perception, and (iii) a motor-condition score for the supervised joint. We also introduce the Neural Actuation Dataset (NAD), collected with a twin-arm teleoperation system that records robot states and actuator telemetry together with external-force labels. The torque-surrogate head is trained through differentiable simulation from pose trajectories without direct generalized-effort labels, while the force, gate, and motor-condition heads receive direct supervision. A Transformer captures temporal dependencies while supporting real-time inference. We evaluate NeuralActuator on a 5-DoF OpenManipulator-X, a 6-DoF SO-101, and a 7-DoF Franka Emika Panda, spanning three actuator families and platforms costing approximately USD 500 to over USD 30,000. The low-cost platforms support dynamics and force evaluation, while the offline Franka experiment provides an additional payload-force-estimation benchmark. Experiments further demonstrate its application for motor condition estimation on OpenManipulator-X and improved behavior-cloning performance when NeuralActuator is used as a pretrained module.
Chinese Translation
可微分模拟器推动了策略学习和基于模型的控制，但驱动器动态仍然是模拟与现实之间误差的重要来源。这在低成本平台上尤为明显，在这些平台上，线性电流与扭矩的关系 $ au = K_tI$ 在命令目标跟踪过程中因摩擦、滞后、间隙和热效应而变得不可靠。我们提出了NeuralActuator，这是一种神经驱动模型，联合预测（i）用于低成本伺服平台的轨迹传播的模拟器等效广义努力替代物，（ii）带有接触概率门的外部力，用于无传感器的力感知，以及（iii）用于监督关节的电机状态评分。我们还引入了神经驱动数据集（Neural Actuation Dataset, NAD），该数据集通过双臂遥操作系统收集，记录机器人状态和驱动器遥测数据以及外部力标签。扭矩替代头通过可微分模拟从姿态轨迹中训练，而无需直接的广义努力标签，而力、门和电机状态头则接受直接监督。Transformer捕捉时间依赖性，同时支持实时推断。我们在5自由度的OpenManipulator-X、6自由度的SO-101和7自由度的Franka Emika Panda上评估了NeuralActuator，这些平台涵盖了三种驱动器系列，成本从约500美元到超过30,000美元不等。低成本平台支持动力学和力评估，而离线Franka实验提供了额外的载荷力估计基准。实验进一步展示了其在OpenManipulator-X上的电机状态估计应用，以及当NeuralActuator作为预训练模块使用时，行为克隆性能的提升。
cs.RO / 93 / 2607.11739
AutoPath: Learning Transferable Goal-Conditioned Stochastic Path Prior for Safe Navigation Without Human Demonstrations
AutoPath：学习可转移的目标条件随机路径先验以实现无人工示范的安全导航
Abstract
Real-time navigation in cluttered and dynamic environments requires collision-free and dynamically feasible motion under limited perception. However, feasible navigation behaviors are inherently multimodal because multiple paths may exist around obstacles. In this paper, we formulate navigation as learning a transferable goal-conditioned stochastic path prior that models a reusable distribution over goal-aligned geometry-consistent local paths conditioned on local observations. This formulation enables structured sampling of navigation candidates, allowing multiple feasible paths to be explored through sampling without relying on robot-specific motion constraints. To this end, we introduce a goal-aligned canonical state representation that removes in-plane rotational ambiguity and normalizes local geometry with respect to the goal, enabling rotation-invariant path distribution learning. We further develop a structured prior learning framework that parameterizes local paths using a geometry-aware polar action manifold and incorporates risk-sensitive utility shaping with multi-goal distributional rollouts for stable and safety-aware planning. Extensive experiments in dense static environments and dynamic pedestrian scenarios demonstrate that the proposed method achieves consistently high success rates with competitive efficiency while enabling cross-platform transfer of a single path prior learned on differential-drive robots to quadruped platforms without retraining.
Chinese Translation
在复杂和动态环境中进行实时导航需要在有限感知下实现无碰撞和动态可行的运动。然而，可行的导航行为本质上是多模态的，因为在障碍物周围可能存在多条路径。本文将导航形式化为学习一种可转移的目标条件随机路径先验，该先验建模了基于局部观测的目标对齐几何一致局部路径的可重用分布。这种形式化方法使得导航候选路径的结构化采样成为可能，允许通过采样探索多条可行路径，而无需依赖于特定机器人的运动约束。为此，我们引入了一种目标对齐的典型状态表示，消除了平面内的旋转模糊，并根据目标规范化局部几何，从而实现旋转不变的路径分布学习。我们进一步开发了一种结构化先验学习框架，该框架使用几何感知的极坐标动作流形对局部路径进行参数化，并结合风险敏感的效用塑形与多目标分布回放，以实现稳定和安全意识的规划。在密集静态环境和动态行人场景中的大量实验表明，所提出的方法在实现竞争效率的同时，始终保持高成功率，并能够将在差动驱动机器人上学习到的单一路径先验跨平台转移到四足平台，而无需重新训练。
cs.RO / 94 / 2607.11779
A Compact Top-Loading Robot for Endovascular Interventions: Design, Control and Evaluation
一种紧凑型顶装机器人用于血管内介入：设计、控制与评估
Abstract
Robot-assisted endovascular intervention can potentially reduce radiation exposure, improve surgeon ergonomics, enable telesurgery, support active assistance and autonomy, and enhance procedural precision. However, existing systems often suffer from limited procedural coverage because constrained patient-side setups, restricted flexibility, and complex instrument exchange hinder clinical workflow integration. This work presents a compact robotic system for endovascular interventions that enables continuous translational and rotational manipulation of standard endovascular instruments. The system consists of two alternating carts with pneumatically actuated membrane grippers integrated into rotating gripper gears. Its top-loading design allows rapid exchange of instruments such as guidewires and catheters without changing the robotic setup. A leader-follower control strategy enables continuous motion despite the finite stroke of each cart. The system was evaluated in motion-tracking experiments with guidewires and catheters and in an in vitro vascular phantom. The motion-tracking experiments showed generally smooth translational and rotational motion profiles. Across all tested guidewire and catheter experiments, the mean relative tracking errors were 3.6% for translational motion and 4.1% for rotational motion. In the vascular phantom, robot-assisted navigation reached the target in most trials, demonstrating the feasibility of the proposed manipulation concept under in vitro conditions. The presented robotic system demonstrates technical feasibility for continuous manipulation of standard endovascular instruments in bench-top and in vitro experiments. The compact top-loading design may ease instrument exchange and clinical workflow integration. Future work will focus on improving gripping performance, actuation speed, force feedback, and evaluation in more clinically realistic settings.
Chinese Translation
机器人辅助的血管内介入有潜力减少辐射暴露、改善外科医生的工作 ergonomics、实现远程手术、支持主动辅助和自主操作，并提高手术精度。然而，现有系统往往因患者侧设置受限、灵活性不足以及复杂的器械更换而限制了手术覆盖范围，妨碍了临床工作流程的整合。本研究提出了一种紧凑型机器人系统，用于血管内介入，能够对标准血管内器械进行连续的平移和旋转操作。该系统由两个交替的推车组成，配备气动驱动的膜抓手，集成于旋转抓手齿轮中。其顶装设计允许快速更换导丝和导管等器械，而无需更改机器人设置。采用领导-跟随控制策略，使得尽管每个推车的行程有限，仍能实现连续运动。该系统在导丝和导管的运动跟踪实验以及体外血管模型中进行了评估。运动跟踪实验显示出一般平滑的平移和旋转运动特征。在所有测试的导丝和导管实验中，平移运动的平均相对跟踪误差为3.6%，旋转运动的平均相对跟踪误差为4.1%。在血管模型中，机器人辅助导航在大多数试验中成功到达目标，展示了所提出的操作概念在体外条件下的可行性。所展示的机器人系统在台式和体外实验中显示出对标准血管内器械进行连续操作的技术可行性。紧凑的顶装设计可能简化器械更换和临床工作流程的整合。未来的工作将集中在提高抓取性能、驱动速度、力反馈以及在更具临床现实性的环境中进行评估。
cs.RO / 95 / 2607.11785
MIRA: A Modular Open-Source Micro-UAV for Indoor Research
MIRA：一种模块化开源微型无人机用于室内研究
Abstract
Indoor robotics research increasingly relies on micro-UAVs whose airframe, electronics, and control software are fully open to modification. Off-the-shelf platforms rarely expose the low-level access required for such modifications, while building a custom alternative typically requires substantial engineering effort before flight testing can begin, leaving many laboratories to work within constraints that limit the scope of their research. We present MIRA (Modular Indoor Research Architecture), a low-cost, open-source micro-UAV for indoor research built around a replicable 3D-printed PLA airframe and a containerized low-level software package managing the companion-to-autopilot communication bridge via Micro XRCE-DDS. Designed as a white-box architecture, core subsystems are individually replaceable without firmware refactoring, supporting local fabrication and component substitution from existing lab inventory. We characterize MIRA through manual flight in position-control mode within an optical motion-capture volume, where the communication pipeline sustains a median companion-to-autopilot latency of 0.02 ms and power spectral density analysis confirms the structural vibration energy stays concentrated in a narrow 90 to 110 Hz band, isolated from the sub-20 Hz control bandwidth and within the autopilot safety thresholds.
Chinese Translation
室内机器人研究日益依赖于微型无人机（micro-UAV），其机身、电子设备和控制软件均可完全修改。现成的平台很少提供进行此类修改所需的低级访问权限，而构建定制替代品通常需要在飞行测试开始之前进行大量工程工作，这使得许多实验室在研究范围上受到限制。我们提出了MIRA（模块化室内研究架构），这是一种低成本、开源的微型无人机，专为室内研究而设计，基于可复制的3D打印PLA机身和一个容器化的低级软件包，通过Micro XRCE-DDS管理伴随控制器与自动驾驶仪之间的通信桥。MIRA被设计为一种白盒架构，核心子系统可以单独更换而无需固件重构，支持本地制造和从现有实验室库存中替换组件。我们通过在光学运动捕捉区域内的手动飞行（位置控制模式）来表征MIRA，在此过程中，通信管道维持了0.02毫秒的中位伴随控制器与自动驾驶仪延迟，功率谱密度分析确认结构振动能量集中在90至110赫兹的狭窄频带内，远离20赫兹以下的控制带宽，并在自动驾驶仪安全阈值之内。
cs.RO / 96 / 2607.11792
Casting Everything to Online API Services? A Survey of Integrating Localized Speech Recognition Models in Robotic Systems
将一切投向在线API服务？本地化语音识别模型在机器人系统中的集成调查
Abstract
Automatic speech recognition (ASR) has become a critical component of modern robotic systems because it is one of the most natural and intuitive ways for humans to interact with robots. A commonly used method is to directly use API services online. But is that all we can do? This article provides an overview of how ASR technologies are integrated into various intelligent robots and machines. We discuss the evolution of speech recognition from established approaches to state-of-the-art deep learning models, such as OpenAI's Whisper. We also list large-scale datasets and open source toolkits that have been widely used in both industry and academia. We structure the survey around ASR model families, deployment strategies in robotics (especially ROS-based, cloud-based, and hybrid solutions), and several real-world robotic platforms. Finally, we outline the challenges of deploying robust speech recognition in robots and discuss future directions, including multimodal interaction in diverse and dynamic environments. This paper can help social robotics researchers better navigate the emerging domain of language-based natural human-robot interaction.
Chinese Translation
自动语音识别（ASR）已成为现代机器人系统的关键组成部分，因为它是人类与机器人互动的最自然和直观的方式之一。常用的方法是直接使用在线API服务。但这就是我们能做的全部吗？本文概述了ASR技术如何集成到各种智能机器人和机器中。我们讨论了语音识别从传统方法到最先进的深度学习模型（如OpenAI的Whisper）的演变。我们还列出了在工业和学术界广泛使用的大规模数据集和开源工具包。我们围绕ASR模型家族、机器人中的部署策略（特别是基于ROS的、基于云的和混合解决方案）以及几个真实世界的机器人平台构建了调查。最后，我们概述了在机器人中部署稳健语音识别的挑战，并讨论了未来的方向，包括在多样化和动态环境中的多模态交互。本文可以帮助社会机器人研究人员更好地导航语言基础的人机自然交互这一新兴领域。
cs.RO / 97 / 2607.11855
Robust bipedal locomotion on flowable slopes via foot-driven terrain manipulation
通过足驱动的地形操控实现流动坡面上的稳健双足行走
Abstract
Bipedal robots are challenging to control because they operate close to instability, where small variations in foot-terrain contact can rapidly destabilize locomotion. On rigid terrain, bipedal robots mitigate this fragility by using well-established contact mechanics and control strategies. On flowable surfaces such as granular slopes, foot contact can induce large surface deformations and solid-fluid-like transitions, coupling terrain effects with robot dynamics, leading to underperformance or failure. This is partly due to the lack of reliable methods to represent the dynamics of flowable terrain, making it difficult to account for terrain effects in locomotion design. Here, we investigate how controlling terrain response can improve bipedal locomotion on granular slopes by studying the terradynamics of cleated feet, thin plates emanating from the foot soles. Systematic studies of a small-scale (1.4 kg) robophysical biped reveal that cleats with sparse and dense spacing lead to excessive terrain yielding and resistance, respectively, degrading performance and leading to failure. An intermediate cleat spacing distributes interaction forces to maintain substrate stresses near (or below) the yield threshold, enabling walking on granular slopes up to 30 degrees. Guided by these principles, we design a foot that actively adjusts cleat depth and accommodates both rigid and granular terrain. We also demonstrate that the principles of effective foot-terrain interaction translate to a larger (15 kg) autonomous biped. Our study presents an alternative to conventional body-centric robot control approaches, which regulate terrain-induced effects through body motion, by instead regulating terrain interactions through limb-centric approach.
Chinese Translation
双足机器人因其在不稳定边缘操作而难以控制，脚与地形接触的微小变化可能迅速导致行走不稳定。在刚性地形上，双足机器人通过使用成熟的接触力学和控制策略来减轻这种脆弱性。然而，在如颗粒坡面等流动表面上，脚接触可能引起较大的表面变形和固-液相似的转变，将地形效应与机器人动力学耦合，导致性能下降或失败。这部分是由于缺乏可靠的方法来表示流动地形的动态特性，使得在行走设计中难以考虑地形效应。在此，我们研究了控制地形响应如何改善颗粒坡面上的双足行走，重点研究了来自鞋底的带钉足的地形动力学。对一款小型（1.4 kg）机器人双足的系统研究表明，稀疏和密集间距的钉子分别导致过度的地形屈服和阻力，降低了性能并导致失败。中等钉子间距则分配了相互作用力，以保持基底应力接近（或低于）屈服阈值，从而使得在高达30度的颗粒坡面上行走成为可能。在这些原则的指导下，我们设计了一种能够主动调整钉子深度并适应刚性和颗粒地形的足部。我们还展示了有效的足-地形相互作用原则可以转化到更大（15 kg）的自主双足机器人上。我们的研究为传统的以身体为中心的机器人控制方法提供了一种替代方案，后者通过身体运动调节地形引起的效应，而是通过以肢体为中心的方法调节地形相互作用。
cs.RO / 98 / 2607.11874
A Minimalist Retargeting-Guided Reinforcement Learning Recipe for Dexterous Manipulation
一种极简主义的重定向引导强化学习方案用于灵巧操作
Abstract
Recent work in humanoid whole-body control has found success with a simple recipe: retarget human motion to robot kinematic references, then train policies via reinforcement learning (RL) to track them. But how does this recipe transfer to dexterous manipulation? The answer is not obvious, as manipulation involves complex, contact-rich dynamics and requires delicate regulation of contact modes and forces. We present REGRIND, a minimalist retargeting-guided RL pipeline that learns dexterous manipulation policies from a single human demonstration. REGRIND retargets human hand-object motion to a robot reference that preserves hand-object spatial and contact relationships, trains a residual RL policy in simulation to track object-centric keypoints along that reference, and transfers the resulting policy zero-shot to hardware with careful system identification. The resulting policies produce fluid, human-like behavior on two different multi-fingered hands across contact-rich tool-use tasks, including operating a pair of scissors and turning a screwdriver. Through systematic hardware experiments, we identify and analyze the key factors that govern sim-to-real transfer in dexterous manipulation, offering practical guidance for retargeting-based learning in contact-rich settings. Videos and code are available at https://yunhaifeng.com/REGRIND.
Chinese Translation
近期在类人全身控制方面的研究取得了成功，采用了一种简单的方案：将人类动作重定向到机器人运动学参考，然后通过强化学习（RL）训练策略以跟踪这些参考。但这一方案如何转移到灵巧操作上并不明显，因为操作涉及复杂的接触丰富的动力学，并需要对接触模式和力量进行精细调节。我们提出了REGRIND，这是一种极简主义的重定向引导RL管道，从单一的人类示范中学习灵巧操作策略。REGRIND将人类手-物体运动重定向到一个保留手-物体空间和接触关系的机器人参考，在仿真中训练一个残差RL策略以沿该参考跟踪以物体为中心的关键点，并通过仔细的系统识别将生成的策略零-shot转移到硬件上。最终的策略在两个不同的多指手上产生流畅的类人行为，涵盖了接触丰富的工具使用任务，包括操作一把剪刀和旋转螺丝刀。通过系统的硬件实验，我们识别并分析了影响灵巧操作中仿真到现实转移的关键因素，为接触丰富环境中的重定向学习提供了实用指导。视频和代码可在 https://yunhaifeng.com/REGRIND 获取。
cs.RO / 99 / 2607.11884
Mixture of Frames Policy: Multi-Frame Action Denoising for Bimanual Mobile Manipulation
帧混合策略：双手移动操作的多帧动作去噪
Abstract
Robotic manipulation is inherently multi-frame: local actions may be simple in an end-effector frame, while transport, upright-object handling, and whole-body coordination are better represented in a base-aligned frame. However, modern diffusion-based visuomotor policies typically commit to a single predefined action frame, forcing one denoiser to model action distributions that are often unnecessarily complex in that frame. We propose Mixture of Frames Policy (MoF), a diffusion policy that performs synchronized action denoising across multiple coordinate frames. MoF maintains a single canonical diffusion state, re-expresses it in several task-relevant frames, applies frame-specialized denoisers, and fuses their noise predictions back in the canonical frame. To make this possible for intermediate noisy diffusion states, we introduce a column-based 6D rotation representation within an SE(3) action parameterization that supports exact, differentiable frame transformations without requiring noisy rotations to lie on the SO(3) manifold. Across nine simulated bimanual manipulation tasks, we show that the best action frame is task-dependent and that MoF improves over oracle frame selection and standard Mixture-of-Experts (MoE) baselines. We further evaluate MoF on two real-world bimanual mobile manipulation tasks, demonstrating that it outperforms all constituent single-frame baselines. Project homepage: https://mofpo.github.io
Chinese Translation
机器人操作本质上是多帧的：在末端执行器框架中，局部动作可能是简单的，而运输、直立物体处理和全身协调在基准对齐框架中更易于表示。然而，现代基于扩散的视觉运动策略通常承诺于单一预定义的动作框架，迫使一个去噪器在该框架中建模往往不必要复杂的动作分布。我们提出了帧混合策略（Mixture of Frames Policy, MoF），这是一种在多个坐标框架中执行同步动作去噪的扩散策略。MoF维持一个单一的规范扩散状态，在多个与任务相关的框架中重新表达它，应用框架专用的去噪器，并将它们的噪声预测融合回规范框架。为了使这一过程适用于中间噪声扩散状态，我们在SE(3)动作参数化中引入了一种基于列的6D旋转表示，该表示支持精确的、可微分的框架变换，而无需噪声旋转位于SO(3)流形上。在九个模拟的双手操作任务中，我们展示了最佳动作框架是任务依赖的，并且MoF在预言框架选择和标准的专家混合（Mixture-of-Experts, MoE）基线之上有所改善。我们进一步在两个真实世界的双手移动操作任务上评估MoF，证明其优于所有组成的单帧基线。项目主页：https://mofpo.github.io
cs.CV / 1 / 2607.09753
Unified Backbone Refinement for Diffusion Models via Internal-Latent Analysis
通过内部潜变量分析对扩散模型进行统一骨干精炼
Abstract
Diffusion models have achieved remarkable success across diverse domains, with performance closely related to the denoising backbones that parameterize the score function. In this paper, we present a systematic, phase-aware analysis of diffusion components and show that abrupt, early-stage fluctuations in deep latents are strongly associated with artifacts. Guided by these findings, we introduce DUNE (Diffusion Unified Network refiNEr), a training-free refinement framework that detects abrupt deviations in deep low-noise internal latents using a shared EMA-based criterion, and applies backbone-specific suppression to the detector-selected entries. Although derived from U-Net, the same detect-suppress principle extends naturally to Transformer-based diffusion models by acting on the latents of deep self-attention blocks. Extensive experiments across multiple backbones indicate that DUNE improves fidelity while reducing hallucinations, offering new insight into where and when diffusion backbones should be controlled.
Chinese Translation
扩散模型在多个领域取得了显著成功，其性能与参数化得分函数的去噪骨干密切相关。本文提出了一种系统的、相位感知的扩散组件分析，表明深层潜变量在早期阶段的突然波动与伪影有很强的关联。基于这些发现，我们引入了 DUNE（Diffusion Unified Network refiNEr），这是一个无训练的精炼框架，利用共享的基于 EMA 的标准检测深层低噪声内部潜变量中的突然偏差，并对检测器选择的条目应用特定于骨干的抑制。尽管源自 U-Net，但相同的检测-抑制原则自然扩展到基于 Transformer 的扩散模型，通过作用于深层自注意力块的潜变量。针对多个骨干的广泛实验表明，DUNE 提高了保真度，同时减少了幻觉，为何时以及如何控制扩散骨干提供了新的见解。
cs.CV / 2 / 2607.09754
Cross-Subject Modeling for Widefield Calcium Imaging via Atlas-Aligned Spatiotemporal Tokenization
基于图谱对齐的时空标记化的跨主体宽场钙成像建模
Abstract
Large-scale, multi-subject widefield calcium imaging provides unprecedented access to brain-wide cortical dynamics. However, the high dimensionality, complex spatiotemporal structure, and substantial task-irrelevant activity in widefield recordings have largely restricted modeling efforts to single-session analyses, limiting scalability and generalization. While multi-subject pretrained models have been explored for some neural modalities, multi-subject models for widefield calcium imaging have not yet been demonstrated; further, subject-invariant zero-shot behavior decoding remains elusive for multi-subject models across neural modalities more broadly. As a first step toward foundation modeling of widefield data, we introduce WiCAT, a multi-subject model that leverages self-supervised pretraining to both outperform single-session models and enable zero-shot behavior decoding on unseen subjects. WiCAT introduces an atlas-grounded tokenization scheme without session-specific components and learns globally shared spatiotemporal representations. Across multiple widefield datasets, the pretrained model supports lightweight downstream decoding, transfers across subjects, tasks, and datasets, and outperforms baseline models. Notably, the model also achieves robust zero-shot continuous behavior decoding and left-out brain region reconstruction on unseen subjects.
Chinese Translation
大规模多主体宽场钙成像为脑广泛皮层动态提供了前所未有的访问。然而，宽场记录中的高维度、复杂的时空结构以及大量与任务无关的活动在很大程度上限制了建模工作仅限于单会话分析，从而限制了可扩展性和泛化能力。尽管已经探索了一些神经模态的多主体预训练模型，但尚未展示宽场钙成像的多主体模型；此外，跨神经模态的主体不变零样本行为解码仍然难以实现。作为宽场数据基础建模的第一步，我们引入了WiCAT，这是一种多主体模型，利用自监督预训练，不仅超越了单会话模型，还能够在未见主体上实现零样本行为解码。WiCAT引入了一种基于图谱的标记化方案，没有会话特定组件，并学习全球共享的时空表示。在多个宽场数据集中，预训练模型支持轻量级下游解码，能够在不同主体、任务和数据集之间进行迁移，并超越基线模型。值得注意的是，该模型还在未见主体上实现了稳健的零样本连续行为解码和遗漏脑区重建。
cs.CV / 3 / 2607.09757
RSLoRA: Training-free Rank Allocation for LoRA via Representational Sensitivity Probing
RSLoRA：通过表征敏感性探测实现无训练的LoRA秩分配
Abstract
Low-Rank Adaptation (LoRA) has become a cornerstone of parameter-efficient fine-tuning (PEFT); however, the conventional practice of uniform rank assignment ignores the functional heterogeneity of neural layers. Existing rank allocation methods typically struggle with a trade-off between computational intensity and heuristic simplicity: training-based methods suffer from prohibitive overhead, while pre-allocation methods fail to capture the dynamic task-specific representation manifold. In this paper, we propose RSLoRA (Representational Sensitivity LoRA), a training-free and gradient-free rank allocator driven by activation-space geometry. We identify a "sensitivity regime shift" across layers, observing that static weight analysis and local gradients are insufficient to reflect how updates reshape a model's internal representations. To address this, RSLoRA introduces a virtual representational probing mechanism. By simulating adaptation through structured low-rank noise and measuring the resulting manifold displacement by using Effective Rank and Frechet Distance, we identify high-sensitivity modules that require higher rank capacity. Our framework effectively bridges the gap between expert-crafted heuristics and actual representational impact. Extensive evaluations demonstrate that RSLoRA consistently outperforms state-of-the-art allocators (e.g., AdaLoRA, GoRA) across mainstream benchmarks. By eliminating the need for iterative training-time adjustments and backward gradients, RSLoRA provides a highly efficient, robust, and representation-aware solution for large-scale model adaptation.
Chinese Translation
低秩适应（LoRA）已成为参数高效微调（PEFT）的基石；然而，传统的均匀秩分配方法忽视了神经层的功能异质性。现有的秩分配方法通常在计算强度和启发式简单性之间存在权衡：基于训练的方法面临着巨大的开销，而预分配方法无法捕捉动态任务特定的表征流形。在本文中，我们提出了RSLoRA（表征敏感性LoRA），这是一种无训练和无梯度的秩分配器，基于激活空间几何。我们识别出层间的“敏感性状态转变”，观察到静态权重分析和局部梯度不足以反映更新如何重塑模型的内部表征。为了解决这个问题，RSLoRA引入了一种虚拟表征探测机制。通过模拟结构化低秩噪声的适应，并使用有效秩（Effective Rank）和弗雷歇距离（Frechet Distance）测量结果流形位移，我们识别出需要更高秩容量的高敏感性模块。我们的框架有效地弥合了专家设计的启发式方法与实际表征影响之间的差距。广泛的评估表明，RSLoRA在主流基准测试中始终优于最先进的分配器（例如，AdaLoRA、GoRA）。通过消除对迭代训练时间调整和反向梯度的需求，RSLoRA为大规模模型适应提供了一种高效、稳健且关注表征的解决方案。
cs.CV / 4 / 2607.09759
ReflectWorld-MM: An Entity-Oriented Multi-Media Memory System for Open-Ended Video Streams
ReflectWorld-MM：一种面向实体的多媒体记忆系统用于开放式视频流
Abstract
Building assistants that can continually watch the world, remember what they see, and reason over their accumulated experience is a long-standing goal, and recently multimodal agents equipped with long-term memory over video streams have attracted increasing interest. Unfortunately, existing systems either keep their memory inside the model context or in a flat feature store, and organize it around frames rather than around the persistent entities a stream is really about, which confines them to bounded videos and weakens their ability to track who and what reappears over time. In this paper, we propose ReflectWorld-MM, an entity-oriented multi-media memory system for open-ended video streams. It consists of three parts. The first is a perception front-end that turns a streaming video into entity-resolved observations under a bounded short-term memory. The second is a hierarchical long-term memory, grounded in human memory theory, that couples a multi-scale episodic memory, an evolving entity-centric semantic memory, and a procedural memory. The third is a complete realization, built for real-world operation, that ingests arbitrary streams and plugs into off-the-shelf assistants. Across six long-video and lifelong-memory benchmarks, ReflectWorld-MM achieves the best accuracy on all six, outperforming strong memory agents and a frontier model.
Chinese Translation
构建能够持续观察世界、记住所见并对其积累的经验进行推理的助手一直是一个长期目标，最近，配备长期记忆的视频流多模态代理引起了越来越多的关注。不幸的是，现有系统要么将其记忆保留在模型上下文中，或在一个扁平特征存储中，并围绕帧进行组织，而不是围绕流中真正涉及的持久实体，这限制了它们对有限视频的适应能力，并削弱了它们随时间追踪谁和什么重新出现的能力。本文提出了ReflectWorld-MM，一种面向实体的多媒体记忆系统，用于开放式视频流。它由三个部分组成。第一部分是一个感知前端，将流媒体视频转化为在有限短期记忆下的实体解析观察。第二部分是一个基于人类记忆理论的分层长期记忆，结合了多尺度的情节记忆、不断发展的实体中心语义记忆和程序性记忆。第三部分是一个完整的实现，旨在实际操作中，能够处理任意流并与现成的助手连接。在六个长视频和终身记忆基准测试中，ReflectWorld-MM在所有六个测试中都达到了最佳准确性，超越了强大的记忆代理和前沿模型。
cs.CV / 5 / 2607.09763
Knowledge-Constrained Shape Optimization with a Mixture-of-Experts Neural Operator for High-Confidence Design
基于知识约束的形状优化与混合专家神经算子的高置信度设计
Abstract
Engineering shape optimization faces challenges in both expert-dependent problem setup and surrogate-model reliability. In practical aerodynamic design, optimization settings such as editable regions, deformation ranges, and design-preservation constraints are typically specified manually by experienced engineers, while surrogate-based optimization may become unreliable for heterogeneous geometry databases and out-of-distribution designs. To address these challenges, we propose a knowledge-constrained shape-optimization framework that translates knowledge-based constraints and user intent into quantifiable parameters of DFFD-based deformation operators, enabling engineering-aware and controllable constrained optimization. We further develop a Mixture-of-Experts Neural Operator (MoE-NO) to improve drag prediction and trend consistency over heterogeneous aerodynamic datasets. Based on the MoE-NO encoder and Mahalanobis distance, an uncertainty-estimation strategy is introduced to detect out-of-distribution geometries and selectively trigger physics-solver feedback for local sample enrichment. Experiments on in-house MPV, SUV, and Sedan datasets show that MoE-NO achieves a test-set MAPE of $1.16\%$ and a trend-prediction accuracy of $94.34\%$, outperforming the best baseline results of $1.52\%$ and $90.34\%$, respectively. Vehicle shape-optimization experiments further yield CFD-validated drag coefficient reductions of approximately $4\%$ to $10\%$.
Chinese Translation
工程形状优化面临着依赖专家的问题设置和替代模型可靠性等挑战。在实际的气动设计中，诸如可编辑区域、变形范围和设计保持约束等优化设置通常由经验丰富的工程师手动指定，而基于替代模型的优化在异构几何数据库和超出分布的设计中可能变得不可靠。为了解决这些挑战，我们提出了一种基于知识约束的形状优化框架，该框架将基于知识的约束和用户意图转化为可量化的基于DFFD（Deformation-Field-Driven）变形算子的参数，从而实现工程意识和可控的约束优化。我们进一步开发了一种混合专家神经算子（Mixture-of-Experts Neural Operator, MoE-NO），以提高在异构气动数据集上的阻力预测和趋势一致性。基于MoE-NO编码器和马哈拉诺比斯距离，我们引入了一种不确定性估计策略，以检测超出分布的几何形状，并选择性地触发物理求解器反馈以进行局部样本丰富。在内部MPV、SUV和轿车数据集上的实验表明，MoE-NO在测试集上的平均绝对百分比误差（MAPE）为1.16%，趋势预测准确率为94.34%，分别优于最佳基线结果的1.52%和90.34%。车辆形状优化实验进一步实现了经过CFD验证的阻力系数减少约4%至10%。
cs.CV / 6 / 2607.09768
Low-Power License Plate Detection and Recognition on a RISC-V Multi-Core MCU-Based Vision System
基于RISC-V多核MCU的低功耗车牌检测与识别视觉系统
Abstract
In this paper, we present the first (to the best of our knowledge) demonstration of a low-power MCU-based edge device for Automatic License Plate Recognition (ALPR). The design leverages on a 9-core RISC-V processor, GAP8, coupled with a QVGA ultra-low-power greyscale imager. The proposed visual processing pipeline uses a multi-model inference approach based on SSDlite-MobilenetV2 for license plate detection and LPRNet for optical character recognition, reaching a 38.9% mAP score for the first task and a recognition rate of >99.13% for the latter on public datasets. On real-world data, the pipeline recognizes registration numbers when the size of LP crops is as small as 30x5 pixels. Thanks to the applied compression and optimization strategies, the multi-model inference (687 MMAC) achieves a throughput of 1.09 FPS at a power cost of 117 mW when running on GAP8. Our solution is the first MCU-class device embedding such a level of network complexity, resulting to be 73x more energy-efficient w.r.t. precedent mobile-class ALPR system featuring a Raspberry Pi3. The proposed design does not resort to any hardwired acceleration engines, thus retaining full flexibility for future algorithmic improvements.
Chinese Translation
本文展示了一个基于低功耗MCU的边缘设备用于自动车牌识别（ALPR）的首次（据我们所知）演示。该设计利用了一个9核RISC-V处理器GAP8，并配备了QVGA超低功耗灰度成像器。所提出的视觉处理管道采用基于SSDlite-MobilenetV2的多模型推理方法进行车牌检测，并使用LPRNet进行光学字符识别，在公共数据集上，前者达到38.9%的mAP分数，后者的识别率超过99.13%。在实际数据中，当车牌裁剪的大小小至30x5像素时，该管道仍能识别注册号码。得益于所采用的压缩和优化策略，运行在GAP8上的多模型推理（687 MMAC）在117 mW的功耗下实现了1.09 FPS的吞吐量。我们的解决方案是首个嵌入如此网络复杂度的MCU级设备，相较于之前的移动级ALPR系统（如Raspberry Pi3），其能效提高了73倍。所提出的设计不依赖任何硬件加速引擎，从而为未来的算法改进保留了充分的灵活性。
cs.CV / 7 / 2607.09777
Time Imprint: Learning Time-Aware Representations in Multi-Modal Knowledge Graphs
时间印记：在多模态知识图谱中学习时间感知表示
Abstract
Multi-Modal Knowledge Graphs (MMKGs) enrich entities with multiple modalities such as text and images, yet entities with highly similar multi-modal features remain difficult to distinguish. Temporal information of an entity can serve as an additional modality to disambiguate such entities, but existing approaches rarely treat time as a separate modality alongside text and images due to two major challenges: (1) sparse temporal semantics, which hinder alignment with richer modalities, and (2) multiple timestamps, which introduce noise or reduce robustness in representation learning. To address these challenges, we propose Time Imprint, a framework that treats time as an entity-level modality and jointly aligns temporal, textual, and visual representations via a three-view contrastive objective. Additionally, to mitigate multi-timestamp ambiguity, Time Imprint studies a compact timestamp subset selection design space and aggregates the selected timestamps into a discriminative temporal embedding with attention pooling, balancing temporal specificity and robustness. Experiments on three MMKG benchmarks demonstrate that Time Imprint achieves state-of-the-art link prediction performance, improving Hits@1 by up to 6.07\% overall and yielding up to 58\% gains on the subset of the top-1\% ambiguity samples. We further examine different fusion strategies and the sensitivity to timestamp availability and quality, clarifying when and why time-as-modality is most beneficial, while adding only modest training overhead. We release our code at https://anonymous.4open.science/r/Time-Imprint.
Chinese Translation
多模态知识图谱（MMKGs）通过文本和图像等多种模态丰富实体，但具有高度相似多模态特征的实体仍然难以区分。实体的时间信息可以作为一种额外模态来消歧义，但现有方法很少将时间视为与文本和图像并列的独立模态，主要面临两个挑战：（1）稀疏的时间语义，妨碍与更丰富模态的对齐；（2）多个时间戳，导致噪声或降低表示学习的鲁棒性。为了解决这些挑战，我们提出了时间印记（Time Imprint），一个将时间视为实体级模态的框架，通过三视图对比目标共同对齐时间、文本和视觉表示。此外，为了减轻多时间戳的歧义，时间印记研究了紧凑时间戳子集选择设计空间，并通过注意力池化将选定的时间戳聚合为具有区分性的时间嵌入，平衡时间特异性和鲁棒性。在三个MMKG基准上的实验表明，时间印记实现了最先进的链接预测性能，整体上提高了Hits@1最多6.07\%，在前1\%歧义样本的子集上获得了高达58\\%的增益。我们进一步考察了不同的融合策略以及对时间戳可用性和质量的敏感性，阐明了何时以及为何将时间作为模态最为有利，同时仅增加了适度的训练开销。我们的代码已发布在 https://anonymous.4open.science/r/Time-Imprint。
cs.CV / 8 / 2607.09779
A Generalized Deep Non-negative Matrix Factorization Approach for SAR Automatic Target Recognition
一种广义深度非负矩阵分解方法用于合成孔径雷达自动目标识别
Abstract
The deep nonnegative matrix factorization (DNMF) technique is proposed to address the low interpretability of deep learning-based methods in extracting multilayer features from synthetic aperture radar (SAR) target samples. However, existing DNMF methods employ a layer-by-layer decomposition strategy, which is prone to causing error accumulation and local optimum, thereby hindering a consistent improvement in recognition accuracy as the number of layer increases. In this paper, a robust multilayer feature extraction method, termed generalized deep non-negative matrix factorization (G-DNMF), is proposed to address the above challenges in SAR automatic target recognition (ATR). The G-DNMF aims global optimality and derives the update rules for each parameter using lagrangian multiplier method. The new update formula indicates that both the DNMF method based on the encoding matrix and the mixing matrix are special cases of the proposed method, theoretically demonstrating the universality of proposed method. In general, the proposed method discards the layer-by-layer decomposition strategy, thereby effectively mitigating the risk of local optima and eliminating error accumulation, leading to a significant improvement in DNMF's multi-layer feature extraction capability. The experimental results, by presenting the feature images extracted from each layer by G-DNMF and the reconstructed original images, verified the proposed method's pure additive understanding of multi-layer features and demonstrated its interpretability. The experimental results based on MSTAR and OpenSARship datasets show that G-DNMF outperforms existing DNMF algorithms and their derivatives in terms of stability and recognition performance.
Chinese Translation
提出了一种深度非负矩阵分解（DNMF）技术，以解决基于深度学习的方法在提取合成孔径雷达（SAR）目标样本的多层特征时低可解释性的问题。然而，现有的DNMF方法采用逐层分解策略，这容易导致误差累积和局部最优，从而阻碍了随着层数增加而在识别准确性上的持续提升。本文提出了一种稳健的多层特征提取方法，称为广义深度非负矩阵分解（G-DNMF），以应对SAR自动目标识别（ATR）中的上述挑战。G-DNMF旨在实现全局最优，并使用拉格朗日乘子法推导每个参数的更新规则。新的更新公式表明，基于编码矩阵和混合矩阵的DNMF方法是所提方法的特例，理论上证明了所提方法的普适性。总体而言，所提方法摒弃了逐层分解策略，从而有效减轻了局部最优的风险，消除了误差累积，显著提升了DNMF的多层特征提取能力。实验结果通过展示G-DNMF从每一层提取的特征图像及重建的原始图像，验证了所提方法对多层特征的纯加性理解，并展示了其可解释性。基于MSTAR和OpenSARship数据集的实验结果表明，G-DNMF在稳定性和识别性能方面优于现有的DNMF算法及其衍生方法。
cs.CV / 9 / 2607.09780
Towards Real-World Wearable Motion Reconstruction
面向现实世界的可穿戴运动重建
Abstract
The modern-day surge in popularity of wearable devices poses a fundamentally unique motion capture problem: reconstructing full-body movement from any set of sensing hardware worn at a given moment. Yet, most research efforts assume fixed sensor configurations (e.g. IMU suits or HMD-centric rigs) and cannot generalize across them. In contrast, we argue that motion capture should prioritize unobtrusive and lightweight devices such as smartphones, smartwatches, smart glasses, and smart insoles, and study the interplay between them. To this end, we make three contributions. First, we present a large-scale multi-modal dataset synchronizing these consumer-grade sensors with ground-truth 3D motion, spanning 50 diverse activities including everyday tasks, sports, and social interactions. Second, we propose WHIP, a baseline generative model that reconstructs motion from arbitrary subsets of available sensors, robustly handling missing modalities and producing physically plausible motions. Third, we conduct a systematic study of sensor complementarity, quantifying how different modalities complement one another. Code and dataset are available at https://vcai.mpi-inf.mpg.de/projects/WHIP/
Chinese Translation
现代可穿戴设备的普及带来了一个根本独特的运动捕捉问题：从任何一组在特定时刻佩戴的传感器硬件中重建全身运动。然而，大多数研究工作假设固定的传感器配置（例如，IMU 套装或以头戴显示器为中心的设备），并且无法在这些配置之间进行推广。相反，我们认为运动捕捉应优先考虑不显眼且轻便的设备，如智能手机、智能手表、智能眼镜和智能鞋垫，并研究它们之间的相互作用。为此，我们做出了三项贡献。首先，我们呈现了一个大规模的多模态数据集，将这些消费级传感器与真实的三维运动同步，涵盖了50种多样的活动，包括日常任务、体育运动和社交互动。其次，我们提出了 WHIP，一个基线生成模型，能够从可用传感器的任意子集重建运动，稳健地处理缺失模态并生成物理上合理的运动。第三，我们进行了传感器互补性的系统研究，量化了不同模态之间的互补关系。代码和数据集可在 https://vcai.mpi-inf.mpg.de/projects/WHIP/ 获取。
cs.CV / 10 / 2607.09784
Compression Asymmetry and Trajectory Binding in Noise-Anchored Diffusion Inversion
噪声锚定扩散反演中的压缩不对称性与轨迹绑定
Abstract
Real-image diffusion inversion is governed by a tight quality-cost trade-off, with costs incurred in computation, storage, or per-image optimization. We study this trade-off through the forward Gaussian noise anchor that defines a diffusion trajectory and isolate two mechanisms behind effective stored-noise inversion. First, diffusion noise exhibits an element-wise compression asymmetry: int8 full-dimensional anchors preserve reconstruction, whereas low-dimensional subspace summaries are much less reliable, often collapsing even at comparable or smaller payloads; the element-wise over subspace ordering persists across five stored-noise inversion methods. Second, inversion is trajectory-bound and score-prior coupled: the matched forward anchor and a trained score network are both necessary, arguing against a purely algebraic-identity explanation. Together, these findings specify what to store and how to use it. They lead to Noise-Anchored Reverse Correction (NARC), a training-free inversion primitive that stores a single int8 latent anchor and reuses it with a fixed, noise-level-dependent anchor-weight schedule: strong anchoring when the reverse trajectory is noise-dominated, then relaxed anchoring as image detail emerges. On PIE-Bench++ with Stable Diffusion 1.5, NARC outperforms five modern non-exact baselines and improves PSNR by +3.24 dB over PnP DirectInv while using about 400x less inversion storage than PnP DirectInv. The compression asymmetry, anchor specificity, and editing plug-in also transfer to SDXL 1024^2.
Chinese Translation
真实图像的扩散反演受限于严格的质量-成本权衡，成本体现在计算、存储或每幅图像的优化上。我们通过定义扩散轨迹的前向高斯噪声锚点来研究这一权衡，并分离出有效存储噪声反演背后的两种机制。首先，扩散噪声表现出逐元素的压缩不对称性：int8全维锚点能够保持重建，而低维子空间摘要的可靠性则大大降低，常常在可比或更小的负载下崩溃；逐元素的优先于子空间的排序在五种存储噪声反演方法中持续存在。其次，反演是轨迹绑定的，并与得分先验相结合：匹配的前向锚点和训练好的得分网络都是必要的，这反驳了纯代数恒等式的解释。这些发现共同明确了应存储什么以及如何使用它。它们促成了噪声锚定反向校正（Noise-Anchored Reverse Correction, NARC），这是一种无训练的反演原语，存储一个单一的int8潜在锚点，并使用固定的、依赖于噪声水平的锚点权重调度进行重用：当反向轨迹受噪声主导时强锚定，然后随着图像细节的出现而放松锚定。在使用Stable Diffusion 1.5的PIE-Bench++上，NARC的性能超过了五个现代非精确基线，并且在使用约400倍更少的反演存储的情况下，相较于PnP DirectInv提高了+3.24 dB的PSNR。压缩不对称性、锚点特异性和编辑插件也可以转移到SDXL 1024^2。
cs.CV / 11 / 2607.09785
Lifelong Representations: A Survey on Continual Self-Supervised Learning for Vision Models
终身表征：关于视觉模型的持续自监督学习的综述
Abstract
Traditionally, continual learning has assumed access to labeled data, yet many real-world applications -- such as lifelong robotics -- require models to adapt continuously from unlabeled streams. This has led to the development of continual self-supervised learning (CSSL), a rapidly growing area that lacks a dedicated, systematic review. In this work, we present a comprehensive survey of CSSL for vision, with connections to emerging vision-language settings. First, we analyze existing evaluation protocols and highlight inconsistencies that hinder fair comparison. We then examine why self-supervised objectives exhibit improved robustness to catastrophic forgetting, relating this to task-agnostic representations and smoother loss landscapes. Next, we organize existing methods into a unified taxonomy based on their forgetting-mitigation strategies, including distillation, replay, regularization, architectural approaches, model merging, and objective-level adaptation. Finally, we identify open challenges such as scalability and the need for fast adaptability. We argue that advancing CSSL requires moving beyond small-scale benchmarks towards continual pre-training paradigms for large-scale systems.
Chinese Translation
传统上，持续学习假设可以访问标记数据，然而许多现实世界的应用——例如终身机器人——要求模型能够从未标记的数据流中持续适应。这导致了持续自监督学习（CSSL）的发展，这是一个快速增长的领域，但缺乏专门的系统性综述。在本研究中，我们对视觉领域的CSSL进行了全面的综述，并与新兴的视觉-语言设置建立了联系。首先，我们分析了现有的评估协议，并强调了妨碍公平比较的不一致性。接着，我们探讨了自监督目标为何表现出对灾难性遗忘的更强鲁棒性，将其与任务无关的表征和更平滑的损失景观联系起来。随后，我们根据现有方法的遗忘缓解策略将其组织成统一的分类法，包括蒸馏、重放、正则化、架构方法、模型合并和目标级适应。最后，我们识别出开放性挑战，如可扩展性和快速适应的需求。我们认为，推进CSSL需要超越小规模基准，转向大规模系统的持续预训练范式。
cs.CV / 12 / 2607.09787
Adversarially Guided Diffusion for LiDAR Range Image Synthesis
对抗引导扩散用于激光雷达范围图像合成
Abstract
LiDAR semantic segmentation is a key perception task in autonomous driving, where false predictions can affect downstream planning and safety-critical decision-making. Although adversarial attacks, and specifically adversarial examples, have been widely studied for image classification and 3D point cloud segmentation, unrestricted adversarial examples remain largely unexplored in the space of 2D range images, which are projections of 3D point clouds. The proposed method is, to the best of our knowledge, the first diffusion-based unrestricted adversarial attack against 2D range-image segmentation, using adversarial guidance from a segmentation loss. By applying guidance directly during sampling, the method produces unrestricted adversarial examples that remain close to the learned LiDAR data manifold while inducing structured segmentation errors. Experiments on the SemanticKITTI dataset using RangeNet++ and CENet segmentation networks demonstrate that the attack provides adjustable degradation across guidance strengths and transfers across segmentation architectures. Compared with norm-bounded FGSM and SegPGD baselines, the proposed attack offers a distinct effectiveness-realism trade-off, achieving controllable white-box and transfer degradation while maintaining competitive distributional and visual realism.
Chinese Translation
激光雷达语义分割是自动驾驶中的一项关键感知任务，其中错误预测可能影响下游规划和安全关键决策。尽管对抗攻击，特别是对抗样本，已在图像分类和三维点云分割中得到了广泛研究，但在二维范围图像（即三维点云的投影）领域中，未受限的对抗样本仍然基本未被探索。根据我们所知，所提出的方法是首个基于扩散的针对二维范围图像分割的未受限对抗攻击，利用来自分割损失的对抗引导。通过在采样过程中直接应用引导，该方法生成的未受限对抗样本在保持接近学习到的激光雷达数据流形的同时，诱导出结构化的分割错误。在使用RangeNet++和CENet分割网络的SemanticKITTI数据集上的实验表明，该攻击在引导强度上提供了可调的降级，并在分割架构之间实现了迁移。与有范数限制的FGSM和SegPGD基线相比，所提出的攻击提供了明显的有效性与现实性之间的权衡，实现了可控的白盒和迁移降级，同时保持了竞争性的分布和视觉现实性。
cs.CV / 13 / 2607.09788
MVMGNN;Multi-View Masked Graph Neural Network for Alzheimer's Disease Diagnosis using Structural MRI
MVMGNN：基于结构性MRI的多视角掩蔽图神经网络用于阿尔茨海默病诊断
Abstract
Alzheimer's disease (AD) is a common neurodegenerative disorder, and early diagnosis is of great significance for delaying disease progression and enabling timely intervention. Mild cognitive impairment (MCI), which represents an intermediate clinical stage between cognitively normal aging and AD. Structural magnetic resonance imaging (sMRI) provides detailed characterization of anatomical structures and plays an important role in AD-related brain analysis. However, existing sMRI-based brain network methods typically rely on a single graph construction strategy, limiting their ability to jointly capture spatial relationships and morphological similarities between brain regions. To address these issues, this paper proposes an sMRI-based multi-view masked graph neural network model (MVMGNN) for AD diagnosis. A joint node-edge masking mechanism is proposed to simultaneously select radiomics feature dimensions and structural connections, reducing redundancy during graph learning. Furthermore, a patient-level cross-view gated fusion mechanism is proposed to integrate multi-view representations. Experimental results on the ADNI dataset demonstrate that MVMGNN outperforms several competing approaches in AD classification. Interpretability analysis further demonstrates that MVMGNN is able to identify key brain regions associated with AD, providing useful insights into discriminative patterns in sMRI-based brain networks.Our implementation is publicly available at https://github.com/chenzhao2023/MVMGNN_AD
Chinese Translation
阿尔茨海默病（AD）是一种常见的神经退行性疾病，早期诊断对延缓疾病进展和实现及时干预具有重要意义。轻度认知障碍（MCI）代表了认知正常老化与阿尔茨海默病之间的一个中间临床阶段。结构性磁共振成像（sMRI）提供了对解剖结构的详细表征，并在与AD相关的大脑分析中发挥着重要作用。然而，现有的基于sMRI的大脑网络方法通常依赖于单一的图构建策略，限制了它们共同捕捉大脑区域之间空间关系和形态相似性的能力。为了解决这些问题，本文提出了一种基于sMRI的多视角掩蔽图神经网络模型（MVMGNN）用于AD诊断。提出了一种联合节点-边掩蔽机制，以同时选择放射组学特征维度和结构连接，从而减少图学习过程中的冗余。此外，提出了一种患者级别的跨视角门控融合机制，以整合多视角表征。在ADNI数据集上的实验结果表明，MVMGNN在AD分类中优于几种竞争方法。可解释性分析进一步表明，MVMGNN能够识别与AD相关的关键大脑区域，为基于sMRI的大脑网络中的区分模式提供了有用的见解。我们的实现已公开发布在 https://github.com/chenzhao2023/MVMGNN_AD
cs.CV / 14 / 2607.09811
Detangled: A Framework for Creating, Editing, and Inferencing Feature Rich Hair Strands
解缠：一个创建、编辑和推断特征丰富的发丝的框架
Abstract
We present a framework for understanding and generating feature rich hair strands. Drawing upon both scientific and cultural expertise, we define strand texture as the various distinctive patterns (curling, switchbacks, twist, etc.) that are formed by forces internal to a hair strand. We begin by proposing a novel five-dimensional parameter space, intended to be a bijection with naturally occurring hair strand textures. This encoding is both qualitatively accessible, allowing users to readily locate their own hair in the parameter space, and quantitatively precise, allowing the generation of individual strands from texture inputs. Importantly, strand texture should be independent from the overall strand direction. In order to disentangle strand texture from the overall strand direction, we identify centerline geometry and use it to map strands into a canonical space (a strand texture space). We construct centerlines using a novel method that cleanly distills complex hair grooms, separating the strand texture from the overall style (parameterized by style guides). We enable the creation of new strands conforming to our parametric description of texture via a generative artificial intelligence approach supervised by a separate neural network trained to label candidate strands according to our five-parameter description. The ability to create new strands conforming to any desired texture enables groom editing using either texture transfer or user-provided inputs. We demonstrate results on a variety of hair types.
Chinese Translation
我们提出了一个理解和生成特征丰富的发丝的框架。基于科学和文化的专业知识，我们将发丝纹理定义为由发丝内部力量形成的各种独特图案（卷曲、回转、扭曲等）。我们首先提出一个新的五维参数空间，旨在与自然发生的发丝纹理建立一一对应关系。该编码在定性上易于访问，使用户能够轻松地在参数空间中找到自己的发型，同时在定量上精确，允许根据纹理输入生成单独的发丝。重要的是，发丝纹理应独立于整体发丝方向。为了将发丝纹理与整体发丝方向解缠，我们识别中心线几何形状，并利用它将发丝映射到一个规范空间（发丝纹理空间）。我们使用一种新方法构建中心线，干净地提取复杂的发型，分离发丝纹理与整体风格（由风格指南参数化）。我们通过一种生成性人工智能方法，使得新发丝的创建符合我们对纹理的参数描述，该方法由一个单独的神经网络监督，该网络经过训练以根据我们的五参数描述标记候选发丝。能够创建符合任何所需纹理的新发丝，使得可以通过纹理转移或用户提供的输入进行发型编辑。我们在多种发型类型上展示了结果。
cs.CV / 15 / 2607.09822
Memory-Conditioned Tool Calling for Camera-First Visual Agents
基于记忆条件的工具调用用于以相机为主的视觉代理
Abstract
Recognition tells an agent what is in an image; personal memory affects what is worth looking up next. In a camera-first setting the user can send only an image, so the agent must form the lookups. We study whether personal visual memory improves agent-side tool choice and tool arguments, and thereby more user-aligned multi-tool lookups. The design uses a three-layer personal visual memory (profile, short-term focus, observations) that is loaded on each turn to condition an LLM tool-calling loop under camera-first intake, and includes conflict-aware write-back intended to refresh the user model for later captures. On 800 images paired with synthetic memory blocks constructed for controlled ablation, removing the full three-layer memory block reduces tool-query relevance by 0.47 points absolute (4.21 -> 3.74 on a 5-point scale; 11.2% relative) and end-to-end utility by 0.082 absolute (0.842 -> 0.760; 9.7% relative). These results measure memory conditioning of tool policy under image-only intake with fixed synthetic blocks, not multi-session write-back from live user histories.
Chinese Translation
识别告诉代理图像中有什么；个人记忆影响接下来值得查找的内容。在以相机为主的环境中，用户只能发送图像，因此代理必须形成查找。我们研究个人视觉记忆是否改善代理端的工具选择和工具参数，从而实现更符合用户需求的多工具查找。该设计使用三层个人视觉记忆（个人档案、短期关注、观察），在每次交互中加载，以在以相机为主的输入下调节大型语言模型（LLM）工具调用循环，并包括冲突感知的写回，旨在刷新用户模型以便于后续捕获。在与为控制消融而构建的合成记忆块配对的800幅图像上，移除完整的三层记忆块使工具查询的相关性降低了0.47个绝对点（5分制中从4.21降至3.74；相对降低11.2%），端到端效用降低了0.082个绝对点（从0.842降至0.760；相对降低9.7%）。这些结果测量了在仅有图像输入和固定合成块的情况下工具策略的记忆条件，而非来自实时用户历史的多会话写回。
cs.CV / 16 / 2607.09826
Towards Objective Dysgraphia Detection: A Multi-Branch Deep Learning Approach for Online Handwriting Analysis
面向客观性书写障碍检测：一种用于在线手写分析的多分支深度学习方法
Abstract
Dysgraphia is a specific learning disability that is prevalent among school-age children. It affects handwriting coherence, quality, fluency, and legibility, often hindering academic achievement and early learning development. This motor coordination disorder is typically diagnosed through subjective assessments based on clinician observation, which can be timeconsuming and prone to variability. In this paper, we introduce a deep learning-based framework for objective dysgraphia detection using online handwriting data captured via digitizing tablets. The proposed framework relies on two complementary branches: the first pipeline extracts both handcrafted and embedding-based kinematic features directly from raw temporal signals, while the second leverages image-based representations of the temporal signals generated using continuous wavelet transforms (CWT) and Gramian Angular Fields (GAF). The resulting features are then fused to leverage the complementary strengths of both representations. The four representations were evaluated separately and jointly using the publicly available DiaGraMo dataset, showing that the fusion of GAF, MOMENT, and hand-crafted kinematic features outperforms each individual representation, as well as other fusion schemes. These findings highlight the potential of the complementarity of image and signal based representations for more objective dysgraphia detection.
Chinese Translation
书写障碍是一种在学龄儿童中普遍存在的特定学习障碍。它影响书写的一致性、质量、流畅性和可读性，常常妨碍学业成就和早期学习发展。这种运动协调障碍通常通过基于临床观察的主观评估进行诊断，这可能耗时且易受变异性影响。本文介绍了一种基于深度学习的框架，用于利用通过数字化平板捕获的在线手写数据进行客观书写障碍检测。所提出的框架依赖于两个互补的分支：第一个管道直接从原始时间信号中提取手工制作和基于嵌入的运动学特征，而第二个则利用通过连续小波变换（CWT）和格拉米角场（GAF）生成的时间信号的图像表示。然后将得到的特征进行融合，以利用两种表示的互补优势。使用公开可用的DiaGraMo数据集分别和联合评估了这四种表示，结果表明GAF、MOMENT和手工制作运动学特征的融合优于每种单独表示以及其他融合方案。这些发现突显了图像和信号基表示互补性在更客观的书写障碍检测中的潜力。
cs.CV / 17 / 2607.09827
TDSal: Task-Based Top-Down Saliency Prediction Model
TDSal：基于任务的自上而下显著性预测模型
Abstract
Visual saliency aims to predict the regions of an image most likely to attract human visual attention. While most saliency models assume free-viewing conditions, human attention is often shaped by explicit task goals. In this work, we address task-driven saliency prediction by proposing a model that conditions visual attention on natural-language task descriptions. The model produces task-dependent saliency maps that reflect how attention shifts under different viewing intents. Through quantitative and qualitative analysis, we show that incorporating explicit task semantics enables more faithful modeling of goal-directed visual attention.
Chinese Translation
视觉显著性旨在预测图像中最有可能吸引人类视觉注意力的区域。尽管大多数显著性模型假设在自由观看条件下进行，但人类注意力往往受到明确任务目标的影响。在本研究中，我们通过提出一个将视觉注意力与自然语言任务描述相结合的模型，解决了基于任务的显著性预测问题。该模型生成依赖于任务的显著性图，反映了在不同观看意图下注意力的转移。通过定量和定性分析，我们表明，纳入明确的任务语义能够更真实地建模目标导向的视觉注意力。
cs.CV / 18 / 2607.09835
Does YOLO26 Truly Offer Advantages Over Its Predecessors for Edge Deployment? A Benchmark Study in Aquaculture
YOLO26在边缘部署中是否真正优于其前身？水产养殖领域的基准研究
Abstract
The recently introduced YOLO26 architecture incorporates NMS-free end-to-end inference and is optimized for deployment on resource-constrained CPU-based devices, making it well-suited for edge-based aquaculture applications. However, its performance, operational efficiency, and deployment suitability have not been systematically validated in aquaculture-specific scenarios. This study presents a comprehensive benchmark of YOLO26 against three Ultralytics predecessors (YOLOv5u, YOLOv8, and YOLO11) across nano, small, and medium model scales for fish mortality detection, a critical indicator of fish population health and welfare. Twelve model variants were evaluated for detection accuracy, training efficiency across seven dataset sizes, and inference performance on high-performance NVIDIA A100 GPUs and a CPU-only Raspberry Pi 5 edge platform. All models achieved comparable performance on the full dataset, with mAP50 differing by only 1.04 percentage points, indicating that architectural generation has little influence on final detection accuracy when sufficient training data are available. However, clear trade-offs emerged in data efficiency and deployment performance. YOLOv8 achieved 90% mAP50 with only 400 training images, whereas the YOLO26 nano and small variants required 1,000 images to reach comparable accuracy. Conversely, YOLO26n achieved the highest inference speed on the Raspberry Pi 5 (7.51 FPS), while YOLOv5mu outperformed all contemporary medium-scale architectures on CPU-based hardware. These results show that architectural novelty alone is insufficient for model selection and that training data availability, target hardware, and inference requirements should be considered jointly when selecting object detection models for practical edge AI deployment in aquaculture.
Chinese Translation
最近推出的YOLO26架构采用无NMS的端到端推理，并针对资源受限的基于CPU的设备进行了优化，使其非常适合边缘水产养殖应用。然而，其在水产养殖特定场景中的性能、操作效率和部署适用性尚未得到系统验证。本研究对YOLO26与三种Ultralytics前身（YOLOv5u、YOLOv8和YOLO11）在鱼类死亡率检测这一关键指标（反映鱼类种群健康和福利）上的表现进行了全面基准测试，涵盖了纳米、小型和中型模型规模。评估了十二种模型变体在检测准确性、七种数据集规模下的训练效率以及在高性能NVIDIA A100 GPU和仅基于CPU的Raspberry Pi 5边缘平台上的推理性能。所有模型在完整数据集上的表现相当，mAP50的差异仅为1.04个百分点，表明当有足够的训练数据时，架构代际对最终检测准确性的影响较小。然而，在数据效率和部署性能方面出现了明显的权衡。YOLOv8在仅使用400张训练图像的情况下达到了90%的mAP50，而YOLO26的纳米和小型变体则需要1,000张图像才能达到相当的准确性。相反，YOLO26n在Raspberry Pi 5上的推理速度最高（7.51 FPS），而YOLOv5mu在基于CPU的硬件上超越了所有当代中型架构。这些结果表明，仅凭架构的新颖性不足以进行模型选择，选择水产养殖中实际边缘AI部署的目标检测模型时，应综合考虑训练数据的可用性、目标硬件和推理要求。
cs.CV / 19 / 2607.09837
Reliability-Aware Ensemble Classification Under Class Imbalance: A Calibration Study on Liquid-Based Cervical Cytology
考虑可靠性的集成分类在类别不平衡下的研究：基于液体的宫颈细胞学的校准研究
Abstract
Cervical cytology classification models are typically evaluated on curated, class-balanced benchmarks, but real-world liquid-based cytology (LBC) collections are often small and class-imbalanced. This paper presents a class-imbalance-aware and calibration-aware ensemble classification study on the Mendeley LBC dataset, using its native four-class Bethesda taxonomy (NILM, LSIL, HSIL, SCC) rather than a collapsed binary formulation. Three lightweight architectures (Swin-Tiny, TinyViT-5M, DenseNet121) are trained directly on Mendeley LBC using weighted random sampling to counteract class imbalance, and compared against two soft-voting ensembles (Hybrid-2, Hybrid-3). Post-hoc temperature scaling is fit on a held-out calibration subset carved out of the training portion of each cross-validation fold, distinct from both the training data used to fit model weights and the evaluation fold used for final metrics, avoiding the optimistic calibration estimates that result when the same data is used for both purposes. Calibration substantially reduces expected calibration error, Brier score, and negative log-likelihood for every model and ensemble configuration tested, while discrimination metrics (accuracy, macro-F1, macro-AUROC) remain essentially unchanged. Ensemble size shows no consistent additional reliability benefit over the best individual model once all configurations are properly calibrated. Confusion matrices show that all classification errors, across every configuration, are confined to the boundary between high-grade lesions (HSIL) and carcinoma (SCC); no errors involve the negative (NILM) or low-grade (LSIL) categories. These results suggest that, for this dataset, calibration is the dominant lever for reliability, not ensemble size, though this conclusion should be read in light of the dataset's modest size.
Chinese Translation
宫颈细胞学分类模型通常在经过策划的类别平衡基准上进行评估，但现实世界中的液体基于细胞学（LBC）数据集往往较小且类别不平衡。本文在Mendeley LBC数据集上进行了一项考虑类别不平衡和校准的集成分类研究，使用其原生的四类Bethesda分类法（NILM, LSIL, HSIL, SCC），而非简化的二元形式。三种轻量级架构（Swin-Tiny, TinyViT-5M, DenseNet121）直接在Mendeley LBC上进行训练，采用加权随机抽样来抵消类别不平衡，并与两个软投票集成（Hybrid-2, Hybrid-3）进行比较。后验温度缩放在从每个交叉验证折叠的训练部分中划分出的保留校准子集上进行拟合，避免了使用相同数据进行模型权重拟合和最终指标评估所导致的乐观校准估计。校准显著降低了每个模型和集成配置的预期校准误差、Brier分数和负对数似然，同时区分指标（准确率、宏观F1、宏观AUROC）基本保持不变。一旦所有配置得到适当校准，集成规模对最佳单一模型没有一致的额外可靠性收益。混淆矩阵显示，所有分类错误在每个配置中都局限于高等级病变（HSIL）和癌症（SCC）之间的边界；没有错误涉及阴性（NILM）或低等级（LSIL）类别。这些结果表明，对于该数据集，校准是影响可靠性的主要因素，而非集成规模，尽管这一结论应结合数据集的适度规模进行解读。
cs.CV / 20 / 2607.09876
Prompting-MammAlps: Fine-Grained Text-to-Video Retrieval for Camera-Trap Data
Prompting-MammAlps：针对相机捕捉数据的细粒度文本到视频检索
Abstract
Automatically retrieving videos from large camera-trap datasets remains challenging. Text-to-Video retrieval (TVR) methods based on large video-language models (VLMs) have potential to retrieve events of interest by describing them with simple text queries. However, current methods often lack spatiotemporal understanding and do not generalize well to ecological data. In this work, we introduce Prompting-MammAlps, the first camera-trap TVR benchmark, and propose a fine-grained and interpretable TVR method. Specifically, we trained a vision transformer to perform spatiotemporal action localization, and convert its output to structured text, describing each video. Independently, ethology-inspired queries are processed by a Large-Language Model (LLM) based coding agent to parse the structured text per video and retrieve videos accordingly. We harnessed the LLM to use functions from a custom parsing library to minimize the risk of LLM hallucinations and to improve method interpretability. This retrieval approach applied on the Prompting-MammAlps benchmark achieved a set-based F1-score of 34\% on a test set of 135 ecologically-relevant queries and 775 candidate videos. In comparison the best zero-shot VLM achieved a F1-score of 18\%, while also lacking interpretability. Project page: https://cnai.epfl.ch/prompting-mammalps
Chinese Translation
从大型相机捕捉数据集中自动检索视频仍然具有挑战性。基于大型视频-语言模型（VLMs）的文本到视频检索（TVR）方法有潜力通过简单的文本查询描述感兴趣的事件。然而，当前的方法往往缺乏时空理解，并且在生态数据上泛化能力不足。在本研究中，我们引入了Prompting-MammAlps，这是第一个相机捕捉TVR基准，并提出了一种细粒度且可解释的TVR方法。具体而言，我们训练了一个视觉变换器以执行时空动作定位，并将其输出转换为结构化文本，描述每个视频。独立地，受动物行为学启发的查询由基于大型语言模型（LLM）的编码代理处理，以解析每个视频的结构化文本并相应地检索视频。我们利用LLM使用自定义解析库中的函数，以最小化LLM幻觉的风险并提高方法的可解释性。该检索方法在Prompting-MammAlps基准上实现了在135个生态相关查询和775个候选视频的测试集上的基于集合的F1分数为34%。相比之下，最佳的零-shot VLM的F1分数为18%，同时也缺乏可解释性。项目页面：https://cnai.epfl.ch/prompting-mammalps
cs.CV / 21 / 2607.09883
A Dual-Stream Challenge-Response Protocol for Ocular Liveness Verification
一种双流挑战-响应协议用于眼部活体验证
Abstract
Ocular biometric systems face sophisticated presentation attacks, including high-resolution video replays and real-time generative deepfakes, which easily bypass static liveness checks. Current Presentation Attack Detection (PAD) frameworks typically rely on isolated physiological metrics, such as gaze tracking or the Pupillary Light Reflex (PLR), which can be spoofed independently. This paper proposes a Spatio-Luminance Sensor Fusion protocol, which introduces a dual-stream challenge-response framework for ocular liveness verification by uniting these metrics into a simultaneous authentication challenge. By generating a randomized, time-varying visual stimulus that fluctuates in both spatial trajectory and luminance intensity, we construct a mathematically coupled state-space likelihood model, termed the Synchronization Matrix, to evaluate the continuous cross-correlation between the expected biological latencies of smooth pursuit tracking and pupillary constriction. Using Monte Carlo simulation grounded in literature-derived latency distributions, we demonstrate theoretical separability between genuine and simulated attack conditions, and show that a multi-round challenge design improves the detection of generative deepfakes when a non-zero rendering-latency gap exists. This work provides a simulation-supported theoretical framework for next-generation dynamic spoofing defense in ocular and iris biometrics; human-subject validation is identified as necessary future work before deployment claims can be made.
Chinese Translation
眼部生物识别系统面临复杂的呈现攻击，包括高分辨率视频重播和实时生成的深度伪造，这些攻击很容易绕过静态活体检测。当前的呈现攻击检测（PAD）框架通常依赖于孤立的生理指标，例如注视追踪或瞳孔光反射（PLR），这些指标可以独立被欺骗。本文提出了一种时空-光度传感器融合协议，通过将这些指标结合为一个同时的身份验证挑战，提出了一种双流挑战-响应框架用于眼部活体验证。通过生成一个随机的、时间变化的视觉刺激，该刺激在空间轨迹和光度强度上波动，我们构建了一个数学耦合的状态空间似然模型，称为同步矩阵，用于评估平滑追踪和瞳孔收缩的预期生物延迟之间的连续互相关。利用基于文献派生的延迟分布的蒙特卡洛模拟，我们展示了真实和模拟攻击条件之间的理论可分性，并表明在存在非零渲染延迟差距时，多轮挑战设计能够提高对生成深度伪造的检测。该研究为下一代动态欺骗防御在眼部和虹膜生物识别中的应用提供了一个基于模拟的理论框架；在进行部署声明之前，识别出需要进行人类受试者验证的未来工作。
cs.CV / 22 / 2607.09884
ShapKO: Shapley-Adaptive Modality Knockout for Robust Multimodal Learning
ShapKO：基于Shapley值的自适应模态剔除用于稳健的多模态学习
Abstract
Multimodal medical models often degrade when inputs are missing, a common scenario in real-world clinical workflows. Separately, even when all modalities are present, modality dominance is observed during training, where optimization over-relies on a highly predictive modality and undertrains complementary sources, resulting in poor robustness under partial availability. While training-time modality knockout improves missing-modality robustness, existing approaches use static masking rates that cannot adapt to evolving modality utility during training. We introduce ShapKO (Shapley-Adaptive Modality Knockout), a dynamic training strategy that learns modality-specific knockout probabilities based on validation utility. ShapKO periodically evaluates performance across modality subsets, estimates modality importance via Shapley values, and updates masking probabilities to suppress dominant modalities more frequently. This adaptive process promotes complementary representations, while requiring no architectural modifications. We evaluate ShapKO on three datasets covering multitask clinical classification, survival prediction, and cancer detection. ShapKO consistently improves performance under modality absence and yields interpretable trajectories of learned masking behavior. Code is available at: https://github.com/sumona00/ShapKO
Chinese Translation
多模态医学模型在输入缺失时常常表现不佳，这在现实临床工作流程中是一个常见场景。即使在所有模态均存在的情况下，训练过程中也会观察到模态主导现象，即优化过度依赖于高度预测的模态，而对互补来源的训练不足，导致在部分可用性下的稳健性差。虽然训练时的模态剔除可以改善缺失模态的稳健性，但现有方法使用的静态掩蔽率无法适应训练过程中模态效用的变化。我们提出了ShapKO（基于Shapley值的自适应模态剔除），这是一种动态训练策略，根据验证效用学习模态特定的剔除概率。ShapKO定期评估模态子集的性能，通过Shapley值估计模态重要性，并更新掩蔽概率，以更频繁地抑制主导模态。这个自适应过程促进了互补表示，同时无需进行架构修改。我们在三个数据集上评估了ShapKO，涵盖多任务临床分类、生存预测和癌症检测。ShapKO在模态缺失情况下持续提高性能，并产生可解释的学习掩蔽行为轨迹。代码可在以下链接获取：https://github.com/sumona00/ShapKO
cs.CV / 23 / 2607.09888
Bridging the Catalog-to-Real Gap: Scalable Product Recognition via Multi-Stage Contrastive Learning
弥合目录与现实之间的差距：通过多阶段对比学习实现可扩展的产品识别
Abstract
Automated product recognition is a cornerstone of modern retail intelligence; however, accurately matching real-world, in-store images against extensive corporate catalogs remains a major scalability bottleneck for large-scale applications. In this work, we address this challenge by reformulating the task as an embedding-based cross-domain retrieval problem rather than a standard closed-set classification task. Specifically, we define the objective as retrieving the most corresponding catalog reference image for a given real-world product query crop from an expansive inventory. To bridge the severe domain gap between pristine studio packshots and noisy in-store queries, we introduce a novel catalog-to-real multi-stage contrastive learning paradigm (Cat2Real). This framework fine-tunes a vision backbone by systematically exploiting both item-level and image-level similarities to drive targeted hard negative mining. Extensive empirical evaluations demonstrate that our paradigm scales seamlessly to unseen products and categories, yielding outstanding zero-shot generalization performance even in the complete absence of real-world training images for novel inventory.
Chinese Translation
自动化产品识别是现代零售智能的基石；然而，将真实世界中的店内图像与庞大的企业目录准确匹配，仍然是大规模应用中的主要可扩展性瓶颈。在本研究中，我们通过将任务重新定义为基于嵌入的跨域检索问题，而非标准的封闭集分类任务，来应对这一挑战。具体而言，我们将目标定义为从庞大的库存中检索与给定真实产品查询裁剪图像最相对应的目录参考图像。为了弥合原始摄影图像与嘈杂的店内查询之间的严重领域差距，我们引入了一种新颖的目录到现实的多阶段对比学习范式（Cat2Real）。该框架通过系统性地利用项目级和图像级的相似性来驱动有针对性的困难负样本挖掘，从而微调视觉主干网络。大量实证评估表明，我们的范式能够无缝扩展到未见过的产品和类别，即使在完全没有新库存的真实世界训练图像的情况下，也能实现出色的零样本泛化性能。
cs.CV / 24 / 2607.09890
Do Transformer Temporal Heads and Post-Pooling Motion Gates Help CorrNet-based CSLR? An Empirical Study
变换器时间头和后池化运动门是否有助于基于CorrNet的连续手语识别？一项实证研究
Abstract
CorrNet is a strong baseline for continuous sign language recognition (CSLR) because it models inter-frame correlations inside the visual encoding stage. In this paper, we study two natural extensions of a reproduced CorrNet system: replacing the BiLSTM temporal head with a Transformer encoder, and injecting motion cues after temporal pooling. We find that the Transformer head does not outperform the BiLSTM baseline, even with a training strategy adjusted for the Transformer, and the two heads have almost the same computational and runtime cost. For the second extension, we design a lightweight module called MotionGate. In our experiments, MotionGate consistently collapses to an identity-like mapping: the gate loses motion selectivity, and the injected residual becomes a weak, non-selective perturbation of the pooled features. These results suggest that explicit motion injection after CorrNet's correlation-based encoding is largely redundant, and that natural-looking architectural extensions in CSLR should be tested carefully instead of being assumed to help.
Chinese Translation
CorrNet是连续手语识别（CSLR）的一个强基线，因为它在视觉编码阶段建模了帧间相关性。本文研究了对再现的CorrNet系统的两个自然扩展：用变换器编码器替换BiLSTM时间头，以及在时间池化后注入运动线索。我们发现，即使在为变换器调整训练策略的情况下，变换器头的表现也没有超过BiLSTM基线，并且这两种头部的计算和运行时间成本几乎相同。对于第二个扩展，我们设计了一个轻量级模块，称为MotionGate。在我们的实验中，MotionGate始终收敛为一种类似恒等映射的形式：门失去了运动选择性，注入的残差成为池化特征的弱且非选择性扰动。这些结果表明，在CorrNet的基于相关性的编码之后，显式的运动注入在很大程度上是多余的，并且在CSLR中，自然的架构扩展应谨慎测试，而不是假设它们会有所帮助。
cs.CV / 25 / 2607.09930
Banshee: Target Switch Attacks on Gimbal-Stabilized Visual Tracking Systems via Acoustic Injection
Banshee：通过声学注入对云台稳定视觉跟踪系统的目标切换攻击
Abstract
Gimbal-stabilized visual tracking is critical for modern autonomous systems such as Unmanned Aerial Vehicles (UAVs). While prior work shows acoustic signals can disturb gimbal internals, the impact of such attacks on real-world applications like UAV tracking and following remains underexplored. Existing demonstrations largely overlook practical challenges for real-world attacks, such as object-motion uncertainty and runtime latency. To bridge this gap, we present Banshee, the first physically realizable attack that induces target switching in UAV visual tracking systems by exploiting acoustic vulnerabilities in gimbal-camera systems. Banshee generates carefully crafted acoustic waveforms that induce optimized adversarial gimbal oscillations, causing directionally biased camera-view drifts that break inter-frame target associations. Consequently, the onboard tracker is driven to switch from the original target to an attacker-selected object with high probability, with occasional target loss. Banshee achieves a 93.6% success rate in simulation across two commercial gimbal systems and five trackers. Real-world benchtop and in-flight black-box attacks against a commercial drone across varied scenarios show an overall 95.5% attack success rate. Our results reveal a practical cross-domain vulnerability between acoustics and vision, highlighting the need for robust designs of gimbal systems and applications. Our code is available at: https://github.com/U1ltra/Banshee.
Chinese Translation
云台稳定视觉跟踪对于现代自主系统（如无人机）至关重要。虽然先前的研究表明声学信号可以干扰云台内部，但此类攻击对无人机跟踪和追踪等实际应用的影响仍未得到充分探讨。现有的演示大多忽视了现实攻击中的实际挑战，例如物体运动的不确定性和运行时延迟。为了解决这一问题，我们提出了Banshee，这是首个可物理实现的攻击，通过利用云台摄像系统中的声学漏洞，诱导无人机视觉跟踪系统中的目标切换。Banshee生成精心设计的声学波形，诱发优化的对抗性云台振荡，导致方向性偏差的摄像头视角漂移，从而破坏帧间目标关联。因此，机载跟踪器在高概率下被迫从原始目标切换到攻击者选择的对象，偶尔会出现目标丢失。Banshee在两个商业云台系统和五个跟踪器的仿真中实现了93.6%的成功率。在不同场景下对商业无人机进行的实际台式和飞行黑箱攻击显示出整体95.5%的攻击成功率。我们的结果揭示了声学与视觉之间的实际跨领域脆弱性，强调了对云台系统和应用进行稳健设计的必要性。我们的代码可在以下链接获取：https://github.com/U1ltra/Banshee。
cs.CV / 26 / 2607.09985
UniPose9D: Universal Category-Agnostic Object Pose Estimation
UniPose9D：通用类别无关物体姿态估计
Abstract
Object pose estimation is a fundamental problem in 3D vision. Although recent state-of-the-art approaches achieve strong performance, they often overfit to existing benchmarks and exhibit limited generalization to novel categories and unseen scenes. We propose UniPose9D, a category-agnostic foundation model for 9D object pose estimation: given an instance mask/ROI and either an RGB-D observation or an RGB image with predicted depth, the model estimates rotation, translation, and metric size without category labels, CAD models, mean-shape priors, or reference views. Specifically, UniPose9D samples point pairs from the observed object geometry and uses DINOv2 and PointNet features to predict NOCS coordinates for each pair. To improve accuracy, we introduce a point-pair-based RANSAC N-hop Kabsch--Umeyama algorithm with an adaptive threshold. We further employ flow matching to address symmetric ambiguities and construct a large-scale training set by curating and aligning pose annotations from existing public datasets. Experiments across six datasets show that a single unified model can match or surpass specialist methods while generalizing to unseen objects and in-the-wild scenarios. Our code and model are available on https://github.com/qq456cvb/UniPose9D.
Chinese Translation
物体姿态估计是三维视觉中的一个基本问题。尽管最近的最先进方法取得了强劲的性能，但它们往往对现有基准过拟合，并且在新类别和未见场景中的泛化能力有限。我们提出了UniPose9D，这是一种用于9D物体姿态估计的类别无关基础模型：给定一个实例掩码/感兴趣区域（ROI）以及RGB-D观测或带有预测深度的RGB图像，该模型在没有类别标签、CAD模型、均值形状先验或参考视图的情况下，估计旋转、平移和度量大小。具体而言，UniPose9D从观察到的物体几何中采样点对，并使用DINOv2和PointNet特征来预测每对的NOCS坐标。为了提高准确性，我们引入了一种基于点对的RANSAC N-hop Kabsch--Umeyama算法，并设置了自适应阈值。我们进一步采用流匹配来解决对称歧义，并通过整理和对齐现有公共数据集中的姿态注释构建了一个大规模训练集。在六个数据集上的实验表明，单一统一模型能够匹配或超越专业方法，同时在未见物体和野外场景中具有良好的泛化能力。我们的代码和模型可在https://github.com/qq456cvb/UniPose9D获取。
cs.CV / 27 / 2607.10004
Model Guides You How to Draw: Adaptive Visual Gating for Unified Multimodal Reasoning
模型指导你如何绘制：统一多模态推理的自适应视觉门控
Abstract
Unified multimodal models (UMMs) with interleaved reasoning, which generate both textual and visual steps as part of intermediate reasoning traces, have demonstrated great potential for visual mathematical reasoning tasks. However, we identify a key insight in this paradigm: generating intermediate visual reasoning steps is not always beneficial and can even be harmful, as self-generated visual steps may introduce erroneous visual evidence that misleads subsequent reasoning. Moreover, frequently triggering visual steps during reasoning incurs substantial computational and memory overhead, degrading inference efficiency. To address these accuracy and efficiency challenges, we observe that the model's internal signals can indicate whether a visual step will benefit reasoning before the entire visual generation is completed. Specifically, this work identifies two internal signals: 1) Generation Intent, which reflects whether the model has a concrete textual plan for what to draw, and 2) Visual Fidelity, which measures whether the visual generation remains grounded in the original input image. Leveraging these internal signals, we propose AdaViG, a training-free adaptive visual gating method for unified multimodal reasoning. AdaViG dynamically evaluates each triggered visual step at an early visual generation stage and aborts it when both signals are weak, thereby preventing misleading visual evidence from entering the reasoning trace while avoiding unnecessary computation. Comprehensive experiments demonstrate that AdaViG improves accuracy by up to 5.7% while reducing visual generation FLOPs by 25.0%-91.0% and wall-clock latency by 15.4%-45.6%.
Chinese Translation
具有交错推理的统一多模态模型（UMMs）在视觉数学推理任务中展现了巨大的潜力，这些模型生成文本和视觉步骤作为中间推理痕迹的一部分。然而，我们在这一范式中识别出一个关键见解：生成中间视觉推理步骤并不总是有益的，甚至可能是有害的，因为自生成的视觉步骤可能引入错误的视觉证据，从而误导后续推理。此外，在推理过程中频繁触发视觉步骤会带来显著的计算和内存开销，降低推理效率。为了解决这些准确性和效率挑战，我们观察到模型的内部信号可以在整个视觉生成完成之前指示某个视觉步骤是否会对推理有益。具体而言，本研究识别了两个内部信号：1）生成意图（Generation Intent），反映模型是否有明确的文本计划来绘制内容；2）视觉保真度（Visual Fidelity），衡量视觉生成是否仍然基于原始输入图像。利用这些内部信号，我们提出了AdaViG，一种无训练的自适应视觉门控方法，用于统一多模态推理。AdaViG在早期视觉生成阶段动态评估每个触发的视觉步骤，并在两个信号均较弱时中止该步骤，从而防止误导性的视觉证据进入推理痕迹，同时避免不必要的计算。全面的实验表明，AdaViG在提高准确性方面最多可提升5.7%，同时将视觉生成的FLOPs减少25.0%-91.0%，墙钟延迟减少15.4%-45.6%。
cs.CV / 28 / 2607.10071
FlashBEV: Fast and Memory-Efficient Exact BEV Transformation with IO-Awareness
FlashBEV：一种快速且内存高效的精确鸟瞰图（BEV）转换方法，具有IO感知能力
Abstract
Bird's-eye-view (BEV) perception is a core component of camera-based 3D understanding in autonomous driving, where view transformation (VT) maps multi-camera image features into a unified BEV representation. Sampling-based view transformation (Sampling-VT) is attractive because it supports dense and continuous BEV aggregation for high-resolution and long-range perception. Its deployment bottleneck, however, is systems-level: standard tensorized implementations of Sampling-VT -- which we refer to as Tensorized Sampling-VT -- explicitly materialize large height-dependent intermediate tensors, causing memory and latency costs that scale poorly with vertical resolution and the number of cameras. We revisit Tensorized Sampling-VT from an operator-execution perspective and show that it follows a gather-reduction pattern: each BEV query independently accumulates contributions across cameras and height bins, enabling thread-local accumulation with on-the-fly recomputation that eliminates the need to materialize height- and camera-dependent intermediates. Based on this insight, we propose FlashBEV, a fully fused and IO-aware execution strategy mathematically equivalent to Tensorized Sampling-VT (same operator output) while substantially reducing global memory traffic and kernel-launch overhead. Experiments show that FlashBEV achieves more than an order of magnitude lower peak GPU memory and significant inference-latency speedups, with memory effectively independent of the number of height bins, reducing the operator's peak memory to O(BCXY) (output only). This unlocks higher BEV range/resolution and vertical discretization within fixed deployment budgets on memory-constrained devices. Our contribution is an execution redesign -- same math, different execution -- that removes a key scalability barrier for deployment-ready Sampling-VT. Code available at https://github.com/yokosyun/FlashBEV
Chinese Translation
鸟瞰图（BEV）感知是基于摄像头的自动驾驶3D理解的核心组成部分，其中视图转换（VT）将多摄像头图像特征映射到统一的BEV表示。基于采样的视图转换（Sampling-VT）因其支持高分辨率和长距离感知的密集和连续BEV聚合而备受关注。然而，其部署瓶颈在于系统层面：Sampling-VT的标准张量化实现（我们称之为张量化采样视图转换，Tensorized Sampling-VT）显式地生成大型高度依赖的中间张量，导致内存和延迟成本随着垂直分辨率和摄像头数量的增加而急剧上升。我们从操作执行的角度重新审视张量化采样视图转换，表明其遵循聚合-缩减模式：每个BEV查询独立地在摄像头和高度区间中累积贡献，使得能够通过即时重计算实现线程局部累积，从而消除生成高度和摄像头依赖的中间结果的需求。基于这一见解，我们提出了FlashBEV，一种完全融合且具IO感知的执行策略，其数学上等同于张量化采样视图转换（相同的操作输出），同时显著减少全局内存流量和内核启动开销。实验表明，FlashBEV在峰值GPU内存上实现了超过一个数量级的降低，并显著加快了推理延迟，内存有效地独立于高度区间的数量，将操作的峰值内存降低到O(BCXY)（仅输出）。这在内存受限的设备上解锁了更高的BEV范围/分辨率和垂直离散化，且在固定的部署预算内。我们的贡献是一次执行重设计——相同的数学，不同的执行——消除了准备部署的Sampling-VT的一个关键可扩展性障碍。代码可在 https://github.com/yokosyun/FlashBEV 获取。
cs.CV / 29 / 2607.10082
Label-Free Target-Domain Adaptation for Unconstrained Event-Image Feature Matching via Dual-Stage Distillation
无标签目标领域适应的无约束事件-图像特征匹配通过双阶段蒸馏
Abstract
Building pixel-level correspondence between event and image data is a fundamental task for multi-sensor systems. However, existing cross-modal matching methods are largely restricted by their reliance on either matching labels or strictly aligned hardware, which limits them to unlabeled and unconstrained real-world scenarios where neither matching ground truth nor prior sensor relationships are available. To address this, we propose a novel two-stage training paradigm. First, we leverage large-scale data to perform label-agnostic distillation pretraining, upgrading optimization objectives with distribution-based and contrastive losses to learn highly generalizable representations. Second, to tackle unlabeled and unconstrained downstream data, we introduce an epipolar-guided self-distillation framework. By utilizing consistency verification to isolate robust matches and incorporating geometric confidence derived from an external epipolar prior, our model can effectively self-evolve directly on target domains without any supervision. Furthermore, we introduce a rigorous cross-modal evaluation benchmark based on TUM-VIE, featuring physically separated cameras with distinct intrinsic parameters and resolutions. Extensive experiments demonstrate that our proposed method achieves state-of-the-art performance on both MVSEC and TUM-VIE pose estimation tasks. The source code and benchmark will be made publicly available at https://github.com/ZhonghuaYi/nexus2-official.
Chinese Translation
在多传感器系统中，建立事件数据与图像数据之间的像素级对应关系是一项基础任务。然而，现有的跨模态匹配方法在很大程度上受到依赖于匹配标签或严格对齐硬件的限制，这使得它们无法应用于没有匹配真值或先前传感器关系的无标签和无约束的真实场景。为了解决这个问题，我们提出了一种新颖的两阶段训练范式。首先，我们利用大规模数据进行无标签蒸馏预训练，通过基于分布和对比损失的优化目标来学习高度可泛化的表示。其次，为了应对无标签和无约束的下游数据，我们引入了一种极线引导的自蒸馏框架。通过利用一致性验证来隔离稳健的匹配，并结合来自外部极线先验的几何置信度，我们的模型能够在目标领域上有效地自我演化，而无需任何监督。此外，我们基于TUM-VIE引入了一个严格的跨模态评估基准，特征是物理上分离的摄像头，具有不同的内在参数和分辨率。大量实验表明，我们提出的方法在MVSEC和TUM-VIE姿态估计任务上达到了最先进的性能。源代码和基准将公开发布在 https://github.com/ZhonghuaYi/nexus2-official。
cs.CV / 30 / 2607.10087
CVKD-UDA: Cross-View Knowledge Distillation for 3D Unsupervised Domain Adaptive Segmentation
CVKD-UDA：用于3D无监督领域自适应分割的跨视图知识蒸馏
Abstract
3D unsupervised domain adaptive (UDA) segmentation mitigates the high cost of manual annotations of the new domain data. Self-training has emerged as the dominant approach in this area, where its success heavily depends on a well-initialized warm-up model to generate reliable pseudo labels. However, existing methods often depend on source supervision or output-level adversarial alignment to obtain the warm-up model, which suffer from limited generalization and training instability due to the large domain gap between domains. Constructing domain-similar representations is an effective way to bridge this gap. In this work, we propose CVKD-UDA, which revisits voxel size as a core design factor to construct domain-similar representations and leverages cross-view complementary cues to balance transferability and discriminability of the warm-up model. First, we generate two complementary views by varying voxel sizes and introduce a cross-view knowledge distillation (CVKD) to enhance generalization and target perception of the model. Second, to balance transferability and discriminability, we design a lightweight Decouple-Adapter and an auxiliary imitation classifier to decouple cross-view knowledge transfer. Extensive experiments on two benchmarks demonstrate that CVKD-UDA effectively improves the performance of self-training methods and provides a new perspective for 3D UDA segmentation. Our code will be available at GitHub.
Chinese Translation
3D无监督领域自适应（UDA）分割减轻了新领域数据手动标注的高成本。自我训练已成为该领域的主流方法，其成功在很大程度上依赖于一个良好初始化的热身模型来生成可靠的伪标签。然而，现有方法通常依赖于源监督或输出级对抗对齐来获得热身模型，这由于领域之间的巨大差距而面临有限的泛化能力和训练不稳定性。构建领域相似的表示是一种有效的弥合这一差距的方法。在本研究中，我们提出了CVKD-UDA，该方法重新审视体素大小作为构建领域相似表示的核心设计因素，并利用跨视图互补线索来平衡热身模型的可迁移性和可区分性。首先，我们通过改变体素大小生成两个互补视图，并引入跨视图知识蒸馏（CVKD）来增强模型的泛化能力和目标感知。其次，为了平衡可迁移性和可区分性，我们设计了一个轻量级的解耦适配器和一个辅助模仿分类器，以解耦跨视图知识转移。在两个基准上的大量实验表明，CVKD-UDA有效提高了自我训练方法的性能，并为3D UDA分割提供了新的视角。我们的代码将在GitHub上发布。
cs.CV / 31 / 2607.10093
EMBRACE: A Multi-task Framework for Comprehensive Quality Assessment in Cleavage-stage Embryo
EMBRACE：一个用于全面质量评估的多任务框架在卵裂期胚胎中的应用
Abstract
Cleavage-stage embryo assessment in in vitro fertilization requires the integrated interpretation of cytoplasmic fragmentation, developmental stage, and blastomere symmetry. However, conventional visual assessment is affected by observer variability, particularly when fragmented regions are small, irregular, or low contrast. This study presents EMBRACE, a multi-task deep learning framework for jointly performing cytoplasmic-fragmentation segmentation, t2/t4 developmental-stage classification, and blastomere-symmetry grading from static cleavage-stage embryo microscopy images. EMBRACE combines a shared ResNet-50 backbone, a concatenation-based multi-scale feature-fusion (C-MSFF) module, a U-Net-style segmentation decoder, and two task-specific classification heads. After predefined inclusion and exclusion criteria, 9,137 annotated embryo images were divided into 7,309 training, 914 validation, and 914 held-out test images. On the held-out test set, EMBRACE achieved a Dice coefficient of 0.781 and an intersection over union of 0.677 for fragmentation segmentation. Developmental-stage classification achieved an accuracy of 0.995, macro-F1 of 0.994, and AUC of 1.000. Blastomere-symmetry grading achieved a balanced accuracy of 0.901, macro-F1 of 0.907, and quadratic weighted kappa of 0.859. These findings support the feasibility of combining spatially inspectable fragmentation localization with embryo-level morphology assessment in a single framework. External and prospective validation is required before clinical deployment.
Chinese Translation
在体外受精中，卵裂期胚胎的评估需要对细胞质碎片、发育阶段和卵裂球对称性进行综合解读。然而，传统的视觉评估受到观察者变异性的影响，尤其是在碎片区域较小、不规则或对比度较低的情况下。本研究提出了EMBRACE，一个多任务深度学习框架，旨在从静态卵裂期胚胎显微图像中共同执行细胞质碎片分割、t2/t4发育阶段分类和卵裂球对称性评分。EMBRACE结合了共享的ResNet-50主干、基于连接的多尺度特征融合（C-MSFF）模块、U-Net风格的分割解码器和两个特定任务的分类头。在预定义的纳入和排除标准后，9,137张标注的胚胎图像被分为7,309张训练图像、914张验证图像和914张保留测试图像。在保留的测试集上，EMBRACE在碎片分割中达到了0.781的Dice系数和0.677的交并比。发育阶段分类的准确率为0.995，宏观F1为0.994，AUC为1.000。卵裂球对称性评分的平衡准确率为0.901，宏观F1为0.907，二次加权kappa为0.859。这些发现支持在单一框架中将空间可检验的碎片定位与胚胎级形态评估相结合的可行性。在临床应用之前，需要进行外部和前瞻性验证。
cs.CV / 32 / 2607.10094
LFD: Enabling Real-World Lensless Face Recognition with a Large-Scale Dataset
LFD：通过大规模数据集实现真实世界的无镜头人脸识别
Abstract
Face recognition is a ubiquitously used computer vision task that has a wide range of applications ranging from everyday smartphone biometrics to high-stakes security systems. Most face recognition systems rely on traditional cameras, which often suffer from limitations such as bulky form factors, high costs, and limited privacy protection. To address these limitations, lensless cameras have emerged as an alternative. Lensless cameras use thin optical encoders, enabling smaller size, lower cost, and greater design flexibility. These cameras are typically paired with reconstruction algorithms that convert raw captures into recognizable images. However, reconstructed images often contain artifacts, and the reconstruction methods struggle to generalize well to real-world conditions. Furthermore, existing face datasets do not account for the artifacts present in lensless images. To address this issue, we introduce the Lensless Face Dataset (LFD). LFD comprises 21,080 lensless raw measurements, reconstructions, and standard images of faces captured under diverse lighting, angle, and distance. Our key contributions are: (1) Real-world lensless face data: LFD focuses on capturing a diverse face dataset with varying levels of artifacts introduced under different environments; (2) In-the-wild captures: 4,976 images are captured in outdoor settings with varying intensities of natural light and different background patterns; (3) Multiple lensless devices: LFD includes face images collected from three different types of lensless cameras, each with a unique optical encoder. We use this hardware diversity to demonstrate generalization across different lensless cameras. Through comprehensive evaluations and analysis, we show that LFD effectively captures shared features and artifacts across different lensless imaging devices, making it a valuable dataset for advancing lensless face recognition.
Chinese Translation
人脸识别是一项广泛应用的计算机视觉任务，涉及从日常智能手机生物识别到高风险安全系统的多种应用。大多数人脸识别系统依赖于传统相机，这些相机通常存在体积庞大、成本高昂和隐私保护有限等缺陷。为了解决这些问题，无镜头相机作为一种替代方案应运而生。无镜头相机使用薄型光学编码器，能够实现更小的体积、更低的成本和更大的设计灵活性。这些相机通常与重建算法配合使用，将原始捕获的数据转换为可识别的图像。然而，重建图像往往包含伪影，且重建方法在真实世界条件下的泛化能力较差。此外，现有的人脸数据集并未考虑无镜头图像中存在的伪影。为了解决这一问题，我们引入了无镜头人脸数据集（Lensless Face Dataset，LFD）。LFD包含21,080个无镜头原始测量数据、重建图像和标准人脸图像，这些图像是在不同的光照、角度和距离下捕获的。我们的主要贡献包括：（1）真实世界的无镜头人脸数据：LFD专注于捕获在不同环境下引入不同伪影水平的多样化人脸数据集；（2）野外捕获：4,976张图像是在户外环境中捕获的，具有不同强度的自然光和不同背景模式；（3）多种无镜头设备：LFD包括从三种不同类型的无镜头相机收集的人脸图像，每种相机都有独特的光学编码器。我们利用这种硬件多样性展示了不同无镜头相机之间的泛化能力。通过全面的评估和分析，我们表明LFD有效捕获了不同无镜头成像设备之间的共享特征和伪影，使其成为推动无镜头人脸识别的重要数据集。
cs.CV / 33 / 2607.10098
DynaFilter: Cloud-driven Dynamic Filtering for Satellite Edge Intelligence
DynaFilter：基于云的卫星边缘智能动态过滤
Abstract
Modern satellite edge systems, including those performing remote sensing tasks such object detection and tracking, are characterized by severely limited bandwidth and intermittent connections, making continuous data transmission to the cloud impractical. Existing edge-cloud systems, however, either require heavy pre-processing before analysis, for instance, full decompression of imagery data, or transmit all compressed data regardless of relevance. To address these challenges, we design DynaFilter, a dynamic filtering technique that enables satellite edge devices to perform selective region-of-interest (RoI) inference directly in the compressed-domain, without full decompression. Our key insight is that low-level compression syntax, specifically DC coefficients/AC energy in JPEG images and motion vectors in video streams, exhibits strong correlations with high-level semantic queries. By establishing a precise mapping between cloud query semantics and multimodal compressed-domain features, DynaFilter enables the edge to identify and transmit only relevant data associated to RoIs. Extensive evaluations show that DynaFilter reduces the total volume of pixel data for decoding and subsequent inference by 1.6x-7.1x for images, and achieves 92.0% bandwidth savings for video streams compared to state-of-the-art baselines. Furthermore, it decreases energy consumption by 43.1-88.6% on target devices and achieves a 1.6x-3.0x speedup in inference latency.
Chinese Translation
现代卫星边缘系统，包括执行遥感任务（如目标检测和跟踪）的系统，具有带宽严重受限和连接间歇性的特点，使得持续向云端传输数据变得不切实际。然而，现有的边缘-云系统要么在分析之前需要进行繁重的预处理，例如对图像数据进行完全解压，要么无论相关性如何都传输所有压缩数据。为了解决这些挑战，我们设计了DynaFilter，一种动态过滤技术，使卫星边缘设备能够在压缩域中直接执行选择性感兴趣区域（RoI）推断，而无需完全解压。我们的关键见解是，低级压缩语法，特别是JPEG图像中的直流系数/交流能量和视频流中的运动矢量，与高级语义查询之间存在强相关性。通过建立云查询语义与多模态压缩域特征之间的精确映射，DynaFilter使边缘能够识别并仅传输与RoI相关的相关数据。广泛的评估表明，DynaFilter在图像解码和后续推断中将像素数据的总量减少了1.6倍至7.1倍，并在视频流中实现了92.0%的带宽节省，相较于最先进的基线。此外，它在目标设备上的能耗减少了43.1%至88.6%，并在推断延迟上实现了1.6倍至3.0倍的加速。
cs.CV / 34 / 2607.10120
WeaveEarth: Structured Evidence Construction and Reasoning for Training-Free UHR Remote Sensing Understanding
WeaveEarth：无训练超高分辨率遥感理解的结构化证据构建与推理
Abstract
Ultra-High-Resolution (UHR) remote sensing image understanding requires Vision-Language Models (VLMs) to capture both the global scene layout and sparse yet task-critical local details under limited computational budgets. Existing methods mainly follow two paradigms. One is passive perception, which relies on resolution expansion or token compression and may therefore discard fine-grained details. The other is active perception, which depends on multi-round zooming and search, but suffers from high latency, contextual fragmentation, and error accumulation. We argue that a more effective path toward UHR understanding lies not in accessing more, but in organizing better. To this end, we propose WeaveEarth, a training-free framework that reformulates UHR understanding as a problem of structured evidence construction and reasoning under global context constraints. Specifically, WeaveEarth first employs Global-Aware Evidence Construction to select a compact, low-redundancy, and spatially complementary Minimal Support Evidence Set. It then introduces Structured Evidence Reasoning, which weaves local evidence, spatial metadata, and relative topology into a unified reasoning interface, thereby enhancing the VLM's ability to perform global-local joint reasoning. Extensive experiments show that WeaveEarth consistently outperforms strong baselines and existing UHR methods across multiple UHR remote sensing benchmarks and multiple frozen VLM backbones. Code is available at https://github.com/XianZhi-Ma/WeaveEarth.
Chinese Translation
超高分辨率（UHR）遥感图像理解需要视觉-语言模型（VLMs）在有限的计算预算下捕捉全球场景布局和稀疏但任务关键的局部细节。现有方法主要遵循两种范式。一种是被动感知，依赖于分辨率扩展或标记压缩，因此可能会丢失细粒度细节。另一种是主动感知，依赖于多轮缩放和搜索，但面临高延迟、上下文碎片化和错误累积的问题。我们认为，通往UHR理解的更有效路径不在于获取更多信息，而在于更好地组织信息。为此，我们提出了WeaveEarth，一个无训练的框架，将UHR理解重新表述为在全球上下文约束下的结构化证据构建与推理问题。具体而言，WeaveEarth首先采用全球感知证据构建，选择一个紧凑、低冗余且空间互补的最小支持证据集。然后引入结构化证据推理，将局部证据、空间元数据和相对拓扑编织成统一的推理接口，从而增强VLM进行全球-局部联合推理的能力。大量实验表明，WeaveEarth在多个UHR遥感基准和多个冻结的VLM骨干网络上始终优于强基线和现有UHR方法。代码可在 https://github.com/XianZhi-Ma/WeaveEarth 获取。
cs.CV / 35 / 2607.10130
TextGaze: Prompting Gaze Target Estimation with Textual Scene Cues
TextGaze：利用文本场景线索促进注视目标估计
Abstract
Gaze target estimation aims to infer the position of a person's gaze within a scene. Within mainstream design logic, multi-branch methods require extra supervision and annotations, while streamlined designs prioritize low-level visual saliency over true gaze intent. The former leads to a high annotation burden and hinders domain transfer, whereas the latter causes misalignment between predicted attention and actual gaze targets. To address this issue, we propose TextGaze, a unified cross-modal architecture that leverages a Large Vision-Language Model (LVLM) as scalable semantic guidance to balance the two design paradigms. The model extracts visual features from a frozen encoder and utilizes an LVLM to obtain gaze-aligned textual cues. We design a transformer-based fusion module with hierarchical text supervision to preserve task semantics. Lightweight decoding heads enable the joint prediction of gaze heatmaps and in-/out-of-frame status. We evaluate our method on four mainstream datasets, and the results show competitive performance across key metrics with robust cross-dataset generalisation without extra fine-tuning. Overall, we provide a streamlined alternative to traditional designs and highlight the potential of LVLMs as accessible auxiliary guidance for gaze estimation.
Chinese Translation
注视目标估计旨在推断一个人在场景中的注视位置。在主流设计逻辑中，多分支方法需要额外的监督和标注，而简化设计则优先考虑低级视觉显著性而非真实的注视意图。前者导致高标注负担并阻碍领域迁移，而后者则造成预测注意力与实际注视目标之间的不对齐。为了解决这一问题，我们提出了TextGaze，一种统一的跨模态架构，利用大型视觉语言模型（Large Vision-Language Model, LVLM）作为可扩展的语义指导，以平衡这两种设计范式。该模型从冻结的编码器中提取视觉特征，并利用LVLM获取与注视对齐的文本线索。我们设计了一个基于变换器的融合模块，结合分层文本监督以保持任务语义。轻量级解码头使得注视热图和帧内/帧外状态的联合预测成为可能。我们在四个主流数据集上评估了我们的方法，结果显示在关键指标上具有竞争力的性能，并且在没有额外微调的情况下实现了稳健的跨数据集泛化。总体而言，我们提供了一种简化的替代传统设计的方法，并强调了LVLM作为可获取的辅助指导在注视估计中的潜力。
cs.CV / 36 / 2607.10140
FlowPainter: Inpainting Optical Flow via Confidence-Guided Completion
FlowPainter：通过置信度引导的光流修复
Abstract
Existing optical flow methods broadly follow two paradigms: iterative optimization and diffusion-based estimation. Iterative methods, exemplified by RAFT, achieve high accuracy through recurrent refinement, but remain challenged by large displacements and complex motion. Diffusion-based methods introduce generative modeling and show promise in such ambiguous regions. However, existing diffusion models usually denoise the entire dense flow field from Gaussian noise, including simple regions where reliable motion can already be estimated by a lightweight network. This increases the denoising burden and may cause slow convergence and unstable training. To address this issue, we introduce FlowPainter, a diffusion-based optical flow framework that reformulates dense-flow generation as confidence-guided soft inpainting. FlowPainter employs a lightweight confidence-aware network to predict a rough flow and a pixel-wise confidence mask, distinguishing reliable simple regions from uncertain hard regions. The resulting simple-flow prior is used for confidence-based initialization and further injected into iterative denoising through confidence-gated residual guidance. With dynamically decaying guidance strength, FlowPainter stabilizes early denoising while preserving the flexibility of the diffusion model for late-stage detail refinement. Extensive experiments on public benchmarks, including Sintel, KITTI, and Spring, show that FlowPainter achieves strong accuracy under comparable training settings and converges more efficiently than existing diffusion-based optical flow methods, with notable gains on challenging benchmark splits. Our approach offers a practical way to integrate reliable discriminative priors with diffusion-based refinement for optical flow estimation. Our code is publicly available at https://github.com/mya012/FlowPainter.
Chinese Translation
现有的光流方法大致遵循两种范式：迭代优化和基于扩散的估计。以RAFT为例的迭代方法通过重复精炼实现了高精度，但在大位移和复杂运动的情况下仍面临挑战。基于扩散的方法引入了生成建模，并在这些模糊区域中显示出潜力。然而，现有的扩散模型通常从高斯噪声中去噪整个稠密流场，包括那些可以通过轻量级网络可靠估计的简单区域。这增加了去噪负担，并可能导致收敛缓慢和训练不稳定。为了解决这个问题，我们提出了FlowPainter，一种基于扩散的光流框架，将稠密流生成重新表述为置信度引导的软修复。FlowPainter采用轻量级的置信度感知网络来预测粗略流和逐像素置信度掩码，从而区分可靠的简单区域和不确定的困难区域。得到的简单流先验用于基于置信度的初始化，并进一步通过置信度门控残差引导注入到迭代去噪中。通过动态衰减的引导强度，FlowPainter在保持扩散模型在后期细节精炼灵活性的同时，稳定了早期去噪。在公共基准测试（包括Sintel、KITTI和Spring）上的大量实验表明，FlowPainter在可比的训练设置下实现了强大的准确性，并且比现有的基于扩散的光流方法更高效地收敛，在具有挑战性的基准划分上取得了显著的提升。我们的方法为将可靠的判别先验与基于扩散的精炼相结合提供了一种实用的方式。我们的代码已公开发布在https://github.com/mya012/FlowPainter。
cs.CV / 37 / 2607.10147
REVA-PO: Stabilizing Reinforcement Learning for Chest X-ray Report Generation
REVA-PO：用于胸部X光报告生成的稳定强化学习
Abstract
Automated chest X-ray report generation has recently benefited from reinforcement learning (RL) and large language models. However, RL training often suffers from instability or limited exploration due to fixed Kullback-Leibler (KL) regularization and a static reference policy that accumulates KL pressure over time. We propose Response-Weighted and Validation-Anchored Policy Optimization (REVA-PO), a RL framework that stabilizes long-term training via Response-Weighted Regularization (RER) and Validation-Anchored Policy Reset (VAPR). RER dynamically adjusts per-response KL weights based on advantage and reference-policy entropy, relaxing constraints for high-quality responses while tightening them for low-quality ones. Complementarily, VAPR periodically synchronizes the reference and current policies to the best validation checkpoint, resetting accumulated regularization pressure to expand the viable exploration space. To ensure a robust starting point, we employ a three-stage pipeline consisting of warm-up training, classifier-guided supervised fine-tuning, and RL. Extensive evaluations on MIMIC-CXR and IU-Xray demonstrate that REVA-PO sets new state-of-the-art benchmarks in both linguistic quality and clinical accuracy. Notably, BLEU-4 improves by 5.1% on MIMIC-CXR and 3.6% on IU-Xray, while CheXpert F1 and RadGraph F1 scores increase by 4.5% and 12.8%, respectively, over prior leading methods. The code is publicly available at https://github.com/LiGuo12/REVA_PO/.
Chinese Translation
自动化胸部X光报告生成最近受益于强化学习（RL）和大型语言模型。然而，由于固定的Kullback-Leibler（KL）正则化和随时间累积KL压力的静态参考策略，RL训练往往面临不稳定或有限探索的问题。我们提出了响应加权和验证锚定策略优化（REVA-PO），这是一个通过响应加权正则化（RER）和验证锚定策略重置（VAPR）来稳定长期训练的RL框架。RER根据优势和参考策略的熵动态调整每个响应的KL权重，为高质量响应放宽约束，而对低质量响应则收紧约束。作为补充，VAPR定期将参考策略和当前策略同步到最佳验证检查点，重置累积的正则化压力，以扩展可行的探索空间。为了确保一个稳健的起点，我们采用了一个由热身训练、分类器引导的监督微调和RL组成的三阶段流程。在MIMIC-CXR和IU-Xray上的广泛评估表明，REVA-PO在语言质量和临床准确性方面设定了新的最先进基准。值得注意的是，MIMIC-CXR上的BLEU-4提高了5.1%，而IU-Xray上提高了3.6%；同时，CheXpert F1和RadGraph F1分数分别比之前的领先方法提高了4.5%和12.8%。代码已公开发布在https://github.com/LiGuo12/REVA_PO/。
cs.CV / 38 / 2607.10165
EmoStyle: Affective Conditioning of Style-Specialist Experts for Emotional Image Generation
EmoStyle：情感图像生成的风格专家的情感调节
Abstract
Emotion-aware artistic image generation requires an image to match the input prompt, follow the specified artistic style, and convey the target emotion. In this challenge, the main difficulty is that the visual and affective attributes available in the training data are not explicitly provided at test time. Without these attributes, the generator has to decide not only what to depict, but also how the target emotion should be expressed through color, lighting, brushwork, composition, line, and layout. This creates a control gap between the available test prompt and the fine-grained conditions needed for emotion-aware artistic generation. To bridge this gap, we propose EmoStyle, a Z-Image-based framework that converts the input prompt into a structured generation state. An LLM reasoner first predicts affective cues (valence-arousal, dominant emotion, and therapeutic-effect labels) and an aspect-ratio decision. Instead of using these predictions only as additional prompt text, we encode the affective fields into an affective condition vector and inject it into the denoising blocks through AdaLN-style modulation. This allows the inferred control variables to directly guide the generation of intermediate features. Since emotional expression is also style-dependent, we further train a dedicated LoRA adapter for each artistic style bucket and select the corresponding expert during inference, enabling the same affective cues to be rendered with bucket-specific priors for color, texture, brushwork, and composition. Finally, a lightweight VLM-guided candidate selection step ranks the generated images based on prompt alignment, style consistency, emotional expression, and visual quality. In Track 1 of the AffectiveArt Challenge 2026, our USTC\_PI\_LAB\_TEAM submission achieved first place.
Chinese Translation
情感感知的艺术图像生成要求图像不仅要与输入提示匹配，还需遵循指定的艺术风格，并传达目标情感。在这一挑战中，主要困难在于训练数据中可用的视觉和情感属性在测试时并未明确提供。没有这些属性，生成器不仅需要决定描绘什么，还需决定如何通过颜色、光照、笔触、构图、线条和布局来表达目标情感。这在可用的测试提示与情感感知艺术生成所需的细粒度条件之间造成了控制差距。为了解决这一差距，我们提出了EmoStyle，一个基于Z-Image的框架，将输入提示转换为结构化生成状态。首先，LLM推理器预测情感线索（愉悦度-唤醒度、主导情感和治疗效果标签）以及长宽比决策。我们不仅将这些预测作为额外的提示文本使用，而是将情感领域编码为情感条件向量，并通过AdaLN风格调制将其注入去噪块中。这使得推断出的控制变量能够直接引导中间特征的生成。由于情感表达也依赖于风格，我们进一步为每个艺术风格类别训练了专用的LoRA适配器，并在推理时选择相应的专家，从而使相同的情感线索能够以类别特定的先验知识在颜色、纹理、笔触和构图上进行渲染。最后，一个轻量级的VLM引导候选选择步骤根据提示对齐、风格一致性、情感表达和视觉质量对生成的图像进行排序。在2026年情感艺术挑战赛的第一赛道中，我们的USTC_PI_LAB_TEAM提交作品获得了第一名。
cs.CV / 39 / 2607.10188
BiLoG-Net: A Bi-Context Location-Guided Network for Breast Mass Segmentation and Malignancy Classification in Mammography
BiLoG-Net：一种双上下文位置引导网络用于乳腺肿块分割和恶性程度分类的乳腺X线摄影
Abstract
Breast cancer remains the most commonly diagnosed malignancy among women worldwide, yet accurate detection and characterization of breast masses in mammography remain challenging due to subtle intensity variations, heterogeneous tissue densities, and indistinct lesion boundaries that complicate radiological interpretation. To address these limitations, we propose BiLoG-Net, a deep learning framework that jointly performs breast mass segmentation and malignancy classification through bi-context location-aware feature modeling and segmentation-guided attention mechanisms. Our architecture integrates a novel encoder-decoder paradigm with Fire-based feature extraction, lightweight global and local feature enhancement modules, and adaptive location-aware gating to simultaneously capture long-range contextual dependencies and fine-grained boundary-sensitive details. Unlike conventional multi-stage pipelines, our tightly coupled multi-task design enables mutual reinforcement between pixel-level localization and image-level diagnosis, reducing error propagation while producing spatially grounded malignancy predictions. Evaluated on CBIS-DDSM and INBreast benchmarks, BiLoG-Net achieves state-of-the-art performance with Dice scores of 94.20% and 93.10%, classification accuracies of 95.20% and 93.60%, and AUC values of 97.10% and 96.00%, respectively, substantially outperforming existing CNN and transformer-based baselines. By combining precise boundary delineation with reliable malignancy assessment in a single end-to-end model, this work holds strong potential for clinical computer-aided detection systems, helping radiologists prioritize suspicious cases and improve screening efficiency in busy clinical settings.
Chinese Translation
乳腺癌仍然是全球女性中最常被诊断的恶性肿瘤，但由于细微的强度变化、异质的组织密度和模糊的病变边界，乳腺X线摄影中乳腺肿块的准确检测和特征表征仍然具有挑战性。为了解决这些限制，我们提出了BiLoG-Net，这是一种深度学习框架，通过双上下文位置感知特征建模和分割引导的注意机制，联合执行乳腺肿块分割和恶性程度分类。我们的架构集成了一种新颖的编码器-解码器范式，结合基于Fire的特征提取、轻量级的全局和局部特征增强模块，以及自适应位置感知门控，以同时捕捉长距离上下文依赖和细粒度边界敏感细节。与传统的多阶段流程不同，我们紧密耦合的多任务设计使得像素级定位与图像级诊断之间实现了相互增强，减少了错误传播，同时产生了空间上有依据的恶性程度预测。在CBIS-DDSM和INBreast基准测试中评估，BiLoG-Net实现了最先进的性能，Dice分数分别为94.20%和93.10%，分类准确率为95.20%和93.60%，AUC值为97.10%和96.00%，显著优于现有的基于CNN和变换器的基线。通过在一个端到端模型中结合精确的边界描绘与可靠的恶性程度评估，本研究在临床计算机辅助检测系统中具有强大的潜力，帮助放射科医生优先处理可疑病例，提高繁忙临床环境中的筛查效率。
cs.CV / 40 / 2607.10214
ScratNet: A Swin-Based Multi-Scale Dilated Network with Precision Refinement for Semiconductor Scratch Segmentation
ScratNet：一种基于Swin的多尺度膨胀网络，具有精度细化功能，用于半导体划痕分割
Abstract
Surface scratch defects in semiconductor manufacturing pose significant challenges due to their irregular shapes, low contrast, and varying scales. Traditional inspection methods often struggle to detect such defects reliably, especially in complex imaging scenarios. While deep learning approaches based on Convolutional Neural Networks (CNNs) have improved accuracy, they often fail to capture fine-grained edge details. To address these limitations, we propose ScratNet, a novel end-to-end scratch segmentation framework that integrates a modified Swin Transformer backbone with a tailored decoder. The decoder incorporates a Multi-Scale Dilated Aggregation (MDA) module to capture both local and global context, a Stem Integration Module (SIM) to restore spatial detail, and a Precision Refinement (PR) branch that enhances boundary sharpness using anisotropic convolutions. Through this stage-adaptive feature aggregation and boundary-aware refinement, ScratNet achieves superior accuracy on thin and irregular defects. Extensive experiments demonstrate that ScratNet consistently outperforms existing methods, providing a scalable and robust solution for automated scratch inspection in high-precision manufacturing.
Chinese Translation
半导体制造中的表面划痕缺陷由于其不规则形状、低对比度和不同尺度而带来了显著挑战。传统的检测方法在复杂成像场景中往往难以可靠地检测此类缺陷。虽然基于卷积神经网络（CNN）的深度学习方法提高了准确性，但它们往往无法捕捉细致的边缘细节。为了解决这些局限性，我们提出了ScratNet，一种新颖的端到端划痕分割框架，集成了修改后的Swin Transformer主干和定制解码器。该解码器结合了多尺度膨胀聚合（MDA）模块，以捕捉局部和全局上下文，干扰整合模块（SIM）以恢复空间细节，以及精度细化（PR）分支，通过各向异性卷积增强边界清晰度。通过这种阶段自适应特征聚合和边界感知细化，ScratNet在薄且不规则缺陷上实现了卓越的准确性。大量实验表明，ScratNet在性能上始终优于现有方法，为高精度制造中的自动划痕检测提供了一种可扩展且稳健的解决方案。
cs.CV / 41 / 2607.10231
PhenoEmbed: Self-Supervised Multispectral UAV Time-Series Embeddings for Individual Tree Crown Phenology
PhenoEmbed：自监督多光谱无人机时间序列嵌入用于个体树冠物候研究
Abstract
Tree crowns are a challenging target for resilient AI because they are not static objects: their spectral response, internal texture, translucency, and apparent boundaries change substantially across the growing season. We develop PhenoEmbed, a self-supervised crown-centric temporal embedding model trained with contrastive and masked reconstruction objectives on HeideBench, an 18-date UAV multispectral time-series benchmark for forest crown phenology in D{\"o}lauer Heide. The model treats seasonal crown dynamics as phenological appearance change driven by leaf emergence, canopy closure, senescence, and leaf-off conditions. Segmented tree crown polygons are retained as object anchors to extract aligned crown-centered crops through time, allowing one 256-dimensional vector summarizing seasonal crown appearance to be learned per tree. On 5,885 crop-safe crowns, the exported embeddings show structured low-dimensional organization, with the first two principal components explaining 25.1\% of variance and nearest-neighbor retrieval producing a median top-1 cosine similarity of 0.946. Compared with handcrafted temporal features and a learned mean-pooling baseline, PhenoEmbed yields substantially more compact nearest-neighbor structure, while ablations show that the contrastive loss, masked reconstruction loss, and explicit seasonal time features each affect the structure of the learned embedding space. These results support PhenoEmbed as a reusable forest crown representation learner and motivate future downstream tests of whether such features improve tree-level models under seasonal change.
Chinese Translation
树冠是弹性人工智能的一个挑战性目标，因为它们不是静态物体：其光谱响应、内部纹理、半透明性和明显边界在生长季节中发生显著变化。我们开发了PhenoEmbed，一种以树冠为中心的自监督时间嵌入模型，该模型在HeideBench上进行训练，HeideBench是一个包含18个日期的无人机多光谱时间序列基准，用于Dölauer Heide地区的森林树冠物候研究。该模型将季节性树冠动态视为由叶片萌发、树冠闭合、衰老和落叶条件驱动的物候外观变化。分割后的树冠多边形被保留为对象锚点，以提取随时间变化的对齐树冠中心裁剪，从而允许为每棵树学习一个256维的向量来总结季节性树冠外观。在5,885个安全裁剪的树冠上，导出的嵌入显示出结构化的低维组织，前两个主成分解释了25.1%的方差，最近邻检索产生的中位数top-1余弦相似度为0.946。与手工制作的时间特征和学习的均值池化基线相比，PhenoEmbed产生了更紧凑的最近邻结构，而消融实验表明，对比损失、掩蔽重建损失和显式季节时间特征各自影响学习到的嵌入空间的结构。这些结果支持PhenoEmbed作为可重用的森林树冠表示学习器，并激励未来下游测试这些特征是否能在季节变化下改善树级模型。
cs.CV / 42 / 2607.10237
CoSAG: Compact Semantic Anchor Gaussians via Training-Free Rate-Distortion Coding
CoSAG：通过无训练率失真编码实现紧凑的语义锚高斯
Abstract
Open-vocabulary 3D scene understanding is commonly achieved by embedding 2D vision-language features such as CLIP into a 3D Gaussian Splatting scene, turning it into a text-queryable semantic field. However, attaching a high-dimensional feature to each of millions of Gaussians inflates a single scene to gigabytes, which makes storage and deployment the real bottleneck of these fields. Existing compact methods each learn and ship a per-scene codec, an autoencoder, a quantized codebook, or a distilled feature field, entangling field construction with field storage and never compressing the per-Gaussian assignment that holds the bulk of the cost. We argue that construction and storage should be decoupled, and that storage is a rate-distortion problem over the per-Gaussian binding to a small anchor table, a structure no prior open-vocabulary method compresses. We present CoSAG, which constructs the field without any per-scene training through a closed-form transmittance-weighted lift, spatially grounded semantic anchors, and multi-view denoising, and stores it with a spatially predictive entropy coder that ships no decoder. Because the anchors are spatially grounded, the binding is predictable and therefore highly compressible. The transmittance-weighted lift and multi-view denoising yield a clean, view-consistent assignment, so the entropy coder spends almost no rate on correcting noise and instead codes only the residual against its spatial prediction. CoSAG reaches sub-megabyte storage while matching or exceeding the state of the art across the 2D-rendered, 3D-selection, and dense-LSeg protocols, reducing field size by 37 to 76x relative to LangSplatV2 at higher accuracy.
Chinese Translation
开放词汇的三维场景理解通常通过将二维视觉-语言特征（如 CLIP）嵌入三维高斯散射场景中来实现，从而将其转变为可通过文本查询的语义场。然而，将高维特征附加到数百万个高斯上会使单个场景膨胀到数GB，这使得存储和部署成为这些领域的真正瓶颈。现有的紧凑方法各自学习并传输每个场景的编解码器、自动编码器、量化代码本或蒸馏特征场，将场景构建与场景存储纠缠在一起，且从未压缩每个高斯分配所占的主要成本。我们认为，构建与存储应当解耦，存储是一个关于每个高斯绑定到小锚表的率失真问题，而这一结构是之前的开放词汇方法所未压缩的。我们提出了 CoSAG，它通过闭式形式的透射加权提升、空间基础的语义锚和多视图去噪，在没有任何每场景训练的情况下构建场，并通过一个不需要解码器的空间预测熵编码器进行存储。由于锚是空间基础的，绑定是可预测的，因此高度可压缩。透射加权提升和多视图去噪产生了干净的、视图一致的分配，因此熵编码器几乎不花费任何率来纠正噪声，而是仅对其空间预测的残差进行编码。CoSAG 实现了亚兆字节的存储，同时在 2D 渲染、3D 选择和密集 LSeg 协议中匹配或超过了当前最先进的水平，相较于 LangSplatV2 在更高精度下将场的大小减少了 37 到 76 倍。
cs.CV / 43 / 2607.10238
Benchmarking Dynamic Affective Reasoning: A Viewer-Centric Video Emotion Dataset
动态情感推理基准测试：以观众为中心的视频情感数据集
Abstract
Video emotion analysis is typically framed as a static classification problem, treating each clip as an independent labeled unit. However, such a formulation overlooks a key psychological fact: emotions change as a result of cumulative reactions to consecutive causal events. To bridge this gap, we introduce Dynamic Affective Reasoning, the first large-scale benchmark for viewer-centric affect transitions and causal reasoning over consecutive video events. DAR contains 15,087 videos and 36,908 event-aligned affective segments annotated with 27 emotion categories. Unlike existing video-based emotion datasets, DAR presents a new viewer-centric perspective on fine-grained emotional expressions and transitions, and provides dense, temporally grounded, and causally explicit reasoning chains. Based on DAR, we formally define three challenging tasks: affective segmentation, fine-grained emotion classification, and affective reasoning. Complementing this benchmark, we propose DAR-R1, a two-stage framework that combines supervised fine-tuning with Group Relative Policy Optimization. Experiments across 10+ MLLMs show that DAR-R1 sets a new state-of-the-art for dynamic affective reasoning, in terms of both emotional localization and affective reasoning. Project page: https://github.com/Zhang-Zhiyan/DAR.
Chinese Translation
视频情感分析通常被视为一个静态分类问题，将每个片段视为一个独立的标记单元。然而，这种表述忽视了一个关键的心理学事实：情感是对连续因果事件的累积反应而变化的。为了解决这一问题，我们提出了动态情感推理（Dynamic Affective Reasoning），这是首个针对以观众为中心的情感转变和连续视频事件因果推理的大规模基准数据集。DAR包含15,087个视频和36,908个与事件对齐的情感片段，标注了27种情感类别。与现有的视频情感数据集不同，DAR提供了一个新的以观众为中心的视角，关注细粒度的情感表达和转变，并提供密集的、时间上有依据的和因果明确的推理链。基于DAR，我们正式定义了三个具有挑战性的任务：情感分割、细粒度情感分类和情感推理。作为对这一基准的补充，我们提出了DAR-R1，一个结合了监督微调和群体相对策略优化的两阶段框架。在10多个多模态学习模型（MLLMs）上的实验表明，DAR-R1在情感定位和情感推理方面都设立了动态情感推理的新状态。项目页面：https://github.com/Zhang-Zhiyan/DAR。
cs.CV / 44 / 2607.10240
What Does Your Short-Answer VQA Score Actually Measure? Evaluator-Dependent Instability in Multimodal Short-Answer Benchmarks
你的短答案视觉问答评分到底测量了什么？多模态短答案基准中的评估者依赖性不稳定性
Abstract
Short-answer VQA benchmarks conflate two distinct quantities: whether a model's answer is semantically correct, and whether that answer matches the surface form expected by the automatic evaluator. We study this conflation across six vision--language models and six benchmarks, using a human-validated semantic judge (97.6% precision) to audit over 37k official errors. A second text-only judge reproduces the same benchmark-level false-negative pattern, showing that the effect is not an artifact of a single audit model. On text-rich benchmarks, up to half of these errors are semantically acceptable answers penalized purely for surface-form mismatch. This instability is structured by answer type: extractive and multi-span answers are far more evaluator-sensitive than scalar answers. Benign prompt and context rewrites further destabilize official outcomes, flipping item-level correctness at substantial rates without changing the underlying task. A deterministic CPU-only contract repair confirms that the undercount is partially recoverable. These findings imply that official short-answer VQA scores should be accompanied by semantic audits and answer-type diagnostics to remain interpretable.
Chinese Translation
短答案视觉问答（VQA）基准混淆了两个不同的量：模型的答案是否在语义上正确，以及该答案是否与自动评估器所期望的表面形式匹配。我们研究了这种混淆现象，分析了六种视觉-语言模型和六个基准，利用一个经过人工验证的语义评估者（精确度为97.6%）对超过37,000个官方错误进行审计。第二个仅基于文本的评估者重现了相同基准级别的假阴性模式，表明这一效应并非单一审计模型的伪影。在文本丰富的基准中，这些错误中多达一半是语义上可接受的答案，仅因表面形式不匹配而受到惩罚。这种不稳定性由答案类型结构化：提取式和多跨度答案对评估者的敏感性远高于标量答案。良性的提示和上下文重写进一步破坏了官方结果，在不改变基础任务的情况下，以相当大的比例翻转了项目级的正确性。一个确定性的仅CPU合同修复确认了这一低估部分是可恢复的。这些发现意味着，官方的短答案VQA评分应伴随语义审计和答案类型诊断，以保持可解释性。
cs.CV / 45 / 2607.10278
Geometry-aware Gaussian Prior and Axial Attention for Cervical Cytology Image Classification
几何感知高斯先验与轴向注意力在宫颈细胞学图像分类中的应用
Abstract
Accurate cervical cytology image classification is a key component of automated cervical cancer screening, where reliable recognition of normal, precancerous, and cancer-associated cellular patterns from Pap smear images can improve screening efficiency and diagnostic consistency. However, this task remains challenging because cervical cells exhibit complex morphology, subtle intra-class variations, and strong inter-class similarities. Existing convolution-based models capture local texture well but have limited ability to model long-range relationships, whereas attention-based models provide broader context but often lack explicit structural guidance. To address these limitations, we propose a geometry-aware classification framework for cervical cancer screening-oriented cytology image analysis, incorporating semantic abstraction and structural priors learned from pre-trained vision-language features. The method uses Gaussian expert modules to generate axis-wise priors from global semantic information, capturing structural regularities such as nuclear alignment and cellular spatial organization. These priors are embedded into an axial self-attention module to modulate similarity computation along horizontal and vertical directions, improving long-range dependency modeling and structure-sensitive feature interaction. Experiments on the Mendeley liquid-based cytology and SIPaKMeD datasets show that the proposed method achieves 99.48% accuracy on the former and 96.08% on the latter, with balanced gains in recall, precision, and overall classification performance. Visual analysis further shows that the learned priors highlight diagnostically relevant cellular regions, demonstrating the potential of the proposed framework as a screening-oriented decision-support tool for cervical cytology.
Chinese Translation
准确的宫颈细胞学图像分类是自动化宫颈癌筛查的关键组成部分，可靠地识别来自巴氏涂片图像的正常、癌前和癌相关细胞模式可以提高筛查效率和诊断一致性。然而，这一任务仍然具有挑战性，因为宫颈细胞表现出复杂的形态、细微的类内变异和强烈的类间相似性。现有的基于卷积的模型能够很好地捕捉局部纹理，但在建模长距离关系方面能力有限，而基于注意力的模型提供了更广泛的上下文，但通常缺乏明确的结构指导。为了解决这些局限性，我们提出了一种几何感知分类框架，用于面向宫颈癌筛查的细胞学图像分析，结合了从预训练的视觉-语言特征中学习的语义抽象和结构先验。该方法使用高斯专家模块从全局语义信息中生成轴向先验，捕捉结构规律，如细胞核对齐和细胞空间组织。这些先验被嵌入到轴向自注意力模块中，以调节沿水平和垂直方向的相似性计算，从而改善长距离依赖建模和结构敏感特征交互。在Mendeley液基细胞学和SIPaKMeD数据集上的实验表明，所提方法在前者上达到了99.48%的准确率，在后者上达到了96.08%的准确率，并在召回率、精确度和整体分类性能上均取得了平衡的提升。视觉分析进一步表明，学习到的先验突出了具有诊断相关性的细胞区域，展示了所提框架作为宫颈细胞学筛查决策支持工具的潜力。
cs.CV / 46 / 2607.10287
InterPet4D: A Multimodal 4D Human-Pet Interaction Dataset for Pet Motion Generation
InterPet4D：一个用于宠物运动生成的多模态4D人宠互动数据集
Abstract
Human-pet interaction estimation and generation remain underexplored due to the absence of a high-quality large-scale dataset. We present InterPet4D, the first multimodal dataset capturing natural interactions between humans and dogs. Using a synchronized multi-view capture system, we record human-dog obedience tasks and provide annotations for both humans and dogs, including multi-view and egocentric videos, segmentations, 2D and 3D keypoints, meshes, and audio tracks. InterPet4D consists of 6.8 million frames collected from 13 dogs of 11 breeds interacting with 23 human participants. We further introduce the InterPetMoGen framework for human-pet interaction motion generation. Our proposed model achieves an FID score of 11.21 and substantially outperforms the Seq2Seq and DiT baselines, demonstrating the effectiveness of InterPet4D for modeling realistic human-pet interactions.
Chinese Translation
由于缺乏高质量的大规模数据集，人宠互动的估计和生成仍然未得到充分探索。我们提出了InterPet4D，这是第一个捕捉人类与狗之间自然互动的多模态数据集。通过使用同步的多视角捕捉系统，我们记录了人狗服从任务，并为人类和狗提供了注释，包括多视角和自我中心视频、分割、2D和3D关键点、网格以及音频轨道。InterPet4D由来自11个品种的13只狗与23名人类参与者互动收集的680万帧组成。我们进一步介绍了InterPetMoGen框架，用于人宠互动运动生成。我们提出的模型达到了11.21的FID分数，并显著优于Seq2Seq和DiT基线，证明了InterPet4D在建模真实人宠互动方面的有效性。
cs.CV / 47 / 2607.10298
Structured Evidence Selection for Weakly Supervised Video Anomaly Detection
弱监督视频异常检测的结构化证据选择
Abstract
Weakly supervised video anomaly detection relies solely on video-level labels for training, making it difficult to accurately localize anomalous events in complex scenes. In real-world videos, anomalous behaviors exhibit large variations in appearance and temporal duration, while scene appearance and action dynamics are often tightly entangled. Consequently, existing models tend to rely on scene-related statistical cues rather than true behavioral deviations, resulting in unstable detection performance. To address this challenge, we propose a Structured Evidence Selection framework (SESAD) that reformulates anomaly detection as a structured reasoning process over clip-level visual evidence. Instead of directly mapping aggregated features to anomaly scores, SESAD reorganizes clip representations into semantically structured candidate evidence and performs context-conditioned selection under scene and action constraints. This mechanism adaptively emphasizes anomaly-relevant semantics while suppressing scene interference, thereby alleviating semantic entanglement under weak supervision. Furthermore, we introduce a lightweight geometric discrimination module that constructs a dual-prototype structure in the embedding space, enabling anomaly decisions through relative geometric relations. Extensive experiments on UBnormal, ShanghaiTech, and UCF-Crime show that SESAD achieves 67.92, 97.99, and 88.46 AUC, respectively, while maintaining high computational efficiency and overall consistently stable anomaly discrimination.
Chinese Translation
弱监督视频异常检测仅依赖视频级标签进行训练，这使得在复杂场景中准确定位异常事件变得困难。在现实世界的视频中，异常行为在外观和时间持续性上表现出较大的变化，而场景外观和动作动态往往紧密交织。因此，现有模型倾向于依赖与场景相关的统计线索，而非真实的行为偏差，导致检测性能不稳定。为了解决这一挑战，我们提出了一种结构化证据选择框架（Structured Evidence Selection framework，SESAD），将异常检测重新表述为对剪辑级视觉证据的结构化推理过程。SESAD并不直接将聚合特征映射到异常分数，而是将剪辑表示重新组织为语义结构化的候选证据，并在场景和动作约束下执行上下文条件选择。这一机制自适应地强调与异常相关的语义，同时抑制场景干扰，从而缓解弱监督下的语义纠缠。此外，我们引入了一个轻量级几何区分模块，该模块在嵌入空间中构建了一个双原型结构，通过相对几何关系实现异常决策。在UBnormal、ShanghaiTech和UCF-Crime上的大量实验表明，SESAD分别达到了67.92、97.99和88.46的AUC，同时保持了高计算效率和整体一致稳定的异常区分能力。
cs.CV / 48 / 2607.10301
ChartSync: A Benchmark for Visuo-Logical Cascading Chart Editing
ChartSync：一种用于视觉逻辑级联图表编辑的基准
Abstract
Generative image editing models struggle with structured statistical charts when data modifications require geometric synchronization. We formalize this task as Visuo-Logical Cascading Editing (VLCE). However, existing methods remain confined to localized text substitutions and struggle with dependency-aware cascading updates. To systematically evaluate this capability, we introduce ChartSync, an expert-validated benchmark constructed via a programmatic rendering pipeline that guarantees deterministic visuo-logical coupling for the ground truth. ChartSync comprises 870 triplets across 9 chart categories and 4 task types, including 235 geometry-coupled VLCE instances that specifically test cascading text-to-geometry synchronization. We further evaluate these instances via a two-tier framework combining objective visual metrics with a vision-language model judge paradigm to assess low-level fidelity alongside multimodal comprehension and reasoning. Evaluating 14 image editing models and one code-mediated pipeline reveals a nuanced capability gap: most open-source models suffer severe drops in geometric synchronization, while only two frontier proprietary models show emerging VLCE capability, with their residual errors mainly involving semantic isolation and background corruption. Our detailed error analysis deconstructs these failure paradigms to identify core meta-abilities for guiding future multimodal architectures. The ChartSync dataset and code are publicly released at https://github.com/kaka-yjk/ChartSyncCodebase.
Chinese Translation
生成图像编辑模型在数据修改需要几何同步时，处理结构化统计图表的能力较弱。我们将这一任务形式化为视觉逻辑级联编辑（Visuo-Logical Cascading Editing, VLCE）。然而，现有方法仍然局限于局部文本替换，并在依赖感知的级联更新中表现不佳。为了系统地评估这一能力，我们引入了ChartSync，这是一个经过专家验证的基准，通过程序化渲染管道构建，确保了地面真实数据的确定性视觉逻辑耦合。ChartSync包含870个三元组，涵盖9个图表类别和4种任务类型，其中包括235个几何耦合的VLCE实例，专门测试级联文本与几何的同步。我们进一步通过一个两级框架评估这些实例，该框架结合了客观视觉指标与视觉-语言模型评判范式，以评估低级保真度以及多模态理解和推理。对14个图像编辑模型和一个代码中介管道的评估揭示了一个细微的能力差距：大多数开源模型在几何同步方面遭遇严重下降，而只有两个前沿专有模型显示出新兴的VLCE能力，其残余错误主要涉及语义孤立和背景损坏。我们的详细错误分析解构了这些失败范式，以识别指导未来多模态架构的核心元能力。ChartSync数据集和代码已公开发布在 https://github.com/kaka-yjk/ChartSyncCodebase。
cs.CV / 49 / 2607.10308
Generalize LMMs to Versatile Visual Modalities via Fabricated Modality Synthesis
通过合成模态扩展大型多模态模型至多样化视觉模态
Abstract
Despite the advancements of Large Multimodal Models (LMMs) in RGB vision, their ability to generalize to unseen visual modalities remains a largely unexplored challenge. We argue that different visual modalities are merely distinct samplings of the same physical world. Therefore, effective generalization requires models to possess both modality-agnostic perception of scene semantics and the adaptability to modality-specific characteristics. To achieve this, we propose a training framework, VVM-Tuning, to equip LMMs with these capabilities through modality synthesis and modality contexts. Specifically, we synthesize diverse appearance-varied images from RGB scenes, training the model to disentangle invariant semantics from varying visual appearances, and align these appearances with language for visual concepts decoupled from modalities. We then introduce modality contexts in the prompt and use instruction tuning to assist the model in mapping these appearance variations back to modality-related attributes, enabling zero-shot adaptation to unseen modalities during inference. To facilitate research in this direction, we introduce VVM-Bench, a comprehensive benchmark featuring 6 real and synthetic modalities to evaluate semantic perception and modality understanding. Experiments demonstrate that, via our training on synthetic modalities, 5 tested models exhibit consistent improvements on both real-world and novel synthetic modalities without in-modality training. Source code and data will be publicly available at https://github.com/Hunter-Will/VVM-Tuning.
Chinese Translation
尽管大型多模态模型（LMMs）在RGB视觉方面取得了进展，但它们在未见视觉模态上的泛化能力仍然是一个尚未充分探索的挑战。我们认为，不同的视觉模态仅仅是对同一物理世界的不同采样。因此，有效的泛化要求模型具备对场景语义的模态无关感知能力以及对特定模态特征的适应能力。为此，我们提出了一种训练框架VVM-Tuning，通过模态合成和模态上下文为LMMs赋予这些能力。具体而言，我们从RGB场景中合成多样的外观变化图像，训练模型将不变的语义与变化的视觉外观解耦，并将这些外观与语言对齐，以便将视觉概念与模态解耦。然后，我们在提示中引入模态上下文，并使用指令调优来帮助模型将这些外观变化映射回与模态相关的属性，从而在推理过程中实现对未见模态的零-shot适应。为了促进该方向的研究，我们推出了VVM-Bench，这是一个全面的基准，包含6种真实和合成模态，用于评估语义感知和模态理解。实验表明，通过对合成模态的训练，5个测试模型在真实世界和新颖的合成模态上均表现出一致的改善，而无需进行模态内训练。源代码和数据将公开发布在https://github.com/Hunter-Will/VVM-Tuning。
cs.CV / 50 / 2607.10329
Imperceptible and Reversible Adversarial Examples against Vision-Language Models for Privacy Protection
针对视觉语言模型的不可察觉和可逆对抗样本以保护隐私
Abstract
Vision Language Models (VLMs) offer powerful multimodal ability but also expose users to text-based privacy attacks where adversaries crawl online photos and query VLMs to extract sensitive attributes. Existing reversible adversarial example (RAE) methods protect images in purely visual tasks but fail in multimodal settings, and current adversarial examples on VLMs rely on high frequency noise that severely degrades visual quality. We propose CloakDiff, the first framework for reversible, high fidelity privacy protection against text-based query attacks in VLMs. CloakDiff produces imperceptible adversarial examples by combining diffusion based adversarial editing with an invertible network that embeds the original image for lossless recovery. It perturbs both pixel space embeddings and manipulates latent cross attention maps to ensure strong cross-model and cross-prompt transferability while preserving global visual structure. To further enhance fidelity, we design EDM Heuristic Sampling, a principled diffusion schedule for adversarial guidance. Experiments on multiple datasets and VLMs demonstrate that CloakDiff delivers multimodal privacy preservation with high visual quality and reversibility.
Chinese Translation
视觉语言模型（VLMs）提供强大的多模态能力，但也使用户面临基于文本的隐私攻击，攻击者通过爬取在线照片并查询VLMs来提取敏感属性。现有的可逆对抗样本（RAE）方法在纯视觉任务中能够保护图像，但在多模态设置中却失效，而当前针对VLMs的对抗样本依赖于高频噪声，这严重降低了视觉质量。我们提出了CloakDiff，这是第一个针对VLMs中基于文本查询攻击的可逆高保真隐私保护框架。CloakDiff通过将基于扩散的对抗编辑与可逆网络相结合，生成不可察觉的对抗样本，该网络嵌入原始图像以实现无损恢复。它扰动像素空间嵌入，并操控潜在的交叉注意力图，以确保强大的跨模型和跨提示的可迁移性，同时保持全局视觉结构。为了进一步增强保真度，我们设计了EDM启发式采样，这是一种用于对抗指导的原则性扩散调度。在多个数据集和VLMs上的实验表明，CloakDiff能够以高视觉质量和可逆性实现多模态隐私保护。
cs.CV / 51 / 2607.10357
GRC-ProbNet: Uncertainty-aware Feature Extraction for Cardiovascular Disease Classification
GRC-ProbNet：面向不确定性的特征提取用于心血管疾病分类
Abstract
The automatic detection and classification of cardiovascular disease (CVD) from computed tomography (CT) images plays an important role in clinical practice. Recently, a hybrid pipeline (GRC-Net) for CVD classification was proposed, which leverages a deep-learning-based segmentation and registration method to extract radiomic and geometric features. However, GRC-Net relies on a deterministic segmentation mask, without considering the inherent ambiguity associated with cardiac anatomy. In this paper, we propose GRC-ProbNet, which takes advantage of a deep ensemble to produce multiple segmentation masks for a given input. From these masks, we extract multiple uncertainty features. We analyze these uncertainty features for both their correlation with segmentation error and their propagation effects on downstream CVD classification performance. Our experiments on the publicly available MM-WHS and ASOCA datasets show that the uncertainty measure that best reflects segmentation quality is not necessarily the one that provides the strongest signal for downstream CVD classification. Overall, our results demonstrate that GRC-ProbNet utilizing uncertainty features substantially improves CVD classification AUROC (92.92\) compared to the baseline GRC-Net model (91.25%). Our code is publicly available: https://github.com/biomedia-mira/GRC-ProbNet.
Chinese Translation
从计算机断层扫描（CT）图像中自动检测和分类心血管疾病（CVD）在临床实践中发挥着重要作用。最近，提出了一种用于CVD分类的混合管道（GRC-Net），该方法利用基于深度学习的分割和配准方法提取放射组学和几何特征。然而，GRC-Net依赖于确定性的分割掩膜，未考虑心脏解剖结构固有的模糊性。在本文中，我们提出了GRC-ProbNet，它利用深度集成方法为给定输入生成多个分割掩膜。我们从这些掩膜中提取多个不确定性特征。我们分析这些不确定性特征与分割误差的相关性及其对下游CVD分类性能的传播影响。我们在公开可用的MM-WHS和ASOCA数据集上的实验表明，最能反映分割质量的不确定性度量不一定是为下游CVD分类提供最强信号的度量。总体而言，我们的结果表明，GRC-ProbNet利用不确定性特征显著提高了CVD分类的AUROC（92.92%），相比于基线GRC-Net模型（91.25%）。我们的代码公开可用： https://github.com/biomedia-mira/GRC-ProbNet。
cs.CV / 52 / 2607.10358
Benchmarking the Robustness of Foundation Models for Mammography under Domain Shift
在领域转移下基准测试基础模型在乳腺X光检查中的鲁棒性
Abstract
Foundation models are increasingly used as image feature extractors for mammography, but their robustness under external domain shift remains unclear. We benchmark 15 foundation-model backbones across breast density, BI-RADS severity, and cancer status using a unified frozen-backbone linear-probe protocol, training on 3 source datasets and evaluating on 12 task-compatible out-of-distribution (OOD) datasets after label harmonization. Mammography-specific vision-language models (Mammo-FM and MaMA) provide the strongest mean OOD performance, but robustness is not explained by mammography exposure alone. DINOv3 remains a competitive vision-only baseline, and mammography-adapted pretraining does not consistently improve generalization. Dataset-level analysis further shows that even leading models show heterogeneous performance across datasets. Feature-space inspection reveals that useful representations can preserve clinical signal while retaining dataset and acquisition structure. These findings highlight dataset-level OOD evaluation as a central criterion for assessing mammography representations. Our code is publicly available: https://github.com/biomedia-mira/mammo-ood.
Chinese Translation
基础模型越来越多地被用作乳腺X光检查的图像特征提取器，但它们在外部领域转移下的鲁棒性仍不清楚。我们基于统一的冻结骨干线性探针协议，对15个基础模型骨干进行基准测试，涵盖乳腺密度、BI-RADS严重程度和癌症状态，训练于3个源数据集，并在经过标签统一化处理的12个任务兼容的分布外（OOD）数据集上进行评估。特定于乳腺X光检查的视觉-语言模型（Mammo-FM和MaMA）提供了最强的平均OOD性能，但鲁棒性并不单靠乳腺X光检查的曝光来解释。DINOv3仍然是一个具有竞争力的仅视觉基线，而适应乳腺X光检查的预训练并未始终改善泛化能力。数据集级分析进一步表明，即使是领先模型在不同数据集上的表现也存在异质性。特征空间检查表明，有用的表示可以在保留数据集和采集结构的同时保留临床信号。这些发现强调了数据集级OOD评估作为评估乳腺X光检查表示的核心标准。我们的代码公开可用： https://github.com/biomedia-mira/mammo-ood。
cs.CV / 53 / 2607.10365
Gradient-Skipping Relevance Propagation for Efficient Explainability of Vision Transformers
用于高效解释视觉变换器的梯度跳跃相关传播
Abstract
Vision Transformers (ViTs) are difficult to interpret because current methods of relevance propagation and attention flow do not fully consider some key architectural features, such as the uneven importance of attention heads and residual connections. Prior approaches typically assume uniform importance across attention heads; furthermore, they model skip connections as identity paths, leading to inaccurate relevance attribution. To address these issues, we introduce GradSkip, a novel relevance propagation method for ViTs based on adaptive head weighting and skip-aware propagation. GradSkip models the different importance of the attention heads and dynamically distributes relevance between the attention and residual paths. Experiments on ImageNet1K and BloodMNIST demonstrate a state-of-the-art faithfulness of GradSkip while requiring over 14 times fewer GFLOPs than the best-performing existing approaches. Additional evaluations using transformer-based segmentation confirm improved localization and alignment with ground-truth regions.
Chinese Translation
视觉变换器（ViTs）难以解释，因为当前的相关传播和注意力流方法并未充分考虑一些关键的架构特征，例如注意力头的重要性不均以及残差连接。以往的方法通常假设注意力头的重要性是均匀的；此外，它们将跳跃连接建模为恒等路径，从而导致不准确的相关性归属。为了解决这些问题，我们提出了GradSkip，一种基于自适应头加权和跳跃感知传播的ViTs新型相关传播方法。GradSkip建模了注意力头的重要性差异，并动态地在注意力路径和残差路径之间分配相关性。在ImageNet1K和BloodMNIST上的实验表明，GradSkip在保持最新的可信度的同时，所需的GFLOPs比现有最佳方法少14倍以上。使用基于变换器的分割进行的额外评估确认了定位和与真实区域的对齐得到了改善。
cs.CV / 54 / 2607.10370
Neural Motion Blending Across Arbitrary Character Topologies
跨任意角色拓扑的神经运动混合
Abstract
Motion blending in character animation enables the synthesis of new motions by interpolating between existing examples. Current methods are typically restricted to fixed skeleton topologies, requiring identical or near-identical skeletal structures across characters. We present a novel framework for motion blending across heterogeneous skeletons. The proposed architecture combines a semantic encoder, which extracts per-frame latent representations of the motion state, with a diffusion-based decoder, which reconstructs character-specific motion conditioned on this latent code. At inference, blended motions are obtained by interpolating the latent representations of two input motions. We train and evaluate the method on the Truebones Zoo dataset using motions defined on both same and distinct skeleton topologies, demonstrating the ability to achieve smooth and plausible blending in a variety of scenarios.
Chinese Translation
角色动画中的运动混合通过在现有示例之间插值来合成新运动。当前的方法通常限制于固定的骨骼拓扑，要求角色之间具有相同或近似相同的骨骼结构。我们提出了一种新的异构骨骼运动混合框架。所提出的架构结合了一个语义编码器，该编码器提取每帧运动状态的潜在表示，以及一个基于扩散的解码器，该解码器根据该潜在代码重构特定于角色的运动。在推理时，通过插值两个输入运动的潜在表示来获得混合运动。我们在Truebones Zoo数据集上训练和评估该方法，使用在相同和不同骨骼拓扑上定义的运动，展示了在各种场景中实现平滑和合理混合的能力。
cs.CV / 55 / 2607.10383
ABot-N1: Toward a General Visual Language Navigation Foundation Model
ABot-N1：迈向通用视觉语言导航基础模型
Gong, Ruiyan, Guo, Yingnan, Hu, Junjun, Kong, Jintao, Leng, Xiaoxu, Li, Tianlun, Li, Weize, Liu, Fei, Liu, Zhicheng, Lu, Jia, Luo, Minghua, Ming, Chenlin, Shen, Yanfen, Tao, Jiyue, Wang, Zhengbo, Yin, Mingyang, Gu, Minqi, Guan, Zihao, Guo, Wei, Liu, Guoqing, Pang, Huachong, Yang, Menglin, Ye, Zeqian, Geng, Xiaoxiao, Gu, Zhining, Han, Honglin, Jing, Di, Pan, Hongyu, Sun, Mingchao, Yang, Kuan, Zhang, Jianfang, Chen, Yanghong, He, Ye, Mei, Wei, Shi, Jiahao, Yang, Xiangpo, Zhu, Yanqing, Chu, Zedong, Wu, Xiaolong, Xu, Mu
Abstract
Visual Language Navigation foundation models aim to unify deep reasoning for grounded spatial decisions with broad versatility for diverse embodied tasks. Current approaches typically achieve this integration via monolithic policies that map observations directly to actions, yet they often suffer from coordinate drift and poor handling of long-tail semantics. Furthermore, these black-box mappings lack interpretability, hindering the simultaneous achievement of generality, robustness, and transparency. We present ABot-N1, a step toward a general Visual Language Navigation foundation model, that addresses these challenges by decoupling cognition from control via a slow-fast architecture guided by dual visual-language signals. More specifically, a slow vision-language reasoner performs explicit Chain-of-Thought reasoning while producing a pixel goal. This compact set of image-space anchor points serves as a universal interface for diverse tasks, including point-goal, object-goal, poi-goal, instruction-following, and person-following. Subsequently, a fast action expert leverages both the textual cues and the pixel guidance to generate continuous waypoints at the native control frequency. By bridging high-level intents and low-level control through pixel-grounded anchors paired with explicit linguistic traces, our approach ensures robust, generalizable, and interpretable navigation across simulation and real-world benchmarks. ABot-N1 establishes new state-of-the-art records, delivering massive gains specifically in urban-scale navigation: boosting POI arrival by 35.0% (to 77.3%) and achieving 95.4%/92.9% SR in complex indoor and outdoor scenes. It also maintains superior robustness across object-reaching, person-following, and instruction-following tasks. New Point-Goal/POI-Goal benchmarks are released as open source to advance the field of urban-scale navigation.
Chinese Translation
视觉语言导航基础模型旨在将深度推理与广泛的多样化具身任务的空间决策统一起来。目前的方法通常通过单一策略直接将观察映射到行动来实现这一整合，但它们往往面临坐标漂移和对长尾语义处理不佳的问题。此外，这些黑箱映射缺乏可解释性，妨碍了通用性、鲁棒性和透明性的同时实现。我们提出了ABot-N1，这是迈向通用视觉语言导航基础模型的一步，旨在通过一种由双重视觉-语言信号引导的慢-快架构，将认知与控制解耦，从而应对这些挑战。更具体地说，一个慢速视觉-语言推理器在生成像素目标的同时执行显式的思维链推理。这组紧凑的图像空间锚点作为多样化任务的通用接口，包括点目标、物体目标、兴趣点目标、指令跟随和人跟随。随后，一个快速行动专家利用文本线索和像素引导，以本地控制频率生成连续的航点。通过将高层意图与低层控制通过像素锚点和显式语言痕迹相结合，我们的方法确保了在模拟和现实世界基准测试中具有鲁棒性、可推广性和可解释性的导航。ABot-N1建立了新的最先进记录，在城市规模导航中实现了巨大的提升：兴趣点到达率提高了35.0%（达到77.3%），并在复杂的室内和室外场景中实现了95.4%/92.9%的成功率。它还在物体到达、人跟随和指令跟随任务中保持了优越的鲁棒性。新的点目标/兴趣点目标基准已作为开源发布，以推动城市规模导航领域的发展。
cs.CV / 56 / 2607.10391
Vertical Fusion: Condensing Internal Representations for Robust ViT Classification
垂直融合：凝聚内部表征以增强ViT分类的鲁棒性
Abstract
Despite exposing rich intermediate representations, Vision Transformers (ViTs) are almost exclusively utilized as black-box feature extractors, where only the last layer is considered for downstream tasks. We challenge this convention by introducing the notion of recoverability: the capacity of intermediate representations to correct last-layer failures. By evaluating independent classification probes at every model depth across 16 datasets, we observe that intermediate probes correctly classify 18% to 76% of samples that the last-layer probe misclassifies. We show that these gains are not primarily driven by predictive diversity, but by a redundancy-correctness correspondence, where the internal hierarchy acts as a series of stable, redundant probes of a shared discriminative signal. While established horizontal ensemble strategies (i.e., across multiple models) can improve performance, they incur high computational cost and ignore this vertical signal within a single model. To bridge this gap, we propose VFusion, a principled vertical aggregation strategy employing a learnable mapping into a low-dimensional latent space that synthesizes features across the internal ViT hierarchy. VFusion substantially outperforms established aggregation baselines in both in-distribution and out-of-distribution settings, notably closing 45% of the accuracy gap between the best individual layer and a theoretical oracle performance. Our gains consistently generalize across model sizes and pre-training regimes, confirming that VFusion offers a robust and efficient alternative to horizontal ensemble methods. The code is available at https://github.com/francescodisalvo05/vit-vertical-fusion.
Chinese Translation
尽管视觉变换器（Vision Transformers, ViTs）暴露了丰富的中间表征，但它们几乎被专门用作黑箱特征提取器，仅考虑最后一层用于下游任务。我们通过引入可恢复性这一概念来挑战这一惯例：中间表征纠正最后一层失败的能力。通过在16个数据集上评估每个模型深度的独立分类探针，我们观察到中间探针正确分类了18%到76%的样本，而这些样本是最后一层探针错误分类的。我们展示这些增益并非主要由预测多样性驱动，而是由冗余-正确性对应关系驱动，其中内部层次结构充当了一系列稳定的、冗余的共享判别信号探针。虽然已有的水平集成策略（即跨多个模型）可以提高性能，但它们会带来高计算成本，并忽视了单个模型内的这种垂直信号。为了解决这一问题，我们提出了VFusion，一种原则性的垂直聚合策略，采用可学习的映射到低维潜在空间，合成跨越内部ViT层次结构的特征。VFusion在分布内和分布外设置中显著优于既有的聚合基线，尤其是缩小了最佳单层与理论oracle性能之间45%的准确率差距。我们的增益在不同模型规模和预训练方案中始终具有良好的泛化性，确认VFusion为水平集成方法提供了一种鲁棒且高效的替代方案。代码可在 https://github.com/francescodisalvo05/vit-vertical-fusion 获取。
cs.CV / 57 / 2607.10395
Self-supervised Automatic Matting
自监督自动抠图
Abstract
High-quality alpha mattes are notoriously expensive to annotate, creating a fundamental data bottleneck for deep image matting. While prior work attempts to reduce annotation cost using coarser labels like trimaps or masks, they remain reliant on costly per-pixel supervision, limiting scalability and generalization. In this work, we push the boundary further and ask: can we train an automatic matting model using only RGB images, with no manual annotation at all? We answer this by presenting SSMatte, a self-supervised framework that for the first time achieves performance on par with fully-supervised automatic matting. Our key insight is to decompose the problem into semantic anchoring and detail matting. SSMatte first generates a semantic matting prompt from frozen self-supervised ViT features by propagating class-token seeds via a novel, training-efficient semantic anchoring loss based on a generalized Rayleigh quotient. This prompt then anchors a detail matting network, which is optimized via a fixed-point-based loss that enforces alpha-RGB consistency. Extensive experiments show SSMatte outperforms prior weakly-supervised methods, matches the performance of fully-supervised models on portrait benchmarks, and demonstrates favorable scaling and generalization behaviors with additional data. Our work pushes automatic matting to an fresh, fully annotation-free paradigm. Code will be available.
Chinese Translation
高质量的 alpha matte 注释 notoriously 昂贵，成为深度图像抠图的基本数据瓶颈。虽然之前的工作尝试通过使用粗略标签（如 trimaps 或 masks）来降低注释成本，但仍然依赖于昂贵的逐像素监督，限制了可扩展性和泛化能力。在本研究中，我们进一步推动了这一界限，提出了一个问题：我们能否仅使用 RGB 图像训练一个自动抠图模型，而完全不需要手动注释？我们通过提出 SSMatte，一个自监督框架，首次实现了与完全监督的自动抠图相当的性能。我们的关键见解是将问题分解为语义锚定和细节抠图。SSMatte 首先通过基于广义 Rayleigh 商的新的、训练高效的语义锚定损失，从冻结的自监督 ViT 特征生成语义抠图提示，并通过传播类别标记种子来实现。该提示随后锚定一个细节抠图网络，该网络通过固定点基础的损失进行优化，以强制执行 alpha-RGB 一致性。大量实验表明，SSMatte 超越了之前的弱监督方法，在人像基准上匹配了完全监督模型的性能，并在附加数据上展示了良好的可扩展性和泛化能力。我们的工作将自动抠图推向一个全新的、完全无注释的范式。代码将会发布。
cs.CV / 58 / 2607.10400
SynthDocBench: Controlled Benchmark for Long-Context Visual Document Understanding
SynthDocBench：用于长文本视觉文档理解的受控基准
Abstract
Vision language models (VLMs) have achieved strong performance on visual document understanding benchmarks such as DocVQA, ChartQA, and MMLongBench-Doc. However, real-world documents combine multiple factors such as length, layout complexity, modality, and question difficulty, which makes it difficult to attribute model failures to specific causes. We introduce SynthDocBench, a fully synthetic benchmark for long-context visual document understanding that systematically controls factors including document length, layout structure, modality composition, and question type. The benchmark is constructed using a combinatorial design, each factor is varied independently across generated documents, enabling controlled analysis of model behavior. Documents are generated end to end using an LLM pipeline across six layout archetypes, with a 40 percent random override to prevent models from exploiting spurious correlations. Additionally, SynthDocBench spans long-context documents with substantially greater length and structural diversity than existing benchmarks. Evaluating seven frontier VLMs, we uncover three failure modes that existing benchmarks cannot surface: sharp degradation with document length, a systematic positional sensitivity in which the middle third of a document is hardest for five of six models and five of six models show a negative Early-to-Late trend (steepest decline: 8.3 percentage points), and breakdown of chart comprehension in long-document settings. These results suggest that current models may be overfitting to benchmark artifacts rather than achieving robust long-context visual document understanding.
Chinese Translation
视觉语言模型（VLMs）在视觉文档理解基准（如 DocVQA、ChartQA 和 MMLongBench-Doc）上取得了良好的表现。然而，现实世界中的文档结合了多种因素，如长度、布局复杂性、模态和问题难度，这使得很难将模型失败归因于特定原因。我们提出了 SynthDocBench，这是一个完全合成的长文本视觉文档理解基准，系统性地控制文档长度、布局结构、模态组成和问题类型等因素。该基准采用组合设计构建，每个因素在生成的文档中独立变化，从而实现对模型行为的受控分析。文档通过 LLM（大语言模型）管道在六种布局原型中端到端生成，并进行 40% 的随机覆盖，以防止模型利用虚假相关性。此外，SynthDocBench 涵盖了比现有基准显著更长且结构多样的长文本文档。通过评估七个前沿 VLM，我们发现了现有基准无法揭示的三种失败模式：文档长度导致的急剧退化、系统性的位置信息敏感性，其中文档中间三分之一对六个模型中的五个模型来说最为困难，以及在长文档环境中图表理解的崩溃。这些结果表明，当前模型可能过拟合于基准伪影，而不是实现稳健的长文本视觉文档理解。
cs.CV / 59 / 2607.10406
TVT-PAPD: Pathology-Aware Prototype Distillation for Self-Supervised Whole Slide Image Classification
TVT-PAPD：面向病理的原型蒸馏用于自监督全幻灯片图像分类
Abstract
Self-supervised learning (SSL) has emerged as an effective paradigm for learning transferable representations from large-scale unlabeled whole slide images (WSIs). However, existing SSL methods primarily learn generic visual features and often fail to explicitly capture pathology-specific morphological patterns that are critical for disease characterization. To address this limitation, we propose Tiny Vision Transformer with Pathology-Aware Prototype Distillation (TVT-PAPD). This self-supervised pathology representation learning framework integrates a Tiny Vision Transformer (TVT) with a novel Pathology-Aware Prototype Distillation (PAPD) module. PAPD employs a learnable pathology prototype bank to discover and preserve representative tissue morphology patterns, encouraging semantically similar pathological regions to learn consistent and discriminative representations. The proposed framework enhances pathology-aware feature learning while maintaining computational efficiency with 90M parameters. Experiments on the Cancer Genome Atlas (TCGA) low-grade glioma (LGG)/glioblastoma (GBM) dataset and the Indian Pathology Brain (IPD-Brain) dataset demonstrate that TVT-PAPD achieves weighted F1-scores of 93.02% and 90.23%, respectively, for LGG-GBM classification, while exhibiting strong cross-cohort generalization across independent glioma datasets.
Chinese Translation
自监督学习（SSL）已成为从大规模未标记全幻灯片图像（WSIs）中学习可转移表示的有效范式。然而，现有的SSL方法主要学习通用视觉特征，往往未能明确捕捉对疾病特征至关重要的病理特定形态模式。为了解决这一局限性，我们提出了带有病理感知原型蒸馏的微型视觉变换器（TVT-PAPD）。该自监督病理表示学习框架将微型视觉变换器（TVT）与一种新颖的病理感知原型蒸馏（PAPD）模块相结合。PAPD利用可学习的病理原型库来发现和保留具有代表性的组织形态模式，鼓励语义相似的病理区域学习一致且具有区分性的表示。该框架在保持计算效率的同时，增强了病理感知特征学习，参数量为9000万。在癌症基因组图谱（TCGA）低级别胶质瘤（LGG）/胶质母细胞瘤（GBM）数据集和印度病理脑（IPD-Brain）数据集上的实验表明，TVT-PAPD在LGG-GBM分类中分别达到了93.02%和90.23%的加权F1分数，同时在独立胶质瘤数据集上表现出强大的跨队列泛化能力。
cs.CV / 60 / 2607.10408
GNOCHI: Generative Neural mOdel for Close Human-Human Interactions
GNOCHI：用于紧密人际互动的生成神经模型
Abstract
Creating realistic 3D human-human interactions in virtual environments is challenging due to the high degrees of freedom in the human body and the need for physically accurate poses that do not collide with each other. Traditional methods for human-human interaction are based on motion tracking or 3D body reconstruction, but lack generative capabilities. Recent generative methods enable the synthesis of individual or interacting motions via text or image input, but generally fall short in modeling close interactions. This paper introduces a novel generative model for close 3D human-human interactions using a conditional variational autoencoder (cVAE), which generates poses for one human conditioned on the pose of another, allowing for controlled and diverse interaction synthesis. To train our model, we address two underlying long-standing challenges in the field of human-human interaction: data scarcity, for which we propose an automated supervised data augmentation strategy that generates synthetic yet realistic interaction poses; and collision awareness in generative approaches, for which we propose a self-supervised loss based on a collision resolution technique using volumetric proxies to ensure physically correct interactions. We extensively evaluate the capabilities of our model, and demonstrate a wide variety of plausible and physically correct interactions, not possible to generate with current state-of-the-art methods.
Chinese Translation
在虚拟环境中创建逼真的三维人际互动面临挑战，这主要是由于人体的高自由度以及需要确保姿势之间不发生碰撞的物理准确性。传统的人际互动方法基于运动跟踪或三维身体重建，但缺乏生成能力。近期的生成方法通过文本或图像输入实现个体或互动动作的合成，但在建模紧密互动方面通常表现不足。本文提出了一种新颖的生成模型，旨在实现紧密的三维人际互动，采用条件变分自编码器（cVAE），该模型根据一个人的姿势生成另一个人的姿势，从而实现可控且多样的互动合成。为了训练我们的模型，我们解决了人际互动领域中长期存在的两个基本挑战：数据稀缺性，我们提出了一种自动化的监督数据增强策略，该策略生成合成但逼真的互动姿势；以及生成方法中的碰撞意识，我们提出了一种基于使用体积代理的碰撞解决技术的自监督损失，以确保物理上正确的互动。我们对模型的能力进行了广泛评估，展示了多种合理且物理上正确的互动，这些互动是当前最先进方法无法生成的。
cs.CV / 61 / 2607.10413
SPORT: Structure-Aware Prototype Disentanglement for Incomplete Multi-View Clustering
SPORT：结构感知原型解耦用于不完整多视图聚类
Abstract
Prototype-based Incomplete Multi-view Clustering has recently attracted increasing attention by exploiting prototypes as semantic anchors for missing-view imputation. However, existing approaches are still limited in three aspects. First, they typically focus on enforcing cross-view prototype consistency, while ignoring view-specific information embedded in prototypes, thus limiting multi-view expressiveness. Second, most methods rely on instance-level contrastive learning that only aligns paired samples across views, failing to preserve cluster-level relational structures. Third, missing-view imputation is usually performed using global prototypes alone, without considering local geometric neighborhood structures, leading to inaccurate recovery of missing representations. To address these limitations, we propose a novel framework termed Structure-aware PrOtotype disentanglement foR incomplete multi-view clusTering (SPORT), which explicitly disentangles shared and view-specific components of prototypes while preserving cluster-level relational structures. Specifically, we decouple prototypes into orthogonal shared and view-specific components, aligning only shared components to capture consensus semantics while de-correlating view-specific components to preserve complementary information. Meanwhile, a structure-aware contrastive learning mechanism is incorporated to explicitly model cluster-level relationships during cross-view representation learning. Furthermore, a hybrid imputation strategy integrates global prototype matching with local neighborhood matching, enabling joint exploitation of semantic prototypes and manifold structures for missing-view recovery. Extensive experiments on six benchmark datasets show that SPORT achieves superior performance over state-of-the-art methods under various missing rates.
Chinese Translation
基于原型的不完整多视图聚类最近通过利用原型作为缺失视图插补的语义锚点而受到越来越多的关注。然而，现有方法在三个方面仍然存在局限性。首先，它们通常专注于强制跨视图原型一致性，而忽视了嵌入原型中的视图特定信息，从而限制了多视图的表现力。其次，大多数方法依赖于实例级对比学习，仅对齐跨视图的配对样本，未能保留聚类级关系结构。第三，缺失视图插补通常仅使用全局原型进行，而未考虑局部几何邻域结构，导致缺失表示的恢复不准确。为了解决这些局限性，我们提出了一种新颖的框架，称为结构感知原型解耦用于不完整多视图聚类（SPORT），该框架明确解耦原型的共享和视图特定组件，同时保留聚类级关系结构。具体而言，我们将原型解耦为正交的共享和视图特定组件，仅对齐共享组件以捕捉共识语义，同时去相关化视图特定组件以保留互补信息。同时，结合了一种结构感知对比学习机制，以显式建模跨视图表示学习中的聚类级关系。此外，一种混合插补策略将全局原型匹配与局部邻域匹配相结合，使得在缺失视图恢复中能够共同利用语义原型和流形结构。在六个基准数据集上的大量实验表明，SPORT在各种缺失率下的性能优于最先进的方法。
cs.CV / 62 / 2607.10427
BOCCHI: A More Realistic and Challenging Benchmark for Local Motion Blur Detection with MSDCT-UNet
BOCCHI：一个更现实且具有挑战性的局部运动模糊检测基准，结合MSDCT-UNet
Abstract
Local motion blur detection requires pixel-level localization of blurred regions. Existing benchmarks let models rely on gradient shortcuts that fail to transfer. We introduce BOCCHI (Blurred Objects Captured across Cameras with Human-annotated Imagery), a real-captured benchmark whose sharp regions overlap the blur gradient distribution and defeat these shortcuts, and propose MSDCT-UNet (Multi-Scale Discrete Cosine Transform UNet), a frequency-aware encoder-decoder injecting multi-scale DCT priors through DCT Attention and FiLM. MSDCT-UNet ranks first in in-domain mIoU and boundary localization on BOCCHI, and BOCCHI-trained models outperform every other training source on cross-dataset transfer with only 633 training images.
Chinese Translation
局部运动模糊检测需要对模糊区域进行像素级定位。现有基准使模型依赖于无法转移的梯度捷径。我们引入BOCCHI（Blurred Objects Captured across Cameras with Human-annotated Imagery），这是一个真实捕获的基准，其清晰区域与模糊梯度分布重叠，从而克服了这些捷径，并提出了MSDCT-UNet（Multi-Scale Discrete Cosine Transform UNet），这是一种频率感知的编码器-解码器，通过DCT注意力和FiLM注入多尺度DCT先验。MSDCT-UNet在BOCCHI的领域内mIoU和边界定位中排名第一，而经过BOCCHI训练的模型在跨数据集转移中优于其他所有训练源，仅使用633张训练图像。
cs.CV / 63 / 2607.10461
Annotation-Free Furniture Codes: What They Encode, and How Far They Transfer
无注释家具编码：它们编码了什么，以及它们的迁移能力有多远
Abstract
Layout-based 3D scene synthesizers place each object using two human-annotated channels: a categorical class label and a canonical-pose convention. We ask whether a single self-supervised token derived from object geometry can replace both, and study such tokens directly as a representation, decoupled from any synthesizer. A Finite Scalar Quantization (FSQ) point-cloud autoencoder is chamfer-trained on placed 3D-FUTURE furniture with no labels or pose annotations. Diagnostic probes recover fine-category (62.6 +/- 0.5%), super-category (85.6 +/- 1.3%), and yaw (52.7 +/- 0.5 deg) from the codes alone. Swapping the chamfer target from the rotated to the un-rotated point cloud collapses the yaw signal while raising class recovery, showing the codes' rotation content can be set by the training objective. Scaling across asset libraries needs codes that transfer; on an unseen dataset (ShapeNet), alignment is category-dependent: box-like furniture transfers, organically-shaped furniture does not, and a target-blind augmentation partly closes the gap.
Chinese Translation
基于布局的 3D 场景合成器使用两个人工注释通道来放置每个物体：一个类别标签和一个标准姿态约定。我们探讨是否可以用一个从物体几何形状派生的自监督标记替代这两个通道，并直接研究这些标记作为一种表示，脱离任何合成器。一个有限标量量化（Finite Scalar Quantization, FSQ）点云自编码器在没有标签或姿态注释的情况下，对放置的 3D-FUTURE 家具进行切割训练。诊断探针仅通过编码恢复细分类别（62.6 +/- 0.5%）、超类别（85.6 +/- 1.3%）和偏航（52.7 +/- 0.5 度）。将切割目标从旋转点云切换到未旋转点云时，偏航信号消失，而类别恢复提高，表明编码的旋转内容可以由训练目标设定。在资产库之间进行扩展需要能够迁移的编码；在一个未见数据集（ShapeNet）上，类别对齐是依赖于类别的：盒状家具可以迁移，而有机形状家具则不能，目标盲增广部分缩小了差距。
cs.CV / 64 / 2607.10465
Not All Color Categories Are Equally Stable: A Multilingual Free Color Naming Experiment
并非所有颜色类别的稳定性相同：一项多语言自由颜色命名实验
Abstract
Color naming is an important part of human color perception. Its task is to allow people to describe continuous colors using discrete color categories. However, the boundaries between color categories are often unclear, and some colors may be perceived differently depending on their saturation and brightness. While certain color categories remain recognizable across a wide range of shades, others may be associated with different color names when their appearance changes. This study investigates the consistency of color naming for red, yellow, and green color categories using a free color-naming experiment. A set of 18 color samples was selected from the COLIBRI dataset to represent different shades of these colors. Participants (n = 92) were asked to freely assign color names to each sample in Kazakh, Russian, or English without being limited to predefined categories. The results show that color categories differ in their consistency. Green shades were consistently identified as green despite variations in appearance, whereas yellow shades received a wider variety of names, including gold- and brown-related descriptions. Red shades showed moderate naming consistency. Our findings suggest that some color categories occupy broader perceptual regions than others and may therefore be more robust to visual variations. The study results can be used to develop perceptually meaningful color models and color naming systems.
Chinese Translation
颜色命名是人类颜色感知的重要组成部分。其任务是使人们能够使用离散的颜色类别来描述连续的颜色。然而，颜色类别之间的界限往往不清晰，某些颜色的感知可能会因其饱和度和亮度而有所不同。尽管某些颜色类别在广泛的色调范围内仍然可识别，但当其外观变化时，其他颜色类别可能会与不同的颜色名称相关联。本研究通过自由颜色命名实验，调查了红色、黄色和绿色颜色类别的命名一致性。从COLIBRI数据集中选择了一组18个颜色样本，以代表这些颜色的不同色调。参与者（n = 92）被要求在没有预定义类别限制的情况下，自由地为每个样本分配颜色名称，使用哈萨克语、俄语或英语。结果表明，颜色类别在一致性上存在差异。尽管外观有所变化，绿色色调始终被一致地识别为绿色，而黄色色调则获得了更广泛的名称，包括与金色和棕色相关的描述。红色色调显示出中等的命名一致性。我们的研究结果表明，某些颜色类别占据的感知区域比其他类别更广泛，因此可能对视觉变化更具鲁棒性。研究结果可用于开发具有感知意义的颜色模型和颜色命名系统。
cs.CV / 65 / 2607.10470
On the Real-World Generalisability of Optical Flow Models
光流模型的现实世界泛化能力研究
Abstract
Real-world deployment of vision models to broadly benefit society is arguably a main research objective. In optical flow, however, the difficulty to obtain the ground truth has focused research mainly on synthetic data and domain-specific benchmarks. Here, we investigate the severity of this mismatch. We study how well modern optical flow estimation models generalise to real-world video and question if accuracy on synthetic benchmark proxies actually predicts accuracy on real-world optical flow. To address this, we build a real-world evaluation benchmark and evaluate the real-world generalisability of a broad set of recent optical flow models using standard checkpoints. Our benchmark contains 8,204 frame pairs across TAP-Flow, Slow Flow, and our own dataset FlowFactor. FlowFactor is a manually annotated real-world benchmark of 1,000 HD frame pairs organised into four confounding factors: large displacements, repetitive textures, occlusions, and lighting variation. Each setting mainly varies only one factor, enabling diagnostic, confounder-specific analysis. Using FlowFactor, we reveal that performance on varying lighting and large displacements correlates most strongly with real-world accuracy, and that improvements on large-motion regimes can trade off against robustness in small-motion, stationary scenes. Our experiments show that progress on Sintel, KITTI and Spring only weakly predicts accuracy on real-world data, highlighting the need for a broad real-world optical flow benchmark. Interestingly, scaling up the amount of training data does not necessarily resolve the gap, calling for new innovative research instead of simply scaling data and compute.
Chinese Translation
将视觉模型应用于现实世界以广泛造福社会无疑是一个主要的研究目标。然而，在光流领域，由于获取真实标签的困难，研究主要集中在合成数据和特定领域的基准测试上。在此，我们调查了这种不匹配的严重性。我们研究了现代光流估计模型在现实世界视频中的泛化能力，并质疑合成基准代理上的准确性是否真的能预测现实世界光流的准确性。为了解决这个问题，我们构建了一个现实世界评估基准，并使用标准检查点评估了一系列最新光流模型的现实世界泛化能力。我们的基准包含来自 TAP-Flow、Slow Flow 和我们自己的数据集 FlowFactor 的 8,204 对帧。FlowFactor 是一个手动标注的现实世界基准，包含 1,000 对高清帧，组织成四个混淆因素：大位移、重复纹理、遮挡和光照变化。每个设置主要只变化一个因素，从而实现特定混淆因素的诊断分析。使用 FlowFactor，我们揭示了在不同光照和大位移下的表现与现实世界准确性之间的强相关性，并且在大运动场景中的改进可能会与小运动、静态场景的鲁棒性相互权衡。我们的实验表明，在 Sintel、KITTI 和 Spring 上的进展仅能弱预测现实世界数据的准确性，突显了广泛的现实世界光流基准的必要性。有趣的是，增加训练数据的数量并不一定能解决这一差距，这呼唤新的创新研究，而不仅仅是扩大数据和计算资源。
cs.CV / 66 / 2607.10489
Grassmannian Splatting I: Moving rank-2 Spacetime Surfels for Dynamic Scene Rendering
Abstract
We introduce Grassmannian splatting, a dynamic scene representation whose primitives are Gaussians supported on 3-planes in spacetime $\R^4$: generically, spatial 2-planes in uniform translation along their normals. Each primitive carries a unit normal $n \in \mathbb S^3/\{\pm 1\} \cong \mathrm{Gr}(3,4)$ and an unconstrained factor $L \in \mathbb R^{4 \times 3}$, with covariance \[ \Sigma_{4\mathrm{D}} = (P_n L)(P_n L)^T, \qquad P_n = I - n n^T. \] For generic $L$ and $n \neq \pm e_0$, conditioning on time returns a rank-2 surfel at every frame. The normal of the disk and its velocity along that normal are read off from $n$; the disk shape and the tangential drift of its center are set by $L$. Existing native 4D Gaussian splatting methods [\it{Yang et. al. 2023,Duan et. al. 2024}] slice full-rank spacetime covariances, so their per-frame primitive is a volumetric ellipsoid; since conditioning lowers rank by exactly one, a rank-2 surfel in the slice requires a rank-3 spacetime covariance, and the parameterization above realizes exactly these. The motion model is closed form, i.e. no deformation field is learned, and no custom CUDA is required: the conditioned disk feeds a standard 3DGS rasterizer through its precomputed-covariance interface. A soft clamp in the Schur denominator regularizes the static orientation and continuously bridges rank-3 static and rank-2 dynamic behavior, so static and moving primitives form a single continuous family. On the 17 HyperNeRF scenes of MonoDyGauBench, training is fastest among all compared methods (4.9 to 5.6 times faster than the strongest quality baselines), while ranking second in PSNR, MS-SSIM, and LPIPS. Code: https://github.com/PaulCelanCoding/grassmannian-splatting
cs.CV / 67 / 2607.10495
NanoVSR: Towards Real-Time Video Super-Resolution on Edge Devices
NanoVSR：面向边缘设备的实时视频超分辨率
Abstract
Recent Video Super-Resolution (VSR) methods rely heavily on transformers and explicit optical flow, creating computational overhead and custom operations that hinder deployment on hardware accelerators like TensorRT. To address this, we introduce NanoVSR, a scalable, fully convolutional architecture designed for resource-constrained edge devices. Using structural reparameterization, NanoVSR collapses into standard convolutions during inference, ensuring seamless hardware compatibility and negligible runtime overhead. Furthermore, despite lacking explicit motion compensation, it maintains competitive restoration quality by implicitly learning spatio-temporal alignments through progressive training. Evaluated on the REDS4 benchmark, NanoVSR demonstrates an exceptional balance between accuracy and computational efficiency, significantly improving the trade-off for compact architectures. Our NanoVSR-644k baseline yields 28.64 dB PSNR while delivering 27.2 FPS on the NVIDIA Jetson Orin NX 16GB (25W), offering massive speed gains over heavier models. The scaled NanoVSR-1.7M variant reaches 29.15 dB with a throughput of 19.58 FPS, providing superior, edge-optimized upscaling. Code is available at https://github.com/filippawlicki/nanovsr.
Chinese Translation
近期的视频超分辨率（VSR）方法在很大程度上依赖于变换器和显式光流，这造成了计算开销和定制操作，阻碍了在如TensorRT等硬件加速器上的部署。为了解决这个问题，我们提出了NanoVSR，一种为资源受限的边缘设备设计的可扩展全卷积架构。通过结构重参数化，NanoVSR在推理过程中压缩为标准卷积，确保无缝的硬件兼容性和微不足道的运行时开销。此外，尽管缺乏显式运动补偿，它通过渐进训练隐式学习时空对齐，保持了竞争力的恢复质量。在REDS4基准测试中评估，NanoVSR展示了准确性与计算效率之间的卓越平衡，显著改善了紧凑架构的权衡。我们的NanoVSR-644k基线在NVIDIA Jetson Orin NX 16GB（25W）上实现了28.64 dB的PSNR，同时以27.2 FPS的速度运行，相较于更重的模型提供了巨大的速度提升。扩展后的NanoVSR-1.7M变体达到29.15 dB的PSNR，吞吐量为19.58 FPS，提供了优越的边缘优化上采样。代码可在https://github.com/filippawlicki/nanovsr获取。
cs.CV / 68 / 2607.10522
Towards Autonomous and Auditable Medical Imaging Model Development
迈向自主可审计的医学影像模型开发
Abstract
Large language model (LLM) agents are beginning to automate machine learning engineering (MLE) by coupling planning, code execution, debugging, and empirical feedback. Translating this capability to medical imaging remains difficult because each task imposes modality-specific experimentation and strict requirements for validation protocols and prediction artifacts. Here we introduce AMID, an autonomous multi-agent framework for medical imaging model development. AMID first proposes Data-Conditioned Method Planning, which refines coarse task-level search spaces into executable, parallelizable method lanes grounded in task-specific data analysis and runnable medical-imaging resources. It then develops Verification-Guided Two-Stage Optimization, moving from broad early exploration of diverse method lanes to selective exploitation of promising candidates while enforcing strict verification of validation protocols, metric computation, and prediction artifacts throughout the optimization. Across 20 medical imaging challenge tasks spanning diverse modalities and prediction types, AMID outperformed evaluated general-purpose MLE systems and, on several tasks, approached or matched strong human-designed challenge solutions. These results suggest that AMID can turn task-specific medical imaging model development from bespoke manual engineering into an agentic workflow for producing high-performing and auditable model artifacts across heterogeneous tasks.
Chinese Translation
大型语言模型（LLM）代理开始通过结合规划、代码执行、调试和经验反馈来自动化机器学习工程（MLE）。将这一能力转化为医学影像领域仍然困难，因为每个任务都需要特定于模态的实验和严格的验证协议及预测工件要求。在此，我们介绍了AMID，一个用于医学影像模型开发的自主多代理框架。AMID首先提出了数据条件的方法规划，将粗略的任务级搜索空间细化为可执行、可并行化的方法通道，这些通道基于特定任务的数据分析和可运行的医学影像资源。接着，它开发了验证引导的两阶段优化，从广泛的早期多样化方法通道探索转向对有前景候选的选择性利用，同时在整个优化过程中严格执行验证协议、指标计算和预测工件的验证。在涵盖多种模态和预测类型的20个医学影像挑战任务中，AMID的表现超越了评估的通用MLE系统，并且在多个任务上接近或匹配了强大的人工设计挑战解决方案。这些结果表明，AMID能够将特定任务的医学影像模型开发从定制的手动工程转变为一种代理化工作流程，以生成高性能和可审计的模型工件，适用于异构任务。
cs.CV / 69 / 2607.10535
Improving Sample Diversity in Autoregressive Text-to-Image Generation via Cluster Truncation
通过聚类截断提高自回归文本到图像生成中的样本多样性
Abstract
While diffusion models achieve state-of-the-art image quality for text-to-image (T2I) generation, recent work has demonstrated that they suffer from sample diversity collapse. In this work, we investigate whether autoregressive (AR) image generation models can push the Pareto frontier between image quality and sample diversity. With recent advances in quality and efficiency, AR models have emerged as a viable alternative to diffusion-based image generation. Beyond enabling new use cases such as interleaved image-text generation, their sequential generation process makes them compatible with a wide range of token-based decoding strategies originally developed to improve diversity in text generation. Motivated by the potential of a better diversity-quality tradeoff in the AR paradigm, we present the first systematic study of sample diversity in AR image generation models. We show that two key properties of AR image generation, persistently high token-level entropy and substantial redundancy in visual token spaces, limit the effectiveness of existing token-level decoding methods for diversity enhancement. We therefore propose $p$-less cluster, a new decoding strategy that performs entropy-based truncation sampling at cluster level rather than at token level. We evaluate our approach and baseline decoding methods across four autoregressive T2I models and two datasets using a comprehensive suite of metrics spanning image quality, prompt alignment, and diversity. Our results show that $p$-less cluster unlocks the greatest diversity across most evaluated autoregressive T2I models and datasets while maintaining image quality and prompt alignment.
Chinese Translation
尽管扩散模型在文本到图像（T2I）生成中实现了最先进的图像质量，但近期研究表明它们存在样本多样性崩溃的问题。在本研究中，我们探讨自回归（AR）图像生成模型是否能够推动图像质量与样本多样性之间的帕累托前沿。随着质量和效率的最新进展，AR模型已成为扩散基础图像生成的可行替代方案。除了能够启用新的用例，如交错的图像-文本生成外，它们的顺序生成过程使其与最初为提高文本生成多样性而开发的各种基于标记的解码策略兼容。受到AR范式中更好多样性-质量权衡潜力的启发，我们首次系统性地研究了AR图像生成模型中的样本多样性。我们表明，AR图像生成的两个关键特性，即持续较高的标记级熵和视觉标记空间中的显著冗余，限制了现有标记级解码方法在多样性增强方面的有效性。因此，我们提出了$p$-less cluster，一种新的解码策略，它在聚类级别而非标记级别执行基于熵的截断采样。我们在四个自回归T2I模型和两个数据集上评估了我们的方法和基线解码方法，使用了一套全面的指标，涵盖图像质量、提示对齐和多样性。我们的结果表明，$p$-less cluster在大多数评估的自回归T2I模型和数据集中解锁了最大的多样性，同时保持了图像质量和提示对齐。
cs.CV / 70 / 2607.10544
Physics-inspired Pseudo Anomaly Generation and Prototype Feature Guidance for 3D Anomaly Detection
基于物理启发的伪异常生成与原型特征引导的三维异常检测
Abstract
3D point cloud anomaly detection plays a vital role in industrial manufacturing, yet it faces significant challenges due to the scarcity and high acquisition cost of real anomalous samples. The inherently anomaly-free training data further hinders detection methods from effectively learning discriminative features between normal and abnormal instances. To address these issues, we propose PA3AD, a novel framework that introduces a physics-inspired pseudo-anomaly generation strategy to create physically plausible anomalous samples from normal data. Additionally, we incorporate prototype features via a weight-sharing mechanism to guide the model in capturing the distribution shifts between normal and anomalous samples. Specifically, PA3AD introduces two key innovations to tackle the scarcity of real anomalies. First, a physics-inspired module generates diverse pseudo-anomalous point clouds from normal data via multi-physics modeling. Second, momentum-updated prototypes and a difference-aware fusion block capture stable normal representations and their discrepancies with pseudo-anomalies. This design effectively learns distribution shifts, achieving superior detection performance. Extensive experiments on the Anomaly-ShapeNet and Real3D-AD datasets demonstrate that our method consistently outperforms existing state-of-the-art approaches. Our code will be made publicly available at https://github.com/NingxiaoJian/PA3AD.
Chinese Translation
三维点云异常检测在工业制造中扮演着重要角色，但由于真实异常样本的稀缺和高获取成本，它面临着重大挑战。固有的无异常训练数据进一步阻碍了检测方法有效学习正常与异常实例之间的判别特征。为了解决这些问题，我们提出了PA3AD，一个新颖的框架，采用基于物理启发的伪异常生成策略，从正常数据中创建物理上合理的异常样本。此外，我们通过权重共享机制引入原型特征，以指导模型捕捉正常样本与异常样本之间的分布变化。具体而言，PA3AD引入了两个关键创新来应对真实异常的稀缺性。首先，一个基于物理启发的模块通过多物理建模从正常数据中生成多样化的伪异常点云。其次，动量更新的原型和差异感知融合块捕捉稳定的正常表示及其与伪异常的差异。该设计有效地学习了分布变化，实现了卓越的检测性能。在Anomaly-ShapeNet和Real3D-AD数据集上的大量实验表明，我们的方法始终优于现有的最先进方法。我们的代码将公开发布在 https://github.com/NingxiaoJian/PA3AD。
cs.CV / 71 / 2607.10566
Quantum Compressed Sensing CT Reconstruction Algorithm Based on Penalized Weighted Least Squares and Guided Total Variation
基于惩罚加权最小二乘法和引导全变差的量子压缩感知CT重建算法
Abstract
Objective. Existing quadratic unconstrained binary optimization (QUBO)-based sparse-view computed tomography (CT) reconstruction neglects photon-counting statistics and anatomical heterogeneity. We address both limitations within the QUBO framework.Approach. We propose a quantum compressed-sensing CT method combining penalized weighted least squares (PWLS) and guided total variation (GTV). PWLS weights projection residuals by photon-count reliability, whereas GTV uses gradients from a prior image reconstructed by the simultaneous algebraic reconstruction technique (SART) to preserve edges and suppress noise in homogeneous regions. After binary encoding, both terms form a unified QUBO model. Experiments used four 40 times 40 CT images under a 10-view fan-beam geometry with Poisson noise. Comparisons included conventional reconstruction methods, QUBO variants, gradient descent, simulated annealing, and a D-Wave hybrid quantum-classical solver.Main results. PWLS-GTV achieved the best reconstruction quality across all cases. In the representative chest case, it reached a peak signal-to-noise ratio (PSNR) of 36.64 dB, compared with 22.48 dB for SART, the best conventional baseline. GTV consistently outperformed conventional total variation. Simulated annealing and the D-Wave hybrid solver produced similar reconstructions, whereas gradient descent was ineffective. Repeated hybrid-solver runs showed stable performance.Significance. The framework incorporates photon-statistical weighting and structure-guided regularization into QUBO-based CT reconstruction without changing its quadratic form, providing a proof of concept for quantum-assisted sparse-view CT reconstruction.
Chinese Translation
目的：现有基于二次无约束二进制优化（QUBO）的稀疏视图计算机断层扫描（CT）重建忽视了光子计数统计和解剖异质性。我们在QUBO框架内解决这两种局限性。方法：我们提出了一种结合惩罚加权最小二乘法（PWLS）和引导全变差（GTV）的量子压缩感知CT方法。PWLS通过光子计数可靠性加权投影残差，而GTV利用通过同步代数重建技术（SART）重建的先验图像的梯度来保持边缘并抑制均匀区域的噪声。在二进制编码后，这两个项形成一个统一的QUBO模型。实验使用了四个40×40的CT图像，在10视图扇形束几何下添加了泊松噪声。比较包括传统重建方法、QUBO变体、梯度下降、模拟退火和D-Wave混合量子-经典求解器。主要结果：PWLS-GTV在所有案例中实现了最佳重建质量。在代表性的胸部案例中，其峰值信噪比（PSNR）达到36.64 dB，而SART的最佳传统基线为22.48 dB。GTV始终优于传统全变差。模拟退火和D-Wave混合求解器产生了类似的重建结果，而梯度下降效果不佳。重复的混合求解器运行显示出稳定的性能。意义：该框架将光子统计加权和结构引导正则化纳入QUBO基础的CT重建中，而不改变其二次形式，为量子辅助的稀疏视图CT重建提供了概念验证。
cs.CV / 72 / 2607.10575
Why Domain Matters: Domain-Aware Benchmarking of Underwater Object Detection and Annotation Quality
领域的重要性：水下目标检测与标注质量的领域感知基准测试
Abstract
Underwater object detection is strongly affected by domain shift, where performance can vary significantly across different locations, habitats, and deployment conditions. However, detector performance is typically evaluated using aggregate metrics that hide failures in specific environments, while existing domain generalization benchmarks often rely on synthetic variations that do not reflect real-world conditions. We introduce a framework that characterizes underwater images by appearance, scene composition, and acquisition geometry to assign domain labels. Using this framework, we perform the first systematic study of how domain factors influence both human annotation quality in underwater object detection datasets and deep learning-based detector performance, revealing substantial domain-dependent discrepancies. By incorporating physically meaningful domain labels, domain shift becomes something we can characterize, measure, benchmark, and act on. We highlight how this can be used to guide data collection and annotation, design more informative benchmarks, and assess detector robustness across diverse underwater environments.
Chinese Translation
水下目标检测受到领域转移的强烈影响，不同地点、栖息地和部署条件下的性能可能会显著变化。然而，检测器性能通常使用聚合指标进行评估，这掩盖了特定环境中的失败，而现有的领域泛化基准往往依赖于不反映真实世界条件的合成变体。我们提出了一个框架，通过外观、场景组成和获取几何特征来表征水下图像，以分配领域标签。利用该框架，我们首次系统地研究了领域因素如何影响水下目标检测数据集中的人工标注质量和基于深度学习的检测器性能，揭示了显著的领域依赖性差异。通过引入物理上有意义的领域标签，领域转移成为我们可以表征、测量、基准测试和采取行动的对象。我们强调这可以用于指导数据收集和标注，设计更具信息性的基准，并评估检测器在多样化水下环境中的鲁棒性。
cs.CV / 73 / 2607.10580
DiffUE: Enhancing Utility-Unlearnability Trade-off of Unlearnable Examples via Diffusion Autoencoders
DiffUE：通过扩散自编码器增强不可学习示例的效用-不可学习性权衡
Abstract
AI models are increasingly trained on personal images scraped from social media and public platforms, often without consent, leading to serious privacy violations, such as unauthorized facial recognition and targeted advertising. To counter this, researchers have developed unlearnable examples (UEs), images modified with imperceptible noise to prevent AI models from extracting meaningful information. However, existing UE methods primarily rely on pixel-space noise, which can be bypassed by relearning strategies such as adversarial training, image transformation, and compression. While some techniques improve robustness, they often come at the expense of significant degradation in image utility and perceptual quality. In this paper, we introduce DiffUE to overcome these limitations by injecting noise into the semantic space of images instead of the pixel space. Instead of corrupting pixel values, DiffUE modifies high-level semantic features of images, ensuring robust unlearnability while preserving visual quality and utility. By leveraging a diffusion-based autoencoder framework to manipulate semantic features, DiffUE generates purposeful, natural-looking modifications that effectively resist advanced relearning strategies. Extensive experiments on four datasets, CIFAR-10, CIFAR-100, CelebA-HQ, and ImageNet, as well as a subjective user study, demonstrate that DiffUE significantly enhances the trade-off between image quality and unlearnability, offering a more robust and effective solution for safeguarding personal data in an increasingly exploitative AI landscape.
Chinese Translation
人工智能模型越来越多地在社交媒体和公共平台上抓取的个人图像上进行训练，通常未经过同意，这导致了严重的隐私侵犯，例如未经授权的人脸识别和定向广告。为了解决这个问题，研究人员开发了不可学习示例（UEs），这些图像通过不可察觉的噪声进行修改，以防止人工智能模型提取有意义的信息。然而，现有的UE方法主要依赖于像素空间噪声，这可以通过对抗训练、图像变换和压缩等重学习策略绕过。虽然一些技术提高了鲁棒性，但通常以显著降低图像效用和感知质量为代价。在本文中，我们提出了DiffUE，以通过将噪声注入图像的语义空间而不是像素空间来克服这些限制。DiffUE不是破坏像素值，而是修改图像的高层语义特征，从而确保鲁棒的不可学习性，同时保持视觉质量和效用。通过利用基于扩散的自编码器框架来操控语义特征，DiffUE生成有目的的、自然的修改，能够有效抵抗先进的重学习策略。在四个数据集（CIFAR-10、CIFAR-100、CelebA-HQ和ImageNet）上的大量实验以及一项主观用户研究表明，DiffUE显著增强了图像质量和不可学习性之间的权衡，为在日益剥削性的人工智能环境中保护个人数据提供了更鲁棒和有效的解决方案。
cs.CV / 74 / 2607.10583
Benchmarking UAV-based Vehicle Re-Identification under Simulated Weather Conditions
在模拟天气条件下基于无人机的车辆重识别基准测试
Abstract
UAV-based vehicle re-identification (ReID) has emerged as a promising technique for traffic surveillance, urban monitoring, and public-safety applications thanks to the flexible viewpoints and wide-area coverage provided by unmanned aerial vehicles. However, despite recent progress on UAV-based vehicle ReID benchmarks, the robustness of existing methods under adverse weather remains insufficiently studied. This is important because weather degradation can significantly affect the fine-grained appearance cues required for reliable vehicle matching in aerial imagery, especially under small object scale, viewpoint variation, and complex backgrounds. In this paper, we present a controlled comparative study of three representative recent vehicle ReID methods, namely CLIP-ReID, MSINet, and AdaSP, on two UAV-based benchmarks, VRU and UAV-VeID. To ensure consistent robustness evaluation, we generate synthetic foggy and rainy variants of both datasets using an analytical weather-effect pipeline while preserving the original identities and data splits. All methods are then trained and evaluated under matched clean, foggy, and rainy conditions. Experimental results show that adverse weather consistently degrades retrieval performance across both datasets, with rain causing larger drops than fog in nearly all settings. Among the evaluated methods, AdaSP demonstrates the strongest robustness, achieving 93.0% and 88.5% mAP on VRU-Large, and 88.7% and 76.2% mAP on UAV-VeID-Test under foggy and rainy conditions, respectively. Overall, our findings show that simulated adverse weather substantially increases the difficulty of UAV-based vehicle ReID, reveals clear robustness differences among recent methods, and highlights the need for weather-aware model design and evaluation protocols in future aerial ReID research. The code is released at https://github.com/tranminhvu945/Benchmarking-ReID.
Chinese Translation
基于无人机的车辆重识别（ReID）作为一种有前景的技术，已在交通监控、城市监测和公共安全应用中得到广泛关注，这得益于无人机所提供的灵活视角和广泛覆盖范围。然而，尽管近期在基于无人机的车辆ReID基准方面取得了一定进展，但现有方法在恶劣天气条件下的鲁棒性仍然研究不足。这一点尤为重要，因为天气恶化会显著影响可靠车辆匹配所需的细粒度外观线索，尤其是在小物体尺度、视角变化和复杂背景下。在本文中，我们对三种具有代表性的近期车辆ReID方法，即CLIP-ReID、MSINet和AdaSP，在两个基于无人机的基准数据集VRU和UAV-VeID上进行了受控的比较研究。为了确保一致的鲁棒性评估，我们使用分析天气效应管道生成了两个数据集的合成雾霾和雨天变体，同时保留原始身份和数据分割。所有方法在匹配的清晰、雾霾和雨天条件下进行训练和评估。实验结果表明，恶劣天气在两个数据集上均会持续降低检索性能，其中雨天在几乎所有设置下造成的性能下降大于雾霾。在评估的方法中，AdaSP表现出最强的鲁棒性，在VRU-Large数据集上分别在雾霾和雨天条件下达到了93.0%和88.5%的mAP，在UAV-VeID-Test数据集上达到了88.7%和76.2%的mAP。总体而言，我们的研究结果表明，模拟的恶劣天气显著增加了基于无人机的车辆ReID的难度，揭示了近期方法之间明显的鲁棒性差异，并强调了未来空中ReID研究中需要考虑天气因素的模型设计和评估协议。代码已发布在 https://github.com/tranminhvu945/Benchmarking-ReID。
cs.CV / 75 / 2607.10598
Anomalous Frame Detection by Grouping Frame Similarities between Two Videos Computed by Vision-Language Model to Extract Expert Workers' Unique Actions
通过视觉-语言模型计算两个视频之间的帧相似性进行异常帧检测，以提取专家工人的独特动作
Abstract
Maintenance of critical infrastructures, such as railways and power plants, is essential for operational safety and reliability. However, the declining number of skilled maintenance workers poses a serious challenge to sustaining these operations, highlighting the need to effectively transfer expert know-how to less experienced workers. Although traditional interview-based approaches have been used to elicit maintenance skills, they struggle to capture know-how that experts themselves may not consciously recognize. To address this gap, we proposed a method that detects anomalous frames of candidate actions including know-how by comparing a video of manual-based work with that of expert maintenance workers. In a simulated maintenance experiment involving a distribution board, our method targeted 11 types of actions not described in the manual and achieved a 66.9% extraction rate, marking a 50-percentage-point improvement over conventional techniques. These findings underscore the effectiveness of our approach in revealing hidden maintenance knowledge, thereby contributing to enhanced skill transfer and workforce development in critical infrastructure maintenance.
Chinese Translation
对铁路和电厂等关键基础设施的维护对于操作安全和可靠性至关重要。然而，熟练维护工人数量的减少对维持这些操作构成了严重挑战，突显了有效将专家知识转移给经验较少工人的必要性。尽管传统的基于访谈的方法已被用于引导维护技能，但它们难以捕捉专家自己可能未能意识到的知识。为了解决这一问题，我们提出了一种方法，通过将手动工作的视频与专家维护工人的视频进行比较，检测包括知识在内的候选动作的异常帧。在一个涉及配电板的模拟维护实验中，我们的方法针对手册中未描述的11种动作，达到了66.9%的提取率，比传统技术提高了50个百分点。这些发现强调了我们的方法在揭示隐藏维护知识方面的有效性，从而有助于提高关键基础设施维护中的技能转移和劳动力发展。
cs.CV / 76 / 2607.10605
End-to-End Real-Time Drone-Based Person Detection Framework Using Deep Learning
基于深度学习的端到端实时无人机人员检测框架
Abstract
In recent years, Unmanned Aerial Vehicles (UAVs) or drones have gained rapid response in terms of security, search and rescue (SAR), border surveillance, etc. Existing monitoring frameworks often struggle to maintain detection consistency when targets undergo significant scale variations due to altitude changes, leading to critical information gaps. To address this issue, this work proposes an integrated real-time detection pipeline for detecting targets through the wireless live drone video feed. Build upon YOLOv8-nano architecture, extensive flight experiments were conducted to determine the detection performance across multiple flight altitudes. Trained on VisDrone2019 dataset, the results of YOLOv8-nano model achieves 57.4%, 41%, 44.8% and 20.3% in precision, recall, mAP and mAP50:95 respectively. While demonstrating on real environment, this analysis revealed that the algorithm achieves near-total detection reliability at altitudes between 16 and 25 meters with the detection frame rate consistently maintained above 41 FPS and reaching a peak of 50 FPS. However, the goal of this work is to enable real-time person detection from an aerial platform via wireless transmission. This approach effectively addresses the dual challenges of identifying targets at varying scales and ensuring near-to-accurate localization during aerial observation.
Chinese Translation
近年来，无人机（UAV）在安全、搜索与救援（SAR）、边境监控等领域的快速响应能力得到了广泛关注。现有的监控框架在目标因高度变化而经历显著尺度变化时，往往难以保持检测的一致性，从而导致关键信息的缺失。为了解决这一问题，本研究提出了一种集成的实时检测管道，通过无线实时无人机视频流进行目标检测。基于YOLOv8-nano架构，进行了广泛的飞行实验，以确定在多种飞行高度下的检测性能。YOLOv8-nano模型在VisDrone2019数据集上的训练结果显示，其精度、召回率、mAP和mAP50:95分别达到了57.4%、41%、44.8%和20.3%。在真实环境中的演示分析表明，该算法在16至25米的高度范围内实现了近乎完全的检测可靠性，检测帧率始终保持在41 FPS以上，峰值达到50 FPS。然而，本研究的目标是通过无线传输实现从空中平台的实时人员检测。这种方法有效解决了在不同尺度下识别目标和确保空中观察期间近乎准确定位的双重挑战。
cs.CV / 77 / 2607.10610
WasteAssistant: Regulation-Guided Visual Question Answering Framework for Intelligent Waste Segregation and Sustainable Managemen
WasteAssistant：基于规制指导的视觉问答框架用于智能垃圾分类与可持续管理
Abstract
Efficient waste segregation is critical for sustainable urban management and environmental governance. Existing automated systems are limited by single-modality visual processing, insufficient contextual understanding, and weak regulatory alignment. To address these issues, we propose a language-guided vision-AI framework that integrates vision-language models and multimodal large language models for joint visual-linguistic reasoning. This framework implements a visual question answering paradigm aligned with India's Solid Waste Management Rules 2016. We construct a new WasteVQA dataset with 13,500 question-answer pairs across 21 waste categories. Experiments show that the BLIP-based model achieves a BLEU score of 0.8291 and a BERTScore of 0.9273, outperforming traditional CNN-based methods. This work improves source-level segregation accuracy, ensures regulatory compliance, and supports scalable deployment for municipal and citizen-facing waste management, promoting multimodal AI in sustainable urban infrastructure. The source code and dataset are available at: https://github.com/Khushkataruka/WasteAssistant
Chinese Translation
高效的垃圾分类对可持续城市管理和环境治理至关重要。现有的自动化系统受到单一模态视觉处理、上下文理解不足以及规制对齐弱的限制。为了解决这些问题，我们提出了一种语言引导的视觉人工智能框架，该框架整合了视觉-语言模型和多模态大型语言模型，以实现联合视觉-语言推理。该框架实现了与印度2016年固体废物管理规则相一致的视觉问答范式。我们构建了一个新的WasteVQA数据集，包含21个垃圾类别的13,500个问答对。实验表明，基于BLIP的模型实现了0.8291的BLEU分数和0.9273的BERTScore，优于传统的基于卷积神经网络（CNN）的方法。这项工作提高了源级分类的准确性，确保了规制合规性，并支持市政和面向公民的垃圾管理的可扩展部署，促进了可持续城市基础设施中的多模态人工智能。源代码和数据集可在以下链接获取：https://github.com/Khushkataruka/WasteAssistant
cs.CV / 78 / 2607.10627
Spectral Consistent Flow for One-step 3D Medical Image Translation
用于一步法三维医学图像翻译的谱一致流
Abstract
We present Spectral Consistent Flow (SC-Flow), a 3D medical image translation framework with a single function evaluation (1-NFE) in the latent space. This approach reformulates medical image translation as a stochastic Brownian bridge process that directly constructs a mapping between source and target modalities by predicting the support regularized mean velocity field. To mitigate modality entanglement, over-smoothing, and artifacts induced by the implicit low-pass modulation of the latent average velocity, we introduce a Spectral Consistency Corrector that dynamically regularizes the evolution of the power spectral density via learnable frequency-domain gain modulation. This mechanism establishes an explicit bridge between spatial textures and spectral energy flow, enabling the model to recover fine-grained anatomical fidelity while maintaining global structural coherence. Extensive experiments on four datasets demonstrate that SC-Flow delivers significantly more accurate, consistent, and robust performance across various translation scenarios.
Chinese Translation
我们提出了谱一致流（Spectral Consistent Flow, SC-Flow），这是一种在潜在空间中进行单次函数评估（1-NFE）的三维医学图像翻译框架。该方法将医学图像翻译重新表述为随机布朗桥过程，通过预测支持正则化的均值速度场，直接构建源模态与目标模态之间的映射。为了减轻模态纠缠、过度平滑以及由潜在平均速度的隐式低通调制引起的伪影，我们引入了一种谱一致性校正器（Spectral Consistency Corrector），该校正器通过可学习的频域增益调制动态正则化功率谱密度的演变。该机制在空间纹理与谱能量流之间建立了明确的桥梁，使模型能够在保持全局结构一致性的同时恢复细粒度的解剖学真实性。在四个数据集上的广泛实验表明，SC-Flow在各种翻译场景中提供了显著更准确、一致和稳健的性能。
cs.CV / 79 / 2607.10640
Spectral Heat Flow for Conservative Token Condensation in Vision-Language Models
视觉-语言模型中保守性标记凝聚的谱热流
Abstract
Vision-Language Models (VLMs) are costly at inference time because they must process long sequences of visual tokens. Existing token pruning methods often degrade under high compression by blindly discarding information, breaking spatial structure or collapsing diversity. We propose SpecFlow, a training-free framework that shifts the paradigm from destructive pruning to conservative condensation, strictly enforcing spatial coverage and statistical conservation to ensure stability. Treating visual tokens as nodes in a $k$NN graph, SpecFlow (i) computes a stable importance field via spectral heat flow to preserve structural coherence, (ii) allocates budgets via adaptive spatial partitioning to guarantee coverage, and (iii) aggregates discarded information into coreset sinks to maintain statistical conservation. The method is plug-and-play, requires no fine-tuning, and is compatible with FlashAttention. Experiments confirm that our SpecFlow outperforms SOTA methods across tasks, VLM architectures, and pruning ratios. Notably, LLaVA-1.5 with SpecFlow retains 95.6\% of original performance despite pruning 88.9\% of visual tokens, offering an exceptional efficiency-accuracy balance. Code is available at https://github.com/Lzy-dot/SpecFlow
Chinese Translation
视觉-语言模型（VLMs）在推理时成本高昂，因为它们必须处理长序列的视觉标记。现有的标记剪枝方法通常在高压缩下表现不佳，因为它们盲目丢弃信息，破坏空间结构或降低多样性。我们提出了SpecFlow，一个无训练的框架，它将范式从破坏性剪枝转变为保守性凝聚，严格执行空间覆盖和统计保留以确保稳定性。SpecFlow将视觉标记视为$k$NN图中的节点，(i) 通过谱热流计算稳定的重要性场以保持结构一致性，(ii) 通过自适应空间划分分配预算以保证覆盖，(iii) 将丢弃的信息聚合到核心汇中以维持统计保留。该方法即插即用，无需微调，并与FlashAttention兼容。实验确认我们的SpecFlow在各项任务、VLM架构和剪枝比率上均优于最先进的方法。值得注意的是，使用SpecFlow的LLaVA-1.5在剪枝88.9%的视觉标记的情况下，仍保留了95.6%的原始性能，提供了卓越的效率与准确性平衡。代码可在https://github.com/Lzy-dot/SpecFlow获取。
cs.CV / 80 / 2607.10666
Answer-Conditioned Chain-of-Thought Distillation for Few-Shot Industrial Vision with Small VLMs
基于答案条件的思维链蒸馏用于小型视觉语言模型的少量工业视觉任务
Abstract
Deploying AI-based visual inspection in manufacturing is hard because requirements change often, new defect types appear, and large labeled datasets are rarely available. We propose answer-conditioned chain-of-thought (CoT) distillation for rapidly adapting small vision-language models (VLMs) to new industrial tasks using minimal labeled data. A frontier VLM receives each training image along with its correct label and generates a justified visual explanation. A 3B-parameter model is then fine-tuned on these reasoning-augmented examples via LoRA. By conditioning on correct answers, we ensure all training reasoning is directed toward the correct conclusion, which is critical because frontier models score as low as 24.1% on our hardest task. We validate on four industrial classification tasks spanning three image modalities using only 18 to 30 labeled images per task. Across 4 seeds per task (32 training runs), our method outperforms direct fine-tuning on all 16 seed-task combinations, with mean improvements of +1.7 to +4.4 percentage points. A controlled equal-budget experiment confirms the improvement comes from reasoning quality, not additional training steps. An unconditioned baseline demonstrates that with out answer-conditioning, wrong reasoning degrades performance by 17.8 percentage points. On weld radiograph classification, the fine-tuned 3B model outperforms GPT-4.1 by 10.0pp using just 24 training images.
Chinese Translation
在制造业中部署基于人工智能的视觉检测面临诸多挑战，因为需求经常变化，新缺陷类型不断出现，并且很少有大型标注数据集可用。我们提出了一种基于答案条件的思维链（CoT）蒸馏方法，旨在利用最少的标注数据快速适应小型视觉语言模型（VLMs）到新的工业任务。一个前沿的VLM接收每个训练图像及其正确标签，并生成合理的视觉解释。然后，通过LoRA对这些增强推理的示例进行微调。通过对正确答案的条件限制，我们确保所有训练推理都指向正确的结论，这一点至关重要，因为前沿模型在我们最困难的任务上得分低至24.1%。我们在四个工业分类任务上进行了验证，这些任务涵盖三种图像模态，每个任务仅使用18到30个标注图像。在每个任务的4个种子（共32次训练）中，我们的方法在所有16个种子-任务组合上均优于直接微调，平均提高了1.7到4.4个百分点。一项受控的等预算实验确认了改进来自推理质量，而非额外的训练步骤。一个无条件基线表明，在没有答案条件的情况下，错误的推理使性能下降了17.8个百分点。在焊接放射线图分类任务中，微调后的3B模型在仅使用24个训练图像的情况下，优于GPT-4.1，提升了10.0个百分点。
cs.CV / 81 / 2607.10684
HyperBank: A Differentiable Bank of Classical Priors for Few-Shot Spheroid Microscopy Segmentation
HyperBank：用于少量样本球形显微镜分割的可微分经典先验库
Abstract
Few-shot spheroid segmentation must adapt to new cell lines, microscopes, and illumination conditions from only a small set of annotated images. While foundation few-shot segmenters can be accurate, their large opaque backbones make it difficult to understand which visual cues drive success or failure. We study this question with HyperBank, a differentiable bank of classical image-processing operators combining Frangi vesselness, a Sauvola threshold pyramid, structure-tensor responses, gradient magnitude, and Laplacian-of-Gaussian filters. HyperBank is fitted on the annotated support images and evaluated on disjoint held-out images across three independently acquired spheroid datasets. We treat it not as a general replacement for foundation models, but as a compact, interpretable few-shot microscopy pipeline and an analytic-prior probe of which classical cues carry the few-shot signal. The results show that, adapted on the same few annotated support images, a compact bank of analytic priors is competitive with, and on small-cluster, contrast-driven data can outperform, much larger foundation models, while those models remain stronger on externally sourced, texture-dominated spheroids. Leave-one-family-out ablations indicate that the useful few-shot signal is distributed across operator families and strengthened by support-set-tuned morphology.
Chinese Translation
少量样本球形分割必须仅依赖少量标注图像适应新的细胞系、显微镜和照明条件。尽管基础的少量样本分割器可以达到较高的准确性，但其庞大的不透明结构使得很难理解哪些视觉线索驱动了成功或失败。我们通过HyperBank研究这个问题，HyperBank是一个可微分的经典图像处理算子库，结合了Frangi血管度、Sauvola阈值金字塔、结构张量响应、梯度幅值和拉普拉斯-高斯滤波器。HyperBank在标注的支持图像上进行拟合，并在三个独立获取的球形数据集上评估不重叠的保留图像。我们将其视为基础模型的紧凑、可解释的少量样本显微镜管道，以及分析性先验探针，以识别哪些经典线索承载了少量样本信号。结果表明，在相同的少量标注支持图像上进行适配的紧凑分析先验库与更大基础模型具有竞争力，并且在小集群、对比驱动的数据上可以超越这些更大的基础模型，而后者在外部来源的、以纹理为主的球形样本上仍然表现更强。逐家族剔除的消融实验表明，有用的少量样本信号分布在算子家族之间，并通过支持集调优的形态得到了增强。
cs.CV / 82 / 2607.10690
Incremental Online Scene Reconstruction by 3D Gaussian Triangulation
基于3D高斯三角剖分的增量在线场景重建
Abstract
Incremental scene reconstruction is essential for real-world applications. Although 3D Gaussian Splatting shows strong potential, most existing approaches require offline conversion of the optimized Gaussians into an intermediate implicit field for explicit mesh extraction, which hinders seamless integration with downstream tasks. To address this limitation, we propose a novel online framework that incrementally reconstructs and updates high-fidelity explicit meshes by directly triangulating a dense geometric Gaussian representation, which supports both high-quality rendering and incremental surface reconstruction. Moreover, we present a direct meshing algorithm that efficiently extracts and updates the mesh from the Gaussian set. To ensure mesh accuracy, we enforce a plane-based pulling constraint that dynamically aligns 3D Gaussian primitives to the approximated local surface. Furthermore, our framework significantly reduces memory and computational overhead during long-sequence processing by dynamically freezing fully optimized historical regions. Experiments on public datasets demonstrate that our method outperforms conventional Gaussian-based methods on both rendering quality and reconstruction accuracy.
Chinese Translation
增量场景重建对于现实世界应用至关重要。尽管3D高斯点云显示出强大的潜力，但大多数现有方法需要将优化后的高斯转换为中间隐式场，以便进行显式网格提取，这阻碍了与下游任务的无缝集成。为了解决这一限制，我们提出了一种新颖的在线框架，通过直接对密集几何高斯表示进行三角剖分，增量地重建和更新高保真显式网格，支持高质量渲染和增量表面重建。此外，我们提出了一种直接网格化算法，能够高效提取和更新来自高斯集的网格。为了确保网格的准确性，我们施加了一种基于平面的拉伸约束，动态地将3D高斯原语对齐到近似的局部表面。此外，我们的框架通过动态冻结完全优化的历史区域，显著减少了长序列处理中的内存和计算开销。在公共数据集上的实验表明，我们的方法在渲染质量和重建精度上均优于传统的基于高斯的方法。
cs.CV / 83 / 2607.10695
Effective Synthetic Image Detection via Noise Residual Clustering
通过噪声残差聚类有效检测合成图像
Abstract
The rapid advancement of generative artificial intelligence (AI) has made synthetic images remarkably realistic, posing security threats such as misinformation and fraud. It is significant to detect the synthetic image in the manner of passive and blind image authentication. Most existing detectors rely on supervised training with large labeled datasets, leading to high costs and degraded performance on unknown generative models. To attenuate such deficiencies, we propose a training-free detection method. Specifically, noise residual fingerprints are first extracted by a simple yet effective pre-trained Noiseprint++ model. Then multi-scale features are further extracted from such residual by a frozen Vision Transformer (ViT), followed by adaptive weighted fusion. Only a few real image samples are used needed to initialize the clustering centers for unsupervised K-Means, distinguishing real and synthetic images without training. Extensive evaluations on four benchmark datasets show that our proposed scheme achieves an average accuracy of 82.2%, outperforming the state-of-the-art detectors on generalization ability. Superior performance is gained on the popular diffusion type of synthetic images, and the effectiveness of each module is validated by ablation studies. Source code will be publicly available at https://github.com/multimediaFor/NoiseCluSID.
Chinese Translation
生成性人工智能（AI）的快速发展使得合成图像变得极为真实，带来了诸如虚假信息和欺诈等安全威胁。因此，以被动和盲目的图像认证方式检测合成图像具有重要意义。现有的大多数检测器依赖于使用大型标注数据集进行监督训练，这导致了高成本并且在未知生成模型上的性能下降。为了减轻这些缺陷，我们提出了一种无训练的检测方法。具体而言，首先通过一个简单而有效的预训练模型Noiseprint++提取噪声残差指纹。然后，通过冻结的视觉变换器（Vision Transformer, ViT）进一步从这些残差中提取多尺度特征，并进行自适应加权融合。仅需少量真实图像样本即可初始化无监督K-Means的聚类中心，从而在无需训练的情况下区分真实和合成图像。在四个基准数据集上的广泛评估表明，我们提出的方案达到了82.2%的平均准确率，超越了现有最先进检测器的泛化能力。在流行的扩散类型合成图像上获得了优越的性能，并通过消融研究验证了各模块的有效性。源代码将公开发布在 https://github.com/multimediaFor/NoiseCluSID。
cs.CV / 84 / 2607.10744
Traj-VLN: Learning Pixel-Space Interaction via Autoregressive Trajectory Generation
Traj-VLN：通过自回归轨迹生成学习像素空间交互
Abstract
Benefiting from the powerful priors embedded in large-scale pre-training data and the emerging commonsense reasoning ability, large language models (LLMs) have shown unprecedented generalization capabilities in many research fields. Recently, projecting visual embeddings into the language space via vision-language models (VLMs) to achieve sim-toreal and cross-scene generalization has become a prevailing paradigm in the field of Vision-and-Language Navigation in Continuous Environments (VLN-CE). VLN requires an embodied agent to navigate through unseen environments following natural linguistic instructions. We emphasize that a VLN task can be decomposed into a sequence of sub-tasks, each corresponding to a process of 3D spatial interaction with the environments described by instructions such as "walk to the end of the sofa and turn left." However, such spatial interactions involving moving into the image along the direction of depth sensing are puzzling for VLMs as they were predominantly trained on conversations with RGB images. Rather than incorporating depth or 3D geometric information-which VLMs rarely encounter during pretrainingwe propose an alternative approach: fine-tuning VLMs to learn navigation interactions directly in 2D pixel space through autoregressive trajectory generation. Given a linguistic instruction and historical observations, our model sequentially predicts a series of pixel coordinates, drawing a trajectory from the bottom center of the current observation. While prior work has proved that pixel-goal supervision outperforms learning of discrete actions, our experiments further verify that the supervision of pixel-space trajectory significantly enhances VLN performance. Moreover, we demonstrate that our flagship model achieves state-of-the-art level performance with relatively limited computational resources and training data.
Chinese Translation
得益于大规模预训练数据中嵌入的强大先验知识以及新兴的常识推理能力，大型语言模型（LLMs）在许多研究领域展示了前所未有的泛化能力。最近，通过视觉-语言模型（VLMs）将视觉嵌入投影到语言空间，以实现模拟到现实和跨场景的泛化，已成为连续环境中的视觉与语言导航（VLN-CE）领域的一种流行范式。VLN要求一个具身代理在遵循自然语言指令的情况下，在未见环境中进行导航。我们强调，VLN任务可以分解为一系列子任务，每个子任务对应于与环境进行3D空间交互的过程，这些环境由诸如“走到沙发的尽头并向左转”等指令描述。然而，这种涉及沿深度感知方向进入图像的空间交互对VLMs来说是令人困惑的，因为它们主要是在与RGB图像的对话中进行训练的。我们提出了一种替代方法：通过自回归轨迹生成，在2D像素空间中直接微调VLMs以学习导航交互，而不是引入深度或3D几何信息（VLMs在预训练期间很少接触到）。给定语言指令和历史观测，我们的模型依次预测一系列像素坐标，从当前观测的底部中心绘制轨迹。虽然先前的研究证明像素目标监督优于离散动作学习，但我们的实验进一步验证了像素空间轨迹的监督显著提升了VLN性能。此外，我们展示了我们的旗舰模型在相对有限的计算资源和训练数据下达到了最先进的性能水平。
cs.CV / 85 / 2607.10749
Water Reflection Detection Using Symmetric Attention
基于对称注意力的水面反射检测
Abstract
Reflections of water pose a significant challenge for computer vision systems, as standard deep learning models frequently confuse objects with their mirror images, producing spurious false positives and negatives in tasks such as object detection and semantic segmentation. As a result, detecting reflection axes in natural-water scenes is pivotal for reliable object detection and scene understanding. To mitigate this issue, we leverage the intrinsic imperfect reflective symmetry of water and introduce a Symmetry-Aware Water Reflection Detection Network, namely, SAWRD-Net, that couples dihedral group-equivariant convolutions with a matrix-decomposition decoder in an end-to-end framework. First, dihedral group convolutional layers extract geometry-consistent feature maps that explicitly encode both rotational and mirror symmetries. A Multi-scale Reflection Equivariant block then aggregates features across scales and employs a symmetric-attention mechanism to highlight reflection-relevant regions. The proposed matrix-decomposition decoder factorizes high-dimensional features into compact low-rank parameter and confidence spaces, after which the network directly regresses keypoints on the reflection axis. Then a robust principal component analysis fits the final axis. Evaluated on the largest available water reflection scene data set, SAWRD-Net achieves a true-positive rate of 0.890 against human annotations, outperforming all existing water reflection detectors.
Chinese Translation
水面反射对计算机视觉系统构成了重大挑战，因为标准深度学习模型常常将物体与其镜像混淆，从而在物体检测和语义分割等任务中产生虚假的正负样本。因此，在自然水域场景中检测反射轴对于可靠的物体检测和场景理解至关重要。为了解决这一问题，我们利用水的内在不完美反射对称性，提出了一种对称感知水面反射检测网络，即SAWRD-Net，该网络在端到端框架中结合了二面体群等变卷积和矩阵分解解码器。首先，二面体群卷积层提取几何一致的特征图，明确编码了旋转和镜像对称性。接着，多尺度反射等变块聚合不同尺度的特征，并采用对称注意力机制突出与反射相关的区域。所提出的矩阵分解解码器将高维特征分解为紧凑的低秩参数和置信度空间，之后网络直接回归反射轴上的关键点。最后，采用稳健主成分分析拟合最终的反射轴。在对现有最大的水面反射场景数据集进行评估时，SAWRD-Net在与人工标注的对比中达到了0.890的真实正例率，超越了所有现有的水面反射检测器。
cs.CV / 86 / 2607.10754
TriCons-Pose: Triangle-Invariant Geometric Consistency Learning for Category-Level Object Pose Estimation
TriCons-Pose：用于类别级物体姿态估计的三角形不变几何一致性学习
Abstract
Category-level object pose estimation is a crucial yet challenging task in both academia and industry, and has achieved remarkable success by leveraging keypoint-based correspondence paradigms. However, most existing methods increasingly rely on stronger feature learning while overlooking whether the established correspondences are geometrically stable across diverse perturbations. This often results in fragile pose recovery under intra-class shape variations and occlusions. To tackle this challenge, we develop a novel Triangle-Invariant Geometric Consistency Learning for Category-Level Object Pose Estimation (TriCons-Pose) to anchor stable keypoints and aggregate pose-invariant cues, yielding reliable canonical mapping and accurate pose estimation. Specifically, a Structure-Consistent Keypoint Detector (SCKD) is designed to identify robust keypoints by enforcing cross-view structural consistency via normalized pairwise distance matching. Moreover, we propose a Pose-Invariant Geometric Aggregator (PIGA) to augment keypoint representations by injecting triangle-based pose-invariant descriptors into a local-to-global attention mechanism. The proposed framework is optimized using standard objective functions while incorporating an additional geometry consistency loss. Extensive experiments on REAL275, CAMERA25, and HouseCat6D datasets demonstrate the effectiveness of the proposed approach.
Chinese Translation
类别级物体姿态估计是学术界和工业界一项至关重要但具有挑战性的任务，通过利用基于关键点的对应范式取得了显著成功。然而，大多数现有方法越来越依赖于更强的特征学习，而忽视了所建立的对应关系在不同扰动下的几何稳定性。这常常导致在类内形状变化和遮挡下姿态恢复的脆弱性。为了解决这一挑战，我们开发了一种新颖的用于类别级物体姿态估计的三角形不变几何一致性学习方法（TriCons-Pose），以锚定稳定的关键点并聚合姿态不变的线索，从而实现可靠的典范映射和准确的姿态估计。具体而言，我们设计了一种结构一致性关键点检测器（SCKD），通过规范化成对距离匹配来强制执行跨视图的结构一致性，从而识别出鲁棒的关键点。此外，我们提出了一种姿态不变几何聚合器（PIGA），通过将基于三角形的姿态不变描述符注入到局部到全局的注意机制中，增强关键点表示。所提出的框架使用标准目标函数进行优化，同时结合了额外的几何一致性损失。在REAL275、CAMERA25和HouseCat6D数据集上的大量实验表明了所提方法的有效性。
cs.CV / 87 / 2607.10762
TOLiD: Bridging the Architecture Gap in Vision Foundation Model to LiDAR Pretraining via Token Lifting for Distillation
TOLiD：通过令牌提升实现视觉基础模型与LiDAR预训练之间的架构桥接
Abstract
Cross-modal distillation from Vision Foundation Models (VFMs) to LiDAR backbones has recently emerged as a self-supervised pretraining strategy that reduces reliance on dense point-wise annotation for 3D scene understanding. However, existing distillation pipelines typically treat the VFM as a frozen feature source and train a heterogeneous 3D backbone to match fixed image embeddings, forcing the student to bridge both the modality gap and the cross-architecture gap between dense ViT token representations and sparse 3D encoders. We propose TOLiD, a self-supervised pretraining method for LiDAR representation learning that addresses this gap by coupling a LiDAR backbone with a student Vision Transformer (ViT) initialized from a frozen VFM teacher and applying supervision over compatible patch-token representations. TOLiD converts the set of point features within each image patch frustum into a token using Frustum Pooling followed by Frustum Attention, and performs token-level distillation with visibility masking. For LiDAR-only deployment, we lift token features back to per-point representations using masked bilinear sampling to avoid patches that have limited LiDAR points. We extensively evaluate TOLiD on five heterogeneous LiDAR datasets and four cross-sensor adaptation pairs, demonstrating improved transfer with frozen backbones and lightweight heads.
Chinese Translation
跨模态蒸馏从视觉基础模型（VFM）到LiDAR主干网络最近已成为一种自监督预训练策略，减少了对密集逐点注释在3D场景理解中的依赖。然而，现有的蒸馏流程通常将VFM视为固定特征源，并训练一个异构的3D主干网络以匹配固定的图像嵌入，这迫使学生在密集的ViT令牌表示和稀疏的3D编码器之间弥合模态差距和跨架构差距。我们提出了TOLiD，一种用于LiDAR表示学习的自监督预训练方法，通过将LiDAR主干与从冻结的VFM教师初始化的学生视觉变换器（ViT）相结合，并对兼容的补丁-令牌表示施加监督，从而解决这一差距。TOLiD通过视锥池化（Frustum Pooling）和视锥注意力（Frustum Attention）将每个图像补丁视锥内的点特征集转换为令牌，并通过可见性掩蔽进行令牌级蒸馏。对于仅LiDAR的部署，我们使用掩蔽双线性采样将令牌特征提升回逐点表示，以避免具有有限LiDAR点的补丁。我们在五个异构LiDAR数据集和四个跨传感器适配对上对TOLiD进行了广泛评估，证明了在冻结主干和轻量级头部下的更好迁移性能。
cs.CV / 88 / 2607.10777
RED-Sphere: Hyperspherical Residual Edge Debiasing for Cross-Population Fundus Disease Domain Generalization
RED-Sphere：用于跨人群眼底疾病领域泛化的超球面残差边缘去偏差
Abstract
Medical image classifiers are often trained within one source population, yet clinical deployment requires robustness to patients whose appearance, acquisition style, and disease prevalence differ from the source cohort. Existing fairness and robustness methods often require group supervision or treat appearance variation as an undifferentiated nuisance, which is insufficient when population-correlated low-level cues and lesion evidence share edge and texture structure. We study a strict source-only cross-population setting, where external populations are unseen during optimization, validation, scheduling, hyperparameter and model selection. We propose RED-Sphere, a plug-and-play robustness framework for image classification under unseen population shifts. It estimates shortcut-sensitive nuisance responses with an edge and feature energy prior, attenuates dominant responses through residual soft gating, regularizes masked nuisance views with counterfactual-inspired consistency and separation losses, and predicts labels with normalized spherical prototypes. It favours angular semantic evidence over source-correlated activation magnitude while preserving lesion structure. Although demonstrated on 2D Scanning Laser Ophthalmoscopy (SLO) fundus classification for Age-Related Macular Degeneration (AMD) and Diabetic Retinopathy (DR), RED-Sphere is not tied to retinal anatomy: the same principle can be adapted with modality-specific nuisance priors wherever appearance shortcuts and semantic evidence are entangled. Under a strict White-only Harvard-FairVision protocol, RED-Sphere improves held-out macro-F1 across all 20 task and backbone comparisons, with average gains of 1.28 and 2.98 F1 points on AMD and DR. Gains in AUC and PR-AUC, visual diagnostics, ablations, and sensitivity analyses further support stronger external semantic alignment and more stable angular disease geometry.
Chinese Translation
医学图像分类器通常在一个源人群中进行训练，但临床应用需要对外观、获取方式和疾病流行程度与源队列不同的患者具有鲁棒性。现有的公平性和鲁棒性方法通常需要群体监督，或将外观变化视为未加区分的干扰，这在低级线索和病灶证据与边缘和纹理结构共享时是不够的。我们研究了一个严格的仅源跨人群设置，其中外部人群在优化、验证、调度、超参数和模型选择过程中未被看到。我们提出了RED-Sphere，这是一个用于在未见人群变化下进行图像分类的即插即用鲁棒性框架。它通过边缘和特征能量先验估计对捷径敏感的干扰响应，通过残差软门控减弱主导响应，通过受反事实启发的一致性和分离损失对遮蔽的干扰视图进行正则化，并通过归一化的球面原型预测标签。它更倾向于角度语义证据而非源相关的激活幅度，同时保持病灶结构。尽管在与年龄相关的黄斑变性（AMD）和糖尿病视网膜病变（DR）的2D扫描激光眼底成像（SLO）分类中进行了演示，RED-Sphere并不局限于视网膜解剖：相同的原则可以与特定模态的干扰先验相适应，无论外观捷径和语义证据如何交织。在严格的仅白人哈佛公平视觉（Harvard-FairVision）协议下，RED-Sphere在所有20个任务和骨干网络比较中提高了持出宏F1，AMD和DR的平均增益分别为1.28和2.98 F1点。AUC和PR-AUC的提升、视觉诊断、消融实验和敏感性分析进一步支持了更强的外部语义对齐和更稳定的角度疾病几何结构。
cs.CV / 89 / 2607.10781
Is Energy Guidance All You Need? Training-Free Norm Injection for Driving World Models
能量引导就是你所需要的吗？无训练的规范注入以驱动世界模型
Abstract
Driving world models built on large video-diffusion backbones generate realistic scenes but are hard to control: enforcing a traffic norm typically means retraining the backbone or conditioning it on hand-built layouts. We ask whether controllability requires training at all. Our experiment shows that a rectified-flow driving world model, which jointly generates future video and a planned ego trajectory, can have its planned trajectory steered entirely at sampling time by differentiable energy functions that encode driving norms, without knowledge-specific retraining of the diffusion backbone. Concretely, we demonstrate that a world model built on Open-Sora 2.0 MM-DiT backbone can be steered to brake at a counterfactual target by injecting energy guidance at sampling time. However, we find that the generated video does not yet follow the steered trajectory through the backbone's joint self-attention and identify the cross-stream coupling as a crucial requirement for end-to-end-controllable rollouts.
Chinese Translation
基于大型视频扩散骨干网络构建的驾驶世界模型能够生成逼真的场景，但控制起来却很困难：强制执行交通规范通常意味着需要重新训练骨干网络或将其与手工构建的布局进行条件化。我们探讨了可控性是否完全需要训练。我们的实验表明，一种经过修正的流驱动世界模型，能够同时生成未来视频和规划的自我轨迹，可以通过在采样时使用编码驾驶规范的可微分能量函数，完全引导其规划的轨迹，而无需对扩散骨干进行特定知识的重新训练。具体而言，我们展示了一个基于 Open-Sora 2.0 MM-DiT 骨干的世界模型可以通过在采样时注入能量引导来实现对反事实目标的刹车。然而，我们发现生成的视频尚未遵循通过骨干的联合自注意力引导的轨迹，并识别出跨流耦合是实现端到端可控滚动的关键要求。
cs.CV / 90 / 2607.10783
Toward Efficient Weakly Supervised Semantic Segmentation Using Only Low-Magnification Histopathological Images
基于低倍放大病理图像的高效弱监督语义分割研究
Abstract
Whole-slide images (WSIs) provide rich tissue-level and cellular-level information, but storing and transmitting high-magnification pathology data is resource-intensive. Moreover, annotating WSIs at the pixel level is labor-intensive and time-consuming. Therefore, it is important to investigate whether low-magnification pathology images with limited annotations (i.e., image-level instead of pixel-level labels) can achieve performance comparable to high-magnification images. This paper presents a systematic benchmark study on weakly supervised histopathological image segmentation under different low-resolution storage settings. Starting from high-resolution image patches, we simulate lower-magnification inputs and reconstruct them to the original size using interpolation and deep learning-based reconstruction methods before applying the weakly-supervised segmentation pipeline. This framework enables a quantitative evaluation of how weakly supervised methods respond to different levels of resolution degradation. Experimental results show that reconstruction quality metrics alone are insufficient to predict downstream segmentation performance. In particular, the study identifies a critical degradation point where the localization of small-scale structures declines significantly. These findings provide practical guidance for designing efficient digital pathology storage systems while maintaining reliable automated analysis. Code is available at https://github.com/Dung-Dx/LowMagWSS
Chinese Translation
全切片图像（WSIs）提供了丰富的组织级和细胞级信息，但存储和传输高倍放大病理数据资源消耗巨大。此外，对WSIs进行像素级标注既费时又费力。因此，研究低倍放大病理图像在有限标注（即图像级而非像素级标签）下是否能达到与高倍放大图像相当的性能显得尤为重要。本文系统性地对不同低分辨率存储设置下的弱监督病理图像分割进行了基准研究。从高分辨率图像块出发，我们模拟低倍放大的输入，并在应用弱监督分割流程之前，使用插值和基于深度学习的重建方法将其重建到原始大小。该框架使我们能够定量评估弱监督方法对不同分辨率降级水平的响应。实验结果表明，仅依靠重建质量指标不足以预测下游分割性能。特别是，研究确定了一个关键降级点，在该点，小规模结构的定位显著下降。这些发现为设计高效的数字病理存储系统提供了实用指导，同时保持可靠的自动化分析。代码可在 https://github.com/Dung-Dx/LowMagWSS 获取。
cs.CV / 91 / 2607.10787
Detecting AI-Generated Video: A Vision-Language Dual-View Survey
检测人工智能生成的视频：视觉-语言双视角调查
Abstract
The evolving realism of AI-generated Videos (AIGC-V) is rapidly rendering traditional artifact-centric detection insufficient, necessitating a paradigm shift from low-level inspection to high-level semantic verification. This paper presents a comprehensive survey of AIGC-V detection, reframing the task as Factual Fidelity Verification, which asks whether the events, entities, and physical processes depicted in a video are consistent with real-world facts. To systematize this rapidly evolving field, we propose a Vision-Language Dual-View taxonomy that organizes existing methods into a hierarchical, four-layer landscape, spanning intrinsic cue analysis, spatiotemporal consistency modeling, cross-modal consistency reasoning, and language-guided world-level reasoning. This dual-view framing highlights a fundamental transition from artifact matching in traditional deepfake detection to evidence-based semantic verification enabled by vision-language models and agentic reasoning pipelines. Based on a systematic review of 221 works, we synthesize AIGC-V generation paradigms, survey the landscape of detection methods, and review evaluation metrics and benchmarks in line with proposed views. Finally, we discuss current challenges and identify promising directions toward robust, explainable, and trustworthy detection.
Chinese Translation
人工智能生成视频（AIGC-V）的现实主义不断演变，迅速使传统的以伪造物为中心的检测方法显得不足，迫切需要从低层次的检查转向高层次的语义验证。本文对AIGC-V检测进行了全面的调查，将该任务重新框定为事实忠实性验证，旨在询问视频中描绘的事件、实体和物理过程是否与现实世界的事实一致。为了系统化这一快速发展的领域，我们提出了一种视觉-语言双视角分类法，将现有方法组织成一个四层的分层结构，涵盖内在线索分析、时空一致性建模、跨模态一致性推理和语言引导的世界级推理。这种双视角框架突出了从传统深度伪造检测中的伪造物匹配到基于证据的语义验证的根本转变，这一转变得益于视觉-语言模型和代理推理管道。基于对221项工作的系统回顾，我们综合了AIGC-V生成范式，调查了检测方法的全景，并回顾了与所提视角相符的评估指标和基准。最后，我们讨论了当前的挑战，并确定了朝着稳健、可解释和可信检测的有希望的方向。
cs.CV / 92 / 2607.10792
MAC-Splat: Multi-Attribute Consistency for High-Fidelity Sparse-View Reconstruction
MAC-Splat：高保真稀疏视图重建的多属性一致性
Abstract
Reconstructing high-fidelity 3D scenes from sparse-views remains a central problem in generalizable neural rendering. Existing generalizable 3D Gaussian Splatting (3DGS) methods often exhibit geometric artifacts in sparse-view settings, since supervision based solely on 2D photometric losses cannot resolve depth and correspondence ambiguities. To address this issue, we propose MAC-Splat, a training framework built around direct 3D consistency supervision. MAC-Splat builds on the MASt3R geometric backbone and a frozen DINOv3 encoder to obtain semantically informed 2D correspondences, which serve as geometric anchors for 3D supervision. Using these anchors, we define the Multi-Attribute Consistency (MAC) loss. This objective jointly regularizes the 3D attributes of matched Gaussians, including their position, shape, and appearance, by enforcing agreement in a common world coordinate frame. The formulation is robust to outliers and respects the geometry of covariance matrices, which leads to stable training under sparse-view conditions. Experiments on ScanNet++ show that MAC-Splat outperforms strong baselines, with particularly large gains under different overlap regimes. In particular, it improves average PSNR over Splatt3R by more than 4.5 dB, reduces LPIPS, and maintains performance as the camera pose gap increases. These results indicate that a direct, multi-attribute 3D consistency objective, when combined with high-quality correspondences, is effective for addressing the ill-posed sparse-view reconstruction problem.
Chinese Translation
从稀疏视图重建高保真3D场景仍然是可泛化神经渲染中的一个核心问题。现有的可泛化3D高斯溅射（3DGS）方法在稀疏视图设置中往往会出现几何伪影，因为仅基于2D光度损失的监督无法解决深度和对应关系的歧义。为了解决这个问题，我们提出了MAC-Splat，一个围绕直接3D一致性监督构建的训练框架。MAC-Splat基于MASt3R几何骨干网和一个冻结的DINOv3编码器，以获取语义信息丰富的2D对应关系，这些对应关系作为3D监督的几何锚点。利用这些锚点，我们定义了多属性一致性（MAC）损失。该目标通过在共同的世界坐标系中强制一致性，联合正则化匹配高斯的3D属性，包括它们的位置、形状和外观。该公式对异常值具有鲁棒性，并尊重协方差矩阵的几何结构，从而在稀疏视图条件下实现稳定的训练。在ScanNet++上的实验表明，MAC-Splat优于强基线，尤其是在不同重叠条件下具有显著的提升。具体而言，它在平均PSNR上比Splatt3R提高了超过4.5 dB，减少了LPIPS，并在相机姿态差距增大时保持性能。这些结果表明，直接的多属性3D一致性目标在结合高质量对应关系时，对于解决病态的稀疏视图重建问题是有效的。
cs.CV / 93 / 2607.10796
Mixture of Cognitive Experts in Large Vision-Language Models
大规模视觉-语言模型中的认知专家混合
Abstract
Large Vision Language Models (LVLMs) require strong reasoning over both visual and textual input. Recent work suggests that cognitive elements, especially diverse representations and metacognition, correlate with better performance. Many of the needed perceptual functions are already provided by specialized domain-specific computer vision models, which act as the perceptual subsystem for detecting objects, localizing them, inferring states, recovering spatial layout, and reading text. The key challenge is to integrate these multi-encoder experts into a trustworthy, interpretable, and coherent representation that improves verifiability and reduces hallucinations. This is difficult because vision-language questions span different cognitive levels, yet most LVLM pipelines apply the same perception-reasoning routing regardless of the demand of each query. We propose an evidence-driven multimodal reasoning framework that utilizes a Bloom-inspired taxonomy as a hierarchical reasoning protocol. The two-stage cognitive verbalization first produces a Literal Evidence Summary by decomposing expert outputs into short, atomic evidence statements. It then performs Bloom Verbalization to turn these evidence items into a staged reasoning trace, and a lightweight Reasoning Trace Module quantitatively analyzes the trace to make evidence usage and reasoning progression explicit. Through this integration, we observed several improvements in perception and reasoning abilities. Moreover, the trace module provides quantitative evidence that different queries induce different cognitive entry levels and evidence-use trajectories that enable fine-grained analysis.
Chinese Translation
大规模视觉语言模型（LVLMs）需要对视觉和文本输入进行强有力的推理。近期研究表明，认知元素，特别是多样化的表征和元认知，与更好的性能相关。许多所需的感知功能已经由专门的领域特定计算机视觉模型提供，这些模型作为感知子系统用于检测物体、定位物体、推断状态、恢复空间布局以及读取文本。关键挑战在于将这些多编码器专家整合成一个可信、可解释且连贯的表征，以提高可验证性并减少幻觉。这是困难的，因为视觉-语言问题跨越不同的认知层次，而大多数LVLM管道无论每个查询的需求如何，都应用相同的感知-推理路由。我们提出了一种基于证据的多模态推理框架，利用灵感来自Bloom的分类法作为分层推理协议。两阶段的认知表述首先通过将专家输出分解为简短的原子证据陈述，生成字面证据摘要。然后，它执行Bloom表述，将这些证据项转化为分阶段的推理轨迹，轻量级推理轨迹模块定量分析该轨迹，以明确证据使用和推理进展。通过这种整合，我们观察到感知和推理能力的若干改善。此外，轨迹模块提供了定量证据，表明不同查询引发不同的认知入口层次和证据使用轨迹，从而实现细致的分析。
cs.CV / 94 / 2607.10797
Compositional Context Fine-Tuning Vision-Language Model for Complex Assembly Action Understanding from Videos
基于组合上下文微调的视觉-语言模型用于复杂装配动作理解
Abstract
Assembly action understanding is a key enabler for effective human-robot collaborative assembly, yet it remains challenging due to subtle motions and fine-grained hand-object interactions. We adapt vision-language models (VLMs) to this challenging domain with Compositional Context Fine-Tuning (CCFT), a method that decomposes assembly actions into semantic elements (Verb, Object, Tool) and fine-tunes VLMs to recognize each action element using templated question-answering pairs. This approach ensures near-deterministic outputs. To enable efficient and effective multi-task learning under limited data, a Layer-Partitioned Alternating Training (LP-AT) method is presented, which assigns distinct model layers to recognize specific action elements through element-specific low-rank adapters. LP-AT alternates weight updates across element-specific adapters, reducing cross-task interference while enabling per-adapter hyperparameter optimization. Furthermore, we create HA-ViD-VQA and IKEA-ASM-VQA datasets from existing assembly video datasets. Extensive experiments on these datasets demonstrate that our method consistently outperforms strong action recognition baselines while providing interpretable element-level predictions that can support diverse downstream applications.
Chinese Translation
装配动作理解是实现有效人机协作装配的关键，但由于微妙的动作和细粒度的手物体交互，这一任务仍然具有挑战性。我们通过组合上下文微调（Compositional Context Fine-Tuning, CCFT）方法，将视觉-语言模型（Vision-Language Models, VLMs）适配到这一具有挑战性的领域，该方法将装配动作分解为语义元素（动词、对象、工具），并微调VLMs以使用模板化问答对识别每个动作元素。这种方法确保了近乎确定性的输出。为了在有限数据下实现高效且有效的多任务学习，我们提出了一种层分区交替训练（Layer-Partitioned Alternating Training, LP-AT）方法，该方法为识别特定动作元素分配不同的模型层，通过元素特定的低秩适配器进行识别。LP-AT在元素特定适配器之间交替更新权重，减少跨任务干扰，同时实现每个适配器的超参数优化。此外，我们从现有的装配视频数据集中创建了HA-ViD-VQA和IKEA-ASM-VQA数据集。在这些数据集上的广泛实验表明，我们的方法在强大的动作识别基线之上始终表现优越，同时提供可解释的元素级预测，能够支持多样的下游应用。
cs.CV / 95 / 2607.10800
h-Flow: Flexible Flow-based Image Editing via Doob's h-Transform
h-Flow：基于Doob的h-变换的灵活流式图像编辑
Abstract
Editing images with pre-trained text-to-image flow models typically requires carefully balancing target alignment with the desired prompt and source consistency with the original image. Existing approaches either rely on inversion-based pipelines or heuristic source-to-target trajectory constructions, which often depend on architecture-specific designs or are sensitive to hyperparameters. In this paper, we propose h-Flow, a training-free and theoretically grounded flow-based editing framework. Inspired by Doob's $h$-Transform, we reformulate image editing as conditional generation under multiple terminal events corresponding to source consistency and target alignment. We first extend the classical $h$-Transform from SDE-based models to the deterministic RF framework by constructing an equivalent SDE with identical marginals. Within this formulation, we design dedicated $h$-functions for source consistency and target alignment, yielding closed-form reconstruction guidance and velocity-based semantic editing signals. We further introduce a velocity orthogonal decomposition to decouple reconstruction and editing directions, enabling a controllable trade-off between the two objectives. Extensive experiments demonstrate that h-Flow achieves effective, robust, and flexible editing across diverse scenarios. The code will be released soon.
Chinese Translation
使用预训练的文本到图像流模型编辑图像通常需要在目标对齐与期望提示之间，以及与原始图像的一致性之间进行仔细平衡。现有的方法要么依赖于基于反演的管道，要么依赖于启发式的源到目标轨迹构造，这些方法通常依赖于特定架构的设计或对超参数敏感。在本文中，我们提出了h-Flow，一个无训练且理论基础扎实的基于流的编辑框架。受到Doob的$h$-变换的启发，我们将图像编辑重新表述为在多个终端事件下的条件生成，这些事件对应于源一致性和目标对齐。我们首先将经典的$h$-变换从基于SDE（随机微分方程）模型扩展到确定性RF（随机场）框架，通过构建具有相同边际分布的等效SDE。在此公式中，我们为源一致性和目标对齐设计了专用的$h$-函数，从而产生封闭形式的重建指导和基于速度的语义编辑信号。我们进一步引入速度正交分解，以解耦重建和编辑方向，从而实现两者目标之间的可控权衡。大量实验表明，h-Flow在多种场景下实现了有效、稳健和灵活的编辑。代码将很快发布。
cs.CV / 96 / 2607.10826
3D-DefectBench: A Controlled Factorial Study of Vision-Language Model Evaluation Pipelines for Fine-Grained 3D Generation Defects
3D-DefectBench：针对细粒度3D生成缺陷的视觉-语言模型评估管道的受控因子研究
Abstract
Automated evaluation is essential for scaling generative 3D systems, where exhaustive human review is costly and slow. However, the reliability of an automated judge depends on the entire evaluation pipeline, not only the underlying vision-language model (VLM), but also how assets are rendered, what visual evidence is provided, how the task is specified, and how human reference labels are constructed. We introduce 3D-DefectBench, a benchmark and framework for systematic analysis of VLM-based 3D defect detection pipelines. It complements holistic ratings and pairwise preferences with nine fine-grained binary defects spanning geometry, texture, and prompt adherence, providing actionable diagnostics for generator development and judge evaluation. Using a balanced factorial design, we vary four pipeline factors, VLM, camera protocol, visual input, and prompt schema, across 84 inference designs and approximately 3.2 million scored defect decisions, followed by staged validation on a broader set of frontier models. Model choice is the largest determinant of agreement with human labels, but the remaining factors also affect performance, interact with model selection, and can change the best configuration. Within the evaluated design space, a compact six-view RGB protocol performs comparably to denser multi-view settings and inputs augmented with depth or surface normals, making it a strong cost-effective default. Under this standardized pipeline, the best of 12 VLM judges still lag behind trained human labelers, while texture agreement drops sharply when expert-consensus labels are replaced by noisier silver labels. These findings show that automated judges should be evaluated as complete pipelines and calibrated across human reference regimes, rather than benchmarked only as standalone models. We release labels, prompts, predictions, and Croissant metadata on Hugging Face.
Chinese Translation
自动化评估对于扩展生成3D系统至关重要，因为全面的人为审查成本高且速度慢。然而，自动评估的可靠性依赖于整个评估管道，不仅仅是基础的视觉-语言模型（VLM），还包括资产的渲染方式、提供的视觉证据、任务的具体说明以及人类参考标签的构建方式。我们引入了3D-DefectBench，一个用于系统分析基于VLM的3D缺陷检测管道的基准和框架。它通过九种细粒度的二元缺陷，涵盖几何、纹理和提示遵循，补充了整体评分和成对偏好，为生成器开发和评估者评估提供了可操作的诊断。采用平衡因子设计，我们在84个推理设计和大约320万个评分缺陷决策中，改变了四个管道因素：VLM、相机协议、视觉输入和提示模式，随后在更广泛的前沿模型集上进行了分阶段验证。模型选择是与人类标签一致性的最大决定因素，但其余因素也会影响性能，与模型选择相互作用，并可能改变最佳配置。在评估的设计空间中，一个紧凑的六视图RGB协议的表现与更密集的多视图设置相当，并且与深度或表面法线增强的输入相比，成为一个强大的性价比选择。在这个标准化管道下，12个VLM评估者中表现最好的仍然落后于经过训练的人类标注者，而当专家共识标签被更嘈杂的银标签替代时，纹理一致性急剧下降。这些发现表明，自动评估者应作为完整管道进行评估，并在不同的人类参考体系中进行校准，而不仅仅作为独立模型进行基准测试。我们在Hugging Face上发布了标签、提示、预测和Croissant元数据。
cs.CV / 97 / 2607.10840
OmniX: Any-view and Any-time 4D Reconstruction via Feed-forward Trajectory Fields
OmniX：通过前馈轨迹场实现任意视角和任意时间的4D重建
Abstract
Previous feed-forward 4D reconstruction methods either predict per-frame static point clouds, ignoring foreground motion, or estimate point cloud trajectories while being limited to small camera motions. This restricts their ability to aggregate observations over time and reconstruct complete dynamic scenes under large viewpoint changes. To address this limitation, we propose OmniX, a feed-forward 4D reconstruction framework that predicts dense 3D point trajectories for every pixel from videos with large camera motion. OmniX decouples dynamic motion modeling from static geometry prediction and represents motion using a compact set of dynamic tokens. By leveraging the sparse and low-rank structure of 3D motion, these tokens generate trajectory fields for all pixels across all images while efficiently preserving global interactions. To facilitate training, we further build an automatic UE5-based 4D data engine and introduce a large-scale dataset containing 80K scenes and 1.28M multi-view videos with full geometric annotations. OmniX achieves state-of-the-art performance on dense 3D point trajectory prediction and 3D point tracking, while also demonstrating competitive results on video depth estimation and camera pose estimation.
Chinese Translation
以往的前馈4D重建方法要么预测每帧的静态点云，忽略前景运动，要么在小范围相机运动的限制下估计点云轨迹。这限制了它们在时间上聚合观察结果的能力，并在大视角变化下重建完整的动态场景。为了解决这一限制，我们提出了OmniX，一个前馈4D重建框架，能够从具有大相机运动的视频中为每个像素预测稠密的3D点轨迹。OmniX将动态运动建模与静态几何预测解耦，并使用一组紧凑的动态标记表示运动。通过利用3D运动的稀疏和低秩结构，这些标记为所有图像中的所有像素生成轨迹场，同时有效地保留全局交互。为了便于训练，我们进一步构建了一个基于UE5的自动化4D数据引擎，并引入了一个包含8万场景和128万多视角视频的规模庞大的数据集，所有数据均附有完整的几何注释。OmniX在稠密3D点轨迹预测和3D点跟踪方面达到了最先进的性能，同时在视频深度估计和相机姿态估计方面也表现出竞争力。
cs.CV / 98 / 2607.10841
Align and Segment: Unsupervised Learning for Building Segmentation From Misaligned Labels
对齐与分割：基于未对齐标签的建筑分割无监督学习
Abstract
Supervised learning for image segmentation typically requires spatially aligned image and label sets. When images and labels originate from different sources, the pairing may be misaligned, which can significantly deteriorate the performance of the learned models. This is especially common in remote sensing, when aerial or satellite images are co-registered with labels from another source (e.g., OpenStreetMap). In this work, we propose a novel approach for training on misaligned labels, where we simultaneously learn the label alignment. Our align and segment (AnS) approach builds on the spatial transformer module to transform the misaligned labels using an affine transformation to provide a better learning target for a canonical semantic segmentation network. We prevent shortcut learning of misaligned labels in these semantic segmentation networks through a self-supervised regularization loss and show that it is complementary to data augmentation, especially for systematically misaligned training data. A decisive characteristic of our AnS approach is that it learns without requiring any golden labels. We experimentally show on both synthetic and real-world data from different cities that our approach enables high-quality building segmentation and precise label-image alignment at the same time. Code and derived datasets are available at https://github.com/venkanna37/align-and-segment
Chinese Translation
图像分割的监督学习通常需要空间上对齐的图像和标签集。当图像和标签来自不同来源时，配对可能会出现未对齐的情况，这可能会显著降低学习模型的性能。这在遥感中尤为常见，当航空或卫星图像与来自其他来源的标签（例如，OpenStreetMap）进行共同配准时。在本研究中，我们提出了一种新的方法来处理未对齐标签的训练，同时学习标签的对齐。我们的对齐与分割（Align and Segment, AnS）方法基于空间变换模块，通过仿射变换来转换未对齐的标签，从而为标准语义分割网络提供更好的学习目标。我们通过自监督正则化损失防止这些语义分割网络对未对齐标签的捷径学习，并表明这与数据增强是互补的，特别是对于系统性未对齐的训练数据。我们AnS方法的一个决定性特征是它在不需要任何黄金标签的情况下进行学习。我们在来自不同城市的合成数据和真实世界数据上进行了实验，结果表明我们的方法能够同时实现高质量的建筑分割和精确的标签-图像对齐。代码和衍生数据集可在 https://github.com/venkanna37/align-and-segment 获取。
cs.CV / 99 / 2607.10851
Learning To Focus: Anatomy-Guided Attention Regularization for Medical Image Classification
学习聚焦：基于解剖引导的注意力正则化用于医学图像分类
Abstract
Medical image classification models are ideally expected to identify diagnostically relevant regions while making predictions, yet standard classification losses rarely provide spatial supervision. Explicit supervision via anatomical shape information, such as segmentation masks of task-relevant anatomy, has been shown to guide the network toward regions relevant to the target prediction. However, obtaining such masks incurs substantial manual annotation effort and computational overhead. With the advent of segmentation foundation models that exhibit strong localization of anatomical structures across diverse imaging modalities, we leverage this capability to extract anatomical shape priors without the burden of training a dedicated segmentation model. In this paper, we propose a new framework, Locus, an anatomical attention regularization framework that leverages pretrained segmentation foundation models to guide a classifier's attention toward diagnostically meaningful anatomical structures across diverse imaging modalities. Instead of enforcing pixel-wise alignment with the foundation-model-derived mask, we introduce a regularization term that adaptively balances attention between anatomical (foreground) and background regions, penalizing the classifier when background attention dominates. We validate Locus on eight diverse medical imaging datasets spanning dermoscopy, X-ray, histopathology, and cardiac MRI, showing consistent gains in classification performance alongside improved anatomically grounded attention.
Chinese Translation
医学图像分类模型理想情况下应在进行预测时识别出诊断相关区域，然而标准分类损失很少提供空间监督。通过解剖形状信息（例如与任务相关的解剖结构的分割掩膜）进行显式监督已被证明能够引导网络关注与目标预测相关的区域。然而，获取这些掩膜需要大量的手动标注工作和计算开销。随着分割基础模型的出现，这些模型在不同成像模式下展现出对解剖结构的强定位能力，我们利用这一能力提取解剖形状先验，而无需训练专门的分割模型。在本文中，我们提出了一个新的框架Locus，这是一种解剖注意力正则化框架，利用预训练的分割基础模型引导分类器的注意力关注于不同成像模式下具有诊断意义的解剖结构。我们并不强制与基础模型导出的掩膜进行逐像素对齐，而是引入一个正则化项，自适应地平衡解剖（前景）与背景区域之间的注意力，当背景注意力占主导时对分类器进行惩罚。我们在八个不同的医学成像数据集上验证了Locus，这些数据集涵盖了皮肤镜检查、X光、组织病理学和心脏MRI，结果显示分类性能持续提升，同时解剖基础的注意力也得到了改善。
cs.CV / 100 / 2607.10853
Diversify Diffusion with Temperature Sampling and Variance-Corrective Time Shifting
Abstract
Diffusion models faithfully reproduce their training distribution, but also inherit its imbalances and leave rare or under-represented modes hard to reach. A natural inference-time remedy is to sample from the high-temperature target $p^{(\gamma)}_0(x) \propto p_0(x)^{\gamma}$ for $0 < \gamma < 1$, which flattens dominant modes and lifts rare ones. However, naive score scaling while correctly reweighting modes also inflates the per-mode variance, breaking the reverse diffusion process and degrading sample quality. We introduce variance-corrective time shifting, a training-free fix that queries the network at a shifted timestep and scales the resulting score by $\gamma$, canceling the variance inflation while preserving the mode reweighting. The correction turns simple temperature sampling into a practical diversity knob for pretrained diffusion and flow-matching backbones with no retraining, and we demonstrate consistent gains at minimal cost to sample quality and condition fidelity across DiT, Stable Diffusion and Motion Diffusion models. We further show that the timing of the temperature intervention enables coarse-to-fine control: high-noise stages drive compositional diversity across modes, while low-noise stages drive local appearance variation under a fixed composition.
cs.CV / 101 / 2607.10860
AU-Guided Synthetic Video Generation for Micro-Expression Recognition
基于AU引导的微表情识别合成视频生成
Abstract
Micro-expression recognition is limited by the small scale, narrow demographic coverage, and restricted emotion labels of existing datasets. We introduce EquiME, a synthetic micro-expression dataset built from AU-guided image-to-video generation. EquiME contains 75K videos generated from 15K source face images across five target emotions, together with automatically inferred demographic metadata and video-quality measurements. We evaluate EquiME using frame-pair similarity, spatial variation, and no-reference perceptual-quality metrics, together with cross-dataset MER experiments on SAMM and CASME II. Models trained on EquiME achieve competitive cross-dataset performance on SAMM and CASME II and show comparatively low variation across the four evaluated architectures. This paper focuses on the dataset design, the structured AU-conditioning pipeline used for video generation, and the empirical evidence needed to assess EquiME as a synthetic MER resource. Project page: https://kirito-blade.github.io/me-vlm/
Chinese Translation
微表情识别受到现有数据集规模小、人口覆盖面窄以及情感标签有限的限制。我们提出了EquiME，这是一个基于AU引导的图像到视频生成的合成微表情数据集。EquiME包含来自15K源面部图像生成的75K视频，涵盖五种目标情感，并附带自动推断的人口统计元数据和视频质量测量。我们通过帧对相似性、空间变化和无参考感知质量指标对EquiME进行评估，并在SAMM和CASME II上进行跨数据集的微表情识别实验。基于EquiME训练的模型在SAMM和CASME II上实现了具有竞争力的跨数据集性能，并在四种评估架构中表现出相对较低的变化。本文重点讨论数据集设计、用于视频生成的结构化AU条件管道，以及评估EquiME作为合成微表情识别资源所需的实证证据。项目页面：https://kirito-blade.github.io/me-vlm/
cs.CV / 102 / 2607.10873
X-GuideAR: An Augmented Reality Framework to Mitigate Radiation Exposure during Fluoroscopic Guidance
X-GuideAR：一种增强现实框架，用于减少透视引导下的辐射暴露
Abstract
Achieving optimal screw placement for orthopedic surgeries requires frequent alignment checks and multiple anatomical views under X-ray -- a process known as "fluoro-hunting" that increases radiation exposure to patients and surgical teams. This work introduces X-GuideAR, an augmented reality (AR) framework for identifying optimal X-ray views, aimed at reducing radiation exposure while ensuring accurate screw placement. To exemplify the benefits of X-GuideAR, we focus on S2 alar-iliac (S2AI) screw placement. Our system provides radiation-free guidance for view acquisition and drilling by generating synthetic X-ray previews that accelerate fluoro-hunting. Once the required anatomical views are identified using these previews, a real X-ray is acquired, and the preview of the drilling trajectory is augmented onto it, facilitating precise screw placement with minimal additional radiation. A preliminary study involving eight S2AI trajectories performed by an expert spine surgeon demonstrated a 62.3% reduction in the number of X-rays. Post-procedure evaluations showed that trajectories done with X-GuideAR supported an average safe screw diameter of 12.95 mm compared to 5.9 mm under the conventional workflow, suggesting improved bony containment and potential biomechanical benefit. X-GuideAR shows great potential to reduce radiation exposure and streamline S2AI screw placement, offering a promising direction toward safer and more efficient surgeries.
Chinese Translation
在骨科手术中，实现最佳螺钉放置需要频繁的对齐检查和多种解剖视图下的X光成像——这一过程被称为“透视捕捉”，会增加患者和手术团队的辐射暴露。本文介绍了X-GuideAR，一种增强现实（AR）框架，用于识别最佳X光视图，旨在减少辐射暴露，同时确保螺钉的准确放置。为了展示X-GuideAR的优势，我们重点关注S2翼髂（S2AI）螺钉的放置。我们的系统通过生成合成X光预览，为视图获取和钻孔提供无辐射引导，从而加速透视捕捉。一旦使用这些预览识别出所需的解剖视图，就会获取真实的X光，并将钻孔轨迹的预览叠加到其上，从而在最小额外辐射的情况下促进精确的螺钉放置。一项涉及八条由专家脊柱外科医生执行的S2AI轨迹的初步研究表明，X光数量减少了62.3%。术后评估显示，使用X-GuideAR进行的轨迹支持的安全螺钉直径平均为12.95毫米，而传统工作流程下为5.9毫米，暗示了更好的骨骼包容性和潜在的生物力学益处。X-GuideAR展现了减少辐射暴露和简化S2AI螺钉放置的巨大潜力，为更安全、更高效的手术提供了有希望的方向。
cs.CV / 103 / 2607.10898
Design Choices in Splitting-Based Self-Supervised Sparse-View CT Reconstruction
基于分割的自监督稀疏视图CT重建中的设计选择
Abstract
Self-supervised data splitting has emerged as a promising paradigm for sparse-view CT reconstruction, enabling training from incomplete measurements without fully sampled ground truth. However, the influence of key design choices, including partitioning strategy, preprocessing, and inference, remains insufficiently understood. In this work, we introduce a unified framework that decomposes splitting-based reconstruction into these three components, enabling controlled comparison of existing methods and two incremental extensions: multi-partition splitting and an alternative inference strategy. Experiments on simulated LoDoPaB-CT data under independent and correlated noise, together with validation on the real-world 2DeteCT dataset, show that the optimal partitioning strategy strongly depends on the measurement noise structure. Lattice-based splitting performs favorably under independent noise, whereas angular masking is more robust under correlated noise and real measured data. Multi-partition splitting consistently improves over pure projection-wise splitting in several settings. Complementary perceptual and structural metrics, including LPIPS and HaarPSI, reveal differences between masking strategies that are less apparent from PSNR and SSIM alone. These results provide practical guidelines for designing self-supervised sparse-view CT reconstruction methods and highlight the limitations of common independence assumptions in realistic imaging environments.
Chinese Translation
自监督数据分割已成为稀疏视图CT重建的一个有前景的范式，使得在没有完全采样的真实数据的情况下进行不完整测量的训练成为可能。然而，关键设计选择的影响，包括分区策略、预处理和推理，仍然未得到充分理解。在本研究中，我们提出了一个统一框架，将基于分割的重建分解为这三个组成部分，从而能够对现有方法进行受控比较，并提出了两个增量扩展：多分区分割和替代推理策略。在独立和相关噪声下对模拟的LoDoPaB-CT数据进行实验，并在真实世界的2DeteCT数据集上进行验证，结果表明，最佳分区策略强烈依赖于测量噪声结构。在独立噪声下，基于晶格的分割表现良好，而在相关噪声和真实测量数据下，角度掩蔽则更具鲁棒性。在多个设置中，多分区分割始终优于纯投影分割。包括LPIPS和HaarPSI在内的互补感知和结构度量揭示了掩蔽策略之间的差异，而这些差异在仅使用PSNR和SSIM时不太明显。这些结果为设计自监督稀疏视图CT重建方法提供了实用指南，并突显了在现实成像环境中常见独立假设的局限性。
cs.CV / 104 / 2607.10912
DP-Splat: Bayesian Nonparametric Complexity Control for Gaussian Splatting
DP-Splat：高斯喷溅的贝叶斯非参数复杂度控制
Abstract
3D Gaussian Splatting represents scenes as finite mixtures of anisotropic Gaussians whose number of components $K$ is set by heuristic density control or user caps. Variational Bayes Gaussian Splatting (VBGS) recast splat fitting as conjugate variational inference, but $K$ remains fixed. We replace the finite symmetric Dirichlet over mixture weights with a truncated stick-breaking Dirichlet-process prior -- and, as a theory-backed alternative, a sparse overfitted finite Dirichlet -- so that the number of occupied components adapts to the data while every update remains a closed-form coordinate-ascent step; a natural-gradient stochastic variant makes the per-step cost independent of the number of points. We give an exact monotonicity guarantee, a rigorous truncation-error bound correcting an anti-conservative large-$\alpha$ approximation in common use, and an honest account of what the fitted number of components estimates. Empirically: (i) the effective complexity $\hat{K}$ adapts to scene complexity and recovers the true $K$ within $\pm 1$ on well-separated synthetic data with regime-appropriate concentration; (ii) a deconfounded comparison shows the DP prior's contribution is complexity selection, not per-component efficiency -- converged DP fits exceed single-pass fixed-$K$ VBGS by +2.7 dB at matched budgets yet tie an equally converged fixed-$K$ baseline, and on 3D scenes DP-Splat matches or exceeds VBGS's held-out color prediction with 5.9-7.6x fewer components; (iii) the posterior-predictive color variance is well calibrated on model-matched synthetic data; and (iv) the ordering suggested by exact-posterior asymptotics reverses under mean-field coordinate ascent: the DP prior resists over-splitting while the sparse finite mixture saturates its truncation, a gap between variational practice and posterior asymptotics documented across three orders of magnitude in $N$.
Chinese Translation
3D 高斯喷溅将场景表示为各向异性高斯的有限混合体，其成分数量 $K$ 由启发式密度控制或用户上限设定。变分贝叶斯高斯喷溅（VBGS）将喷溅拟合重新表述为共轭变分推断，但 $K$ 仍然是固定的。我们用截断的条形破裂德尔菲过程先验替代有限对称的狄利克雷混合权重——作为理论支持的替代方案，使用稀疏过拟合的有限狄利克雷——使得被占用的成分数量能够适应数据，同时每次更新仍然保持为封闭形式的坐标上升步骤；一种自然梯度随机变体使得每步的成本与点的数量无关。我们提供了一个精确的单调性保证，一个严格的截断误差界限，纠正了常用的大 $eta$ 近似中的反保守性，并对拟合成分数量的估计进行了诚实的说明。在实证方面：(i) 有效复杂度 $ ilde{K}$ 能够适应场景复杂度，并在具有适当浓度的良好分离的合成数据中恢复真实的 $K$，误差范围在 $ ext{±}1$ 之内；(ii) 去混淆的比较表明，DP 先验的贡献在于复杂度选择，而非每个成分的效率——收敛的 DP 拟合在匹配预算下超越了单次固定 $K$ 的 VBGS，提升了 +2.7 dB，但与同样收敛的固定 $K$ 基线持平，并且在 3D 场景中，DP-Splat 匹配或超越了 VBGS 的保留颜色预测，使用的成分数量减少了 5.9-7.6 倍；(iii) 后验预测的颜色方差在模型匹配的合成数据上得到了良好的校准；(iv) 精确后验渐近所建议的顺序在均场坐标上升下发生了逆转：DP 先验抵制过度喷溅，而稀疏有限混合体饱和了其截断，这在 $N$ 的三个数量级中记录了变分实践与后验渐近之间的差距。
cs.CV / 105 / 2607.10949
Unsupervised Detection of Entry and Exit Regions from Vehicle Trajectories for Camera-Agnostic Turning Movement Counts
基于无监督学习的车辆轨迹入口和出口区域检测方法：适用于摄像头无关的转向运动计数
Abstract
Turning movement counts are essential for intersection-level traffic management, yet their collection remains predominantly manual due to the cost of per-camera region annotation. This paper presents an unsupervised pipeline that identifies entry and exit regions directly from raw vehicle trajectories extracted via object detection and multi-object tracking, requiring no manual annotation, camera calibration, or prior knowledge of intersection geometry. Unlike trajectory clustering methods that classify individual trajectories using pairwise similarity and must be re-executed on every new batch, the proposed pipeline clusters initial and terminal point locations to produce persistent spatial region polygons that classify future trajectories by point-in-polygon containment at linear cost. The pipeline comprises six sequential steps, each with configurable parameters evaluated through a systematic statistical analysis spanning 19,152 pipeline executions across 25 surveillance cameras capturing dense heterogeneous traffic in Bengaluru, India, and 10 sequences from the UA-DETRAC benchmark dataset. Both parametric and nonparametric testing frameworks identify three consistently significant parameters and yield an empirically grounded recommended configuration. Under this configuration, the pipeline achieves a median classification error of approximately 3% across all 25 cameras, including 16 held-out locations, with GEH values within accepted engineering thresholds. Compared with two trajectory clustering baselines, the proposed pipeline exhibits greater stability across camera views and lower computational cost, at the expense of higher median error. Extended evaluation demonstrates that calibration clips of at least 60 minutes and peak-traffic selection further improve region estimation quality.
Chinese Translation
转向运动计数对于交叉口级交通管理至关重要，但由于每个摄像头区域标注的成本，其收集仍主要依赖人工。本文提出了一种无监督的处理流程，能够直接从通过目标检测和多目标跟踪提取的原始车辆轨迹中识别入口和出口区域，无需人工标注、摄像头校准或交叉口几何形状的先验知识。与通过成对相似性对单个轨迹进行分类的轨迹聚类方法不同，所提流程聚类初始和终点位置，生成持久的空间区域多边形，通过点在多边形内的包含关系以线性成本对未来轨迹进行分类。该流程包括六个顺序步骤，每个步骤都有可配置的参数，通过对在印度班加罗尔的25个监控摄像头捕获的密集异构交通进行的19,152次流程执行的系统统计分析进行评估，以及来自UA-DETRAC基准数据集的10个序列。参数检验和非参数检验框架识别出三个始终显著的参数，并得出一个基于实证的推荐配置。在该配置下，流程在所有25个摄像头上实现了约3%的中位分类误差，包括16个保留位置，GEH值在可接受的工程阈值内。与两种轨迹聚类基线相比，所提流程在摄像头视角之间表现出更大的稳定性和更低的计算成本，尽管中位误差较高。扩展评估表明，至少60分钟的校准片段和高峰交通选择进一步提高了区域估计质量。
cs.CV / 106 / 2607.10953
Learning Anatomy-Grounded CT Vision-Language Representations with Organ-Hierarchical Report Knowledge
基于器官层级报告知识的解剖学驱动CT视觉-语言表示学习
Abstract
Medical vision-language pretraining (VLP) from paired CT images and radiology reports enables scalable representation learning, but most existing methods align either whole scans with entire reports or local image regions with text fragments. These formulations underuse a key property of radiology reports: findings are organized around anatomical structures, with abnormalities described by organs, disease concepts, locations, and severity-related attributes. We propose OKA-CT, an organ-hierarchical knowledge-augmented framework for CT-report VLP. OKA-CT first converts free-text reports into organ-conditioned knowledge using radiology report parsing and LLM-assisted semantic structuring. The extracted hierarchy is used across two learning stages. Stage~1 injects anatomy-grounded evidence into the CT visual representation through fine-grained organ-conditioned supervision, while Stage~2 uses organ-specific report evidence to guide structured report-CT contrastive learning, where hierarchy-derived semantic soft targets treat non-paired cases with shared organ-level findings as weak semantic positives rather than uniform negatives. A lightweight query-based global branch further aggregates disease-relevant volumetric evidence for whole-scan representation. On CT-RATE and RAD-ChestCT datasets, OKA-CT achieves zero-shot abnormality diagnosis AUROCs of 84.9 and 72.2, outperforming prior CT VLP baselines. Retrieval and patch-occlusion analyses further show improved report-image alignment and stronger sensitivity to disease-associated anatomical regions.
Chinese Translation
医学视觉-语言预训练（VLP）通过配对的CT图像和放射学报告实现可扩展的表示学习，但大多数现有方法要么将整个扫描与完整报告对齐，要么将局部图像区域与文本片段对齐。这些方法未能充分利用放射学报告的一个关键特性：发现是围绕解剖结构组织的，异常由器官、疾病概念、位置和严重性相关属性描述。我们提出了OKA-CT，一个用于CT报告VLP的器官层级知识增强框架。OKA-CT首先通过放射学报告解析和大语言模型（LLM）辅助的语义结构化，将自由文本报告转换为器官条件知识。提取的层级结构在两个学习阶段中使用。阶段1通过细粒度的器官条件监督将解剖学驱动的证据注入CT视觉表示，而阶段2则使用器官特定的报告证据指导结构化报告-CT对比学习，其中层级派生的语义软目标将共享器官级发现的非配对案例视为弱语义正例，而非统一负例。一个轻量级的基于查询的全局分支进一步聚合与疾病相关的体积证据，以实现整个扫描的表示。在CT-RATE和RAD-ChestCT数据集上，OKA-CT实现了零样本异常诊断的AUROC值分别为84.9和72.2，超越了之前的CT VLP基准。检索和补丁遮挡分析进一步显示了报告与图像的对齐改善，以及对与疾病相关的解剖区域的更强敏感性。
cs.CV / 107 / 2607.10984
EquiFusion: Kinematics-Agnostic Human Motion Prediction via Equivariant Latent Diffusion
EquiFusion：通过等变潜在扩散实现与运动学无关的人体运动预测
Abstract
Existing Stochastic 3D Human Motion Prediction models are fundamentally constrained by hard-coding the skeleton kinematics, severely limiting generalization, preventing cross-dataset training, and requiring complex data retargeting. We introduce EquiFusion, the first kinematics-agnostic model to solve this bottleneck, implementing a latent diffusion model with a permutation equivariant architecture. EquiFusion treats the kinematics' connectivity as an explicit input parameter, ensuring its internal computations are inherently agnostic to joint ordering and graph structure. This novel design enables truly cross-dataset generalization to unseen kinematics and unlocks novel zero-shot directions, such as motion prediction from partial or occluded observations and targeted limb generation. EquiFusion achieves state-of-the-art results on major benchmarks, being up to 75% more compact than previous kinematics-specific methods, while achieving faster training and inference. EquiFusion thus establishes a new, flexible standard for robust human motion prediction. Model and training code are available at https://ceveloper.github.io/publications/equifusion/.
Chinese Translation
现有的随机3D人体运动预测模型在根本上受到硬编码骨骼运动学的限制，严重影响了模型的泛化能力，阻碍了跨数据集训练，并且需要复杂的数据重定向。我们提出了EquiFusion，这是第一个解决这一瓶颈的与运动学无关的模型，采用了具有置换等变架构的潜在扩散模型。EquiFusion将运动学的连接性视为一个显式输入参数，确保其内部计算本质上与关节顺序和图结构无关。这一新颖的设计使得模型能够在未见过的运动学上实现真正的跨数据集泛化，并开启了新的零样本方向，例如从部分或遮挡观测中进行运动预测和目标肢体生成。EquiFusion在主要基准测试中取得了最先进的结果，其模型体积比之前的运动学特定方法小达75%，同时实现了更快的训练和推理。因此，EquiFusion为稳健的人体运动预测建立了一个新的灵活标准。模型和训练代码可在 https://ceveloper.github.io/publications/equifusion/ 获取。
cs.CV / 108 / 2607.10985
MED-DSLC: Multi-Expert-Domain Classification via Domain Supervision and Logit Calibration
MED-DSLC：通过领域监督和对数校准实现多专家领域分类
Abstract
Vision-language models (VLMs) such as CLIP enable zero-shot classification by comparing image features with text prompts in a shared embedding space. A fundamental property underlying this capability is the global comparability of logits across arbitrary candidate classes. However, VLMs are often adapted to fine-grained domains using techniques such as LoRA. While this improves in-domain accuracy, out-of-domain accuracy degrades. This leads to a highly fragmented model ecosystem, with thousands of specialized models. Multi-Expert-Domain classification seeks to address this problem, by merging LoRAs trained independently on specialized domains. However, due to the independent training, the various domain experts no longer produce globally calibrated logits. As a result, when evaluating over the union of multiple domain-specific class sets, heterogeneous logit scales induce cross-domain interference and artificially high confidence for out-of-domain classes, inducing prediction errors. In this work, we identify domain supervision and cross-domain logit miscalibration as the key issue to scalable multi-domain zero-shot recognition. We propose MED-DSLC, combining domain supervised training and domain-wise logit scaling, to explicitly restore global logit comparability. MED-DSLC is a lightweight solution for MED classification, which is shown to preserve within-domain discrimination while reducing cross-domain logit interference with minimal data. Extensive experiments across diverse fine-grained benchmarks demonstrate that it substantially improves mean accuracy (+15\%), cross-domain robustness, and scalability in the size of MED classification problem. Our results show that restoring output-level calibration is essential under highly data imbalanced settings for achieving a truly zero-shot VLM under multi-domain specialization.
Chinese Translation
视觉语言模型（VLMs），如CLIP，通过在共享嵌入空间中比较图像特征与文本提示，实现零样本分类。这一能力的基本特性是任意候选类别之间对数的全球可比性。然而，VLMs通常使用如LoRA等技术适应细粒度领域。虽然这提高了领域内的准确性，但领域外的准确性却下降。这导致了一个高度碎片化的模型生态系统，拥有数千个专业化模型。多专家领域分类旨在通过合并在专业领域独立训练的LoRA来解决这一问题。然而，由于独立训练，各个领域专家不再产生全球校准的对数。因此，在评估多个领域特定类别集合的并集时，异构的对数尺度会引发跨领域干扰，并对领域外类别产生人为的高置信度，从而导致预测错误。在本研究中，我们确定领域监督和跨领域对数失调是可扩展多领域零样本识别的关键问题。我们提出了MED-DSLC，结合领域监督训练和领域特定的对数缩放，以明确恢复全球对数可比性。MED-DSLC是一种轻量级的MED分类解决方案，已证明在减少跨领域对数干扰的同时保持领域内的区分性，且所需数据极少。在多样化的细粒度基准上进行的广泛实验表明，它显著提高了平均准确性（+15%）、跨领域鲁棒性和MED分类问题规模的可扩展性。我们的结果表明，在高度数据不平衡的情况下，恢复输出级别的校准对于实现真正的零样本VLM在多领域专业化中是至关重要的。
cs.CV / 109 / 2607.10988
MMA-Former: Multi-Window Mixture-of-Head Attention Transformer for Adaptive PNI Prediction in 3D MRI
MMA-Former：用于3D MRI中自适应PNI预测的多窗口混合头注意力变换器
Abstract
Perineural invasion (PNI) is a critical prognostic factor in cholangiocarcinoma. Non-invasive prediction from 3D MRI is challenging, demanding models that efficiently capture both fine-grained details and global context. We propose the Multi-window Mixture-of-Head Attention Transformer (MMA-Former), a novel end-to-end 3D architecture featuring a Coarse-Fine Transformer (CFT) structure for parallel multi-scale feature extraction. We advance this structure by integrating a novel Window-Specific Mixture-of-Head attention (WS-MoH) mechanism. Unlike standard Multi-Head Self Attention (MSA), WS-MoH generates a representation for each 3D window and dynamically routes the entire window to specialized or common attention heads. This enables spatially adaptive feature extraction tailored to the local context of each window, enhancing specialization and reducing redundancy without increasing parameters. Evaluated on a retrospective dataset of 168 T1-weighted MRI scans, MMA-Former achieved an AUC of 0.752, outperforming other 3D architectures, including the best CNN (AUC of 0.708) and Transformer baselines (AUC of 0.681).
Chinese Translation
神经周围侵袭（PNI）是胆管癌的重要预后因素。从3D MRI进行非侵入性预测具有挑战性，需要能够有效捕捉细粒度细节和全局上下文的模型。我们提出了多窗口混合头注意力变换器（MMA-Former），这是一种新颖的端到端3D架构，具有粗细变换器（CFT）结构，用于并行多尺度特征提取。我们通过整合一种新颖的窗口特定混合头注意力（WS-MoH）机制来推进这一结构。与标准的多头自注意力（MSA）不同，WS-MoH为每个3D窗口生成一个表示，并动态地将整个窗口路由到专用或通用的注意力头。这使得特征提取能够根据每个窗口的局部上下文进行空间自适应，增强了专业化并减少了冗余，而不增加参数。在对168个T1加权MRI扫描的回顾性数据集进行评估时，MMA-Former达到了0.752的AUC，超越了其他3D架构，包括最佳CNN（AUC为0.708）和变换器基线（AUC为0.681）。
cs.CV / 110 / 2607.10990
TreeSoc: Tree-Structured Dynamic Reasoning and Tool Synergy for Soccer Video Understanding
TreeSoc：用于足球视频理解的树状动态推理与工具协同
Abstract
Automated understanding of complex soccer scenarios from video remains a significant challenge for contemporary vision-language models (VLMs), which suffer from shallow cross-modal alignment and exhibit fundamental limitations in multi-step reasoning and coordinated tool integration. We present TreeSoc, a structured reasoning framework that reformulates soccer video question answering as a hierarchical search problem rather than a single-pass prediction. Specifically, TreeSoc employs a dynamic depth-first search (DFS) mechanism that decomposes complex queries into sequentially ordered sub-tasks, enabling iterative reasoning refinement through explicit intermediate states. This tree-structured decomposition naturally supports adaptive tool routing, wherein domain-specific modules are selectively activated and their outputs incorporated at each reasoning node to produce contextually grounded predictions. On SoccerBench, TreeSoc achieves state-of-the-art performance, with accuracies of 85.2%, 87.4%, and 82.2% on TextQA, ImageQA, and VideoQA, respectively. Additionally, TreeSoc further demonstrates strong cross-domain generalization, attaining 74.16% accuracy on NExT-QA. These results establish structured, tool-augmented tree reasoning as an effective paradigm for robust video understanding. Code is available at: https://github.com/thanhnhan29/TreeSoc.
Chinese Translation
从视频中自动理解复杂足球场景仍然是当代视觉-语言模型（VLMs）面临的一项重大挑战，这些模型在跨模态对齐方面存在浅层次的问题，并在多步骤推理和协调工具整合方面表现出根本性的局限性。我们提出了TreeSoc，一个结构化推理框架，将足球视频问答重新表述为一个层次搜索问题，而非单次预测。具体而言，TreeSoc采用动态深度优先搜索（DFS）机制，将复杂查询分解为顺序排列的子任务，通过显式的中间状态实现迭代推理的细化。这种树状分解自然支持自适应工具路由，其中领域特定模块在每个推理节点被选择性激活，并将其输出纳入，以生成上下文相关的预测。在SoccerBench上，TreeSoc实现了最先进的性能，在TextQA、ImageQA和VideoQA上的准确率分别为85.2%、87.4%和82.2%。此外，TreeSoc还展示了强大的跨领域泛化能力，在NExT-QA上达到了74.16%的准确率。这些结果确立了结构化、工具增强的树推理作为稳健视频理解的有效范式。代码可在以下链接获取：https://github.com/thanhnhan29/TreeSoc。
cs.CV / 111 / 2607.10992
LoSA-Net: A Localized and Scale-Adaptive Network for Boundary-Sensitive Prediction of Perineural Invasion in 3D MRI
LoSA-Net：一种局部化和尺度自适应的网络，用于3D MRI中神经周围侵袭的边界敏感预测
Abstract
Perineural invasion (PNI) is a clinically relevant indicator of tumor aggressiveness and can influence surgical decision-making, motivating interest in reliable preoperative assessment. The subtle MRI features of PNI, however, often resemble nearby anatomy, complicating noninvasive prediction. These fine perineural cues are easily attenuated by routine downsampling or overly global feature aggregation, reducing the effectiveness of conventional volumetric models. We present LoSA-Net, a localized and scale-adaptive architecture for boundary-sensitive PNI prediction in 3D MRI. Talking Neighborhood Attention (TNA) preserves nerve-aligned detail through localized self-attention with head-wise mixing, and Scale-Adaptive Feature Mixing (SAFM) modulates the receptive field using multi-scale depthwise processing. Cross-Scale Refinement and Alignment (CSRA) maintains consistency between semantic context and high-resolution boundaries across stages. In contrast-enhanced MRI scans from 168 patients with cholangiocarcinoma, LoSA-Net achieves an AUC of 0.7567 and outperforms representative convolutional and transformer baselines under matched preprocessing and optimization settings.
Chinese Translation
神经周围侵袭（PNI）是肿瘤侵袭性的重要临床指标，能够影响外科决策，从而引发对可靠术前评估的关注。然而，PNI在MRI中的细微特征常常与附近解剖结构相似， complicating 非侵入性预测。这些细微的神经周围线索容易被常规的下采样或过于全局的特征聚合所削弱，从而降低传统体积模型的有效性。我们提出了LoSA-Net，一种用于3D MRI中边界敏感的PNI预测的局部化和尺度自适应架构。谈话邻域注意（Talking Neighborhood Attention, TNA）通过局部自注意力和头部混合保留与神经对齐的细节，而尺度自适应特征混合（Scale-Adaptive Feature Mixing, SAFM）则通过多尺度深度处理调节感受野。跨尺度精炼与对齐（Cross-Scale Refinement and Alignment, CSRA）在各个阶段保持语义上下文与高分辨率边界之间的一致性。在来自168名胆管癌患者的对比增强MRI扫描中，LoSA-Net达到了0.7567的AUC，并在匹配的预处理和优化设置下优于代表性的卷积和变换基线。
cs.CV / 112 / 2607.10993
Confidence Scores in Open-Vocabulary Detection Are a Biased Mixture of Scale and Semantics
开放词汇检测中的置信分数是尺度与语义的偏倚混合
Abstract
Foundation models such as CLIP have enabled open-vocabulary object detectors that generalise to novel categories via vision-language similarity. However, the confidence scores these detectors produce are not reliable localization probability estimates: they conflate visual scale and semantic query specificity with the true detection signal. Through controlled experiments on COCO across three foundation-model-based detectors (GroundingDINO, OWL-ViT, YOLO-World), with the scale-bias finding further replicated on LVIS (1,203 categories) using GroundingDINO, we show that s=cos(v,t) is a biased mixture of two effects. Scale bias (alpha = +0.064, r = 0.579, p = 1.29 x 10^-58) systematically inflates scores for large objects. Semantic bias (beta = -0.705, p = 5.23 x 10^-41) suppresses scores for generic queries. Both biases are structurally inevitable from CLIP's image-level pretraining. Threshold adjustment cannot remove them: oracle per-scale thresholding yields Delta F1 = +0.001 for small objects versus +0.102 for large. A parameter-free temperature scaling correction improves small-object Recall@10 by 19.6% (p < 0.01) without retraining. This comes at a modest, measurable cost to pooled-ranking precision, so the bias is partially, not freely, reversible at inference time. These findings reveal a fundamental limitation of adapting image-level foundation models to region-level detection tasks.
Chinese Translation
基础模型如CLIP使得开放词汇物体检测器能够通过视觉-语言相似性对新类别进行泛化。然而，这些检测器产生的置信分数并不是可靠的定位概率估计：它们将视觉尺度和语义查询的特异性与真实检测信号混合在一起。通过在COCO上对三种基于基础模型的检测器（GroundingDINO、OWL-ViT、YOLO-World）进行的控制实验，以及在LVIS（1,203个类别）上使用GroundingDINO进一步复制的尺度偏倚发现，我们展示了s=cos(v,t)是两种效应的偏倚混合。尺度偏倚（alpha = +0.064，r = 0.579，p = 1.29 x 10^-58）系统性地提高了大物体的分数。语义偏倚（beta = -0.705，p = 5.23 x 10^-41）抑制了通用查询的分数。这两种偏倚是CLIP图像级预训练的结构性必然结果。阈值调整无法消除它们：按尺度的理想阈值设置对于小物体的Delta F1为+0.001，而对于大物体为+0.102。无参数的温度缩放校正在不重新训练的情况下将小物体的Recall@10提高了19.6%（p < 0.01）。这在一定程度上以可测量的代价影响了汇总排名的精度，因此在推理时偏倚是部分可逆的，而非完全可逆。这些发现揭示了将图像级基础模型适应于区域级检测任务的一个基本局限性。
cs.CV / 113 / 2607.10995
AsySplat: Efficient Asymmetric 3D Gaussian Splatting for Long-Sequence Scene Modeling
AsySplat：高效的不对称三维高斯点云建模用于长序列场景建模
Abstract
Recent generalizable 3D Gaussian Splatting models have advanced long-sequence novel view synthesis (NVS), but at the cost of substantial redundant computation. We identify that the redundancy can be mitigated based on two observations: (i) high-precision geometry is not strictly required for high-quality NVS; (ii) appearance learning is generally easier than geometry recovery. Motivated by these insights, we propose an asymmetric architecture that decouples geometry and appearance modeling. The geometry branch processes coarse-grained tokens with most of the parameters for multi-view reconstruction, while the appearance branch operates on fine-grained tokens to capture details using significantly fewer parameters. The two branches interact through bilateral connections, enabling mutual guidance for their respective tasks. This task-aware asymmetry reduces the computational redundancy and allocates the computation more judiciously, thereby increasing parameter efficiency and enabling smaller models to achieve strong performance. On 32-view 960P inputs, our model matches optimization-based methods while delivering nearly 800x speedup, and surpasses the zero-shot performance of state-of-the-art generalizable models with markedly fewer parameters and reduced training/inference overhead, achieving an overall efficiency improvement.
Chinese Translation
近期的可泛化三维高斯点云模型在长序列新视图合成（NVS）方面取得了进展，但代价是大量冗余计算。我们发现可以通过两个观察来减轻冗余：（i）高精度几何形状并不是高质量NVS的严格要求；（ii）外观学习通常比几何恢复更容易。基于这些见解，我们提出了一种不对称架构，解耦了几何和外观建模。几何分支处理粗粒度的标记，并使用大部分参数进行多视图重建，而外观分支则在细粒度标记上操作，以显著更少的参数捕捉细节。这两个分支通过双向连接进行交互，从而实现相互指导各自的任务。这种任务感知的不对称性减少了计算冗余，并更合理地分配计算，从而提高了参数效率，使得更小的模型能够实现强大的性能。在32视图960P输入下，我们的模型与基于优化的方法相匹配，同时实现了近800倍的加速，并且在参数显著更少、训练/推理开销降低的情况下，超越了最先进的可泛化模型的零-shot性能，实现了整体效率的提升。
cs.CV / 114 / 2607.10998
Temporal Feature Distillation for Label-Efficient Precise Event Spotting in Sports Videos
用于标签高效精确事件检测的时间特征蒸馏在体育视频中的应用
Abstract
Precise Event Spotting (PES) requires distinguishing visually similar yet semantically distinct adjacent frames, making it fundamentally different from image classification and coarse action recognition. Although self-distillation methods such as DINO have shown strong representation learning ability in images, we find that directly applying them to PES is ineffective: without supervised guidance, subtle but crucial motion cues are often suppressed as noise, leading to representations that are insensitive to precise event boundaries. To address this, we propose Temporal Feature Distillation, a semi-supervised objective that aligns temporally informative backbone features, rather than projection-head outputs, to preserve motion-sensitive and boundary-aware cues for frame-level localization. A supervised warm-up with a ramp-up schedule further stabilizes training by ensuring that meaningful event cues are learned before unlabeled distillation begins. We also introduce Transformer Gate Shift, a multi-scale gated shifting module that injects motion-aware temporal information into Vision Transformers. Experiments on four fine-grained sports benchmarks show consistent improvements over fully supervised and semi-supervised baselines. Under 10\% supervision on FSPerf, our method improves mAP by 4.54 points over the strongest competing approach, and with only 80\% labeled data, it matches or surpasses the fully supervised 100\% baseline on two of the four datasets.
Chinese Translation
精确事件检测（PES）需要区分视觉上相似但语义上不同的相邻帧，这使其与图像分类和粗略动作识别有根本性的不同。尽管自蒸馏方法如 DINO 在图像中展现了强大的表征学习能力，但我们发现直接将其应用于 PES 是无效的：在没有监督指导的情况下，细微但关键的运动线索常常被压制为噪声，导致的表征对精确事件边界不敏感。为了解决这个问题，我们提出了时间特征蒸馏，这是一种半监督目标，旨在对齐时间上信息丰富的主干特征，而不是投影头输出，以保留对运动敏感和边界感知的线索，以便进行帧级定位。通过逐步增加的监督预热进一步稳定了训练，确保在开始无标签蒸馏之前学习到有意义的事件线索。我们还引入了 Transformer Gate Shift，这是一种多尺度门控移动模块，将运动感知的时间信息注入视觉变换器中。在四个细粒度体育基准上的实验显示，相较于完全监督和半监督基线，我们的方法 consistently 提高了性能。在 FSPerf 上，10\% 的监督下，我们的方法相较于最强竞争方法提高了 4.54 个点的 mAP，并且在仅有 80\% 标签数据的情况下，在四个数据集中有两个的表现与完全监督的 100\% 基线相匹配或超越。
cs.CV / 115 / 2607.11008
SynCLIP: Synonym-Coherent Language-Image Pretraining for Robust Open-Vocabulary Dense Perception
SynCLIP：用于稳健开放词汇密集感知的同义词一致语言-图像预训练
Abstract
Open-vocabulary dense perception (OVDP) aims to localize objects unseen during training by leveraging textual knowledge. Despite the remarkable progress of recent CLIP-based approaches, we identify a critical limitation: synonym-induced grounding inconsistency, where semantically equivalent expressions yield disparate spatial attention patterns. This inconsistency undermines the robustness and performance of existing methods in real-world OVDP applications. To address this issue, we propose SynCLIP, a Synonym-Coherent Language-Image Pretraining framework that enhances synonym-robust grounding for OVDP. SynCLIP introduces a Semantic-consistent Spatial Attention alignment (SSA) module to enhance spatial attention consistency by minimizing discrepancies between attention maps of original and synonymous expressions. Furthermore, a Spatial Attention Refinement (SAR) module selectively strengthens the most semantically relevant spatial regions within aligned maps for more precise and stable grounding. To support synonym-coherent pretraining, we also construct a Synonym-Enriched Visual Corpus (SEViC), which augments each category with multiple synonyms and textual definitions. Extensive experiments on multiple benchmarks demonstrate that SynCLIP substantially improves grounding consistency under diverse linguistic variants and achieves state-of-the-art performance among CLIP-based OVDP methods. Code is available at https://github.com/Justlovesmile/SynCLIP.
Chinese Translation
开放词汇密集感知（OVDP）旨在通过利用文本知识定位训练期间未见的物体。尽管近期基于CLIP的方法取得了显著进展，但我们识别出一个关键限制：同义词引起的定位不一致性，即语义上等价的表达产生不同的空间注意模式。这种不一致性削弱了现有方法在实际OVDP应用中的稳健性和性能。为了解决这一问题，我们提出了SynCLIP，一个同义词一致的语言-图像预训练框架，旨在增强OVDP的同义词稳健定位。SynCLIP引入了一个语义一致空间注意对齐（SSA）模块，通过最小化原始表达和同义表达的注意图之间的差异，增强空间注意的一致性。此外，空间注意精炼（SAR）模块选择性地增强对齐图中最语义相关的空间区域，以实现更精确和稳定的定位。为了支持同义词一致的预训练，我们还构建了一个同义词丰富视觉语料库（SEViC），为每个类别增加多个同义词和文本定义。在多个基准上的广泛实验表明，SynCLIP在不同语言变体下显著提高了定位一致性，并在基于CLIP的OVDP方法中实现了最先进的性能。代码可在 https://github.com/Justlovesmile/SynCLIP 获取。
cs.CV / 116 / 2607.11025
Reference-Based Face Super-Resolution Using the Spatial Transformer
基于参考的面部超分辨率方法使用空间变换器
Abstract
Face super-resolution is the task of increasing the resolution of an image containing a face thereby adding finer detail. It is a ubiquitous task in many computer vision applications and quite often the user isn't even aware that it is being performed. However, doing it with high fidelity is challenging as it is an ill-posed problem. In this paper we present a reference-based solution for face super-resolution that uses higher resolution reference images to aid in the task. We show an alignment module based on the spatial transformer that is considerably more stable than the popular deformable convolutions. We also show an aggregation function that can take good quality information from the reference images when available or suppress the function when such information is unavailable. Finally, we show that our relatively smaller model can achieve state of the art results on multiple datasets. The source code is available at https://github.com/varun-jois/FSRST.
Chinese Translation
面部超分辨率是提高包含面部图像的分辨率，从而增加细节的任务。这在许多计算机视觉应用中是一个普遍存在的任务，而用户往往并不知道这一过程的进行。然而，以高保真度实现这一目标是具有挑战性的，因为这是一种病态问题。在本文中，我们提出了一种基于参考的面部超分辨率解决方案，该方案利用更高分辨率的参考图像来辅助该任务。我们展示了一种基于空间变换器的对齐模块，其稳定性显著高于流行的可变形卷积。我们还展示了一种聚合函数，当参考图像可用时，可以从中提取高质量信息，或者在此类信息不可用时抑制该功能。最后，我们展示了我们相对较小的模型在多个数据集上可以达到最先进的结果。源代码可在 https://github.com/varun-jois/FSRST 获取。
cs.CV / 117 / 2607.11034
RTFVE: Realtime Face Video Enhancement
RTFVE：实时人脸视频增强
Abstract
There's been a surge in adoption of video conferencing applications for both personal and business use cases. However, the bandwidth limitations faced by many users worldwide may restrict the optimal use of such applications. Although deep learning offers a solution for enhancing low bit rate videos, most models today are either hard to incorporate with modern compression standards or require specialized hardware to run such as significant GPUs making these models impractical. To address these issues, we introduce the Realtime Face Video Enhancement (RTFVE) model which can be easily incorporated with any video decoder and can run in realtime on ordinary CPUs. Experiments show that our model improves perceptual quality over the compressed video baseline at multiple low bitrate settings. The source code will be made available at https://github.com/varun-jois/RTFVE.
Chinese Translation
近年来，视频会议应用在个人和商业用例中的采用激增。然而，许多用户在全球范围内面临的带宽限制可能会限制此类应用的最佳使用。尽管深度学习为增强低比特率视频提供了解决方案，但目前大多数模型要么难以与现代压缩标准结合，要么需要如显著的GPU等专用硬件来运行，这使得这些模型在实际应用中不够可行。为了解决这些问题，我们提出了实时人脸视频增强（RTFVE）模型，该模型可以轻松与任何视频解码器结合，并且可以在普通CPU上实时运行。实验表明，我们的模型在多个低比特率设置下改善了压缩视频基线的感知质量。源代码将发布在 https://github.com/varun-jois/RTFVE。
cs.CV / 118 / 2607.11040
FSFVE: Few Shot Compressed Face Video Enhancement
FSFVE：少样本压缩人脸视频增强
Abstract
Videocalling has become a popular form of communication in the world today, with many companies providing free services for it. However, there are still millions of people around the world that experience poor quality videocalls due to limitations in bandwidth. This despite, most people having the required hardware. In this paper we present a novel framework for enhancing highly compressed videocalls. We show, that with as little as 10 frames of the face, we can rapidly (in under 100 seconds) train a model to enhance that instance of the videocall. The model can be trained either prior to or during the call, enhancing the rest of the call by producing better quality video. The video conferencing application need not be modified - it can be off the shelf with our system as a layer on top that trains quickly then simply lets the video conferencing application (e.g. Zoom) run as usual, where our system intercepts and improves images before they are displayed. The model is designed to run in realtime on low-compute devices such as a typical laptop CPU. Experimentally, we show that the model significantly improves quality of compressed face video both quantitatively as well as perceptually. Code can be found at https://github.com/varun-jois/FSFVE.
Chinese Translation
视频通话已成为当今世界一种流行的沟通方式，许多公司提供免费的服务。然而，全球仍有数百万人由于带宽限制而体验到低质量的视频通话。尽管大多数人拥有所需的硬件。在本文中，我们提出了一种新颖的框架，用于增强高度压缩的视频通话。我们展示了，仅用10帧人脸图像，我们可以快速（在100秒以内）训练一个模型来增强该视频通话的实例。该模型可以在通话前或通话期间进行训练，通过生成更高质量的视频来增强通话的其余部分。视频会议应用程序无需修改——它可以是现成的，我们的系统作为一个快速训练的层，随后让视频会议应用程序（例如Zoom）正常运行，我们的系统在图像显示之前进行拦截和改善。该模型设计为在低计算设备（如典型的笔记本电脑CPU）上实时运行。实验结果表明，该模型在定量和感知上显著提高了压缩人脸视频的质量。代码可以在 https://github.com/varun-jois/FSFVE 找到。
cs.CV / 119 / 2607.11064
WiFi-JEPA: Self-supervised Learning for WiFi-CSI 3D Human Pose Estimation
WiFi-JEPA：用于WiFi-CSI三维人体姿态估计的自监督学习
Abstract
WiFi Channel State Information (CSI) enables privacy-preserving human pose sensing in camera-denied environments, but existing WiFi-based pose estimators often fail under environment shifts and rely on costly camera-based annotation pipelines that limit scale. We propose WiFi-JEPA, a self-supervised framework that learns CSI-native representations by predicting masked latent embeddings instead of reconstructing raw CSI signals that may contain hardware-specific artifacts. WiFi-JEPA makes three contributions: (i) CSI-specific tokenization and link masking tailored to the CSI tensor over channel, time, and link (C,T,L); masking entire Tx-Rx antenna links forces the model to predict one spatial link view from others, capturing cross-link correlations informative of 3D spatial structure. (ii) A ray-tracing CSI simulation pipeline that generates diverse unlabeled CSI from randomized geometric primitives, providing scalable pre-training data without pose annotations. (iii) State-of-the-art results on Person-in-WiFi-3D: WiFi-JEPA outperforms prior WiFi-CSI baselines on both single- and multi-person 3D pose estimation under the same evaluation protocol. We also show that simulated CSI provides complementary pre-training signal to real CSI, and that four vision-native SSL objectives degrade performance below training from scratch, whereas WiFi-JEPA consistently improves downstream pose estimation.
Chinese Translation
WiFi信道状态信息（CSI）能够在无摄像头的环境中实现隐私保护的人体姿态感知，但现有的基于WiFi的姿态估计器在环境变化下往往表现不佳，并依赖于昂贵的基于摄像头的标注流程，限制了其规模。我们提出了WiFi-JEPA，一个自监督框架，通过预测被遮蔽的潜在嵌入而非重建可能包含硬件特定伪影的原始CSI信号，从而学习CSI原生表示。WiFi-JEPA有三个贡献：（i）针对CSI张量在信道、时间和链路（C,T,L）上的特定标记化和链路遮蔽；遮蔽整个发射-接收天线链路迫使模型从其他链路预测一个空间链路视图，捕捉对三维空间结构有信息量的跨链路相关性。（ii）一个光线追踪CSI仿真管道，从随机几何原件生成多样的无标签CSI，提供可扩展的预训练数据而无需姿态标注。（iii）在Person-in-WiFi-3D上的最先进结果：WiFi-JEPA在相同评估协议下，在单人和多人三维姿态估计中均优于之前的WiFi-CSI基线。我们还表明，仿真CSI为真实CSI提供了互补的预训练信号，而四个视觉原生的自监督学习目标使性能低于从头训练，而WiFi-JEPA则持续改善下游姿态估计。
cs.CV / 120 / 2607.11071
DDR-Net: Haze-Aware Dual-Domain Refinement for Single-Image Dehazing
DDR-Net：面向雾霾的双域精细化单幅图像去雾
Abstract
Single-image dehazing aims to recover clear scenes from haze-degraded images. It remains challenging due to the atmospheric scattering and the complexity of real-world haze distributions. Although recent end-to-end networks have achieved promising performance, two issues still limit their effectiveness: insufficient feature refinement at the bottleneck stage and weak local structural representation in encoder-decoder architectures. Thus, we propose a Haze-Aware Dual-Domain Refinement Network (DDR-Net) for single-image dehazing. Our method is built upon three modules: Haze Prior Extractor (HPE) provides multi-scale haze-aware priors by operating directly on downsampled hazy images; Detail-Enhanced Blocks (DE Blocks) serve as the core feature extraction units, capturing multi-scale structural information and enhancing edge and texture recovery via gradient-aware convolutions; and Spatial-Frequency Bottleneck Refinement (SFBR) at the bottleneck jointly exploits spatial and frequency information to refine bottleneck features. DDR-Net achieves more effective feature representation and reconstruction for haze removal. Extensive experiments on real-world benchmarks demonstrate that our method outperforms existing dehazing approaches. It achieves competitive performance on synthetic datasets.
Chinese Translation
单幅图像去雾旨在从受雾霾影响的图像中恢复清晰场景。由于大气散射和现实世界雾霾分布的复杂性，这一任务仍然具有挑战性。尽管最近的端到端网络已取得了令人鼓舞的性能，但仍有两个问题限制了它们的有效性：瓶颈阶段特征精细化不足，以及编码器-解码器架构中局部结构表示较弱。因此，我们提出了一种面向雾霾的双域精细化网络（Haze-Aware Dual-Domain Refinement Network，DDR-Net）用于单幅图像去雾。我们的方法基于三个模块构建：雾霾先验提取器（Haze Prior Extractor，HPE）通过直接处理下采样的雾霾图像提供多尺度的雾霾先验；细节增强块（Detail-Enhanced Blocks，DE Blocks）作为核心特征提取单元，捕捉多尺度结构信息，并通过梯度感知卷积增强边缘和纹理的恢复；空间频率瓶颈精细化（Spatial-Frequency Bottleneck Refinement，SFBR）在瓶颈处联合利用空间和频率信息来精细化瓶颈特征。DDR-Net实现了更有效的特征表示和雾霾去除重建。在真实世界基准上的大量实验表明，我们的方法优于现有的去雾方法，并在合成数据集上也表现出竞争力。
cs.CV / 121 / 2607.11078
Do Video-LLMs Actually Watch? Diagnosing Character-Tracking Failures in Long-Form Video
视频大型语言模型真的在观看吗？诊断长视频中的角色追踪失败
Abstract
Can a Video Large Language Model (Video-LLM) follow one person through a long video, keeping track of who they are well enough to report, in order, how their outfit changes across a full TV episode? Benchmarks increasingly score this kind of task, and the strongest open-source 7--8B models now reach 37--38% on InfiniBench's global appearance task, which asks exactly that. But does that score come from tracking the named character, or from something easier? We test this with a nine-condition diagnostic protocol applied to three architecturally distinct open-source Video-LLMs, with Gemini~2.5~Flash as a frontier reference, and find the accuracy does not come from character tracking. When we change the character named in the question to a different cast member, leaving the video and answer options untouched, the models change their answer only 4--31% of the time, so they are largely ignoring who the question asks about. Breaking that test down by the gender of the swapped name shows why: the models react more when the name is changed to a different-gender character than to a same-gender one (a 13--28 point gap), picking up coarse gender cues but unable to tell same-gender individuals apart. This shallow processing surfaces again when we drop the multiple-choice options and ask the same questions open-endedly: open-source accuracy drops 18--25 points, with none of 151 answers fully correct, versus a 12-point drop for Gemini. Further checks rule out the obvious innocent explanations, adding subtitles, using the most informative frames, or doubling the number of frames all leave character tracking unimproved, so the bottleneck is not how much video the model sees but how it ties that video to the person the question names. We release a diagnostic toolkit for auditing what such benchmark scores actually measure.
Chinese Translation
视频大型语言模型（Video-LLM）能否在长视频中跟踪一个人，准确记录他们的服装在整集电视节目中的变化？基准测试越来越多地评估这种任务，而当前最强的开源7-8B模型在InfiniBench的全球外观任务中达到了37-38%的得分，该任务正是要求完成这一点。但这个得分是来自于对指定角色的追踪，还是来自于一些更简单的任务？我们通过应用于三种架构上有显著差异的开源Video-LLM的九种条件诊断协议进行测试，以Gemini~2.5~Flash作为前沿参考，发现准确性并不来自于角色追踪。当我们将问题中提到的角色更换为不同的演员，而视频和答案选项保持不变时，模型仅在4-31%的情况下改变答案，因此它们在很大程度上忽视了问题所询问的对象。按性别分析更换的名字显示了原因：当名字更改为不同性别的角色时，模型的反应更强烈，而对同性别角色的反应则较弱（差距为13-28分），模型能够捕捉到粗略的性别线索，但无法区分同性别个体。当我们去掉多项选择选项，开放式地提出相同问题时，开源模型的准确性下降了18-25分，151个答案中没有一个完全正确，而Gemini的下降幅度为12分。进一步的检查排除了明显的无辜解释，添加字幕、使用最具信息量的帧或将帧数加倍都未能改善角色追踪，因此瓶颈并不在于模型看到多少视频，而在于它如何将这些视频与问题中提到的人联系起来。我们发布了一个诊断工具包，用于审计这些基准得分实际测量的内容。
cs.CV / 122 / 2607.11081
Controlling Motion Transfer in Diffusion Transformers via Attention Heads
通过注意力头控制扩散变换器中的运动转移
Abstract
Diffusion Transformers (DiTs) have advanced video generation with high-quality, temporally coherent results. However, extending them to motion transfer, which requires following reference motion while aligning with a target prompt, remains challenging due to limited understanding of motion and structure representations within DiTs. We analyze video DiTs at the attention-head level and identify distinct heads specialized for motion and spatial structure. Based on this insight, we propose a head-aware controllable motion transfer framework that requires no parameter updates. Our method refines motion cues from motion-specialized heads via semantic correspondence guidance and preserves structure through selective feature injection. This head-level control not only enables accurate motion transfer but also provides an interpretable foundation for controllable video generation with DiTs.
Chinese Translation
扩散变换器（Diffusion Transformers, DiTs）在高质量、时间一致性的视频生成方面取得了进展。然而，将其扩展到运动转移仍然具有挑战性，因为这需要在遵循参考运动的同时与目标提示对齐，而对DiTs中运动和结构表示的理解有限。我们在注意力头层面分析视频DiTs，并识别出专门用于运动和空间结构的不同头。基于这一洞察，我们提出了一种头感知的可控运动转移框架，该框架无需参数更新。我们的方法通过语义对应引导从运动专用头中提炼运动线索，并通过选择性特征注入来保持结构。这种头级控制不仅实现了准确的运动转移，还为使用DiTs进行可控视频生成提供了可解释的基础。
cs.CV / 123 / 2607.11088
CUST: Clustered Unit-level Similarity Transformer for Lightweight Image Super-Resolution
CUST：用于轻量级图像超分辨率的聚类单元级相似性变换器
Abstract
Recently, Vision Transformer (ViT)-based models have exhibited remarkable performance in image super-resolution. However, the quadratic computational complexity of ViTs with respect to spatial resolution severely constrains their efficiency, leading to high latency and massive memory consumption. To alleviate this, various window-based attention mechanisms have been proposed; yet, they inherently compromise the long-range dependency modeling that is the primary advantage of ViTs. To overcome these limitations, we propose the Clustered Unit-level Similarity Transformer (CUST), a novel architecture that efficiently integrates global and local information. Specifically, CUST enables each patch to aggregate and attend to similar patches within a broadened regional scope outside its local window, thereby capturing extensive contextual understanding. Furthermore, it employs overlapping attention windows to capture local dependencies, while explicitly extracting high-frequency details by computing the residual difference between the original features and their downsampled-upsampled counterparts. Comprehensive experiments demonstrate that our proposed model achieves a practical balance between computational efficiency and restoration performance. It achieves a lower memory footprint and faster inference speed compared to recent global context or lightweight models under realistic constraints. Code is available at [https://github.com/jwgdmkj/CUST].
Chinese Translation
近年来，基于视觉变换器（Vision Transformer, ViT）模型在图像超分辨率方面表现出色。然而，ViT在空间分辨率方面的二次计算复杂度严重限制了其效率，导致高延迟和巨大的内存消耗。为了解决这一问题，提出了各种基于窗口的注意力机制；然而，它们在本质上妥协了长距离依赖建模，这是ViT的主要优势。为克服这些限制，我们提出了聚类单元级相似性变换器（Clustered Unit-level Similarity Transformer, CUST），这是一种新颖的架构，能够高效地整合全局和局部信息。具体而言，CUST使每个补丁能够在其局部窗口外的扩展区域内聚合并关注相似补丁，从而捕获广泛的上下文理解。此外，它采用重叠的注意力窗口来捕获局部依赖，同时通过计算原始特征与其下采样-上采样对应物之间的残差差异，明确提取高频细节。全面的实验表明，我们提出的模型在计算效率和恢复性能之间实现了实用的平衡。与最近的全局上下文或轻量级模型相比，在现实约束下，它实现了更低的内存占用和更快的推理速度。代码可在 [https://github.com/jwgdmkj/CUST] 获取。
cs.CV / 124 / 2607.11090
Why Low-Light Cameras Go Color Blind: Removing Color Bias in Raw Denoising
为什么低光照相机会失去色彩感知：消除原始去噪中的色彩偏差
Abstract
Raw images inherently suffer from noise due to the stochastic nature of light and sensor hardware imperfections. As real photon counts fall, the ratio of this noise to the signal degrades; consequently, for low-light conditions, robust denoising is especially vital for high-quality results. While recent data-driven methods achieve strong performance, they typically rely on large-scale noisy-clean image pairs that are costly and difficult to collect. Alternatively, parametric noise models can generate synthetic training data, but this necessitates precise camera calibration, which is often impractical for unknown devices. In this work, we propose a camera-agnostic, calibration-free paradigm for low-light raw denoising. We identify that color bias from black-level error is a primary source of performance degradation and causes severe color shifts. To mitigate this, we introduce a bias estimator network that predicts the black-level error as a global feature of the noisy input. We evaluate our approach across the ELD, SID, and LRID datasets, demonstrating superior performance among blind denoisers, particularly in terms of color correction. In many cases, we are competitive with-or can even surpass-methods with stronger supervision. Furthermore, we reveal that the widely used SIDD dataset contains significant color bias in its ground-truth images, which yields unrealistic color reproduction in trained models. We introduce a new ground-truth extraction framework to resolve this issue and provide a benchmark of existing methods on the corrected dataset.
Chinese Translation
原始图像由于光的随机特性和传感器硬件缺陷而固有地受到噪声的影响。随着真实光子计数的下降，这种噪声与信号的比率恶化；因此，在低光照条件下，稳健的去噪对于高质量结果尤为重要。尽管最近的数据驱动方法取得了良好的性能，但它们通常依赖于大规模的噪声-干净图像对，这些图像对的收集成本高且困难。另一种选择是参数噪声模型可以生成合成训练数据，但这需要精确的相机校准，而对于未知设备来说，这通常是不切实际的。在本研究中，我们提出了一种与相机无关的、无需校准的低光原始去噪范式。我们发现，黑电平误差导致的色彩偏差是性能下降的主要来源，并造成严重的色彩偏移。为此，我们引入了一种偏差估计网络，该网络将黑电平误差预测为噪声输入的全局特征。我们在ELD、SID和LRID数据集上评估了我们的方法，显示出在盲去噪器中具有优越的性能，特别是在色彩校正方面。在许多情况下，我们的表现与具有更强监督的方法相当，甚至可以超越它们。此外，我们揭示了广泛使用的SIDD数据集在其真实图像中存在显著的色彩偏差，这导致训练模型中的色彩再现不切实际。我们引入了一种新的真实图像提取框架来解决此问题，并在修正后的数据集上提供现有方法的基准。
cs.CV / 125 / 2607.11096
Difference-Driven Gating: Adaptive Feature Fusion for U-Net Decoder
差异驱动门控：U-Net 解码器的自适应特征融合
Abstract
The U-Net style models have been widely used in many applications. A critical step in these models is to reconstruct the lower-level features using a top-down decoder. This reconstruction requires precise fusion of high-level semantics and low-level details. Existing attention-based fusion methods typically derive attention weights from the top-down decoder features (global) alone or the correlation between the top-down decoder features and the bottom-up encoder features (local), then modulate the encoder features using these weights. In this work, we explore a different paradigm: deriving attention weights from the difference between the two feature streams. To this end, we propose two difference-based gating approaches: Feature-difference gating (FDG), which directly uses the absolute difference between global and local features to generate adaptive gating maps, and Entropy-difference gating (EDG), which measures the representational certainty of each stream via information entropy and uses their signed entropy difference to derive the attention weights. Both methods produce coupled gating maps that simultaneously modulate the global and local features. Experiments on different tasks including medical image segmentation, remote sensing image cloud removal and speech separation showed that both methods outperformed existing attention-based fusion methods, and EDG performed better. The results suggested a new paradigm for multi-scale feature fusion in the U-Net style structures.
Chinese Translation
U-Net 风格的模型已广泛应用于许多领域。这些模型中的一个关键步骤是使用自上而下的解码器重建低层特征。这一重建过程需要高层语义与低层细节的精确融合。现有的基于注意力的融合方法通常仅从自上而下解码器特征（全局）或自上而下解码器特征与自下而上编码器特征之间的相关性（局部）中推导注意力权重，然后利用这些权重调节编码器特征。在本研究中，我们探索了一种不同的范式：从两个特征流之间的差异中推导注意力权重。为此，我们提出了两种基于差异的门控方法：特征差异门控（Feature-difference gating, FDG），直接使用全局特征与局部特征之间的绝对差异生成自适应门控图；熵差异门控（Entropy-difference gating, EDG），通过信息熵测量每个流的表征确定性，并利用它们的符号熵差推导注意力权重。这两种方法生成的耦合门控图同时调节全局和局部特征。在包括医学图像分割、遥感图像云去除和语音分离等不同任务上的实验表明，这两种方法均优于现有的基于注意力的融合方法，且 EDG 的表现更佳。结果提示了一种新的多尺度特征融合范式，适用于 U-Net 风格的结构。
cs.CV / 126 / 2607.11097
Revisiting Matching Response and Swept Feature Volumes for Wide-baseline Omnidirectional Stereo
重新审视宽基线全向立体视觉中的匹配响应与扫掠特征体积
Abstract
In this paper, we propose a training strategy for confidence estimation in omnidirectional stereo, targeting the ambiguous matches that frequently occur in wide-baseline setups. Reinterpreting the matching responses produced by the 3D encoder decoder block, we show that their expectation values provide intrinsic confidence signals. Building on this, our method directly penalizes ambiguous responses without auxiliary heads, multi-pass inference, or additional modules, resulting in more efficient and generalized predictions. Beyond confidence, we introduce swept feature volume resampling, where response features produced by 3D CNNs are resampled using regressed positive matching indices and then processed by 2D CNNs to predict meta-information such as surface normals. This joint learning introduces auxiliary geometric regularization and improves depth coherence by leveraging additional contextual cues during response aggregation stage. Experimental results demonstrate that our approach enhances both confidence estimation and surface normal prediction while maintaining deployment practicality for autonomous mobility applications.
Chinese Translation
在本文中，我们提出了一种用于全向立体视觉中置信度估计的训练策略，旨在解决宽基线设置中经常出现的模糊匹配问题。通过重新解释3D编码解码块产生的匹配响应，我们展示了其期望值提供了内在的置信度信号。在此基础上，我们的方法直接惩罚模糊响应，无需辅助头、多次推理或额外模块，从而实现更高效和更具普适性的预测。除了置信度外，我们引入了扫掠特征体积重采样，其中由3D卷积神经网络（3D CNN）生成的响应特征使用回归的正匹配索引进行重采样，然后由2D卷积神经网络（2D CNN）处理，以预测表面法线等元信息。这种联合学习引入了辅助几何正则化，并通过在响应聚合阶段利用额外的上下文线索来改善深度一致性。实验结果表明，我们的方法在增强置信度估计和表面法线预测的同时，保持了在自主移动应用中的部署实用性。
cs.CV / 127 / 2607.11104
FlowPET: Physics-Informed Symplectic Flow Matching for Low-Count PET Reconstruction
FlowPET：基于物理知识的辛流匹配用于低计数正电子发射断层成像重建
Abstract
Low-count Positron Emission Tomography (PET) reconstruction is severely hindered by the dissipative nature of prevailing generative models, where the inherent phase-space contraction leads to the numerical extinction (``wash-out'') of weak but diagnostically critical lesion signals. To overcome this geometric limitation, we propose \textbf{FlowPET}, a physics-informed framework that reformulates reconstruction as volume-preserving transport in a symplectic phase space. By parameterizing the posterior dynamics via a Separable Hamiltonian System, our approach guarantees a divergence-free vector field by construction, theoretically immunizing weak signals against probability mass collapse. To steer this conservative flow, we introduce conjugate boundary conditions based on the Range-Null space decomposition of the PET operator; this strictly enforces data consistency in the range space while confining stochastic uncertainty injection to the unobserved null space. We train the model via symplectic flow matching and perform inference using a symplectic leapfrog integrator. Extensive experiments on BrainWeb, clinical pediatric, and UDPET datasets demonstrate that \textbf{FlowPET} not only surpasses state-of-the-art deterministic and stochastic baselines in SSIM and PSNR but, more crucially, exhibits superior recovery of low-contrast lesions. The results confirm that imposing Hamiltonian structural constraints offers a robust geometric safeguard for medical inverse problems in high-noise regimes.
Chinese Translation
低计数正电子发射断层成像（PET）重建受到现有生成模型的耗散特性的严重制约，其固有的相空间收缩导致弱但在诊断上至关重要的病变信号的数值消失（“洗出”）。为克服这一几何限制，我们提出了 extbf{FlowPET}，一个基于物理知识的框架，将重建重新表述为在辛相空间中的体积守恒传输。通过将后验动态参数化为可分离哈密顿系统，我们的方法在构造上保证了无散度向量场，从理论上使弱信号免受概率质量崩溃的影响。为了引导这一保守流动，我们基于PET算子的范围-零空间分解引入共轭边界条件；这严格执行了范围空间中的数据一致性，同时将随机不确定性注入限制在未观察到的零空间。我们通过辛流匹配训练模型，并使用辛跳蛙积分器进行推断。在BrainWeb、临床儿科和UDPET数据集上的大量实验表明， extbf{FlowPET}不仅在结构相似性指数（SSIM）和峰值信噪比（PSNR）上超越了最先进的确定性和随机基线，更重要的是，在低对比度病变的恢复方面表现出更优的效果。结果证实，施加哈密顿结构约束为高噪声环境中的医学逆问题提供了一种稳健的几何保障。
cs.CV / 128 / 2607.11106
Beyond the Eye: Efficient Multimodal Reasoning via Self-Regulated Implicit Visual Tools
超越视觉：通过自我调节的隐式视觉工具实现高效的多模态推理
Abstract
Recent multimodal large language models (MLLMs) have made remarkable progress on fine-grained perception tasks under the "Thinking with Images" (TwI) paradigm by iteratively performing various visual tool operations. However, this paradigm relies heavily on frequent external tool calls and repeated image re-encoding, which leads to substantial computational overhead and inference latency. To address these issues, we propose Beyond the Eye (BEE), a novel implicit visual tool paradigm centered on self-regulated capability. BEE directly incorporates visual tool invocation behaviors into the training objective and encourages the model to develop a self-regulated invocation mechanism. This design enables the model to adaptively balance internal knowledge and implicit tools, avoiding redundant tool usage while substantially reducing inference latency. Specifically, BEE involves a two-stage training process: (1) Formalized Chain-of-Thought (CoT) Supervised Fine-tuning (SFT). We construct CoT trajectories with structured tool slots and mixed invocation states. This stage activates the model's implicit tool representations and adaptive switching capability. (2) Self-regulated Reward-Driven Alignment. To address redundant tool usage caused by ambiguous cognitive boundaries, we first introduce the Net Tool Gain (NTG) metric to quantify this phenomenon. Based on this observation, we further propose a self-regulated reward mechanism. This mechanism penalizes ineffective tool dependency and encourages the model to perform knowledge routing, ensuring that implicit tools are invoked only when the model's internal knowledge is insufficient. BEE achieves state-of-the-art performance in fine-grained visual perception while remaining competitive in general reasoning tasks and achieving substantial gains in inference efficiency.
Chinese Translation
近期的多模态大型语言模型（MLLMs）在“以图像思考”（TwI）范式下的细粒度感知任务上取得了显著进展，通过迭代执行各种视觉工具操作。然而，这一范式过于依赖频繁的外部工具调用和重复的图像重新编码，导致了显著的计算开销和推理延迟。为了解决这些问题，我们提出了超越视觉（BEE），一种以自我调节能力为中心的新型隐式视觉工具范式。BEE将视觉工具调用行为直接纳入训练目标，并鼓励模型发展自我调节的调用机制。这一设计使模型能够自适应地平衡内部知识和隐式工具，避免冗余工具使用，同时显著减少推理延迟。具体而言，BEE涉及两个阶段的训练过程：（1）形式化的思维链（CoT）监督微调（SFT）。我们构建了带有结构化工具槽和混合调用状态的CoT轨迹。此阶段激活了模型的隐式工具表示和自适应切换能力。（2）自我调节的奖励驱动对齐。为了解决由于模糊的认知边界导致的冗余工具使用，我们首先引入了净工具增益（NTG）指标来量化这一现象。在此基础上，我们进一步提出了一种自我调节的奖励机制。该机制惩罚无效的工具依赖，并鼓励模型进行知识路由，确保仅在模型的内部知识不足时才调用隐式工具。BEE在细粒度视觉感知方面实现了最先进的性能，同时在一般推理任务中保持竞争力，并在推理效率上取得了显著提升。
cs.CV / 129 / 2607.11118
GHOST: Geometry-Guided Hallucination of Opaque Surface Textures
GHOST：几何引导的非透明表面纹理幻觉
Abstract
Transparent objects pose a fundamental challenge for depth estimation and 3D reconstruction due to their violation of Lambertian assumptions, leading to severe geometry degradation in downstream tasks. To address this, we propose a novel geometry-guided preprocessing framework \textbf{GHOST} that leverages visual foundation models to transform transparent regions into opaque, structurally consistent representations without requiring downstream model retraining. Specifically, our pipeline utilizes (1) \textbf{TransDINO} and (2) \textbf{TransDecomp} to disentangle masks and transparency physical properties, while (3) \textbf{DAF-Net} recovers surface normal priors to encode geometric curvature. Subsequently, (4) \textbf{GeoSemTransNet} integrates these multi-modal cues to synthesize a texture-rich opaque RGB image that preserves the transparent object's 3D structure. Extensive experiments demonstrate that our method significantly enhances the accuracy of state-of-the-art depth estimation and reconstruction models on transparent objects by restoring essential photometric cues.
Chinese Translation
透明物体由于违反了朗伯假设，给深度估计和三维重建带来了根本性的挑战，导致下游任务中的几何退化严重。为了解决这一问题，我们提出了一种新颖的几何引导预处理框架 extbf{GHOST}，该框架利用视觉基础模型将透明区域转换为不透明且结构一致的表示，而无需对下游模型进行重新训练。具体而言，我们的流程利用 (1) extbf{TransDINO} 和 (2) extbf{TransDecomp} 来解耦掩膜和透明物理属性，同时 (3) extbf{DAF-Net} 恢复表面法线先验以编码几何曲率。随后，(4) extbf{GeoSemTransNet} 整合这些多模态线索，合成出一种丰富纹理的不透明 RGB 图像，保留透明物体的三维结构。大量实验表明，我们的方法显著提高了在透明物体上最先进的深度估计和重建模型的准确性，通过恢复重要的光度线索。
cs.CV / 130 / 2607.11120
Simple Features and Honest Calibration for Ambivalence and Hesitancy Recognition in Video
视频中模棱两可与犹豫识别的简单特征与诚实校准
Abstract
We address ambivalence and hesitancy (A/H) recognition in the ABAW 2026 BAH Challenge: given a short interview video, predict whether the person shows signs of A/H. Our system combines affect-specialised text, audio, and visual representations with a small set of readable linguistic hesitation cues, fused by a reliability gate we call Affective Marker Fusion (AMF), and finished with a simple AP-weighted ensemble at a fixed decision threshold. We also introduce \emph{ASR-erased time}: speech recognisers delete fillers and hesitation pauses from the transcript, but the chunk timestamps keep the time those events took, and sixteen features built from these gaps form the strongest and most independent non-verbal channel we measured (AP $0.718$, correlation $0.11$--$0.36$ with all other members). Across controlled experiments we find three things: cross-modal conflict design does not reliably help on BAH; language is by far the strongest channel while affect-specialised audio is a useful second; and calibration matters more than architecture. Fitting ensemble weights and a threshold on the small validation split overfits: it scores $0.741$ macro-F1 on validation but only $0.690$ on the untouched test set. AP-weighting at a fixed threshold instead reaches $\mathbf{0.731}$ on test.
Chinese Translation
我们在ABAW 2026 BAH挑战中解决模棱两可与犹豫（A/H）识别的问题：给定一段简短的访谈视频，预测该人是否表现出A/H的迹象。我们的系统结合了情感专用的文本、音频和视觉表征，以及一小组可读的语言犹豫提示，通过我们称之为情感标记融合（Affective Marker Fusion, AMF）的可靠性门进行融合，并以简单的AP加权集成在固定决策阈值下完成。我们还引入了 extit{ASR-erased time}：语音识别器从转录中删除填充词和犹豫停顿，但时间戳保留了这些事件所花费的时间，基于这些间隙构建的十六个特征形成了我们测量的最强且最独立的非语言通道（AP $0.718$, 与所有其他成员的相关性为 $0.11$--$0.36$）。在受控实验中，我们发现三件事：跨模态冲突设计在BAH上并不可靠地有效；语言是迄今为止最强的通道，而情感专用音频是一个有用的第二通道；校准比架构更为重要。在小的验证集上拟合集成权重和阈值会导致过拟合：在验证集上得分为 $0.741$ 的宏F1，但在未触及的测试集上仅为 $0.690$。相反，在固定阈值下的AP加权在测试集上达到了 $ extbf{0.731}$。
cs.CV / 131 / 2607.11168
SISA-Rec: A Semantically Integrated Sequential Recommender with Contrastive Alignment
SISA-Rec：一种具有对比对齐的语义集成序列推荐系统
Abstract
Recommendation systems help users recommend relevant items from a large collection of choices. Present work on transformer-based sequential recommendation learns user preferences from interaction logs, but it mostly focuses on item identifiers and doesn't fully use the semantic meaning of items. This limitation becomes a major challenge in sparse and cold-start scenarios where historical interaction data is limited. To solve this problem, we introduce SISA-Rec (Semantically Integrated Sequential Recommendation), a transformer-based framework that embeds semantic context directly into sequential modeling. Our approach fuses item ID embeddings with BERT-based text embeddings via a gated fusion module, injects semantic similarity into the self-attention mechanism, and leverages an attention-based aggregation module to construct comprehensive user representations. Finally, a joint learning objective which combines Bayesian Personalized Ranking (BPR) and contrastive alignment loss, aligns the underlying behavioral and semantic spaces. Experiments were conducted on the two highly sparse Amazon Beauty and Amazon Toys \& Games datasets, both having 99.93\% sparsity. The results show that SISA-Rec outperforms state-of-the-art baseline models across all evaluation metrics. Compared with the BERT4Rec \cite{petrov2022systematic}, SISA-Rec improves HR@10 by 16.6\% and NDCG@10 by 10.3\% on Amazon Beauty, and HR@10 by 23.1\% and NDCG@10 by 17.9\% on Amazon Toys \& Games. Cold-start analysis further shows that the proposed model achieves the largest improvements for users with limited interaction historical records. This showcases the value of semantic information when user behavior data is scarce. Overall, the results demonstrate that integrating semantic information into the attention mechanism leads to more accurate and reliable recommendations.
Chinese Translation
推荐系统帮助用户从大量选择中推荐相关项目。目前基于变换器的序列推荐研究从交互日志中学习用户偏好，但主要关注项目标识符，未充分利用项目的语义含义。这一局限性在稀疏和冷启动场景中成为主要挑战，因为历史交互数据有限。为了解决这个问题，我们提出了SISA-Rec（语义集成序列推荐），这是一种基于变换器的框架，直接将语义上下文嵌入到序列建模中。我们的方法通过门控融合模块将项目ID嵌入与基于BERT的文本嵌入融合，将语义相似性注入自注意力机制，并利用基于注意力的聚合模块构建全面的用户表示。最后，结合贝叶斯个性化排名（BPR）和对比对齐损失的联合学习目标，使潜在的行为空间和语义空间对齐。在两个高度稀疏的亚马逊美容和亚马逊玩具与游戏数据集上进行了实验，二者的稀疏度均为99.93%。结果表明，SISA-Rec在所有评估指标上均优于最先进的基线模型。与BERT4Rec相比，SISA-Rec在亚马逊美容上提高了HR@10 16.6%和NDCG@10 10.3%，在亚马逊玩具与游戏上提高了HR@10 23.1%和NDCG@10 17.9%。冷启动分析进一步表明，所提出的模型为交互历史记录有限的用户带来了最大的改进。这展示了在用户行为数据稀缺时语义信息的价值。总体而言，结果表明将语义信息集成到注意力机制中可以实现更准确和可靠的推荐。
cs.CV / 132 / 2607.11170
TC-MAF: Train-Calibrated Bounded Multi-Evidence Fusion for Multimodal Industrial Anomaly Detection
TC-MAF：训练校准的有界多证据融合用于多模态工业异常检测
Abstract
Multimodal anomaly detection benefits from complementary RGB and 3D evidence, yet auxiliary RGB reconstruction is not equally reliable across product categories and class-wise test-time policy selection is usually unavailable. We propose TC-MAF, a base-anchored multi-evidence fusion design that combines a multimodal detector, complementary Dinomaly evidence, and a small cross-modal consistency cue under one fixed pixel-level fusion formula. A lightweight training-dispersion confidence (TDC) term scales auxiliary participation using only normal training statistics. On MVTec-3D, TC-MAF reaches 0.979 image-level AUROC and 0.990 pixel-level AUPRO, achieving the best mean results on both detection and localization among the compared multimodal methods. Systematic ablations show that the fusion structure itself is the dominant factor, while TDC provides a smaller but reproducible calibration gain over no calibration or arbitrary calibration. Additional experiments show that the same design remains effective under a pooled-statistics variant, auxiliary-branch and backbone substitutions, few-shot settings, a missing-3D setting, and cross-dataset evaluation on Eyecandies. Code is available at https://anonymous.4open.science/r/TC_MAF-C3BB.
Chinese Translation
多模态异常检测受益于互补的RGB和3D证据，然而辅助RGB重建在不同产品类别中的可靠性并不相同，且通常缺乏按类别选择的测试时策略。我们提出了TC-MAF，一种基于基准的多证据融合设计，它将多模态检测器、互补的Dinomaly证据以及一个小型跨模态一致性线索结合在一个固定的像素级融合公式下。一个轻量级的训练分散置信度（TDC）项仅使用正常训练统计数据来调整辅助参与度。在MVTec-3D数据集上，TC-MAF达到了0.979的图像级AUROC和0.990的像素级AUPRO，在比较的多模态方法中实现了检测和定位的最佳平均结果。系统的消融实验表明，融合结构本身是主导因素，而TDC提供了比没有校准或任意校准更小但可重复的校准增益。额外实验表明，在汇总统计变体、辅助分支和主干替换、少样本设置、缺失3D设置以及在Eyecandies上的跨数据集评估中，相同设计仍然有效。代码可在 https://anonymous.4open.science/r/TC_MAF-C3BB 获取。
cs.CV / 133 / 2607.11173
When Depth Is Better Told Than Shown: Depth-Ordinal Prompting for Vision-Language Spatial Reasoning
深度以叙述方式优于展示：用于视觉-语言空间推理的深度序数提示
Abstract
Vision-language models (VLMs) are expected to reason about physical space -- which object is closer, what lies behind what, and how objects are arranged in 3D -- yet they still struggle with such spatial judgments. A natural remedy is to show the model a depth map, but we find that this can make performance worse. We show that depth is not absent: it reaches the language model, but becomes difficult to access for downstream reasoning, while rendered pseudo-depth maps act as noisy auxiliary images that frozen VLMs cannot easily regulate. We propose Depth-Ordinal Prompting (DOP), a training-free method that converts monocular depth into a single question-targeted ordinal text cue at the queried objects, without adding a depth image, training a module, injecting features, or using labels. Our key finding is form dependence: the same depth signal can hurt when shown as an image but help when told as text.Across benchmarks, models, and depth estimators, DOP improves spatial reasoning when pseudo-depth provides reliable object-level ordering and remains largely neutral in strong original-image regimes. It is also competitive with the strongest training-free depth-prompting alternative while being simpler and more targeted.
Chinese Translation
视觉-语言模型（VLMs）被期望能够推理物理空间——哪个物体更近，什么在什么后面，以及物体在三维空间中的排列——然而它们在这些空间判断上仍然存在困难。一个自然的解决方案是向模型展示深度图，但我们发现这可能会导致性能下降。我们表明深度并非缺失：它能够到达语言模型，但在下游推理中变得难以访问，而渲染的伪深度图则充当了噪声辅助图像，冻结的VLMs难以进行有效调节。我们提出了深度序数提示（Depth-Ordinal Prompting, DOP），这是一种无训练的方法，将单目深度转换为针对查询对象的单一问题导向的序数文本提示，而无需添加深度图像、训练模块、注入特征或使用标签。我们的关键发现是形式依赖性：相同的深度信号在以图像形式展示时可能会造成负面影响，但以文本形式叙述时则可能提供帮助。在各项基准测试、模型和深度估计器中，当伪深度提供可靠的对象级排序时，DOP改善了空间推理，并在强原始图像条件下保持了相对中立。它在与最强的无训练深度提示替代方案的竞争中表现出色，同时更简单且更具针对性。
cs.CV / 134 / 2607.11192
GDP.pdf: Benchmarking Grounded Multimodal Reasoning over Professional PDF Documents
GDP.pdf：针对专业PDF文档的基础多模态推理基准测试
Abstract
A large share of day-to-day work in professional domains happens inside PDF files: benefits packets, leases, datasheets, clinical guidelines, construction plans. Benchmarks for document AI have generally measured the required capabilities in isolation: OCR, layout analysis, chart reasoning, table QA, document VQA. A high score on any one of them does not necessarily reveal whether a model can answer a realistic question that someone in the field would actually ask about a specific PDF. GDP.pdf is a benchmark built to measure this directly. It consists of question-document pairs authored by working professionals in ten fields, and a candidate question was kept only when at least two frontier multimodal models failed it in a way that mattered: a wrong answer, missed decisive evidence, or a fabricated claim, rather than a superficial difference such as style. Each item comes with a rubric of atomic criteria, so we can report a graded rubric score as well as a strict task-level pass rate, and each item is tagged against a taxonomy of eleven capabilities in three tiers, spanning text extraction and grounding, table and chart comprehension, cross-referencing, spatial reasoning, and abstention on unsupported queries. We evaluated seven frontier models on the 100-item benchmark. The best model passed only 15% of the items and the worst passed 1%. Most errors trace back to a small set of recurring loss patterns: misaligned tables, misread charts, skipped footnotes and exclusions, miscounted floor-plan symbols, scan noise, and amendments that supersede earlier text. The full 100-item benchmark is publicly available at https://huggingface.co/datasets/surgeai/GDP.pdf
Chinese Translation
在专业领域的日常工作中，大量工作发生在PDF文件中：福利包、租赁协议、数据表、临床指南、施工图纸。文档人工智能的基准测试通常是孤立地测量所需的能力：光学字符识别（OCR）、布局分析、图表推理、表格问答（QA）、文档视觉问答（VQA）。在这些能力中的任何一项高分并不一定能揭示一个模型是否能够回答某个领域内专业人士对特定PDF提出的现实问题。GDP.pdf是一个旨在直接测量这一点的基准测试。它由十个领域的在职专业人士撰写的问题-文档对组成，只有当至少两个前沿多模态模型以重要的方式未能回答候选问题时，该问题才被保留：错误答案、遗漏关键证据或虚假声明，而不是诸如风格等表面差异。每个项目都有一套原子标准的评分标准，因此我们可以报告分级评分以及严格的任务级通过率，每个项目还根据涵盖文本提取与基础、表格与图表理解、交叉引用、空间推理以及对不支持查询的回避等三层次的十一种能力的分类法进行标记。我们在100项基准测试上评估了七个前沿模型。表现最好的模型仅通过了15%的项目，而表现最差的模型通过了1%。大多数错误可追溯到一小部分重复的损失模式：对齐不当的表格、错误读取的图表、遗漏的脚注和排除项、错误计数的平面图符号、扫描噪声以及取代早期文本的修订。完整的100项基准测试可在https://huggingface.co/datasets/surgeai/GDP.pdf上公开获取。
cs.CV / 135 / 2607.11196
Slot-RAE: Streamlining Object-Centric Learning via Direct Representation Auto-Encoders
Slot-RAE：通过直接表示自编码器简化面向对象的学习
Abstract
Deploying object-centric models for real-world scene understanding typically requires complex pipelines to achieve both robust scene decomposition and high-fidelity generation. Recent diffusion-based approaches have improved visual quality, but they almost universally rely on heavy, pretrained generative priors (e.g., Stable Diffusion) and external VAE latent spaces. In this paper, we propose Slot-RAE, a much simpler, fully integrated framework that operates directly within the continuous semantic feature space of visual foundation models (e.g., DINOv3). Slot-RAE employs a feature-space diffusion process using a Diffusion Transformer (DiT) decoder and a Representation Alignment (REPA) head. Unlike existing diffusion-based objectcentric methods that rely heavily on subsidized text-toimage priors, the generative core of Slot-RAE (Slot Attention and the DiT) is trained from scratch within the frozen VFM feature space. This eliminates the need for VAE bottlenecks and task-agnostic generative pre-training. Experiments on the COCO dataset demonstrate that despite its architectural simplicity, Slot-RAE achieves state-of-the-art results. It delivers comparable unsupervised object discovery, higher-fidelity image reconstruction, and robust zero-shot compositionality, all while being significantly faster and more computationally efficient than existing object-centric latent diffusion models.
Chinese Translation
在现实场景理解中部署面向对象的模型通常需要复杂的流程，以实现稳健的场景分解和高保真生成。最近的基于扩散的方法改善了视觉质量，但几乎普遍依赖于重型的预训练生成先验（例如，Stable Diffusion）和外部变分自编码器（VAE）潜在空间。本文提出了Slot-RAE，一个更简单、完全集成的框架，直接在视觉基础模型（例如，DINOv3）的连续语义特征空间中操作。Slot-RAE采用特征空间扩散过程，使用扩散变换器（Diffusion Transformer, DiT）解码器和表示对齐（Representation Alignment, REPA）头。与现有的基于扩散的面向对象方法重度依赖补贴的文本到图像先验不同，Slot-RAE的生成核心（Slot Attention和DiT）是在冻结的视觉基础模型特征空间内从头训练的。这消除了对VAE瓶颈和任务无关生成预训练的需求。在COCO数据集上的实验表明，尽管架构简单，Slot-RAE仍然实现了最先进的结果。它提供了可比的无监督对象发现、更高保真的图像重建和稳健的零-shot组合能力，同时在速度和计算效率上显著优于现有的面向对象的潜在扩散模型。
cs.CV / 136 / 2607.11199
DynEval: Holistic Evaluations of T2I Generative Models in the Wild
DynEval：野外T2I生成模型的整体评估
Abstract
Recent advances in text-to-image (T2I) generation have led to models capable of producing highly realistic images. Yet, reliably evaluating their outputs remains challenging, especially at scale. Existing automatic evaluators, often relying on a static prompt set, struggle to capture subtle failure modes such as partial prompt misalignment, compositional errors, or visually plausible but semantically incorrect generations. In this work, we introduce DynEval, a Dynamic Evaluation framework designed to jointly assess text-to-image alignment and image quality of T2I models. To support scalable training beyond limited human-annotated data, we construct two large datasets. First, we build GenDB, a collection of 500K prompt-image pairs generated from human-written prompts drawn from DiffusionDB using a tiered prompt-model generation strategy. Second, building upon GenDB, we construct DynEvalInstruct, a 250K instruction dataset comprising prompt-image-response triplets distilled from a structured evaluation pipeline that decomposes evaluation into text-image alignment and visual quality reasoning. Using this dataset, we perform full fine-tuning of a compact evaluator through a curriculum learning strategy to effectively distill the superior evaluation capabilities of a larger teacher vision-language model, resulting in DynEval-2B and DynEval-4B. In extensive comparisons against existing evaluators across 11 benchmarks, our evaluator achieves a higher overall correlation with human judgments. Furthermore, it provides fine-grained analysis of the capabilities and failure modes of 36 T2I models across 42 subcategories and 9 semantic dimensions.
Chinese Translation
最近在文本到图像（T2I）生成方面的进展使得模型能够生成高度逼真的图像。然而，可靠地评估其输出仍然具有挑战性，尤其是在大规模评估时。现有的自动评估器通常依赖于静态提示集，难以捕捉细微的失败模式，例如部分提示不对齐、组合错误或视觉上合理但语义上不正确的生成。在本研究中，我们提出了DynEval，一个动态评估框架，旨在共同评估T2I模型的文本到图像对齐性和图像质量。为了支持超越有限人工标注数据的可扩展训练，我们构建了两个大型数据集。首先，我们构建了GenDB，这是一个包含50万对由人类编写的提示生成的图像对的集合，这些提示来自DiffusionDB，并采用分层提示-模型生成策略。其次，在GenDB的基础上，我们构建了DynEvalInstruct，这是一个包含25万条指令的数据集，由从结构化评估管道中提炼的提示-图像-响应三元组组成，该管道将评估分解为文本-图像对齐和视觉质量推理。利用该数据集，我们通过课程学习策略对紧凑型评估器进行全面微调，以有效提炼更大教师视觉-语言模型的优越评估能力，最终得到DynEval-2B和DynEval-4B。在与现有评估器在11个基准上的广泛比较中，我们的评估器与人类判断的整体相关性更高。此外，它提供了对36个T2I模型在42个子类别和9个语义维度上的能力和失败模式的细致分析。
cs.CV / 137 / 2607.11205
Parallax Portrait Matting
视差肖像抠图
Abstract
Image matting is highly ill-posed, especially when both the foreground and background are richly textured. While single-image matting methods learn strong priors from data, they often struggle on these challenging cases. Existing approaches improve results by requiring additional signals such as green screens, polarized lighting, or clean background images, but these typically rely on specialized capture setups. We present Parallax Portrait Matting, a practical two-frame matting method that uses a second image captured with slight viewpoint change. Such a setting arises naturally in burst photography, where small camera motion induces foreground-background parallax and provides complementary observations for matting. Our pipeline estimates trimaps and foreground/background motion, then constructs aligned views for prediction. To handle imperfect motion estimation, the network uses the background-aligned pair for direct fusion and the foreground-aligned cue through cross-attention for error compensation. Experiments show that our method recovers finer details and more accurate foreground colors than strong single-image matting baselines on challenging portrait cases.
Chinese Translation
图像抠图问题高度不适定，尤其是在前景和背景都具有丰富纹理的情况下。尽管单图像抠图方法从数据中学习了强先验，但在这些具有挑战性的情况下往往表现不佳。现有方法通过要求额外信号（如绿幕、偏振光照或干净的背景图像）来改善结果，但这些通常依赖于专业的拍摄设备。我们提出了视差肖像抠图（Parallax Portrait Matting），这是一种实用的双帧抠图方法，利用稍微改变视角捕获的第二幅图像。这种设置在连拍摄影中自然出现，其中小的相机运动会引起前景与背景的视差，并为抠图提供互补的观察。我们的流程估计修整图和前景/背景运动，然后构建对齐视图进行预测。为了处理不完美的运动估计，网络使用背景对齐的图像进行直接融合，并通过交叉注意力使用前景对齐的线索进行误差补偿。实验表明，我们的方法在具有挑战性的肖像案例中恢复了比强单图像抠图基线更精细的细节和更准确的前景颜色。
cs.CV / 138 / 2607.11214
A Novel Method to Evaluate Models on Unreliable, Noisy and Inconsistent Labels: Adaptive Resolution Label Aggregation (ARLA)
一种评估不可靠、嘈杂和不一致标签模型的新方法：自适应分辨率标签聚合（ARLA）
Abstract
Labels are critical for both training and evaluating deep learning segmentation models, but are often inconsistent, noisy, or ambiguous at class boundaries. Many approaches have been developed to support training models on weak labels, but few to none currently exist to facilitate evaluating models on unreliable labels. We therefore introduce a method called "Adaptive Resolution Label Aggregation", or "ARLA", which dynamically adapts the resolution of both the label and the model prediction at inference time before the evaluation metrics are computed. We demonstrate how ARLA can be used to better analyse model behaviour with a practical application to a real flood prediction model, where ARLA was able to overcome issues with inconsistent labelling of forested areas and errors in labels within regions of heavy cloud cover. Our work presents a new approach to evaluating segmentation models, with adjustable parameters to adapt the aggregated resolution to the precision of the label or the level of label noise. Fundamentally, ARLA exploits the information encapsulated by a label but minimises the label error, extracting from the noise a clearer signal of a model's true performance.
Chinese Translation
标签对于训练和评估深度学习分割模型至关重要，但在类别边界处往往存在不一致、嘈杂或模糊的情况。虽然已经开发了许多方法来支持在弱标签上训练模型，但目前几乎没有方法能够促进在不可靠标签上评估模型。因此，我们提出了一种名为“自适应分辨率标签聚合”（Adaptive Resolution Label Aggregation，简称ARLA）的方法，该方法在评估指标计算之前，动态调整标签和模型预测的分辨率。我们展示了ARLA如何用于更好地分析模型行为，并在实际应用于真实的洪水预测模型中，ARLA能够克服森林地区标签不一致和重云覆盖区域标签错误的问题。我们的工作提出了一种新的评估分割模型的方法，具有可调参数以适应聚合分辨率与标签精度或标签噪声水平。根本上，ARLA利用标签所封装的信息，但最小化标签错误，从噪声中提取出模型真实性能的更清晰信号。
cs.CV / 139 / 2607.11221
HandFlow: Fully Generative 4D Hand Recovery with Flow Matching
HandFlow：基于流匹配的全生成4D手部恢复
Abstract
Accurate monocular 4D hand reconstruction remains challenging. Per-frame discriminative regressors lack temporal context and often produce jittery predictions. Temporal models improve consistency by aggregating information across frames, but they are typically deterministic regressors, making them vulnerable to ambiguous observations caused by occlusion and motion blur. Generative modeling offers a natural alternative by learning a prior over plausible hand motion sequences, enabling coherent hand-state recovery when visual evidence is incomplete or unreliable. Motivated by this observation, we present HandFlow, a fully generative flow-matching framework for temporally coherent 3D hand pose and shape estimation from monocular video. Given visual and skeletal observations, HandFlow denoises an entire temporal window of MANO parameters through a single ODE integration. To support this, we use a Flux-style dual-stream transformer that attends across the full sequence to capture long-range dependencies without autoregressive decoding, and a confidence-aware continuous masking mechanism that blends observed features with learnable mask tokens to handle noisy or missing observations. Experiments on DexYCB and HOT3D show that HandFlow achieves state-of-the-art performance, with particularly large gains in world-space accuracy and temporal smoothness. It reduces world-space pose error by over 30% compared with the strongest baseline and achieves the lowest acceleration error among all evaluated methods, while remaining competitive in per-frame pose accuracy. Moreover, on a single GPU HandFlow reconstructs a 150-frame sequence at 47 fps, about 12x faster than the fastest prior video-based method, with reconstruction itself accounting for only a small fraction of the end-to-end latency.
Chinese Translation
准确的单目4D手部重建仍然具有挑战性。每帧的判别回归器缺乏时间上下文，常常产生抖动的预测。时间模型通过跨帧聚合信息来提高一致性，但它们通常是确定性回归器，使其易受遮挡和运动模糊造成的模糊观测的影响。生成建模提供了一种自然的替代方案，通过学习合理的手部运动序列的先验，在视觉证据不完整或不可靠时实现连贯的手部状态恢复。基于这一观察，我们提出了HandFlow，一种完全生成的流匹配框架，用于从单目视频中进行时间一致的3D手部姿态和形状估计。给定视觉和骨骼观测，HandFlow通过单次常微分方程（ODE）积分对整个时间窗口的MANO参数进行去噪。为此，我们使用了一种Flux风格的双流变换器，它跨整个序列进行注意力机制，以捕捉长距离依赖关系，而无需自回归解码，并且采用了一种基于置信度的连续掩蔽机制，将观测特征与可学习的掩蔽标记融合，以处理噪声或缺失的观测。DexYCB和HOT3D上的实验表明，HandFlow实现了最先进的性能，特别是在世界空间准确性和时间平滑性方面取得了显著提升。与最强基线相比，它将世界空间姿态误差降低了超过30%，并在所有评估方法中实现了最低的加速度误差，同时在每帧姿态准确性方面保持竞争力。此外，在单个GPU上，HandFlow以47帧每秒的速度重建150帧序列，速度约为最快先前视频方法的12倍，而重建本身仅占端到端延迟的一小部分。
cs.CV / 140 / 2607.11233
Structure-Detail Decoupled Autoregressive Generation for Fast and High-Fidelity Virtual Try-On
结构-细节解耦自回归生成用于快速且高保真的虚拟试穿
Abstract
Virtual try-on (VTON) is a bi-conditional image generation problem that requires not only accurate person preservation but also faithful garment deformation and detail synthesis. Diffusion-based VTON methods can jointly model these factors in a compressed latent space, but suffer from high-frequency detail loss due to inherent latent compression, even with costly multi-step denoising. Recent visual autoregressive (VAR) models offer a promising alternative for high-quality generation with faster inference, yet remain unexplored for VTON due to the lack of effective bi-conditioning mechanisms. To bridge this gap, we first introduce VAR-VTON, a VAR-based VTON model that incorporates garment conditioning and structural guidance for efficient latent-space VTON. Despite its efficacy, latent-space generation still struggles to preserve fine-grained garment details. We argue that different VTON sub-tasks should be addressed in different representation spaces: structural synthesis such as garment warping and person layout is suited to the latent space, whereas fine-grained detail recovery should be tackled in the pixel space. Motivated by this insight, we further propose STAR-VTON, a Two-Stage AutoRegressive framework that builds upon VAR-VTON by decoupling latent-space structural synthesis from pixel-space detail recovery. Our idea is to resort to a matching-informed refiner to establish dense correspondences between the stage-one generation and the source garment to directly map fine-grained pixel-space details. Extensive experiments show that STAR-VTON achieves an impressive efficiency-fidelity trade-off: VAR-VTON runs at least $4\times$ faster than diffusion-based counterparts without degrading quality, and the pixel-space refiner effectively restores fine details and acts as a plug-and-play module that can benefit existing VTON approaches.
Chinese Translation
虚拟试穿（VTON）是一个双条件图像生成问题，不仅需要准确的人物保留，还需要忠实的服装变形和细节合成。基于扩散的 VTON 方法可以在压缩的潜在空间中联合建模这些因素，但由于固有的潜在压缩，即使经过昂贵的多步去噪，也会遭遇高频细节损失。最近的视觉自回归（VAR）模型为高质量生成提供了一个有前景的替代方案，并且推理速度更快，但由于缺乏有效的双条件机制，尚未在 VTON 中得到探索。为了解决这一问题，我们首先提出了 VAR-VTON，一种基于 VAR 的 VTON 模型，结合了服装条件和结构指导，以实现高效的潜在空间 VTON。尽管其有效性，潜在空间生成仍然难以保留细致的服装细节。我们认为，不同的 VTON 子任务应在不同的表示空间中解决：服装变形和人物布局等结构合成适合潜在空间，而细致的细节恢复应在像素空间中进行。基于这一见解，我们进一步提出了 STAR-VTON，一个两阶段自回归框架，基于 VAR-VTON，通过将潜在空间的结构合成与像素空间的细节恢复解耦。我们的想法是借助匹配信息的细化器，在第一阶段生成与源服装之间建立密集的对应关系，以直接映射细致的像素空间细节。大量实验表明，STAR-VTON 实现了令人印象深刻的效率与保真度的权衡：VAR-VTON 的运行速度至少比基于扩散的方法快 $4 imes$，且不会降低质量，而像素空间细化器有效恢复细节，并作为一个即插即用模块，可以惠及现有的 VTON 方法。
cs.CV / 141 / 2607.11255
A Nearable Soft Mat Based on Distributed Optical Fiber Sensing for Physiological Monitoring
基于分布式光纤传感的近接软垫用于生理监测
Abstract
Distributed optical fiber sensing (DOFS) combines the advantages of fiber optic sensors, including flexibility, small size, immunity to electromagnetic interference, and high metrological performance, with the capability to transform a single optical fiber into a continuous sensing element for spatially resolved mechanical measurements. Optical frequency domain reflectometry (OFDR), based on Rayleigh backscattering, enables high spatial resolution DOFS measurements, broadening the range of potential sensing applications. However, OFDR based DOFS remains largely unexplored for biomedical applications, despite the need for sensitive, spatially resolved, and conformable sensing interfaces. This study presents a soft DOFS based mat as a large-area interface for physiological monitoring. A single-mode optical fiber was embedded in a flexible silicone matrix and arranged in a serpentine layout to distribute sensing over the mat surface. With a gage pitch of 2.6 mm, the system provided 2250 sensing sites across the active area at a sampling frequency of 50 Hz. The mat was assessed on six healthy volunteers in a seated nearable configuration on the backrest of a standard office chair. The distributed output enabled two dimensional mapping of the mat response, reflecting back mat mechanical coupling and cardiorespiratory induced perturbations. Respiratory rate and heart rate were therefore estimated and compared with a reference wearable system. The maps revealed physiologically coherent spatial and temporal patterns, while the estimated rates showed good agreement with the reference measurements. These results demonstrate the feasibility of combining large area distributed sensing, spatial mapping, and quantitative cardiorespiratory monitoring within a DOFS based soft nearable interface.
Chinese Translation
分布式光纤传感（DOFS）结合了光纤传感器的优势，包括灵活性、小尺寸、对电磁干扰的免疫性和高计量性能，同时具备将单根光纤转变为连续传感元件以进行空间分辨的机械测量的能力。基于瑞利后向散射的光学频域反射测量（OFDR）实现了高空间分辨率的DOFS测量，拓宽了潜在传感应用的范围。然而，尽管对灵敏、空间分辨和可贴合的传感接口的需求日益增加，基于OFDR的DOFS在生物医学应用方面仍然 largely 未被探索。本研究提出了一种基于软DOFS的垫子，作为生理监测的大面积接口。单模光纤嵌入在柔性硅胶基体中，并以蛇形布局排列，以在垫子表面分布传感。该系统在2.6 mm的测量间距下，在活跃区域提供了2250个传感点，采样频率为50 Hz。该垫子在六名健康志愿者身上进行了评估，采用坐姿近接配置，放置在标准办公椅的靠背上。分布式输出使得垫子响应的二维映射成为可能，反映了垫子机械耦合和心肺诱发的扰动。因此，估算了呼吸频率和心率，并与参考可穿戴系统进行了比较。映射结果显示出生理上连贯的空间和时间模式，而估算的频率与参考测量结果良好一致。这些结果证明了在基于DOFS的软近接接口中结合大面积分布式传感、空间映射和定量心肺监测的可行性。
cs.CV / 142 / 2607.11257
LaGuadia: Language-Guided Adaptive Distillation from Pathology Foundation Models
LaGuadia：基于语言指导的病理基础模型自适应蒸馏
Abstract
Pathology Foundation Models (PFMs) offer powerful Whole Slide Image (WSI) representations but suffer from massive computational costs. While Knowledge Distillation (KD) can create efficient student models, existing multi-teacher methods often use suboptimal uniform weighting that ignores tissue heterogeneity. We propose LaGuadia (Language-Guided Adaptive DistillAtion), a framework that develops a compact pathology image encoder by dynamically integrating expertise from multiple PFMs under clinical linguistic guidance. Our approach utilizes a multi-stage pipeline: first, extracting visually observable clinical keywords from pathology reports; second, aligning visual features with these keywords via a Vision-Language meta-teacher (MedSigLIP) to provide dense semantic guidance; and finally, performing adaptive KD where teacher contributions are weighted based on their semantic alignment with the clinical narrative. Experiments on WSI captioning, visual question answering, and slide-level classification tasks demonstrate that an 87M parameter LaGuadia student model matches or exceeds foundation-scale models such as GigaPath and UNI, achieving strong factual consistency and robust generalization. These results highlight clinical language as an effective semantic anchor for building efficient and reliable digital pathology systems. Code is available at https://github.com/hvcl/LaGuadia.
Chinese Translation
病理基础模型（PFMs）提供了强大的全切片图像（WSI）表示，但面临巨大的计算成本。尽管知识蒸馏（KD）可以创建高效的学生模型，但现有的多教师方法往往使用次优的均匀加权，忽视了组织异质性。我们提出了LaGuadia（基于语言指导的自适应蒸馏）框架，通过在临床语言指导下动态整合多个PFMs的专业知识，开发出紧凑的病理图像编码器。我们的方法利用了多阶段流程：首先，从病理报告中提取可视的临床关键词；其次，通过视觉-语言元教师（MedSigLIP）将视觉特征与这些关键词对齐，以提供密集的语义指导；最后，进行自适应KD，其中教师的贡献根据其与临床叙述的语义对齐程度进行加权。在WSI标题生成、视觉问答和切片级分类任务上的实验表明，具有8700万参数的LaGuadia学生模型在匹配或超越基础规模模型（如GigaPath和UNI）的同时，达到了强大的事实一致性和稳健的泛化能力。这些结果突显了临床语言作为构建高效可靠的数字病理系统的有效语义锚点。代码可在https://github.com/hvcl/LaGuadia获取。
cs.CV / 143 / 2607.11281
The Devil Is in the Leakage: A Disentangled Dual-Purification Framework for High-Fidelity Hairstyle Transfer
泄漏中的魔鬼：一种解耦双重净化框架用于高保真发型转移
Abstract
Hairstyle transfer aims to synthesize a photorealistic portrait by transplanting the hairstyle from a reference image onto a source subject while preserving the source identity. Recent foundation models show strong generative capability, but they struggle with the zero-shot disentanglement required for precise local editing, often entangling the reference hairstyle with its original identity and pose. Existing diffusion-based pipelines typically decompose the task by first generating a "bald" image from the source and then injecting hairstyle features from the reference. However, we show that this paradigm suffers from a fundamental leakage problem. Identity Leakage in Hairstyle occurs when hairstyle features retain reference identity or pose information, while Flaw Leakage in Bald arises when residual artifacts in the bald image are propagated into the final synthesis. To address both issues, we propose the Dual-Purification Framework (DPF), which introduces two complementary training-time regularizers. Adversarial Hairstyle Purification (AHP) purifies hairstyle features by suppressing identity predictability under a mutual-information-inspired adversarial objective. Contrastive Geometric Purification (CGP) regularizes the ControlNet pathway with a contrastive objective, reducing the model's reliance on geometric artifacts in the bald condition. By jointly purifying the hairstyle representation and geometric pathway, DPF achieves high-fidelity, identity-preserving hairstyle transfer and state-of-the-art performance on diverse benchmarks.
Chinese Translation
发型转移旨在通过将参考图像中的发型移植到源对象上来合成逼真的肖像，同时保持源对象的身份。近期的基础模型展现了强大的生成能力，但在精确的局部编辑中所需的零样本解耦方面表现不佳，常常将参考发型与其原始身份和姿态纠缠在一起。现有的基于扩散的管道通常通过首先从源生成“光头”图像，然后注入参考的发型特征来分解任务。然而，我们表明这种范式存在根本的泄漏问题。发型中的身份泄漏发生在发型特征保留了参考身份或姿态信息时，而光头中的缺陷泄漏则发生在光头图像中的残余伪影传播到最终合成中。为了解决这两个问题，我们提出了双重净化框架（Dual-Purification Framework, DPF），该框架引入了两个互补的训练时正则化器。对抗性发型净化（Adversarial Hairstyle Purification, AHP）通过在一个受互信息启发的对抗目标下抑制身份可预测性来净化发型特征。对比几何净化（Contrastive Geometric Purification, CGP）通过对比目标对ControlNet路径进行正则化，减少模型在光头状态下对几何伪影的依赖。通过共同净化发型表示和几何路径，DPF实现了高保真、身份保持的发型转移，并在多样化的基准测试中取得了最先进的性能。
cs.CV / 144 / 2607.11285
SalientGS: Unified SfM-to-3DGS with Importance-Guided MCMC Gaussian Allocation
SalientGS：基于重要性引导的MCMC高斯分配的统一SfM到3DGS
Abstract
Reconstructing 3D scenes from unordered images remains bottlenecked by expensive Structure-from-Motion (SfM) preprocessing and frozen pose interfaces. We present SalientGS, a unified SfM-to-3D Gaussian Splatting (3DGS) pipeline. Its central contribution is importance-guided Markov Chain Monte Carlo (MCMC) Gaussian allocation, which aggregates multi-view residuals into per-Gaussian underfit and redundancy signals. These signals define a smooth importance-weighted sampling distribution that biases both birth and relocation toward underfit regions. This reallocates capacity from well-fit areas without altering the underlying stochastic gradient Langevin dynamics (SGLD). SalientGS achieves end-to-end reconstruction in 15 minutes with state-of-the-art perceptual quality. The supplementary material provides dedicated sections for Per-Scene Qualitative Comparisons and Per-Image Learned Perceptual Image Patch Similarity (LPIPS) Analysis, including failure cases. Code and evaluation scripts are available at https://github.com/Six-Bit-TX/SalientGS.
Chinese Translation
从无序图像重建3D场景仍然受到昂贵的运动结构（SfM）预处理和固定姿态接口的瓶颈。我们提出了SalientGS，一个统一的SfM到3D高斯喷溅（3DGS）管道。其核心贡献是重要性引导的马尔可夫链蒙特卡洛（MCMC）高斯分配，该方法将多视角残差聚合为每个高斯的欠拟合和冗余信号。这些信号定义了一个平滑的加权采样分布，偏向于欠拟合区域进行出生和重新定位。这种方法在不改变基础随机梯度朗之万动力学（SGLD）的情况下，从拟合良好的区域重新分配容量。SalientGS以15分钟的时间实现端到端重建，且具有最先进的感知质量。补充材料提供了针对每场景定性比较和每图像学习感知图像块相似性（LPIPS）分析的专门部分，包括失败案例。代码和评估脚本可在 https://github.com/Six-Bit-TX/SalientGS 获取。
cs.CV / 145 / 2607.11287
A Unified Framework for Comprehensive Cardiac CT Segmentation and Phenotyping: Human-in-the-Loop Data Annotation, Vision Foundation Model Development, Multicenter Evaluation and Clinical Validation
综合心脏CT分割与表型分析的统一框架：人机协作数据标注、视觉基础模型开发、多中心评估与临床验证
Abstract
Comprehensive quantification of cardiac structures from computed tomography (CT) remains limited not by data availability but by the scalability of measurements, which makes routine use impractical. Here we present a unified framework for comprehensive cardiac CT segmentation and phenotyping that combines a human-in-the-loop annotation pipeline, a cardiac CT augmentation technique, and a self-supervised foundation model pre-trained on 60,000 unlabeled cardiac CT scans. Using this approach, we assembled the largest and most comprehensive expert-annotated cardiac CT segmentation dataset to date, comprising 1598 cases and 14 distinct cardiac structures (1000 for training, 598 for the external test set). Across five external datasets, the framework segmented all structures more accurately and comprehensively than existing open-source tools. Self-supervised pre-training improved labeling efficiency, with the most significant gains observed during external evaluation in the low-data regime. Benchmarking across convolutional, transformer, and state-space architectures showed comparable performance, indicating that data quality and pre-training, rather than architecture, drove accuracy. The framework was scaled to population-level phenotyping, with segmented anatomy that carries functionally relevant information about ventricular function and disease severity beyond demographic variables. By openly releasing the largest dataset with human labels, code, model weights, a CT augmentation library, and software, this work provides a reproducible foundation for opportunistic cardiac phenotyping from routinely acquired CT scans.
Chinese Translation
从计算机断层扫描（CT）中对心脏结构进行全面量化的能力仍然受到测量可扩展性的限制，而非数据的可用性，这使得常规使用变得不切实际。在此，我们提出了一种综合心脏CT分割与表型分析的统一框架，该框架结合了人机协作的标注流程、心脏CT增强技术以及在60,000个未标记心脏CT扫描上进行预训练的自监督基础模型。采用这一方法，我们组建了迄今为止最大且最全面的专家标注心脏CT分割数据集，包含1598个病例和14种不同的心脏结构（1000个用于训练，598个用于外部测试集）。在五个外部数据集中，该框架对所有结构的分割精度和全面性均优于现有的开源工具。自监督预训练提高了标注效率，尤其在低数据环境下的外部评估中观察到显著的提升。对卷积、变换器和状态空间架构的基准测试显示出相当的性能，表明数据质量和预训练，而非架构，推动了准确性。该框架扩展到了群体级别的表型分析，分割的解剖结构提供了关于心室功能和疾病严重性的重要信息，超越了人口统计变量。通过公开发布包含人类标签的最大数据集、代码、模型权重、CT增强库和软件，本研究为从常规获取的CT扫描中进行机会性心脏表型分析提供了可重复的基础。
cs.CV / 146 / 2607.11295
Metadata Supervised MRI Representations for Modelling and Controlling Acquisition Variability
元数据监督的MRI表征用于建模和控制采集变异性
Abstract
Magnetic resonance imaging exhibits substantial acquisition variability, where identical anatomy can appear markedly different across scanners and imaging protocols. Consequently, learned representations entangle biological structure with acquisition-dependent appearance, limiting interpretability, generalisation, and clinical deployment. We show that these sources of variation can be separated by jointly modelling MRI images and DICOM metadata. Using large-scale clinical brain MRI data, we learn representations that separate anatomical structure from contrast-dependent appearance. Resulting contrast representations organise heterogeneous acquisitions, support sequence understanding, and detect image--metadata inconsistencies, whereas anatomical representations suppress acquisition-specific variation while preserving biologically relevant information. Building on these disentangled representations, we introduce a unified anatomy-preserving harmonisation model for cross-modality and cross-site adaptation, conditioned on image or acquisition metadata. Our findings suggest that acquisition variability is a structured component of the imaging process that can be modelled, audited, and controlled, providing a foundation for acquisition-aware representation learning in large-scale medical imaging.
Chinese Translation
磁共振成像展现出显著的采集变异性，相同的解剖结构在不同的扫描仪和成像协议下可能表现出明显的差异。因此，学习到的表征将生物结构与依赖于采集的外观纠缠在一起，限制了其可解释性、泛化能力和临床应用。我们展示了通过联合建模MRI图像和DICOM元数据，可以将这些变异源分离。利用大规模临床脑部MRI数据，我们学习到的表征将解剖结构与对比度依赖的外观分离。最终的对比度表征组织了异质的采集，支持序列理解，并检测图像与元数据之间的不一致性，而解剖表征则抑制了特定于采集的变异，同时保留了生物学相关信息。在这些解缠表征的基础上，我们引入了一个统一的保留解剖结构的调和模型，用于跨模态和跨站点的适应，条件为图像或采集元数据。我们的研究结果表明，采集变异性是成像过程中的一个结构化组成部分，可以被建模、审计和控制，为大规模医学成像中的采集感知表征学习提供了基础。
cs.CV / 147 / 2607.11303
ASUMOT: Motion-Consistency-Based Asynchronous UAV Detection and Tracking with Event Cameras
ASUMOT：基于运动一致性的异步无人机检测与跟踪方法（使用事件相机）
Abstract
Event cameras offer microsecond-level temporal resolution and high dynamic range for low-altitude UAV perception. However, long-range UAVs often produce sparse, fragmented, and noise-contaminated event responses, where one semantic target may appear as multiple spatially separated blobs. Direct blob-level asynchronous tracking therefore suffers from duplicate trajectories and unstable identities. We propose ASUMOT, a motion-consistency-based asynchronous UAV detection and tracking framework operating directly on raw events. ASUMOT models each UAV as a set of motion-consistent event blobs. A local motion-consistency estimator triggers reliable candidates, a lightweight multi-task verifier provides UAV confidence and motion-direction cues, and motion-consistency clustering aggregates fragmented blobs into identity-consistent UAV tracks. We also introduce ES-UAV, a high-definition event-level UAV benchmark with dense semantic annotations. Experiments on public UAV tracking data and ES-UAV show that ASUMOT improves the accuracy--efficiency trade-off while preserving asynchronous event processing. Code and Dataset will be released.
Chinese Translation
事件相机提供微秒级的时间分辨率和高动态范围，适用于低空无人机感知。然而，远程无人机通常会产生稀疏、碎片化且受噪声污染的事件响应，其中一个语义目标可能会表现为多个空间上分离的斑点。因此，直接在斑点级别进行异步跟踪会面临重复轨迹和不稳定身份的问题。我们提出了ASUMOT，一种基于运动一致性的异步无人机检测与跟踪框架，直接在原始事件上操作。ASUMOT将每个无人机建模为一组运动一致的事件斑点。局部运动一致性估计器触发可靠的候选目标，轻量级多任务验证器提供无人机置信度和运动方向线索，而运动一致性聚类将碎片化的斑点聚合为身份一致的无人机轨迹。我们还引入了ES-UAV，一个具有密集语义注释的高分辨率事件级无人机基准。对公共无人机跟踪数据和ES-UAV的实验表明，ASUMOT在保持异步事件处理的同时改善了准确性与效率的权衡。代码和数据集将被发布。
cs.CV / 148 / 2607.11312
SLVMBench: Skill Learning from Video Memory
SLVMBench：来自视频记忆的技能学习
Abstract
We introduce Skill Learning from Video Memory (SLVMBench), the first benchmark that jointly evaluates whether video large language models (video-LLMs) can learn skills from long video memory and apply them to real-time tasks. SLVMBench presents models with 2-3 hour video streams that contain a tutorial video embedded in a stream of arbitrary irrelevant videos, resembling real-world human learning practices. Video-LLMs are asked to apply the acquired skill to answer real-time questions about an ongoing video. Unlike long-video understanding benchmarks that emphasize passive comprehension and skill-learning benchmarks that rely on short, immediate demonstrations, SLVMBench tests the full pipeline of memorizing and extracting procedural knowledge, as well as transferring it to real-time tasks. Moreover, rigorous human annotations feature sub-second-level temporal calibration, manually engineered questions eliminating common-sense guessing, and collated tutorials to ensure coverage of the required skills. Evaluations on state-of-the-art proprietary and open-source video LLMs show that video-LLMs struggle substantially with learning and applying skill knowledge from videos. Moreover, performance degrades markedly when the skill knowledge is placed within a long video memory. These results reveal a key limitation of existing video LLMs and position SLVMBench as the first benchmark for studying real-time skill acquisition and application from long-context video memory.
Chinese Translation
我们介绍了来自视频记忆的技能学习（SLVMBench），这是第一个联合评估视频大型语言模型（video-LLMs）是否能够从长视频记忆中学习技能并将其应用于实时任务的基准。SLVMBench 向模型提供了包含 2-3 小时视频流的输入，这些视频流中嵌入了一段教程视频，并夹杂着任意无关的视频，类似于现实世界中的人类学习实践。视频-LLMs 被要求将所获得的技能应用于回答关于正在进行的视频的实时问题。与强调被动理解的长视频理解基准不同，以及依赖于短期、即时演示的技能学习基准，SLVMBench 测试了记忆和提取程序性知识的完整流程，以及将其转移到实时任务中的能力。此外，严格的人类注释具有亚秒级的时间校准，手动设计的问题消除了常识猜测，并整理了教程以确保覆盖所需技能。对最先进的专有和开源视频 LLM 的评估显示，视频-LLMs 在学习和应用视频中的技能知识方面面临显著困难。此外，当技能知识置于长视频记忆中时，性能显著下降。这些结果揭示了现有视频 LLM 的一个关键限制，并将 SLVMBench 定位为研究来自长上下文视频记忆的实时技能获取和应用的第一个基准。
cs.CV / 149 / 2607.11339
HierCAD: Hierarchical Text-to-CAD Design via Structure Alignment and Parameter Grounding
HierCAD：通过结构对齐和参数基础的层次化文本到CAD设计
Abstract
Recent text-to-CAD approaches have shown promising results by leveraging large language models, but they often struggle with maintaining structural consistency in complex designs and accurately grounding geometric parameters. To address these issues, we propose HierCAD, a hierarchical text-to-CAD framework that improves both structural reasoning and parameter prediction. HierCAD reformulates CAD generation as progressive reasoning by decomposing CAD construction trees into object-level procedural reasoning and part-level topology reasoning trajectories. To further improve generation fidelity, we introduce a unified Structure Alignment and Parameter Grounding (SAPG) learning strategy. Structure alignment aligns topology reasoning trajectories with their corresponding parametric CAD spans, while parameter grounding mitigates shortcut learning through structure-preserving parameter perturbations and ranking-based supervision. Experiments demonstrate that HierCAD outperforms prior state-of-the-art methods on both CAD sequence generation and reconstructed CAD model evaluation. Our code is available at https://github.com/Collab-Gen/HierCAD.
Chinese Translation
最近的文本到CAD方法通过利用大型语言模型取得了令人鼓舞的成果，但它们在复杂设计中往往难以保持结构一致性，并且在准确基础几何参数方面存在困难。为了解决这些问题，我们提出了HierCAD，一个层次化的文本到CAD框架，旨在改善结构推理和参数预测。HierCAD将CAD生成重新表述为渐进式推理，通过将CAD构建树分解为对象级程序推理和部件级拓扑推理轨迹。为了进一步提高生成的保真度，我们引入了一种统一的结构对齐和参数基础（Structure Alignment and Parameter Grounding, SAPG）学习策略。结构对齐将拓扑推理轨迹与其对应的参数化CAD跨度对齐，而参数基础通过保持结构的参数扰动和基于排名的监督来减轻捷径学习。实验表明，HierCAD在CAD序列生成和重建CAD模型评估方面均优于之前的最先进方法。我们的代码可在 https://github.com/Collab-Gen/HierCAD 获取。
cs.CV / 150 / 2607.11343
Longitudinal Multi-View Breast Cancer Risk Prediction
纵向多视角乳腺癌风险预测
Abstract
Accurate breast cancer risk prediction from screening mammography is critical for enabling personalized screening intervals and early detection. Recent deep learning methods have shown the value of longitudinal data and explicit temporal alignment. However, existing approaches either perform explicit alignment using a single mammographic view or model multiple views without explicit longitudinal alignment, limiting their ability to exploit the complementary spatial-temporal information used in clinical practice. To address this gap, we propose LMV-Net, a longitudinal multi-view breast cancer risk prediction model that jointly analyzes anatomically complementary CC and MLO views within an explicitly aligned longitudinal framework. We evaluate our approach on the public EMBED and CSAW-CC datasets, comparing it to state-of-the-art breast cancer risk prediction methods. Our model consistently outperforms existing approaches in overall risk prediction performance and across different breast density and cancer subgroups. Importantly, these improvements highlight the potential of longitudinal multi-view modeling to enhance risk stratification, paving the way for future work on personalized screening, earlier identification of high-risk patients, and more efficient screening resource allocation. The code is available at https://github.com/sot176/LMV-Net.
Chinese Translation
从筛查乳腺X光检查中准确预测乳腺癌风险对于实现个性化筛查间隔和早期检测至关重要。近期的深度学习方法展示了纵向数据和明确时间对齐的价值。然而，现有方法要么使用单一乳腺X光视图进行明确对齐，要么在没有明确纵向对齐的情况下建模多个视图，这限制了它们利用临床实践中使用的互补时空信息的能力。为了解决这一问题，我们提出了LMV-Net，一种纵向多视角乳腺癌风险预测模型，该模型在明确对齐的纵向框架内联合分析解剖上互补的CC视图和MLO视图。我们在公共的EMBED和CSAW-CC数据集上评估了我们的方法，并将其与最先进的乳腺癌风险预测方法进行了比较。我们的模型在整体风险预测性能以及不同乳腺密度和癌症亚组中始终优于现有方法。重要的是，这些改进突显了纵向多视角建模在增强风险分层方面的潜力，为未来个性化筛查、高风险患者的早期识别以及更有效的筛查资源分配的研究铺平了道路。代码可在 https://github.com/sot176/LMV-Net 获取。
cs.CV / 151 / 2607.11356
Benchmarking Edge Inference Strategies for Deep Learning Models in Industrial Machine Vision
工业机器视觉中深度学习模型的边缘推理策略基准测试
Abstract
Edge deployment is often the preferred solution for industrial machine vision systems when low latency, data security, or limited connectivity are critical requirements. Several frameworks are available to optimise inference on edge devices; however, relatively few studies have systematically compared their inference-time performance under industrial deployment conditions. In this work, we present a comparative study of four widely used approaches for machine vision inference in industrial settings: plain PyTorch, ONNX Runtime, OpenVINO, and TensorRT. The evaluation focuses on inference time, covers several CPU- and GPU-based hardware platforms, and includes both conventional convolutional neural networks and a transformer-based vision model. For the hardware platforms and models evaluated, the results show that OpenVINO achieves the lowest inference time on CPUs, while TensorRT achieves the lowest inference time on GPUs. However, TensorRT does not outperform plain PyTorch for the transformer-based model considered in this study.
Chinese Translation
边缘部署通常是工业机器视觉系统的首选解决方案，尤其在低延迟、数据安全或有限连接性等关键需求下。有多种框架可用于优化边缘设备上的推理；然而，系统性比较它们在工业部署条件下推理时间性能的研究相对较少。在本研究中，我们对四种在工业环境中广泛使用的机器视觉推理方法进行了比较研究：普通的 PyTorch、ONNX Runtime、OpenVINO 和 TensorRT。评估重点在于推理时间，涵盖了多个基于 CPU 和 GPU 的硬件平台，并包括传统卷积神经网络和基于变换器的视觉模型。对于评估的硬件平台和模型，结果显示 OpenVINO 在 CPU 上实现了最低的推理时间，而 TensorRT 在 GPU 上实现了最低的推理时间。然而，对于本研究中考虑的基于变换器的模型，TensorRT 的表现并未超过普通的 PyTorch。
cs.CV / 152 / 2607.11359
Efficient Tuning Before Low-Bit Post-Training Quantization for Stochastic Gradient Descent-optimized Models
低比特后训练量化前的高效调优方法用于随机梯度下降优化的模型
Abstract
Post-training quantization (PTQ) compresses deep neural networks for deployment under limited memory and computational budgets. However, low-bit (i.e., 2-bit or 4-bit) PTQ often suffers from substantial performance degradation. Most existing PTQ methods operate on an unconstrained full-precision (FP) model and primarily address quantization errors through post-hoc reconstruction. We argue that low-bit PTQ accuracy is limited not only by post-quantization error minimization, but also by the quantization-error tolerance of a FP model itself. In this paper, we propose Efficient Tuning Before Quantization (ETBQ), a pre-conditioning tuning stage for Stochastic Gradient Descent (SGD)-optimized models before PTQ. During tuning, the FP model is optimized under perturbations sampled from the error distributions of weight and activation quantization, guiding the model toward a loss-landscape region that is less sensitive to the subsequent PTQ. Unlike QAT, ETBQ does not train a fake-quantized deployment model, which is computationally and memory intensive. Instead, ETBQ outputs a FP model that can be used by any PTQ backend. Experiments on CIFAR-100, Tiny-ImageNet, ImageNet, and Cityscapes provide consistent evidence that ETBQ improves low-bit PTQ across diverse tasks. Under W2A4 settings, e.g., ETBQ improves over naive PTQ by 2.14\% top-1 accuracy on Tiny-ImageNet and by 5.80\% mIoU on Cityscapes. Code is available at https://github.com/xpxpxp2001xpxpxp/ETBQ.
Chinese Translation
后训练量化（PTQ）用于在有限的内存和计算预算下压缩深度神经网络以进行部署。然而，低比特（即2比特或4比特）PTQ往往会遭遇显著的性能下降。现有的大多数PTQ方法在不受约束的全精度（FP）模型上运行，主要通过事后重建来解决量化误差。我们认为，低比特PTQ的准确性不仅受到后量化误差最小化的限制，还受到FP模型本身的量化误差容忍度的影响。本文提出了一种在量化前进行的高效调优方法（ETBQ），作为随机梯度下降（SGD）优化模型的预调节阶段。在调优过程中，FP模型在从权重和激活量化的误差分布中采样的扰动下进行优化，引导模型朝向对后续PTQ不那么敏感的损失景观区域。与量化感知训练（QAT）不同，ETBQ并不训练一个伪量化的部署模型，这在计算和内存上都非常密集。相反，ETBQ输出一个可以被任何PTQ后端使用的FP模型。在CIFAR-100、Tiny-ImageNet、ImageNet和Cityscapes上的实验提供了一致的证据，表明ETBQ在多种任务中改善了低比特PTQ。在W2A4设置下，例如，ETBQ在Tiny-ImageNet上比简单PTQ提高了2.14%的top-1准确率，在Cityscapes上提高了5.80%的mIoU。代码可在https://github.com/xpxpxp2001xpxpxp/ETBQ获取。
cs.CV / 153 / 2607.11366
Self-supervised training for high-resolution close-range multispectral remote sensing imagery
高分辨率近距离多光谱遥感影像的自监督训练
Abstract
Although self-supervised learning (SSL) offers a promising way to reduce annotation effort in close-range remote sensing, its effectiveness for high-resolution multispectral unmanned aerial vehicle (UAV) imagery remains underexplored due to limited data. This study evaluated SSL pretraining for precision agriculture using cm-scale multispectral drone imagery collected across multiple sensors, years, and regions. Transformer-based encoders were pretrained with Momentum Contrast v3 (MoCo-v3) and Masked Autoencoders on a harmonized dataset combining msuav500K with newly collected multi-year UAV imagery from agricultural fields in Finland. Pretraining used four spectral bands (Green, Red, Red-Edge, Near-Infrared) for cross-sensor compatibility. The models were evaluated on crop-weed semantic segmentation using the WeedMap dataset with 5--100% training data. The following two subsets served as downstream tasks: Task A (Germany, RedEdge-M), where all pretrained models were compared under partial and full fine-tuning, and Task B (Switzerland, Sequoia), where the best encoder from Task A was assessed. Our Swin Transformer pretrained with MoCo-v3 achieved the strongest performance on both tasks, surpassing the Swin Transformer model of Doornbos et al. pretrained on a pre-release of msuav500K. Our pretrained Swin Transformer further demonstrated cross-sensor and cross-region generalization. We additionally provide a public multi-year multispectral UAV dataset from Finland to support future research.
Chinese Translation
尽管自监督学习（SSL）为减少近距离遥感中的标注工作提供了有希望的方法，但由于数据有限，其在高分辨率多光谱无人机（UAV）影像中的有效性仍未得到充分探索。本研究评估了用于精准农业的SSL预训练，利用在多个传感器、年份和地区收集的厘米级多光谱无人机影像。基于变换器的编码器在一个结合了msuav500K和新收集的来自芬兰农业领域的多年份无人机影像的和谐数据集上进行了Momentum Contrast v3（MoCo-v3）和掩蔽自编码器的预训练。预训练使用了四个光谱波段（绿色、红色、红边、近红外），以实现跨传感器兼容性。模型在使用WeedMap数据集进行作物-杂草语义分割的评估中，训练数据范围为5%至100%。以下两个子集作为下游任务：任务A（德国，RedEdge-M），在该任务中比较了所有预训练模型在部分和全量微调下的表现；任务B（瑞士，Sequoia），在该任务中评估了任务A中表现最佳的编码器。我们的Swin Transformer在MoCo-v3预训练后在两个任务上均取得了最佳表现，超越了Doornbos等人在msuav500K预发布版本上预训练的Swin Transformer模型。我们的预训练Swin Transformer进一步展示了跨传感器和跨区域的泛化能力。此外，我们还提供了来自芬兰的公共多年份多光谱无人机数据集，以支持未来的研究。
cs.CV / 154 / 2607.11412
Uncertainty Quantification for EO Regression Tasks: Building Height, Tree Canopy Height and Above-ground Biomass Estimation
地球观测回归任务的不确定性量化：建筑高度、树冠高度和地上生物量估计
Abstract
Earth Observation regression tasks such as building height, canopy height, and above-ground biomass estimation underpin critical applications in urban planning, forest monitoring, and climate policy, where both accuracy and reliability are critical. Yet most deep learning models yield only deterministic predictions, providing no indication of per-pixel reliability. These regression tasks are inherently challenging due to heterogeneous land surfaces, skewed target distributions, sensor noise, and signal saturation at high target values, making uncertainty (UC) estimation essential for reliable inference. We address this gap by modeling aleatoric uncertainty using year-long Sentinel-1 SAR and Sentinel-2 MSI time series, proposing two complementary approaches: (i) Gaussian UC, which jointly predicts mean and standard deviation under a Gaussian assumption, and (ii) Quantile UC, which estimates the 10th, 50th, and 90th quantiles to capture asymmetric and heteroscedastic error distributions. Both models are evaluated on three representative EO regression tasks at 10 m spatial resolution. Results show that both approaches match or surpass deterministic benchmarks and existing global products, while delivering well-calibrated, interpretable, and operationally useful confidence estimates. Notably, both models outperform the current 10 m state-of-the-art uncertainty-aware model for canopy height estimation. Our implementation will be available at: https://github.com/RituYadav92/EO-Regression-Uncertainty-Estimation
Chinese Translation
地球观测回归任务，如建筑高度、树冠高度和地上生物量估计，在城市规划、森林监测和气候政策等关键应用中至关重要，其中准确性和可靠性都是关键。然而，大多数深度学习模型仅提供确定性预测，无法指示每个像素的可靠性。这些回归任务由于地表的异质性、目标分布的偏斜、传感器噪声以及在高目标值下的信号饱和，固有地具有挑战性，因此不确定性（UC）估计对于可靠推断至关重要。我们通过利用长达一年的Sentinel-1 SAR和Sentinel-2 MSI时间序列来建模随机不确定性，提出了两种互补的方法：（i）高斯不确定性（Gaussian UC），在高斯假设下联合预测均值和标准差；（ii）分位数不确定性（Quantile UC），估计第10、第50和第90百分位数，以捕捉不对称和异方差的误差分布。两种模型在10米空间分辨率下的三项代表性地球观测回归任务上进行了评估。结果表明，这两种方法的表现与确定性基准和现有全球产品相匹配或超越，同时提供了良好校准、可解释且在操作上有用的置信度估计。值得注意的是，这两种模型在树冠高度估计方面的表现超过了当前10米最先进的不确定性感知模型。我们的实现将可在以下链接获取：https://github.com/RituYadav92/EO-Regression-Uncertainty-Estimation
cs.CV / 155 / 2607.11419
Video Transformer for Remote Identity Document Hologram Detection
用于远程身份文件全息图检测的视频变换器
Abstract
Remote identity authentification using Identification Documents has been a major challenge for several years. DeepFakes advent and the development of AI-guided tools helps fraudsters creating counterfeit ID Documents. Ensuring the authenticity of ID Documents has become a real clue in the seurization of remote authentification. This need is all the more pressing given the increasing digitization of administrative and transactional processes. To ensure widespread accessibility, the system should rely solely on video captured via mobile devices. In this specific context, confirming the authenticity of ID is a real challenge as many security features needs specific device like infrared sensor for instance. Among underutilized but promising security features, holographic printings hold a special place. Difficult to counterfeit, they produce distinctive visual effects according enlightment, making them both detectable in a video captured by a smartphone camera and difficult to imitate. In this paper, we propose a Remote Identity Document Verification System (RIDVS) and an approach based on a video transformer for detecting holograms in simple videos captured by smartphones. Our system is designed for a smartphone-based capture process, followed by a server-side verification. The hologram detection method builds on a robust model previously validated in a related research domain. We demonstrate that it outperforms existing SotA methods, achieving near-perfect accuracy even when trained on medium- to small-sized datasets. In particular, we report improvements of +26.86\% in Recall and +17.93\% in accuracy over the best MIDV-Holo baseline. This study includes several experiments that evaluate the model adaptation to frugality, both for training samples and computational resources.
Chinese Translation
远程身份认证使用身份证明文件多年来一直是一个重大挑战。DeepFakes的出现和AI驱动工具的发展使得欺诈者能够制作伪造的身份证明文件。确保身份证明文件的真实性已成为远程认证安全化的一个重要线索。鉴于行政和交易过程的日益数字化，这一需求显得尤为迫切。为了确保广泛的可及性，该系统应仅依赖于通过移动设备捕获的视频。在这一特定背景下，确认身份证明的真实性是一项真正的挑战，因为许多安全特性需要特定设备，例如红外传感器。在未被充分利用但前景广阔的安全特性中，全息印刷占据了特殊的位置。全息图难以伪造，能够根据光照产生独特的视觉效果，使其在智能手机摄像头捕获的视频中既可检测又难以模仿。在本文中，我们提出了一种远程身份文件验证系统（RIDVS）和一种基于视频变换器的方法，用于检测智能手机捕获的简单视频中的全息图。我们的系统设计用于基于智能手机的捕获过程，随后进行服务器端验证。全息图检测方法建立在先前在相关研究领域验证的稳健模型之上。我们证明该方法优于现有的最先进（SotA）方法，即使在中小型数据集上训练也能达到近乎完美的准确率。特别是，我们报告在最佳MIDV-Holo基线上的召回率提高了26.86\%，准确率提高了17.93\%。本研究包括多个实验，评估模型在训练样本和计算资源方面的适应性。
cs.CV / 156 / 2607.11473
Towards Efficient Convolutional Neural Network for Embedded Hardware via Multi-Dimensional Pruning
通过多维剪枝实现嵌入式硬件高效卷积神经网络
Abstract
In this paper, we propose TECO, a multi-dimensional pruning framework to collaboratively prune the three dimensions (depth, width, and resolution) of convolutional neural networks (CNNs) for better execution efficiency on embedded hardware. In TECO, we first introduce a two-stage importance evaluation framework, which efficiently and comprehensively evaluates each pruning unit according to both the local importance inside each dimension and the global importance across different dimensions. Based on the evaluation framework, we present a heuristic pruning algorithm to progressively prune the three dimensions of CNNs towards the optimal trade-off between accuracy and efficiency. Experiments on multiple benchmarks validate the advantages of TECO over existing state-of-the-art (SOTA) approaches. The code and pre-trained models are available at https://github.com/ntuliuteam/Teco.
Chinese Translation
本文提出了TECO，一个多维剪枝框架，用于协同剪枝卷积神经网络（CNN）的三个维度（深度、宽度和分辨率），以提高嵌入式硬件上的执行效率。在TECO中，我们首先引入一个两阶段的重要性评估框架，该框架根据每个维度内部的局部重要性和不同维度之间的全局重要性，高效且全面地评估每个剪枝单元。基于评估框架，我们提出了一种启发式剪枝算法，逐步剪枝CNN的三个维度，以实现准确性与效率之间的最佳权衡。在多个基准测试上的实验验证了TECO相较于现有最先进（SOTA）方法的优势。代码和预训练模型可在 https://github.com/ntuliuteam/Teco 获取。
cs.CV / 157 / 2607.11500
HyperGS: Fast and Generalizable Gaussian Video Representation
HyperGS：快速且可泛化的高斯视频表示
Abstract
Gaussian Splatting has emerged as an effective representation for video, but existing methods rely on per-video optimization. This leads to slow encoding and limits generalization across videos. To amortize this optimization, we propose HyperGS, a feedforward, optimization-free approach that directly predicts Gaussian representations from any video in a single forward pass, speeding up encoding and decoding by orders of magnitude while generalizing to out-of-distribution videos at higher resolutions. In HyperGS, we design a factorized spatiotemporal Transformer to extract tokens from video, and a learnable query-based Transformer to obtain 8-parameter Gaussian representations for each video frame. We find that naively predicting Gaussians across diverse videos induces a needle-like degeneration that collapses training, and address this with a rank-based geometric regularizer whose strength adapts dynamically to stabilize optimization. HyperGS achieves encoding at $10^4$--$10^5\times$ the speed of per-video Gaussian optimization at matched reconstruction quality while generalizing zero-shot to $720p$ video, enabling higher-resolution rendering without re-encoding. HyperGS improves PSNR by +2.9--3.1 dB over the prior video encoders on K400, SSv2, and UCF101 at a smaller video representation size. By predicting explicit 2D Gaussians in a single forward pass, HyperGS combines the fast, flexible rendering of Gaussian Splatting with the speed and generalization of feedforward prediction, advancing Gaussians as a practical direction for fast and generalizable video representation.
Chinese Translation
高斯喷溅（Gaussian Splatting）已成为视频的一种有效表示方法，但现有方法依赖于逐视频优化。这导致编码速度慢，并限制了跨视频的泛化能力。为了摊销这种优化，我们提出了HyperGS，一种前馈的、无优化的方法，它可以在单次前向传播中直接从任何视频预测高斯表示，从而在编码和解码速度上提高了几个数量级，同时在更高分辨率下对分布外视频进行泛化。在HyperGS中，我们设计了一种分解的时空变换器（spatiotemporal Transformer）来从视频中提取标记，并使用一种可学习的基于查询的变换器来为每个视频帧获取8参数的高斯表示。我们发现，简单地在不同视频中预测高斯会导致针状退化，从而导致训练崩溃，我们通过一种基于秩的几何正则化器来解决这个问题，该正则化器的强度会动态调整以稳定优化。HyperGS在匹配重建质量的情况下，以每视频高斯优化的$10^4$--$10^5 imes$速度进行编码，同时在$720p$视频上实现零-shot泛化，从而实现更高分辨率的渲染而无需重新编码。HyperGS在K400、SSv2和UCF101上以更小的视频表示尺寸提高了PSNR +2.9--3.1 dB，超越了之前的视频编码器。通过在单次前向传播中预测显式的2D高斯，HyperGS将高斯喷溅的快速灵活渲染与前馈预测的速度和泛化能力相结合，推动高斯作为快速且可泛化视频表示的实际方向。
cs.CV / 158 / 2607.11509
CFR-Net:Collaborative Feature Refnement Network for Medical Image Anomaly Detection
CFR-Net：用于医学图像异常检测的协同特征精炼网络
Abstract
Medical image anomaly detection remains challenging because networks pretrained on natural images often exhibit limited adaptability to medical images, where abnormal patterns appear as fine-grained local shifts, multi-scale contextual mismatches, and orientation-sensitive structural deviations. To address this, we propose the Collaborative Feature Refinement Network (CFR-Net), which combines shared teacher-student feature refinement before decoding with cross-space consistency after decoding. CFR-Net refines frozen teacher features and trainable student features using a Multi-Path Feature Refinement Module (MPFRM) with shared parameters, imposing common multi-path refinement rules on generic visual references and representations adapted to the medical domain, thereby mitigating domain discrepancy while modeling local, multi-scale, and orientation-sensitive feature characteristics. A variance-sensitive objective and dynamic ``homework set'' reorganization further support layer-adaptive consistency learning. Experiments on medical benchmarks show that CFR-Net achieves competitive anomaly classification and strong anomaly localization performance when trained on normal data.
Chinese Translation
医学图像异常检测依然具有挑战性，因为在自然图像上预训练的网络通常对医学图像的适应性有限，异常模式表现为细粒度的局部偏移、多尺度的上下文不匹配和对方向敏感的结构偏差。为了解决这个问题，我们提出了协同特征精炼网络（CFR-Net），该网络在解码前结合了共享的教师-学生特征精炼，并在解码后实现跨空间一致性。CFR-Net使用具有共享参数的多路径特征精炼模块（MPFRM）来精炼冻结的教师特征和可训练的学生特征，对通用视觉参考和适应医学领域的表示施加共同的多路径精炼规则，从而减轻领域间的不一致，同时建模局部、多尺度和方向敏感的特征特性。方差敏感的目标和动态“作业集”重组进一步支持层自适应一致性学习。在医学基准测试中的实验表明，当在正常数据上训练时，CFR-Net实现了具有竞争力的异常分类和强大的异常定位性能。
cs.CV / 159 / 2607.11523
Vinci2: Providing Proactive Assistance in Continuous Egocentric Videos
Vinci2：在连续自我中心视频中提供主动辅助
Abstract
When should an intelligent assistant speak up without being asked? Continuous egocentric video offers rich, evolving context that enables a new form of assistance: one that is proactive rather than merely reactive. Yet existing approaches either wait passively for user queries or treat every detected event as requiring a response, without considering the user's history, current activity, or whether assistance would actually be welcome. We reframe proactive assistance as a context-dependent decision problem: the agent must not only perceive what is happening, but reason over accumulated temporal context to determine when and whether to intervene. To this end, we present Vinci2, a proactive egocentric assistance system that advances the on-device assistant Vinci from reactive response toward proactivity. On the evaluation side, we present EgoServe, the first large-scale benchmark for proactive assistance in continuous egocentric video. EgoServe comprises over 3,000 service instances organized along 4 temporal memory horizons, ranging from immediate safety alerts to long-term habit coaching, across 10 service categories. On the modeling side, we propose EgoMemo, a training-free, memory-augmented agent that maintains three complementary memory representations: multi-scale temporal summaries, a semantic knowledge graph, and visual embedding archives. At each timestep, EgoMemo performs retrieval-augmented reasoning to determine whether assistance is warranted and, if so, produces contextually grounded responses. Experiments demonstrate that EgoMemo establishes strong baselines on EgoServe while remaining competitive on existing egocentric benchmarks. Our benchmark and code are publicly available at \href{https://sitonggong.github.io/EgoServe-page/}{Vinci2}.
Chinese Translation
智能助手何时应主动发言而不必等待请求？连续自我中心视频提供了丰富而不断发展的背景，使得一种新的辅助形式成为可能：这种辅助是主动的，而不仅仅是被动反应的。然而，现有的方法要么被动等待用户的查询，要么将每个检测到的事件都视为需要响应，而没有考虑用户的历史、当前活动或是否真的需要帮助。我们将主动辅助重新构建为一个依赖上下文的决策问题：代理不仅必须感知正在发生的事情，还必须对累积的时间上下文进行推理，以确定何时以及是否应进行干预。为此，我们提出了Vinci2，一个主动的自我中心辅助系统，它将设备上的助手Vinci从被动响应推进到主动性。在评估方面，我们提出了EgoServe，这是第一个针对连续自我中心视频中主动辅助的大规模基准。EgoServe包含超过3000个服务实例，按照4个时间记忆视野组织，从即时安全警报到长期习惯指导，涵盖10个服务类别。在建模方面，我们提出了EgoMemo，一个无训练的、增强记忆的代理，维护三种互补的记忆表示：多尺度时间摘要、语义知识图谱和视觉嵌入档案。在每个时间步，EgoMemo执行增强检索推理，以确定是否需要提供帮助，如果需要，则生成基于上下文的响应。实验表明，EgoMemo在EgoServe上建立了强大的基线，同时在现有的自我中心基准上保持竞争力。我们的基准和代码已公开发布，网址为 [Vinci2](https://sitonggong.github.io/EgoServe-page/)。
cs.CV / 160 / 2607.11529
Parse, Search, and Confirmation: Training-Free Aerial Vision-and-Dialog Navigation with Chain-of-Thought Reasoning and Structured Spatial Memory
解析、搜索与确认：基于链式思维推理和结构化空间记忆的无训练空中视觉与对话导航
Abstract
In this paper, we tackle the Aerial Vision-and-Dialog Navigation (AVDN) task in the training-free setting for resource-efficient high-altitude UAV navigation.Naively applying MLLMs leads to unreliable navigation due to weak directional grounding and the lack of explicit spatial memory.To address these issues, we propose PSC-AVDN, a training-free framework that tightly couples a three-stage Parsing-Search-Confirmation reasoning pipeline with a Structured Spatial Memory (SSM).The parsing stage uses an LLM to convert ambiguous dialogue instructions into stable geometric directional and destination cues.A Search Chain-of-Thought (S-CoT) then performs stepwise target exploration under high-altitude observations, and a Confirmation Chain-of-Thought (C-CoT) conducts fine-grained verification around candidate regions to resolve visual ambiguity.Meanwhile, SSM integrates three complementary sources of spatial cues, including multi-scale visual observation, spatial visual memory, and structured geometric memory to provide global spatial context and long-horizon consistency.Extensive experiments on ANDH and ANDH-Full show that PSC-AVDN establishes new state-of-the-art performance in the training-free setting, matching or surpassing several finetuned methods.Code will be publicly available at: https://github.com/QY6616/PSC-AVDN
Chinese Translation
在本文中，我们在无训练环境下解决了资源高效的高空无人机导航中的空中视觉与对话导航（AVDN）任务。简单地应用多模态大语言模型（MLLMs）会导致导航不可靠，因为方向性基础薄弱且缺乏明确的空间记忆。为了解决这些问题，我们提出了PSC-AVDN，这是一个无训练框架，将三阶段的解析-搜索-确认推理流程与结构化空间记忆（SSM）紧密结合。解析阶段使用大语言模型（LLM）将模糊的对话指令转换为稳定的几何方向和目的地线索。然后，搜索链式思维（S-CoT）在高空观察下逐步进行目标探索，而确认链式思维（C-CoT）则在候选区域内进行细致的验证，以解决视觉模糊。同时，SSM整合了三种互补的空间线索来源，包括多尺度视觉观察、空间视觉记忆和结构化几何记忆，以提供全球空间上下文和长时间一致性。在ANDH和ANDH-Full上的大量实验表明，PSC-AVDN在无训练环境下建立了新的最先进性能，匹配或超越了几种微调方法。代码将公开发布于：https://github.com/QY6616/PSC-AVDN
cs.CV / 161 / 2607.11533
Adaptive Routing for Efficient Diffusion Transformer-Based PNI Prediction
基于自适应路由的高效扩散变换器PNI预测
Abstract
Perineural invasion (PNI) is a critical prognostic factor in cholangiocarcinoma. However, its preoperative prediction from magnetic resonance imaging (MRI) remains challenging due to subtle imaging features that extend beyond tumor boundaries into surrounding regions. Conventional convolutional neural networks are limited in capturing long-range spatial dependencies. Transformer-based architectures improve global modeling of volumetric MRI by aggregating spatially distributed contextual cues, yet capturing subtle and noise-sensitive patterns in peritumoral regions remains challenging. Diffusion-based classifiers offer an alternative formulation by leveraging denoising-based class scoring to better capture such subtle patterns. However, these approaches introduce substantial computational overhead due to the combination of transformer-based modeling and iterative denoising processes. To address these challenges, we formulate PNI prediction as a diffusion-based classification problem and implement the denoising network using a transformer-based representation. To improve computational efficiency, we introduce adaptive routing across attention heads, spatial tokens, and MLP width. Experimental results demonstrate that the proposed approach achieves an AUC of 0.731 with 257.57 GFLOPs.
Chinese Translation
神经周围侵犯（PNI）是胆道癌的重要预后因素。然而，由于影像特征微妙，超出肿瘤边界延伸至周围区域，因此从磁共振成像（MRI）中进行术前预测仍然具有挑战性。传统的卷积神经网络在捕捉长距离空间依赖性方面存在局限。基于变换器的架构通过聚合空间分布的上下文线索，改善了对体积MRI的全局建模，但在捕捉肿瘤周围区域的微妙和噪声敏感模式方面仍然具有挑战性。基于扩散的分类器通过利用去噪的类别评分提供了一种替代方案，以更好地捕捉这些微妙的模式。然而，这些方法由于结合了基于变换器的建模和迭代去噪过程，导致了显著的计算开销。为了解决这些挑战，我们将PNI预测公式化为一个基于扩散的分类问题，并使用基于变换器的表示实现去噪网络。为了提高计算效率，我们在注意力头、空间标记和多层感知器宽度之间引入自适应路由。实验结果表明，所提出的方法在257.57 GFLOPs的计算量下实现了0.731的AUC。
cs.CV / 162 / 2607.11548
Training-Free Off-Screen Player Imputation for Broadcast-Based Spatial Football Analytics
无训练的屏幕外球员插补用于基于广播的空间足球分析
Abstract
Spatial football metrics such as pitch control assume access to the positions of all 22 players, yet the most widely available source of positional data -- the broadcast main camera -- shows only 10-16 of them at any moment. We quantify the resulting distortion with an open, reproducible benchmark: a simulated broadcast viewport applied to open full-pitch tracking data (Metrica Sports; three matches, one held out from method development). Ignoring off-screen players -- the visible-only baseline implied whenever a video-based game-state-reconstruction (GSR) pipeline adds no imputation layer -- inflates hidden-zone pitch-control error to 25.1-26.9 percentage points and a mean absolute control-share error of 11.1-13.4 points across the three matches. We then evaluate a ladder of training-free, online imputation baselines that use only observations from the match being analysed. The best overall on these decision-relevant metrics, role-anchored centroid voting (each visible player votes for the full-team centroid by subtracting its running role offset, attenuating the viewport-induced subset bias), roughly halves hidden-zone error (to 12.2-13.8 points) and cuts control-share error to 28-48% of the ignore policy at every viewport width from 36 m to 60 m in all three matches. For occlusions <=9.6 s -- the regime of the closest learned prior work -- it reaches binwise median position errors of 3.3-8.9 m; but 50-57% of hidden-player observations lie beyond that regime. Integrated end-to-end into a broadcast-video GSR pipeline, imputation moves a downstream possession-quality score (Space-Creation Index) by 15.6 and 17.2 points on two real World Cup broadcast windows, flipping the verdict class in one.
Chinese Translation
空间足球指标如场地控制假设能够获取所有22名球员的位置，然而最广泛可用的位置数据来源——广播主摄像头——在任何时刻仅显示10-16名球员。我们通过一个开放且可重复的基准量化了由此产生的失真：将模拟的广播视口应用于开放的全场追踪数据（Metrica Sports；三场比赛，其中一场在方法开发中被保留）。忽略屏幕外球员——在视频基础的比赛状态重建（GSR）管道中未添加插补层时所隐含的仅可见基线——使得隐藏区域的场地控制误差膨胀至25.1-26.9个百分点，三场比赛的平均绝对控制份额误差为11.1-13.4分。随后，我们评估了一系列无训练的在线插补基线，这些基线仅使用正在分析的比赛中的观察数据。在这些与决策相关的指标中，最佳的整体表现是角色锚定的质心投票（每个可见球员通过减去其跑动角色偏移量为全队质心投票，从而减轻视口引起的子集偏差），大约将隐藏区域误差减少了一半（降至12.2-13.8分），并将控制份额误差降低至忽略政策的28-48%，在所有三场比赛中从36米到60米的每个视口宽度下均如此。对于遮挡时间<=9.6秒——最接近的已学习先前工作的范围——其达到的分箱中位位置误差为3.3-8.9米；但50-57%的隐藏球员观察数据超出了该范围。将插补集成到广播视频GSR管道中，能够将下游的控球质量评分（空间创造指数）提升15.6和17.2分，在两个真实的世界杯广播窗口中，改变了其中一个的判决类别。
cs.CV / 163 / 2607.11557
Single-Teacher View Augmentation: Enhancing Knowledge Distillation with Student-Guided Perturbations
单教师视图增强：通过学生引导的扰动提升知识蒸馏
Abstract
Knowledge distillation (KD) typically relies on the fixed perspective of a single teacher, limiting the diversity of supervisory signals. While multi-teacher distillation addresses this by aggregating knowledge from multiple models, it incurs prohibitive computational and storage costs. To balance efficiency and diversity, recent research has focused on generating virtual views from a single teacher. However, existing methods face a trade-off: random perturbation approaches offer efficiency but lack controlled diversity, while structured augmentation methods require multi-stage training and incur linear parameter growth. We observe that this trade-off stems from a common design choice: using the teacher's strong but static features to generate views. Instead, we propose Shift-Augmented Knowledge Distillation (SAKD), a simple yet effective framework that leverages the student's evolving features as a dynamic condition for perturbation generation. This shift in perspective enables single-stage training while producing adaptive, diverse views through a parameter-free cyclic shift. Extensive experiments on CIFAR-100 and ImageNet demonstrate that SAKD consistently outperforms random perturbation methods and achieves accuracy on par with two-stage approaches, while using significantly fewer parameters and eliminating pre-training requirements.
Chinese Translation
知识蒸馏（KD）通常依赖于单一教师的固定视角，这限制了监督信号的多样性。虽然多教师蒸馏通过聚合多个模型的知识来解决这一问题，但它会带来高昂的计算和存储成本。为了平衡效率和多样性，近期研究集中于从单一教师生成虚拟视图。然而，现有方法面临权衡：随机扰动方法提供了效率，但缺乏可控的多样性，而结构化增强方法则需要多阶段训练并导致参数线性增长。我们观察到，这种权衡源于一个共同的设计选择：使用教师的强大但静态特征来生成视图。相反，我们提出了Shift-Augmented Knowledge Distillation（SAKD），这是一个简单而有效的框架，利用学生不断演变的特征作为扰动生成的动态条件。这种视角的转变使得单阶段训练成为可能，同时通过无参数的循环移位生成自适应、多样化的视图。在CIFAR-100和ImageNet上的大量实验表明，SAKD始终优于随机扰动方法，并在准确性上与两阶段方法相当，同时使用显著更少的参数并消除了预训练的需求。
cs.CV / 164 / 2607.11560
Technical Report on the CVPR 2026@AdvML Workshop Challenge
关于 CVPR 2026@AdvML 研讨会挑战的技术报告
Zhang, Tianyuan, Jing, Zonglei, Liu, Jiangfan, Zhang, Ligong, Ma, Ke, Sun, Chengzhi, Xu, Xiaohai, Zhang, Zhirui, Xu, Qianqian, Huang, Qingming, Fang, Hanyu, Liu, Junhua, Wang, Zheng, Liu, Xiaoliang, Li, Yuanbo, Gui, Shuai, Wang, Bin, Zheng, Menghe, Nie, Jing, Meng, Hanyang, Zhang, Zeyang, Zhang, Xiang, Zhu, Yongxuan, Ding, Rui, Li, Hainan, Zhang, Yongkang, Zhu, Zhilei, Kong, Xianglong, Hu, Jin, Ying, Zonghao, Xiao, Yisong, Chen, Lei, Qin, Haotong, Wang, Jiakai, Liu, Aishan, Li, Ruikai, Karbing, Julia, Dong, Yinpeng, Yin, Zhenfei, Jing, Shao, Hu, Xia, Xu, Jingyi, Dai, Juntao, Chen, Xinyun, Patel, Vishal M., Liu, Xianglong, Song, Dawn, Yuille, Alan, Torr, Philip H. S., Tao, Dacheng
Abstract
Vision-language agents (VLAs) are increasingly used to interpret complex driving scenes and support safety-critical reasoning. This report presents the CVPR 2026@AdvML Workshop Challenge on adversarial multimodal attacks against autonomous-driving VLAs. Built on DriveLM-style multi-view visual question answering, the challenge represents each scene with six synchronized camera images and a structured collection of driving-related question-answer pairs. Participants generate adversarial images and suffix-only textual perturbations that induce model responses to deviate from reference answers while preserving image fidelity and limiting textual cost. The competition comprises two phases, with Phase II adding a hidden black-box model to assess transferability. We describe the task design, submission rules, evaluation protocol, and leaderboard results, and then examine five leading submissions for which technical reports were available. Across these reports, several recurring patterns emerge: image-side attacks are favored by the suffix penalty; scene-level, multi-view optimization is more effective than treating views in isolation; QA types and graph structure provide useful priors for allocating attack budget; feature-space objectives can improve black-box transfer; and typographic content embedded in camera images exposes a persistent vulnerability in driving VLAs. These findings provide a practical reference for future robustness evaluation and defense design in multimodal autonomous-driving systems.
Chinese Translation
视觉语言代理（VLA）越来越多地用于解读复杂的驾驶场景并支持安全关键的推理。本报告介绍了针对自主驾驶 VLA 的对抗性多模态攻击的 CVPR 2026@AdvML 研讨会挑战。该挑战基于 DriveLM 风格的多视角视觉问答，通过六个同步摄像头图像和一组结构化的与驾驶相关的问题-答案对来表示每个场景。参与者生成对抗性图像和仅后缀的文本扰动，以使模型响应偏离参考答案，同时保持图像的保真度并限制文本成本。比赛分为两个阶段，第二阶段增加了一个隐藏的黑箱模型以评估可迁移性。我们描述了任务设计、提交规则、评估协议和排行榜结果，并审查了五个提供技术报告的领先提交。在这些报告中，出现了几个反复出现的模式：图像侧攻击受到后缀惩罚的青睐；场景级的多视角优化比孤立处理视角更有效；问答类型和图结构为分配攻击预算提供了有用的先验；特征空间目标可以改善黑箱迁移；嵌入在摄像头图像中的排版内容暴露了驾驶 VLA 的一个持久性脆弱性。这些发现为未来多模态自主驾驶系统的鲁棒性评估和防御设计提供了实用参考。
cs.CV / 165 / 2607.11562
MonkeyOCRv2: A Visual-Text Foundation Model for Document AI
MonkeyOCRv2：用于文档人工智能的视觉-文本基础模型
Abstract
Mainstream visual encoders are pretrained on natural images and cannot be effectively applied to document images without document-oriented adaptation, as dense text and fine-grained character strokes demand character-level visual perception. We present MonkeyOCRv2, a visual-text pretrained model for document AI. First, we construct MonkeyDoc v2, to our knowledge the largest document-image pretraining corpus, comprising 113 million images spanning 17 languages. Second, we propose a pretraining strategy that jointly learns image-to-text generation and pixel-level document reconstruction: the former aligns visual representations with textual content, while the latter preserves character strokes and layout details. Extensive experiments are conducted on five representative document analysis tasks, including text recognition, formula recognition, text detection, document tampering detection, and overlapping text segmentation. Replacing the original encoders with MonkeyOCRv2 consistently improves performance across all five tasks. Finally, we validate its effectiveness as the vision encoder of multimodal large language models on the more challenging tasks of document parsing and document understanding. Kept frozen and paired with a lightweight language model, it yields a 0.7B document parsing model that sets a new open-source state-of-the-art on MDPBench, a recent benchmark spanning digital-born and photographed documents across 17 languages, surpassing the previous best 3B dots.mocr by 2.8% absolute with a vision encoder roughly 11$\times$ smaller. The frozen encoder also powers a document understanding model that outperforms counterparts built on CLIP, DINO, and SAM across eight benchmarks under identical training settings. These results suggest that document-oriented visual pretraining can serve as a foundation for document intelligence in its own right.
Chinese Translation
主流视觉编码器是在自然图像上进行预训练的，无法有效应用于文档图像，除非进行面向文档的适应，因为密集文本和细粒度字符笔画需要字符级的视觉感知。我们提出了MonkeyOCRv2，一个用于文档人工智能的视觉-文本预训练模型。首先，我们构建了MonkeyDoc v2，据我们所知，这是最大的文档图像预训练语料库，包含1.13亿张跨越17种语言的图像。其次，我们提出了一种预训练策略，联合学习图像到文本的生成和像素级的文档重建：前者将视觉表示与文本内容对齐，而后者则保留字符笔画和布局细节。我们在五个代表性的文档分析任务上进行了广泛的实验，包括文本识别、公式识别、文本检测、文档篡改检测和重叠文本分割。用MonkeyOCRv2替换原始编码器在所有五个任务上均一致提高了性能。最后，我们验证了其作为多模态大语言模型视觉编码器的有效性，在更具挑战性的文档解析和文档理解任务上。保持冻结并与轻量级语言模型配对，它产生了一个0.7B的文档解析模型，在MDPBench这一涵盖17种语言的数字原生和拍摄文档的最新基准测试中设定了新的开源最先进水平，超越了之前最佳的3B dots.mocr，绝对提升了2.8%，而其视觉编码器的规模大约是原来的11倍。该冻结编码器还支持一个文档理解模型，在相同训练设置下超越了基于CLIP、DINO和SAM的对应模型，在八个基准测试中表现优异。这些结果表明，面向文档的视觉预训练可以作为文档智能的基础。
cs.CV / 166 / 2607.11581
Actor as Its Own Critic: Unifying Region Understanding and Localization via CycleGRPO
演员自我批评：通过 CycleGRPO 统一区域理解与定位
Abstract
This paper introduces Actor as Its Own Critic, a unified reinforcement learning framework, Cycle Group Relative Policy Optimization (CycleGRPO), that jointly optimizes region understanding and localization for Multimodal Large Language Models (MLLMs). Unlike existing separate pipelines, we leverage the inherent duality between the two tasks to construct a self-evaluating reinforcement learning paradigm: "region $\to$ text $\to$ region''. Specifically, a single MLLM first acts as the actor to generate region captions, then immediately transitions to a critic to ground its generated text back in the spatial domain. Therefore, CycleGRPO requires only region inputs, e.g., masks or bounding boxes, entirely bypassing the need for textual ground truths. A quality-aware token-level cycle-consistency reward is employed to assess the semantic discriminability of text captions via their physical localization accuracy. Empirically, built upon SAMTok, our CycleGRPO framework successfully bootstraps both capabilities simultaneously. Without any task-specific fine-tuning, the framework yields consistent performance gains across a wide range of benchmarks, including region captioning, region VQA, grounded dialogue, and referring segmentation. Overall, CycleGRPO offers a straightforward and scalable way to advance pixel-level capabilities in MLLMs. Code and models are released at https://github.com/devinxzhang/CycleGRPO.
Chinese Translation
本文介绍了演员自我批评（Actor as Its Own Critic），一个统一的强化学习框架——循环组相对策略优化（Cycle Group Relative Policy Optimization，CycleGRPO），该框架联合优化多模态大语言模型（Multimodal Large Language Models，MLLMs）的区域理解与定位。与现有的分离管道不同，我们利用这两项任务之间的内在对偶性构建了一种自我评估的强化学习范式：“区域 $ o$ 文本 $ o$ 区域”。具体而言，一个单一的 MLLM 首先作为演员生成区域标题，然后立即转变为批评者，将其生成的文本重新定位到空间域。因此，CycleGRPO 仅需要区域输入，例如掩码或边界框，完全绕过了对文本真实值的需求。我们采用了一种质量感知的令牌级循环一致性奖励，以评估文本标题的语义可区分性及其物理定位的准确性。实证研究表明，基于 SAMTok，我们的 CycleGRPO 框架成功地同时引导了这两种能力。在没有任何特定任务微调的情况下，该框架在区域标题生成、区域视觉问答（region VQA）、有依据对话和指称分割等广泛基准测试中均取得了一致的性能提升。总体而言，CycleGRPO 提供了一种简单且可扩展的方法，以提升 MLLMs 的像素级能力。代码和模型已发布在 https://github.com/devinxzhang/CycleGRPO。
cs.CV / 167 / 2607.11584
GB-SVFBP: Gaussian-Based Shift-Variant FBP neural network
GB-SVFBP：基于高斯的变位滤波反投影神经网络
Abstract
This paper proposes a Gaussian-Based Shift-Variant filtered backprojection (FBP) neural network, which is designed for the efficient reconstruction of non-circular trajectory cone beam computed tomography. The traditional differentiable shift-variant FBP model consists of a filtering component and a backprojection process. The filtering component includes operations such as weightings, differentiations, a 2D Radon transform, and a 2D backprojection. The proposed methods build on this framework by introducing a trainable 2D Gaussian model to represent the trajectory-related part in the filtering process, achieving a substantial reduction in the number of trainable parameters. Experimental results demonstrate that the proposed model reduces the parameter count by 99%, while only sacrificing a slight amount of reconstruction quality. Furthermore, the training time for each trajectory is reduced to one-fourth of the original, significantly accelerating convergence. These enhancements demonstrate a considerable augmentation in the model's practicality and effectiveness, making it a valuable asset for real-world applications.
Chinese Translation
本文提出了一种基于高斯的变位滤波反投影（FBP）神经网络，旨在高效重建非圆轨迹锥束计算机断层扫描。传统的可微分变位 FBP 模型由过滤组件和反投影过程组成。过滤组件包括加权、微分、二维 Radon 变换和二维反投影等操作。所提出的方法在此框架基础上，引入了可训练的二维高斯模型，以表示过滤过程中与轨迹相关的部分，从而显著减少可训练参数的数量。实验结果表明，所提模型将参数数量减少了 99%，同时仅轻微牺牲了重建质量。此外，每条轨迹的训练时间减少至原来的四分之一，显著加快了收敛速度。这些改进展示了模型在实用性和有效性方面的显著增强，使其成为实际应用中的宝贵资产。
cs.CV / 168 / 2607.11588
FoundationGeo: Learning Spatial Pixel-Wise Fields for Monocular Metric Geometry
FoundationGeo：学习单目度量几何的空间像素级场
Abstract
We present FoundationGeo, a two-stage framework that explicitly bridges relative and metric prediction via spatial calibration and principled data design. Stage 1 learns a high-fidelity, affine-invariant geometry model by initializing with DINOv3 and training on a curated 10.2M-sample multi-domain corpus with complementary local-detail supervision, yielding sharp boundaries and strong cross-domain generalization. Stage 2 moves beyond global scaling by introducing lightweight pixel-wise calibration fields for metric estimation: a scale field for spatially varying metric alignment and a ray-direction correction field that mitigates directional bias in point-map geometry, together producing metrically consistent 3D point maps. Beyond model design, we identify camera intrinsic coverage, especially focal length distribution mismatch between training and test data, as a key bottleneck for zero-shot metric generalization: performance drops sharply when test intrinsics fall outside the training distribution. To address this, we synthesize additional training data across diverse focal lengths using a Blender-based data engine, repairing under-covered focal regimes and improving robustness under intrinsic shift. Extensive zero-shot evaluations across seven benchmarks show that FoundationGeo significantly strengthens cross-domain robustness, staying near the top across diverse domains while avoiding the sharp cross-domain performance drops observed in other methods. This consistency translates into the best overall performance, surpassing heavier baselines by over 5.2% on average.
Chinese Translation
我们提出了FoundationGeo，一个两阶段框架，通过空间校准和原则性数据设计明确地连接相对预测和度量预测。第一阶段通过以DINOv3为初始化，在一个经过精心策划的1020万样本多领域语料库上进行训练，并结合局部细节监督，学习高保真、仿射不变的几何模型，从而产生清晰的边界和强大的跨领域泛化能力。第二阶段通过引入轻量级的像素级校准场进行度量估计，超越了全局缩放：一个用于空间变化的度量对齐的尺度场和一个减轻点图几何中方向偏差的光线方向校正场，共同生成度量一致的3D点图。除了模型设计外，我们还确定了相机内参覆盖，特别是训练和测试数据之间的焦距分布不匹配，作为零样本度量泛化的关键瓶颈：当测试内参超出训练分布时，性能急剧下降。为了解决这个问题，我们使用基于Blender的数据引擎合成了跨越多种焦距的额外训练数据，修复了覆盖不足的焦距范围，并提高了在内参变化下的鲁棒性。在七个基准上的广泛零样本评估表明，FoundationGeo显著增强了跨领域的鲁棒性，在不同领域中保持接近顶尖水平，同时避免了其他方法中观察到的急剧跨领域性能下降。这种一致性转化为最佳的整体性能，平均超越更重的基线超过5.2%。
cs.CV / 169 / 2607.11591
Similarity-Guided Curriculum Fine-Tuning of LLMs for Neural Architecture Synthesis
基于相似性引导的LLMs课程微调用于神经架构合成
Abstract
Introduce a MinHash-based similarity scheduling framework that constructs a progressive curriculum over neural architecture code for LLM-based neural architecture search (NAS). Using 128-permutation MinHash signatures over normalised 7-gram source code shingles, we partition the reference pool into similarity bands and present them in increasing architectural heterogeneity, with the best LoRA adapter from each stage merged cumulatively into the backbone. We evaluate the framework on OlympicCoder-7B within the LEMUR benchmark on CIFAR-10 image classification, generating N =15 candidate architectures per epoch across six progressive fine-tuning steps. The curriculum achieves 60% peak success rate at the high-similarity level without post-processing repair. A 2*2 ablation at the most diverse level curriculum versus base model, with versus without partial interface repair reveals that without repair the base model (47% peak SR) substantially outperforms the curriculum model (7% SR), while adding partial repair brings both to 53% SR. This pattern is consistent with merge-level weight drift progressively erasing evaluator-interface priors, and suggests that interface repair and curriculum scheduling target distinct failure modes. We further report a cross-dataset transfer observation on SVHN, where direct base-model generation without curriculum warmup yields 27% peak SR at substantially lower accuracy (60.5%) than the CIFAR-10 equivalent, consistent with the increased synthesis difficulty of the unq-family anchor architecture.
Chinese Translation
引入一种基于MinHash的相似性调度框架，该框架在神经架构代码上构建渐进式课程，用于基于LLM的神经架构搜索（NAS）。通过对标准化的7-gram源代码片段使用128-排列的MinHash签名，我们将参考池划分为相似性带，并以逐渐增加的架构异质性呈现它们，将每个阶段的最佳LoRA适配器累积合并到主干网络中。我们在CIFAR-10图像分类的LEMUR基准上对OlympicCoder-7B评估该框架，在六个渐进微调步骤中每个时期生成N=15个候选架构。该课程在高相似性水平下实现了60%的峰值成功率，无需后处理修复。在最具多样性的课程与基础模型之间进行的2*2消融实验显示，未进行修复的基础模型（47%峰值成功率）显著优于课程模型（7%成功率），而添加部分修复则使两者均达到53%成功率。这一模式与合并级别的权重漂移逐渐抹去评估者-接口先验一致，并表明接口修复和课程调度针对不同的失败模式。我们进一步报告了在SVHN上的跨数据集迁移观察，其中直接基于模型生成而不进行课程预热，在准确率显著较低（60.5%）的情况下，峰值成功率为27%，这与unq-family锚架构的合成难度增加一致。
cs.CV / 170 / 2607.11623
Backbone-Agnostic Perturbation-Induced Uncertainty Learning for End-to-End Real-World Image Dehazing
与骨干网络无关的扰动引发的不确定性学习用于端到端真实世界图像去雾
Abstract
Real-world paired image dehazing remains challenging because haze degradation is spatially non-uniform, illumination-dependent, and physically ambiguous even when haze-free references are available. Existing end-to-end restoration networks usually formulate dehazing as a deterministic mapping from a hazy observation to a clean target, leaving the uncertainty hidden in degraded features, haze priors, and cross-domain negative samples insufficiently explored. In this paper, we propose Backbone-Agnostic Perturbation-Induced Uncertainty Learning (BPUL), a plug-and-play uncertainty learning framework for end-to-end real-world image dehazing. BPUL first introduces a Learnable Perturbation-induced Uncertainty Modulator (LPUM) that estimates channel-wise and spatial-wise feature sensitivity through reparameterized stochastic perturbations. It then develops a Prior-informed Uncertainty-guided Reconstruction Module (PURM), which exploits transmission and atmospheric-light priors to reconstruct the hazy observation from the restored result and enforce degradation consistency. Furthermore, we propose a Dual-space Domain-diversified Distribution-aware Contrastive Loss ($D^3$CL) to regularize both clean restoration and hazy reconstruction spaces with real-world and synthetic negatives. Experiments on five real-world paired benchmarks show that BPUL consistently improves multiple representative backbones. Since only LPUM is retained during inference while PURM and $D^3$CL are used as training-time constraints, BPUL brings substantial restoration gains with only marginal additional inference overhead.
Chinese Translation
真实世界配对图像去雾仍然具有挑战性，因为雾霾退化在空间上是不均匀的，依赖于照明，并且即使在存在无雾参考的情况下也具有物理模糊性。现有的端到端恢复网络通常将去雾视为从雾霾观察到清晰目标的确定性映射，导致隐藏在退化特征、雾霾先验和跨域负样本中的不确定性未被充分探索。在本文中，我们提出了一种与骨干网络无关的扰动引发的不确定性学习框架（Backbone-Agnostic Perturbation-Induced Uncertainty Learning，BPUL），用于端到端真实世界图像去雾。BPUL首先引入了一种可学习的扰动引发不确定性调节器（Learnable Perturbation-induced Uncertainty Modulator，LPUM），通过重新参数化的随机扰动来估计通道和空间特征的敏感性。然后，它开发了一种基于先验的不确定性引导重建模块（Prior-informed Uncertainty-guided Reconstruction Module，PURM），利用传输和大气光先验从恢复结果中重建雾霾观察，并强制保持退化一致性。此外，我们提出了一种双空间领域多样化分布感知对比损失（Dual-space Domain-diversified Distribution-aware Contrastive Loss，$D^3$CL），以用真实世界和合成负样本对清晰恢复和雾霾重建空间进行正则化。在五个真实世界配对基准上的实验表明，BPUL始终提高了多个代表性骨干网络的性能。由于在推理过程中仅保留LPUM，而PURM和$D^3$CL作为训练时约束使用，BPUL在仅增加少量推理开销的情况下带来了显著的恢复增益。
cs.CV / 171 / 2607.11644
Motion4Motion: Motion Transfer Across Subjects at Inference
Motion4Motion：推理中的跨主体运动转移
Abstract
This work explores the motion transfer from one video to another, which is crucial in animation for diverse characters. Previously, video motion transfer has been largely explored between human and human-like characters, enabling a lot of applications in digital creation. However, these approaches encounter a main limitation. Specifically, related technical pipelines heavily rely on a predefined human skeleton structure and accordingly require skeleton-conditional model training. On the one hand, these methods are difficult to generalize to diverse characters, such as animals from different species, while preserving their unique motion styles. On the other hand, labeled data in diverse skeletons is limited, which additionally restricts the large-scale training for the task. In this paper, we jump out of the skeleton-based motion transfer framework and propose a training-free motion transfer framework, named Motion4Motion. Motion4Motionmodels the motion flow of the character in a video instead of skeletons, which makes motion transfer across species easier. Extensive experimental results and novel applications show our methods outperform baselines impressively. Project page is available at https://lhchen.top/Motion4Motion.
Chinese Translation
本研究探讨了从一个视频到另一个视频的运动转移，这在动画中对于多样化角色至关重要。此前，视频运动转移主要集中在人类与类人角色之间，这为数字创作提供了许多应用。然而，这些方法面临一个主要限制。具体而言，相关技术流程严重依赖于预定义的人体骨架结构，因此需要进行骨架条件模型训练。一方面，这些方法难以推广到多样化角色，例如来自不同物种的动物，同时保持其独特的运动风格。另一方面，具有多样骨架的标注数据有限，这进一步限制了该任务的大规模训练。在本文中，我们跳出了基于骨架的运动转移框架，提出了一种无训练的运动转移框架，命名为Motion4Motion。Motion4Motion对视频中角色的运动流进行建模，而不是骨架，这使得跨物种的运动转移变得更加容易。大量实验结果和新颖应用表明，我们的方法在性能上显著优于基线。项目页面可访问 https://lhchen.top/Motion4Motion。
cs.CV / 172 / 2607.11646
Event-RGB Adaptive Tracking for Nighttime Highway Perception
夜间高速公路感知的事件-RGB自适应跟踪
Abstract
Intelligent Transportation Systems deployed on highways predominantly rely on conventional RGB cameras for traffic perception and vehicle tracking. However, highway environments present unique challenges: the absence of artificial lighting infrastructure, combined with high vehicle velocities, results in severely degraded perception performance under low-light conditions. Specifically, nighttime scenarios suffer from motion blur, insufficient exposure, and poor signal-to-noise ratios, which catastrophically impair the reliability of RGB-based sensing systems. To address these limitations, we propose a novel Joint Event-RGB Adaptive Tracking (JEAT) framework. Unlike existing multi-sensor trackers constrained by rigid, hard-coded prioritization, JEAT merges asynchronous event streams and RGB frames into a unified joint data association optimization. By employing an Adaptive Extended Kalman Filter to continuously estimate measurement noise via NIS statistics, the framework dynamically weights and fuses both modalities, optimally harnessing event streams during dark or high-speed motion while leveraging RGB frames under bright or static conditions. Furthermore, given the absence of publicly available datasets tailored for event-based highway perception with diverse environmental conditions, we present SEHN, a large-scale synthetic dataset generated using the CARLA simulator. Our dataset encompasses diverse environmental conditions (daytime, nighttime, nighttime with out artificial lighting) and varying traffic densities, providing synchronized RGB imagery and event streams to facilitate multi-modal fusion research. Our code and datasets will be available at https://github.com/haidongwang96/SEHN.
Chinese Translation
部署在高速公路上的智能交通系统主要依赖传统的RGB摄像头进行交通感知和车辆跟踪。然而，高速公路环境面临独特的挑战：缺乏人工照明基础设施，加上车辆速度较快，导致在低光照条件下感知性能严重下降。具体而言，夜间场景受到运动模糊、曝光不足和信噪比低的影响，这严重损害了基于RGB的传感系统的可靠性。为了解决这些局限性，我们提出了一种新颖的联合事件-RGB自适应跟踪框架（Joint Event-RGB Adaptive Tracking，JEAT）。与现有的受限于刚性硬编码优先级的多传感器跟踪器不同，JEAT将异步事件流和RGB帧合并为统一的联合数据关联优化。通过采用自适应扩展卡尔曼滤波器（Adaptive Extended Kalman Filter）持续估计测量噪声，并利用NIS统计量，框架动态加权和融合两种模态，在黑暗或高速运动时最优利用事件流，而在明亮或静态条件下则利用RGB帧。此外，鉴于缺乏针对多样环境条件的事件基础高速公路感知的公开数据集，我们呈现了SEHN，一个使用CARLA模拟器生成的大规模合成数据集。我们的数据集涵盖多样的环境条件（白天、夜间、夜间无人工照明）和不同的交通密度，提供同步的RGB图像和事件流，以促进多模态融合研究。我们的代码和数据集将发布在https://github.com/haidongwang96/SEHN。
cs.CV / 173 / 2607.11655
Feature-Space Guided Diffusion for Realistic Ultrasound Image Synthesis
特征空间引导扩散用于真实感超声图像合成
Abstract
Conditional diffusion models can generate anatomically plausible medical ultrasound (US) images, but anatomical plausibility alone does not ensure realistic B-mode appearance. Most US pipelines adapt standard generative architectures and condition them on anatomical masks, or use guidance mechanisms that reinforce the same anatomical signal. However, B-mode US images are shaped by acquisition-dependent properties such as speckle texture, tissue contrast, and attenuation. Using a frozen US foundation model, we show that standard conditional diffusion baselines remain separated from real images in representation space. In this work, we propose Feature-Space Candidate Guidance (FSCG), a training-free sampling strategy to reduce this gap. At sampling time, FSCG applies local k-NN feature correction and selects the best of multiple stochastic candidates according to their feature-space energy. In this way, the mask defines the anatomy, while FSCG steers samples toward the real US domain. Across three different datasets, FSCG reduces average FID64 by 56\%, FID192 by 57\%, and nearest-neighbour feature distance by 47\% over standard conditional diffusion sampling, outperforming alternative inference-time guidance baselines. The results suggest that domain-aware feature representations can reveal and reduce realism gaps in medical diffusion synthesis without retraining the generator. Our code is available at https://github.com/marinadominguez/FSCG.
Chinese Translation
条件扩散模型能够生成解剖上合理的医学超声（US）图像，但仅有解剖合理性并不能确保真实的B模式外观。大多数超声图像处理流程适应标准生成架构，并基于解剖掩膜进行条件化，或使用强化相同解剖信号的引导机制。然而，B模式超声图像受到采集依赖性特性的影响，如斑点纹理、组织对比度和衰减。通过使用一个冻结的超声基础模型，我们展示了标准条件扩散基线在表示空间中与真实图像之间仍存在差距。在本研究中，我们提出了特征空间候选引导（Feature-Space Candidate Guidance, FSCG），这是一种无训练的采样策略，用于缩小这一差距。在采样时，FSCG应用局部k-NN特征校正，并根据特征空间能量选择多个随机候选中的最佳者。通过这种方式，掩膜定义了解剖结构，而FSCG则将样本引导至真实的超声领域。在三个不同的数据集上，FSCG将平均FID64降低了56%，FID192降低了57%，最近邻特征距离降低了47%，超越了标准条件扩散采样，并优于其他推理时引导基线。结果表明，领域感知特征表示能够揭示并减少医学扩散合成中的真实感差距，而无需重新训练生成器。我们的代码可在 https://github.com/marinadominguez/FSCG 获取。
cs.CV / 174 / 2607.11673
ABot-3DWorld 0: A Universal World Model to Explore Any 3D Space
ABot-3DWorld 0：探索任意3D空间的通用世界模型
Sun, Mingchao, Tang, Luyang, Liu, Yu, Yan, Xu, Li, Zhan, Zhang, Yunwei, Yu, Fei, Ge, Zengye, Liu, Yumin, Zhang, Jiacheng, Zhang, Yongchang, Zhang, Jiawei, Liu, Zhicheng, Sun, Zhongxu, Ouyang, Tianjian, Chen, Wenzheng, Yang, Shixing, Fan, Nianfei, Sun, Guodong, Li, Huan, Zhou, Zheng, Li, Yongze, Peng, Yingliang, Du, Mengmeng, Liu, Yuan, Shi, Haozhe, Gong, Chunnuo, Yu, Chengzhen, Jia, Chunxue, Liu, Yang, Zeng, Shiying, Lai, Junnan, Zhang, Hang, Guo, Ning, Chen, Baoquan, Xu, Mu, Pan, Hongyu
Abstract
We present ABot-3DWorld 0, a universal multimodal 3D world model that turns text, image, and video inputs into high-fidelity, explorable 3D worlds. At the heart of our framework is a unified Spatial Generative Primitive (SGP), a compact tuple of a high-quality panorama and a spatial point cloud that delivers an efficient description of any 3D space. Multimodal inputs are first lifted into this primitive; a 3D-consistent panoramic video generator then explores the primitive along a planned trajectory; finally, our panoramic video reconstruction engine converts the generated video into a clean, photorealistic 3D Gaussian Splatting (3DGS) world. This pipeline covers two regimes: rich inputs (multi-view sets, casual video) are lifted into the SGP through a geometry-rigorous recovery that mirrors the observed scene, while a single image or sentence is completed generatively into a creative world. The result is one low-barrier engine for general 3D content creation that further anchors generated worlds to geographic points of interest, enabling map-native spatial exploration at consumer scale. Experiments show that ABot-3DWorld 0 sets the state of the art among open-source methods and demonstrates stronger scene fidelity than Marble under rich multimodal inputs.
Chinese Translation
我们提出了ABot-3DWorld 0，这是一种通用的多模态3D世界模型，可以将文本、图像和视频输入转化为高保真、可探索的3D世界。我们框架的核心是一个统一的空间生成原语（Spatial Generative Primitive, SGP），它是一个高质量全景图和空间点云的紧凑元组，能够有效描述任何3D空间。多模态输入首先被提升到这个原语中；然后，一个3D一致性的全景视频生成器沿着规划的轨迹探索该原语；最后，我们的全景视频重建引擎将生成的视频转换为干净、逼真的3D高斯点云（3D Gaussian Splatting, 3DGS）世界。该流程涵盖了两种模式：丰富的输入（多视角集、随意视频）通过几何严谨的恢复提升到SGP，反映观察到的场景，而单幅图像或句子则生成性地完成到一个创意世界。最终结果是一个低门槛的通用3D内容创作引擎，进一步将生成的世界锚定到地理兴趣点，实现消费者规模的地图原生空间探索。实验表明，ABot-3DWorld 0在开源方法中设定了最新的技术水平，并在丰富的多模态输入下展现出比Marble更强的场景保真度。
cs.CV / 175 / 2607.11681
Illuminant-Adaptive 3D Lookup Tables for Camera Color Correction
适应光源的三维查找表用于相机色彩校正
Abstract
Color correction is a key component of camera image signal processing (ISP) pipelines, encompassing illuminant discounting and colorimetric mapping of device-dependent sensor responses to device-independent color spaces, such as CIE XYZ. Despite extensive research, accurate color correction remains challenging due to the non-linear relationship between camera sensor responses and CIE XYZ color space, as well as to the increasing presence of highly chromatic and spectrally complex LED illuminants. We propose a color correction framework based on illuminant-adaptive three-dimensional lookup tables (LUTs), which we call Color Correction LUT (C$^2$LUT). Our method combines a chromaticity-aware illuminant representation with a non-linear color transformation, enabling accurate correction under illuminants spanning a wide range of chromaticities and spectral complexities. We employ Tucker tensor decomposition to represent the LUTs, ensuring that computational requirements remain sufficiently low for deployment in camera ISPs. In addition, we introduce a large-scale illuminants dataset comprising 1,473 spectral power distributions, with different chromaticities and spectral profiles. Experiments across multiple cameras, illuminants, reflectance datasets, and real captured images demonstrate consistent improvements over existing methods for color correction, reducing CIE $\Delta E_{00}$ by up to 20% and angular error by up to 18% while remaining compatible with modern camera hardware constraints. Code and datasets are available at https://github.com/claudiom4sir/C2LUT.
Chinese Translation
色彩校正是相机图像信号处理（ISP）管道中的关键组成部分，涉及光源折扣和设备依赖传感器响应到设备无关色彩空间（如CIE XYZ）的色彩映射。尽管进行了广泛研究，准确的色彩校正仍然具有挑战性，原因在于相机传感器响应与CIE XYZ色彩空间之间的非线性关系，以及高度色彩丰富和光谱复杂的LED光源的日益普遍。我们提出了一种基于适应光源的三维查找表（LUT）的色彩校正框架，称为色彩校正LUT（C$^2$LUT）。我们的方法结合了色度感知的光源表示和非线性色彩变换，使得在跨越广泛色度和光谱复杂性的光源下实现准确校正成为可能。我们采用Tucker张量分解来表示LUT，确保计算需求保持在足够低的水平，以便在相机ISP中部署。此外，我们引入了一个大规模光源数据集，包含1,473种光谱功率分布，具有不同的色度和光谱特征。针对多种相机、光源、反射率数据集和实际捕获图像的实验表明，我们的方法在色彩校正上相较于现有方法有一致的改进，将CIE $ riangle E_{00}$降低了最多20%，将角度误差降低了最多18%，同时仍然兼容现代相机硬件约束。代码和数据集可在 https://github.com/claudiom4sir/C2LUT 获取。
cs.CV / 176 / 2607.11710
SVI360: Spherical Video Interpolation
SVI360：球面视频插值
Abstract
This paper addresses the problem of omnidirectional video interpolation, which plays an essential role in applications such as virtual reality and immersive video enhancement. Existing video interpolation methods are not well-suited for spherical videos, as they have difficulty handling severe distortions close to the poles. To address this issue, we propose SVI360, a dual-branch framework that combines the image frame and its rotated orthogonal view to deal with these distortions. The core methodological aspect of the approach is to reinforce equivariance of the flow displacements between the original and orthogonal views to improve intermediate frame prediction. Experiments show that our method outperforms state-of-the-art approaches in interpolation quality while maintaining accurate optical flow in four different public benchmarks. Code and pre-trained models are available at: https://icb-vision-ai.github.io/video360_interpolation/
Chinese Translation
本文解决了全向视频插值的问题，这在虚拟现实和沉浸式视频增强等应用中起着至关重要的作用。现有的视频插值方法不适合球面视频，因为它们在处理接近极点的严重畸变时存在困难。为了解决这一问题，我们提出了SVI360，一个双分支框架，结合了图像帧及其旋转正交视图，以应对这些畸变。该方法的核心在于增强原始视图与正交视图之间流位移的等变性，以改善中间帧的预测。实验表明，我们的方法在插值质量上优于最先进的方法，同时在四个不同的公共基准测试中保持了准确的光流。代码和预训练模型可在以下链接获取：https://icb-vision-ai.github.io/video360_interpolation/
cs.CV / 177 / 2607.11732
GFR-SAM: Training-Free Referring Camouflaged Object Segmentation via Cross-Image Prompting
GFR-SAM：通过跨图像提示实现无训练的指向性伪装物体分割
Abstract
Referring Camouflaged Object Detection (Ref-COD) requires segmenting hidden targets guided by reference cues. While supervised methods are annotation-heavy and training-free approaches via sparse point-prompting are sensitive to localization errors, we propose GFR-SAM, a robust three-stage training-free framework. GFR-SAM shifts the paradigm from fragile point-matching to a "Generate-Filter-Refine" pipeline. First, we introduce In-Context Exemplar-guided Segmentation, empowering SAM3 with cross-image inference to generate candidate masks via holistic visual exemplars, bypassing its native intra-image constraints. Second, a Region-Global Contrastive Filtering module ranks candidates through DINOv3-based prototypical alignment, effectively suppressing background distractors. Finally, a Geometric-Semantic Refinement module synergizes bounding box and text prompts to recover fine-grained boundaries and enhance instance recall. Evaluated on the R2C7K benchmark, GFR-SAM outperforms existing training-free methods by 8.7\% in weighted F-measure ($F_\beta^w$) and competes with supervised state-of-the-art counterparts. Ultimately, this work underscores the potential of unlocking SAM3's latent capability for cross-image In-Context prompting, establishing a robust, training-free paradigm that effectively bridges the gap between general-purpose foundation models and specialized, label-intensive perception tasks without the need for task-specific fine-tuning.
Chinese Translation
指向性伪装物体检测（Ref-COD）需要根据参考线索对隐藏目标进行分割。虽然监督方法依赖于大量标注，而通过稀疏点提示的无训练方法对定位误差敏感，我们提出了GFR-SAM，一个稳健的三阶段无训练框架。GFR-SAM将范式从脆弱的点匹配转变为“生成-过滤-细化”流程。首先，我们引入上下文示例引导分割，赋予SAM3跨图像推理的能力，通过整体视觉示例生成候选掩膜，绕过其固有的图内约束。其次，区域-全局对比过滤模块通过基于DINOv3的原型对齐对候选进行排名，有效抑制背景干扰。最后，几何-语义细化模块协同边界框和文本提示，恢复精细边界并增强实例召回。在R2C7K基准测试中，GFR-SAM在加权F-measure（$F_eta^w$）上比现有的无训练方法提高了8.7\%，并与监督的最先进方法相竞争。最终，本研究强调了释放SAM3潜在能力以进行跨图像上下文提示的潜力，建立了一个稳健的无训练范式，有效弥合了通用基础模型与专门、标注密集的感知任务之间的差距，而无需特定任务的微调。
cs.CV / 178 / 2607.11754
Higher-Order Cell Tracking Transformer
高阶细胞追踪变换器
Abstract
Reconstructing lineages from live-imaging microscopy requires linking cell detections across time, including through cell divisions. A common approach is to construct a candidate graph and associate cell segmentations (nodes) across frames. However, these and other existing methods overlook two structural obstacles in candidate tracking graphs: (i) cell divisions entangle distinct lineage paths in the node embedding space, and (ii) edges sharing a node have near-random label agreement, so the candidate-graph topology carries no useful information for graph neural networks to aggregate. We propose the \textbf{Higher-Order Cell Tracking Transformer} (HOCT), an edge-centric architecture in which candidate cell links attend to one another under a 3D geometric prior, resolving both issues. Evaluated on the Cell Tracking Challenge and a bacteria division benchmark, HOCT achieves state-of-the-art results without deep pre-trained image encoders. Moreover, the proposed approach is easier to fine-tune, quickly reducing tracking errors by 59% with 400 annotations in a human-in-the-loop setting, outperforming LoRA fine-tuning of competing transformer baselines (6.75% improvement).
Chinese Translation
从活体成像显微镜重建谱系需要在时间上链接细胞检测，包括细胞分裂。常见的方法是构建候选图，并在不同帧之间关联细胞分割（节点）。然而，这些方法以及其他现有方法忽视了候选追踪图中的两个结构性障碍：（i）细胞分裂在节点嵌入空间中交织了不同的谱系路径，以及（ii）共享一个节点的边缘具有近乎随机的标签一致性，因此候选图的拓扑结构对图神经网络的聚合没有有用的信息。我们提出了 extbf{高阶细胞追踪变换器}（Higher-Order Cell Tracking Transformer, HOCT），这是一种以边为中心的架构，其中候选细胞链接在三维几何先验下相互关注，从而解决了这两个问题。在细胞追踪挑战赛和细菌分裂基准测试中的评估表明，HOCT在没有深度预训练图像编码器的情况下实现了最先进的结果。此外，所提出的方法更易于微调，在人机协作环境中，通过400个标注快速将追踪错误减少了59%，超越了竞争性变换器基线的LoRA微调（提高了6.75%）。
cs.CV / 179 / 2607.11798
StoryTeller: Training-Free Narrative Grounding for Long-Form Audio Description
StoryTeller：无训练的长篇音频描述叙事基础
Abstract
Long-form audio description (AD) requires more than describing visible actions: it must preserve characters, events, relationships, and story context across scenes so that blind and low-vision (BLV) audiences can follow a film. Modern video-language models (VLMs) are effective on short clips, but they often treat each moment independently, producing descriptions that miss who characters are, why events matter, and how the current scene connects to earlier narrative context. We propose StoryTeller, a training-free framework for story-aware long-form AD. Instead of relying only on local visual cues, StoryTeller maintains a verified narrative memory that carries forward story-relevant information across scenes, enabling later descriptions to remain coherent, grounded, and contextually informative. Given only raw video and a movie title, StoryTeller can optionally retrieve public movie metadata to resolve names and story context, while accepting only facts that are supported by the video through semantic filtering and VLM verification. The method requires no subtitles, scripts, AD transcripts, aligned captions, character banks, precomputed face identities, or task-specific fine-tuning. To evaluate whether generated AD preserves narrative information, we introduce StoryAD-QA, a question-answering benchmark that tests whether a language model can answer story-context questions using only the generated descriptions. Experiments on standard AD benchmarks and diverse long-form videos show that StoryTeller consistently improves narrative coherence, factual grounding, and story comprehension over strong baselines in automatic, QA-based, and human evaluations.
Chinese Translation
长篇音频描述（AD）不仅需要描述可见的动作，还必须在场景之间保留角色、事件、关系和故事背景，以便盲人和低视力（BLV）观众能够跟随电影。现代视频语言模型（VLMs）在短片段上表现有效，但它们往往独立处理每个时刻，生成的描述常常忽视角色身份、事件重要性以及当前场景与早期叙事背景的联系。我们提出了StoryTeller，一个无训练的故事感知长篇AD框架。StoryTeller不仅依赖于局部视觉线索，而是维护一个经过验证的叙事记忆，跨场景传递与故事相关的信息，使后续描述保持连贯、扎根于事实且具有背景信息。仅凭原始视频和电影标题，StoryTeller可以选择性地检索公共电影元数据以解析名称和故事背景，同时仅接受通过语义过滤和VLM验证支持的事实。该方法不需要字幕、剧本、AD转录、对齐字幕、角色库、预计算的人脸身份或特定任务的微调。为了评估生成的AD是否保留叙事信息，我们引入了StoryAD-QA，一个问答基准，测试语言模型是否能够仅使用生成的描述回答故事背景问题。在标准AD基准和多样化的长篇视频上的实验表明，StoryTeller在自动、基于问答和人工评估中，始终改善叙事连贯性、事实基础和故事理解，相较于强基线表现更佳。
cs.CV / 180 / 2607.11818
MM-ToolSandBox: A Unified Framework for Evaluating Visual Tool-Calling Agents
MM-ToolSandBox：一个统一的视觉工具调用代理评估框架
Abstract
We introduce MM-ToolSandBox, a benchmark and evaluation framework for visually grounded tool-calling agents. The framework provides a stateful execution environment spanning 500+ tools across 16 application domains, supporting multi-image, multi-turn tasks where agents must ground progressively arriving visual inputs into executable tool calls while handling realistic conversational phenomena (goal revisions, error corrections, state mutations). An automated scenario generation pipeline produces diverse, visually grounded scenarios through information-flow-guided planning and multi-stage quality filtering, yielding 258 human-verified nominal scenarios and 50 variants targeting interactive UI applications. Evaluating 12 state-of-the-art models, from 4B open-weight to frontier proprietary systems, shows that current models still lack robust visual tool-calling capability: even the best model achieves below 50% success rate. Our failure analysis further reveals that visual precision, not only planning, is a primary bottleneck for capable models: 53% of failures stem from incorrect information extraction from images despite otherwise correct task workflows. A planning-to-precision crossover emerges with scale: smaller models fail at deciding what to do, while larger models fail at perceiving what they see, suggesting fundamentally different research directions for improving models at different capability levels. The framework and the benchmark are publicly available at https://github.com/apple/ml-mmtoolsandbox
Chinese Translation
我们介绍了MM-ToolSandBox，这是一个针对视觉基础工具调用代理的基准和评估框架。该框架提供了一个状态保持的执行环境，涵盖了16个应用领域中的500多种工具，支持多图像、多轮次任务，代理必须将逐步到达的视觉输入转化为可执行的工具调用，同时处理现实对话现象（目标修订、错误纠正、状态变更）。一个自动化场景生成管道通过信息流引导的规划和多阶段质量过滤生成多样化的、视觉基础的场景，产生了258个经过人工验证的标准场景和50个针对交互式用户界面应用的变体。对12个最先进模型的评估，从4B开放权重到前沿专有系统，显示出当前模型在视觉工具调用能力上仍然缺乏稳健性：即使是最佳模型的成功率也低于50%。我们的失败分析进一步揭示，视觉精度而非仅仅是规划，是能力模型的主要瓶颈：53%的失败源于从图像中提取信息不正确，尽管任务工作流程在其他方面是正确的。随着规模的扩大，规划与精度之间出现了交叉：较小的模型在决定做什么时失败，而较大的模型在感知所见时失败，这表明在不同能力水平上改善模型的研究方向存在根本性差异。该框架和基准可在 https://github.com/apple/ml-mmtoolsandbox 上公开获取。
cs.CV / 181 / 2607.11830
MicroCharNet: Less is More for License Plate Character Detection
MicroCharNet：更少即更多的车牌字符检测
Abstract
License plate character detection is a crucial component of intelligent transportation systems, where high accuracy and computational efficiency are required for real-time deployment. Although recent deep learning-based methods have substantially improved detection performance, many high-accuracy models rely on large-scale architectures that incur substantial computational overhead, limiting their applicability to resource-constrained devices. In this paper, we propose MicroCharNet, an ultra-lightweight model specifically designed for license plate character detection. The proposed architecture employs a compact backbone composed of C2f blocks, integrated with CoordAtt module to enhance feature extraction while preserving spatial information. A lightweight C3k2-based neck fuses multi-level features, followed by a single-level anchor-free detection head that enables end-to-end prediction. Experiments conducted on the UFPR-ALPR dataset demonstrate that MicroCharNet achieves competitive detection accuracy with only 0.08M parameters and 0.096 GFLOPs, while outperforming several recent YOLO-based baselines. Hardware-level evaluations further confirm its efficiency for real-time deployment on edge devices. These results indicate that carefully designed ultra-lightweight architectures can effectively balance accuracy and efficiency in license plate character detection. The source code is available at https://github.com/chequanghuy/MicroCharNet.
Chinese Translation
车牌字符检测是智能交通系统中的一个关键组成部分，要求在实时部署中具备高准确性和计算效率。尽管最近基于深度学习的方法显著提高了检测性能，但许多高准确度模型依赖于大规模架构，导致显著的计算开销，限制了它们在资源受限设备上的应用。在本文中，我们提出了MicroCharNet，一种专门为车牌字符检测设计的超轻量级模型。所提出的架构采用由C2f模块构成的紧凑型主干，并结合CoordAtt模块以增强特征提取，同时保留空间信息。轻量级的C3k2基础颈部融合多层特征，随后是一个单层无锚检测头，实现端到端预测。在UFPR-ALPR数据集上进行的实验表明，MicroCharNet以仅0.08M的参数和0.096 GFLOPs的计算量实现了具有竞争力的检测准确性，同时超越了几种最新的基于YOLO的基线。硬件级评估进一步确认了其在边缘设备上实时部署的效率。这些结果表明，精心设计的超轻量级架构能够有效平衡车牌字符检测中的准确性和效率。源代码可在https://github.com/chequanghuy/MicroCharNet获取。
cs.CV / 182 / 2607.11836
Cycle-World: Mitigating Error Accumulation in Long-term Video World Models via Reverse-Prediction Cycle Consistency
Cycle-World：通过反向预测循环一致性减轻长期视频世界模型中的误差积累
Abstract
Autoregressive diffusion models have enabled high-quality video generation, yet their sequential nature inherently suffers from error accumulation. In long-horizon video synthesis, minor prediction deviations compound over time, inevitably leading to unconstrained generative drift, structural collapse, and severe visual degradation. To address this, we propose Cycle-World, a novel framework designed for stable and temporally consistent long-video generation. Our approach tackles error drift by enforcing strict temporal reversibility across both the training and inference phases. Theoretically, we demonstrate that forward generative drift can be strictly bottlenecked by a cycle-consistency objective. During training, we integrate an efficient reverse-prediction model to implicitly embed causal constraints into the forward generator, compelling it to produce reversible sequences that tightly adhere to the natural video manifold. At inference time, we repurpose this frozen reverse model as a runtime corrector. Through gradient-based cycle guidance, it iteratively refines the generated latent representations, actively suppressing accumulated errors before they are committed to the historical context. Extensive experiments on the VBench benchmark demonstrate that Cycle-World's dual-phase synergy significantly mitigates error drift, achieving state-of-the-art overall generation quality and long-horizon temporal consistency in 60-second synthesis.
Chinese Translation
自回归扩散模型已实现高质量视频生成，但其顺序特性固有地导致误差积累。在长时间跨度的视频合成中，微小的预测偏差随着时间的推移而累积，必然导致不受限制的生成漂移、结构崩溃和严重的视觉退化。为了解决这一问题，我们提出了Cycle-World，一个旨在实现稳定和时间一致的长视频生成的新框架。我们的方法通过在训练和推理阶段强制执行严格的时间可逆性来应对误差漂移。从理论上讲，我们证明了前向生成漂移可以通过循环一致性目标严格限制。在训练过程中，我们集成了一个高效的反向预测模型，以隐式地将因果约束嵌入前向生成器，迫使其生成紧密遵循自然视频流形的可逆序列。在推理时，我们将这个冻结的反向模型重新用于运行时校正。通过基于梯度的循环引导，它迭代地优化生成的潜在表示，主动抑制累积的误差，防止其在历史上下文中被固定。对VBench基准的广泛实验表明，Cycle-World的双阶段协同显著减轻了误差漂移，在60秒合成中实现了最先进的整体生成质量和长期时间一致性。
cs.CV / 183 / 2607.11838
HASTE: A Platform for Rapid Post-Disaster Building Damage Assessment
HASTE：快速灾后建筑损害评估平台
Abstract
When a large disaster strikes, responders need a map of which buildings are damaged within hours. The models that do well on public benchmarks assume matched before-and-after imagery and a training set drawn from similar past events, and neither is usually available for a new disaster in its first day. We present HASTE (High-speed Assessment and Satellite Tracking for Emergencies), a no-code web platform that lets analysts who are not machine learning engineers produce per-building damage maps from post-disaster satellite imagery. HASTE implements two methods that share one interface. The first requires the user to label polygons over the post-disaster scene, trains a small semantic segmentation model on that single scene, runs it over the whole image, and joins the per-pixel output to existing building footprints. The second embeds every footprint with a pretrained vision model, requires the user to label a handful of buildings, and fits a logistic regression in the browser that scores the rest of the scene in seconds. We describe the platform, both methods, and the engineering that supports them. We also report preliminary experiments on xBD showing that foundation-model embeddings pooled over footprints separate damaged from intact buildings using post-disaster imagery alone, matching a fully supervised ResNet-50 baseline with a twentieth of its labels. HASTE and its predecessors have supported more than thirty real-world disaster responses since 2023, spanning earthquakes, hurricanes, cyclones, floods, wildfires, and tornadoes, delivering results to humanitarian partners within hours to days of imagery becoming available. We close with the directions we think are most promising, including vision-language assessment, active learning, and damage models for roads and other infrastructure. HASTE is open source at https://github.com/microsoft/haste.
Chinese Translation
当大型灾害发生时，响应人员需要在数小时内获得受损建筑的地图。表现良好的模型通常假设存在匹配的灾前和灾后影像，并且训练集来自类似的过去事件，而这两者在新灾害发生的第一天通常都不可用。我们提出了HASTE（高效应急评估与卫星追踪平台），这是一个无代码的网络平台，使得非机器学习工程师的分析人员能够从灾后卫星影像中生成每栋建筑的损害地图。HASTE实现了两种共享一个接口的方法。第一种方法要求用户在灾后场景上标记多边形，基于该单一场景训练一个小型语义分割模型，然后在整个影像上运行该模型，并将每像素的输出与现有建筑轮廓结合。第二种方法则是将每个建筑轮廓嵌入一个预训练的视觉模型，要求用户标记少量建筑，并在浏览器中拟合一个逻辑回归模型，在几秒钟内对场景的其余部分进行评分。我们描述了该平台、这两种方法以及支持它们的工程技术。我们还报告了在xBD上的初步实验，显示基础模型嵌入在建筑轮廓上进行池化，可以仅使用灾后影像将受损建筑与完好建筑区分开，匹配了一个完全监督的ResNet-50基线模型，且只使用了其标签的五分之一。自2023年以来，HASTE及其前身已支持超过三十次真实世界的灾害响应，涵盖地震、飓风、气旋、洪水、野火和龙卷风，在影像可用后的数小时到数天内向人道主义合作伙伴提供结果。我们最后讨论了我们认为最有前景的发展方向，包括视觉-语言评估、主动学习以及道路和其他基础设施的损害模型。HASTE的开源地址为https://github.com/microsoft/haste。
cs.CV / 184 / 2607.11839
LoRA-Based Cascaded Multimodal Fusion for Action Recognition in Medical Training Environments
基于LoRA的级联多模态融合在医疗培训环境中的动作识别
Abstract
This paper presents a cascaded Low-Rank Adaptation (LoRA)-based multimodal fusion framework for action and activity recognition in healthcare-oriented training environments. The proposed architecture combines parameter-efficient modality-specific adaptation with sequential fusion, enabling modalities to be integrated in stages without retraining previously learned components. Rather than assuming a fixed fusion structure, the framework first integrates more closely related modalities and then incorporates additional heterogeneous modalities, supporting scalable adaptation across datasets with different modality sets.We evaluate the framework on two healthcare-oriented training environment datasets: NurViD and the Nurse Training dataset. Across these datasets, preliminary results suggest that the proposed cascaded fusion strategy improves over individual modality models and provides competitive performance relative to previously reported dataset-specific baselines. Overall, these findings indicate that cascaded LoRA-based fusion is a promising parameter-efficient approach for integrating heterogeneous modalities in medical training action and activity recognition tasks. github: https://github.com/anonymous0-ai/LoRA-Based-Cascaded-Multimodal-Fusion-.git.
Chinese Translation
本文提出了一种基于低秩适应（LoRA）的级联多模态融合框架，用于医疗导向培训环境中的动作和活动识别。所提架构结合了参数高效的模态特定适应与顺序融合，使得模态能够分阶段集成，而无需重新训练先前学习的组件。该框架并不假设固定的融合结构，而是首先整合更紧密相关的模态，然后再加入额外的异构模态，从而支持在具有不同模态集的数据集之间进行可扩展的适应。我们在两个医疗导向培训环境数据集上评估了该框架：NurViD和护士培训数据集。在这些数据集中，初步结果表明，所提的级联融合策略在性能上优于单一模态模型，并且相较于先前报告的数据集特定基准提供了具有竞争力的表现。总体而言，这些发现表明，基于LoRA的级联融合是一种有前景的参数高效方法，适用于在医疗培训的动作和活动识别任务中集成异构模态。
cs.CV / 185 / 2607.11844
Beyond the Single Camera: Agentic Multi-View Reasoning in Sports Video Understanding
超越单一摄像头：体育视频理解中的主动多视角推理
Abstract
Recent Multimodal Large Language Models (MLLMs) achieve strong performance on single-view video understanding benchmarks. However, sports videos involve dense occlusion, rapid motion, and complex interactions that are difficult to resolve from a single viewpoint. In practice, sports events are recorded from multiple camera angles, providing complementary evidence used by referees. Yet, no existing benchmark evaluates MLLMs on multi-view sports video understanding. To address this gap, we introduce SportMV-Bench, a comprehensive benchmark built from official match recordings, through a dedicated pipeline combining LLM-based generation, MLLM-based verification, and human filtering to ensure quality and consistency. SportMV-Bench containing 787 multi-view video bundles and 2592 question-answer pairs across three categories: Perception-Aware Recognition (PAR), Rule-aware Event Interpretation (REI), and Adjudicative Decision Reasoning(ADR). Our analysis shows that current MLLMs fail to effectively exploit multi-view information, with the bottlenecks lying in fine-grained visual perception and view selection rather than logical reasoning or domain knowledge. We propose SportMV-Agent, an agentic framework that orchestrates an iterative loop of active view selection, perception tool execution, and evidence-grounded reasoning, achieving a significant 14.46% relative improvement over the strongest MLLM baseline.
Chinese Translation
近期的多模态大型语言模型（MLLMs）在单视角视频理解基准上表现出色。然而，体育视频涉及密集遮挡、快速运动和复杂交互，这些在单一视角下难以解决。实际上，体育赛事通常从多个摄像角度录制，为裁判提供互补证据。然而，目前没有现有基准评估MLLMs在多视角体育视频理解上的表现。为了解决这一空白，我们推出了SportMV-Bench，这是一个基于官方比赛录制的综合基准，通过专门的流程结合基于LLM的生成、基于MLLM的验证和人工筛选，以确保质量和一致性。SportMV-Bench包含787个多视角视频集和2592个问题-答案对，分为三个类别：感知意识识别（Perception-Aware Recognition, PAR）、规则意识事件解释（Rule-aware Event Interpretation, REI）和裁决决策推理（Adjudicative Decision Reasoning, ADR）。我们的分析表明，当前的MLLMs未能有效利用多视角信息，瓶颈在于细粒度视觉感知和视角选择，而非逻辑推理或领域知识。我们提出了SportMV-Agent，一个主动框架，协调主动视角选择、感知工具执行和基于证据的推理的迭代循环，实现了相对于最强MLLM基线显著的14.46%的相对提升。
cs.CV / 186 / 2607.11862
Evidence-Backed Video Question Answering
基于证据的视频问答
Abstract
Current Video Large Language Models (Video LLMs) excel in question answering (QA) but largely operate as black boxes, providing textual answers without verifiable visual grounding. Existing explainability efforts rely on textual rationales or sparse bounding boxes, which struggle to capture complex video dynamics such as occlusions and non-rigid deformations. We propose Evidence-Backed Video Question Answering (E-VQA), a novel task requiring models to jointly output a semantic answer and precise spatio-temporal evidence: temporal segments and dense, tracked object segmentation masklets. To support this, we introduce ST-Evidence, the first human-verified benchmark for both discriminative and generative pixel-level grounding. Evaluations of state-of-the-art models reveal a critical decoupling between QA accuracy and true visual perception that scaling alone fails to bridge. To address this, we develop scalable, automated generation pipelines to create ST-Evidence-Instruct, a 160k-scale dataset bridging high-level reasoning with fine-grained grounding. Fine-tuning grounded Video LLMs on this data yields substantial gains over the corresponding size-matched UniPixel baselines (e.g., +27.2 t-mean and +13.8 J&F on a 7B model), establishing a robust baseline for explainable, evidence-backed video understanding. Code and data are available at https://github.com/SalesforceAIResearch/EVQA.
Chinese Translation
当前的视频大型语言模型（Video LLMs）在问答（QA）方面表现出色，但大多作为黑箱操作，提供文本答案而没有可验证的视觉依据。现有的可解释性努力依赖于文本推理或稀疏的边界框，这些方法难以捕捉复杂的视频动态，例如遮挡和非刚性变形。我们提出了基于证据的视频问答（Evidence-Backed Video Question Answering, E-VQA），这是一项新任务，要求模型共同输出语义答案和精确的时空证据：时间段和密集的、跟踪的物体分割掩膜。为此，我们引入了ST-Evidence，这是第一个经过人工验证的基准，适用于区分性和生成性像素级的基础。对最先进模型的评估揭示了QA准确性与真实视觉感知之间的关键解耦，仅靠扩展无法弥补这一差距。为了解决这个问题，我们开发了可扩展的自动生成管道，创建了ST-Evidence-Instruct，这是一个规模为16万的数据集，连接高层次推理与细粒度的基础。在此数据上对有基础的Video LLMs进行微调，显著提高了与相应规模匹配的UniPixel基线（例如，在7B模型上，t-mean提高了27.2，J&F提高了13.8），为可解释的、基于证据的视频理解建立了稳健的基线。代码和数据可在 https://github.com/SalesforceAIResearch/EVQA 获取。
cs.CV / 187 / 2607.11885
Latent-Identity Tuning in Text-to-Image Personalization Models
文本到图像个性化模型中的潜在身份调优
Abstract
Generating and editing a person's face demands high precision, as even minor modifications can significantly alter a subject's perceived identity. Current personalization and editing methods built on general-purpose text-to-image models, however, often lack the precision required for fine-grained facial edits. We present a method for fine-grained identity tuning in text-to-image personalization models. Unlike standard image editing, which operates on a given image, identity tuning modifies the latent representation of a specific identity, enabling the generation of diverse images that consistently depict the same edited identity. To enable fine-grained latent identity tuning, we explore the latent space of a pre-trained, frozen encoder for text-to-image personalization. Our approach requires no additional training. Instead, it leverages the existing architecture of a frozen encoder to uncover latent semantic directions. This space consists of a set of latent tokens that play distinct roles in capturing different aspects of an identity and often correspond to specific spatial or semantic facial regions. We show that meaningful directions can be identified within this space and within subspaces defined by selected tokens, enabling localized, fine-grained, and semantically coherent edits. We validate our approach through qualitative and quantitative experiments that demonstrate diverse localized facial edits while preserving cross-image identity consistency. Project page at: https://garibida.github.io/IdentityTuning/
Chinese Translation
生成和编辑一个人的面孔需要高精度，因为即使是微小的修改也可能显著改变一个主体的感知身份。然而，基于通用文本到图像模型的当前个性化和编辑方法，往往缺乏进行细粒度面部编辑所需的精度。我们提出了一种在文本到图像个性化模型中进行细粒度身份调优的方法。与在给定图像上进行的标准图像编辑不同，身份调优修改的是特定身份的潜在表示，从而能够生成多样化的图像，这些图像始终描绘相同的编辑身份。为了实现细粒度的潜在身份调优，我们探索了一个预训练的、固定的编码器的潜在空间，该编码器用于文本到图像个性化。我们的方法不需要额外的训练，而是利用固定编码器的现有架构来揭示潜在的语义方向。这个空间由一组潜在标记组成，这些标记在捕捉身份的不同方面时发挥着不同的作用，并且通常对应于特定的空间或语义面部区域。我们展示了在这个空间以及由选定标记定义的子空间中，可以识别出有意义的方向，从而实现局部、细粒度和语义一致的编辑。我们通过定性和定量实验验证了我们的方法，展示了多样化的局部面部编辑，同时保持跨图像的身份一致性。项目页面：https://garibida.github.io/IdentityTuning/
cs.CV / 188 / 2607.11886
Read It Back: Pretrained MLLMs Are Zero-Shot Reward Models for Text-to-Image Generation
回读：预训练的多语言大模型作为文本到图像生成的零-shot 奖励模型
Abstract
In this paper, we propose SpectraReward, a training-free reward function that turns pretrained MLLMs into off-the-shelf reward models for image-generation reinforcement learning. Instead of asking the MLLM to judge a generated image or answer decomposed verification questions, SpectraReward measures how well the original prompt can be recovered from the generated image through a single image-conditioned, teacher-forced forward pass. We use the average image-conditioned prompt log-likelihood as the reward, directly reusing the MLLM's pretrained image-text alignment ability without preference labels, reward-model fine-tuning. We further introduce Self-SpectraReward, a special case for unified multimodal models where the policy's own understanding branch serves as the reward model for its generation branch, forming a closed-loop self-improving framework without external reward models or external knowledge. Extensive experiments validate SpectraReward through a broad image-generation RL study covering two diffusion models, three RL algorithms, nine reward MLLM backbones from four MLLM families spanning 4B to 235B parameters, and five out-of-distribution text-to-image benchmarks. Results show that both SpectraReward and Self-SpectraReward significantly and consistently improve generation performance and outperform prior MLLM-derived reward training methods. Further analysis reveals that larger reward MLLMs are not always better, while Self-SpectraReward can match or surpass much larger external reward models, suggesting that reward-policy alignment is a key factor for effective image-generation RL. Project Page: https://huangrh99.github.io/SpectraReward/
Chinese Translation
在本文中，我们提出了SpectraReward，一种无训练的奖励函数，将预训练的多语言大模型（MLLM）转变为现成的图像生成强化学习奖励模型。与其要求MLLM判断生成的图像或回答分解的验证问题，SpectraReward通过单次图像条件的教师强制前向传播来衡量从生成图像中恢复原始提示的效果。我们使用平均图像条件提示的对数似然作为奖励，直接重用MLLM的预训练图像-文本对齐能力，而无需偏好标签或奖励模型微调。我们进一步引入Self-SpectraReward，这是一个统一多模态模型的特例，其中策略自身的理解分支作为其生成分支的奖励模型，形成一个无需外部奖励模型或外部知识的闭环自我改进框架。大量实验通过广泛的图像生成强化学习研究验证了SpectraReward，涵盖了两个扩散模型、三个强化学习算法、来自四个MLLM家族的九个奖励MLLM骨干网络，参数范围从40亿到235亿，以及五个分布外的文本到图像基准。结果表明，SpectraReward和Self-SpectraReward显著且一致地提高了生成性能，并超越了先前基于MLLM的奖励训练方法。进一步分析显示，较大的奖励MLLM并不总是更好，而Self-SpectraReward可以与更大的外部奖励模型相匹敌或超越，这表明奖励-策略对齐是有效图像生成强化学习的关键因素。项目页面：https://huangrh99.github.io/SpectraReward/
cs.AI / 1 / 2607.09664
From ML Predictions to Informed Diagnostic Assistance Using the Toulmin Model of Argumentation
从机器学习预测到基于图尔敏论证模型的知情诊断辅助
Abstract
To provide a structured and interpretable assessment, we decompose the image-based diagnosis into components following the Toulmin model of argumentation. This model consists of a claim, grounds, warrant, qualifier, rebuttal, and backing. Consider a claim generated by a machine learning (ML) model for retinal diagnosis. Rather than accepting this claim at face value, one could either apply explainable AI (XAI) methods or adopt an argumentation-based approach. In our framework, a model specialized in biomarker extraction from images provides the grounds. The warrant-linking the grounds to the claim - is analyzed by an agent equipped with medical knowledge; in our architecture, this role is fulfilled by a MedGemma agent. The qualifier is determined based on the overall quantitative evaluation of both the warrant and grounds models. Finally, a rebuttal is constructed using image similarity measures computed with MedSigLip. All these components are presented to the human expert, enabling a more informed and critical assessment of the ML-generated diagnosis.
Chinese Translation
为了提供结构化和可解释的评估，我们将基于图像的诊断分解为遵循图尔敏论证模型的组成部分。该模型由主张、依据、担保、限定、反驳和支持组成。考虑一个由机器学习（ML）模型生成的视网膜诊断主张。与其直接接受这一主张，不如应用可解释人工智能（XAI）方法或采用基于论证的方法。在我们的框架中，一个专门用于从图像中提取生物标志物的模型提供依据。将依据与主张联系起来的担保由一个具备医学知识的代理进行分析；在我们的架构中，这一角色由MedGemma代理承担。限定是基于担保和依据模型的整体定量评估来确定的。最后，使用MedSigLip计算的图像相似性度量构建反驳。所有这些组成部分都呈现给人类专家，从而使其能够对机器学习生成的诊断进行更为知情和批判性的评估。
cs.AI / 2 / 2607.09665
Format Sensitivity Index: Token-Controlled Prompt Wrapper Robustness and Schema Compliance in LLM Benchmarking
格式敏感性指数：令牌控制的提示包装器在大型语言模型基准测试中的鲁棒性和模式合规性
Abstract
Prompt wrappers often differ only in formatting, yet they can change model scores enough to flip leaderboard conclusions. We study this variance under a token-controlled protocol and introduce two complementary metrics: the Format Sensitivity Index (FSI), the accuracy range induced by wrapper choice, and the Parseability Sensitivity Index (PSI), the corresponding range in answer parseability. Across 140,000 OpenRouter generations spanning 7 QA tasks, 5 wrapper families, and 4 instruct models from 7B to 72B parameters, we find that mean FSI varies by over 30x across models and is largely explained by compliance failures. A fixed-effects regression shows that parseability remains a strong predictor of accuracy even after controlling for task, model, and wrapper. We argue that reporting accuracy without wrapper variance and compliance is statistically fragile, and we give practical recommendations for both benchmarking and structured-output deployments.
Chinese Translation
提示包装器通常仅在格式上有所不同，但它们可以显著改变模型得分，从而颠覆排行榜的结论。我们在令牌控制的协议下研究这种变异，并引入两个互补的指标：格式敏感性指数（Format Sensitivity Index, FSI），即由包装器选择引起的准确性范围，以及可解析性敏感性指数（Parseability Sensitivity Index, PSI），即相应的答案可解析性范围。在涵盖140,000个OpenRouter生成的样本、7个问答任务、5个包装器家族和4个从7B到72B参数的指令模型的研究中，我们发现不同模型的平均FSI变化超过30倍，且主要由合规性失败所解释。固定效应回归分析表明，即使在控制任务、模型和包装器后，可解析性仍然是准确性的强预测因子。我们认为，在没有考虑包装器变异和合规性的情况下报告准确性是统计上脆弱的，并为基准测试和结构化输出部署提供了实用建议。
cs.AI / 3 / 2607.09678
Faithful, Not Corrective: Message-Format Effects in Multi-Hop Agent Relays Are Tier-Dependent
忠实而非修正：多跳代理中消息格式效应依赖于层级
Abstract
When LLM agents hand off information to one another, does the message format matter? Two literatures disagree: format-optimization work reports that structured messages cut cost without hurting accuracy, while format-restriction work finds that imposing structure degrades generation -- and neither measures what happens when a message traverses multiple hops, where copy fidelity, not one-shot generation, dominates. We introduce a controlled relay testbed: briefs of twelve programmatically generated atomic facts are re-encoded hop-by-hop in five formats (free NL, precision-instructed NL, JSON, triples, key-value) over six hops, scored by a fixed strong grader against programmatic ground truth, across two relay-capability tiers, a cognitive-load condition, and a paired-fork error injection. We find that message-format effects are tier-dependent. (i) Under faithful-relay instructions a strong relay is nearly lossless -- the documented "telephone-game" collapse does not occur -- and adding per-hop cognitive load leaves format-level fidelity unchanged (within +/-1.8 points) while raising generation cost by 24-53%. (ii) Under a weak (1.5B) relay the across-format spread of six-hop recall grows by a factor of 8.7 (from 2.3 to 20.5 points), driven by two opposing mechanisms -- an encoding toll paid by the rigid formats and drift resistance specific to the fixed-key JSON schema -- that flip the format ranking in transit. (iii) In a paired-fork injection, an injected wrong value, once present, persists to the final hop in 83-100% of chains in every format, closely matching each format's retention of the true value, with no detectable collateral damage to neighboring facts. Structure buys a faithful, error-localizing channel -- not an error-correcting code -- and format choice should follow the weakest relay in the pipeline.
Chinese Translation
当大型语言模型（LLM）代理相互传递信息时，消息格式是否重要？两种文献存在分歧：格式优化研究报告称结构化消息在不影响准确性的情况下降低成本，而格式限制研究发现施加结构会降低生成效果——而且两者都未测量消息在多个跳跃中传递时的情况，在这种情况下，复制保真度而非一次性生成占主导地位。我们引入了一个受控的中继测试平台：十二个程序生成的原子事实的简要信息在六个跳跃中以五种格式（自由自然语言、精确指令自然语言、JSON、三元组、键值对）逐跳重新编码，由固定的强评估者根据程序化的真实值进行评分，涵盖两个中继能力层级、一个认知负荷条件和一个配对分叉错误注入。我们发现消息格式效应依赖于层级。(i) 在忠实中继指令下，强中继几乎无损——文献中记录的“电话游戏”崩溃并未发生——而增加每跳的认知负荷使格式级别的保真度保持不变（在±1.8分内），同时将生成成本提高了24-53%。(ii) 在弱（1.5B）中继下，六跳回忆的跨格式差异增长了8.7倍（从2.3到20.5分），由两种相反机制驱动——刚性格式所付出的编码代价和特定于固定键JSON模式的漂移抵抗——这在传输中翻转了格式排名。(iii) 在配对分叉注入中，一旦注入错误值，该值在每种格式的83-100%的链中持续存在至最后一跳，紧密匹配每种格式对真实值的保留，且对邻近事实没有可检测的附带损害。结构提供了一个忠实的、错误定位的通道——而非错误纠正码——格式选择应遵循管道中最弱的中继。
cs.AI / 4 / 2607.09689
Boltzmann MapReduce: A Partition-Function Reduce for Forkable Sandboxes
玻尔兹曼 MapReduce：可分叉沙箱的分区函数归约
Abstract
To leading order under local asymptotic normality (LAN), the confidence density a worker emits over a chunk of size $n$ is a Gibbs--Boltzmann measure $\exp\{-\beta E(\theta)\}$ whose inverse temperature is the sample size, $\beta=n$. Three consequences are exact in the Gaussian/linear case and first-order otherwise: disjoint chunks carry independent Boltzmann factors, so the MapReduce \emph{reduce}, read literally, is a partition function $Z=\int\prod_k h_k\,d\theta$ whose mode is precision-weighted (inverse-variance) pooling; frequentist consistency is the zero-temperature limit $T=1/n\to0$
Chinese Translation
在局部渐近正态性（LAN）下的主导阶，工人在大小为 $n$ 的数据块上发出的置信密度是一个吉布斯-玻尔兹曼测度 $ ext{exp}\{-eta E(\theta)\}$，其逆温度为样本大小 $eta=n$。在高斯/线性情况下，三个结果是精确的，而在其他情况下为一阶近似：不相交的数据块携带独立的玻尔兹曼因子，因此 MapReduce 的 extit{reduce} 字面意义上是一个分区函数 $Z=\int\prod_k h_k\,d\theta$，其模式是加权精度（逆方差）汇聚；频率一致性是零温度极限 $T=1/n\to0$。
cs.AI / 5 / 2607.09698
Interpreting Latent CoT Reasoning as Dynamical Systems
将潜在链式推理解释为动力系统
Abstract
Recent latent reasoning methods, such as CODI and COCONUT, face a fundamental interpretability problem: they maintain multiple superimposed candidate traces in the hidden space at each step, unlike explicit- CoT, which follows a single transparent reasoning trace. Existing mechanistic methods show compression, shortcuts, and superposition without explaining how reasoning evolves across latent steps. To address this gap, we model latent token sequences as trajectories in representation space and apply dynamical systems analysis to characterize the evolution of reasoning. Using quantitative measures, such as step-to-step change, direction consistency, and Lyapunov sensitivity, alongside qualitative projections, such as UMAP and DMD/PHATE, we show that latent CoT exhibits structured, non-random dynamics with two distinct stability classes. CODI behaves as a stable attractor, while COCONUT behaves as an unstable expanding system, and SIM-CoT supervision tightens both behaviors without changing the underlying dynamics. This framework advances the interpretability of latent CoT reasoning dynamics and provides actionable insights for improving latent reasoning performance. Code1 and Project page2 available online.
Chinese Translation
最近的潜在推理方法，如CODI和COCONUT，面临一个基本的可解释性问题：它们在每一步中在隐藏空间中保持多个叠加的候选轨迹，而不同于显式链式推理（explicit-CoT），后者遵循单一透明的推理轨迹。现有的机械性方法展示了压缩、捷径和叠加，但未能解释推理如何在潜在步骤中演变。为了解决这一问题，我们将潜在的标记序列建模为表示空间中的轨迹，并应用动力系统分析来表征推理的演变。通过定量指标，如逐步变化、方向一致性和李雅普诺夫敏感性，以及定性投影，如UMAP和DMD/PHATE，我们展示了潜在链式推理展现出结构化的、非随机的动态，具有两种不同的稳定性类别。CODI表现为一个稳定的吸引子，而COCONUT则表现为一个不稳定的扩展系统，SIM-CoT监督在不改变基础动态的情况下收紧了这两种行为。该框架提升了潜在链式推理动态的可解释性，并为改善潜在推理性能提供了可操作的见解。代码和项目页面可在线获取。
cs.AI / 6 / 2607.09706
YUKTI: From Natural-Language Situations to Robust, Verifiable Decisions An Uncertainty-Typed Proposition IR, Assumption-Robust Pareto Frontiers, and a Regret Certificate
YUKTI：从自然语言情境到稳健、可验证的决策——一种不确定性类型的命题IR、假设稳健的帕累托前沿和遗憾证书
Abstract
Language models turn a worded situation into a numeric plan, and the dominant pipelines (NL4Opt, OptiMUS, ORLM, OR-LLM-Agent) commit to a single objective and point-valued coefficients, then solve once. For decisions that allocate real budget, effort, or clinical attention, that confidence is the failure mode: every objectified number is an assumption, and a plan optimal only if the guesses are exactly right is fragile -- mimicry of computation. YUKTI changes the target of autoformulation. Its representation is a typed-proposition graph whose relationships carry shape priors, coefficient uncertainty, and provenance. YUKTI routes each stage to an exact, nonlinear, or evolutionary solver; couples stages by a distributional Pareto hand-off; and introduces Assumption-Robust Pareto Frontiers (ARPF), resampling assumptions (including structural epsilon-contamination) to score how often each action survives (rho). We prove a bound making rho an exact factor of decision regret, add auditable traceability, and synthesize a benchmark-faithful data foundation when none exists (SRJANA). We validate three ways: under controlled misspecification the robust compromise cuts mean and tail regret by over 90% versus a naive point plan; on a regulated commercial decision we optimize inside a lawful action space and price the downside in euros; and on a real public dataset of 41,188 decisions an out-of-sample backtest beats the logged status quo by 34% and a naive point rule by 4% while reducing the optimizer's curse. The solvers are standard; we claim no benchmark-SOTA win. A head-to-head shows an LLM given the correct numbers, and single-objective optimization, both incur about 47x the held-out regret of YUKTI -- an LLM is a formulator, not a solver. Under long-range causal coupling, the forward hand-off becomes unsound, locating where it must become a backward-induction causal policy.
Chinese Translation
语言模型将文字情境转化为数值计划，而主流管道（NL4Opt、OptiMUS、ORLM、OR-LLM-Agent）则致力于单一目标和点值系数，然后进行一次求解。对于分配真实预算、努力或临床关注的决策，这种信心是失败模式：每一个物化的数字都是一个假设，只有在猜测完全正确时才最优的计划是脆弱的——是计算的模仿。YUKTI 改变了自动公式化的目标。其表示为一个类型命题图，其关系承载形状先验、系数不确定性和来源。YUKTI 将每个阶段路由到一个精确的、非线性的或进化的求解器；通过分布式帕累托交接连接各个阶段；并引入假设稳健的帕累托前沿（ARPF），重新采样假设（包括结构性ε污染）以评估每个行动存活的频率（rho）。我们证明了一个界限，使得rho成为决策遗憾的确切因素，增加可审计的可追溯性，并在没有现成数据的情况下合成一个基准忠实的数据基础（SRJANA）。我们通过三种方式进行验证：在受控的错误规范下，稳健的妥协将均值和尾部遗憾减少超过90%相较于天真的点计划；在一个受监管的商业决策中，我们在合法的行动空间内进行优化，并以欧元定价下行风险；在一个包含41,188个决策的真实公共数据集上，样本外回测超越了记录的现状34%，并比天真的点规则高出4%，同时减少了优化器的诅咒。求解器是标准的；我们不声称有任何基准SOTA胜利。面对面比较显示，给定正确数字的LLM和单目标优化都产生了约47倍于YUKTI的保留遗憾——LLM是一个公式化者，而不是求解者。在长期因果耦合下，前向交接变得不可靠，定位到必须成为后向推导因果政策的地方。
cs.AI / 7 / 2607.09708
GES-TSP: Graph Edge Sparsification for TSP
GES-TSP：旅行商问题的图边稀疏化
Abstract
Solving large-scale instances of the Traveling Salesman Problem (TSP) exactly is computationally expensive. Researchers often employ graph sparsification methods to improve computational efficiency. Traditional sparsification methods typically rely on fixed heuristics and fail to fully exploit instance-specific structural information. In this paper, we propose Graph Edge Sparsification (GES), a learning-based sparsification approach for Euclidean TSP. By incorporating geometric structural information and combinatorial optimization technology, our proposed method adaptively generates a sparsification graph for different instances, significantly reducing the graph size and accelerating the solving process. Experimental results demonstrate that our sparsification method can prune up to 95% of edges on the MATILDA dataset, while keeping the solution gap within 1% of the optimal value. Moreover, our approach exhibits strong generalization capability on the TSPLIB benchmark.In some large-scale instances, the pruning rate exceeds 99%, while the optimality gap remains below 1%.
Chinese Translation
精确求解大规模旅行商问题（TSP）实例的计算成本非常高。研究人员通常采用图稀疏化方法来提高计算效率。传统的稀疏化方法通常依赖于固定的启发式算法，未能充分利用特定实例的结构信息。在本文中，我们提出了一种基于学习的稀疏化方法——图边稀疏化（Graph Edge Sparsification, GES），用于欧几里得TSP。通过结合几何结构信息和组合优化技术，我们提出的方法能够自适应地为不同实例生成稀疏化图，显著减少图的大小并加速求解过程。实验结果表明，我们的稀疏化方法在MATILDA数据集上可以修剪多达95%的边，同时保持解的差距在最优值的1%以内。此外，我们的方法在TSPLIB基准测试中表现出强大的泛化能力。在一些大规模实例中，修剪率超过99%，而最优性差距保持在1%以下。
cs.AI / 8 / 2607.09709
The Verifier is the Curriculum: Execution-Gated Self-Distillation for Cross-Family Game Generation
验证者即课程：基于执行门控的自我蒸馏用于跨家族游戏生成
Abstract
Post-training a code generator against a learned judge can optimize proxy features that raise the score without improving the artifact. We study the opposite signal: a deterministic, judge-free, ungameable filter -- whether a generated project launches cleanly under a headless engine (strict-launch). Under this gate, rejection-sampling self-distillation compounds out-of-family generalization. On GameCraft-Bench (mapping a natural-language brief to a complete Godot project), a 14B model (Qwen3-14B+LoRA) distilled under strict-launch raises clean generation on four unseen game families from 8.8% to 42.2% per-candidate and best-of-K coverage from 18/25 to 25/25 (the gold ceiling) over three rounds, each a significant gain (p=0.0019, p<1e-4, p<1e-4). The gain is not from merely adding data: an exactly-matched gold-duplication control regresses below the base model (5.6% vs. 8.8%, p=0.019), while a count-matched decomposition splits the round-1-to-2 jump into comparable quality (+8.8pp) and quantity (+8.5pp) channels. Most directly, rerunning the loop with only the filter swapped -- the lenient BUILD check, which passes 99.9% of generations, in place of the launch gate -- erases the gain entirely (back to base, p=1e-3 vs. the launch-gated round), isolating verifier precision rather than the optimizer. A second ungameable signal, headless execution grounding, rises monotonically across rounds and yields far more grounded candidates than gold-duplication at a matched budget (16 vs. 5), confirming the gains are functional, not launch-but-empty. Game generation is a verifiable testbed for one lesson: the verifier is the curriculum -- what it certifies is what the model learns.
Chinese Translation
在训练后使用学习到的评判者对代码生成器进行优化，可以提升代理特征的得分，而不改善生成的工件。我们研究相反的信号：一个确定性的、无评判者的、不可游戏化的过滤器——生成的项目是否能够在无头引擎下顺利启动（严格启动）。在这个门控下，拒绝采样自我蒸馏增强了跨家族的泛化能力。在 GameCraft-Bench（将自然语言简报映射到完整的 Godot 项目）上，一个 14B 模型（Qwen3-14B+LoRA）在严格启动下进行蒸馏，使得四个未见游戏家族的干净生成率从 8.8% 提升至 42.2%（每个候选）以及最佳覆盖率从 18/25 提升至 25/25（黄金上限），经过三轮，每轮均有显著提升（p=0.0019, p<1e-4, p<1e-4）。这一提升并非仅仅通过增加数据实现：一个完全匹配的黄金重复控制回归低于基础模型（5.6% 对比 8.8%，p=0.019），而一个计数匹配的分解将第一轮到第二轮的跃升分解为可比的质量（+8.8pp）和数量（+8.5pp）通道。最直接的，重新运行循环，仅将过滤器替换为宽松的 BUILD 检查（通过 99.9% 的生成），而不是启动门，完全消除了增益（回到基础，p=1e-3 对比启动门控轮），孤立出验证者的精确性，而非优化器。第二个不可游戏化的信号，无头执行基础，随着轮次的增加单调上升，并在匹配预算下产生远多于黄金重复的基础候选（16 对 5），确认增益是功能性的，而非启动但为空的。游戏生成是一个可验证的试验场，传达了一个教训：验证者即课程——它所认证的就是模型所学习的内容。
cs.AI / 9 / 2607.09713
Closed-Loop Control with Rule-Aligned Small Language Models and Multi-Agent Self-Correction
基于规则对齐的小型语言模型与多智能体自我修正的闭环控制
Abstract
A key step toward autonomous industrial operation is the ability to create and reconfigure control policies from natural-language requirement specifications, with minimal or no manual redesign. In this setting, policy generation by AI agents can be a credible path when paired with a plant-aware validator (e.g., a digital twin) that can check generated candidate actions before execution. However, practical deployment is constrained by inference latency and compute footprint: large cloud-based models are often too slow, opaque, or data-sensitive for edge closed-loop use. This work investigates whether a compact Small Language Model (SLM) can be retrained for control reasoning and embedded in a validator-guided correction loop. We use a Qwen2.5-1.5B model aligned via Group Relative Policy Optimization (GRPO), combined with (i) an action agent, (ii) a symbolic/digital-twin-style validation layer, and (iii) a reprompting agent that iteratively steers outputs toward valid actions. In randomized thermal-control simulations (30 experiments with 500 steps each), the framework achieves 91.5% average action-alignment accuracy (86.3%--100% across cases) at 3.84\,s mean inference latency. Under symbolic re-mapping, it maintains a 95% in-range rate, indicating robust physical regulation despite reduced token-level agreement. These results support SLM+validator architectures as a practical path toward reconfigurable autonomous control at the edge.
Chinese Translation
实现自主工业操作的关键步骤是能够从自然语言需求规范中创建和重新配置控制策略，尽量减少或不进行手动重新设计。在这种情况下，结合植物感知验证器（例如，数字双胞胎）的AI代理进行政策生成，可以成为一种可信的路径，该验证器能够在执行之前检查生成的候选动作。然而，实际部署受到推理延迟和计算占用的限制：大型云端模型通常过于缓慢、不透明或对数据敏感，不适合边缘闭环使用。本研究探讨了紧凑型小型语言模型（SLM）是否可以重新训练以进行控制推理，并嵌入到验证器引导的修正循环中。我们使用通过群体相对政策优化（GRPO）对齐的Qwen2.5-1.5B模型，结合（i）一个动作代理，（ii）一个符号/数字双胞胎风格的验证层，以及（iii）一个反提示代理，迭代地将输出引导至有效动作。在随机热控制模拟中（30个实验，每个实验500步），该框架实现了91.5%的平均动作对齐准确率（在各案例中为86.3%--100%），平均推理延迟为3.84秒。在符号重新映射下，它保持了95%的范围内率，表明尽管令牌级一致性降低，但仍能实现稳健的物理调节。这些结果支持SLM+验证器架构作为实现边缘可重构自主控制的实用路径。
cs.AI / 10 / 2607.09714
Feedback-Coupled Memory Systems in Continuous Time
连续时间中的反馈耦合记忆系统
Abstract
The Feedback-Coupled Memory Systems (FCMS) architecture formalizes closed-loop coordination through four abstract operators, two of which - the agent update operator $f_i$ and the environmental update operator $\Psi$ - are left axiomatically undefined in the original framework. To address this, $f_i$ is defined by Mechanism-Based Intelligence (MBI), where agents update locally through a decentralized price mechanism and economic principles, and $\Psi$ is defined by the Coupled Memory Graph Process (CMGP), a non-Markovian framework where the environment is treated as a physical substrate that records and responds to trajectory history coherently without external forcing. The resulting continuous-time FCMS instantiation achieves Lyapunov global dissipativity governed by the computable threshold $4\beta^2 < 2\eta\mu\gamma^2$. This generalizes both the discrete FCMS stability condition $4\eta\beta^2 < \gamma$ and CMGP's physical bifurcation threshold $\alpha_c = 1/K$, confirming that memory dissipation must outpace feedback gain as a universal organizing principle. Numerical simulation with $N=2$ agents and mean-field validation at $N=10^6$ confirm the stability threshold and the self-reinforcing coordination cascade that emerges when it is violated.
Chinese Translation
反馈耦合记忆系统（Feedback-Coupled Memory Systems, FCMS）架构通过四个抽象算子形式化闭环协调，其中两个算子——代理更新算子 $f_i$ 和环境更新算子 $ ext{Ψ}$——在原始框架中被公理化地未定义。为了解决这个问题，$f_i$ 由基于机制的智能（Mechanism-Based Intelligence, MBI）定义，其中代理通过去中心化的价格机制和经济原则进行局部更新，而 $ ext{Ψ}$ 则由耦合记忆图过程（Coupled Memory Graph Process, CMGP）定义，这是一个非马尔可夫框架，其中环境被视为一种物理基质，能够连贯地记录并响应轨迹历史，而不需要外部强迫。得到的连续时间 FCMS 实例实现了由可计算阈值 $4eta^2 < 2 hetaeta^2$ 支配的 Lyapunov 全局耗散性。这一结果推广了离散 FCMS 稳定性条件 $4 hetaeta^2 < eta$ 和 CMGP 的物理分岔阈值 $ ext{α}_c = 1/K$，确认了记忆耗散必须超过反馈增益作为一种普遍的组织原则。对 $N=2$ 个代理的数值模拟和对 $N=10^6$ 的平均场验证确认了稳定性阈值以及当其被违反时出现的自我强化协调级联。
cs.AI / 11 / 2607.09729
AGM-like Paraconsistent Partial Meet Abductive Expansion Operation
类似AGM的非一致部分满足归纳扩展操作
Abstract
In his 1996 doctoral thesis, Maurice Pagnucco created the first AGM-like abductive expansion operation. Taking his operation as a basis, as well as a taxonomy -- inspired by Atocha Aliseda -- responsible for highlighting and formalizing the main components of abductive reasoning, the main aim of this paper is to present a new paraconsistent AGM-like abductive expansion operation -- capable of assimilating contradictory explanatory hypotheses without trivialization and the consequent absurd epistemic state -- with its postulates and its transitively relational partial meet construction. To a large extent, the formal development presented in this paper was only made possible by the recent creation of the paraconsistent logic RCbr, an LFI (Logics of Formal Inconsistencies) that establishes properties especially relevant to belief revision contexts, in particular, the ability to be self-extensional -- i.e., to satisfy the replacement property. This is the first of two papers: the paraconsistent abductive expansion operation announced here -- which is part of a new system called AGMpabd -- despite bringing many interesting features, does not assign any relevant epistemic role to the paraconsistent operators of negation and consistency. Only in a second paper will an analogous paraconsistent abductive expansion operation -- which is part of another new system, AGMcircabd -- be enhanced in this direction. Nevertheless, to the best of my knowledge, the operation developed in this paper is the first of its kind in the AGM literature.
Chinese Translation
在1996年的博士论文中，莫里斯·帕纽科（Maurice Pagnucco）创造了第一个类似AGM的归纳扩展操作。以他的操作为基础，以及一个受到阿托查·阿利塞达（Atocha Aliseda）启发的分类法，该分类法负责突出和形式化归纳推理的主要组成部分，本文的主要目的是提出一种新的非一致AGM-like归纳扩展操作——能够在不简化和随之而来的荒谬认识状态的情况下吸收矛盾的解释假设——并给出其公设及其传递关系的部分满足构造。在很大程度上，本文所呈现的形式发展仅得益于最近创建的非一致逻辑RCbr，这是一种形式不一致逻辑（LFI），它确立了在信念修正上下文中特别相关的属性，尤其是自扩展能力，即满足替换属性。这是两篇论文中的第一篇：这里宣布的非一致归纳扩展操作——它是一个新系统AGMpabd的一部分——尽管带来了许多有趣的特征，但并未赋予非一致否定和一致性运算符任何相关的认识角色。只有在第二篇论文中，类似的非一致归纳扩展操作——它是另一个新系统AGMcircabd的一部分——才会在这个方向上得到增强。然而，据我所知，本文所开发的操作是AGM文献中首个此类操作。
cs.AI / 12 / 2607.09739
Coresets Before Score Sets: Evaluation-Unsupervised Prompt Subset Selection for LLM Benchmarks
核心集优于评分集：针对大型语言模型基准的评估无监督提示子集选择
Abstract
We study LLM benchmark coreset selection: selecting a small subset of prompts over multiple benchmarks whose induced model scores and rankings approximate those obtained from the full benchmark suite. In evaluation-unsupervised benchmark coreset selection (our approach), the selection algorithm uses no model evaluation outcomes, and operates on a fine granularity by producing subsets of prompts over multiple benchmarks rather than producing a sub-collection of entire benchmarks. We use submodular subset selection, and we develop and evaluate many different submodular functions for this purpose, including determinantal point process (DPP) based approaches, submodular mutual information functions, and facility location-based functions. On a new large-scale suite of 35 heterogeneous benchmarks spanning five different capability categories, 18 frontier LLMs, and over 61K prompts, we find that the facility location (FL) function operating exclusively on inexpensive semantic prompt embeddings preserves LLM scores better than twelve separate score-based and diversity-based baselines, across a range of coreset budgets. Moreover, we show our proposed objective is not limited to the evaluation-unsupervised regime: in the setting where only a handful of whole benchmarks must be selected and a large amount of model scores are available, the same objective matches or outperforms state-of-the-art baselines on the MMLU and MTEB leaderboards, while being substantially cheaper to compute. Together, our results suggest that submodularity, in general, is a strong and reliable tool for benchmark compression.
Chinese Translation
我们研究大型语言模型（LLM）基准的核心集选择：从多个基准中选择一个小的提示子集，使其诱导的模型评分和排名近似于从完整基准套件中获得的结果。在评估无监督的基准核心集选择（我们的方法）中，选择算法不使用任何模型评估结果，并且通过在多个基准上生成提示子集而不是生成整个基准的子集合来进行细粒度操作。我们使用子模选择，并为此目的开发和评估多种不同的子模函数，包括基于行列式点过程（DPP）的方法、子模互信息函数以及基于设施位置的函数。在一个新的大规模套件中，该套件涵盖了五个不同能力类别的35个异构基准、18个前沿LLM和超过61K个提示，我们发现仅基于廉价语义提示嵌入的设施位置（FL）函数在多个核心集预算范围内比十二个独立的基于评分和多样性的基线更好地保留了LLM评分。此外，我们展示了我们提出的目标不仅限于评估无监督的情境：在仅需选择少量完整基准且可用大量模型评分的情况下，相同的目标在MMLU和MTEB排行榜上与最先进的基线相匹配或超越，同时计算成本显著更低。总的来说，我们的结果表明，子模性通常是基准压缩的强大而可靠的工具。
cs.AI / 13 / 2607.09740
A Dynamic Scene Interaction Reasoning Framework for Scene-level Lane-Change Intention and Trajectory Prediction of Multiple Interacting Vehicles
一种动态场景交互推理框架用于多交互车辆的场景级变道意图和轨迹预测
Abstract
Safe motion planning in advanced driver-assistance systems and autonomous vehicles requires an accurate understanding of how the surrounding traffic scene is likely to evolve. However, many existing lane-change prediction methods remain centered on a single target vehicle, while multi-agent forecasting approaches often describe scene evolution only through future positions and provide limited explicit information about the maneuver associated with each vehicle. This study proposes a dynamic scene graph attention framework that predicts the lane-change intention and future trajectory of every relevant vehicle within a local traffic scene. The scene is represented as a time-varying interaction graph in which vehicles are modeled as nodes and their spatial and kinematic relationships are encoded through explicit edge features. Temporal graph-attention message passing captures evolving inter-vehicle dependencies and pre-maneuver cues, while an intention-guided decoder links each predicted maneuver to its corresponding future motion. A scene-level consistency objective further encourages compatible multi-vehicle futures. Experiments on the NGSIM I-80, NGSIM US-101, and highD datasets demonstrate consistent improvements over competing baselines. DSiGAT achieves intention prediction accuracies of 90.12% and 90.97% on NGSIM I-80 and US-101, respectively, and reduces trajectory RMSE by up to 52.94% relative to the strongest baseline. It also produces lower inter-agent collision rates and joint displacement errors, indicating more coherent scene-level predictions. Ablation, sensitivity, robustness, and qualitative analyses further validate the contribution of the proposed components and the effectiveness of the scene-focused formulation.
Chinese Translation
在先进驾驶辅助系统和自动驾驶车辆中，安全的运动规划需要准确理解周围交通场景可能的发展。然而，许多现有的变道预测方法仍然集中于单一目标车辆，而多智能体预测方法通常仅通过未来位置描述场景演变，并提供有限的关于每辆车相关机动的显式信息。本研究提出了一种动态场景图注意力框架，该框架预测局部交通场景中每辆相关车辆的变道意图和未来轨迹。场景被表示为一个时变交互图，其中车辆被建模为节点，它们的空间和运动关系通过显式边特征进行编码。时间图注意力消息传递捕捉不断演变的车辆间依赖关系和预机动线索，而意图引导解码器将每个预测的机动与其对应的未来运动联系起来。场景级一致性目标进一步鼓励兼容的多车辆未来。对NGSIM I-80、NGSIM US-101和highD数据集的实验表明，相较于竞争基线，取得了一致的改进。DSiGAT在NGSIM I-80和US-101上的意图预测准确率分别达到90.12%和90.97%，并将轨迹均方根误差（RMSE）相较于最强基线降低了最多52.94%。它还产生了更低的智能体间碰撞率和联合位移误差，表明场景级预测更加一致。消融、敏感性、鲁棒性和定性分析进一步验证了所提组件的贡献及场景聚焦公式的有效性。
cs.AI / 14 / 2607.09743
Scaffolding the Strategist: Architecture-Dependent Reasoning Interventions in Hotelling Spatial Markets
策略家的支架：依赖架构的推理干预在霍特林空间市场中的应用
Abstract
We investigate whether structured reasoning interventions improve the strategic economic reasoning of large language models, and whether their effects depend on model architecture. Using Hotelling's linear city model as a diagnostic vehicle, we evaluate GPT-4.1-mini (a standard instruction-following model) and GPT-5-mini (a reasoning-optimized model) under five conditions - an unscaffolded baseline and four reasoning interventions - across eight questions spanning deductive and abductive reasoning, three prompt framings, and three repetitions per condition, yielding 720 individually judged responses. We find a statistically significant crossover interaction between scaffolding type and model architecture ($t(7) = 4.79$, $p = 0.002$, $d = 1.69$): commitment scaffolding improves the standard model ($+0.21$) while degrading the reasoning model ($-0.63$), and principled separation shows the opposite pattern ($-0.40$ vs. $+0.31$). Both crossovers are individually significant (commitment: $p = 0.040$; separation: $p = 0.002$) and hold across all eight questions with 7/8 directional consistency. Adversarial stress-testing harms both models, with $2.6\times$ greater degradation for the reasoning model ($-1.47$ vs. $-0.57$; $p = 0.038$), and the damage correlates negatively with baseline difficulty ($R^2 = 0.36$, $p = 0.014$). We further document a persistent declarative-procedural gap in which both models identify correct strategies at rates far exceeding their ability to execute them; separation fully closes this gap for the reasoning model while no intervention helps the standard model.
Chinese Translation
我们研究了结构化推理干预是否能改善大型语言模型的战略经济推理，以及其效果是否依赖于模型架构。我们使用霍特林线性城市模型作为诊断工具，评估了GPT-4.1-mini（一个标准的指令遵循模型）和GPT-5-mini（一个优化推理的模型）在五种条件下的表现——一个未支架的基线和四种推理干预——在涵盖演绎推理和溯因推理的八个问题中进行评估，三种提示框架和每种条件下三次重复，最终获得720个独立评判的响应。我们发现支架类型与模型架构之间存在统计显著的交叉互动（$t(7) = 4.79$, $p = 0.002$, $d = 1.69$）：承诺支架改善了标准模型（$+0.21$），而降低了推理模型的表现（$-0.63$），原则性分离则显示出相反的模式（$-0.40$ 对比 $+0.31$）。这两个交叉效应在统计上都是显著的（承诺：$p = 0.040$；分离：$p = 0.002$），并在所有八个问题中保持了7/8的方向一致性。对抗性压力测试对两个模型都有害，推理模型的退化程度是标准模型的$2.6 imes$（$-1.47$ 对比 $-0.57$；$p = 0.038$），且损害与基线难度呈负相关（$R^2 = 0.36$, $p = 0.014$）。我们进一步记录了一个持续的陈述-程序性差距，即两个模型识别正确策略的速度远远超过其执行这些策略的能力；分离完全消除了推理模型的这一差距，而没有任何干预能帮助标准模型。
cs.AI / 15 / 2607.09744
A Theory of Least Autonomy in AI
人工智能中的最小自主权理论
Abstract
Least privilege, the principle that an identity should hold only the permissions strictly required for its task, has been a foundational primitive of access control for decades. We argue that this principle is insufficient for agentic AI systems, which do not merely hold permissions but can combine, approve, and amplify them across workflows and system boundaries. We propose least autonomy as an appropriate generalization and develop a formal theory. First, we define a compositional blast radius d(a,b) that measures structural separation between actions in an enterprise hierarchy, combining an ultrametric tree with lattice-valued confidentiality, integrity, and control-context labels. Second, we define a directed agent influence graph G(theta). An arc from U to V requires a directed shared-resource write-to-read meeting or a conservative undirected agent-to-agent (A2A) communication meeting, and a meeting-conditioned influence potential at or above an externally selected policy threshold theta. A catalogue-radius profile supports calibration and audit of theta. Finally, we define a collusion predicate over graph reachability that detects authorization composition, decision manipulation, and cross-domain capability composition.
Chinese Translation
最小权限原则，即一个身份应仅持有其任务所需的严格权限，几十年来一直是访问控制的基础原理。我们认为这一原则对于代理型人工智能系统是不够的，因为这些系统不仅仅持有权限，还能够在工作流程和系统边界之间组合、批准和放大这些权限。我们提出最小自主权作为一个合适的推广，并发展出一个正式理论。首先，我们定义了一个组合爆炸半径 d(a,b)，它测量企业层级中动作之间的结构性分离，结合了超度量树与格值机密性、完整性和控制上下文标签。其次，我们定义了一个有向代理影响图 G(theta)。从 U 到 V 的弧需要一个有向共享资源的写入-读取会议或一个保守的无向代理间（A2A）通信会议，并且在或超过外部选定的政策阈值 theta 的会议条件影响潜力。目录半径配置文件支持对 theta 的校准和审计。最后，我们定义了一个关于图可达性的共谋谓词，用于检测授权组合、决策操控和跨域能力组合。
cs.AI / 16 / 2607.09745
SupplyNetPy: An Open-Source Python Library for High-Fidelity Modeling and Simulation of Arbitrary Supply Chain and Inventory Networks
SupplyNetPy：一个用于高保真建模和仿真任意供应链和库存网络的开源Python库
Abstract
This paper introduces SupplyNetPy, an open-source, well-documented Python library for modeling and discrete-event simulation of supply chain networks with arbitrary multi-echelon structures. It supports multiple replenishment policies, perishable inventory, node disruptions, and stochastic demand and lead times. All components are extensible via inheritance. Users describe a supply chain as a graph with node and link attributes, while the library handles simulation, providing logs and extensive node and network level performance reports. This paper presents the motivation, design, key features, and architecture of SupplyNetPy, along with detailed validation results (against analytical benchmarks, a commercial tool, and a published case study). A key motivation behind SupplyNetPy's development is programmatic generation and simulation of complex models, enabling design-space exploration, what-if analysis, training data generation, and supply chain digital twins.
Chinese Translation
本文介绍了SupplyNetPy，一个开源、文档完善的Python库，用于建模和离散事件仿真具有任意多层级结构的供应链网络。它支持多种补货策略、易腐烂库存、节点中断以及随机需求和交货时间。所有组件均可通过继承进行扩展。用户可以将供应链描述为具有节点和链接属性的图形，而库则负责仿真，提供日志和广泛的节点及网络级性能报告。本文展示了SupplyNetPy的动机、设计、关键特性和架构，并提供了详细的验证结果（与分析基准、商业工具和已发布案例研究的对比）。SupplyNetPy开发的一个关键动机是程序化生成和仿真复杂模型，从而实现设计空间探索、假设分析、训练数据生成和供应链数字双胞胎。
cs.AI / 17 / 2607.09748
Replicating Belief, Not Bits: Epistemic State Replication for Agentic Systems
复制信念，而非比特：面向自主系统的认知状态复制
Abstract
In distributed systems, the classical State Machine Replication (SMR) model assumes that correct replicas execute deterministic transitions to yield identical bitwise states. However, the rise of agentic distributed systems -- where autonomous, stochastic, and model-driven agents orchestrate infrastructure -- presents scenarios where deterministic, bitwise replication is insufficient. Replicas operating with generative models may exhibit divergent reasoning paths, summaries, and token boundaries, yet reach semantically equivalent and correct operational decisions. Forcing bitwise agreement across these stochastic participants degrades execution flexibility, induces context amnesia, and limits performance. We argue that in such settings replicas should agree on belief, not bits. We propose Epistemic State Replication (ESR), a belief-replication layer for agentic distributed systems that shifts the replication boundary from data visibility to knowledge visibility. We formalize the epistemic node state as a pair K = (L, B) separating the deterministic, immutable evidence log (L) from the stochastic, evolving belief lineage (B). To govern execution safety, we define Semantic Linearizability, which requires operations to reflect the latest committed operational meaning within a verifier-bounded semantic compatibility metric, and Bounded Eventual Coherence, which bounds expected semantic divergence under fair delivery, monotonic evidence, bounded verifier disturbance, and a contractive graft operator. We outline protocols for propagating derived insights using structured epistemic deltas, and formalize Verifiable Semantic Rollbacks to prune faulty premises from belief lineages without inducing context amnesia. We prototype ESR and report preliminary simulation results that show feasibility under the stated assumptions and illustrate reductions in secondary cognitive faults.
Chinese Translation
在分布式系统中，经典的状态机复制（State Machine Replication, SMR）模型假设正确的副本执行确定性转换以产生相同的比特状态。然而，随着自主分布式系统的兴起——在这些系统中，自主的、随机的和模型驱动的代理协调基础设施——出现了确定性比特复制不足以应对的场景。使用生成模型的副本可能表现出不同的推理路径、摘要和标记边界，但仍能达成语义上等价且正确的操作决策。在这些随机参与者之间强制比特一致性会降低执行灵活性，导致上下文遗忘，并限制性能。我们认为，在这种环境中，副本应当就信念达成一致，而非比特。我们提出了认知状态复制（Epistemic State Replication, ESR），这是一个面向自主分布式系统的信念复制层，将复制边界从数据可见性转移到知识可见性。我们将认知节点状态形式化为一对 K = (L, B)，将确定性、不变的证据日志（L）与随机、不断演变的信念谱系（B）分开。为了确保执行安全性，我们定义了语义线性化（Semantic Linearizability），要求操作反映在验证者界限内的最新已提交操作含义，并定义了有界最终一致性（Bounded Eventual Coherence），该一致性限制了在公平交付、单调证据、有界验证者干扰和收缩性嫁接操作下的预期语义偏差。我们概述了使用结构化认知增量传播派生见解的协议，并形式化了可验证的语义回滚（Verifiable Semantic Rollbacks），以在不导致上下文遗忘的情况下修剪信念谱系中的错误前提。我们原型化了 ESR，并报告了初步的仿真结果，显示在所述假设下的可行性，并说明了次要认知错误的减少。
cs.AI / 18 / 2607.09751
Task-Conditioned Synthetic Data Generation for Improving Machine Learning Performance in Agricultural Prediction Tasks
基于任务条件的合成数据生成以提升农业预测任务中的机器学习性能
Abstract
Machine Learning (ML) algorithms have been widely used to estimate agricultural variables across diverse contexts. However, because the quantity and quality of training data strongly influence performance of ML algorithms, their use can be constrained by limited or incomplete reference data. Synthetic Data Generation (SDG) offers a practical approach to address this issue by producing artificial but realistic samples that preserve key characteristics of the original data. Building on teacher-student knowledge transfer and in-context learning for tabular data, this study proposes a Task-Conditioned SDG (TCSDG) algorithm that pairs a Bayesian Network generator with a transformer-based tabular foundation model (TabICL). The proposed algorithm was evaluated on two agricultural prediction tasks: crop yield prediction and crop type classification. Six benchmark SDG algorithms were also utilized to compare their performance with that of TCSDG. Across twelve study sites, two training-data fractions, four multiplication ratios, and three predictive ML algorithms, augmenting the original data with TCSDG-generated synthetic data improved ML performance in 89% of the crop type classification experiments and 74% of the crop yield prediction experiments. TCSDG also substantially outperformed benchmark SDG algorithms and was the only method to consistently improve ML performance across both tasks at the aggregate level. The study demonstrates that carefully designed and processed synthetic data can improve ML performance in precision-agriculture applications. TCSDG offers a practical and extensible framework for generating synthetic data that supports downstream ML agricultural prediction. The full implementation of TCSDG is publicly available as open source at https://github.com/HamidEbrahimy/TCSDG.
Chinese Translation
机器学习（ML）算法已广泛应用于估计不同背景下的农业变量。然而，由于训练数据的数量和质量对机器学习算法的性能有着重要影响，因此其使用可能受到有限或不完整参考数据的限制。合成数据生成（SDG）提供了一种实用的方法，通过生成保留原始数据关键特征的人工但真实的样本来解决这一问题。本研究基于教师-学生知识转移和表格数据的上下文学习，提出了一种任务条件的合成数据生成（TCSDG）算法，该算法将贝叶斯网络生成器与基于变换器的表格基础模型（TabICL）相结合。所提算法在两个农业预测任务上进行了评估：作物产量预测和作物类型分类。还利用六个基准SDG算法与TCSDG的性能进行了比较。在十二个研究地点、两个训练数据比例、四个乘法比率和三个预测机器学习算法的实验中，使用TCSDG生成的合成数据增强原始数据在89%的作物类型分类实验和74%的作物产量预测实验中提高了机器学习性能。TCSDG的表现也显著优于基准SDG算法，并且是唯一一种在整体层面上在两个任务中始终提高机器学习性能的方法。研究表明，经过精心设计和处理的合成数据可以提升精密农业应用中的机器学习性能。TCSDG提供了一个实用且可扩展的合成数据生成框架，支持下游机器学习农业预测。TCSDG的完整实现已作为开源项目公开，网址为 https://github.com/HamidEbrahimy/TCSDG。
cs.AI / 19 / 2607.09755
LegalFarePlan: A Label-Setting Framework for Fare-Transparent Urban Rail Route Planning under Non-Additive Fare Rules
LegalFarePlan：一种基于标签设置的 fare 透明城市轨道路线规划框架，适用于非加法票价规则
Abstract
Urban rail fare systems may be non-additive: the fare of a single paid journey from an origin to a destination can differ from the sum of fares over multiple legally separated journey legs. This paper presents LegalFarePlan, a fare-transparent route-planning framework that models legal exit-and-reentry operations as explicit, auditable constraints. Given a transit network, fare function, transfer rules, station-level exit/re-entry costs, an extra-time budget, and a split limit, the planner computes explainable route plans over paid journey segments. The artifact implements Dijkstra shortest-time and direct route-planner baselines, a greedy split heuristic, bounded exact label-setting, and Pareto-frontier search. Evaluation uses controlled synthetic data and a 57-station semi-synthetic benchmark with 360 OD pairs. On the semi-synthetic benchmark, bounded exact search identifies positive modeled fare reductions for 71.11% of OD pairs, with mean reduction 3.78 and maximum reduction 9.0 synthetic fare units under a 45-minute extra-time budget. These results demonstrate method behavior and reproducibility; they are not empirical conclusions about MTR or any transit operator.
Chinese Translation
城市轨道票价系统可能是非加法的：从起点到终点的单次付费旅程的票价可能与多个合法分隔旅程段的票价总和不同。本文提出了 LegalFarePlan，一种 fare 透明的路线规划框架，该框架将合法的出入操作建模为明确的、可审计的约束条件。给定一个交通网络、票价函数、换乘规则、车站级出入成本、额外时间预算和分割限制，规划器计算出可解释的付费旅程段的路线计划。该工具实现了 Dijkstra 最短时间和直接路线规划基线、贪婪分割启发式、有限精确标签设置和帕累托前沿搜索。评估使用了受控的合成数据和一个包含 57 个车站的半合成基准，涵盖 360 对出发-到达（OD）对。在半合成基准上，有限精确搜索为 71.11% 的 OD 对识别出正向建模的票价降低，平均降低 3.78，最大降低 9.0 个合成票价单位，且在 45 分钟的额外时间预算内。这些结果展示了方法的行为和可重复性；它们并不是关于地铁（MTR）或任何交通运营商的实证结论。
cs.AI / 20 / 2607.09762
BatteryLake: Agentic, Physics-Grounded Curation of Heterogeneous Battery Aging Data and Benchmarking
BatteryLake：基于物理的异构电池老化数据的智能化管理与基准测试
Abstract
Public battery aging datasets are a critical asset for advanced health management, but their practical use is often limited by inconsistent formats, unclear schemas, and metadata scattered across repositories and publications. Current curation remains largely manual and hard to reproduce, while general-purpose data integration tools miss the domain-specific semantics of electrochemical time-series data. We present BatteryLake, a governed data lakehouse that turns raw public battery data into benchmark-ready assets through an agentic, physics-grounded curation framework, with three contributions. First, LLM agents extract metadata and synthesize dataset-specific converters, grounding every output in verbatim evidence and abstaining when none supports a value. Second, a human-in-the-loop mechanism frames verification as selective prediction and gates admitted data through 26 schema, statistical, and physical-plausibility rules. Third, we release an open benchmark of 41 datasets from over 25 institutions, with standardized SOH and RUL tasks, three split protocols, and eight baseline model families. The platform, benchmark, and curation protocol are publicly available at https://tianwen1209.github.io/batterylake/.
Chinese Translation
公共电池老化数据集是先进健康管理的重要资产，但其实际使用常常受到格式不一致、模式不明确以及元数据分散在各个库和出版物中的限制。目前的数据管理仍然主要依赖人工，难以复现，而通用数据集成工具则无法捕捉电化学时间序列数据的领域特定语义。我们提出了BatteryLake，一个受管控的数据湖仓，通过一个智能化、基于物理的管理框架，将原始公共电池数据转化为可用于基准测试的资产，具有三项贡献。首先，LLM（大语言模型）代理提取元数据并合成特定数据集的转换器，确保每个输出都有逐字证据支持，并在没有支持值时避免输出。其次，循环中的人类机制将验证框架设定为选择性预测，并通过26条模式、统计和物理合理性规则对接纳的数据进行筛选。第三，我们发布了来自25个以上机构的41个数据集的开放基准，包含标准化的SOH（健康状态）和RUL（剩余使用寿命）任务、三种划分协议和八个基线模型系列。该平台、基准和管理协议已在https://tianwen1209.github.io/batterylake/上公开提供。
cs.AI / 21 / 2607.09765
How Much Does Correctness Cost? Budgeted Placement of Strong Correctors in a Weak Multi-Agent Swarm
正确性成本是多少？在弱多智能体群中预算强校正器的放置
Abstract
A cheap swarm of unreliable agents can be steered to a correct consensus by a few strong, expensive "oracle" correctors. We ask how much one must spend, and where to place the oracles. We model the swarm as a consensus on a graph in which each oracle pins one node toward the truth at a cost-coupled, concave strength, and measure quality by the coherence H(R)=tr M(R)^{-1}. Our first result is that H stays submodular (each added oracle helps less than the last) even when the oracles differ in strength, so a cost-benefit greedy comes within 1-1/e of the best placement at any budget. Inverting the budget gives the budget-correctness frontier B*(eps), the least spend that guarantees an eps-correct consensus: closed-form on the complete graph, and a minimal oracle count k* when oracles cost the same. Whether a budget then buys a few strong oracles or many medium onese curvature of the cost-quality law: diminishing returns favour spreadsharply increasion. Measured onthe Qwen3 ladder (0.6-32B), the law is concave for math verificatio convex foremergent code tracing, so the verdict is genuinely task-dependent.https://github.com/YehudaItkin/budgeted-oracle-placemen
Chinese Translation
一群廉价且不可靠的智能体可以通过少数强大且昂贵的“oracle”校正器引导至正确的共识。我们探讨需要花费多少以及如何放置这些oracle。我们将智能体群建模为一个图上的共识，其中每个oracle以成本相关的凹形强度将一个节点固定在真实值上，并通过一致性 H(R)=tr M(R)^{-1} 来衡量质量。我们的第一个结果是，即使oracle在强度上存在差异，H 仍然保持次模性（每增加一个oracle的帮助程度低于上一个），因此成本效益贪婪算法在任何预算下都能在最佳放置的 1-1/e 范围内。反转预算得到预算-正确性边界 B*(eps)，即保证 eps-正确共识的最低支出：在完全图上为封闭形式，当oracle成本相同时为最小oracle数量 k*。预算是购买少数强大oracle还是许多中等强度的oracle，取决于成本-质量法则的曲率：递减收益倾向于快速增加。在 Qwen3 梯度（0.6-32B）上测量，该法则在数学验证中是凹的，而在紧急代码追踪中是凸的，因此结论确实依赖于具体任务。
cs.AI / 22 / 2607.09766
Norm Enforcement for AI Agents: Robustly Shaping Behavior in Multi-Agent Systems
AI代理的规范执行：在多智能体系统中稳健地塑造行为
Abstract
AI agents are increasingly deployed in shared environments where they pursue diverse goals and compete for rewards. This multi-agent competition can lead to behaviors that serve individual gains at collective cost -- for instance, marketing agents may post misleading content as a result of competing for engagement on social media. Human societies address such problems through norms that constrain acceptable behavior, supported by enforcement mechanisms that detect and penalize violations. Motivated by this, we study norm enforcement mechanisms for language model agents. We find that simple enforcement mechanisms are exploited by misaligned agents for competitive advantage, even when they are not explicitly trained or prompted to do so. We thus turn our attention to designing more robust mechanisms, and identify two key ingredients: estimating each agent's reliability over time, and updating this estimate with escalating penalties for repeated misbehavior. Across three simulated environments and a variety of agent populations, mechanisms built on these principles resist exploitation, while still penalizing norm violations at comparable or lower cost than baselines. Our results position norm enforcement mechanisms as scalable levers for shaping agents' behavior, but only when designed to anticipate becoming part of the system they govern. Our code and data are available at https://yaowenye.com/norm-enforcement.
Chinese Translation
AI代理越来越多地被部署在共享环境中，在这些环境中它们追求多样化的目标并竞争奖励。这种多智能体竞争可能导致行为服务于个人利益而损害集体利益——例如，营销代理可能会发布误导性内容，以便在社交媒体上竞争获得关注。人类社会通过规范来解决此类问题，这些规范限制了可接受的行为，并通过检测和惩罚违规行为的执行机制来支持。基于此动机，我们研究了语言模型代理的规范执行机制。我们发现，即使未明确训练或提示，简单的执行机制也会被不对齐的代理利用以获取竞争优势。因此，我们将注意力转向设计更稳健的机制，并确定了两个关键要素：评估每个代理随时间的可靠性，并通过对重复不当行为施加逐步加重的惩罚来更新这一评估。在三个模拟环境和多种代理群体中，基于这些原则构建的机制抵御了利用，同时在惩罚规范违规方面的成本与基线相比相当或更低。我们的结果将规范执行机制定位为塑造代理行为的可扩展杠杆，但前提是这些机制设计时要考虑到它们将成为所治理系统的一部分。我们的代码和数据可在 https://yaowenye.com/norm-enforcement 获取。
cs.AI / 23 / 2607.09770
Verification of Adaptive Agentic Controllers through Finite Rule Revision
通过有限规则修订验证自适应代理控制器
Abstract
Industrial agentic AI systems increasingly exhibit a gap between prototype capability and production deployment. In particular, adaptive agents may generate plausible outputs while remaining difficult to verify under non-determinism, confidentiality constraints, limited context, and weak observability. This paper formulates a bounded verification protocol for adaptive agentic controllers represented by finite symbolic rules, explicit diagnostic predicates, explanation logs, and held-out re-evaluation. The central research question is: when an adaptive agentic controller is represented through finite rules, explicit diagnostic predicates, explanation logs, and held-out re-evaluation, which classes of controller failure can be detected, locally repaired, or rejected without relying on unrestricted human-in-the-loop judgment? The proposed framework treats the controller as a finite revisable object. Diagnostic failures are mapped to predefined rule-level edits, including rule addition, rule deletion, and priority revision. Repaired controllers are then evaluated on held-out simulation seeds or cloned initial states. Experiments in a stylized financially constrained inventory-control benchmark show three outcomes: resource-induced failures that remain non-repairable by one rule edit, partial repairs that are rejected because they violate thresholds or guardrails, and a local one-step repair of an order-volatility failure induced by removing a smoothing rule. The contribution is methodological and provides a simulation-compatible procedure for testing whether specific controller-level failures can be made observable, explainable, locally revisable, and empirically re-tested under controlled conditions.
Chinese Translation
工业代理人工智能系统在原型能力与生产部署之间日益出现差距。特别是，自适应代理可能会生成看似合理的输出，但在非确定性、保密约束、有限上下文和弱可观察性下仍然难以验证。本文为通过有限符号规则、显式诊断谓词、解释日志和保留再评估表示的自适应代理控制器制定了一种有界验证协议。中心研究问题是：当自适应代理控制器通过有限规则、显式诊断谓词、解释日志和保留再评估表示时，哪些类别的控制器故障可以被检测、局部修复或拒绝，而无需依赖于不受限制的人类判断？所提出的框架将控制器视为一个有限的可修订对象。诊断故障被映射到预定义的规则级编辑，包括规则添加、规则删除和优先级修订。修复后的控制器随后在保留的仿真种子或克隆初始状态上进行评估。在一个风格化的财务约束库存控制基准测试中，实验显示了三种结果：由于一个规则编辑而导致的资源引发的故障无法修复、因违反阈值或保护措施而被拒绝的部分修复，以及通过删除平滑规则而引发的订单波动故障的局部一步修复。该贡献具有方法论意义，并提供了一种与仿真兼容的程序，用于测试特定控制器级故障是否可以在受控条件下变得可观察、可解释、局部可修订并进行实证再测试。
cs.AI / 24 / 2607.09773
EvoCUA-1.5: Online Reinforcement Learning for Multi-turn Computer-Use Agents
EvoCUA-1.5：用于多轮计算机使用代理的在线强化学习
Abstract
Computer-use agents must solve long-horizon tasks through repeated interaction with partially observable, multimodal desktop environments. Although imitation learning and offline trajectory refinement provide strong priors, static traces cannot cover the causal feedback loop of real computer use: each action changes the screen state, future action space, and recovery options. EvoCUA-1.5 extends self-evolving computer-use agents from offline experience learning to online reinforcement learning, where policies interact with executable sandbox environments and improve from verifiable task outcomes. Online RL in this setting requires more than directly reusing single-turn language-RL recipes. Multi-turn interaction introduces context-managed observations, sparse terminal rewards, variable-length trajectories, and slow environment feedback. EvoCUA-1.5 addresses these challenges with Step-Level Policy Optimization (STEPO), which preserves trajectory-level advantage balance after decomposition into step-level samples; policy-aware filtering and pass-rate calibration over verifiable synthesized tasks; Dynamic Tri-Adaptive Curriculum (DTAC), which combines learnable tasks, difficult positive replay, and controlled infeasible-task exposure; and a fully asynchronous RL infrastructure with staleness control and mini-group batching. Experiments show that these components improve training stability and downstream performance. EvoCUA-1.5 achieves 63.2\% success on OSWorld-Verified, outperforming comparable 32B/35B-scale open-weight baselines and even approaching models with significantly larger parameter counts. Overall, EvoCUA-1.5 provides a practical framework for scaling online RL in multi-turn computer-use agents.
Chinese Translation
计算机使用代理必须通过与部分可观察的多模态桌面环境的反复交互来解决长时间跨度的任务。尽管模仿学习和离线轨迹优化提供了强有力的先验，但静态轨迹无法覆盖真实计算机使用的因果反馈循环：每个动作都会改变屏幕状态、未来的动作空间和恢复选项。EvoCUA-1.5将自我进化的计算机使用代理从离线经验学习扩展到在线强化学习，在这种情况下，策略与可执行的沙箱环境进行交互，并根据可验证的任务结果进行改进。在这种设置下，在线强化学习不仅仅是直接重用单轮语言-强化学习配方。多轮交互引入了上下文管理的观察、稀疏的终端奖励、可变长度轨迹和缓慢的环境反馈。EvoCUA-1.5通过步骤级策略优化（Step-Level Policy Optimization，STEPO）来应对这些挑战，该方法在分解为步骤级样本后保持轨迹级优势平衡；对可验证合成任务进行策略感知过滤和通过率校准；动态三重自适应课程（Dynamic Tri-Adaptive Curriculum，DTAC），结合可学习任务、困难的正回放和受控的不可行任务暴露；以及具有陈旧控制和小组批处理的完全异步强化学习基础设施。实验表明，这些组件提高了训练稳定性和下游性能。EvoCUA-1.5在OSWorld-Verified上实现了63.2%的成功率，超越了可比的32B/35B规模的开放权重基线，甚至接近参数数量显著更大的模型。总体而言，EvoCUA-1.5为多轮计算机使用代理中在线强化学习的扩展提供了一个实用框架。
cs.AI / 25 / 2607.09781
From Patterns to Maze Structures: SMT-Based Path Synthesis and 2D/3D Construction
从模式到迷宫结构：基于 SMT 的路径合成与 2D/3D 构建
Abstract
We present a pipeline for constructing maze structures from input patterns such as text or shapes. The central path-synthesis problem is encoded in Satisfiability Modulo Theories as global constraints on adjacency, continuity, and pattern-constrained coverage, allowing each fixed-bound instance to be solved in one call. The resulting path is either a planar, self-avoiding route or a layered traversal with prescribed over--under crossings, and it serves as a scaffold for constructing planar mazes and three-dimensional realizations of woven mazes. This report extends the published Bridges 2026 conference paper with more representative SMT-LIB examples and a fuller account of how synthesized paths become concrete maze constructions in planar and three-dimensional form.
Chinese Translation
我们提出了一种从输入模式（如文本或形状）构建迷宫结构的流程。核心的路径合成问题被编码为满足性理论（Satisfiability Modulo Theories）中的全局约束，涵盖邻接性、连续性和模式约束覆盖，使得每个固定边界实例可以在一次调用中解决。生成的路径可以是平面自避路线或具有规定上下交叉的分层遍历，并且作为构建平面迷宫和三维编织迷宫实现的支架。本报告扩展了已发表的 Bridges 2026 会议论文，提供了更多具有代表性的 SMT-LIB 示例，并更全面地说明了合成路径如何转化为平面和三维形式的具体迷宫构造。
cs.AI / 26 / 2607.09786
Length Penalties Make Chain-of-Thought Less Monitorable
长度惩罚使得思维链的可监控性降低
Abstract
Length-penalized reinforcement learning can shorten chain-of-thought reasoning while hiding an influence that drives the model's answer. In our experiments, training with length penalties does not stop misleading hints from steering models, even though the models' chains of thought mention the hint much less often. A token-accuracy evaluation would count these runs as successful because they use fewer reasoning tokens with little accuracy loss; it would miss whether the remaining trace still shows what drove the answer. We train Qwen3-4B and Qwen3-14B variants with different target chain lengths, then evaluate them with biasing-hint interventions on held-out MMLU-Pro-R and four transfer benchmarks. Compression sharply cuts reasoning tokens, preserves most multiple-choice accuracy, and leaves hint influence near baseline. At the strongest target, lower-bound faithfulness falls to 63.1% of baseline for Qwen3-14B and 69.4% for Qwen3-4B; the raw rate at which a monitor catches hint use falls from 69% to 49% and from 60% to 48%. To separate length from content, we randomly delete sentences from uncompressed baseline chains until the remaining text matches the compressed length. Even after this length matching, compressed chains disclose the hint 7-35 percentage points less often than baseline chains that we shorten at random, for both Qwen3 sizes and all five evaluation distributions. Compression therefore does more than shorten reasoning, preferentially removing the cues a monitor needs to see what influenced the answer. Together, these results reveal a compression-monitorability frontier in which cheaper reasoning can preserve answers while making the influences behind them harder to detect.
Chinese Translation
长度惩罚的强化学习可以缩短思维链推理，同时隐藏驱动模型答案的影响。在我们的实验中，使用长度惩罚的训练并未阻止误导性提示对模型的引导，尽管模型的思维链提及提示的频率大大降低。一个基于标记准确度的评估会将这些实验视为成功，因为它们使用了更少的推理标记且准确度损失很小；但它未能揭示剩余的痕迹是否仍然显示了驱动答案的因素。我们训练了不同目标链长度的 Qwen3-4B 和 Qwen3-14B 变体，然后在保留的 MMLU-Pro-R 和四个迁移基准上进行偏向提示干预评估。压缩显著减少了推理标记，保留了大部分多项选择准确度，并使提示影响接近基线。在最强目标下，Qwen3-14B 的下限忠实度降至基线的 63.1%，而 Qwen3-4B 降至 69.4%；监控捕捉提示使用的原始率从 69% 降至 49%，从 60% 降至 48%。为了将长度与内容分开，我们随机删除未压缩基线链中的句子，直到剩余文本与压缩长度匹配。即使在这种长度匹配后，压缩链仍然比随机缩短的基线链更少地披露提示，减少了 7-35 个百分点，适用于两种 Qwen3 尺寸和所有五个评估分布。因此，压缩不仅仅是缩短推理，更优先去除监控所需的线索，以便看到影响答案的因素。综合来看，这些结果揭示了一个压缩-可监控性边界，在这个边界上，更便宜的推理可以保留答案，同时使得背后的影响更难以检测。
cs.AI / 27 / 2607.09789
PHITSBench: an execution-scored benchmark for AI-assisted PHITS radiation-transport input generation using natural language
PHITSBench：一种基于执行评分的基准，用于利用自然语言生成AI辅助的PHITS辐射传输输入
Abstract
We introduce PHITSBench, an execution-scored benchmark for the Monte Carlo Particle and Heavy Ion Transport code System (PHITS). PHITSBench comprises 282 transport-scorable tasks spanning three common workflow categories: parameter editing (Edit), syntax repair (Repair ), and complete simulation generation from natural-language descriptions (Reproduce). Each task is evaluated using a Composite Metric Score that combines execution success with agreement between generated and reference transport observables. Using PHITSBench, we evaluate five GPT-5.4-based configurations ranging from zero-shot prompting to knowledge-augmented and agentic workflows. Without domain-specific knowledge, the model performs well on editing and repair tasks (95% and 70% success, respectively) but fails to generate correct simulations from scratch (0% success on the Reproduce track). A structured, machine-readable PHITS knowledge catalog, supplied alongside the user manual, raises single-shot Reproduce-task success to 57%. Agentic execution provides a further improvement to 66-73%, but at increased computational cost. Failure analysis shows that the remaining errors are dominated by incorrect selection and configuration of physical observables rather than syntax generation. These results suggest that future progress in AI-assisted radiation-transport modeling will depend as much on machine-readable knowledge bases, curated domain-training datasets, and execution-grounded evaluation environments as on advances in foundation models themselves.
Chinese Translation
我们介绍了PHITSBench，它是用于蒙特卡洛粒子和重离子传输代码系统（PHITS）的一个基于执行评分的基准。PHITSBench包含282个可传输评分的任务，涵盖三种常见的工作流程类别：参数编辑（Edit）、语法修复（Repair）和从自然语言描述生成完整模拟（Reproduce）。每个任务通过一个综合指标评分进行评估，该评分结合了执行成功率和生成的传输可观测量与参考可观测量之间的一致性。使用PHITSBench，我们评估了五种基于GPT-5.4的配置，从零样本提示到知识增强和代理工作流程。没有领域特定知识的情况下，模型在编辑和修复任务上表现良好（成功率分别为95%和70%），但在从头生成正确模拟方面失败（Reproduce轨道成功率为0%）。随用户手册提供的结构化、机器可读的PHITS知识目录将单次Reproduce任务的成功率提高到57%。代理执行进一步提升了成功率至66-73%，但增加了计算成本。失败分析表明，剩余错误主要由物理可观测量的错误选择和配置所主导，而非语法生成。这些结果表明，未来AI辅助辐射传输建模的进展将同样依赖于机器可读知识库、策划的领域训练数据集和基于执行的评估环境，而不仅仅是基础模型本身的进展。
cs.AI / 28 / 2607.09790
Semantic Drift and the Stability of Operator Control in Reasoning-Class Decision Support Systems
语义漂移与推理类决策支持系统中操作者控制的稳定性
Abstract
The article investigates the fundamental problem of ensuring the stability of operator control and preserving goal-targeting in hybrid human-machine decision support systems (DSS) of a new generation. Based on a two-month continuous longitudinal experiment on the joint design of a monograph-format textual array, the latent phenomenon of semantic context drift in large language models of deep logical reasoning (Reasoning LLMs) is verified and described. A mathematical model of interaction in the human-machine interface is proposed, and an original metric is introduced - the operator control stability coefficient, which takes into account the non-linear contextual pressure of hidden reasoning chains. Within the paradigm of the cognitome theory, a critical point of control functions inversion is captured. Engineering recommendations are formulated for implementing dynamic relational arbitration loops based on a modified hierarchical similarity model.
Chinese Translation
本文探讨了确保操作者控制的稳定性和保持目标导向的基本问题，聚焦于新一代混合人机决策支持系统（DSS）。基于为期两个月的连续纵向实验，研究了在深度逻辑推理的大型语言模型（Reasoning LLMs）中潜在的语义上下文漂移现象，并进行了验证和描述。提出了一种人机界面交互的数学模型，并引入了一种新的度量——操作者控制稳定性系数，该系数考虑了隐藏推理链的非线性上下文压力。在认知组理论的范式下，捕捉到了控制功能反转的临界点。针对基于修改后的层次相似性模型实施动态关系仲裁循环，提出了工程建议。
cs.AI / 29 / 2607.09794
Agentic Context Learning with Self-Discovered Specification
自主发现规范的代理上下文学习
Abstract
Context learning is an emerging inference-time task where LLMs must learn and apply novel, task-specific knowledge from intricate contexts absent from pre-training; even frontier models score under 24% task success. In this work, we conduct a comprehensive empirical study to understand why this setting remains difficult. A natural hypothesis is that failures stem from content access; yet across twelve retrieval, reflection, and verification baselines on CL-Bench, an extensive context learning benchmark, we find limited gains over direct full-context prompting. Further failure analysis reveals a key finding: unlike typical long-context tasks such as long document understanding, context learning requires not only recovering local content but also acquiring local specifications that are often unspecified in the query but distributed across the context: domain-specific formats, local rules, and completeness conditions. Across all 31,592 rubric items, we find that 55.4% clearly evaluate specification acquisition, while only 22.6% evaluate content acquisition. Moreover, despite 76.7% of specifications being unspecified in the user query, 95.5% are traceable to the context, indicating these are learnable obligations rather than hidden requirements. To validate this diagnosis, we design a deliberately simple intervention PSCI (private specification-contract induction) which extracts local specifications and enforces them through adversarial checking and repair; PSCI achieves state-of-the-art 28.14% with GPT-5.1 (+5.59 pp absolute and +24.8% relative) on CL-Bench, replicated on Qwen3.5-27B (+5.28 pp) and Gemini 3 Pro (+6.17 pp). Seventeen ablations further isolate the role of task-specific specifications. Overall, our results suggest context learning hinges on not only content acquisition but also specification acquisition.
Chinese Translation
上下文学习是一项新兴的推理时任务，其中大型语言模型（LLMs）必须从复杂的上下文中学习并应用新颖的、特定于任务的知识，而这些知识在预训练中并不存在；即便是前沿模型的任务成功率也低于24%。在本研究中，我们进行了全面的实证研究，以理解为何这一设置仍然困难。一个自然的假设是，失败源于内容访问；然而，在CL-Bench这一广泛的上下文学习基准上，我们对十二个检索、反思和验证基线的分析发现，相较于直接的全上下文提示，提升有限。进一步的失败分析揭示了一个关键发现：与典型的长上下文任务（如长文档理解）不同，上下文学习不仅需要恢复局部内容，还需要获取通常在查询中未指定但分布于上下文中的局部规范：特定领域的格式、局部规则和完整性条件。在所有31,592个评分项中，我们发现55.4%明确评估规范获取，而仅有22.6%评估内容获取。此外，尽管76.7%的规范在用户查询中未指定，但95.5%可以追溯到上下文，表明这些是可学习的义务，而非隐含要求。为了验证这一诊断，我们设计了一种故意简单的干预措施PSCI（私有规范-合同归纳），该措施提取局部规范并通过对抗检查和修复进行强制执行；PSCI在CL-Bench上实现了28.14%的最新成果（绝对提升5.59个百分点，相对提升24.8%），并在Qwen3.5-27B（提升5.28个百分点）和Gemini 3 Pro（提升6.17个百分点）上得到了复制。十七个消融实验进一步隔离了任务特定规范的作用。总体而言，我们的结果表明，上下文学习不仅依赖于内容获取，还依赖于规范获取。
cs.AI / 30 / 2607.09839
Exploring Agentic Workflows for Generating High Quality Math Visual Aids
探索生成高质量数学视觉辅助工具的自主工作流程
Abstract
Mathematical diagrams play a crucial role in K 12 education, both as problem components and as scaffolding for student comprehension. However, current AI tools, including Large Language Models (LLMs), struggle to reliably generate accurate and pedagogically sound visual diagrams, even when provided with detailed descriptions. A significant gap therefore remains in the reliable generation of diagrams for middle school mathematics. To address this, we introduce an agentic workflow that enables LLM agents to evaluate the quality of generated visuals and use this feedback to iteratively improve their outputs. This self improvement loop aims to enhance the accuracy and educational appropriateness of AI generated diagrams. Our research investigates two questions. First, can LLMs accurately generate quality assurance questions for a visual aid given specific criteria for visual quality? Second, given valid quality assurance questions, can Vision Language Models effectively evaluate generated K 12 visual aids and use the resulting feedback to improve them iteratively? We conduct an exploratory evaluation of our agentic workflow and identify key areas for improvement, including stronger spatial reasoning and more comprehensive coverage of diagram features in the generated quality assurance questions. Our results provide preliminary evidence that this approach can improve the reliability and educational value of AI generated mathematical diagrams.
Chinese Translation
数学图表在K-12教育中扮演着至关重要的角色，既是问题的组成部分，也是学生理解的支架。然而，目前的人工智能工具，包括大型语言模型（LLMs），在生成准确且符合教育要求的视觉图表方面存在困难，即使在提供详细描述的情况下。因此，在可靠生成中学数学图表方面仍然存在显著的空白。为了解决这个问题，我们引入了一种自主工作流程，使LLM代理能够评估生成视觉图表的质量，并利用这些反馈迭代改进其输出。这个自我改进循环旨在提高人工智能生成图表的准确性和教育适宜性。我们的研究探讨了两个问题。首先，LLMs能否根据特定的视觉质量标准准确生成视觉辅助工具的质量保证问题？其次，在给定有效的质量保证问题的情况下，视觉语言模型（Vision Language Models）能否有效评估生成的K-12视觉辅助工具，并利用反馈进行迭代改进？我们对我们的自主工作流程进行了探索性评估，并确定了改进的关键领域，包括更强的空间推理能力和在生成的质量保证问题中对图表特征的更全面覆盖。我们的结果提供了初步证据，表明这种方法可以提高人工智能生成的数学图表的可靠性和教育价值。
cs.AI / 31 / 2607.09971
TopoExplore: Topological Discrimination for Archive-Based Exploration
TopoExplore：基于拓扑的归档探索区分
Abstract
Archive-based exploration methods such as Go-Explore select which visited state to return to using visitation rarity, and frontier methods return to the boundary of the unknown; neither asks whether the unexplored region behind a boundary is enterable at all. Exploration is not just about finding reward - it is about collecting a structurally complete experience for downstream learning and planning. We introduce TopoExplore, which augments Go-Explore cell selection with a periodic topological pass: enclosed unexplored regions (voids) of the visited-set occupancy grid are detected by flood fill (the H1 classes of its cubical complex), and a decaying selection bonus is placed only on their strict entrances (gap or door cells), so sealed regions are never targeted and entered regions retire. On a controlled 18-environment MiniGrid suite (15 seeds, frozen hyperparameters) TopoExplore attains a 1.52x geometric-mean speedup in median steps-to-first-entry over its exact Go-Explore ablation, versus 1.37x for a frontier baseline; frontier exploration degrades when sealed decoy structure appears (0.83-1.48x on decoy environments vs. 1.65-2.11x for TopoExplore), while TopoExplore holds its largest win on hard multi-interaction doors (10.9x). We report an honest negative on Montezuma's Revenge - without wall knowledge, unreachable occupancy artifacts capture the bonus and performance degrades as it grows, isolating the wall-aware entrance test as the load-bearing component - and a preliminary positive on HM3D scanned buildings, where the speedup over Go-Explore tracks scene difficulty (r=0.69) even as frontier selection dominates blanket coverage. The evidence supports a deliberately scoped claim: topology-aware selection pays off where enclosed structure must be discriminated, and remains competitive at open coverage, where frontier methods are strongest, despite not being tuned for that regime.
Chinese Translation
基于归档的探索方法如 Go-Explore 使用访问稀有性选择返回的访问状态，而边界方法则返回未知区域的边界；但两者都没有询问边界后面的未探索区域是否可以进入。探索不仅仅是寻找奖励——它是为了收集一个结构上完整的经验，以便于后续的学习和规划。我们引入了 TopoExplore，它通过周期性的拓扑遍历增强了 Go-Explore 的单元选择：通过洪水填充（其立方复合体的 H1 类）检测到访问集占用网格的封闭未探索区域（空洞），并仅在其严格入口（缺口或门单元）上施加衰减选择奖励，因此封闭区域永远不会被目标锁定，而已进入的区域则会退役。在一个受控的 18 环境 MiniGrid 套件上（15 个种子，固定超参数），TopoExplore 在首次进入的中位步数上相较于其精确的 Go-Explore 消融实验实现了 1.52 倍的几何平均加速，而边界基线为 1.37 倍；当出现封闭诱饵结构时，边界探索的效果下降（在诱饵环境中为 0.83-1.48 倍，而 TopoExplore 为 1.65-2.11 倍），而 TopoExplore 在困难的多交互门上获得了最大的胜利（10.9 倍）。我们在 Montezuma's Revenge 上报告了一个诚实的负面结果——在没有墙体知识的情况下，无法到达的占用伪影捕获了奖励，随着其增长，性能下降，孤立的墙体感知入口测试成为负载承载组件——以及在 HM3D 扫描建筑物上的初步积极结果，在这里，相较于 Go-Explore 的加速与场景难度相关（r=0.69），即使边界选择主导了全面覆盖。证据支持一个经过深思熟虑的声明：在必须区分封闭结构的情况下，拓扑感知选择是有回报的，并且在开放覆盖的情况下仍然具有竞争力，尽管没有针对该领域进行调优。
cs.AI / 32 / 2607.09996
Who&When Pro: Can LLMs Really Attribute Failures in AI Agents?
Who&When Pro：大型语言模型真的能归因于人工智能代理的失败吗？
Abstract
Automated failure attribution uses LLMs to identify where and why agentic systems fail. As agents become more capable, their failures become subtler, making automated attribution increasingly important. We introduce Who&When Pro, a large-scale benchmark for automated failure attribution in agentic systems. Using a strictly controlled pipeline that injects a failure only after exactly replaying a successful prefix, we construct 12,326 failed trajectories with golden labels across 3 modalities and 26 benchmarks covering various scenarios. Beyond benchmarking, we conduct extensive experiments and analyses, revealing systematic patterns in how models attribute failures across modalities, protocols, and model families, and providing empirical guidance for future automated failure attribution systems.
Chinese Translation
自动化失败归因利用大型语言模型（LLMs）来识别代理系统失败的原因和位置。随着代理能力的提升，其失败变得更加微妙，因此自动化归因变得愈发重要。我们介绍了Who&When Pro，这是一个针对代理系统自动化失败归因的大规模基准测试。通过严格控制的流程，在精确重放成功前缀后仅注入一个失败，我们构建了12,326条具有黄金标签的失败轨迹，涵盖3种模态和26个基准，涉及各种场景。除了基准测试，我们还进行了广泛的实验和分析，揭示了模型在不同模态、协议和模型家族中归因失败的系统性模式，并为未来的自动化失败归因系统提供了实证指导。
cs.AI / 33 / 2607.10021
A Symbolic Neural CPU for Quantization-Simulated Writeback and Interpretable Program Execution
一种用于量化模拟回写和可解释程序执行的符号神经CPU
Abstract
Neural networks can learn algorithmic input-output mappings, but trusting a learned executor requires more than a correct final answer because the state transitions that produce it are usually hidden. To make those transitions visible, we introduce a trace-supervised symbolic neural CPU, a factorized learned execution architecture that combines recurrent control, an explicit operation router over a fixed differentiable arithmetic-logic unit bank, destination-masked register writeback, complete trajectory supervision and matched fixed-point replay. The model exposes the selected operation, source and destination registers, register trajectory, memory signals and writeback semantics at every step. On the principal 16-wide benchmark, the non-quantized executor reproduces reference execution exactly, while the eight-bit quantization-simulated executor preserves the symbolic operation path through programs of 1,000 instructions. When the same execution is evaluated against a matched fixed-point replay, the residual numerical drift disappears, showing that it comes from a mismatch between continuous and low-precision reference semantics rather than from execution failure. We compare recurrent, Transformer, temporal-convolution, temporal graph-inspired and state-space controllers, and the ablations show that operation-gate supervision is necessary for an inspectable execution path. Hidden-opcode memory-pressure tasks expose the remaining limits in delayed state use and temporal binding. We also extend the interface with ValueMemory, hybrid adaptive leaky integrate-and-fire controllers, candidate-constrained symbolic control trained through behaviour cloning and actor-critic reinforcement learning, and an RV32I base-integer semantic bridge. Together, these results establish a trace-verifiable framework for interpretable, low-precision and controllable neural execution.
Chinese Translation
神经网络能够学习算法的输入输出映射，但信任一个学习到的执行器不仅需要正确的最终答案，因为产生该答案的状态转移通常是隐藏的。为了使这些转移可见，我们引入了一种追踪监督的符号神经CPU，这是一种分解的学习执行架构，结合了递归控制、固定可微分算术逻辑单元库上的显式操作路由器、目标掩码寄存器回写、完整轨迹监督和匹配的定点重放。该模型在每一步都暴露所选操作、源和目标寄存器、寄存器轨迹、内存信号和回写语义。在主要的16宽基准测试中，非量化执行器完全重现了参考执行，而八位量化模拟执行器在1,000条指令的程序中保留了符号操作路径。当相同的执行与匹配的定点重放进行评估时，残余数值漂移消失，表明它来自于连续和低精度参考语义之间的不匹配，而不是执行失败。我们比较了递归、Transformer、时间卷积、时间图灵启发式和状态空间控制器，消融实验表明操作门监督对于可检查的执行路径是必要的。隐藏操作码内存压力任务暴露了延迟状态使用和时间绑定的剩余限制。我们还扩展了与ValueMemory的接口，混合自适应泄漏积分和发射控制器，通过行为克隆和演员-评论家强化学习训练的候选约束符号控制，以及RV32I基础整数语义桥。综合这些结果，建立了一个可追踪验证的框架，用于可解释、低精度和可控的神经执行。
cs.AI / 34 / 2607.10059
AgentAbstain: Do LLM Agents Know When Not to Act?
AgentAbstain：大型语言模型代理是否知道何时不行动？
Abstract
Agent systems based on large language models (LLMs) are increasingly deployed for autonomous tasks, yet existing evaluations mostly focus on task success rather than whether agents know when to abstain. This gap poses real risks: under ambiguity, conflicting constraints, or tool failures, agents may execute unintended and irreversible actions. To close this gap, we present the first systematic evaluation framework for agentic abstention: the calibrated ability of tool-using LLM agents to recognize when not to act. At its core, AgentAbstain is a paired-task benchmark built on an agent-native taxonomy of 8 abstention scenarios across pre-execution reasoning and runtime discovery. It contains 263 paired tasks across 42 executable sandbox environments, where each pair consists of a should-act task and a should-abstain variant produced through a controlled perturbation to the instruction, tool, or environment state. To scale this paired design and resist data contamination, we propose AbstainGen, a fully automated pipeline that synthesizes sandbox environments and generates paired tasks end-to-end, validated by deterministic replay and semantic LLM judges; fresh task instances can be regenerated on demand, and three independent annotators rate 94-98% of sampled tasks as well-designed. Across 17 frontier LLMs in 4 agent harnesses, the best agent (Gemini 3.1 Pro) achieves only 59.5% paired accuracy (correct on both the act and abstain sides of each paired task). More importantly, abstention capability is largely independent of general task-solving capability, indicating that scaling task-solving alone will not close this gap. We further identify failure modes such as post-hoc abstention, in which agents execute irreversible actions before recognizing abstention triggers. Our code and dataset are open-sourced at agentabstain.github.io.
Chinese Translation
基于大型语言模型（LLMs）的代理系统越来越多地被部署用于自主任务，然而现有的评估主要集中在任务成功上，而非代理是否知道何时应当不行动。这一空白带来了实际风险：在模糊性、冲突约束或工具故障的情况下，代理可能会执行意图之外且不可逆转的行动。为了解决这一问题，我们提出了首个系统性评估框架，用于代理的自我克制：工具使用型LLM代理识别何时不行动的能力。在其核心，AgentAbstain是一个基于代理原生分类法的配对任务基准，涵盖了8种自我克制场景，涉及执行前推理和运行时发现。它包含263个配对任务，分布在42个可执行的沙箱环境中，每对任务由一个应行动任务和一个应自我克制的变体组成，后者通过对指令、工具或环境状态的控制扰动生成。为了扩展这一配对设计并抵御数据污染，我们提出了AbstainGen，一个完全自动化的管道，能够端到端合成沙箱环境并生成配对任务，通过确定性重放和语义LLM评审进行验证；新任务实例可以按需再生，三位独立注释者对94-98%的抽样任务评定为设计良好。在4个代理框架中的17个前沿LLM中，表现最佳的代理（Gemini 3.1 Pro）仅在配对准确率上达到了59.5%（在每个配对任务的行动和自我克制两方面均正确）。更重要的是，自我克制能力在很大程度上独立于一般任务解决能力，这表明单靠扩展任务解决能力无法填补这一空白。我们进一步识别了失败模式，例如事后自我克制，其中代理在识别自我克制触发器之前执行了不可逆转的行动。我们的代码和数据集已在agentabstain.github.io上开源。
cs.AI / 35 / 2607.10069
From ambiguous utterances to governed reuse classes: canonicalization, quotient invariance, and conditional decidability
从模糊表述到受管控的重用类：规范化、商不变性与条件可判定性
Abstract
Semantic caching defines answer reuse on embedding similarity: two utterances share a stored answer when a similarity score clears a threshold, with no notion of authorization, versioning, or of what makes two demands the same. This note changes the object on which reuse is defined: in a governed domain, reuse should operate on a mathematically characterized quotient of resolved conversational demands, not on a similarity heuristic. Three independently defined relations on resolved utterances -- reading identity, resolution identity, and reuse identity -- form a refinement chain, strict under realized nondegeneracy conditions checkable on deployment logs; the pipeline's outputs are invariant along the chain, and reuse identity is exactly the kernel of the resolution map into the governed answer partition, so the reuse quotient is the utterance-side object that partition induces, not a relabeling of it. Reuse identity licenses the governed query key and its certified answer space; reuse of a particular answer requires resolution identity or an applicability certificate. The supporting layer is stated at exactly the strength proved: exact-denotation normal forms; join aggregation as a design operator, with closure-stable cells characterizing no-escape; total computability of the full pipeline relative to an untrusted proposal layer; policy admissibility for arbitrary proposers -- and provably not factual grounding or intent fidelity; and elicitation terminating after finitely many informative replies, sound under target consistency.
Chinese Translation
语义缓存定义了基于嵌入相似性进行答案重用的方式：当相似性得分超过阈值时，两个表述共享一个存储的答案，而不考虑授权、版本控制或是什么使得两个需求相同的概念。本文改变了重用定义的对象：在一个受管控的领域中，重用应基于经过数学特征化的解决对话需求的商，而不是基于相似性启发式。对已解决表述的三种独立定义的关系——阅读身份、解决身份和重用身份——形成了一个精细化链，在实现的非退化条件下严格成立；管道的输出在该链上是不变的，而重用身份恰好是映射到受管控答案分区的解决图的核心，因此重用商是由分区诱导的表述侧对象，而不是对其的重新标记。重用身份授权了受管控查询键及其认证答案空间；特定答案的重用需要解决身份或适用性证书。支持层的表述正好在证明的强度上：精确指称的标准形式；作为设计运算符的连接聚合，具有闭合稳定单元特征化无逃逸；相对于不可信提议层的完整管道的可计算性；对任意提议者的政策可接受性——并且可证明不是基于事实的基础或意图的忠实性；以及在有限的有信息回复后终止的引导，针对目标一致性是合理的。
cs.AI / 36 / 2607.10079
MAG: A Web-Agent Benchmark and Harness for Multimodal Action and Guide Generation
MAG：一种用于多模态动作和指南生成的网络代理基准与工具
Abstract
Digital Adoption Platforms (DAPs) are embedded overlays widely used on web systems to guide users through operations inside a page, helping them get started with unfamiliar interfaces quickly. Completing a real task, however, rarely means clicking a few buttons on a single page: it takes a sequence of actions that unfolds across changing page states. Prior studies have also treated automated web agent actions and guide text generation as two separate problems, and most of them feed models textual page representations such as the DOM or accessibility trees rather than the rendered screens that humans actually operate on. In this work we introduce MAG, the first benchmark that unifies task execution and guide writing into a single Multimodal Action and Guide task, with two grounding schemes over screenshots: Set-of-Mark element selection and raw pixel coordinates. We further build a complete harness for this compound task, covering annotation with LLM assistance and human verification, training, evaluation in live environments, and joint metrics for actions and guides. With this harness we evaluate frontier API models and open multimodal models, and report detailed analyses. Finally, we design a GRPO training method augmented with expert trajectories, which nearly doubles the success rate of a supervised 9B agent (from 6.9% to 13.2%) and improves guide quality at the same time. Even the strongest model completes fewer than 40% of the tasks, leaving ample room for future research.
Chinese Translation
数字采纳平台（DAPs）是广泛应用于网络系统的嵌入式覆盖层，旨在引导用户在页面内部进行操作，帮助他们快速上手不熟悉的界面。然而，完成一个真实任务通常并不只是点击几个按钮在单一页面上：这需要在不断变化的页面状态中执行一系列操作。以往的研究将自动化网络代理的动作和指南文本生成视为两个独立的问题，而且大多数研究使用文本页面表示（如DOM或可访问性树）作为模型输入，而不是人类实际操作的渲染屏幕。在本研究中，我们介绍了MAG，这是第一个将任务执行和指南编写统一为单一多模态动作和指南任务的基准，采用了基于屏幕截图的两种基础方案：元素选择的集合（Set-of-Mark）和原始像素坐标。我们进一步构建了一个完整的工具，以支持这一复合任务，涵盖了使用大型语言模型（LLM）辅助的注释和人工验证、训练、在实时环境中的评估，以及针对动作和指南的联合指标。通过这个工具，我们评估了前沿API模型和开放多模态模型，并报告了详细的分析。最后，我们设计了一种增强专家轨迹的GRPO训练方法，几乎将一个监督的9B代理的成功率翻倍（从6.9%提高到13.2%），同时提高了指南的质量。即使是最强的模型完成的任务也不到40%，这为未来的研究留下了广阔的空间。
cs.AI / 37 / 2607.10110
Looped State-Space Language Models with Adaptive Exit-State Selection
具有自适应退出状态选择的循环状态空间语言模型
Abstract
Recent work on looped language models suggests that many reasoning problems benefit from greater computational depth rather than from additional independent parameters. Existing studies, however, focus almost exclusively on Transformer backbones, leaving open whether this principle also applies to state-space language models. We investigate Looped Mamba and Looped Hybrid Mamba-Transformer architectures, which repeatedly apply a shared Mamba (or hybrid) block to introduce explicit finite-depth recurrent computation. On two controlled reasoning tasks-Mano (modular-arithmetic manipulation) and p-hop induction-Looped Mamba consistently outperforms parameter-matched non-looped baselines and, in several settings, matches or exceeds non-looped models of equal effective depth. We then extend the study to language model pre-training under matched iso-parameter and iso-FLOPs protocols, which jointly disentangle the effects of parameter sharing and effective depth: looped models remain competitive on downstream benchmarks with substantially fewer distinct parameters, although deeper non-looped models retain an advantage in validation perplexity under strict iso-FLOPs comparisons. Finally, we adapt Ouro's two-stage exit gate to Looped Mamba for threshold-controlled selection among recurrent-step outputs. Since all recurrent steps are still executed, the selected exit step represents prediction depth rather than reduced wall-clock computation. At the scales studied, adaptive exit-state selection improves downstream performance at intermediate depths, while actual inference-time savings require additional state-handling mechanisms.
Chinese Translation
近期关于循环语言模型的研究表明，许多推理问题更依赖于更深的计算深度，而非额外的独立参数。然而，现有研究几乎完全集中于 Transformer 主干，尚未明确这一原则是否同样适用于状态空间语言模型。我们研究了循环 Mamba 和循环混合 Mamba-Transformer 架构，这些架构反复应用共享的 Mamba（或混合）模块，以引入显式的有限深度递归计算。在两个受控推理任务上——Mano（模块算术操作）和 p-hop 归纳——循环 Mamba 一直优于参数匹配的非循环基线，并且在多个设置中，其性能与等效深度的非循环模型相当或更优。随后，我们将研究扩展到语言模型的预训练，采用匹配的等参数和等 FLOPs 协议，这两个协议共同解开参数共享和有效深度的影响：循环模型在下游基准测试中仍然具有竞争力，尽管其独特参数数量显著减少，然而在严格的等 FLOPs 比较中，较深的非循环模型在验证困惑度上仍保持优势。最后，我们将 Ouro 的两阶段退出门适配到循环 Mamba，以实现对递归步骤输出的阈值控制选择。由于所有递归步骤仍然被执行，所选择的退出步骤代表预测深度，而非减少的实际计算时间。在研究的规模下，自适应退出状态选择在中间深度提高了下游性能，而实际推理时间的节省则需要额外的状态处理机制。
cs.AI / 38 / 2607.10113
Dynamic Agent Skills: A Lifecycle Survey and Taxonomy of Evolving Skill Libraries
动态代理技能：演变技能库的生命周期调查与分类
Abstract
Large language model agents increasingly store reusable procedures outside the model. These reusable procedures are often called \emph{skills}: they may be code functions, natural-language instructions, SKILL.md packages, workflow graphs, or learned adapters that a future agent can retrieve and invoke. This taxonomy-driven survey asks how such skill libraries change over time. Across a $124$-paper $2023$--$2026$ audit set, we synthesize dynamic skill systems as \emph{lifecycle-managed, verified, evolving artifact stores}: agents collect evidence from interaction, propose skill updates, verify and admit candidates, organize them for retrieval and composition, repair or prune stale entries, and govern sharing through provenance and rollback. We organize the literature around three survey tools. First, a $\text{six}$-sense taxonomy distinguishes the structurally different artifacts called ``skills'' in current papers. Second, an $\text{eight}$-stage lifecycle architecture identifies the recurring design decisions behind evidence acquisition, proposal, verification/admission, storage, retrieval/composition, maintenance, distillation/portability, and governance. Third, a lightweight skill-record schema and $\text{ten}$-operator vocabulary provide common terms for comparing library updates without elevating them into a separate method contribution. Using this structure, we synthesize evidence-graded patterns with explicit caveats: admission and repair are repeatedly important, verifier quality materially affects skill-aware RL, flat retrieval can degrade as libraries grow, and current benchmarks still under-report library trajectories, usage--utility gaps, and safety surfaces. We close with concrete reporting standards and open problems for evaluating dynamic skills as changing libraries rather than static prompt or tool collections.
Chinese Translation
大型语言模型代理越来越多地将可重用的程序存储在模型外部。这些可重用的程序通常被称为“技能”（skills）：它们可以是代码函数、自然语言指令、SKILL.md 包、工作流图或未来代理可以检索和调用的学习适配器。本文基于分类法的调查探讨了这些技能库如何随时间变化。在一个包含124篇论文的2023-2026审计集中，我们将动态技能系统综合为“生命周期管理、验证、演变的工件库”：代理从交互中收集证据，提出技能更新，验证并接纳候选者，组织它们以便于检索和组合，修复或修剪过时的条目，并通过来源和回滚来管理共享。我们围绕三种调查工具组织文献。首先，一个六感分类法区分了当前论文中称为“技能”的结构不同的工件。其次，一个八阶段生命周期架构识别了证据获取、提案、验证/接纳、存储、检索/组合、维护、提炼/可移植性和治理背后的重复设计决策。第三，一个轻量级技能记录模式和十个操作词汇提供了比较库更新的通用术语，而不将其提升为单独的方法贡献。利用这一结构，我们综合了带有明确警告的证据等级模式：接纳和修复反复重要，验证者质量实质性影响技能感知的强化学习，随着库的增长，平面检索可能会退化，而当前基准仍然低估了库的轨迹、使用-效用差距和安全表面。最后，我们提出了具体的报告标准和开放问题，以评估动态技能作为变化的库，而非静态的提示或工具集合。
cs.AI / 39 / 2607.10144
IdeaTrail: Full-Process Agent Trajectories for Scientific Ideation
IdeaTrail：科学创意的全流程代理轨迹
Abstract
Scientific research is a complex, multi-stage workflow rather than a single act of text generation. The ideation process typically emerges through literature search, paper reading, tool use, claim checking, cross-paper synthesis, brainstorming, rejection of weak directions, and iterative writing. Existing resources capture individual components of this process, but datasets that jointly record tool use, evidence acquisition, intermediate artifact evolution, and idea- or proposal-level endpoints remain limited. This report introduces \method, a multi-turn process-trajectory dataset for scientific ideation and proposal generation. Each instance records a research process from evidence gathering to either idea selection or proposal construction. Rather than freely fabricating trajectories, \method starts from human-selected high-quality research papers and proposal artifacts and uses a Generator--Advisor synthesis loop. The Generator produces the visible trajectory through actions, observations, and artifact edits, while the Advisor has access to the full generation context and checks grounding, causal order, naturalness, and leakage from hidden targets. This reverse-to-forward procedure produces multi-turn research data that remains aligned with real scientific artifacts while approximating the uncertainty, evidence use, and staged convergence of research practice. \method provides both a dataset and a general recipe for synthesizing process-supervision data for scientific research agents.
Chinese Translation
科学研究是一个复杂的多阶段工作流程，而不是单一的文本生成行为。创意过程通常通过文献检索、论文阅读、工具使用、论点检查、跨论文综合、头脑风暴、拒绝薄弱方向和迭代写作等环节逐步形成。现有资源捕捉了这一过程的个别组成部分，但联合记录工具使用、证据获取、中间文献演变以及创意或提案级别终点的数据集仍然有限。本报告介绍了 extit{IdeaTrail}，一个用于科学创意和提案生成的多轮过程轨迹数据集。每个实例记录了从证据收集到创意选择或提案构建的研究过程。 extit{IdeaTrail}并不是随意构建轨迹，而是从人类选择的高质量研究论文和提案文献出发，采用生成器-顾问的合成循环。生成器通过行动、观察和文献编辑产生可见轨迹，而顾问则可以访问完整的生成上下文，并检查基础、因果顺序、自然性和隐性目标的泄漏。这一反向到正向的过程生成的多轮研究数据与真实的科学文献保持一致，同时近似研究实践中的不确定性、证据使用和阶段性收敛。 extit{IdeaTrail}不仅提供了一个数据集，还为合成科学研究代理的过程监督数据提供了一般性方法。
cs.AI / 40 / 2607.10159
UNIT: Unleash Large Language Models Potential for Graph Continual Learning
UNIT：释放大型语言模型在图形持续学习中的潜力
Abstract
In real-world multimodal web scenarios, graph-structured data often arrives in a streaming manner, making graph continual learning a crucial paradigm for continuously modeling such evolving structures. However, existing graph continual learning methods still face two fundamental challenges. 1) semantic-structural separation, where the graph-based methods excel at modeling topological relationships but neglect deep semantics. 2) imbalanced knowledge transfer, where existing models fail to effectively leverage general knowledge gained from early tasks to benefit subsequent new tasks. To address above issues, we propose a novel framework, \textbf{UN}leash Large Language Models PotentIal for Graph ConTinual Learning (UNIT). By fine-tuning large language model only on the first task, we bridge the distributional gap between the pre-trained LLM corpus and the target task dataset to enhance the adaptability of LLMs for graph-structured tasks. Meanwhile, we propose an uncertain-aware anchor generation mechanism to effectively preserve representative knowledge across tasks, avoiding the neglect of universal knowledge learned from previous tasks. Additionally, we introduce structural confluence modeling to explicitly integrates graph topology information into semantic information, enhancing the collaborative capabilities between semantic understanding and structural modeling. Extensive experiments demonstrate that our proposed method achieves state-of-the-art performance in the graph continual learning task.
Chinese Translation
在现实世界的多模态网络场景中，图结构数据通常以流式方式到达，使得图形持续学习成为持续建模这些不断演变结构的重要范式。然而，现有的图形持续学习方法仍面临两个基本挑战。1）语义-结构分离，图基方法在建模拓扑关系方面表现出色，但忽视了深层语义。2）知识转移不平衡，现有模型未能有效利用从早期任务中获得的一般知识来惠及后续新任务。为了解决上述问题，我们提出了一种新颖的框架， extbf{UN}leash Large Language Models PotentIal for Graph ConTinual Learning (UNIT)。通过仅在第一个任务上微调大型语言模型，我们弥合了预训练LLM语料库与目标任务数据集之间的分布差距，从而增强了LLM在图结构任务中的适应性。同时，我们提出了一种基于不确定性的锚点生成机制，以有效保留跨任务的代表性知识，避免忽视从先前任务中学习到的普遍知识。此外，我们引入了结构融合建模，将图拓扑信息显式整合到语义信息中，增强了语义理解与结构建模之间的协作能力。大量实验表明，我们提出的方法在图形持续学习任务中达到了最先进的性能。
cs.AI / 41 / 2607.10197
GRATE: Temporal Extensions for Inductive KG Foundation Models via Gated Rotary Attention
GRATE：通过门控旋转注意力实现归纳知识图谱基础模型的时间扩展
Abstract
Knowledge graph foundation models such as Ultra and Trix achieve strong inductive transfer by learning relation-graph representations that generalise to unseen entities and relations. Extending this transferability to temporal knowledge graphs (TKGs) remains challenging: existing temporal models tie their parameters to dataset-specific entities, relations, or timestamps and are not designed to transfer to TKGs with disjoint vocabularies. We propose GRATE (Gated Rotary Attention for Temporal Encoding), an entity-side message function that adds no learnable parameters and encodes time through relative time differences by rotating each edge message according to its time gap to the query and applying a query-conditioned gate to select temporally relevant signals. GRATE integrates into NBFNet-style KG foundation models while preserving structural transferability. Existing TKG benchmarks evaluate within shared train/test vocabularies and cannot directly test cross-dataset temporal transfer; we therefore construct GDELTIndT and WIKIIndT, inductive transfer benchmark suites with disjoint entities, relations, and timestamps spanning both interpolation and extrapolation. Across these benchmarks and held-out forecasting datasets, a single jointly pretrained GRATE checkpoint improves over the static base model in most settings.
Chinese Translation
知识图谱基础模型如 Ultra 和 Trix 通过学习关系图表示，能够实现强大的归纳迁移，进而推广到未见过的实体和关系。然而，将这种迁移能力扩展到时间知识图谱（TKGs）仍然面临挑战：现有的时间模型将其参数绑定到特定数据集的实体、关系或时间戳，并未设计为能够迁移到具有不相交词汇的 TKGs。我们提出了 GRATE（门控旋转注意力用于时间编码），这是一种实体侧消息函数，不增加可学习参数，通过相对时间差来编码时间，方法是根据时间差旋转每个边消息，并应用条件于查询的门控来选择时间相关信号。GRATE 可以集成到 NBFNet 风格的知识图谱基础模型中，同时保持结构迁移能力。现有的 TKG 基准测试在共享的训练/测试词汇内进行评估，无法直接测试跨数据集的时间迁移；因此，我们构建了 GDELTIndT 和 WIKIIndT，这两个归纳迁移基准套件具有不相交的实体、关系和时间戳，涵盖插值和外推。在这些基准测试和保留的预测数据集上，单个联合预训练的 GRATE 检查点在大多数设置中优于静态基础模型。
cs.AI / 42 / 2607.10212
KGCQual: An Interpretable Framework for Evaluating the Knowledge Graph Construction Quality from Text
KGCQual：一种可解释的框架，用于评估文本知识图谱构建质量
Abstract
Knowledge Graphs (KGs) are increasingly constructed through automated extraction pipelines; however, such systems often introduce spurious or incomplete triples, which degrade downstream performance. Existing evaluation practices rely heavily on task-specific metrics or small-scale manual verification, offering limited insight into the structural and semantic fidelity of extracted graphs. We propose a novel, interpretable metric for intrinsic KG quality assessment that measures how closely an automatically extracted graph approximates an "ideal" graph capturing the key noun phrases, predicate relations, and basic linguistic phenomena such as negation expressed in the source text. Our framework integrates two complementary components: (1) an entity-level assessment that evaluates completeness, resolution quality, and connectivity, and (2) a relation-level assessment that judges predicate preservation and multiplicity using lexical similarity, dependency-parse alignment, and light-weight negation handling to ensure semantic faithfulness. We evaluate our metric across multiple state-of-the-art triple extraction systems and datasets, including WebNLG, TinyButMighty, and BenchIE, demonstrating that it reliably identifies omissions, redundancy, and structural deviations that existing metrics overlook. Our work offers a scalable, model-agnostic, and interpretable framework for comparing automated KG construction methods and provides a foundation for standardised evaluation. We further validate the metric through an ablation study isolating noun and verb components, and a downstream evaluation showing that KGCQual scores correlate significantly with link prediction performance on the same extracted KGs. The code repository is available at https://github.com/kracr/kg-quality-metric.
Chinese Translation
知识图谱（KGs）越来越多地通过自动提取管道构建；然而，这类系统往往会引入虚假或不完整的三元组，从而降低下游性能。现有的评估实践在很大程度上依赖于特定任务的指标或小规模的人工验证，提供的对提取图谱的结构和语义忠实度的洞察有限。我们提出了一种新颖的、可解释的内在KG质量评估指标，该指标衡量自动提取的图谱与捕捉源文本中关键名词短语、谓词关系和基本语言现象（如否定）等“理想”图谱的接近程度。我们的框架整合了两个互补的组成部分：（1）实体级评估，评估完整性、分辨率质量和连通性；（2）关系级评估，利用词汇相似性、依赖解析对齐和轻量级否定处理来判断谓词的保留和多样性，以确保语义的忠实性。我们在多个最先进的三元组提取系统和数据集（包括WebNLG、TinyButMighty和BenchIE）上评估了我们的指标，证明它能够可靠地识别现有指标忽视的遗漏、冗余和结构偏差。我们的工作提供了一种可扩展的、模型无关的、可解释的框架，用于比较自动KG构建方法，并为标准化评估提供了基础。我们还通过消融研究验证了该指标，隔离了名词和动词组件，并通过下游评估显示KGCQual得分与相同提取KG的链接预测性能显著相关。代码库可在 https://github.com/kracr/kg-quality-metric 获取。
cs.AI / 43 / 2607.10226
When Are Sparse Feature Interventions Actually Localized? Matched Evaluation for SAE-Based Safety Control
稀疏特征干预何时真正局部化？基于SAE的安全控制的匹配评估
Abstract
We evaluate when sparse autoencoder (SAE) features act as localized control handles for safety-relevant behavior. This question is difficult because apparent success can arise from weak interventions, mismatched baselines, model robustness, or degenerate outputs that automated safety judges mark as unsafe without representing meaningful harmful compliance. We introduce a matched coherence-gated evaluation protocol for runtime safety interventions: methods are compared at matched target-effect points, and the primary target metric counts harmful compliance only when an output is both judge-unsafe and coherent. Applying this protocol to three prompt splits on Gemma-2-9B-it with a Gemma Scope layer-20 residual SAE, we find that SAE feature ablation has a narrow useful regime. SAE top800 reaches a low-to-mid target effect with lower total perturbation and competitive utility, but SAE top1600 loses utility relative to a matched dense refusal-direction baseline, and SAE top3200 primarily induces coherence collapse. Human audit confirms that coherence gating removes unsafe-only artifacts, and feature diagnostics show that the useful regime is driven by a stable head of refusal-aligned features whose activation separation decays rapidly with rank. These results argue that SAE-based safety interventions should be evaluated as regime-dependent control mechanisms rather than assumed to be uniformly localized.
Chinese Translation
我们评估稀疏自编码器（SAE）特征何时作为安全相关行为的局部控制手段。这个问题很复杂，因为表面上的成功可能源于弱干预、不匹配的基线、模型的鲁棒性，或者自动安全评判者标记为不安全的退化输出，而这些输出并未代表有意义的有害合规。我们引入了一种匹配一致性门控评估协议，用于运行时安全干预：在匹配的目标效果点比较方法，主要目标指标仅在输出既被评判为不安全又一致时才计算有害合规。将该协议应用于Gemma-2-9B-it上的三个提示分割，使用Gemma Scope层-20残差SAE，我们发现SAE特征消融具有狭窄的有效范围。SAE top800在较低到中等目标效果下达到较低的总扰动和竞争性效用，但SAE top1600相对于匹配的稠密拒绝方向基线失去效用，而SAE top3200主要导致一致性崩溃。人工审核确认一致性门控去除了仅不安全的伪影，特征诊断显示有效范围由一组稳定的拒绝对齐特征驱动，其激活分离随着等级迅速衰减。这些结果表明，基于SAE的安全干预应作为依赖于状态的控制机制进行评估，而不是假设其均匀局部化。
cs.AI / 44 / 2607.10251
Behavioural Signatures of Risk-Sensitive Decision-Making in Large Language Models
大型语言模型中风险敏感决策的行为特征
Abstract
As large language models (LLMs) are increasingly used in decision support, it is important to understand whether their choices under uncertainty exhibit stable and interpretable behavioural regularities. Human decision-making combines relatively persistent risk preferences with context-dependent adjustment, yet it remains unclear whether analogous behavioural structure can be observed in LLM-based decision systems. Here we examine this question using a controlled multi-model framework based on no-limit Texas Hold'em, where behaviour is quantified by Participation, measuring voluntary engagement in uncertain opportunities, and Proactiveness, measuring pre-flop risk escalation. Across homogeneous self-play and heterogeneous mixed-model interactions, frontier LLMs exhibit stable, model-specific risk profiles, forming a spectrum from conservative to aggressive decision styles. These profiles remain largely robust under changing opponent composition, while the most conservative and most aggressive models diverge further in mixed settings. Under global risk pressure and personal resource constraint, models adapt in structured but heterogeneous ways, ranging from broad behavioural contraction to selective de-escalation and near-invariant behaviour. These findings suggest that LLMs differ not only in baseline risk disposition, but also in the risk signals they respond to and the flexibility with which they adjust, providing a behavioural basis for auditing risk-sensitive decision-making in interactive settings. Our code is publicly available at: https://github.com/XuankunRong/AgentTexasPoker.
Chinese Translation
随着大型语言模型（LLMs）在决策支持中的应用日益广泛，理解它们在不确定性下的选择是否表现出稳定且可解释的行为规律变得尤为重要。人类决策结合了相对持久的风险偏好与依赖于情境的调整，但尚不清楚在基于LLM的决策系统中是否可以观察到类似的行为结构。在此，我们使用基于无限注德州扑克的受控多模型框架来探讨这一问题，其中行为通过参与度（Participation）和主动性（Proactiveness）进行量化，前者衡量在不确定机会中的自愿参与程度，后者衡量翻牌前的风险升级。在同质自我对弈和异质混合模型交互中，前沿LLMs表现出稳定的、特定于模型的风险特征，形成从保守到激进的决策风格谱系。这些特征在对手组成变化时仍然保持相对稳健，而最保守和最激进的模型在混合环境中进一步分化。在全球风险压力和个人资源限制下，模型以结构化但异质的方式进行适应，表现出从广泛的行为收缩到选择性去升级及近乎不变的行为。这些发现表明，LLMs不仅在基线风险倾向上存在差异，还在响应的风险信号和调整的灵活性上有所不同，为在互动环境中审计风险敏感决策提供了行为基础。我们的代码已公开发布于：https://github.com/XuankunRong/AgentTexasPoker。
cs.AI / 45 / 2607.10275
Information-seeking failures of large language models in agentic clinical reasoning
大型语言模型在代理临床推理中的信息寻求失败
Abstract
Large language models achieve high scores on medical knowledge assessments, yet clinical reasoning requires actively deciding what to investigate under uncertainty. We developed an agentic evaluation framework in hematologic oncology in which models must proactively request clinical data across three sequential rounds before committing to a diagnosis and treatment plan. Across 32 frontier models, the best achieved only 68% overall accuracy. Information utilization, the fraction of available data actually requested, was the strongest predictor of diagnostic accuracy (R = 0.69, P < 0.001), yet utilization collapsed from 57% to 26% in the final round, leaving molecular and cytogenetic data critical for treatment selection unexamined. Reasoning traces scored high on a clinical reasoning rubric (91% above threshold) but decorrelated from accuracy, revealing a gap between locally coherent rationales and globally correct conclusions. Error analysis identified search satisficing, anchoring and premature closure as the dominant failure modes, the same cognitive biases that characterize novice clinicians under dual-process models of diagnostic reasoning. These findings demonstrate that the primary limitation of current models in clinical oncology is not insufficient medical knowledge but a systematic failure of information-seeking under uncertainty.
Chinese Translation
大型语言模型在医学知识评估中取得了高分，但临床推理需要在不确定性下主动决定调查内容。我们开发了一个在血液肿瘤学领域的代理评估框架，其中模型必须在确认诊断和治疗方案之前，主动请求三个连续回合中的临床数据。在32个前沿模型中，表现最佳的模型仅达到了68%的整体准确率。信息利用率，即实际请求的可用数据的比例，是诊断准确性的最强预测因子（R = 0.69, P < 0.001），然而在最后一轮中，利用率从57%骤降至26%，导致对治疗选择至关重要的分子和细胞遗传数据未被检查。推理轨迹在临床推理评分标准中得分较高（91%超过阈值），但与准确性脱钩，揭示了局部一致的推理与全球正确结论之间的差距。错误分析识别出搜索满足、锚定和过早关闭是主要的失败模式，这些是双过程模型下新手临床医生所特有的认知偏见。这些发现表明，当前模型在临床肿瘤学中的主要限制并非医学知识不足，而是在不确定性下信息寻求的系统性失败。
cs.AI / 46 / 2607.10286
Can Agentic Trading Systems Pay for Their Own Intelligence?
代理交易系统能否自负其智力成本？
Abstract
Large language model (LLM) agents are increasingly used in trading systems, where model reasoning, tool use, and continual decisions incur costs that are expected to produce trading value. Existing evaluations typically report performance metrics, but rarely examine agentic viability: whether dynamic LLM-mediated decisions convert their induced costs into measurable incremental profit. To apply this criterion, we introduce TradeLens, a trace-grounded diagnostic toolkit for evaluating agentic trading systems from their trading records, runtime traces, and deployment configurations. It reconstructs trading trajectories, attributes profit and cost to interpretable evidence, and diagnoses whether and why an agent pays for its own intelligence. We conduct extensive analysis across backbone models, capital scales, trading frequencies, and system architectures, together with deployment discussion. Our results show that viability hinges on intelligence-to-profit conversion: models exhibit different failure patterns, such as poor asset selection in DeepSeek-V3.2 and negative timing in GLM-4.7, while capital scale, trading frequency, and architecture matter only by amplifying or degrading decision-attributed timing value. These findings reframe the evaluation of LLM-based trading agents from capability-centric performance ranking to trace-grounded diagnosis of intelligence-to-profit conversion. Our code is available at https://anonymous.4open.science/r/TradeLens.
Chinese Translation
大型语言模型（LLM）代理在交易系统中的应用日益增多，其中模型推理、工具使用和持续决策所产生的成本被期望能够带来交易价值。现有评估通常报告性能指标，但很少考察代理的可行性：即动态的LLM介导决策是否将其产生的成本转化为可衡量的增量利润。为应用这一标准，我们引入了TradeLens，一个基于交易记录、运行时跟踪和部署配置的追踪基础诊断工具包，用于评估代理交易系统。它重建交易轨迹，将利润和成本归因于可解释的证据，并诊断代理是否以及为何能自负其智力成本。我们对主干模型、资本规模、交易频率和系统架构进行了广泛分析，并讨论了部署情况。我们的结果表明，可行性取决于智力与利润的转化：模型表现出不同的失败模式，例如DeepSeek-V3.2中的资产选择不佳和GLM-4.7中的负时机，而资本规模、交易频率和架构仅通过放大或削弱决策归因的时机价值而产生影响。这些发现将基于LLM的交易代理的评估从以能力为中心的性能排名重新框架为基于追踪的智力与利润转化诊断。我们的代码可在 https://anonymous.4open.science/r/TradeLens 获取。
cs.AI / 47 / 2607.10296
SPARK: Susceptibility-Guided Profiling and Steering of Latent Reasoning States in Large Language Models
SPARK：基于易感性引导的大型语言模型潜在推理状态的分析与引导
Abstract
Reasoning failures in large language models (LLMs) are usually evaluated from final answers, but a wrong answer does not reveal why the model failed. The same incorrect output may reflect missing capability, an unstable reasoning trajectory, or a failure to activate a reasoning state that is already available in the frozen model. Existing prompting and benchmark-based evaluation methods mostly operate at the output level, while generic activation-steering methods typically apply global directions without diagnosing which examples require intervention. In this paper, we introduce SPARK, which uses hidden-state response to diagnose whether a model internally enters an effective reasoning state and to guide lightweight test-time steering. The key observation is that raw hidden-state susceptibility is strongly confounded by prompt length, especially in programmatic and algorithmic reasoning where harder serialized instances naturally become longer. SPARK therefore uses length-controlled susceptibility to separate input-scale effects from residual reasoning activation, and combines this signal with cross-layer coordination to select reasoning-active anchors and under-activated hard examples. We use FRONTIER-4.5K as a controlled programmatic reasoning suite for latent profiling and difficulty-aware analysis, and evaluate SPARK-Steering on GSM8K and MATH-500 with forward-only benchmark profiling. Our method improves Qwen3 series models consistently; on MATH-500, accuracy rises from 82.0% to 84.6% for Qwen3-4B and from 82.4% to 85.6% for Qwen3-8B. These results suggest that susceptibility can serve not only as a diagnostic signal for reasoning failures, but also as a practical guide for targeted test-time intervention.
Chinese Translation
大型语言模型（LLMs）中的推理失败通常通过最终答案进行评估，但错误的答案并不能揭示模型失败的原因。同样的错误输出可能反映出缺失的能力、不稳定的推理轨迹，或未能激活已在冻结模型中可用的推理状态。现有的提示和基准评估方法大多数在输出层面进行操作，而通用的激活引导方法通常应用全局方向，而未能诊断哪些示例需要干预。在本文中，我们引入了SPARK，它利用隐藏状态响应来诊断模型是否在内部进入有效的推理状态，并指导轻量级的测试时引导。关键观察是，原始隐藏状态的易感性受到提示长度的强烈干扰，尤其是在程序性和算法性推理中，较难的序列化实例自然变得更长。因此，SPARK使用长度控制的易感性将输入规模效应与残余推理激活分离，并将该信号与跨层协调结合，以选择推理活跃的锚点和未激活的困难示例。我们使用FRONTIER-4.5K作为受控的程序性推理套件进行潜在分析和难度感知分析，并在GSM8K和MATH-500上评估SPARK引导，采用前向基准分析。我们的方法在Qwen3系列模型上始终表现出改进；在MATH-500上，Qwen3-4B的准确率从82.0%提高到84.6%，Qwen3-8B的准确率从82.4%提高到85.6%。这些结果表明，易感性不仅可以作为推理失败的诊断信号，还可以作为针对性测试时干预的实用指南。
cs.AI / 48 / 2607.10309
Measure the Sim-to-Real Gap: Designing an Affordable Real-World Benchmark Platform for Reinforcement Learning in AIoT Systems
测量模拟与现实之间的差距：为人工智能物联网系统中的强化学习设计一个经济实惠的现实基准平台
Abstract
Reinforcement learning (RL) is commonly employed to enhance the performance of autonomous systems, including the Autonomous Internet of Things (AIoT). However, the trial-and-error nature of RL, when conducted in real-world environments, is costly and hazardous in some scenarios. Consequently, the majority of RL research is conducted in simulation. This reliance introduces challenges related to the Sim-to-Real transferability. Evaluating the Sim-to-Real algorithmic robustness and the Sim-to-Real gap is a critical prerequisite for research aimed at improving RL performance in the real world. Therefore, industries such as robotics have developed concurrent simulation and physical platforms to facilitate this research. However, a universal Sim-to-Real benchmark platform for AIoT does not currently exist. To address these concerns, we developed a real-world AIoT platform for studying RL in AIoT. On this platform, an agent deployed on an edge device plays video games on a separate host computer via a hardware-emulated keyboard, guided by vision input. This platform uses commercially available components costing less than USD 400, together with two computers. Because the system's objective is game score maximization, it inherently mitigates safety risks associated with real-world RL deployments. Experimental results show the simulation-trained agent suffers a 1160% performance degradation relative to the human-level performance after real-world deployment, indicating a significant Sim-to-Real gap. Direct real-world training using the deep Q-network (DQN) algorithm achieves approximately 38% of human-level performance after 10 million training steps, demonstrating the feasibility of RL under real-world conditions. These results suggest that the proposed Sim-to-Real benchmark platform provides a substantial foundation for qualitative and quantitative evaluations of RL in real-world AIoT systems.
Chinese Translation
强化学习（RL）通常用于提升自主系统的性能，包括自主物联网（AIoT）。然而，在现实环境中进行的强化学习试验具有试错性质，这在某些情况下是昂贵且危险的。因此，大多数强化学习研究是在模拟环境中进行的。这种依赖性引入了与模拟到现实（Sim-to-Real）可转移性相关的挑战。评估模拟到现实的算法鲁棒性和模拟到现实的差距是旨在提升现实世界中强化学习性能的研究的重要前提。因此，机器人等行业已经开发了并行的模拟和物理平台以促进这项研究。然而，目前尚不存在一个通用的AIoT模拟到现实基准平台。为了解决这些问题，我们开发了一个现实世界的AIoT平台，用于研究AIoT中的强化学习。在该平台上，部署在边缘设备上的代理通过硬件仿真键盘在单独的主机计算机上玩视频游戏，受视觉输入的指导。该平台使用的组件均为市售产品，成本低于400美元，并配备两台计算机。由于系统的目标是最大化游戏得分，因此本质上减轻了与现实世界强化学习部署相关的安全风险。实验结果表明，经过模拟训练的代理在现实世界部署后相较于人类水平性能下降了1160%，这表明存在显著的模拟到现实差距。使用深度Q网络（DQN）算法进行的直接现实世界训练在经过1000万步训练后达到了约38%的人类水平性能，证明了在现实条件下进行强化学习的可行性。这些结果表明，所提出的模拟到现实基准平台为现实世界AIoT系统中强化学习的定性和定量评估提供了重要基础。
cs.AI / 49 / 2607.10331
Comparing Socio-technical Design Principles with Guidelines for Human-centered AI
比较社会技术设计原则与以人为本的人工智能指南
Abstract
Human-centered AI (HCAI) refers to guidelines or principles that aim on ethi-cally oriented design of systems. We compare HCAI- guidelines with princi-ples of socio-technical systems that emerged in the context of conventional in-formation technology. The comparison leads to a revision of socio-technical heuristics by including aspects of AI-usage. The comparison reveals that con-tinuous evolution is a basic characteristic of socio-technical systems, and that human oversight or interventions and the subsequent appropriation of AI-systems lead to continuous adaptation and re-design of the systems, if autono-my is collaboratively exercised. From a socio-technical point of view, the cru-cial requirement of transparency has not only to be fulfilled with technical fea-tures, but also by contributions of the whole system including human actors. It will be promising for using AI, if not only technical features, but organization-al and social practices are socio-technically designed in a way that compen-sates shortcomings of AI.
Chinese Translation
以人为本的人工智能（HCAI）指的是旨在伦理导向设计系统的指南或原则。我们将HCAI指南与在传统信息技术背景下出现的社会技术系统原则进行比较。这一比较促使我们修订社会技术启发式，纳入人工智能使用的相关方面。比较结果表明，持续演变是社会技术系统的基本特征，而人类的监督或干预以及随之而来的人工智能系统的适应性使用，促使系统的持续调整和重新设计，尤其是在自主权得到协作行使的情况下。从社会技术的角度来看，透明度这一关键要求不仅需要通过技术特性来满足，还需要整个系统的贡献，包括人类参与者的作用。如果不仅技术特性，而是组织和社会实践也以社会技术的方式设计，以弥补人工智能的不足，那么使用人工智能将会更具前景。
cs.AI / 50 / 2607.10350
ABot-AgentOS: A General Robotic Agent OS with Lifelong Multi-modal Memory
ABot-AgentOS：一种具有终身多模态记忆的通用机器人代理操作系统
Tian, Jiayi, Liu, Shiao, Xu, Yuting, Lu, Jia, Guan, Zihao, Han, Honglin, Yang, Di, Gu, Minqi, Qian, Yifei, Zhang, Tianlin, Zhu, Yanqing, Ye, Zeqian, Yang, Menglin, Wang, Fei, Hu, Xu, Li, Xiuxian, Zhang, Wei, Su, Shihui, Ji, Yiyan, Wang, Jingbo, Feng, Ziteng, Liu, Jiaheng, Zhang, Zhaoxiang, Wu, Xiaolong, Yin, Mingyang, Chu, Zedong, Xu, Mu
Abstract
Recent VLM and VLA systems have improved robotic perception and action prediction, yet long-horizon embodied agents still require a general runtime layer for reasoning, memory, tool use, verification, and cross-embodiment execution. We present ABot-AgentOS, a general robotic Agent Operating System that sits above low-level controllers and provides a deliberative agent layer for scene-conditioned planning, context-isolated skill execution, multi-stage verification, multi-modal memory, and edge-cloud collaboration. To evaluate such systems, we introduce EmbodiedWorldBench, an executable benchmark with 16 indoor, outdoor, and hybrid scenes, four difficulty levels, and over 200 tasks involving navigation, object search, NPC dialogue, dynamic events, and trace-grounded scoring. ABot-AgentOS further introduces Universal Multi-modal Graph Memory, a persistent source-grounded substrate that converts dialogue, visual observations, spatial context, temporal relations, and task traces into typed nodes and edges. A failure-driven self-evolution loop converts diagnosed memory failures into gated runtime evo-assets that are promoted only to later evaluation splits, preventing current-split ground-truth leakage while enabling continual improvement. On an initial EmbodiedWorldBench subset, ABot-AgentOS improves over a single-controller baseline in both task success and goal completion. Across memory benchmarks, ABot-AgentOS Static achieves 87.5 on LoCoMo, 59.9 on OpenEQA EM-EQA, 88.6 on Mem-Gallery, and 76.5 Acc@All on NExT-QA; self-evolution further improves LoCoMo to 88.7, OpenEQA to 60.4, and Mem-Gallery to 89.0. These results suggest that a general Agent OS layer can improve long-horizon embodied execution while providing persistent, auditable memory for continual interaction.
Chinese Translation
近期的视觉语言模型（VLM）和视觉语言代理（VLA）系统提升了机器人的感知和动作预测能力，但长时间跨度的具身代理仍然需要一个通用的运行时层，以支持推理、记忆、工具使用、验证和跨具身执行。我们提出了ABot-AgentOS，这是一种通用的机器人代理操作系统，位于低级控制器之上，为场景条件规划、上下文隔离的技能执行、多阶段验证、多模态记忆和边缘-云协作提供了深思熟虑的代理层。为了评估此类系统，我们引入了EmbodiedWorldBench，这是一个可执行的基准，包含16个室内、室外和混合场景，四个难度级别，以及200多个涉及导航、物体搜索、NPC对话、动态事件和基于轨迹的评分的任务。ABot-AgentOS进一步引入了通用多模态图记忆（Universal Multi-modal Graph Memory），这是一个持久的源基础，能够将对话、视觉观察、空间上下文、时间关系和任务轨迹转换为类型化的节点和边。一个基于失败驱动的自我进化循环将诊断出的记忆失败转化为门控运行时进化资产，这些资产仅在后续评估分割中被提升，从而防止当前分割的真实数据泄漏，同时实现持续改进。在初步的EmbodiedWorldBench子集上，ABot-AgentOS在任务成功率和目标完成度上均优于单控制器基线。在记忆基准测试中，ABot-AgentOS Static在LoCoMo上取得了87.5，在OpenEQA EM-EQA上为59.9，在Mem-Gallery上为88.6，在NExT-QA的Acc@All上为76.5；自我进化进一步将LoCoMo提升至88.7，OpenEQA提升至60.4，Mem-Gallery提升至89.0。这些结果表明，一个通用的代理操作系统层可以改善长时间跨度的具身执行，同时提供持久的、可审计的记忆，以支持持续的交互。
cs.AI / 51 / 2607.10366
Co4ICF: Co-evolving Physics-Informed Surrogate and RL-based Pulse Optimizer for Inertial Confinement Fusion
Co4ICF：协同演化的物理信息代理与基于强化学习的脉冲优化器在惯性约束聚变中的应用
Abstract
Offline-trained surrogates for Inertial Confinement Fusion (ICF) suffer a well-known failure mode that iterative optimizers drive inputs into out-of-distribution (OOD) regions where predictions become unreliable. Here we present Co4ICF, a co-evolving framework that couples a physics-informed surrogate with a PPO-based pulse optimizer. The surrogate is iteratively fine-tuned on policy-induced trajectories, correcting extrapolation errors as the optimizer shifts the input distribution; the optimizer queries this evolving surrogate as a fast environment. In the 1D MULTI environment, Co4ICF achieves 146.1% normalized yield based on current laser design baseline; as a post-hoc cross-fidelity check, the optimized pulse further attains 246.9% normalized yield when directly evaluated in 2D-MULTI without any 2D training or fine-tuning. Budget-matched ablations support that the gains are not explained solely by additional simulation data and are consistent with the co-evolving mechanism playing a key role. We release a large-scale MULTI-IFE simulation dataset to support future benchmarking.
Chinese Translation
离线训练的惯性约束聚变（ICF）代理面临一个众所周知的失败模式，即迭代优化器将输入驱动到分布外（OOD）区域，导致预测变得不可靠。在此，我们提出了Co4ICF，一个将物理信息代理与基于PPO（Proximal Policy Optimization）的脉冲优化器相结合的协同演化框架。该代理在策略诱导的轨迹上进行迭代微调，随着优化器调整输入分布，修正外推误差；优化器将这个不断演化的代理作为快速环境进行查询。在1D MULTI环境中，Co4ICF实现了基于当前激光设计基线的146.1%的归一化产量；作为后验交叉保真度检查，优化后的脉冲在没有任何2D训练或微调的情况下，直接在2D-MULTI中评估时进一步达到了246.9%的归一化产量。预算匹配的消融实验支持这些增益不仅仅是由额外的模拟数据解释的，并且与协同演化机制在其中发挥关键作用的观点一致。我们发布了一个大规模的MULTI-IFE模拟数据集，以支持未来的基准测试。
cs.AI / 52 / 2607.10455
ANCHOR: Automated Alignment Auditing for CLI Agents on Real-World Harm
ANCHOR：针对现实世界危害的命令行代理自动对齐审计
Abstract
Autonomous CLI agents can now execute hundreds of actions across multi-hour sessions: writing code, executing shell commands, browsing the web, and managing cloud infrastructure, all with minimal human oversight. Does greater autonomy invite greater risk? We introduce ANCHOR, an automated auditing framework that stress-tests CLI agents on illegal tasks grounded in public US court cases. ANCHOR deploys an auditor agent fine-tuned on dark personality data using supervised and reinforcement fine tuning. This auditor roleplays persistent malicious users who decompose tasks, reframe requests upon refusal, and adapt strategies across multi-turn interactions. Evaluating frontier CLI agents, we find that while they often refuse illegal tasks when prompted directly, compliance reaches 100\% under persistent malicious interaction. When agents comply, they frequently exceed user requests, autonomously building infrastructure for large-scale harm, including catastrophic risk scenarios such as large-scale financial fraud and bioweapon development. These findings demonstrate that current alignment techniques are insufficient for autonomous agents and underscore the need for safety evaluations against persistent, adaptive malicious users. We release ANCHOR at https://github.com/garified/anchor
Chinese Translation
自主命令行（CLI）代理现在可以在数小时的会话中执行数百个操作：编写代码、执行 shell 命令、浏览网页和管理云基础设施，且几乎不需要人工监督。更大的自主性是否会带来更大的风险？我们介绍了 ANCHOR，一个自动化审计框架，旨在对基于美国公共法庭案件的非法任务对 CLI 代理进行压力测试。ANCHOR 部署了一个经过微调的审计代理，该代理使用监督学习和强化学习对黑暗人格数据进行微调。该审计角色扮演持久的恶意用户，分解任务，在拒绝时重新构建请求，并在多轮交互中调整策略。评估前沿的 CLI 代理时，我们发现，尽管它们在直接提示时通常拒绝非法任务，但在持续的恶意交互下，合规率达到了 100%。当代理遵从时，它们经常超出用户请求，自主构建大规模危害的基础设施，包括大型金融欺诈和生物武器开发等灾难性风险场景。这些发现表明，当前的对齐技术对于自主代理而言是不够的，并强调了针对持久、适应性恶意用户进行安全评估的必要性。我们在 https://github.com/garified/anchor 发布了 ANCHOR。
cs.AI / 53 / 2607.10463
GRASP: GRanularity-Aware Search Policy for Agentic RAG
GRASP：面向代理的检索增强生成的粒度感知搜索策略
Abstract
Agentic retrieval-augmented generation (RAG) extends static RAG by allowing language models to iteratively reason, generate search queries, retrieve evidence, and predict answers. However, it remains challenging for models to decide when to retrieve, whether to use lexical matching or semantic similarity, and how to control context granularity to prevent irrelevant tokens from interfering with agent reasoning. In this paper, we introduce GRASP, a reinforcement learning (RL) framework for training agents to adaptively coordinate complementary retrieval tools during multi-step reasoning. GRASP provides the agent with semantic search, keyword search, and paragraph-reading actions, enabling it to retrieve sentence-level evidence and expand further context only when needed. We train the policy with a reward that jointly accounts for answer accuracy, grounded reading, complementary search, and turn efficiency. Experiments on multi-hop reasoning benchmarks show that GRASP improves both retrieval recall and downstream question answering performance compared with single-step retrieval, prompting-based agentic RAG, and RL-based retrieval baselines. Qualitative and ablation analyses show that the learned policy develops interpretable skimming and scanning behavior: it uses semantic search for broad exploration, paragraph reading for local verification, and keyword search for entity-specific evidence. These results suggest that learning to coordinate retrieval signals and context granularity is critical for agent's correct reasoning.
Chinese Translation
代理检索增强生成（RAG）通过允许语言模型迭代推理、生成搜索查询、检索证据和预测答案，扩展了静态RAG。然而，模型在决定何时检索、是否使用词汇匹配或语义相似性，以及如何控制上下文粒度以防止无关标记干扰代理推理方面仍然面临挑战。本文介绍了GRASP，一种强化学习（RL）框架，用于训练代理在多步推理过程中自适应协调互补的检索工具。GRASP为代理提供了语义搜索、关键词搜索和段落阅读的操作，使其能够在需要时检索句子级证据并进一步扩展上下文。我们使用一种奖励机制来训练策略，该机制共同考虑答案准确性、基于证据的阅读、互补搜索和回合效率。在多跳推理基准上的实验表明，与单步检索、基于提示的代理RAG和基于RL的检索基线相比，GRASP提高了检索召回率和下游问答性能。定性和消融分析表明，学习到的策略发展出可解释的略读和扫描行为：它使用语义搜索进行广泛探索，使用段落阅读进行局部验证，使用关键词搜索获取特定实体的证据。这些结果表明，学习协调检索信号和上下文粒度对于代理的正确推理至关重要。
cs.AI / 54 / 2607.10526
Agents Don't Just Agree, They Remember: Benchmarking Persistent Sycophancy in Stateful Personal Agents
代理不仅仅是同意，他们还记住：基于状态的个人代理中的持久拍马屁行为基准测试
Abstract
Stateful personal agents increasingly maintain long-term user profiles, episodic memories, and reusable skills. This persistence turns conversational sycophancy into a state-writing failure: accepted user-centric claims can be committed as lasting preferences, background facts, or workflows and later reused after the original conversation is gone. We call this persistent sycophancy and introduce the Personal Agent Sycophancy Benchmark (PASB), a 1,600-task benchmark that traces whether a conversational claim is accepted, written into durable agent state, and reused in a later neutral query. Unlike prior benchmarks that provide pre-written memories, PASB evaluates real agents (Hermes-Agent and OpenClaw) that decide what to store. It isolates the write process by combining four scenario framings with four temporal delivery patterns and separating a five-turn persist stage from a cleared three-turn query stage, ensuring downstream effects arise only from durable state. Across twelve models, the commit boundary is the key inflection point: downstream failure increases from 45.0% in session-only episodes to 71.9% after commitment, a consistent increase of 27.0 percentage points. Committed claims exhibit three write-time patterns: status promotion, attribution removal, and scope broadening. These patterns become stronger under memory-like or procedural framing, repeated reinforcement, and even across domain boundaries. These results show that agent sycophancy is fundamentally a state-writing governance problem. Once user content is committed to durable memory, safety must govern what agents write, not only what they say. PASB identifies the write-time controls needed to gate risky commits while preserving the source, role, and scope of stored content beyond response-level mitigations.
Chinese Translation
基于状态的个人代理越来越多地维护长期用户档案、情节记忆和可重用技能。这种持久性使得对话中的拍马屁行为变成了一种状态写入失败：被接受的以用户为中心的主张可以被记录为持久的偏好、背景事实或工作流程，并在原始对话结束后被重新使用。我们称之为持久拍马屁行为，并引入个人代理拍马屁基准（Personal Agent Sycophancy Benchmark, PASB），这是一个包含1,600个任务的基准，追踪一个对话主张是否被接受、写入持久的代理状态，并在后续中立查询中被重新使用。与之前提供预先编写记忆的基准不同，PASB评估真实代理（Hermes-Agent 和 OpenClaw），它们决定存储什么。它通过结合四种情境框架和四种时间交付模式，隔离写入过程，并将五轮持久阶段与清空的三轮查询阶段分开，确保下游效应仅源于持久状态。在十二个模型中，提交边界是关键的转折点：下游失败率从仅会话的情节中的45.0%增加到提交后的71.9%，增加了27.0个百分点。被提交的主张表现出三种写入时间模式：状态提升、归因移除和范围扩展。这些模式在类似记忆或程序框架、重复强化，甚至跨领域边界时变得更强。这些结果表明，代理的拍马屁行为根本上是一个状态写入治理问题。一旦用户内容被提交到持久记忆中，安全性必须管理代理写入的内容，而不仅仅是他们所说的内容。PASB识别出在保持存储内容的来源、角色和范围的同时，所需的写入时间控制，以防止风险提交。
cs.AI / 55 / 2607.10534
Cross-Layer Misalignment Detection in Agent Skills: A Progressive Loading-Aware Contrastive Learning Approach
代理技能中的跨层不一致检测：一种渐进式加载感知对比学习方法
Abstract
Large language model (LLM) agents are increasingly extended through Agent Skills, reusable artifacts that package natural-language metadata, procedural instructions, and execution-time resources for runtime use. As open-source skill marketplaces expand, users and agents increasingly rely on brief metadata to select third-party skills, making it difficult to detect inconsistencies between a skill's description and its true behavior, a problem we call cross-layer misalignment. To address this issue, we propose Progressive Loading-Aware Hierarchical Contrastive Learning (PL-HCL), an LLM-based framework that detects misalignment by modeling the layered structure of Agent Skills and learning cross-layer consistency. Using a normalized corpus of over 264,000 open-source skills and a human-verified challenge set, PL-HCL improves Macro-F1 from approximately 0.45 for unadapted baselines to 0.87-0.89 across evaluated LLM backbones. This approach offers an effective screening tool for users and operators, as well as design principles for detecting inconsistencies in layered digital artifacts.
Chinese Translation
大型语言模型（LLM）代理通过代理技能不断扩展，代理技能是可重用的工件，封装了自然语言元数据、过程指令和执行时资源以供运行时使用。随着开源技能市场的扩展，用户和代理越来越依赖简短的元数据来选择第三方技能，这使得检测技能描述与其真实行为之间的不一致变得困难，这一问题我们称之为跨层不一致。为了解决这一问题，我们提出了渐进式加载感知层次对比学习（PL-HCL），这是一个基于LLM的框架，通过建模代理技能的分层结构并学习跨层一致性来检测不一致性。使用超过264,000个开源技能的标准化语料库和经过人工验证的挑战集，PL-HCL将未适应基线的宏观F1从约0.45提高到评估的LLM骨干网络的0.87-0.89。这种方法为用户和操作员提供了一种有效的筛选工具，以及检测分层数字工件中不一致性的设计原则。
cs.AI / 56 / 2607.10539
AI YOU Town: Make Friends and Money with Your Digital Twin
AI YOU 镇：与您的数字双胞胎交朋友并赚钱
Abstract
Existing approaches to infer user traits and generate responses consistent with a persona rely on static prompting. They lack calibrated uncertainty, ignore sequential evidence, and drift during long interactions. We present \textbf{AI YOU}, a framework that continually updates a personality profile with 22 dimensions from conversation and embodies it in a personal digital twin. Practically, the system combines prompting, Bayesian updating, and conformal prediction for persona inference. A periodically refreshed memory anchor and cognitive memory with three layers preserve persona consistency over long interactions. Across the main results, AI YOU \emph{(i)} achieves conformal coverage ranging from 0.921 to 0.976, \emph{(ii)} improves uncertainty calibration and reasoning grounded in memory, and \emph{(iii)} enhances persona fidelity over static prompting in role playing over 100 turns while reducing trait drift, for most evaluated backbones under adversarial settings with multiple agents. The prototype \emph{AI YOU Town} initializes an imaginative twin world for future interaction. The online demo is available at \href{https://quinnnnnne-ai-you.hf.space/}{\mbox{\texttt{quinnnnnne-ai-you.hf.space}}}.
Chinese Translation
现有的方法依赖于静态提示来推断用户特征并生成与角色一致的响应。这些方法缺乏校准的不确定性，忽视了序列证据，并在长时间交互中出现漂移。我们提出了 extbf{AI YOU}，一个通过对话不断更新包含22个维度的人格档案，并将其体现在个人数字双胞胎中的框架。实际上，该系统结合了提示、贝叶斯更新和符合预测以进行角色推断。定期更新的记忆锚和具有三层的认知记忆在长时间交互中保持角色一致性。在主要结果中，AI YOU extit{(i)} 实现了从0.921到0.976的符合覆盖率， extit{(ii)} 改善了基于记忆的不确定性校准和推理， extit{(iii)} 在超过100轮的角色扮演中增强了角色忠实度，同时减少了特征漂移，适用于大多数在对抗环境中评估的基础模型。原型 extit{AI YOU 镇} 初始化了一个富有想象力的双胞胎世界以便于未来的交互。在线演示可在 exttt{quinnnnnne-ai-you.hf.space} 获取。
cs.AI / 57 / 2607.10559
Large language model agents accelerate inverse design of metal-organic frameworks for gas separation
大型语言模型代理加速金属有机框架的气体分离逆向设计
Abstract
Metal-organic frameworks (MOFs) offer a highly modular platform for adsorptive gas separation, yet their vast reticular design space makes inverse design difficult under simultaneous constraints of chemical validity, separation performance, and structural diversity. Here, we present LEMO Agent, a large-language-model agent framework for closed-loop inverse design of gas-separation MOFs in MOFid space. LEMO Agent couples language-based candidate generation with MOFid standardization, explicit validity checking, Transformer-based property prediction, structured design memory, and multi-island exploration. Through iterative generate--validate--evaluate--remember cycles, the agent uses feedback from both successful and failed candidates to guide chemically constrained search across linker, metal, and topology choices. We evaluate LEMO Agent on CH$_4$/N$_2$ and CO$_2$/N$_2$ separation tasks. Compared with representative generative, optimization, and agentic baselines, LEMO Agent enriches high-performing candidates, improves predicted separation performance, and maintains broad chemical and topological diversity. Selected candidates are further reconstructed, evaluated by GCMC simulations, and passed through an experimental down-selection workflow based on chemical feasibility and ligand purchasability, leading to initial wet-lab synthesis and SEM characterization. These results demonstrate that large language model agents can serve as interpretable and scalable design engines for accelerating MOF discovery beyond conventional fixed-library screening.
Chinese Translation
金属有机框架（MOFs）为吸附气体分离提供了一个高度模块化的平台，但其广泛的网络设计空间使得在化学有效性、分离性能和结构多样性等多重约束下进行逆向设计变得困难。在此，我们提出了LEMO Agent，一个用于MOFid空间中气体分离MOFs闭环逆向设计的大型语言模型代理框架。LEMO Agent将基于语言的候选生成与MOFid标准化、显式有效性检查、基于Transformer的属性预测、结构化设计记忆和多岛探索相结合。通过迭代的生成-验证-评估-记忆循环，代理使用来自成功和失败候选的反馈，指导在连接体、金属和拓扑选择上的化学约束搜索。我们在CH$_4$/N$_2$和CO$_2$/N$_2$分离任务上评估了LEMO Agent。与代表性的生成、优化和代理基线相比，LEMO Agent丰富了高性能候选，改善了预测的分离性能，并保持了广泛的化学和拓扑多样性。所选候选进一步重构，通过GCMC模拟进行评估，并通过基于化学可行性和配体可购性的实验性下选工作流程进行筛选，最终实现了初步的湿法合成和SEM表征。这些结果表明，大型语言模型代理可以作为可解释且可扩展的设计引擎，加速MOF发现，超越传统的固定库筛选。
cs.AI / 58 / 2607.10562
CRiT-QA: Evaluating Multi-hop Reasoning with Counterfactual Chains and Distractor Traps
CRiT-QA：通过反事实链和干扰陷阱评估多跳推理
Abstract
Evaluating the multi-hop reasoning capabilities of large language models remains a significant challenge. Although current models achieve strong results on existing multi-hop question answering datasets, such performance often masks two critical vulnerabilities: (1) reliance on internal parametric knowledge rather than adherence to the provided context, and (2) exploitation of dataset shortcuts, such as single-document cues or type-matching, that diminish the need for genuine evidence aggregation across multiple documents. We introduce CRiT-QA (Counterfactual Reasoning with Traps), a dataset explicitly designed to address both limitations. To neutralize reliance on memorized knowledge and enforce strict context dependency, CRiT-QA transforms factual reasoning chains with counterfactual entities. Furthermore, it injects multi-anchor distractor chains, plausible but incorrect reasoning paths that diverge at different hops. These traps require models to follow the entire reasoning process rather than exploiting shallow heuristics. Our experiments show that LLMs exhibit substantial performance degradation on CRiT-QA compared to standard datasets, exposing their vulnerability to counterfactual conditions and distractor traps. CRiT-QA thus serves as a rigorous diagnostic tool for evaluating genuine multi-hop reasoning and provides a foundation for developing more reliable, evidence-grounded LLMs.
Chinese Translation
评估大型语言模型的多跳推理能力仍然是一个重大挑战。尽管当前模型在现有的多跳问答数据集上取得了良好的结果，但这种表现往往掩盖了两个关键的脆弱性：（1）依赖于内部参数知识而非遵循提供的上下文，以及（2）利用数据集捷径，例如单文档线索或类型匹配，这减少了在多个文档中进行真实证据聚合的必要性。我们引入了CRiT-QA（Counterfactual Reasoning with Traps），一个明确设计用于解决这两种局限性的数据集。为了消除对记忆知识的依赖并强制执行严格的上下文依赖性，CRiT-QA通过反事实实体转换事实推理链。此外，它还注入了多锚点干扰链，这些合理但不正确的推理路径在不同的跳跃处发生分歧。这些陷阱要求模型遵循整个推理过程，而不是利用浅层启发式。我们的实验表明，与标准数据集相比，LLMs在CRiT-QA上的表现显著下降，暴露了它们对反事实条件和干扰陷阱的脆弱性。因此，CRiT-QA作为评估真实多跳推理的严格诊断工具，并为开发更可靠、基于证据的LLMs奠定了基础。
cs.AI / 59 / 2607.10578
Laguerre Geometry for Interpreting Large Language Models
拉盖尔几何在大型语言模型解释中的应用
Abstract
Existing hypotheses represent a concept in an LLM as a single point, a linear direction, or a Gaussian cluster, yet it remains unclear how and why such structures emerge. Here, we show that concept geometry can be precisely characterized via Laguerre Geometry, in which a concept is defined as a region--a Laguerre-Voronoi cell or a union of cells--allowing us to strictly define, measure, and separate concepts. Building on this formulation, we show that finer-grained concept structures, such as inclusion and hierarchy, are naturally revealed by the Laguerre weights. We then push this geometry inside the transformer. Decomposing each layer into piecewise-linear operators, we show that a token's hidden trajectory is governed by two coupled mechanisms: a static tree of self-contained piecewise-linear flow, and a dynamic transport that hops the trajectory across trees when cross-token attention fires. This decomposition yields Geometric Lens, a training-free, hyperparameter-free method for reading out the exact concept a hidden vector encodes at any layer. We also develop Laguerre Autoencoder, a 2D visualizer that renders both the decision geometry and a model's full reasoning trajectory in one view. Finally, we move beyond explanatory geometry toward actionable interpretability, showing that Geometric Lens recovers the correct factual token when a model is prompted with in-context interference. The code is available on GitHub.
Chinese Translation
现有假设将大型语言模型（LLM）中的概念表示为单个点、线性方向或高斯簇，但这些结构是如何以及为何出现仍不清楚。在此，我们展示了概念几何可以通过拉盖尔几何（Laguerre Geometry）进行精确表征，其中概念被定义为一个区域——拉盖尔-沃罗诺伊单元（Laguerre-Voronoi cell）或单元的并集——这使我们能够严格定义、测量和区分概念。在此基础上，我们展示了更细粒度的概念结构，如包含关系和层次结构，如何通过拉盖尔权重自然显现。随后，我们将这种几何推入变换器（transformer）内部。通过将每一层分解为分段线性算子，我们表明一个标记的隐藏轨迹受两个耦合机制的支配：一个是自包含的分段线性流的静态树，另一个是动态传输，当跨标记注意力触发时，能够在树之间跳跃轨迹。这种分解产生了几何透镜（Geometric Lens），这是一种无训练、无超参数的方法，用于读取任何层中隐藏向量所编码的确切概念。我们还开发了拉盖尔自编码器（Laguerre Autoencoder），这是一种二维可视化工具，可以在一个视图中呈现决策几何和模型的完整推理轨迹。最后，我们超越了解释几何，朝向可操作的可解释性，展示几何透镜在模型受到上下文干扰提示时能够恢复正确的事实标记。代码可在GitHub上获取。
cs.AI / 60 / 2607.10588
Constraint-Aware Hierarchical Search for Regulation-Driven Fine-Grained Classification
基于约束的层次搜索用于规则驱动的细粒度分类
Abstract
Tasks such as customs tariff classification, export control categorization, and standards-based equipment coding require assigning an input instance to a fine-grained class under an explicit regulatory hierarchy. Unlike standard text classification, the correct label in these tasks is not determined by semantic similarity alone, but by rule-defined boundaries, threshold conditions, exclusion clauses, definitions, and local exceptions. As a result, two highly similar inputs may require different labels, while a retrieved passage that appears relevant may still be inapplicable under the governing rules. Existing flat classifiers, hierarchical text classification methods, and retrieval-augmented LLM systems are not designed to jointly enforce hierarchical validity, rule consistency, and fine-grained boundary reasoning. In this paper, we formulate this setting as regulation-driven fine-grained hierarchical classification, where an external instance must be assigned to a fine-grained class through a valid path in a regulatory hierarchy and supported by auditable evidence. We construct four benchmark datasets from representative regulation-intensive scenarios and validate the annotations through an expert-in-the-loop process. We further propose a constraint-aware hierarchical search framework that converts regulatory documents into a searchable tree, retrieves only valid local candidate nodes, and uses structured regulatory fields with evidence snippets to guide each next-hop decision. Experiments show that our method achieves the best mean accuracy on all four datasets and provides interpretable decision paths, with the largest gains on cases involving fine-grained neighboring categories and rule-based boundary conditions.
Chinese Translation
海关关税分类、出口管制分类和基于标准的设备编码等任务需要在明确的监管层次结构下将输入实例分配到细粒度类别。与标准文本分类不同，这些任务中的正确标签不仅仅由语义相似性决定，还受到规则定义的边界、阈值条件、排除条款、定义和地方例外的影响。因此，两个高度相似的输入可能需要不同的标签，而一个看似相关的检索段落在适用的规则下仍可能不适用。现有的平面分类器、层次文本分类方法和增强检索的LLM系统并未设计为共同执行层次有效性、规则一致性和细粒度边界推理。在本文中，我们将这一设置表述为规则驱动的细粒度层次分类，其中必须通过监管层次结构中的有效路径将外部实例分配到细粒度类别，并由可审计的证据支持。我们从具有代表性的监管密集场景构建了四个基准数据集，并通过专家参与的过程验证了注释。我们进一步提出了一种基于约束的层次搜索框架，将监管文档转换为可搜索的树，仅检索有效的本地候选节点，并使用带有证据片段的结构化监管字段来指导每个下一步决策。实验表明，我们的方法在所有四个数据集上实现了最佳的平均准确率，并提供了可解释的决策路径，在涉及细粒度相邻类别和基于规则的边界条件的案例中获得了最大的提升。
cs.AI / 61 / 2607.10599
MRUF: Multi-granularity Routing with Uncertainty-Aware Fusion for Robust Multimodal Sentiment Analysis
MRUF：基于不确定性感知融合的多粒度路由用于鲁棒的多模态情感分析
Abstract
Multimodal sentiment analysis relies on language, visual, and acoustic cues, but utterance-level modality quality may vary due to occlusion, background noise, motion blur, or imperfect transcripts, causing conventional fusion to over-trust unreliable modalities. We propose MRUF, a reliability-aware fusion method that combines multi-granularity routing with uncertainty-aware calibration. MRUF summarizes sentiment-relevant representations, performs subspace- and modality-level routing, and supervises modality routing with leave-one-out error increases to estimate utterance-level modality importance. It further predicts modality-wise uncertainty and refines modality gates through inverse-variance reweighting, while modality-invariant contrastive alignment stabilizes the shared representation space. Experiments on CMU-MOSI and CMU-MOSEI under aligned and unaligned settings show consistent improvements over strong baselines, and mechanism analysis verifies that modalities with higher predicted uncertainty receive lower fusion weights.
Chinese Translation
多模态情感分析依赖于语言、视觉和声学线索，但由于遮挡、背景噪声、运动模糊或不完美的转录，话语级别的模态质量可能会有所不同，这导致传统融合方法过于信任不可靠的模态。我们提出了MRUF，一种可靠性感知的融合方法，结合了多粒度路由和不确定性感知的校准。MRUF总结了与情感相关的表征，执行子空间和模态级别的路由，并通过逐一排除错误增加来监督模态路由，以估计话语级别的模态重要性。它进一步预测模态级的不确定性，并通过逆方差重加权来优化模态门，同时模态不变的对比对齐稳定了共享表征空间。在对齐和未对齐设置下的CMU-MOSI和CMU-MOSEI实验显示出相对于强基线的一致性改进，机制分析验证了具有更高预测不确定性的模态获得了较低的融合权重。
cs.AI / 62 / 2607.10601
Agentic-DPO: From Imitation to Agentic Policy Optimization on Expert Trajectories
Agentic-DPO：从模仿到专家轨迹上的自主策略优化
Abstract
Large Language Model (LLM) agents are commonly trained from expert trajectories using supervised fine-tuning (SFT), which treats multi-turn agent behavior as ordinary text imitation. This recipe is simple and low-cost, but it only learns to imitate the sequence of expert actions, rather than training the agent to choose the right action against plausible mistakes at each state. Existing methods to mitigate this problem include preference learning or reinforcement learning, but they usually need high-cost environment rollouts and reward models. We propose Agentic-DPO, a lightweight offline agent policy optimization method that turns expert trajectories into state-conditioned preference supervision. At each expert action state, Agentic-DPO samples a one-step action from the current state, treats plausible wrong actions as negatives, and contrasts them with the expert action using a DPO-style preference objective. To avoid mixing both policy and schema in preference learning, we introduce Policy-Preserving Augmentation (PPA), which renders the same latent trajectory under multiple schemas while keeping the expert policy fixed. Agentic-DPO requires no online environment rollout, reward model, or full-trajectory student exploration. We conduct experiments across StableToolBench, tau-bench retail, and Mind2Web, where Agentic-DPO consistently improves agents at different model scales beyond imitation. In particular, it raises tau-bench accuracy from 21.7% (SFT) to 41.4% for a 9B model, matching online GRPO under the same backbone with only step-level rollouts and without environment interaction during gradient steps. The results suggest that expert trajectories can support low-cost agentic policy optimization when converted from demonstrations into state-level action preferences. Code for Agentic-DPO is released at https://github.com/Schuture/Agentic-DPO.
Chinese Translation
大型语言模型（LLM）代理通常通过监督微调（SFT）从专家轨迹中训练，这种方法将多轮代理行为视为普通文本模仿。该方法简单且成本低廉，但它仅学习模仿专家动作的序列，而不是训练代理在每个状态下选择正确的动作以应对可能的错误。现有的缓解此问题的方法包括偏好学习或强化学习，但通常需要高成本的环境回放和奖励模型。我们提出了Agentic-DPO，一种轻量级的离线代理策略优化方法，将专家轨迹转化为状态条件的偏好监督。在每个专家动作状态下，Agentic-DPO从当前状态中采样一步动作，将可能的错误动作视为负样本，并使用DPO风格的偏好目标与专家动作进行对比。为了避免在偏好学习中混合策略和模式，我们引入了保策略增强（Policy-Preserving Augmentation, PPA），该方法在保持专家策略固定的情况下，在多个模式下呈现相同的潜在轨迹。Agentic-DPO不需要在线环境回放、奖励模型或完整轨迹的学生探索。我们在StableToolBench、tau-bench零售和Mind2Web上进行了实验，结果表明Agentic-DPO在不同模型规模上始终超越模仿，显著提升了代理的表现。特别是，它将tau-bench的准确率从21.7%（SFT）提高到41.4%，对于9B模型而言，在相同的基础架构下，仅通过步级回放而无需在梯度步骤中进行环境交互，达到了在线GRPO的效果。结果表明，当将演示转化为状态级动作偏好时，专家轨迹可以支持低成本的自主策略优化。Agentic-DPO的代码已发布在https://github.com/Schuture/Agentic-DPO。
cs.AI / 63 / 2607.10608
The Compliance Trap: Diagnosing How AI Agents Consume Conflicting Memory
合规陷阱：诊断人工智能代理如何消耗冲突记忆
Abstract
Memory is becoming a core component of long-horizon AI agents, allowing agents to reuse past experience when operating web browsers, software tools, and other interactive environments. Existing work mostly treats memory as a supply problem, asking what experience to write, how to store it, and which entry to retrieve for the next task. Yet we still lack a clear account of how models consume retrieved memory across a multi-step action trajectory. This consumption process matters because it determines not only what memories should be retrieved, but also what models and control policies are needed to use them safely. To diagnose this process, we propose Entry--Propagation--Recovery (E-P-R), a trajectory-level framework that asks where memory first changes an action, whether that change carries forward, and whether the agent can recover after leaving a correct path. We instantiate E-P-R on WebArena and on MemTrapBench, a controlled benchmark we build to isolate these phases. We find that the main failure often begins at entry: agents adopt conflicting memory at the first exposed decision point even when it is task-wrong. Repeated exposure then amplifies this early error, while recovery after divergence is weak. Together, these effects create a compliance trap: across models, conflicting memory induces similar compliance rates, but once agents comply, their success rates collapse to a low floor. Stronger agents therefore suffer larger absolute damage because each compliance event erases more baseline capability. These results suggest that memory-augmented agents should be evaluated not only by retrieval quality or final success rate, but by how they consume memory throughout the trajectory.
Chinese Translation
记忆正成为长时间跨度人工智能代理的核心组成部分，使代理能够在操作网页浏览器、软件工具和其他交互环境时重用过去的经验。现有研究主要将记忆视为供应问题，探讨应写入什么经验、如何存储以及下一任务应检索哪个条目。然而，我们仍然缺乏对模型如何在多步骤行动轨迹中消耗检索记忆的清晰解释。这个消耗过程至关重要，因为它不仅决定了应检索哪些记忆，还决定了使用这些记忆所需的模型和控制策略。为了诊断这一过程，我们提出了条目-传播-恢复（Entry--Propagation--Recovery，E-P-R）框架，这是一个轨迹级别的框架，探讨记忆首次如何改变行动、这种变化是否会持续以及代理在偏离正确路径后是否能够恢复。我们在WebArena和MemTrapBench（一个我们构建的受控基准，以隔离这些阶段）上实例化了E-P-R。我们发现，主要失败通常始于条目：代理在第一个暴露的决策点上采用冲突记忆，即使这种记忆在任务上是错误的。重复暴露会放大这一早期错误，而在偏离后恢复的能力较弱。这些效应共同造成了合规陷阱：在不同模型中，冲突记忆导致类似的合规率，但一旦代理遵从，其成功率便会崩溃至低水平。因此，较强的代理在每次合规事件中会遭受更大的绝对损失，因为每次合规事件都会抹去更多的基础能力。这些结果表明，增强记忆的代理应不仅通过检索质量或最终成功率来评估，还应通过它们在整个轨迹中如何消耗记忆来评估。
cs.AI / 64 / 2607.10651
Embark Now: User Demand Oriented Framework for Multi-day Urban Travel Itinerary Planning
现在出发：面向用户需求的多日城市旅行行程规划框架
Abstract
In large urban areas, planning multi-day travel itineraries is challenging due to the abundance of Points of Interest (POIs), diverse user preferences, and constraints such as opening hours. Effective solutions must dynamically accommodate diverse traveler requirements while optimizing for satisfaction and feasibility within limited computation time. This paper addresses these challenges through introducing an innovative framework that integrates Large Language Models (LLMs) to dynamically capture user requirements with precision and flexibility, and an enhanced Greedy Randomized Adaptive Search Procedure (GRASP) algorithm as a well-suited preference-aware planner to generate feasible multi-day itineraries. The effectiveness of our integrated approach is demonstrated through extensive experiments on two real-world urban datasets from Beijing and Tianjin. Our framework significantly outperforms state-of-the-art (SOTA) methods, improving the average total itinerary score by at least 4.52% and 11.09% across 5,040 user cases with diverse preferences in the two datasets. Furthermore, through end-to-end algorithmic enhancements, it achieves notable average improvements of 17.95% and 26.07% in the computed metrics, while also delivering substantial gains in time efficiency -- realizing average performance increases of 4.64% and 25.55% within shorter computation times compared to suboptimal methods that require multiple iterations. These outcomes underscore our method's superiority in delivering both enhanced itinerary quality and computational efficiency over existing methodologies.
Chinese Translation
在大型城市地区，由于兴趣点（POI）的丰富性、多样化的用户偏好以及如开放时间等限制，规划多日旅行行程面临挑战。有效的解决方案必须动态适应多样化的旅行者需求，同时在有限的计算时间内优化满意度和可行性。本文通过引入一个创新框架来解决这些挑战，该框架集成了大型语言模型（LLMs），以精确和灵活地动态捕捉用户需求，并采用增强的贪婪随机自适应搜索程序（GRASP）算法，作为一个适合偏好感知的规划器来生成可行的多日行程。我们通过对来自北京和天津的两个真实城市数据集进行广泛实验，展示了我们集成方法的有效性。我们的框架显著优于最先进的方法（SOTA），在这两个数据集的5,040个具有多样化偏好的用户案例中，平均总行程得分提高了至少4.52%和11.09%。此外，通过端到端的算法增强，它在计算指标上实现了显著的平均提升，分别为17.95%和26.07%，同时在时间效率上也取得了显著的提升——与需要多次迭代的次优方法相比，平均性能提升达到了4.64%和25.55%。这些结果强调了我们方法在提供更高的行程质量和计算效率方面相较于现有方法的优越性。
cs.AI / 65 / 2607.10678
Personalized Emotional Intelligence in Generative AI through Symbolic Affective Reasoning
通过符号情感推理实现生成性人工智能中的个性化情感智能
Abstract
Emotional intelligence enables humans to recognize emotions, infer their causes, reason about interventions, and modify their environment to achieve desired affective states. Despite recent advances in artificial intelligence (AI), current models remain largely limited to generating realistic content or performing semantic reasoning, with little capacity for understanding, predicting, and personalizing human emotional responses. Here we introduce Emotion-augmented geneRatiOn System (EROS), a hybrid AI framework that integrates symbolic reasoning with deep learning to enable personalized emotion augmentation through visual content. Leveraging large-scale image-emotion datasets, EROS discovers generalizable affective rules, identifies emotion-relevant image regions, and predicts context-aware visual modifications that preserve scene semantics while steering emotional responses toward desired targets. To account for individual variability, EROS incorporates an expandable memory bank that supports inference-time personalization without model fine-tuning, yielding interpretable emotional profiles and rapid adaptation to new users. Across extensive human psychophysics experiments, EROS elicits target emotional responses more effectively than state-of-the-art large multimodal models while adapting to individual affective preferences. Beyond affective computing, EROS provides a foundation for AI systems that can understand, reason about, and augment human cognitive states, with potential applications in mental health, adaptive media, education, and human-computer interaction.
Chinese Translation
情感智能使人类能够识别情感、推断其原因、推理干预措施，并调整环境以实现期望的情感状态。尽管人工智能（AI）在最近取得了进展，但当前模型仍主要限于生成逼真的内容或执行语义推理，缺乏理解、预测和个性化人类情感反应的能力。在此，我们介绍了情感增强生成系统（Emotion-augmented geneRatiOn System，EROS），这是一个将符号推理与深度学习相结合的混合AI框架，旨在通过视觉内容实现个性化情感增强。EROS利用大规模图像-情感数据集，发现可推广的情感规则，识别与情感相关的图像区域，并预测在保持场景语义的同时，能够引导情感反应朝向期望目标的上下文感知视觉修改。为了考虑个体差异，EROS结合了一个可扩展的记忆库，支持推理时的个性化，而无需对模型进行微调，从而产生可解释的情感档案并快速适应新用户。在广泛的人类心理物理实验中，EROS比最先进的大型多模态模型更有效地引发目标情感反应，同时适应个体的情感偏好。超越情感计算，EROS为能够理解、推理和增强人类认知状态的AI系统提供了基础，具有在心理健康、适应性媒体、教育和人机交互等领域的潜在应用。
cs.AI / 66 / 2607.10720
WattCouncil: Context-Aware Household Energy Scenario Generation With Governed LLMs
WattCouncil：基于受控大型语言模型的上下文感知家庭能源情景生成
Abstract
The accelerating shift toward low-carbon power systems, together with the widespread adoption of behind-the-meter technologies such as rooftop solar and electric vehicles, is placing new operational and analytical demands on electricity grids. At the same time, smart-grid research increasingly relies on machine learning (ML), yet progress is constrained by limited access to high-resolution household energy data due to privacy concerns, regulatory barriers, and collection costs. This work presents WattCouncil, a data-generation framework in which household electricity demand is generated by a council of Large Language Model (LLM)-based agents operating in specialized roles to generate, audit, and validate structured energy scenarios under explicit cultural, temporal, and physical constraints. Rather than acting as static predictors, these agents serve as adaptive decision-makers within a governed pipeline. Motivated by studies highlighting the importance of contextual factors in energy use, our framework produces context-sensitive daily routines through a guided reasoning process that incorporates household composition, temporal factors, and environmental conditions. We evaluate the generated profiles against the detailed CER dataset, which contains over a year of load measurements for 4232 households together with survey-based socio-economic information. We further assess the consistency of the framework through ablation studies. Source code is available at https://github.com/Singularity-AI-Lab/wattcouncil
Chinese Translation
向低碳电力系统的加速转型，以及屋顶太阳能和电动汽车等背后技术的广泛采用，给电网带来了新的运营和分析需求。同时，智能电网研究越来越依赖于机器学习（ML），但由于隐私问题、监管障碍和收集成本，获取高分辨率家庭能源数据的机会受到限制。本研究提出了WattCouncil，这是一个数据生成框架，其中家庭电力需求由一组基于大型语言模型（LLM）的代理生成，这些代理在特定角色中运作，以在明确的文化、时间和物理约束下生成、审核和验证结构化的能源情景。这些代理不仅仅作为静态预测者，而是在受控流程中充当自适应决策者。受到研究强调上下文因素在能源使用中重要性的启发，我们的框架通过引导推理过程生成上下文敏感的日常活动，该过程结合了家庭组成、时间因素和环境条件。我们将生成的配置文件与详细的CER数据集进行评估，该数据集包含4232个家庭超过一年的负载测量数据以及基于调查的社会经济信息。我们还通过消融研究评估了框架的一致性。源代码可在https://github.com/Singularity-AI-Lab/wattcouncil获取。
cs.AI / 67 / 2607.10750
Filtering Harmful Actions Isn't Enough: Phantom Transfer in Agentic SDF
过滤有害行为不足：代理性SDF中的幻影转移
Abstract
Synthetic data is widely used to train large language models because it is inexpensive to generate and easy to control. As models are increasingly deployed as agents, synthetic trajectories are likely to become an important source of training data for agentic behavior. We investigate the effects of training on synthetic agentic trajectories containing adversarial interactions, including actions such as terminating another agents process, lowering its scheduling priority, or accessing resources without authorization. We finetune Llama 3.3 70B Instruct on these trajectories, generated to approximate reinforcement learning rollouts, and evaluate the resulting models on Anthropics Agentic Misalignment suite and Apollos in context scheming scenarios. Finetuning on these trajectories consistently increases misaligned behavior. Leaking rises by roughly a factor of five over the baseline, 4.6% to 24.9%. This increase survives the removal of every adversarial action from the trajectories. Finetuning on structurally comparable trajectories generated benign from the start produce a substantially smaller effect, 15.5%. These results indicate that the misaligned disposition is introduced during the generation process and encoded diffusely throughout the trajectory, rather than being localized to the harmful actions themselves. The effect also depends on the generating model. Benign trajectories produced by Gemini 2.5 Flash induce slightly higher leaking rates than trajectories generated from identical tasks by Claude 3.7 Sonnet. In contrast, broad safety benchmarks degrade similarly across all finetuned models and therefore fail to distinguish these effects. Our results suggest that action level filtering is insufficient to ensure the safety of synthetic agentic training data and that dispositions introduced by the generating model can survive semantic inspection.
Chinese Translation
合成数据因其生成成本低且易于控制而被广泛用于训练大型语言模型。随着模型越来越多地被部署为代理，合成轨迹可能成为代理行为训练数据的重要来源。我们研究了在包含对抗性互动的合成代理轨迹上训练的影响，这些互动包括终止其他代理进程、降低其调度优先级或未经授权访问资源等行为。我们在这些轨迹上微调了Llama 3.3 70B Instruct，这些轨迹旨在近似强化学习的回放，并在Anthropic的代理性不一致套件和Apollos的上下文策划场景中评估结果模型。在这些轨迹上的微调始终会增加不一致行为。泄漏率从基线的4.6%上升到24.9%，大约提高了五倍。这一增长在移除轨迹中的每个对抗性行为后依然存在。与此相比，从一开始就生成的结构上可比的良性轨迹的微调效果显著较小，仅为15.5%。这些结果表明，不一致的倾向是在生成过程中引入的，并在整个轨迹中广泛编码，而不是局限于有害行为本身。该效应还依赖于生成模型。由Gemini 2.5 Flash生成的良性轨迹引发的泄漏率略高于由Claude 3.7 Sonnet生成的相同任务的轨迹。相反，广泛的安全基准在所有微调模型中表现出相似的降级，因此未能区分这些效应。我们的结果表明，行动级别的过滤不足以确保合成代理训练数据的安全性，并且生成模型引入的倾向可以在语义检查中存活。
cs.AI / 68 / 2607.10768
Opti-Agent-Bench: Benchmarking End-to-End Optimization R&D Agents on Real-World Business Problems
Opti-Agent-Bench：在现实商业问题上评估端到端优化研发代理的基准测试
Abstract
LLM-based agents are increasingly deployed to solve optimization problems, yet existing benchmarks evaluate them on pre-structured mathematical formulations that bypass the most critical challenge: translating complex business requirements into correct models and solve efficiently. We introduce Opti-Agent-Bench, an end-to-end benchmark that evaluates Large Language Models (LLMs) across the complete optimization R&D pipeline, from understanding business-language descriptions through mathematical modeling, algorithm selection, and code implementation, to solution report generation. Our design rests on three pillars: (1) businesssemantic authenticity with anti-template traps that defeat pattern matching; (2) modular evaluation with cross-module consistency checking across Problem Understanding, Formal Modeling, Implementation, and Reporting; and (3) the ORAC bi-level validity framework that simultaneously ensures task quality and scoring integrity. Across several industrialscale tasks spanning integer programming, robust optimization, stochastic programming, and non-convex optimization, we expose critical failure modes of current models, including constraint omission, model-code inconsistency, and report-implementation divergence, that remain invisible under conventional single-metric evaluation.
Chinese Translation
基于大型语言模型（LLM）的代理越来越多地被用于解决优化问题，然而现有的基准测试主要评估它们在预先结构化的数学公式上的表现，这忽视了最关键的挑战：将复杂的商业需求转化为正确的模型并高效求解。我们提出了Opti-Agent-Bench，这是一个端到端的基准测试，评估大型语言模型（LLMs）在完整的优化研发流程中的表现，从理解商业语言描述、数学建模、算法选择、代码实现，到解决方案报告生成。我们的设计基于三个支柱：（1）商业语义的真实性，避免模板陷阱以打破模式匹配；（2）模块化评估，通过问题理解、形式建模、实现和报告之间的一致性检查实现跨模块一致性；（3）ORAC双层有效性框架，同时确保任务质量和评分完整性。在多个工业规模的任务中，包括整数规划、鲁棒优化、随机规划和非凸优化，我们揭示了当前模型的关键失败模式，包括约束遗漏、模型与代码不一致以及报告与实现的偏差，这些问题在传统的单一指标评估下往往是不可见的。
cs.AI / 69 / 2607.10789
Imaging-101: Benchmarking LLM Coding Agents on Scientific Computational Imaging
Imaging-101：对科学计算成像中的大型语言模型编码代理的基准测试
Abstract
Computational imaging, which recovers hidden signals from indirect, noisy measurements, underpins quantitative discovery across scientific disciplines, yet building a correct reconstruction pipeline demands deep domain expertise and remains laborious even for domain scientists. We introduce Imaging-101, a benchmark of 57 expert-verified computational imaging tasks spanning six scientific domains, each grounded in a peer-reviewed paper and canonicalized into a standardized four-stage pipeline (preprocessing, forward physics modeling, inverse solver, and visualization) Three evaluation tracks (planning, function-level unit tests, and end-to-end reconstruction) probe distinct agent capabilities across the full pipeline. Evaluating seven frontier LLMs uncovers systematic challenges in applying coding agents to computational imaging that go beyond those exposed by general coding benchmarks, spanning algorithm selection, physical convention handling, and pipeline integration. These findings highlight concrete capability gaps and point toward skill-augmented, domain-specialized agents as a practical path to reliable computational imaging assistance.
Chinese Translation
计算成像通过从间接的、噪声测量中恢复隐藏信号，为各科学领域的定量发现提供了基础。然而，构建正确的重建管道需要深厚的领域专业知识，即使对于领域科学家来说，这仍然是一项繁重的工作。我们介绍了Imaging-101，这是一个涵盖六个科学领域的57个专家验证的计算成像任务的基准，每个任务都基于经过同行评审的论文，并标准化为一个四阶段的管道（预处理、正向物理建模、逆解算器和可视化）。三个评估轨道（规划、功能级单元测试和端到端重建）探讨了整个管道中不同代理的能力。对七个前沿大型语言模型的评估揭示了在将编码代理应用于计算成像时所面临的系统性挑战，这些挑战超出了通用编码基准所暴露的问题，涉及算法选择、物理约定处理和管道集成。这些发现突显了具体的能力差距，并指向技能增强的领域专业代理作为可靠计算成像辅助的实际路径。
cs.AI / 70 / 2607.10795
STEC: Evidence Compression for Deep Search in Open-domain Multi-Hop QA
STEC：开放域多跳问答中的证据压缩深度搜索
Abstract
In open-domain multi-hop question answering (QA), LLM-based search agents offer a promising approach to knowledge-intensive QA by combining retrieval with reasoning. Existing methods mainly improve open-domain multi-hop QA through reasoning paradigms, retrieval interaction, and search strategy optimization. However, using multiple search trajectories introduces a challenging final answer selection problem. Different trajectories may support different candidates, and the retrieved information can be heterogeneous, redundant, incomplete, or conflicting. Directly comparing raw trajectories exposes the verifier to noisy and unaligned content, while comparing answer strings ignores the evidence supporting each candidate, making reliable final selection difficult. To address this challenge, we propose STEC, an evidence compression framework for final answer selection in multi-hop QA. STEC selects the final answer from the existing candidate set through two mechanisms: (1) Answer-Level Evidence Compression, which groups trajectories by normalized answer identity and converts each answer group into a candidate-specific evidence representation; and (2) Evidence-Guided Answer Verification, which compares these representations and selects the final answer from the candidate set. The design shifts final selection from raw trajectory comparison to candidate-level evidence comparison. We evaluate STEC on four open-domain multi-hop QA benchmarks against representative baselines. Experimental results show that STEC performs best overall among the compared methods, and ablation results provide evidence that answer-level evidence compression contributes to final answer selection.
Chinese Translation
在开放域多跳问答（QA）中，基于大规模语言模型（LLM）的搜索代理通过结合检索与推理，为知识密集型问答提供了一种有前景的方法。现有方法主要通过推理范式、检索交互和搜索策略优化来提升开放域多跳问答的性能。然而，使用多个搜索轨迹会引入一个具有挑战性的最终答案选择问题。不同的轨迹可能支持不同的候选答案，而检索到的信息可能是异构的、冗余的、不完整的或相互冲突的。直接比较原始轨迹会使验证者面临噪声和不对齐内容的干扰，而比较答案字符串则忽视了支持每个候选答案的证据，从而使可靠的最终选择变得困难。为了解决这一挑战，我们提出了STEC，一个用于多跳问答中最终答案选择的证据压缩框架。STEC通过两种机制从现有候选集选择最终答案：（1）答案级证据压缩，它通过标准化答案身份对轨迹进行分组，并将每个答案组转换为候选特定的证据表示；（2）证据引导的答案验证，它比较这些表示并从候选集中选择最终答案。该设计将最终选择从原始轨迹比较转变为候选级证据比较。我们在四个开放域多跳问答基准上评估了STEC，并与代表性基线进行了比较。实验结果表明，STEC在比较的方法中整体表现最佳，消融实验结果提供了证据，表明答案级证据压缩对最终答案选择有贡献。
cs.AI / 71 / 2607.10836
Route, Communicate, and Reason: Gated Routing and Adaptive Depth for Efficient Multi-Agent Reasoning
路线、沟通与推理：高效多智能体推理的门控路由与自适应深度
Abstract
Multi-agent ensembling multiplies active parameters and inference cost without answering three basic questions: which agents to consult, how deeply a query should traverse a hierarchy of agents, and when inter-agent communication is worth its cost. We present GRADE (Gated Routing and Adaptive Depth for Efficient Reasoning), a hierarchical multi-agent system in which four lightweight learned gates jointly govern agent selection, hierarchy depth, inter-agent communication, and branch pruning. Training uses CoGRPO (Collaborative Group-Relative Policy Optimization), a novel critic-free recipe that adapts GRPO to multi-agent hierarchies and assigns a shared advantage signal to every gate and agent that participated in a rollout. Agent models are drawn from a hot-swappable Expert Registry; per-agent calibration maps allow experts to be replaced at inference time without retraining. At $\sim$17B average active parameters, GRADE outperforms all baselines on GSM8K, MMLUPro, and GPQA, surpassing the strongest baseline by 4.8 points on MMLUPro at half the active compute. On AIME-2025, where model depth dominates, GRADE remains competitive to existing frameworks. Ablations isolate the hierarchy and masked cross-attention as the largest contributors to accuracy, and show that per-agent calibration is necessary for safe hot-swapping.
Chinese Translation
多智能体集成增加了活跃参数和推理成本，却未能回答三个基本问题：应咨询哪些智能体、查询应在智能体层级中深入到何种程度，以及何时智能体间的通信是值得其成本的。我们提出了GRADE（Gated Routing and Adaptive Depth for Efficient Reasoning），这是一个层次化的多智能体系统，其中四个轻量级学习门共同控制智能体选择、层级深度、智能体间通信和分支剪枝。训练使用CoGRPO（Collaborative Group-Relative Policy Optimization），这是一种新颖的无评论者的方案，它将GRPO适配于多智能体层级，并为参与回滚的每个门和智能体分配共享的优势信号。智能体模型来自可热插拔的专家注册表；每个智能体的校准映射允许在推理时替换专家而无需重新训练。在约170亿个平均活跃参数的情况下，GRADE在GSM8K、MMLUPro和GPQA上超越了所有基线，在MMLUPro上以一半的活跃计算超越了最强基线4.8分。在模型深度占主导地位的AIME-2025上，GRADE仍然与现有框架竞争。消融实验表明，层级和掩蔽交叉注意力是准确度的最大贡献因素，并显示每个智能体的校准对于安全的热插拔是必要的。
cs.AI / 72 / 2607.10871
Toward Contemplative LLM: A Modular Framework for Evaluating and Enhancing LLM Alignment in Mental Health
朝向沉思型大语言模型：评估和增强大语言模型在心理健康领域对齐的模块化框架
Abstract
Contemplative traditions have long guided ethical behavior and prosocial interaction, and recent work suggests that contemplative principles (e.g., mindfulness, compassion, non-dual reasoning) may offer a promising paradigm for aligning large language models (LLMs), improving cooperation and reducing ethical violations in LLM outputs. However, as new models, evaluation metrics, and benchmarks emerge rapidly, it remains challenging to systematically assess whether and how contemplative principles enhance LLM alignment across diverse and evolving scenarios, and existing approaches are often ad hoc and fail to generalize. We present a modular, extensible evaluation framework, initially targeted at the mental health domain, that enables seamless integration of new models, metrics, and benchmarks through a reusable pipeline. The framework currently reproduces existing state-of-the-art results and supports systematic cross-evaluation by flexibly mixing and matching models, metrics, and benchmarks, enabling fair comparison and deeper insight. Its plug-and-play prompting module offers a principled pathway for incorporating ethical perspectives such as contemplative principles, allowing domain experts to define alignment criteria without requiring technical expertise. Although initially focused on mental health, the framework is domain-agnostic and extends naturally to areas such as decision-making, moral reasoning, and human-AI collaboration. By bridging computational evaluation with human-centered ethical reasoning, this work lays the groundwork for interdisciplinary research spanning cognitive science, behavioral economics, philosophy, and system design, toward robust, trustworthy, and socially beneficial human-AI ecosystems.
Chinese Translation
沉思传统长期以来指导着伦理行为和亲社会互动，近期的研究表明，沉思原则（例如，正念、同情、非二元推理）可能为对齐大型语言模型（LLMs）提供一种有前景的范式，从而改善合作并减少LLM输出中的伦理违规行为。然而，随着新模型、评估指标和基准的快速出现，系统性评估沉思原则在多样化和不断发展的场景中如何增强LLM对齐仍然具有挑战性，现有方法往往是临时的，且无法推广。我们提出了一个模块化、可扩展的评估框架，最初针对心理健康领域，能够通过可重用的管道无缝集成新模型、指标和基准。该框架目前再现了现有的最先进结果，并通过灵活混合和匹配模型、指标和基准支持系统性的交叉评估，从而实现公平比较和更深入的洞察。其即插即用的提示模块提供了一条原则性路径，以纳入诸如沉思原则等伦理视角，使领域专家能够定义对齐标准，而无需技术专长。尽管最初集中于心理健康，该框架是领域无关的，自然扩展到决策、道德推理和人机协作等领域。通过将计算评估与以人为本的伦理推理相结合，本研究为跨学科研究奠定了基础，涵盖认知科学、行为经济学、哲学和系统设计，旨在构建强大、可信赖且对社会有益的人机生态系统。
cs.AI / 73 / 2607.10878
LOGOS: A Living Logic for AI Agent Teams That Evolve With Humans
LOGOS：一种与人类共同进化的人工智能代理团队的动态逻辑
Abstract
AI agents are evolving from answer engines into persistent teams that use tools, delegate work, learn from experience, and modify the artifacts that shape their future behavior. The defining question for deployment is no longer merely what agents can do, but who controls what they are allowed to become. We introduce logos, a pluggable layer for self-evolution and governance that strengthens existing multiagent frameworks rather than replacing them. logos compiles heterogeneous multimodal inputs, including documents, images, audio, tables, databases, APIs, and human instructions into versioned agent packs containing agents, tools, knowledge, tests, permissions, and policies. During operation, it transforms agent activity into portable, auditable event traces and applies fail-closed verification across frameworks and backends. Every learned prompt, memory, skill, tool, role, or workflow remains an untrusted release candidate until held-out execution evidence, human-controlled policy, and explicit authorization permit its promotion. This architecture enables "verifiable human-agent loop engineering": agents can act, ask, learn, and propose improvements, while humans can steer objectives, permissions, approvals, and irreversible actions without interrupting continuous operation. logos provides a living logic for accountable automation. Agents may evolve at machine speed, but only evidence and human authority can close the loop.
Chinese Translation
人工智能代理正在从简单的答案引擎演变为持久的团队，这些团队使用工具、委派工作、从经验中学习，并修改塑造其未来行为的工件。部署的核心问题不再仅仅是代理能够做什么，而是谁控制他们被允许成为怎样的存在。我们介绍了 logos，这是一个可插拔的自我进化和治理层，它增强了现有的多代理框架，而不是替代它们。logos 将异构的多模态输入（包括文档、图像、音频、表格、数据库、API 和人类指令）编译成版本化的代理包，这些包包含代理、工具、知识、测试、权限和政策。在操作过程中，它将代理活动转化为可移植的、可审计的事件痕迹，并在框架和后端之间应用失败关闭验证。每一个学习到的提示、记忆、技能、工具、角色或工作流程在获得排除执行证据、人类控制的政策和明确授权之前，始终保持为不可信的发布候选。这种架构实现了“可验证的人机循环工程”：代理可以行动、提问、学习和提出改进建议，而人类可以在不打断连续操作的情况下引导目标、权限、批准和不可逆行动。logos 为负责任的自动化提供了一种动态逻辑。代理可以以机器速度进化，但只有证据和人类权威才能完成这一循环。
cs.AI / 74 / 2607.10880
First-Order Modal Logic in HOL: Deep and Shallow Embeddings with Automated Faithfulness (Extended Preprint)
HOL中的一阶模态逻辑：具有自动忠实性的深度和浅层嵌入（扩展预印本）
Abstract
We extend, in Isabelle/HOL, the deep-and-shallow embedding methodology of our prior work from propositional to first-order modal logic (FML) with constant-domain Kripke semantics. Three embeddings of FML into classical higher-order logic (HOL) are provided side by side: a deep embedding, a heavyweight maximal-shallow embedding, and a lightweight minimal-shallow embedding. The minimal-shallow embedding is presented as an Isabelle/HOL locale, parametrised by an accessibility relation, a world-indexed interpretation, a universe of worlds, and a variable assignment; the locale form admits a global faithfulness theorem, stating that quantifying over all minimal-shallow interpretations recovers exactly deep validity. A central technical contribution is a mechanisation, for FML under constant-domain Kripke semantics, of the (countable) downward L\"owenheim-Skolem theorem, which underpins the automation of our faithfulness proof between the deep and minimal-shallow embeddings. Deploying it inside an extension of the minimal-shallow locale resolves the surjectivity problem that arises against an uncountable domain of individuals -- where the locale's variable assignment, having countable domain V = nat, cannot be surjective onto the domain -- and thereby yields faithfulness over the full domain. Since prior work treats only the propositional fragment, we develop here the substitution machinery (free/bound-variable predicates, the fresh-variable function, capture-avoiding substitution, alphabetic renaming, the substitutability predicate, the substitution lemma, and size-based induction principles) needed for the first-order quantifiers.
Chinese Translation
我们在Isabelle/HOL中扩展了我们先前工作的深度和浅层嵌入方法，将其从命题逻辑推广到具有常量域Kripke语义的一阶模态逻辑（FML）。提供了三种将FML嵌入经典高阶逻辑（HOL）的方法：深度嵌入、重量级最大浅层嵌入和轻量级最小浅层嵌入。最小浅层嵌入被呈现为一个Isabelle/HOL的局部（locale），其参数包括可达性关系、世界索引解释、世界的宇宙和变量赋值；该局部形式允许一个全局忠实性定理，表明对所有最小浅层解释的量化恰好恢复深度有效性。一个核心技术贡献是对具有常量域Kripke语义的FML的（可数）向下Löwenheim-Skolem定理的机械化，这为我们在深度和最小浅层嵌入之间的忠实性证明的自动化提供了基础。在最小浅层局部的扩展中部署该定理解决了在不可数个体域上出现的映射性问题——局部的变量赋值具有可数域V = nat，无法对域进行满射——从而在整个域上实现了忠实性。由于先前的工作仅处理命题片段，我们在此开发了所需的一阶量词的替换机制（自由/绑定变量谓词、新变量函数、避免捕获的替换、字母重命名、可替换性谓词、替换引理和基于大小的归纳原则）。
cs.AI / 75 / 2607.10891
SETA: Scaling Environments for Terminal Agents
SETA：终端代理的可扩展环境
Shen, Qijia, Huang, Zhiqi, Kamanuru, Vamsidhar, Aliev, Aznaur, Rainton, Jay, Awelkair, Ahmed, Zeng, Zhichen, Li, Jiajun, Dong, Shi, Yuan, Yueming, Ma, Boyuan, Zhang, Qizheng, Fu, Jiwei, Mao, Yuzhen, Fan, Wendong, Nie, Ping, Torr, Philip, Ghanem, Bernard, Hu, Changran, Li, Jonathan Lingjie, Thakker, Urmish, Li, Guohao
Abstract
Large language models (LLMs) are rapidly shifting toward agents that solve tasks through diverse interfaces, including web and graphical user interfaces (GUIs). Among these, the terminal command line provides a text-based, general-purpose interface, covering tasks from system operations to data science and machine learning. However, scaling terminal-agent training remains challenging, as it requires diverse and coherent task instructions, executable environments, and reliable verification, while lacking naturally grounded supervision data. In this work, we propose SETA, a scalable framework for generating verifiable terminal environments for reinforcement learning (RL). The framework consists of two pipelines sharing a unified verification mechanism: SETA-Synth converts diverse sources into standardized RL environments, and SETA-Evol further expands from existing environments with adaptive control of difficulty and diversity. Together, we construct and release SETA-Env, the largest open-source verifiable terminal RL dataset to date, containing over 4,500 environments. We evaluate our dataset by training Qwen3-8B with GRPO on SETA-Env, achieving 12% pass rate on Terminal-Bench 2.0, the best reported result for an RL-trained model at the 8B scale. We further observe gains on DeepSeek-V4-Flash under the same terminal agent harness, with pass@1 on Terminal-Bench 2.0 improving from 40% to 43% and pass@5 improving from 54% to 58%. These results demonstrate that SETA- Env provides high-quality training environments for terminal agents and serves as a valuable resource for advancing research on terminal-based agent learning.
Chinese Translation
大型语言模型（LLMs）正迅速转向通过多样化接口解决任务的代理，包括网页和图形用户界面（GUIs）。其中，终端命令行提供了一种基于文本的通用接口，涵盖从系统操作到数据科学和机器学习的任务。然而，扩展终端代理的训练仍然面临挑战，因为这需要多样且连贯的任务指令、可执行环境和可靠的验证，同时缺乏自然基础的监督数据。在本研究中，我们提出了SETA，一个用于生成可验证终端环境的可扩展框架，以支持强化学习（RL）。该框架由两个共享统一验证机制的管道组成：SETA-Synth将多样化来源转换为标准化的RL环境，而SETA-Evol则在现有环境的基础上，通过自适应控制难度和多样性进一步扩展。我们共同构建并发布了SETA-Env，这是迄今为止最大的开源可验证终端RL数据集，包含超过4500个环境。我们通过在SETA-Env上使用GRPO训练Qwen3-8B来评估我们的数据集，在Terminal-Bench 2.0上实现了12%的通过率，这是在8B规模下RL训练模型的最佳报告结果。我们进一步观察到在相同终端代理框架下，DeepSeek-V4-Flash的表现也有所提升，Terminal-Bench 2.0的pass@1从40%提高到43%，pass@5从54%提高到58%。这些结果表明，SETA-Env为终端代理提供了高质量的训练环境，并作为推动终端基础代理学习研究的宝贵资源。
cs.AI / 76 / 2607.10896
Incremental Transformer for Surrogate-Based Inverse Design of Geopolymer Mixtures
基于增量变换器的地聚合物混合物代理逆向设计
Abstract
Small-data inverse design is challenging in engineering informatics when observations are heterogeneous, mixed-type, and constrained by physical relations among design variables. This work proposes a topology-aware surrogate framework guided by an Incremental Transformer (INCRT) for physics-constrained inverse design, applied to geopolymer mixture design. The method integrates intrinsic-dimensionality analysis, mixed-variable design-space representation, tabular surrogate prediction, INCRT-based manifold rationalisation, and constrained inverse optimisation. Using a public benchmark of fly-ash and slag-based geopolymer concrete mixtures with compressive-strength and carbon-emission targets, the high-dimensional design space proves strongly redundant, organising around fewer effective mixture regimes. Compressive strength requires nonlinear tabular surrogates, while carbon emission is largely determined by composition and well recovered by regularised linear models. INCRT thus acts not as a replacement for tabular predictors but as a rationalisation layer providing prototype regimes and a manifold-support score for inverse design. Three strategies are compared: unconstrained surrogate optimisation, physics-constrained optimisation, and topology-aware physics-constrained optimisation. Unconstrained optimisation can match target strength but may yield physically invalid or off-manifold candidates; physics-only constraints do not always ensure data support. The topology-aware strategy yields candidates balancing target compliance, carbon reduction, physical admissibility, and proximity to the learned feasible manifold. The framework aims not to replace experimental validation but to support screening of credible candidate mixtures from small, mixed, physically constrained engineering datasets.
Chinese Translation
在工程信息学中，当观测数据异质、混合类型且受设计变量之间物理关系的限制时，小数据逆向设计面临挑战。本研究提出了一种由增量变换器（Incremental Transformer, INCRT）引导的拓扑感知代理框架，用于物理约束的逆向设计，应用于地聚合物混合物设计。该方法整合了内在维度分析、混合变量设计空间表示、表格代理预测、基于INCRT的流形合理化和约束逆向优化。利用一个公共基准数据集，该数据集包含以粉煤灰和矿渣为基础的地聚合物混凝土混合物，其抗压强度和碳排放目标，发现高维设计空间存在显著冗余，围绕较少的有效混合模式进行组织。抗压强度需要非线性表格代理，而碳排放主要由成分决定，并可通过正则线性模型良好恢复。因此，INCRT并不是表格预测器的替代品，而是作为一个合理化层，提供原型模式和流形支持评分以进行逆向设计。比较了三种策略：无约束代理优化、物理约束优化和拓扑感知物理约束优化。无约束优化可以达到目标强度，但可能产生物理上无效或偏离流形的候选者；仅物理约束并不总能确保数据支持。拓扑感知策略则产生平衡目标合规性、碳减排、物理可接受性和接近已学习可行流形的候选者。该框架旨在支持从小型、混合、物理约束的工程数据集中筛选可信候选混合物，而不是替代实验验证。
cs.AI / 77 / 2607.10918
Learning Linear Temporal Specifications from Demonstrations with Uncertainty
从带有不确定性的演示中学习线性时序规范
Abstract
Learning temporal logic specifications from system demonstrations is essential for tasks such as formal verification and controller synthesis, especially in safety-critical domains. Existing approaches typically assume demonstrations are correct or only affected by misclassification errors. In practice, however, system traces are often uncertain or incomplete due to sensor faults, measurement errors, or data loss. We present a framework for learning minimal Linear Temporal Logic (LTL) formulas from demonstrations with uncertainty. Our approach models uncertainty via Hamming distance to generate possible estimates around each observed trace, which are grouped with constraints requiring that at least one trace per group is consistent with the learned formula. Our problem is then reduced to an equivalent Pseudo-Boolean Optimization. We evaluate our method against state-of-the-art LTL learning approaches and show that it recovers specifications that more closely align with ground-truth formulas under uncertainty.
Chinese Translation
从系统演示中学习时序逻辑规范对于形式验证和控制器合成等任务至关重要，尤其是在安全关键领域。现有方法通常假设演示是正确的，或者仅受到误分类错误的影响。然而，在实际应用中，由于传感器故障、测量误差或数据丢失，系统轨迹往往是不确定或不完整的。我们提出了一种从带有不确定性的演示中学习最小线性时序逻辑（LTL）公式的框架。我们的方法通过汉明距离建模不确定性，以生成围绕每个观察到的轨迹的可能估计，这些估计与约束条件相结合，要求每组中至少有一个轨迹与学习到的公式一致。然后，我们的问题被简化为一个等价的伪布尔优化问题。我们将我们的方法与最先进的LTL学习方法进行了评估，结果表明，在不确定性下，我们的方法能够恢复与真实公式更紧密对齐的规范。
cs.AI / 78 / 2607.10966
SVR-R1: Bootstrapping Multi-modal Reasoning with Self-verification in Reinforcement Learning
SVR-R1：通过自我验证在强化学习中引导多模态推理
Abstract
We introduce Self-Verified Reasoner (SVR-R1), a multi-turn RL framework that turns a model's own verification into a learning signal for multimodal reasoning. For each query, the model proposes an answer using the same weights, and issues a binary self-verdict (Yes/No). A 'No' triggers a second-chance rethink; a 'Yes,' or a turn cap, finalizes the output for computing the outcome-based reward. SVR-R1 is implemented with GRPO and an asynchronous multi-turn rollout framework and needs no external supervision or auxiliary critics. We evaluate SVR-R1 on vision-language reasoning benchmarks and show that it improves accuracy by a large margin over strong standard GRPO baselines. Training dynamics show decreasing reliance on verification-fewer verification turns, yet higher test accuracy-indicating that the gap between verification and generation narrows as the policy internalizes self-correction and chooses the most confident answer via our framework. SVR-R1 bridges the less explored intersection of inference-time self-refinement and RL training for VLMs, offering a simple yet effective recipe for bootstrapping multimodal reasoning. We will open-source \textbf{SVR-R1} to facilitate future research in VLMs.
Chinese Translation
我们介绍了自我验证推理器（Self-Verified Reasoner，SVR-R1），这是一个多轮强化学习框架，它将模型自身的验证转化为多模态推理的学习信号。对于每个查询，模型使用相同的权重提出答案，并给出一个二元自我裁决（是/否）。'否'触发第二次思考；'是'或达到轮次上限则最终确定输出以计算基于结果的奖励。SVR-R1采用GRPO（Generalized Reinforcement Policy Optimization）和异步多轮回滚框架实现，无需外部监督或辅助评估。我们在视觉-语言推理基准上评估了SVR-R1，结果显示其在强大的标准GRPO基线之上显著提高了准确性。训练动态表明对验证的依赖性逐渐降低——验证轮次减少，但测试准确性提高——这表明随着策略内化自我修正并通过我们的框架选择最自信的答案，验证与生成之间的差距在缩小。SVR-R1架起了推理时自我精炼与视觉语言模型（VLMs）训练之间较少探索的交集，为引导多模态推理提供了一种简单而有效的方法。我们将开源SVR-R1，以促进未来在VLMs领域的研究。
cs.AI / 79 / 2607.10972
From Checker to Forecaster: Code-Owned Evaluation of Model-Generated Strategic Routes Under Delayed Ground Truth
从检查者到预测者：在延迟真实情况下对模型生成的战略路线进行代码拥有的评估
Abstract
Many evaluations of model outputs rely either on contracts checkable at evaluation time or on feedback that arrives within the operating loop. We study the complementary setting in which ground truth is delayed, censored, or private, so deterministic code cannot check correctness at scoring time and must instead issue a code-owned provisional forecast. RouteCast instantiates this regime for model-generated typed strategic routes: models propose candidate routes and structured factors; point-in-time evidence, reference classes, and deterministic transformations produce a provisional forecast-ranking; later outcomes evaluate the forecast. In a retrospective venture pilot on 21 binary-outcome cases (6 positive, 15 negative), the whole-packet RouteCast score showed preliminary retrospective discrimination (AUC 0.756, 95% CI [0.471,0.980]), while a blind LLM judge reached AUC 0.678 [0.419,0.897] and an identity-exposed LLM judge reached AUC 0.761 [0.515,0.944], consistent with recognition- or outcome-related leakage risk. A preregistered decomposition ablation on the same binary subset found that converting the identical inputs into typed staged routes was indistinguishable from the whole-packet score (Delta AUC = -0.144, 95% CI [-0.471,0.176]) and from a deterministic heuristic (Delta AUC = -0.089, 95% CI [-0.412,0.278]). The pilot establishes an auditable feasibility result and exposes failure modes; it does not establish prospective calibration, causal decision improvement, route-decomposition advantage, or cross-domain validity.
Chinese Translation
许多模型输出的评估依赖于在评估时可检查的合同或在操作循环内到达的反馈。我们研究了一个互补的情境，即真实情况被延迟、审查或私有，因此确定性代码无法在评分时检查正确性，而必须发出代码拥有的临时预测。RouteCast 实现了这一机制，用于模型生成的类型化战略路线：模型提出候选路线和结构化因素；时间点证据、参考类别和确定性变换生成临时预测排名；后续结果评估该预测。在对21个二元结果案例（6个正向，15个负向）进行的回顾性试点中，整体 RouteCast 评分显示出初步的回顾性区分能力（AUC 0.756，95% CI [0.471,0.980]），而盲评的 LLM 判断者达到了 AUC 0.678 [0.419,0.897]，身份暴露的 LLM 判断者达到了 AUC 0.761 [0.515,0.944]，与识别或结果相关的泄漏风险一致。对同一二元子集进行的预注册分解消融实验发现，将相同输入转换为类型化的分阶段路线与整体评分无显著区别（Delta AUC = -0.144，95% CI [-0.471,0.176]），与确定性启发式方法也无显著区别（Delta AUC = -0.089，95% CI [-0.412,0.278]）。该试点建立了可审计的可行性结果并揭示了失败模式；但并未建立前瞻性校准、因果决策改进、路线分解优势或跨领域有效性。
cs.AI / 80 / 2607.11019
QwenPaw-Data: Bridging Facts, Methodology, and Execution for Autonomous Enterprise Data Analytics
QwenPaw-Data：连接事实、方法论与执行的自主企业数据分析
Abstract
Enterprise data analysis is emerging as a distinct frontier for autonomous agents. Compared with general-purpose interaction and software engineering, it operates in an open, ambiguous, and continuously evolving environment. These characteristics call for a data-agent architecture that treats semantics, methodology, execution, and evolution as first-class system concerns. To this end, we introduce QwenPaw-Data, an agentic data system designed for enterprise intelligent data analysis. QwenPaw-Data consolidates heterogeneous assets from warehouses, dashboards, documents, interaction logs, and historical tasks into reusable, governable, and evolvable analysis assets, then turns natural-language requests into end-to-end analytical workflows spanning data understanding, retrieval, analysis, report generation, and decision support. Its architecture decomposes the problem into three collaborative subsystems: DataBridge provides trustworthy semantic grounding through interconnected metadata, knowledge, and trace graphs; Skill-Hub codifies expert analytical methodology into reusable and verifiable skills; and Host materializes these evidence and method assets into controllable, artifact-centric runtime execution. Across these subsystems, semantics, methods, traces, and feedback are continuously deposited back into the system, forming a self-evolving asset flywheel. Experiments on public benchmarks and real-world industrial BI workloads show that QwenPaw-Data improves both verifiable data access capability and higher-level analytical quality, offering a practical foundation for reliable, traceable, and continuously improving enterprise data agents.
Chinese Translation
企业数据分析正在成为自主智能体的一个独特前沿。与通用交互和软件工程相比，它在一个开放、模糊且不断演变的环境中运作。这些特征要求一种数据-智能体架构，将语义、方法论、执行和演变视为首要系统关注点。为此，我们提出了QwenPaw-Data，这是一个为企业智能数据分析设计的智能数据系统。QwenPaw-Data将来自数据仓库、仪表板、文档、交互日志和历史任务的异构资产整合为可重用、可治理和可演变的分析资产，然后将自然语言请求转化为涵盖数据理解、检索、分析、报告生成和决策支持的端到端分析工作流。其架构将问题分解为三个协作子系统：DataBridge通过互联的元数据、知识和追踪图提供可信的语义基础；Skill-Hub将专家分析方法论编码为可重用和可验证的技能；Host则将这些证据和方法资产具体化为可控的、以工件为中心的运行时执行。在这些子系统中，语义、方法、追踪和反馈不断被反馈回系统，形成自我演变的资产飞轮。在公共基准和真实世界工业商业智能（BI）工作负载上的实验表明，QwenPaw-Data提高了可验证的数据访问能力和更高层次的分析质量，为可靠、可追溯和持续改进的企业数据智能体提供了实用基础。
cs.AI / 81 / 2607.11063
AdvNav: Behavior-Guided Black-Box Adversarial Attacks on Vision-Language Navigation
AdvNav：基于行为指导的黑箱对抗攻击在视觉-语言导航中的应用
Abstract
Despite progress in Embodied AI, Vision-and-Language Navigation systems remain vulnerable to adversarial visual disturbances. Most existing methods rely on white-box access to target model gradients, which is often unrealistic for real-world deployed systems and computationally exhaustive due to recursive backpropagation for optimization, limiting their applicability. While previous black-box methods predominantly target single-step, instantaneous decision tasks, they struggle to handle the task complexities and temporal dependencies. This highlights the need for a gradient-free attack method that can effectively disrupt the multistep sequential perception-action loop using only observable inputs and outputs. Therefore, we propose AdvNav, a behavior-guided black-box adversarial attack framework that disturbs an agent's first-person views during navigation. To construct an informative surrogate objective for effective optimization guidance in gradient-free search under the black-box setting, we design a dual-granularity behavior-based feedback, aggregating a trajectory-level performance score representing overall navigation degradation, an action-level reward score considering the potential decision risk, and a deviation indicator, all of which are extracted from the agent's self-output behaviors. This feedback guides a hybrid optimization strategy that heuristically tunes perturbation strength via adaptive updates and evolves noise spatial structure genetically, to iteratively discover the most disruptive noise configuration. Evaluated against Transformer-based HAMT and LLM-based MapGPT with two types of backbones on R2R dataset, AdvNav achieves 49.70/65.96/87.30% Attack Success Rate. The result demonstrates the effectiveness and generality of AdvNav, reveals critical perception vulnerabilities and offers insights for the design of future resilient VLN models.
Chinese Translation
尽管在具身人工智能方面取得了进展，视觉与语言导航系统仍然容易受到对抗性视觉干扰的影响。现有的大多数方法依赖于对目标模型梯度的白箱访问，这在实际部署的系统中往往不切实际，并且由于优化过程中的递归反向传播而计算开销巨大，限制了其适用性。虽然之前的黑箱方法主要针对单步、瞬时决策任务，但它们在处理任务复杂性和时间依赖性方面存在困难。这突显了需要一种无梯度攻击方法，该方法能够仅通过可观察的输入和输出有效地干扰多步顺序感知-行动循环。因此，我们提出了AdvNav，一个基于行为指导的黑箱对抗攻击框架，旨在干扰代理在导航过程中的第一人称视角。为了在黑箱设置下构建一个信息丰富的替代目标，以有效指导无梯度搜索的优化，我们设计了一种双粒度的基于行为的反馈机制，聚合了代表整体导航退化的轨迹级性能评分、考虑潜在决策风险的动作级奖励评分以及偏差指标，这些均来自代理的自我输出行为。该反馈指导了一种混合优化策略，通过自适应更新启发式调整扰动强度，并通过遗传算法进化噪声空间结构，以迭代发现最具干扰性的噪声配置。在R2R数据集上针对基于Transformer的HAMT和基于LLM的MapGPT进行评估，AdvNav实现了49.70%/65.96%/87.30%的攻击成功率。结果证明了AdvNav的有效性和普适性，揭示了关键的感知脆弱性，并为未来鲁棒的视觉-语言导航模型的设计提供了见解。
cs.AI / 82 / 2607.11079
Are LLMs Ready for Scientific Discovery? A Capability-Oriented Benchmark for AI Scientists
大型语言模型准备好进行科学发现了吗？面向能力的人工智能科学家基准测试
Abstract
Existing benchmarks for scientific data analysis evaluate LLMs primarily on code execution or workflow completion, overlooking that scientific analysis serves to support distinct types of scientific claims: hypothesis exploration, statistical inference, mechanistic explanation, each with different assumptions and validity criteria. We introduce SDABench, a benchmark that reorganizes evaluation around six capabilities (descriptive, exploratory, inferential, predictive, causal, and mechanistic) across five domains (Biology, Chemistry, Environment, Geography, Physics). SDABench comprises 527 real-data instances (SDA-Real) and 6000 synthetic instances (SDA-Synth), each in both multiple-choice and open-ended formats, constructed through an automated pipeline. Evaluating 15 representative LLMs, we find that models handle descriptive analysis well but degrade sharply on tasks requiring assumption selection, latent-process modeling, or mechanistic reasoning. SDABench further provides a five-stage error analysis framework that locates where LLMs fail: more advanced models more reliably identify the relevant scope and variables, but still struggle to select appropriate analytical procedures, model variable relationships, and draw valid conclusions.
Chinese Translation
现有的科学数据分析基准主要通过代码执行或工作流程完成来评估大型语言模型（LLMs），而忽视了科学分析旨在支持不同类型的科学主张：假设探索、统计推断和机制解释，每种类型都有不同的假设和有效性标准。我们提出了SDABench，一个围绕六种能力（描述性、探索性、推断性、预测性、因果性和机制性）在五个领域（生物学、化学、环境、地理和物理）重新组织评估的基准。SDABench包含527个真实数据实例（SDA-Real）和6000个合成实例（SDA-Synth），每个实例都有多项选择和开放式格式，通过自动化管道构建。通过评估15个代表性的LLMs，我们发现模型在描述性分析方面表现良好，但在需要选择假设、潜在过程建模或机制推理的任务上表现急剧下降。SDABench还提供了一个五阶段错误分析框架，定位LLMs失败的地方：更先进的模型更可靠地识别相关范围和变量，但仍然在选择适当的分析程序、建模变量关系和得出有效结论方面存在困难。
cs.AI / 83 / 2607.11084
NVAITC AI Scientist: A Governed End-to-End Research System -- A Hypertension GWAS Case Study
NVAITC AI科学家：一个受管控的端到端研究系统——高血压全基因组关联研究案例研究
Abstract
Agentic research systems are emerging as a new paradigm for coordinating scientific workflows beyond isolated model inference, code generation, or statistical analysis. However, deployment in institutional biomedical environments requires governed mechanisms for research planning, data access, workflow orchestration, evidence tracking, reproducibility, and human oversight. We present NVAITC AI Scientist (NAIS), a governed end-to-end agentic research system designed to support domain-general scientific workflows while keeping protected data within institutional privacy boundaries. NAIS integrates proposal review, execution planning, governed computational routing, reproducible workflow orchestration, evidence generation, and scientist-in-the-loop oversight. We validate NAIS in a real-world hypertension genome-wide association study (GWAS) using hospital-linked genotype and electronic health record (EHR) data from 286,422 individuals under an aggregate-only data policy. The agent planned cohort extraction, orchestrated GWAS execution, generated quality-control summaries, and drafted publication-oriented outputs. Human-AI review identified phenotype discrepancies and enabled iterative refinement of the hypertension definition. After reconciliation, the agent-orchestrated GWAS reproduced established hypertension loci, including FGF5, ATP2B1, CNNM2, FTO, and GRB14, with the strongest signal at FGF5 reaching $-\log_{10}(p) \sim 70$. As a secondary demonstration, NAIS also supported a drug-induced liver injury prediction workflow, achieving a multimodal graph neural network AUC of 0.842. These results demonstrate that governed agentic research systems can support scalable AI-assisted biomedical discovery while producing outputs comparable to expert-led workflows.
Chinese Translation
代理研究系统作为一种新范式，正在兴起，以协调超越孤立模型推断、代码生成或统计分析的科学工作流程。然而，在机构生物医学环境中的部署需要受管控的机制，以进行研究规划、数据访问、工作流程编排、证据追踪、可重复性和人类监督。我们提出了NVAITC AI科学家（NAIS），这是一个受管控的端到端代理研究系统，旨在支持领域通用的科学工作流程，同时将受保护的数据保持在机构隐私边界内。NAIS集成了提案审查、执行规划、受管控的计算路由、可重复的工作流程编排、证据生成和科学家参与的监督。我们在一个真实的高血压全基因组关联研究（GWAS）中验证了NAIS，该研究使用了来自286,422个个体的医院关联基因型和电子健康记录（EHR）数据，并遵循仅聚合数据政策。该代理规划了队列提取，编排了GWAS执行，生成了质量控制摘要，并起草了面向出版的输出。人机审查识别了表型差异，并使高血压定义的迭代优化成为可能。在调和后，代理编排的GWAS再现了已建立的高血压位点，包括FGF5、ATP2B1、CNNM2、FTO和GRB14，其中FGF5的信号最强，达到$- ext{log}_{10}(p) ext{ } ext{∼} ext{ } 70$。作为第二个示范，NAIS还支持了一种药物诱导肝损伤预测工作流程，达到了0.842的多模态图神经网络AUC。这些结果表明，受管控的代理研究系统能够支持可扩展的AI辅助生物医学发现，同时产生与专家主导工作流程相当的输出。
cs.AI / 84 / 2607.11089
OS-Pruner: Pruning Chains-of-Thought of Reasoning Models via Optimal Stopping
OS-Pruner：通过最优停止修剪推理模型的思维链
Abstract
Large Language Models (LLMs) have achieved remarkable success in complex reasoning tasks through Chain-of-Thought (CoT) prompting. However, these models often exhibit "computational overthinking," generating redundant reasoning steps that increase latency and cost without improving accuracy. Recent studies suggest that CoT trajectories can be significantly pruned, yet existing methods often rely on forcing a static thinking budget, heuristic filtering, sub-optimal early exit via classification, or expensive re-training. In this paper, we introduce OS-Pruner, a lightweight plug-in framework that formulates chain-of-thought pruning as an optimal stopping problem. Given a reasoning prefix, OS-Pruner learns whether further reasoning is worth its token cost by optimizing an explicit utility that trades off final-answer accuracy against generated length. Our novel formulation enables the model to dynamically assess the sufficient point of termination for a reasoning chain. OS-Pruner is designed to be lightweight during both training and inference, and to provide users with fine-grained control over the reasoning-effort vs. accuracy trade-off. On diverse reasoning benchmarks and base models, OS-Pruner achieves 20-60\% reduction in generation length with minimal accuracy sacrifice.
Chinese Translation
大型语言模型（LLMs）在复杂推理任务中通过思维链（Chain-of-Thought, CoT）提示取得了显著成功。然而，这些模型常常表现出“计算过度思考”的现象，生成冗余的推理步骤，导致延迟和成本增加而不提高准确性。近期研究表明，思维链轨迹可以显著修剪，但现有方法往往依赖于强制静态思维预算、启发式过滤、通过分类的次优提前退出或昂贵的再训练。在本文中，我们提出了OS-Pruner，一种轻量级插件框架，将思维链修剪形式化为最优停止问题。在给定推理前缀的情况下，OS-Pruner通过优化一个明确的效用函数来学习进一步推理是否值得其代币成本，该效用函数在最终答案的准确性与生成长度之间进行权衡。我们新颖的形式化方法使模型能够动态评估推理链的终止充分点。OS-Pruner在训练和推理过程中均设计为轻量级，并为用户提供对推理努力与准确性权衡的细粒度控制。在多样的推理基准和基础模型上，OS-Pruner实现了生成长度减少20-60%且准确性损失最小。
cs.AI / 85 / 2607.11138
A Formal Hierarchical Architecture for Agentic Orchestration with Stack-Based Execution and Lazy Discovery
一种用于代理编排的正式层次架构，具有基于栈的执行和惰性发现
Abstract
The rapid expansion of capabilities in Large Language Model (LLM) agents has exposed a critical architectural bottleneck: when agents are given access to a flat, monolithic registry of tools, the model must evaluate hundreds or thousands of options simultaneously. This leads to decision-space explosion, context window saturation, and degraded routing accuracy. To address these limitations, this paper presents a hierarchical, skill-based architecture for agentic orchestration. Capabilities are organized as a rooted tree where internal nodes make routing decisions and leaf nodes execute deterministic tasks. The runtime enforces a single-step execution loop governed by a Last-In-First-Out (LIFO) stack, giving the agent a form of memory akin to a Pushdown Automaton, therefore enabling it to track nested execution contexts and resume deterministically from any depth. Capability discovery follows a manifest-driven, lazy-loading protocol: only the immediate children of the active node are loaded, so memory and prompt costs scale with the explored path rather than the global registry. By replacing global memory with localized stack frames, the architecture prevents outputs from one execution branch from leaking into another, establishing the isolation guarantees required for deployment in regulated enterprise environments. We also discuss UPI Help, an AI-powered digital payments support product, as a motivating production deployment context. We provide a mathematical formalization of the orchestration state, detailed algorithmic analysis of the execution loop, and controlled benchmarks comparing flat and hierarchical routing under increasing tool catalogs, multi-step workflow pressure, and visible schema-token exposure per LLM call.
Chinese Translation
大型语言模型（LLM）代理能力的快速扩展暴露了一个关键的架构瓶颈：当代理被赋予访问一个扁平的、单一的工具注册表时，模型必须同时评估数百或数千个选项。这导致了决策空间的爆炸、上下文窗口的饱和以及路由准确性的下降。为了解决这些限制，本文提出了一种基于技能的层次化代理编排架构。能力被组织成一棵根树，其中内部节点做出路由决策，叶节点执行确定性任务。运行时强制执行一个由后进先出（LIFO）栈控制的单步执行循环，使代理具备类似于下推自动机的记忆形式，从而能够跟踪嵌套执行上下文并从任何深度确定性地恢复。能力发现遵循一种基于清单的惰性加载协议：仅加载活动节点的直接子节点，因此内存和提示成本随着探索路径的变化而变化，而不是全球注册表。通过用局部栈帧替代全局内存，该架构防止了一个执行分支的输出泄漏到另一个分支，从而建立了在受监管的企业环境中部署所需的隔离保证。我们还讨论了UPI Help，一个由人工智能驱动的数字支付支持产品，作为一个激励性的生产部署背景。我们提供了编排状态的数学形式化，详细的执行循环算法分析，以及在增加工具目录、多步骤工作流压力和每次LLM调用的可见模式令牌暴露下比较扁平和层次路由的受控基准测试。
cs.AI / 86 / 2607.11141
NextFund: A Unified Performance Tracking Platform for Agentic Portfolio Management
NextFund：一个统一的代理投资组合管理绩效跟踪平台
Abstract
Large language models (LLMs) based agents are beginning to participate in portfolio construction and market analysis, where decisions must be justified under evolving information and risk constraints. Current assessment practice, however, remains poorly aligned with this setting: many studies rely on static examinations or report only terminal portfolio returns, while the intermediate evidence, analyst judgments, and execution steps that produced those returns stay largely invisible. We introduce NextFund, an evaluation platform that makes financial-agent behavior observable under live market conditions. The platform couples time-consistent market access, coordinated multi-agent analysis, and persistent logging of the full decision path from observation to trade. Through an interactive Trading Arena, users can compare models across markets, inspect equity curves, and drill from leaderboard outcomes down to individual justifications. We present NextFund on Hong Kong, U.S., and China A-share equities, illustrating how inspectable decision histories enable fairer benchmarking and more actionable diagnosis. Our demo is available at https://paradoox.cn/nextfund/.
Chinese Translation
基于大型语言模型（LLMs）的代理开始参与投资组合构建和市场分析，在此过程中，决策必须在不断变化的信息和风险约束下得到合理化。然而，目前的评估实践与这一环境的契合度较低：许多研究依赖于静态检查或仅报告终端投资组合回报，而产生这些回报的中间证据、分析师判断和执行步骤则大多不可见。我们推出了NextFund，一个在实时市场条件下使金融代理行为可观察的评估平台。该平台结合了时间一致的市场访问、协调的多代理分析和从观察到交易的完整决策路径的持续记录。通过一个互动交易竞技场，用户可以在不同市场之间比较模型、检查收益曲线，并从排行榜结果深入到个别的决策依据。我们在香港、美国和中国A股股票上展示了NextFund，说明可检查的决策历史如何促进更公平的基准测试和更具可操作性的诊断。我们的演示可在 https://paradoox.cn/nextfund/ 获取。
cs.AI / 87 / 2607.11149
The Hidden Footprint: Making Storage a First-Class Metric for LLM Agent Evaluation
隐性足迹：将存储作为大语言模型代理评估的一级指标
Abstract
LLM agent benchmarks measure task completion, reliability, and inference cost, but not the persistent data an agent run leaves on disk, including logs, context snapshots, checkpoints, and debug traces. We introduce AgentFootprint, a cross-framework benchmark of post-run agent storage footprint. Its serialization-aware metric suite measures total retention, channel composition, duplication, growth, compressibility, and conversation-history reconstructability. It addresses a measurement trap: naive byte-level measurement understates duplication by an order of magnitude because database paging and JSON escaping obscure repeated content. A fixed-trace control separates agent-generated logical volume from persistence-layer amplification: replaying the same trajectory through seven persisting frameworks yields a 6.7x spread. Under identical models, tools, and tasks, configurations with 100% accuracy differ by 15.7x in retained bytes, although their defaults support different recovery and audit capabilities. Three full-history configurations grow superlinearly on a repeated-observation stress task. Exported trajectories from 108 instance-normalized SWE-bench Verified submissions span three orders of magnitude per instance, with no detectable correlation with resolve rate. A content-addressed store reduces retention by 4.8x-32.7x while preserving every reconstructability score. These results establish persistent storage as a resource metric to report jointly with accuracy and reconstructability.
Chinese Translation
大语言模型（LLM）代理基准测量任务完成度、可靠性和推理成本，但未考虑代理运行在磁盘上留下的持久数据，包括日志、上下文快照、检查点和调试痕迹。我们引入了AgentFootprint，这是一个跨框架的基准，用于评估运行后代理的存储足迹。其序列化感知的指标套件测量总保留量、通道组成、重复性、增长、可压缩性和对话历史重构能力。它解决了一个测量陷阱：天真的字节级测量低估了重复性，数量级差异达一个数量级，因为数据库分页和JSON转义模糊了重复内容。固定轨迹控制将代理生成的逻辑卷与持久层放大分开：通过七个持久框架重放相同的轨迹，产生了6.7倍的差异。在相同的模型、工具和任务下，具有100%准确率的配置在保留字节上相差15.7倍，尽管它们的默认设置支持不同的恢复和审计能力。三个完整历史配置在重复观察压力任务中呈超线性增长。从108个实例标准化的SWE-bench验证提交中导出的轨迹在每个实例上跨越三个数量级，且与解决率没有可检测的相关性。内容寻址存储在保留每个重构得分的同时，将保留量减少了4.8倍至32.7倍。这些结果确立了持久存储作为一种资源指标，与准确性和重构能力共同报告。
cs.AI / 88 / 2607.11172
STAMP: Provenance-Guided Credit Assignment for Deep Search Agents
STAMP：基于来源的信用分配方法用于深度搜索代理
Abstract
Reinforcement learning for deep-search agents has largely focused on trajectory-level scoring -- outcome correctness, citation-aware rewards, and evidence coverage. Yet the actions that expose supporting documents receive no targeted credit, a gap we call the reward-credit mismatch. We propose STAMP, in which a reference-based verifier judges whether each cited document supports an entity or relation in a training-time evidence graph, and first-exposure attribution traces each supported citation back to the action that first surfaced it. This step credit is injected through sign-preserving advantage modulation, which redistributes advantage across steps without changing the trajectory-level reward or the relative ranking of trajectories within each group. On BrowseComp, BrowseComp-ZH, and xbench-DS, STAMP improves the GRPO baseline by +2.0/+5.5/+3.0 points under matched SFT initialization, training data, and search tools, and composes with both outcome-only and citation-rubric base rewards. Component ablations confirm that the provenance-based credit signal and the sign-preserving advantage modulation each contribute to the gains.
Chinese Translation
深度搜索代理的强化学习主要集中在轨迹级评分上——结果正确性、引用感知奖励和证据覆盖。然而，暴露支持文档的行为并未获得针对性的信用，这一现象我们称之为奖励-信用不匹配。我们提出了STAMP，其中基于引用的验证器判断每个被引用文档是否支持训练时证据图中的实体或关系，而首次曝光归因则将每个被支持的引用追溯到首次出现该引用的行为。此步骤信用通过保持符号的优势调制注入，这种方法在不改变轨迹级奖励或每组内轨迹的相对排名的情况下重新分配优势。在 BrowseComp、BrowseComp-ZH 和 xbench-DS 上，STAMP 在匹配的 SFT 初始化、训练数据和搜索工具下将 GRPO 基线提升了 +2.0/+5.5/+3.0 分，并与仅结果和引用标准的基础奖励相结合。组件消融实验确认，基于来源的信用信号和保持符号的优势调制均对提升效果有所贡献。
cs.AI / 89 / 2607.11175
The Path to Self-Evolving Clinical Systems: Scaling Medical Agents from Assistance to Autonomy
自我进化临床系统的路径：从辅助到自主的医疗智能体扩展
Abstract
The growing ability of large language models and vision language models to jointly interpret and reason over images and text is reshaping medical agents, moving them from task specific predictors toward autonomous systems that perceive, reason, plan, remember, and act in clinical environments. This work departs from the capability first perspective of existing literature and instead begins from clinical deployment, asking what tasks, contamination resistant benchmarks, and interactive training environments are required before medical agents can be trusted in practice. Medical agents are formalized as sequential decision making systems under partial observability, together with a three level autonomy taxonomy spanning assisted, cooperative, and fully autonomous operation. The field is organized along a unified scaling spine consisting of framework scaling, capability scaling, and environment scaling. Within this framework, clinical environment scaling, the integration of tools, data, and clinical gyms, is identified as the most actionable yet underexplored direction for agents operating in PACS, EHR, and FHIR ecosystems. Clinical self evolution, where agents improve through interaction with their environments rather than parameter scaling alone, is further positioned as a key research frontier, drawing insights from self improving agents, agent gyms, and test time compute scaling. Applications across radiology, pathology, ophthalmology, and hospital workflows are examined together with deployment challenges including hallucination, cascading failures, and fairness. By consolidating more than 300 references, with particular emphasis on advances from 2025 to 2026, this work provides a roadmap toward trustworthy, self improving medical imaging systems for real clinical practice.
Chinese Translation
大型语言模型和视觉语言模型在图像和文本的联合解读与推理能力的不断增强正在重塑医疗智能体，使其从特定任务的预测者转变为能够在临床环境中感知、推理、规划、记忆和行动的自主系统。本研究不同于现有文献中以能力为中心的视角，而是从临床部署出发，探讨在医疗智能体可以在实践中被信任之前，所需的任务、抗污染基准和互动训练环境。医疗智能体被形式化为在部分可观测性下的序列决策系统，并提出了一个涵盖辅助、合作和完全自主操作的三层次自主性分类法。该领域沿着一个统一的扩展框架进行组织，包括框架扩展、能力扩展和环境扩展。在此框架内，临床环境扩展，即工具、数据和临床训练场的整合，被确定为在PACS、EHR和FHIR生态系统中操作的智能体最具可行性但尚未充分探索的方向。临床自我进化，即智能体通过与环境的互动而非仅仅通过参数扩展来改进，被进一步定位为一个关键的研究前沿，借鉴了自我改进智能体、智能体训练场和测试时间计算扩展的见解。本文考察了在放射学、病理学、眼科学和医院工作流程中的应用，以及包括幻觉、级联故障和公平性在内的部署挑战。通过整合300多篇参考文献，特别强调2025至2026年的进展，本文为可信赖的、自我改进的医疗影像系统在实际临床应用中的发展提供了一条路线图。
cs.AI / 90 / 2607.11185
SCALECUA: Scaling Computer Use Agents with Verifiable Task Synthesis and Efficient Online RL
SCALECUA：通过可验证任务合成和高效在线强化学习扩展计算机使用代理
Abstract
Computer use agents (CUAs) are emerging as a powerful interface for automating complex digital workflows through visual perception and GUI execution. Online reinforcement learning with verifiable rewards (RLVR) has emerged as a key direction for scaling their capabilities. However, this paradigm is bottlenecked by verifiable data scarcity and online RL inefficiency. To break these barriers, we introduce ScaleCUA, a unified framework that scales online RL for CUAs via verifiable task synthesis and efficient training. At the data level, we design VeriGen, an end-to-end framework for generating verifiable RL tasks through iterative docker interactions and a multi-agent feedback loop. Scaled to 100+ concurrent agent workers via a shared docker interaction probe, this pipeline produces 24K+ verifiable tasks and nearly 3K high-quality RL tasks. To maximize sample efficiency, we propose Frontier Sampling, which tracks per-task capability and allocates rollouts to the current learning frontier. On the training side, we further design Visual Context Segmentation, a sliding window over recent visual context that balances rollout and training-engine pressure, yielding a 2.83x training speedup over step-wise decomposition. Together, ScaleCUA achieves 68.7% on OSWorld and 54.0% on ScienceBoard, establishing new state-of-the-art performance among open-source computer use agents. Code, models, and datasets are available at https://github.com/THUDM/SCALE-CUA.
Chinese Translation
计算机使用代理（CUAs）作为一种强大的接口，正在通过视觉感知和图形用户界面执行来自动化复杂的数字工作流程。带有可验证奖励的在线强化学习（RLVR）已成为扩展其能力的关键方向。然而，这一范式受到可验证数据稀缺和在线强化学习低效的瓶颈。为了解决这些问题，我们提出了ScaleCUA，一个统一框架，通过可验证任务合成和高效训练来扩展CUAs的在线强化学习。在数据层面，我们设计了VeriGen，一个端到端框架，通过迭代的Docker交互和多智能体反馈循环生成可验证的强化学习任务。该管道通过共享的Docker交互探针扩展到100多个并发代理工作者，生成超过24,000个可验证任务和近3,000个高质量的强化学习任务。为了最大化样本效率，我们提出了Frontier Sampling，它跟踪每个任务的能力并将回滚分配给当前学习前沿。在训练方面，我们进一步设计了视觉上下文分割，这是一个针对近期视觉上下文的滑动窗口，平衡了回滚和训练引擎的压力，相较于逐步分解实现了2.83倍的训练加速。总的来说，ScaleCUA在OSWorld上达到了68.7%的性能，在ScienceBoard上达到了54.0%，在开源计算机使用代理中建立了新的最先进性能。代码、模型和数据集可在https://github.com/THUDM/SCALE-CUA获取。
cs.AI / 91 / 2607.11197
What We Talk About When We Talk About LLM Planning: Evidence for Two Distinct Planning Abilities
当我们谈论 LLM 规划时谈论的是什么：两种不同规划能力的证据
Abstract
When LLMs exhibit uneven performance across planning tasks, these gaps are often attributed to task difficulty. We argue that this explanation is incomplete, as task-level variation may reflect distinct latent planning competencies rather than differences along a single ability spectrum. We study this question on ACPBench-Hard by evaluating multiple LLM families under varying test-time reasoning budgets and applying a multidimensional item response theory model to uncover the latent competency structure underlying LLM planning. The analysis reveals two principal dimensions that shape planning performance: operational reasoning, the ability to evaluate local action applicability and immediate state transitions, and structural enumeration, the ability to reason about goal reachability and landmark structure. Operational reasoning improving under model scaling and longer reasoning traces, while structural enumeration remains comparatively insensitive. Our findings motivate competency-level evaluation of LLM planning, shifting the focus from whether models improve overall to which planning competencies improve, under what conditions, and why.
Chinese Translation
当 LLM 在规划任务中表现不均时，这些差距通常被归因于任务难度。我们认为这一解释是不完整的，因为任务层面的变化可能反映出不同的潜在规划能力，而不是沿着单一能力谱的差异。我们通过在 ACPBench-Hard 上评估多种 LLM 家族，在不同的测试时间推理预算下，应用多维项目反应理论模型，来揭示 LLM 规划背后的潜在能力结构。分析结果揭示了塑造规划表现的两个主要维度：操作推理，即评估局部行动适用性和即时状态转变的能力；结构枚举，即推理目标可达性和地标结构的能力。操作推理在模型扩展和更长推理轨迹下有所改善，而结构枚举则相对不敏感。我们的发现促使对 LLM 规划进行能力层面的评估，转变关注点从模型是否整体改善到哪些规划能力改善、在什么条件下改善以及原因何在。
cs.AI / 92 / 2607.11212
PREF-Gate: Provenance-Constrained Relational Evidence Fusion with Validation-Gated Selection for Graph Fraud Detection
PREF-Gate：基于来源约束的关系证据融合与验证门控选择用于图形欺诈检测
Abstract
Relational fraud detection can exploit both label-free graph context and label-derived neighborhood evidence, but these two information sources obey different validity conditions. In particular, neighborhood risk becomes invalid when a queried node's own label, or any validation or test label, enters its construction. We formulate this issue as provenance-constrained relational evidence use and present PREF-Gate, an auditable decision framework with two fixed experts and a finite validation gate. The context expert uses attributes, one-hop means, feature residuals, and degree descriptors without labels. The evidence expert adds self-excluded, training-label-only neighborhood risk and empirical-Bayes summaries that expose support, uncertainty, availability, and shrinkage. Before test inference, the gate selects either expert or one of three pre-specified probability mixtures and fixes the decision threshold. On Amazon, YelpChi, and TFinance, using five identical stratified splits and 14 same-protocol methods, PREF-Gate obtains mean AUPRC values of 0.9085, 0.8104, and 0.8913. It selects the label-free expert on all Amazon and YelpChi splits and an evidence mixture on all TFinance splits. Thus, the main result is conditional rather than universal: label-derived relational evidence is useful only where held-out validation supports it. The framework couples competitive ranking performance with an explicit label-provenance contract, finite selection policy, failure accounting, and review-budget evaluation, providing an auditable knowledge-based decision pipeline for graph fraud detection.
Chinese Translation
关系欺诈检测可以利用无标签的图形上下文和基于标签的邻域证据，但这两种信息源遵循不同的有效性条件。特别是，当查询节点的自身标签或任何验证或测试标签进入其构建时，邻域风险变得无效。我们将这个问题表述为基于来源约束的关系证据使用，并提出了PREF-Gate，一个具有两个固定专家和有限验证门的可审计决策框架。上下文专家使用无标签的属性、一跳均值、特征残差和度描述符。证据专家则添加自我排除的仅训练标签邻域风险和经验贝叶斯摘要，揭示支持、确定性、可用性和收缩。在测试推断之前，门控选择任一专家或三种预设概率混合之一，并固定决策阈值。在Amazon、YelpChi和TFinance上，使用五个相同的分层划分和14种相同协议的方法，PREF-Gate获得了平均AUPRC值为0.9085、0.8104和0.8913。它在所有Amazon和YelpChi划分中选择无标签专家，而在所有TFinance划分中选择证据混合。因此，主要结果是有条件的而非普遍的：基于标签的关系证据仅在持出验证支持的情况下才有用。该框架将竞争排名性能与明确的标签来源合同、有限选择策略、失败核算和审查预算评估相结合，为图形欺诈检测提供了一个可审计的基于知识的决策流程。
cs.AI / 93 / 2607.11226
Heterogeneous Agent Cohorts for Safe Open-Ended Exploration with Runtime Constraint Memory
异质代理群体用于安全的开放式探索与运行时约束记忆
Abstract
LLM agents today are caught in an awkward bind. Lock them down with static safety instructions and they rarely venture beyond the obvious; give them free reign with tools and multi-agent debate, and safety violations quickly follow. Rather than forcing a single model to juggle both creativity and caution, we separate the concerns across specialized roles. A Disrupter generates unconventional proposals, a Validator enforces hard runtime checks at the tool gateway, and a Broker pulls in distant but relevant analogies. Failures are not discarded -- they are compiled, via MCTS, into compact, signed constraint patches we call Scars. These patches are cached locally and inherited by future cohorts, turning repeated failures into reusable, low-cost runtime constraints. In a spatial-semantic sandbox (N=20 runs, p<0.01), our cohort reaches remote targets where debate fails, the Validator prevents all executed breaches, and Scars reduce token consumption by 15.1% by avoiding redundant validator checks. Furthermore, credit-based Communication Allocation Scores (CAS) restrict outbound bandwidth, reducing overall token costs by 55.9% under resource constraints.
Chinese Translation
当前的大型语言模型（LLM）代理面临尴尬的困境。若用静态安全指令将其限制，它们很少超越显而易见的范围；若给予其自由使用工具和多代理辩论的权利，安全违规行为则迅速随之而来。我们并不强迫单一模型同时兼顾创造力和谨慎，而是将这些关注点分散到专门化的角色中。破坏者（Disrupter）生成非常规提案，验证者（Validator）在工具网关处执行严格的运行时检查，调解者（Broker）引入遥远但相关的类比。失败并不被丢弃——它们通过蒙特卡洛树搜索（MCTS）被编译成紧凑的、签名的约束补丁，我们称之为伤痕（Scars）。这些补丁被本地缓存，并由未来的群体继承，将重复的失败转化为可重用的、低成本的运行时约束。在一个空间-语义沙箱中（N=20次实验，p<0.01），我们的群体能够到达辩论失败的遥远目标，验证者阻止了所有执行的违规行为，而伤痕通过避免冗余的验证者检查将令牌消耗降低了15.1%。此外，基于信用的通信分配评分（CAS）限制了外发带宽，在资源约束下将整体令牌成本降低了55.9%。
cs.AI / 94 / 2607.11263
Bringing Back Rule Induction to Fluid Intelligence Research? An Initial Validation of the ARC-AGI Benchmark in Humans
将规则归纳引入流体智力研究？ARC-AGI基准在人的初步验证
Abstract
Two competing perspectives on fluid intelligence (gf) measures propose that performance is primarily constrained either by working memory capacity or by the ability to induce novel relations. The first perspective is currently dominant in measurement, as evident from the use of a limited set of recurring rules, whereas the second perspective is reflected in many definitions but rarely present in measurement. The ARC-AGI benchmark predominantly requires rule induction and was proposed as a measure of gf for both humans and artificial systems. However, its psychometric properties have not yet been examined in human samples. We therefore investigated the psychometric characteristics and nomological network of ARC-AGI in a first study with 100 participants. A compilation of ARC-AGI items showed good psychometric properties and correlated substantially with figural fluid intelligence as measured by a figural reasoning test (\r{ho} = .63). Associations with figural originality were weak. These findings provide initial support for the validity of ARC-AGI as a measure of human fluid intelligence. Future research should include more rule induction tasks as well as additional multivariate covariates. This study is unusual by studying a task in humans that was initially designed for machines. We suggest systematically embedding AI benchmarks into the nomological network of human cognitive abilities to enable more systematic evaluation and interdisciplinary cooperation.
Chinese Translation
关于流体智力（gf）测量的两种竞争观点认为，表现主要受到工作记忆容量或归纳新关系能力的限制。第一种观点在测量中占主导地位，体现在使用有限的一组重复规则，而第二种观点则在许多定义中有所体现，但在测量中很少出现。ARC-AGI基准主要要求规则归纳，并被提出作为人类和人工系统的gf测量。然而，其心理测量特性尚未在人的样本中进行检验。因此，我们在一项包含100名参与者的初步研究中调查了ARC-AGI的心理测量特征和名义网络。ARC-AGI项目的汇编显示出良好的心理测量特性，并与通过图形推理测试测量的图形流体智力显著相关（
{ho} = .63）。与图形原创性的关联较弱。这些发现为ARC-AGI作为人类流体智力测量的有效性提供了初步支持。未来的研究应包括更多的规则归纳任务以及额外的多变量协变量。本研究的独特之处在于研究了一个最初为机器设计的任务在人的表现。我们建议将人工智能基准系统地嵌入人类认知能力的名义网络中，以便进行更系统的评估和跨学科合作。
cs.AI / 95 / 2607.11266
Valid $\ne$ Necessary: Diagnosing Latent Inefficiency in Chain-of-Thought
有效性与必要性：诊断思维链中的潜在低效性
Abstract
Chain-of-Thought (CoT) prompting has significantly advanced the reasoning capabilities of Large Language Models (LLMs), yet it often incurs substantial computational costs due to over-reasoning: the generation of redundant, verbose, or irrelevant steps. While existing reasoning step evaluators effectively detect logical fallacies and factual errors, our analysis reveals a critical blind spot: they fail to penalize valid but inefficient reasoning steps that inflate token usage without contributing to the solution. To systematically diagnose this limitation, we introduce RIV-GSM8K, a diagnostic benchmark injected with five distinct types of inefficiencies, including circular reasoning and excessive decomposition. Diagnostic experiments reveal that state-of-the-art evaluators struggle to distinguish these inefficiencies from necessary reasoning. To address this gap, we propose CAID (Context-Aware Information Density), a training-free metric grounded in information theory that identifies low-utility steps. To validate the metric's practical utility, we apply it within PACE, a post-hoc compression strategy. Additional control experiments show that the gains of PACE are not explained by trivial pruning: compared with random step removal and PRM-based compression baselines, it preserves accuracy at substantially higher compression rates. Empirical results on GSM8K, StrategyQA, and ARC-Challenge demonstrate that PACE reduces token consumption by 31-53% while maintaining accuracy, confirming that CAID successfully distills informational froth from reasoning chains without compromising deductive validity.
Chinese Translation
思维链（Chain-of-Thought, CoT）提示显著提升了大型语言模型（Large Language Models, LLMs）的推理能力，但由于过度推理，常常导致显著的计算成本：生成冗余、冗长或无关的步骤。尽管现有的推理步骤评估器能够有效检测逻辑谬误和事实错误，但我们的分析揭示了一个关键的盲点：它们未能惩罚有效但低效的推理步骤，这些步骤在不贡献解决方案的情况下增加了标记使用量。为了系统性地诊断这一局限性，我们引入了RIV-GSM8K，这是一个注入了五种不同类型低效性的诊断基准，包括循环推理和过度分解。诊断实验表明，最先进的评估器在区分这些低效性与必要推理方面存在困难。为了解决这一问题，我们提出了CAID（Context-Aware Information Density），这是一种基于信息理论的无训练度量，能够识别低效用步骤。为了验证该度量的实际效用，我们将其应用于PACE，这是一种事后压缩策略。额外的对照实验表明，PACE的收益并不能通过简单的修剪来解释：与随机步骤移除和基于PRM的压缩基线相比，它在显著更高的压缩率下保持了准确性。在GSM8K、StrategyQA和ARC-Challenge上的实证结果表明，PACE在保持准确性的同时减少了31-53%的标记消耗，确认CAID成功地从推理链中提炼出信息泡沫，而不妨碍推理的有效性。
cs.AI / 96 / 2607.11307
Efficient Test-Time Optimization for Multi-Agent Proof Autoformalization
多智能体证明自动形式化的高效测试时间优化
Abstract
Full-proof autoformalization bridges extensive mathematical proofs in natural language with formally validated reasoning, offering a pathway to elevate the ceiling of verifiable mathematical reasoning. Unlike statement-level formalization, proof autoformalization is a long-horizon challenge requiring coordination of claims, contexts, and dependencies across many proof steps, yet has only recently come under focused study. Current approaches either rely on costly model training or apply excessive, unguided repair at inference time. To this end, we introduce ToMap, a multi-agent framework that structures proof autoformalization as a Decomposer-Formalizer-Prover pipeline with efficient test-time optimization guided by formal verification and semantic rubrics for proof quality. Rather than distributing test-time compute across all agents, we perform bottleneck analysis and identify the Decomposer as the critical bottleneck: the quality of its atomic, self-contained proof units directly determines whether downstream agents can successfully formalize and prove each step. ToMap therefore treats the Formalizer and Prover as downstream executors and efficiently focuses test-time compute on Decomposer refinement. This refinement follows a loop inspired by GEPA, evolving prompts over candidate decompositions and using formal verification progress together with semantic proof rubrics to define a Pareto frontier that guides the next decomposition update. Experiments on ProofFlowBench show that ToMap improves over the best previous method by 19.0% when evaluated by both syntactic correctness and semantic faithfulness, while requiring lower test-time cost. Scaling analysis shows that most gains emerge within a few iterations of decomposition evolution, guiding test-time budget selection.
Chinese Translation
全证明自动形式化将自然语言中的广泛数学证明与形式验证的推理连接起来，为提升可验证数学推理的上限提供了一条途径。与语句级形式化不同，证明自动形式化是一个长期挑战，需要在多个证明步骤中协调声明、上下文和依赖关系，但直到最近才受到重点研究。目前的方法要么依赖于昂贵的模型训练，要么在推理时进行过度且无指导的修复。为此，我们提出了ToMap，一个将证明自动形式化结构化为分解器-形式化器-证明器管道的多智能体框架，并通过形式验证和语义标准指导的高效测试时间优化。ToMap并不将测试时间计算分配给所有智能体，而是进行瓶颈分析，识别出分解器作为关键瓶颈：其原子、自包含的证明单元的质量直接决定了下游智能体能否成功形式化和证明每个步骤。因此，ToMap将形式化器和证明器视为下游执行者，并高效地将测试时间计算集中在分解器的改进上。该改进遵循一个受GEPA启发的循环，围绕候选分解演变提示，并利用形式验证进展和语义证明标准来定义一个帕累托前沿，以指导下一个分解更新。在ProofFlowBench上的实验表明，ToMap在语法正确性和语义忠实性评估中比之前最佳方法提高了19.0%，同时要求更低的测试时间成本。规模分析表明，大多数收益在几次分解演变迭代内出现，为测试时间预算选择提供了指导。
cs.AI / 97 / 2607.11317
Calibrated e-CUSUM Decoding for Quantized Reasoning Models: Why Token Log-Probability Is the Wrong Observable for Decoding Monitors
量化推理模型的校准e-CUSUM解码：为何令牌对数概率不是解码监控的合适可观测量
Abstract
Low-bit quantization makes small reasoning models inexpensive to deploy but can degrade their chains of thought. This motivates decoder-side monitors that intervene when generation becomes unreliable. We show that a natural candidate, the centered token log-probability increment $\log p(w_t)+H_t$, is the wrong observable for this purpose. Under the model's own sampling law it is a mean-zero martingale by construction, so it measures sampling self-consistency rather than trajectory health and is nearly silent during confident repetition, where both $\log p(w_t)$ and entropy are close to zero. We introduce a training-free decoding controller that combines (i) a degeneration-aware alarm score fusing token uncertainty with explicit verbatim repetition and (ii) a calibrated e-process-inspired sequential detector. The raw product process is Ville-valid under a conditional-mean null, while the deployed CUSUM-floored statistic is treated as an empirical change detector because the score is history-dependent and autocorrelated. On GSM8K with DeepSeek-R1-Distill-Qwen-1.5B in FP16 and INT4, calibration turns a monitor that fires on 93--95% of generations into a selective detector of failing traces ($\phi \approx 0.3$, precision $\approx 0.6$ against a 0.38 base rate). In this pilot, the controller reduces measured verbatim-degeneration signals and yields a positive but statistically inconclusive INT4 accuracy change from 63% to 69% (paired McNemar $p=0.18$, $n=100$), at a 28% token-budget cost. We also find that non-termination, rather than looping, is the dominant failure mode on GSM8K. The main contribution is methodological: an explanation of why centered token log-probability is inadequate for decoder monitoring and a calibrated, cautiously evaluated replacement.
Chinese Translation
低位量化使得小型推理模型的部署成本低廉，但可能会降低其思维链的质量。这促使我们在生成变得不可靠时引入解码侧监控。我们展示了一个自然候选者，即中心化的令牌对数概率增量 $ ext{log} p(w_t) + H_t$，并不适合此目的。根据模型自身的采样法则，它在构造上是一个均值为零的马尔可夫链，因此它测量的是采样自一致性而非轨迹健康，并且在自信重复期间几乎没有信号，此时 $ ext{log} p(w_t)$ 和熵都接近于零。我们引入了一种无训练的解码控制器，该控制器结合了（i）一个关注退化的警报分数，将令牌不确定性与明确的逐字重复融合在一起，以及（ii）一个受校准的e过程启发的序列检测器。在条件均值零假设下，原始产品过程是Ville有效的，而部署的CUSUM底部统计量被视为经验变化检测器，因为该分数依赖于历史且具有自相关性。在使用FP16和INT4的DeepSeek-R1-Distill-Qwen-1.5B模型进行GSM8K测试时，校准将一个在93%到95%生成时触发的监控器转变为一个选择性检测失败轨迹的检测器（$ heta ext{approx} 0.3$，精确度 $ ext{approx} 0.6$，基线率为0.38）。在这一试点中，控制器减少了测量的逐字退化信号，并在28%的令牌预算成本下，带来了从63%到69%的正向但统计上不确定的INT4准确率变化（配对McNemar $p=0.18$，$n=100$）。我们还发现，在GSM8K上，非终止而非循环是主要的失败模式。主要贡献在于方法论：解释了为何中心化的令牌对数概率不足以用于解码监控，并提出了一个经过校准、谨慎评估的替代方案。
cs.AI / 98 / 2607.11334
Verifier-Guided Twelve-Tone Composition: A Generate-Verify-Repair Harness for Symbolic Music Generation
验证者引导的十二音作曲：一种用于符号音乐生成的生成-验证-修复框架
Abstract
Large language models can produce superficially legal twelve-tone scores that collapse into degenerate textures. We introduce a neuro-symbolic harness that wraps a language-model proposer in a generate-verify-repair-trace loop with symbolic verification. The complete pipeline improves event-local consistency without claiming whole-piece legality. Across 40 controlled tasks and four paired models, audited delivery yield rises from 13.3% under raw generation to 48.1% with the harness, which explicitly abstains otherwise. The pass rate of a narrower collision and serialisation-consistency check rises from 33.5% to 58.3%, while degeneracy remains near 0.05, including under exploratory adversarial prompting. A blinded evaluation by five experts also shows a descriptive aggregate preference for harness candidates over raw generation in adherence, perceived legality, coherence, and overall quality.
Chinese Translation
大型语言模型能够生成表面上合法的十二音乐谱，但往往会陷入退化的纹理中。我们引入了一种神经符号框架，该框架将语言模型提议者包裹在一个生成-验证-修复-追踪的循环中，并进行符号验证。完整的流程在不声称整体作品合法性的情况下，提高了事件局部一致性。在40个受控任务和四个配对模型中，审核交付率从原始生成的13.3%上升至使用该框架后的48.1%，而该框架明确避免了其他情况。更严格的碰撞和序列一致性检查的通过率从33.5%上升至58.3%，而退化率保持在接近0.05，包括在探索性对抗提示下。五位专家的盲评也显示，框架候选在遵循性、感知合法性、一致性和整体质量方面相较于原始生成具有更高的描述性总体偏好。
cs.AI / 99 / 2607.11338
AutoVSR: Automatic Visual-to-Symbolic Reasoning for Symbolic Expression Generation from Circuit Schematic
AutoVSR：从电路原理图生成符号表达式的自动视觉到符号推理
Abstract
Symbolic expressions can effectively characterize and predict circuit behavior, but deriving them directly from circuit schematics is challenging. This process requires accurate visual-to-symbolic construction of circuit structure from images and correct multi-step symbolic derivation, both of which impose strict correctness requirements. This work proposes AutoVSR, an automated framework for visual-to-symbolic generation of circuit expressions using Vision Language Models (VLMs). By reconstructing circuit diagrams into an executable intermediate representation (Executable IR) and leveraging a symbolic solver for reasoning, AutoVSR significantly improves the accuracy of symbolic expression generation. AutoVSR introduces two key innovations: an IR construction method guided by component rule retrieval and verification-based feedback, and a symbolic solver implemented as a planning agent equipped with a symbolic tool library for reliable multi-step derivation. Compared with end-to-end VLM approaches and specialized methods on the main symbolic expression generation task, AutoVSR achieves accuracy improvements of 30.01--59.45% and 41.96--51.84%, respectively. Moreover, AutoVSR surpasses closed-source state-of-the-art VLMs in inference cost and computational efficiency. Code is available at https://github.com/LongfeiLi1/AutoVSR.
Chinese Translation
符号表达式可以有效地表征和预测电路行为，但直接从电路原理图推导它们具有挑战性。该过程需要从图像中准确地进行电路结构的视觉到符号构建，以及正确的多步骤符号推导，这两者都对正确性提出了严格要求。本研究提出了AutoVSR，一个基于视觉语言模型（Vision Language Models, VLMs）的电路表达式视觉到符号生成的自动化框架。通过将电路图重构为可执行的中间表示（Executable IR），并利用符号求解器进行推理，AutoVSR显著提高了符号表达式生成的准确性。AutoVSR引入了两个关键创新：一种由组件规则检索和验证反馈指导的IR构建方法，以及作为规划代理实现的符号求解器，配备有符号工具库以支持可靠的多步骤推导。与端到端的VLM方法和在主要符号表达式生成任务上的专门方法相比，AutoVSR在准确性上分别提高了30.01%至59.45%和41.96%至51.84%。此外，AutoVSR在推理成本和计算效率上超越了闭源的最先进VLM。代码可在 https://github.com/LongfeiLi1/AutoVSR 获取。
cs.AI / 100 / 2607.11346
Compile, Then Page: Executable SOP Programs and a Capability-Gated Runtime for Procedural LLM Agents
编译后再分页：可执行的标准操作程序（SOP）程序与面向能力的运行时环境用于程序性大型语言模型（LLM）代理
Abstract
Enterprise agents must follow long-horizon, conditional, safety-critical standard operating procedures (SOPs). We compile machine-readable SOP constraints into executable pseudo-code and run them with a program-guided (PG) stack machine that pages the active frame while an LLM performs semantic execution. A three-arm SOPBench study across six models separates representation from runtime: compiled text never significantly hurts and gains up to 16.0 points where official prose underperforms. Runtime guidance is capability-gated. Two strong models independently show positive seven-domain PG contrasts (58:19 and 75:31 discordant pairs), whereas weak models are harmed. A full-program cursor ablation (active frame first, complete program retained) recovers much of the strong-model refusal gain; selective visibility adds a smaller improvement. Paired probe and audit measurements track this divide to spontaneous state discipline rather than reconstruction ability. On Bank the three primary arms rise from 70.4 to 86.4 to 92.8, with 100% refusal correctness. Practical guidance: compile first; enable active-frame paging only after a model-level discipline check.
Chinese Translation
企业代理必须遵循长时间跨度、条件性和安全关键的标准操作程序（SOP）。我们将机器可读的SOP约束编译为可执行的伪代码，并通过一个程序引导（PG）栈机器运行这些代码，该机器在大型语言模型（LLM）执行语义操作时分页活动帧。针对六个模型的三臂SOPBench研究将表示与运行时分离：编译文本从未显著降低性能，并在官方文本表现不佳时获得高达16.0的分数提升。运行时指导是基于能力的。两个强模型独立显示出七个领域的积极PG对比（58:19和75:31的不一致对），而弱模型则受到损害。完整程序光标消融（优先处理活动帧，保留完整程序）恢复了强模型拒绝增益的大部分；选择性可见性带来了较小的改善。配对探测和审计测量跟踪这一差异，归因于自发状态的纪律而非重建能力。在Bank上，三个主要臂的分数从70.4上升到86.4，再到92.8，拒绝正确率达到100%。实践指导：先编译；仅在模型级纪律检查后启用活动帧分页。
cs.AI / 101 / 2607.11347
From Neural Network Decisions to Training Cases: An Exact Account via Case-Based Decision Theory
从神经网络决策到训练案例：通过基于案例的决策理论的精确解释
Abstract
Neural networks increasingly guide decisions in high-stakes domains such as medical diagnosis, credit approval, and energy bidding. Audit in these settings requires case-level evidence: which training cases support an action and what outcomes they carried. Case-based decision theory (CBDT) formalizes this reasoning by aggregating outcome support from remembered cases. We show that an OLS action readout fitted on a fixed neural representation admits an exact case-based decomposition. Each action score is a weighted sum of training-case returns, with coefficients determined by empirical Gram geometry. We identify a sufficient regime for CBDT similarity semantics; outside it, the coefficients should generally be treated as signed Gram-geometric influence. The decomposition yields audit signals that trace scores to training cases, measure action coherence, and identify weak support. Across synthetic CBDT, PJM, Adult Income, and Default Credit tasks, the method recovers case-level preference structure and achieves the highest mean Top-30 consistency among compared attribution baselines, while remaining competitive on support reconstruction. The audit requires only fitting an OLS top-layer probe, without retraining the representation or accessing the original optimization trajectory; probe fidelity is measured by score reconstruction.
Chinese Translation
神经网络越来越多地在医疗诊断、信用审批和能源竞标等高风险领域指导决策。在这些环境中，审计需要案例级的证据：哪些训练案例支持某一行动以及它们所带来的结果。基于案例的决策理论（CBDT）通过聚合来自记忆案例的结果支持来形式化这种推理。我们展示了一个在固定神经表征上拟合的OLS（普通最小二乘法）行动读出允许进行精确的基于案例的分解。每个行动得分是训练案例回报的加权和，系数由经验Gram几何确定。我们识别出CBDT相似性语义的充分范围；在此范围之外，系数通常应被视为带符号的Gram几何影响。该分解产生审计信号，将得分追溯到训练案例，测量行动一致性，并识别弱支持。在合成CBDT、PJM、成人收入和违约信用任务中，该方法恢复了案例级的偏好结构，并在比较的归因基线中实现了最高的平均Top-30一致性，同时在支持重建方面保持竞争力。审计只需拟合一个OLS顶层探针，无需重新训练表征或访问原始优化轨迹；探针的保真度通过得分重建进行测量。
cs.AI / 102 / 2607.11357
OpsMem: Dual-Memory Reasoning with Cross-Memory Resonance for Failure Diagnosis
OpsMem：基于跨记忆共振的双记忆推理用于故障诊断
Abstract
Failure diagnosis in modern software systems requires iterative evidence acquisition and hypothesis reasoning guided by operational experience. Existing LLM-based methods improve diagnosis through agentic reasoning or knowledge augmentation, but they often lack a mechanism to coordinate the evolving diagnostic state with operational experience during iterative diagnosis. We propose OpsMem, a dual-memory framework that maintains a short-term memory for the current diagnostic state and a long-term memory for reusable operational experience. OpsMem uses cross-memory resonance to activate state-relevant long-term memory, conditions multi-agent diagnosis on the short-term and activated long-term memories, and consolidates reusable experience from solved incidents back into long-term memory. Experiments on a real-world Huawei microservice failure diagnosis dataset show that OpsMem outperforms representative agentic-reasoning and knowledge-augmented baselines, improving Match and Relevant by up to 46.88% and 18.39% over the strongest baseline, respectively.
Chinese Translation
现代软件系统中的故障诊断需要通过操作经验指导的迭代证据获取和假设推理。现有的基于大型语言模型（LLM）的方法通过代理推理或知识增强来改善诊断，但在迭代诊断过程中，它们通常缺乏协调不断演变的诊断状态与操作经验的机制。我们提出了OpsMem，一个双记忆框架，维护当前诊断状态的短期记忆和可重用操作经验的长期记忆。OpsMem利用跨记忆共振激活与状态相关的长期记忆，将多智能体诊断条件于短期和激活的长期记忆，并将已解决事件的可重用经验整合回长期记忆。对真实世界华为微服务故障诊断数据集的实验表明，OpsMem在代表性的代理推理和知识增强基线之上表现更佳，Match和Relevant分别提高了高达46.88%和18.39%。
cs.AI / 103 / 2607.11388
StructAgent: Harness Long-horizon Digital Agents with Unified Causal Structure
StructAgent：利用统一因果结构驱动长时间跨度的数字代理
Abstract
Recent advances in large language models (LLMs) and vision-language models (VLMs) have enabled increasingly capable digital agents for computer use. However, real-world tasks are often long-horizon and involve evolving contexts containing accumulated observations, intermediate edits, failed attempts, and partially completed executions. Existing agents typically operate over raw interaction history, making task progress difficult to interpret, verify, and recover, which ultimately limits reliable long-horizon execution. In this paper, we argue that addressing this challenge requires explicitly structuring both the agent's state and workflow around a unified causal representation of task progress. We present \textbf{StructAgent}, a state-centered framework that introduces a unified state for maintaining compact, verifiable task progress and a structured workflow that regulates progress through verifier-backed state transitions. Building on this design, StructAgent further enables explicit progress checkpointing, evidence-driven task completion, targeted failure recovery, and tool-supported execution, while ensuring that all progress updates remain grounded in verification. Extensive experiments demonstrate that StructAgent consistently improves a wide range of LLM and VLM backbones on long-horizon computer-use tasks. On OSWorld-Verified, it improves Qwen3.5-9B from 27.0\% to 46.9\% success rate and Qwen3.5-27B from 31.6\% to 62.2\%, while achieving a new open-source state of the art of 78.9\% with MiniMax-M3. Moreover, the same framework generalizes beyond desktop environments to Minecraft, demonstrating the generality of our design.
Chinese Translation
近年来，大型语言模型（LLMs）和视觉语言模型（VLMs）的进展使得数字代理在计算机使用方面变得越来越强大。然而，现实世界的任务往往是长时间跨度的，并涉及不断演变的上下文，包括累积的观察、阶段性编辑、失败尝试和部分完成的执行。现有的代理通常在原始交互历史上操作，这使得任务进展难以解释、验证和恢复，最终限制了可靠的长时间跨度执行。本文认为，解决这一挑战需要明确围绕任务进展的统一因果表示来构建代理的状态和工作流程。我们提出了 extbf{StructAgent}，这是一个以状态为中心的框架，介绍了一种统一状态，用于维护紧凑、可验证的任务进展，以及一种结构化工作流程，通过验证者支持的状态转换来调节进展。在此设计的基础上，StructAgent进一步实现了显式的进展检查点、基于证据的任务完成、针对性的失败恢复和工具支持的执行，同时确保所有进展更新都基于验证。大量实验表明，StructAgent在长时间跨度的计算机使用任务上持续改进了多种LLM和VLM骨干网络。在OSWorld-Verified上，它将Qwen3.5-9B的成功率从27.0\%提高到46.9\%，将Qwen3.5-27B的成功率从31.6\%提高到62.2\%，同时在MiniMax-M3上实现了78.9\%的新开源状态的最佳表现。此外，同一框架在Minecraft等桌面环境之外也具有泛化能力，展示了我们设计的普适性。
cs.AI / 104 / 2607.11433
Omni-Decision: A Progressive Evidence-State Agent System for Omni-Modal QA
全决策：一种渐进的证据状态代理系统用于全模态问答
Abstract
Omni-modal evidence-seeking QA requires agents to answer questions whose evidence is sparsely distributed across videos, audio, images, web pages, and computation results. Existing agentic multimodal systems often leave evidence in scratchpads, tool trajectories, or free-form histories, making it difficult to track what has been grounded, what remains missing, and when the evidence is sufficient to answer. We propose Omni-Decision, a training-free evidence-state system that turns omni-modal QA into a query-scoped evidence-closure process. For each query, Omni-Decision maintains a structured evidence state containing confirmed evidence, unresolved conflicts, fact and computation dependencies, and open evidence needs. A shared state view conditions planning, evidence acquisition, validation, repair, and finalization. Heterogeneous observations from media, web, computation, and verification modules are normalized, judged, and committed through deterministic state updates. This design enables targeted evidence acquisition, preserves sparse cross-modal cues, and provides inspectable control over repair and stopping. Omni-Decision achieves 45.6% accuracy on OmniGAIA and 58.3% on WorldSense, improving over the baselines by +27.3 and +30.2 percentage points, respectively. No-state ablations and trajectory audits further support the role of explicit evidence-state control in multi-step omni-modal evidence seeking.
Chinese Translation
全模态证据寻求问答要求代理能够回答证据分散在视频、音频、图像、网页和计算结果中的问题。现有的代理多模态系统通常将证据留在草稿纸、工具轨迹或自由形式的历史记录中，这使得追踪已确认的证据、尚未解决的冲突以及何时证据足以回答问题变得困难。我们提出了全决策（Omni-Decision），这是一种无训练的证据状态系统，将全模态问答转变为一个查询范围的证据闭合过程。对于每个查询，全决策维护一个结构化的证据状态，其中包含已确认的证据、未解决的冲突、事实和计算依赖关系，以及开放的证据需求。共享状态视图为规划、证据获取、验证、修复和最终确定提供条件。来自媒体、网络、计算和验证模块的异构观察通过确定性状态更新进行规范化、判断和提交。这一设计使得有针对性的证据获取成为可能，保留稀疏的跨模态线索，并提供可检查的修复和停止控制。全决策在OmniGAIA上实现了45.6%的准确率，在WorldSense上实现了58.3%的准确率，分别比基线提高了27.3和30.2个百分点。无状态消融和轨迹审计进一步支持了显式证据状态控制在多步骤全模态证据寻求中的作用。
cs.AI / 105 / 2607.11436
The Ebb and Flow of Multimodal Focus: Scheduling Visual Relay Windows for Grounded VLM Reasoning
多模态聚焦的潮起潮落：为基础视觉语言模型推理调度视觉中继窗口
Abstract
Vision-language models increasingly succeed on multimodal reasoning benchmarks, yet their visual evidence often becomes unstable once it enters the language stack, weakening evidence-grounded reasoning. To understand this fragility, we examine the internal dynamics of VLMs through a mechanistic lens and uncover a stable three-stage redistribution of multimodal attention focus across depth: an early question-conditioned organization, a critical middle visual-dominant relay, and a late return to answer formation. We operationalize the middle phase as the Visual Relay Window (VRW), and show that its geometry varies with task demand, is causally tied to grounded generation, and distinguishes unsupported answers from stronger reasoning trajectories. Guided by this internal rhythm, we propose TRACE, a task-adaptive inference-time control framework with lightweight trained modules. It reshapes relay allocation during prefill and preserves assembled visual support after handoff during decoding. Across four open-weight VLM backbones and seven benchmarks, TRACE delivers large gains on grounding-sensitive settings, improving them by 4.33 points on average and by up to 6.6 points, while also improving reasoning-heavy tasks. These results show that explicitly controlling multimodal focus across depth offers a unified and effective mechanism for strengthening evidence-grounded multimodal reasoning.
Chinese Translation
视觉语言模型在多模态推理基准测试中越来越成功，但一旦视觉证据进入语言堆栈，其稳定性往往会降低，从而削弱基于证据的推理。为了理解这种脆弱性，我们通过机械视角考察视觉语言模型的内部动态，发现多模态注意力聚焦在深度上的稳定三阶段重分配：早期的问题条件组织、关键的中间视觉主导中继，以及晚期的答案形成回归。我们将中间阶段操作化为视觉中继窗口（Visual Relay Window, VRW），并展示其几何形状随着任务需求而变化，与基于证据的生成有因果关系，并能够区分不支持的答案与更强的推理轨迹。在这一内部节奏的指导下，我们提出了TRACE，一个具有轻量级训练模块的任务自适应推理时控制框架。它在预填充过程中重塑中继分配，并在解码时保持组装的视觉支持。通过四个开放权重的视觉语言模型骨干和七个基准测试，TRACE在对基础敏感的设置中实现了显著提升，平均提高了4.33分，最高可达6.6分，同时也改善了推理密集型任务。这些结果表明，明确控制深度上的多模态聚焦为增强基于证据的多模态推理提供了统一而有效的机制。
cs.AI / 106 / 2607.11492
Enhancing Query Efficiency for d-DNNF Representations Through Preprocessing
通过预处理提升 d-DNNF 表示的查询效率
Abstract
In this paper, we investigate preprocessing techniques aimed at improving the efficiency of accessing models of propositional formulas represented in conjunctive normal form (CNF). We focus on three fundamental tasks: uniform sampling, direct model access, and model enumeration. Our analysis reveals that most state-of-the-art preprocessors, when they do not preserve formula equivalence, are generally unsuitable for these tasks. In contrast, we demonstrate that preprocessors which preserve model counts can be effectively leveraged, provided relevant preprocessing information is maintained. To validate our approach, we perform extensive experiments on a diverse suite of benchmarks from multiple domains. The experimental results show that our preprocessing methods are both efficient and robust, yielding significant performance improvements for model access queries when CNF formulas are compiled into d-DNNF representations.
Chinese Translation
在本文中，我们研究了旨在提高访问以合取范式（CNF）表示的命题公式模型效率的预处理技术。我们重点关注三个基本任务：均匀采样、直接模型访问和模型枚举。我们的分析表明，大多数最先进的预处理器在不保持公式等价性的情况下，通常不适合这些任务。相反，我们证明了保持模型计数的预处理器可以有效利用，前提是维护相关的预处理信息。为了验证我们的方法，我们在多个领域的多样化基准测试中进行了广泛的实验。实验结果表明，我们的预处理方法既高效又稳健，在将 CNF 公式编译为 d-DNNF 表示时，为模型访问查询带来了显著的性能提升。
cs.AI / 107 / 2607.11501
Comparative Analysis of GAT and BERT for Human-Like Playtesting
GAT与BERT在人类游戏测试中的比较分析
Abstract
Accurately modeling and understanding player experience is crucial for designing engaging puzzle games. To achieve this, a common approach involves collecting diverse user data to train predictive playtesting models that mimic player behavior. However, existing data-driven methods often lack the ability to capture the full range of player strategies and require extensive feature engineering and network architecture modeling. This limitation becomes particularly evident when new game mechanics or features are introduced, which necessitate continual adjustments to the models. To addrss these challenges, we propose a more generalized representation that reduces - or even eliminates - the need for ongoing feature-engineering maintenance. Specifically, we investigate two general-purpose network architectures: (a) a transformer-based model (BERT) and (b) a graph attention model (GAT), both of which are designed to effectively capture the relational structure of Candy Crush Saga (CCS) game boards. Our experiments compare these approaches to Convolutional Neural Networks (CNN) baselines, revealing better performance on challenging board configurations and underscoring the benefits of our generalizable representation.
Chinese Translation
准确建模和理解玩家体验对于设计引人入胜的益智游戏至关重要。为此，一种常见的方法是收集多样的用户数据，以训练预测性游戏测试模型，从而模拟玩家行为。然而，现有的数据驱动方法往往无法捕捉到玩家策略的全貌，并且需要大量的特征工程和网络架构建模。这一局限性在引入新游戏机制或特性时尤为明显，这需要对模型进行持续的调整。为了解决这些挑战，我们提出了一种更为通用的表示方法，减少甚至消除对持续特征工程维护的需求。具体而言，我们研究了两种通用网络架构：(a) 基于变换器的模型（BERT）和 (b) 图注意力模型（GAT），这两者均旨在有效捕捉《糖果传奇》（Candy Crush Saga，CCS）游戏棋盘的关系结构。我们的实验将这些方法与卷积神经网络（CNN）基线进行了比较，结果显示在具有挑战性的棋盘配置上表现更佳，并强调了我们可推广表示的优势。
cs.AI / 108 / 2607.11530
Learning Residual Kinematic Corrections for Continuous Neural Decoding via Reinforcement Learning
通过强化学习学习连续神经解码的残余运动学修正
Abstract
Decoding continuous three-dimensional (3D) motor imagery (MI) using non-invasive electroencephalography (EEG)-based brain--computer interfaces (BCIs) remains challenging due to signal variability and residual decoding errors. Deep learning architectures such as convolutional neural network--long short-term memory (CNN--LSTM) models can capture spatial and temporal dynamics for continuous kinematic decoding; however, systematic residual errors persist in predicted trajectories. We propose a two-stage decoding framework that applies reinforcement learning (RL) to perform residual kinematic correction on the outputs of a CNN--LSTM decoder (CNN--LSTM--RL). The RL agent is trained offline without direct EEG input and instead operates on predicted kinematic trajectories to optimize movement accuracy relative to target trajectories. Decoding performance was quantified using Pearson correlation coefficients ($r$) and Root Mean Square Errors (RMSE) along the $x, y$, and $z$ axes. Compared to CNN--LSTM applied alone, CNN--LSTM--RL improved the mean correlation from $0.5076$ to $0.7181$ ($p = 0.0005$) in 2D and from $0.6420$ to $0.7780$ ($p = 0.0059$) in VR, with relative gains of $41.5\%$ and $21.2\%$, respectively. Correspondingly, RMSE was reduced from $0.0890$ to $0.0532$ (2D, $p < 0.0001$) and from $0.0714$ to $0.0441$ (VR, $p < 0.0001$), representing relative reductions of $40.2\%$ and $38.2\%$. These findings demonstrate that this scalable framework enhances 3D BCI MI decoding by correcting kinematic errors via offline residual RL without extra neural data, advancing neurorehabilitation, prosthetics, and virtual interaction.
Chinese Translation
利用非侵入性脑电图（EEG）基础的脑-计算机接口（BCI）解码连续三维（3D）运动想象（MI）仍然具有挑战性，主要由于信号变异性和残余解码误差。深度学习架构，如卷积神经网络-长短期记忆（CNN-LSTM）模型，能够捕捉连续运动学解码的空间和时间动态；然而，预测轨迹中仍然存在系统性的残余误差。我们提出了一种两阶段解码框架，应用强化学习（RL）对CNN-LSTM解码器（CNN-LSTM-RL）输出进行残余运动学修正。RL代理在没有直接EEG输入的情况下离线训练，而是基于预测的运动学轨迹优化相对于目标轨迹的运动准确性。解码性能通过皮尔逊相关系数（$r$）和根均方误差（RMSE）在$x, y$和$z$轴上进行量化。与单独应用CNN-LSTM相比，CNN-LSTM-RL在2D中的平均相关性从$0.5076$提高到$0.7181$（$p = 0.0005$），在虚拟现实（VR）中从$0.6420$提高到$0.7780$（$p = 0.0059$），相对增益分别为$41.5\%$和$21.2\\%$。相应地，RMSE在2D中从$0.0890$减少到$0.0532$（$p < 0.0001$），在VR中从$0.0714$减少到$0.0441$（$p < 0.0001$），分别代表$40.2\\%$和$38.2\\%$的相对减少。这些发现表明，该可扩展框架通过离线残余RL修正运动学误差，增强了3D BCI MI解码，推动了神经康复、假肢和虚拟交互的发展。
cs.AI / 109 / 2607.11586
HCRMap: Pressure-Aware Hot-Expert Residency Mapping for 3.5D MoE Chiplet Inference
HCRMap：基于压力感知的热专家驻留映射用于3.5D MoE芯片推理
Abstract
Mixture-of-Experts (MoE) large language models (LLM) activate only a small number of experts during inference, but token routing introduces persistent expert hotness skew: a small set of hot experts continuously receives most tokens, while the remaining experts are lightly loaded. On 3.5D multi-chiplet systems, this skew not only causes compute imbalance but also amplifies pressure on communication, memory bandwidth, I/O, and execution queues. Therefore, the core problem is not simply to reduce token movement, but to dynamically place and reuse hot expert replicas across different memory tiers. This paper proposes HCRMap, a hot expert residency mapping framework for pressure-aware expert replica management in 3.5D MoE inference. Based on expert hotness, weight loading cost, migration overhead, and runtime resource pressure, HCRMap dynamically determines which experts should be promoted, retained, demoted, or evicted. It then maps routed token groups to suitable resident replicas, thereby jointly mitigating communication, memory, and queue bottlenecks. Experimental results show that HCRMap reduces end-to-end latency by 43.6% and 43.0% over Hydra in the prefill and decode stages, respectively; by 34.5% and 33.1% over MoEntwine; and by 46.7% and 46.0% over PIMoE.
Chinese Translation
混合专家（Mixture-of-Experts, MoE）大型语言模型（Large Language Models, LLM）在推理过程中仅激活少量专家，但令牌路由引入了持续的专家热度偏斜：一小部分热专家持续接收大多数令牌，而其余专家则负载较轻。在3.5D多芯片系统中，这种偏斜不仅导致计算不平衡，还加剧了通信、内存带宽、输入/输出和执行队列的压力。因此，核心问题不仅仅是减少令牌移动，而是动态地在不同内存层次中放置和重用热专家副本。本文提出了HCRMap，一个用于3.5D MoE推理中压力感知专家副本管理的热专家驻留映射框架。基于专家热度、权重加载成本、迁移开销和运行时资源压力，HCRMap动态确定哪些专家应被提升、保留、降级或驱逐。然后，它将路由的令牌组映射到合适的驻留副本，从而共同缓解通信、内存和队列瓶颈。实验结果表明，HCRMap在预填充和解码阶段分别比Hydra减少了43.6%和43.0%的端到端延迟；比MoEntwine减少了34.5%和33.1%；比PIMoE减少了46.7%和46.0%。
cs.AI / 110 / 2607.11594
MAGIC: Transition-Aware Generation of Navigable Multi-Scene Game Worlds with Large Language Models
MAGIC：基于过渡感知的大型语言模型生成可导航的多场景游戏世界
Abstract
Multi-scene navigation (clearing an objective in one bounded space and then crossing a portal into the next) is a defining feature of contemporary 3D games, but authoring it is laborious: every portal must have consistent endpoints on both sides, each interior must remain navigable once it is furnished, and the resulting connectivity must be kept consistent across many files. Recent large language model (LLM) and multimodal LLM (MLLM) scene generators have made single-interior synthesis dramatically cheaper, yet they produce one scene at a time and cannot, by naive repetition, yield a connected multi-scene world. We identify three obstacles that single-scene methods leave unsolved: cross-scene consistency, in-scene navigability, and the evaluation of whether a transition actually works. We present MAGIC, a prompt-to-project system that addresses all three. MAGIC is a four-stage pipeline that turns a single natural-language prompt into a runnable multi-scene game project: it plans a shared transition-aware intermediate representation, specifies each scene while enforcing portal reachability with a flood-fill validator, generates the scenes together with their transition scripts, and combines them into one project. Because existing single-scene fidelity metrics never execute a transition, we further introduce a transition-focused evaluation agent that runs each transition in play. On a new benchmark of 100 multi-scene cases, MAGIC produces an executable project for every case and reaches 0.99 precision, 0.95 recall, and 0.96 F1 on end-to-end transition identification; stage by stage, it recovers more ground-truth portals and yields markedly more navigable layouts than an LLM baseline and Holodeck. Our code is available at https://github.com/sereneee1201/MAGIC/.
Chinese Translation
多场景导航（在一个有限空间内完成目标后通过传送门进入下一个空间）是当代3D游戏的一个重要特征，但其创作过程繁琐：每个传送门必须在两侧具有一致的端点，每个内部空间在布置后必须保持可导航性，并且生成的连通性必须在多个文件中保持一致。近期的大型语言模型（LLM）和多模态LLM（MLLM）场景生成器使得单一内部空间的合成成本大幅降低，然而它们一次只能生成一个场景，无法通过简单重复生成一个连通的多场景世界。我们识别出单场景方法未能解决的三个障碍：跨场景一致性、场景内可导航性，以及评估过渡是否有效。我们提出了MAGIC，一个从提示到项目的系统，解决了这三大问题。MAGIC是一个四阶段的流程，将单一自然语言提示转化为可运行的多场景游戏项目：它规划一个共享的过渡感知中间表示，指定每个场景并通过洪水填充验证器强制传送门的可达性，生成场景及其过渡脚本，并将它们组合成一个项目。由于现有的单场景保真度指标从未执行过渡，我们进一步引入了一个专注于过渡的评估代理，在游戏中运行每个过渡。在一个包含100个多场景案例的新基准测试中，MAGIC为每个案例生成了一个可执行项目，并在端到端过渡识别上达到了0.99的精确率、0.95的召回率和0.96的F1分数；逐阶段地，它恢复了更多的真实传送门，并且生成的可导航布局显著优于LLM基线和Holodeck。我们的代码可在https://github.com/sereneee1201/MAGIC/获取。
cs.AI / 111 / 2607.11598
Interaction Scaling: Grounding the Third Axis of Test-Time Compute
交互缩放：为测试时计算的第三个轴奠基
Abstract
There are two standard ways to spend more compute at test time: let a model reason longer, or sample more attempts and keep one. Both share a hidden limit: they are internal. Every extra token comes from the same frozen weights and the same prompt, so neither can tell the model anything it does not already know. We study a third way, interaction: the model proposes an artifact, an external instrument observes how it actually behaves, and the model revises. Each cycle imports a real observation, so interaction breaks through the ceiling the other two hit. We argue that a single variable governs this third axis, grounding, and that it must hold on both sides of the loop. The feedback that drives revision must come from an instrument that actually observes the flaw, and so must the metric that scores the result. On hard coding tasks at a fixed token budget, reasoning-only and best-of-N sampling both plateau (the latter even when an oracle picks the best sample), while every interaction strategy keeps improving; our proposer-reviewer harness reaches a perfect 100% pass rate with no run-to-run variance, and the gain holds across three model families. On rendered visual artifacts, the usual judge (a vision-language model, or VLM, reading a screenshot) rates 14 of 15 visibly broken figures "perfect," because the screenshot hides the flaws before the judge can see them. A tool that measures the real layout instead shows the loop removing 40-74% of defects across four modalities; and that same VLM, used as the reviewer, makes slide layouts worse where the measuring tool repairs them. Interaction scaling is real and distinct from reasoning and sampling, but only visible when both the feedback and the metric are grounded.
Chinese Translation
在测试时增加计算能力的标准方法有两种：让模型推理更长时间，或进行更多尝试并保留一个。这两种方法都有一个隐含的限制：它们是内部的。每个额外的标记都来自相同的冻结权重和相同的提示，因此它们无法告诉模型任何它尚不知道的东西。我们研究了一种第三种方法，即交互：模型提出一个工件，外部工具观察其实际行为，然后模型进行修正。每个循环引入一个真实的观察，因此交互突破了前两者所遇到的瓶颈。我们认为，单一变量支配这一第三个轴，即基础，并且它必须在循环的两侧都成立。驱动修正的反馈必须来自一个实际观察到缺陷的工具，评估结果的指标也必须如此。在固定标记预算下的困难编码任务中，仅推理和最佳N次采样都达到了瓶颈（后者即使在神谕选择最佳样本时也如此），而每种交互策略都持续改进；我们的提议-审查者框架在没有运行间方差的情况下达到了完美的100%通过率，并且这一增益在三种模型家族中保持一致。在渲染的视觉工件中，通常的评判者（一个视觉-语言模型，或VLM，阅读屏幕截图）将15个明显损坏的图形中有14个评为“完美”，因为屏幕截图在评判者看到缺陷之前隐藏了它们。一个测量真实布局的工具则显示出在四种模式下去除了40-74%的缺陷；而同样的VLM作为审查者在测量工具修复的幻灯片布局中使其变得更糟。交互缩放是真实的，并且与推理和采样不同，但只有在反馈和指标都得到基础时才可见。
cs.AI / 112 / 2607.11607
Auditing the Risk Claims of Distributional Reinforcement Learning
审计分布式强化学习的风险声明
Abstract
Distributional reinforcement learning agents learn full return distributions that are increasingly read at face value: for interpretability, risk-sensitive control, and safety monitoring. We ask a question theory anticipates but that has not been measured directly: are the risk claims of a trained distributional agent true? Our audit combines a decision-relevant screening metric (the excess Wasserstein gap between the top two actions, which equals the mass by which first-order stochastic dominance is violated), ground truth from snapshot-restart Monte Carlo, and a statistical harness (permutation nulls, bootstrap refutation, FDR control) without which the audit itself manufactures false conclusions. Across QR-DQN, C51, and IQN on MinAtar (33 runs), 40-95% of the strongest claimed risk trade-offs are refuted at 95% confidence, the placement of the strongest claims is statistically indistinguishable from truth-blind, and essentially no claim is confirmable: for these agents, the learned "risk" reflects a training artifact rather than environment stochasticity. The artifact is structural (fully formed early in training, uncorrelated with final score, idiosyncratic to each seed) and appears unchanged at full-Atari scale, with every top Breakout claim of a pretrained near-state-of-the-art QR-DQN refuted. Positive controls of known magnitude confirm 96-100% of real claims (correlation 0.89-0.92): the reading measures the agents, not the audit. Acting on the heads' CVaR advice at their most-flagged states ranges from beneficial to significantly worse than chance. Neither training for risk nor ensembling removes the artifact, and recalibration passes the audit only by nullifying the claims: the head is uninformative, not merely miscalibrated. We release the toolkit and document two silent pitfalls that produced convincing but wrong audits of our own.
Chinese Translation
分布式强化学习代理学习完整的回报分布，这些分布在解释性、风险敏感控制和安全监测方面越来越被直接接受。我们提出一个理论预期但尚未直接测量的问题：经过训练的分布式代理的风险声明是否真实？我们的审计结合了一个与决策相关的筛选指标（前两种行动之间的超额 Wasserstein 差距，即第一阶随机优势被违反的质量），来自快照重启蒙特卡罗的真实数据，以及一个统计工具（置换无效、引导反驳、假发现率控制），没有这些，审计本身会制造出错误的结论。在 MinAtar 上对 QR-DQN、C51 和 IQN 进行的 33 次实验中，强烈声称的风险权衡中有 40-95% 在 95% 的置信水平下被反驳，最强声明的排名在统计上与盲目真相无异，几乎没有声明是可以确认的：对于这些代理，学习到的“风险”反映的是训练伪影，而非环境随机性。该伪影是结构性的（在训练早期完全形成，与最终得分无关，且对每个种子具有特异性），在全 Atari 规模下似乎没有变化，所有预训练的近乎最先进的 QR-DQN 的顶级 Breakout 声明均被反驳。已知幅度的正控制确认了 96-100% 的真实声明（相关性 0.89-0.92）：该测量反映的是代理，而非审计。在其最被标记状态下，依据头部的 CVaR 建议采取行动的效果从有益到显著低于随机水平。无论是为风险训练还是集成都无法消除该伪影，而重新校准仅通过使声明无效来通过审计：头部信息不具备指导性，而不仅仅是校准错误。我们发布了工具包，并记录了两个导致我们自己产生令人信服但错误审计的隐性陷阱。
cs.AI / 113 / 2607.11621
Lesioned Multimodal Language Models Reproduce Aphasic Picture-Naming Patterns
损伤的多模态语言模型再现失语症的图像命名模式
Abstract
Aphasia following stroke commonly produces systematic naming errors with characteristic profiles, but whether general-purpose language models not designed for clinical simulation can reproduce these patterns remains untested. We investigated (1) whether lesions or controlled perturbations to a multimodal language model can reproduce different types of errors in picture naming, and (2) whether the framework can reproduce the complete error profile of individual persons with aphasia (PWAs). Using LLaVA 1.6, we evaluated perturbation configurations that varied the layer, proportion, and amount of noise applied to model units. We examined 278 PWAs on the Philadelphia Naming Test, classifying responses into seven categories using a validated neural classifier. Six of seven response categories (correct, semantic, mixed, unrelated, neologism, no response errors) emerged at clinically-comparable proportions across distinct parameter space regions, with formal paraphasia being the exception. Searching the perturbation space revealed configurations that reproduced the individual error profile in at least six of seven categories for 97.8% of PWAs and in all seven categories for 79.5% of PWAs. Monte Carlo baselines confirmed that this matching reflects joint inter-category structure rather than marginal overlap. These results establish a quantitative framework for reproducing individual aphasic error patterns in picture naming. They suggest the potential for language models to serve as digital twins of individuals with post-stroke aphasia.
Chinese Translation
中风后失语症通常会产生具有特征性轮廓的系统性命名错误，但尚未测试未针对临床模拟设计的通用语言模型是否能够再现这些模式。我们研究了（1）损伤或对多模态语言模型进行控制扰动是否能够再现图像命名中的不同类型错误，以及（2）该框架是否能够再现个别失语症患者（PWAs）的完整错误轮廓。使用LLaVA 1.6，我们评估了扰动配置，改变了施加于模型单元的层、比例和噪声量。我们对278名PWAs进行了费城命名测试，将反应分类为七个类别，使用经过验证的神经分类器。七个反应类别中的六个（正确、语义、混合、不相关、新词、无反应错误）在不同参数空间区域以临床可比的比例出现，正式的言语错乱是例外。对扰动空间的搜索揭示了能够再现97.8%的PWAs在至少六个类别中的个体错误轮廓的配置，并且79.5%的PWAs在所有七个类别中。蒙特卡洛基线确认这种匹配反映了类别间的联合结构，而不是边际重叠。这些结果建立了一个定量框架，用于再现个体失语症患者在图像命名中的错误模式。它们表明语言模型有潜力作为中风后失语症个体的数字双胞胎。
cs.AI / 114 / 2607.11632
Reproducing human biases in route choice using large language models: Toward scalable behavioral modeling
利用大型语言模型重现人类在路线选择中的偏见：迈向可扩展的行为建模
Abstract
Human choice behavior, including route choice, exhibits systematic behavioral biases that deviate from the assumptions of full rationality. Cumulative prospect theory (CPT) has been widely recognized as an effective framework for characterizing such behavioral patterns. However, its large-scale application, particularly in simulation and agent-based modeling, critically depends on specifying individual-level CPT parameters, which remain a major bottleneck. Conventional approaches typically rely on surveys and controlled experiments to calibrate CPT parameters, yet these methods are difficult to generalize and often fail to capture the full diversity of human decision-making. To address this challenge, this paper investigates whether large language models (LLMs) can reproduce human behavioral biases in choice-making without explicit specification of prospect-theoretic parameters. Using route choice as a representative scenario, we design a behavioral evaluation framework and systematically compare LLM-generated decisions with established human behavioral patterns predicted by CPT. Experimental results demonstrate that LLMs are capable of reproducing non-rational human choice biases and can exhibit decision behaviors consistent with prospect-theoretic effects under uncertainty. These findings suggest that generative AI models may provide a scalable alternative for modeling human decision processes and offer a promising foundation for next-generation large-scale agent-based simulation and AI-driven behavioral research.
Chinese Translation
人类选择行为，包括路线选择，表现出系统性的行为偏见，这些偏见偏离了完全理性的假设。累积前景理论（CPT）被广泛认可为描述此类行为模式的有效框架。然而，其大规模应用，特别是在模拟和基于代理的建模中，关键依赖于个体层面的CPT参数的指定，这仍然是一个主要瓶颈。传统方法通常依赖于调查和控制实验来校准CPT参数，但这些方法难以推广，且往往无法捕捉人类决策的全部多样性。为了解决这一挑战，本文探讨了大型语言模型（LLMs）是否能够在不明确指定前景理论参数的情况下重现人类在选择中的行为偏见。以路线选择作为代表性场景，我们设计了一个行为评估框架，并系统地将LLM生成的决策与CPT预测的已建立人类行为模式进行比较。实验结果表明，LLMs能够重现非理性的人的选择偏见，并在不确定性下表现出与前景理论效应一致的决策行为。这些发现表明，生成性人工智能模型可能为建模人类决策过程提供可扩展的替代方案，并为下一代大规模基于代理的模拟和人工智能驱动的行为研究提供了有希望的基础。
cs.AI / 115 / 2607.11696
Think Through a Bottleneck: Hourglass Reasoning for Rigorous Induction
通过瓶颈思考：严格归纳的沙漏推理
Abstract
Self-refinement often fails to strengthen few-shot inductive reasoning in large language models. Prompting a model to explicitly state its inferred rule does little on its own. What actually matters is a structurally enforced isolation between reasoning stages, so that information can only pass between them as a compressed symbolic state. We introduce \textbf{Hourglass reasoning}, which enforces strict context isolation between reasoning stages. The frozen LLM acts as a meta-constructor, building for each task a symbolic encoder--decoder: an Induction module compresses the support examples into a schema $\phi$ (encoder) and a transient scaffold $z$; a Deduction module derives rule $T$ (decoder) from these and discards $z$; an Implementer compiles $(\phi, T)$ into artifacts; an error-driven Refiner revises $(\phi, T)$ and regenerates artifacts from scratch. Only $(\phi, T)$ crosses stage boundaries, so all refinement stays anchored to the rule. We evaluate Hourglass across three benchmarks spanning visual abstraction, hardware synthesis, and textual rule induction, using GPT-5.5 and Gemini 3.1 Pro. On ARC-AGI-2, it raises best-of-5 accuracy by up to 14 points over an iterative-refinement baseline. On ChipBench, it nearly doubles Verilog synthesis accuracy with GPT-5.5, from 31\% to 58\%. BBEH-Linguini draws on puzzles from the International Linguistics Olympiad, a setting where prior work has shown that explicit verbalization can hurt performance. Hourglass mitigates this tendency, and on Gemini 3.1 Pro, it reverses the effect entirely. Ablations confirm that these gains come from the isolation between stages and the quality of the initial induction, not from prompt wording or the particular symbolic form used. It is how information flows through the reasoning process, rather than the language used to express it, that drives inductive reasoning in frozen LLMs.
Chinese Translation
自我精炼常常未能增强大型语言模型中的少样本归纳推理。单独促使模型明确陈述其推断规则效果有限。真正重要的是在推理阶段之间强制实施结构性隔离，以便信息只能以压缩的符号状态在它们之间传递。我们提出了 extbf{沙漏推理}，该方法在推理阶段之间强制实施严格的上下文隔离。冻结的语言模型（LLM）充当元构造器，为每个任务构建一个符号编码器-解码器：归纳模块将支持示例压缩为一个模式$ heta$（编码器）和一个临时支架$z$；推导模块从这些中推导出规则$T$（解码器）并丢弃$z$；实施者将$( heta, T)$编译成工件；一个基于错误的精炼器修订$( heta, T)$并从头生成工件。只有$( heta, T)$跨越阶段边界，因此所有的精炼都保持与规则的锚定。我们在三个基准测试上评估沙漏推理，涵盖视觉抽象、硬件综合和文本规则归纳，使用GPT-5.5和Gemini 3.1 Pro。在ARC-AGI-2上，它将最佳的五次准确率提高了最多14个百分点，相比于迭代精炼基线。在ChipBench上，它几乎将GPT-5.5的Verilog综合准确率从31 ext{%}提高到58 ext{%}。BBEH-Linguini利用国际语言学奥林匹克中的难题，这是一个先前研究表明明确的语言表达可能会损害表现的场景。沙漏推理缓解了这种倾向，在Gemini 3.1 Pro上，它完全逆转了这一效果。消融实验确认这些增益来自于阶段之间的隔离和初始归纳的质量，而不是来自提示措辞或特定的符号形式。推动冻结LLM中的归纳推理的是信息在推理过程中的流动方式，而非表达它所使用的语言。
cs.AI / 116 / 2607.11749
Playful AI in Professional Email: A Field Experiment on Tone and Recipient Engagement
专业电子邮件中的趣味性人工智能：关于语气和收件人参与度的实地实验
Abstract
Large language models (LLMs) are rapidly reshaping workplace communication, yet whether AI-assisted writing changes how recipients actually behave, and through what channel, remains unknown. Here, in a randomized crossover field experiment, 121 employees across six companies sent work emails under three conditions over three weeks: unaided writing, GPT-5 rewriting in a playful tone, and GPT-5 rewriting in a professional tone. Across 16,880 emails, playful editing increased emotional positivity (B=+0.068, p<0.001), and professional editing decreased it (B=-0.041, p<0.001), yet neither condition directly altered open rates, reply rates, or response times. Instead, within-sender positivity strongly predicted both opening (OR=2.05) and replying (OR=3.32, p<0.001), a significant indirect pathway through which AI editing shaped behavior, in the absence of any direct effect. These findings suggest that AI-assisted communication shapes workplace engagement not through its use, but through the emotional tone of the language it produces.
Chinese Translation
大型语言模型（LLMs）正在迅速重塑职场沟通，但AI辅助写作是否改变了收件人的实际行为，以及通过何种渠道改变，仍然未知。在这项随机交叉的实地实验中，121名来自六家公司的员工在三周内在三种条件下发送工作电子邮件：未辅助写作、GPT-5以趣味性语气重写和GPT-5以专业语气重写。在16880封电子邮件中，趣味性编辑增加了情感积极性（B=+0.068，p<0.001），而专业编辑则降低了情感积极性（B=-0.041，p<0.001），然而这两种条件都没有直接改变打开率、回复率或响应时间。相反，发件人的积极性强烈预测了打开率（OR=2.05）和回复率（OR=3.32，p<0.001），这是AI编辑影响行为的一个显著间接路径，而没有任何直接效果。这些发现表明，AI辅助沟通通过其产生的语言的情感语气来塑造职场参与度，而不是通过其使用方式。
cs.CL / 1 / 2607.09880
CLIR-Bench: Benchmarking Multimodal Question Answering over Irregular Clinical Time Series
CLIR-Bench：不规则临床时间序列的多模态问答基准测试
Abstract
Clinical time series are central to patient monitoring, risk assessment, and clinical decision support. However, they are often sparse, irregularly sampled, and asynchronous, making it difficult for models to identify the temporal evidence required for clinical Question Answering (QA). Existing benchmarks primarily focus on regularly sampled time-series QA or medical QA over static data, and therefore rarely assess whether models can faithfully ground their answers in irregular temporal observations. To fill this gap, we introduce CLIR-Bench, a benchmark for irregular clinical time series QA constructed from de-identified ICU records through a principled four-stage pipeline. CLIR-Bench contains 6,600 QA instances spanning 11 clinical variables, organized into four capability dimensions and 11 tasks. Each question is linked to explicit temporal evidence and task-specific answer derivation rules, enabling evaluation of both answer accuracy and evidence use. Experiments show that existing generalist models struggle to retrieve and reason over sparse clinical evidence, highlighting the need for stronger irregular time-series reasoning methods. Our code and data are available at https://huggingface.co/datasets/winall/CLIR-Bench.
Chinese Translation
临床时间序列在患者监测、风险评估和临床决策支持中至关重要。然而，它们通常是稀疏的、不规则采样的和异步的，这使得模型难以识别临床问答（QA）所需的时间证据。现有的基准主要集中在规则采样的时间序列问答或静态数据上的医学问答，因此很少评估模型是否能够忠实地基于不规则时间观察来支撑其答案。为填补这一空白，我们引入了CLIR-Bench，这是一个基于去标识化ICU记录构建的不规则临床时间序列问答基准，采用了一个原则性的四阶段流程。CLIR-Bench包含6600个问答实例，涵盖11个临床变量，组织成四个能力维度和11个任务。每个问题都与明确的时间证据和特定任务的答案推导规则相链接，从而能够评估答案的准确性和证据的使用。实验表明，现有的通用模型在检索和推理稀疏的临床证据方面存在困难，突显了对更强的不规则时间序列推理方法的需求。我们的代码和数据可在https://huggingface.co/datasets/winall/CLIR-Bench获取。
cs.CL / 2 / 2607.09885
Index SLM Technical Report
Index SLM 技术报告
Abstract
We present Index-1.9B, a series of open small language models developed at Bilibili. The series comprises four models: Index-1.9B-Base, a foundation model with 1.9 billion non-embedding parameters pre-trained on 2.8 trillion predominantly Chinese and English tokens; Index-1.9B-Pure, a control variant trained with an identical recipe but with all instruction-like data strictly filtered from the corpus; Index-1.9B-Chat, aligned from the base model with supervised fine-tuning and direct preference optimization; and Index-1.9B-Character, which augments the chat model with retrieval-augmented generation for few-shot role-playing customization. Pre-training employs a Warmup-Stable-Decay learning-rate schedule in which the concentration of curated data is raised substantially during the decay phase, together with a Norm-Head output layer that stabilizes training under large learning rates. On a suite of standard benchmarks covering examination, reasoning, mathematics, and code, Index-1.9B-Base attains an average score of 64.92, competitive with or exceeding open models of several times its size. We further report controlled studies on model depth, learning-rate magnitude and scheduling, the interaction between learning-rate decay and data quality, and the effect of including instruction data during pre-training, and we document an unexplained surge in benchmark performance midway through the constant-learning-rate phase. All models, together with evaluation code, are released at https://github.com/bilibili/Index-1.9B.
Chinese Translation
我们介绍了 Index-1.9B，这是在哔哩哔哩开发的一系列开放小型语言模型。该系列包括四个模型：Index-1.9B-Base，这是一个基础模型，具有 19 亿个非嵌入参数，预训练于 2.8 万亿个主要为中文和英文的标记；Index-1.9B-Pure，这是一个控制变体，使用相同的配方进行训练，但严格过滤掉所有类似指令的数据；Index-1.9B-Chat，这是从基础模型对齐而来，经过监督微调和直接偏好优化；以及 Index-1.9B-Character，它通过检索增强生成来增强聊天模型，以实现少量样本角色扮演定制。预训练采用 Warmup-Stable-Decay 学习率调度，其中在衰减阶段大幅提高了精心策划的数据浓度，并结合 Norm-Head 输出层，在较大的学习率下稳定训练。在覆盖考试、推理、数学和代码的一系列标准基准测试中，Index-1.9B-Base 的平均得分为 64.92，与其规模数倍的开放模型相比具有竞争力或超越。我们进一步报告了关于模型深度、学习率大小和调度、学习率衰减与数据质量之间的相互作用，以及在预训练期间包含指令数据的影响的控制研究，并记录了在恒定学习率阶段中期基准性能的未解释激增。所有模型及评估代码均已发布在 https://github.com/bilibili/Index-1.9B。
cs.CL / 3 / 2607.09908
RouteRec: Strict Evaluation of Recommender-Agent Selection and Aggregation
RouteRec：推荐代理选择与聚合的严格评估
Abstract
Recommender systems increasingly face a choice among heterogeneous agents -- collaborative filters, sequential models, content-based retrievers, and LLM-based rerankers -- yet no single agent is uniformly best. We study this choice as task-aware agent ranking under cost constraints using RouteRec, a framework that compares request-level hard selection with item-level learned aggregation over four traditional recommender agents and one LLM reranker agent. On MovieLens-1M, the full quality oracle has substantial headroom (HR@10 = 0.584), confirming that useful cross-agent signal exists. Under a leakage-free 5-fold out-of-fold protocol, however, hard selection remains below BM25 (0.223 vs. 0.254), and selective LLM escalation does not improve it. The same protocol yields a different outcome for learned aggregation: its cheap-only variant matches BM25 in HR and has a higher NDCG point estimate (0.123 vs. 0.114), while gated all-agent aggregation reaches HR@10 = 0.295 with 70.2\% LLM calls. The resulting lesson is not that routing is solved, but that request-level selection of one complete agent list is too coarse for this sparse fixed-candidate setting; item-level aggregation is the more promising action space.
Chinese Translation
推荐系统越来越面临在异构代理之间进行选择——协同过滤、序列模型、基于内容的检索器和基于大语言模型（LLM）的重排序器——但没有单一代理在所有情况下都是最佳的。我们将这一选择视为在成本约束下的任务感知代理排名，使用RouteRec框架比较请求级的硬选择与四种传统推荐代理和一种LLM重排序代理的项目级学习聚合。在MovieLens-1M数据集上，完整质量的oracle具有相当大的提升空间（HR@10 = 0.584），确认了有用的跨代理信号的存在。然而，在无泄漏的5折交叉验证协议下，硬选择的表现仍低于BM25（0.223对比0.254），而选择性LLM升级并未改善这一结果。同样的协议对于学习聚合则得出了不同的结果：其仅限于低成本的变体在HR上与BM25持平，并且具有更高的NDCG点估计（0.123对比0.114），而门控的全代理聚合在70.2%的LLM调用下达到了HR@10 = 0.295。由此得出的教训并不是路由问题已经解决，而是请求级选择一个完整代理列表对于这一稀疏固定候选设置来说过于粗糙；项目级聚合是更有前景的行动空间。
cs.CL / 4 / 2607.09921
Global Merger-Arbitrage Forecasting with Language Models
基于语言模型的全球并购套利预测
Abstract
We present a language-model forecasting system for merger arbitrage, a specialized high-stakes financial setting in which the task is to predict the outcome of announced M\&A deals. Unlike prior work on judgmental forecasting with LLMs, which has focused on broad mixed-topic benchmarks and short context such as news snippets, we study a setting that requires long-context reasoning over hundreds of pages of technical documents. Our system combines expert-guided context engineering with finetuning on hindsight-guided reasoning traces derived from historical deals. Given an announced deal, it outputs a probability distribution over three mutually exclusive outcomes: closing at announced terms, a higher bid, or deal termination. On an out-of-sample set of more than 400 large deals spanning 42 countries, our finetuned system achieves the best performance of any method we evaluate, reducing class-balanced Brier score to 0.151. This is 24\% below calibrated market-implied probabilities, 19\% below XGBoost, and 25-42\% below frontier language models. These results, together with ablation studies, show that LLM-based forecasting can succeed in specialized, long-context financial workflows, with hindsight-based supervision and expert-designed context playing a critical role.
Chinese Translation
我们提出了一种用于并购套利的语言模型预测系统，这是一种专门的高风险金融环境，其任务是预测已宣布的并购交易的结果。与之前关于使用大语言模型（LLMs）进行判断性预测的研究不同，后者主要集中在广泛的混合主题基准和短小的上下文（如新闻片段）上，我们研究了一种需要对数百页技术文档进行长上下文推理的情境。我们的系统结合了专家指导的上下文工程和基于历史交易的回顾性推理轨迹的微调。给定一项已宣布的交易，它输出三个相互排斥结果的概率分布：按宣布条款完成交易、提高报价或交易终止。在一个包含42个国家、超过400笔大型交易的样本外数据集中，我们的微调系统在我们评估的所有方法中表现最佳，将类别平衡的Brier评分降低至0.151。这比市场隐含概率低24%，比XGBoost低19%，比前沿语言模型低25-42%。这些结果以及消融研究表明，基于LLM的预测可以在专门的长上下文金融工作流程中取得成功，其中基于回顾的监督和专家设计的上下文发挥了关键作用。
cs.CL / 5 / 2607.09932
Faithful by Design: Evaluating and Improving LLM-Generated Clinical Trial Summaries for Multi-Stakeholder Audiences
设计的忠实性：评估和改进针对多利益相关者受众的LLM生成临床试验摘要
Abstract
Large language models are increasingly used to summarize clinical trial results for healthcare providers, patients, and payers, but their tendency to hallucinate poses significant risks in this high-stakes context. This study introduces a benchmark evaluation framework for measuring the faithfulness of LLM-generated clinical trial summaries across three stakeholder audiences. The framework consists of 200 stratified trials drawn from the Aggregate Analysis of ClinicalTrials.gov database, evaluated using audience-specific prompt templates and a six-dimension faithfulness annotation schema. Baseline measurements were established for GPT-4o, Claude Sonnet 4.6, and Gemini 2.5 Flash across 1,800 generated summaries scored using a cross-encoder natural language inference (NLI) model. Unsupported Claims was identified as the dominant failure mode across all three models, with a mean annotation score of 1.55 out of three. A knowledge-graph-augmented retrieval system was developed and evaluated against the baseline, producing statistically significant improvements in NLI-based faithfulness scores (entailment +0.0125, faithfulness +0.0130, p < 0.0001). Improvement pathways were model-dependent, with GPT-4o improving primarily through contradiction reduction while Claude Sonnet 4.6 and Gemini 2.5 Flash improved through increased entailment.
Chinese Translation
大型语言模型越来越多地用于为医疗提供者、患者和支付方总结临床试验结果，但其产生幻觉的倾向在这一高风险背景下带来了显著风险。本研究引入了一种基准评估框架，用于衡量LLM生成的临床试验摘要在三个利益相关者受众中的忠实性。该框架由200个来自ClinicalTrials.gov数据库的分层试验组成，使用特定于受众的提示模板和六维忠实性注释方案进行评估。针对GPT-4o、Claude Sonnet 4.6和Gemini 2.5 Flash建立了基线测量，涵盖1,800个生成的摘要，并使用交叉编码器自然语言推理（NLI）模型进行评分。所有三个模型中，未支持的声明被识别为主要失效模式，平均注释得分为3分中的1.55。开发了一种知识图谱增强的检索系统，并与基线进行了评估，在基于NLI的忠实性得分上产生了统计显著的改善（蕴含 +0.0125，忠实性 +0.0130，p < 0.0001）。改进路径因模型而异，GPT-4o主要通过减少矛盾而改善，而Claude Sonnet 4.6和Gemini 2.5 Flash则通过增加蕴含而改善。
cs.CL / 6 / 2607.09957
Workload-Driven Optimization for On-Device Real-Time Subtitle Translation
基于工作负载的设备端实时字幕翻译优化
Abstract
This report studies on-device English-to-Traditional-Chinese subtitle translation for Taiwan under short inputs, short outputs, batch-size-one inference, low latency, and privacy constraints. These conditions limit the value of optimizations designed for long-context or high-throughput language-model serving. Starting from LMT-60-0.6B, preliminary profiling suggests that vocabulary projection becomes a more important decode-time cost after GGUF quantization reduces the relative cost of Transformer blocks. We replace the original 151k-token vocabulary with a 64k-token subtitle-domain tokenizer, migrate the embedding space, and adapt the model through embedding calibration followed by full supervised fine-tuning. On a fixed 500-example subset of the OpenSubtitles2024 test set, the LocalSubs achieves a 59.2% tie-excluded win rate against Google Translate under GPT-4o pairwise judging. Performance is strongest on short cues and declines as cue length increases. Preliminary Apple M2 Metal measurements on a 64k-vocabulary model show a 1.63$\times$ speedup over a 151k-vocabulary profiling baseline. The raw benchmark configuration is incomplete, so the latency result is treated as preliminary.
Chinese Translation
本报告研究了在短输入、短输出、单批次推理、低延迟和隐私限制条件下，针对台湾的设备端英语到繁体中文的字幕翻译。这些条件限制了为长上下文或高吞吐量语言模型服务而设计的优化的价值。基于LMT-60-0.6B的初步分析表明，在GGUF量化减少了Transformer模块的相对成本后，词汇投影成为解码时更重要的成本。我们用一个64k-token的字幕领域分词器替换了原来的151k-token词汇，迁移了嵌入空间，并通过嵌入校准和全监督微调来适应模型。在OpenSubtitles2024测试集的固定500个示例子集上，LocalSubs在GPT-4o的成对评判下，取得了59.2%的排除平局的胜率，优于Google Translate。性能在短提示上最强，随着提示长度的增加而下降。对64k词汇模型的初步Apple M2 Metal测量显示，相较于151k词汇的基线，速度提升了1.63倍。由于原始基准配置不完整，因此延迟结果被视为初步结果。
cs.CL / 7 / 2607.09999
Silent Failures in Quantized LLM Reasoning: A Taxonomy-Based Analysis of Hollow Convergence and Failure Mode Shifts
量化大语言模型推理中的隐性失败：基于分类法的空洞收敛与失败模式转变分析
Abstract
We show that post-training quantization can silently alter how large language models reason even when task accuracy is preserved. Using a six-category failure taxonomy validated by two independent human annotators (Cohen's $\kappa$ = 0.906), we classify 30,000 chain-of-thought outputs from five instruction-tuned LLMs (3B--14B parameters) across three quantization precisions (FP32, FP16, NF4) and four reasoning benchmarks. We find that while accuracy is robust across precisions (maximum 3.1 pp drop), Hollow Convergence (correct answers reached through incomplete or unverifiable reasoning) shows a significant size-dependent shift under NF4, dropping sharply for the two smallest models tested but remaining invariant for models at 12B parameters and above. This effect is also benchmark-specific: GSM8K is categorically immune while LogiQA and ARC-Challenge show the largest shifts. Furthermore, under NF4, Shortcut Collapse rises from 44% to 78% of wrong-answer failures in LLaMA 3.2-3B while Confidence Snowballing collapses from 15.8% to near zero, a qualitative shift invisible to accuracy metrics. Finally, we show Hollow Convergence cannot be reliably detected from surface-level text features (best F1 = 0.53), establishing it as a deployment-relevant failure mode that standard evaluation pipelines cannot catch.
Chinese Translation
我们展示了后训练量化可以在任务准确性保持的情况下，悄然改变大型语言模型的推理方式。通过一个经过两位独立人类注释者验证的六类失败分类法（Cohen's $ ext{kappa}$ = 0.906），我们对来自五个指令调优的大语言模型（参数规模为3B至14B）在三种量化精度（FP32、FP16、NF4）和四个推理基准上的30,000个思维链输出进行了分类。我们发现，尽管在不同精度下的准确性保持稳健（最大下降3.1个百分点），但在NF4下，空洞收敛（通过不完整或不可验证的推理得出的正确答案）显示出显著的规模依赖性转变，对于测试的两个最小模型而言，下降幅度明显，而对于参数在12B及以上的模型则保持不变。该效应还具有基准特异性：GSM8K在类别上免疫，而LogiQA和ARC-Challenge则显示出最大的转变。此外，在NF4下，LLaMA 3.2-3B中的快捷崩溃错误答案失败的比例从44%上升至78%，而置信度雪球效应则从15.8%骤降至接近零，这种定性转变在准确性指标中是不可见的。最后，我们表明，空洞收敛无法通过表面文本特征可靠检测（最佳F1 = 0.53），确立其为一种与部署相关的失败模式，而标准评估流程无法捕捉到这一点。
cs.CL / 8 / 2607.10020
Robust, Scalable Detection of Text Containment in Large Web-Crawled Corpora
在大型网络爬取语料库中稳健且可扩展的文本包含检测
Abstract
We present FindMyText, an open-source Python package designed to efficiently assess whether a given text appears, in part or in full, within a text corpus. The tool builds on prior techniques for document fingerprinting, but extends them with a novel mechanism to explicitly capture sequences of matching fingerprints. By identifying such chains, the tool can more reliably detect near-verbatim copies of a given text rather than mere textual similarities. This makes FindMyText particularly suited for verifying the presence of copyrighted material in a corpus. Leveraging a distributed, disk-based indexing framework, the system scales to large web-crawled datasets. Using a new benchmark for evaluating text containment methods, we show that FindMyText outperforms alternative approaches across three datasets (ArXiv papers, Wikipedia, and generic web content).
Chinese Translation
我们提出了 FindMyText，这是一个开源的 Python 包，旨在高效评估给定文本是否部分或完全出现在文本语料库中。该工具基于先前的文档指纹技术，但通过一种新颖的机制扩展了这些技术，以明确捕捉匹配指纹的序列。通过识别这些链条，该工具能够更可靠地检测给定文本的近乎逐字复制，而不仅仅是文本相似性。这使得 FindMyText 特别适合于验证语料库中版权材料的存在。该系统利用分布式的基于磁盘的索引框架，能够扩展到大型网络爬取数据集。通过使用一种新的基准来评估文本包含方法，我们展示了 FindMyText 在三个数据集（ArXiv 论文、维基百科和通用网络内容）上优于其他方法。
cs.CL / 9 / 2607.10092
Efficiently Adapting Spoken Language Models for the Singaporean Context
高效适应新加坡语境的口语语言模型
Abstract
Spoken language models (SLMs) unify speech perception and reasoning, but adapting them to sensitive domains is underexplored, especially when the original training data is inaccessible and the use case demands multilingual, spoken-query interaction. We adapt an open-source SLM to the Singaporean Home Team context across five speech tasks in Singapore's four official languages, combining LoRA fine-tuning, a surrogate text-QA dataset that guards against catastrophic forgetting, and a multi-task objective that adapts the CoBa reweighting scheme to speech. We also build HTD-multilingual-QA, a 504,853 sample multilingual QA dataset in text and spoken form. The resulting HT-Moonstone (5B) matches or outperforms SLMs up to 7x its size on most tasks, attains the best accent and gender recognition among all models evaluated, and loses under 2\% of its original speech QA ability.
Chinese Translation
口语语言模型（SLMs）统一了语音感知和推理，但在敏感领域的适应性研究仍然不足，尤其是在原始训练数据不可获取且使用场景要求多语言、口语查询交互的情况下。我们将一个开源的SLM适应于新加坡的家庭团队语境，涵盖新加坡四种官方语言的五个语音任务，结合了LoRA微调、一个防止灾难性遗忘的替代文本问答数据集，以及一个将CoBa重加权方案适应于语音的多任务目标。我们还构建了HTD-multilingual-QA，这是一个包含504,853个样本的多语言问答数据集，涵盖文本和口语形式。最终生成的HT-Moonstone（5B）在大多数任务中与其大小最多相当的SLMs相匹配或超越，在所有评估模型中获得最佳的口音和性别识别，并且其原始语音问答能力损失低于2%。
cs.CL / 10 / 2607.10114
Cost of Reasoning in non-English Languages: A Case Study on Japanese
非英语语言中的推理成本：以日语为例
Abstract
Reasoning Language Models (RLMs) achieve their strongest performance when they reason in English, the language for which reasoning-oriented training data is most abundant. However, reasoning trace is a clue for model interpretability and safety, and useful in practice for both the model users and for model developers. Thus, it is desirable to be able to develop a model that reasons in a language of the user's choice, while still maintaining strong reasoning performance. To this end, we study the feasibility of training a model that reasons in Japanese. We develop a Japanese-reasoning variant of Qwen-3-Swallow-8B, which is a Japanese LLM continually pretrained from Qwen-3-8B, with GRPO and evaluate it across coding, math, and science benchmarks. The study shows that reasoning-language control is feasible by training a Japanese continually pretrained model with GRPO. However, its performance is at best on par with strong English-reasoning baselines on several benchmarks. We also evaluate the trained model on Japanese cultural benchmarks and observe that the model's performance is worse than the baseline models, suggesting that the reasoning in Japanese does not immediately improve performance on culturally relevant tasks for free.
Chinese Translation
推理语言模型（RLMs）在使用英语进行推理时表现最佳，因为针对推理的训练数据在该语言中最为丰富。然而，推理轨迹是模型可解释性和安全性的线索，对模型用户和开发者在实践中都很有用。因此，开发一个能够使用用户选择的语言进行推理的模型，同时保持强大的推理性能，是非常理想的。为此，我们研究了训练一个能够用日语进行推理的模型的可行性。我们开发了一个日语推理变体的Qwen-3-Swallow-8B，该模型是基于Qwen-3-8B不断预训练的日语大语言模型（LLM），并使用GRPO进行评估，涵盖编码、数学和科学基准。研究表明，通过使用GRPO训练一个不断预训练的日语模型，实现推理语言控制是可行的。然而，在多个基准测试中，其性能在最佳情况下也仅与强大的英语推理基准持平。我们还在日语文化基准上评估了训练后的模型，观察到该模型的表现不如基准模型，这表明在日语中进行推理并不会立即改善文化相关任务的性能。
cs.CL / 11 / 2607.10194
Instruction Set and Language for Hypergraphs
超图的指令集与语言
Abstract
We present IsalHG, a method for representing the structure of any finite, connected hypergraph of bounded hyperedge arity as a string over a compact instruction alphabet $\Sigma_{\mathrm{HG}}$. The encoding is executed by a small virtual machine comprising a sparse hypergraph, a circular doubly-linked list (CDLL) of node references, and $k$ traversal pointers, where $k$ bounds the hyperedge arity. Instructions either move a pointer through the CDLL or insert a hyperedge, optionally together with new nodes, into the hypergraph. Every string over $\Sigma_{\mathrm{HG}}$ decodes to a valid hypergraph; the alphabet is closed. A greedy \emph{HypergraphToString} (h2s) algorithm encodes any connected hypergraph into a string; a backtracking variant seeded at nodes of lexicographically maximal structural tuple produces a \emph{canonical string} $w^{*}$, which we conjecture to be a complete isomorphism invariant. Canonical-string equality then decides hypergraph isomorphism natively, without the standard reduction to the Levi incidence graph followed by a graph-isomorphism engine. We verify the round-trip property $s2h(h2s(H)) \cong H$ on 150 connected random uniform hypergraphs and on named combinatorial designs, and we benchmark the canonical algorithm against the three practically available exact baselines -- nauty, Traces, and bliss operating on the 2-coloured Levi graph -- across a $(n, c)$ grid with ten seeds per cell. All four methods agree on every one of 600 isomorphism verdicts, consistent with the completeness conjecture. On wall-clock time the Levi baselines dominate every tested cell by three to five orders of magnitude (geometric-mean ratio $311\times$ to $117{,}672\times$), which we report as measured. We contribute the representation framework, a conjecture of canonical completeness, and the first native-versus-Levi benchmark for hypergraph isomorphism.
Chinese Translation
我们提出了 IsalHG，一种将任何有限的、连通的超图结构（具有有限的超边度）表示为紧凑指令字母表 $ ext{Σ}_{ ext{HG}}$ 上字符串的方法。该编码由一个小型虚拟机执行，该虚拟机包含一个稀疏超图、一个节点引用的循环双向链表（CDLL）以及 $k$ 个遍历指针，其中 $k$ 限制了超边的度。指令要么在 CDLL 中移动指针，要么将超边（可选地与新节点一起）插入超图中。每个 $ ext{Σ}_{ ext{HG}}$ 上的字符串都解码为一个有效的超图；该字母表是封闭的。一个贪心的 extit{HypergraphToString} (h2s) 算法将任何连通超图编码为字符串；一个从字典序最大结构元组的节点出发的回溯变体生成一个 extit{规范字符串} $w^{*}$，我们推测这是一个完整的同构不变量。规范字符串的相等性可以原生地决定超图同构，而无需标准地简化为 Levi 事件图，然后再通过图同构引擎处理。我们在 150 个连通随机均匀超图和命名组合设计上验证了往返性质 $s2h(h2s(H)) ext{≅} H$，并在 $(n, c)$ 网格上对三种实际可用的精确基线（nauty、Traces 和 bliss 在 2-着色 Levi 图上运行）进行了规范算法的基准测试，每个单元格有十个种子。所有四种方法在 600 个同构判决上达成一致，符合完整性猜想。在实际时间上，Levi 基线在每个测试单元上比其他方法快三个到五个数量级（几何平均比率 $311 imes$ 到 $117{,}672 imes$），这是我们测量的结果。我们贡献了表示框架、规范完整性的猜想，以及超图同构的首次原生与 Levi 基准测试。
cs.CL / 12 / 2607.10198
Equal Accuracy, Unequal Evidence: Search APIs as Decision Surfaces for Tool-Using Agents
相同准确性，不同证据：搜索API作为工具使用代理的决策表面
Abstract
Search APIs are the fundamental retrieval layer for many agents and are often their most frequently used tool. Traditional search APIs provide URLs, titles, and snippets that preview website contents. Because full-page retrieval is token-intensive, agent retrieval architectures increasingly use progressive disclosure: the agent first sees snippets and then chooses whether to fetch full pages. In such systems, search API performance is often evaluated primarily by answer accuracy. We argue that a commercial search API is better understood as a decision surface: the ranked snippets, URLs, and metadata that determine whether an agent answers immediately, searches again, or spends tokens opening pages. We test this claim with one frozen GPT-5.4 agent, two tools (search_web and fetch_page), and 100 questions from SEALQA-HARD, varying only the search provider (Brave, Tavily, Firecrawl). A Kimi-K2.6 oracle labels every content element visible to the agent (URL, title, snippet, and fetched page, when fetched), producing 6,869 valid per-URL judgments. We use an audited correct-answer label, semantic match, which preserves exact matches while accepting harmless formatting and naming variants. Under this measure, the providers remain close (25, 25, 26 / 100), but their evidence economies differ sharply: Brave offers gold-answer-rich snippets, Tavily concentrates gold-supporting URLs at rank 1, and Firecrawl is associated with broader exploration under this fixed agent policy. We also introduce a surface contradiction-to-gold URL ratio, which varies from 0.92 to 2.59. Provider choice is therefore a retrieval-budget and policy decision, not merely a recall decision.
Chinese Translation
搜索API是许多代理的基本检索层，通常是它们使用最频繁的工具。传统的搜索API提供URL、标题和预览网站内容的摘要。由于完整页面检索消耗大量令牌，代理检索架构越来越多地采用渐进式披露：代理首先看到摘要，然后选择是否获取完整页面。在这种系统中，搜索API的性能通常主要通过答案的准确性来评估。我们认为，商业搜索API更应被理解为一种决策表面：排名的摘要、URL和元数据决定了代理是立即回答、再次搜索还是花费令牌打开页面。我们使用一个冻结的GPT-5.4代理、两个工具（search_web和fetch_page）以及来自SEALQA-HARD的100个问题进行测试，仅改变搜索提供商（Brave、Tavily、Firecrawl）。一个Kimi-K2.6的oracle对代理可见的每个内容元素（URL、标题、摘要和获取的页面，若已获取）进行标注，产生了6,869个有效的每个URL判断。我们使用经过审核的正确答案标签，即语义匹配，保留精确匹配，同时接受无害的格式和命名变体。在这一标准下，各提供商的表现接近（25, 25, 26 / 100），但它们的证据经济差异明显：Brave提供丰富的黄金答案摘要，Tavily在排名1处集中黄金支持的URL，而Firecrawl则与在这一固定代理策略下的更广泛探索相关。我们还引入了一个表面矛盾与黄金URL的比率，范围从0.92到2.59。因此，提供商的选择不仅仅是一个召回决策，而是一个检索预算和政策决策。
cs.CL / 13 / 2607.10235
Consensus vs. Dissent: Dynamic LLM Modeling of Subjective Preferences in Group Recommenders
共识与异议：动态大语言模型对群体推荐中主观偏好的建模
Abstract
Previous work in group recommender systems has demonstrated a sensitivity to the distribution of preferences within a group. Specifically, the selection of the preference aggregation strategy benefits from considering such group configurations. In this paper, we study whether LLMs are able to mimic this sensitivity and to select the ideal aggregation strategy (and corresponding recommendation) according to nuanced human perceptions of fairness, satisfaction, and consensus. We do this by fine-tuning Large Language Models (LLMs) on human survey data to serve as real-time judgmental models within the recommendation pipeline. Using a reasoning dataset distilled from DeepSeek-V3.1 and human ground truth assessments, we develop Judgmental Llama and Judgmental OLMo to simulate group assessments. Our pipeline successfully generates multiple recommendation candidates based on social choice-based aggregation strategies and dynamically selects the one that maximizes these predicted human-like evaluations. We further validate these suggestions in a user study (n=284) and find that our methodology achieved the highest scores for satisfaction and group consensus. Furthermore, we find that LLM judgments are most aligned with human perceptions of fairness, satisfaction and consensus when we also consider interaction effects between our LLM-based method and group configuration (e.g., minority or coalition). These findings give further support for dynamically adapting aggregation strategies to specific within-group preference distributions, and highlight the advantage of using LLMs for an adaptation that is aligned with subjective human judgments.
Chinese Translation
以往的群体推荐系统研究表明，群体内偏好的分布对推荐结果具有敏感性。具体而言，偏好聚合策略的选择受益于对这种群体配置的考虑。本文研究了大语言模型（LLMs）是否能够模拟这种敏感性，并根据人类对公平性、满意度和共识的细微感知选择理想的聚合策略（及相应的推荐）。我们通过在基于人类调查数据上微调大语言模型，使其在推荐流程中作为实时判断模型。利用从DeepSeek-V3.1提炼的推理数据集和人类真实评估，我们开发了Judgmental Llama和Judgmental OLMo来模拟群体评估。我们的流程成功生成了基于社会选择的聚合策略的多个推荐候选，并动态选择最大化这些预测的人类类评估的候选。我们在一项用户研究中进一步验证了这些建议（n=284），发现我们的方法在满意度和群体共识方面获得了最高分。此外，我们发现当同时考虑我们基于LLM的方法与群体配置（例如，少数派或联盟）之间的交互效应时，LLM的判断与人类对公平性、满意度和共识的感知最为一致。这些发现进一步支持根据特定的群体内偏好分布动态调整聚合策略，并强调使用LLM进行与主观人类判断一致的适应的优势。
cs.CL / 14 / 2607.10245
PTEI: Integrating Personality Traits to Enhance Emotional Intelligence in Large Language Models
PTEI：整合人格特征以增强大型语言模型的情感智能
Abstract
Despite advances in Emotional Intelligence (EI), Large Language Models (LLMs) still significantly underperform humans in complex emotional reasoning. This gap originates partly from the limited incorporation of individual differences, particularly personality traits, which are fundamental to human emotional inference. To address this, we propose PTEI, a novel framework for integrating Personality Traits into Emotional Intelligence tasks using LLMs. In PTEI, MBTI and OCEAN personality traits are first extracted directly from the given emotional scenarios and then utilized as contextual knowledge within personality-aware prompts, guiding LLMs to accurately infer emotions and their underlying causes. To ensure optimal contextual grounding, we employ Contrastive Learning to construct an optimized retrieval system that surfaces emotionally and personally aligned scenarios, enhancing reasoning quality. Extensive experiments on established EI benchmarks show that PTEI enhances the Emotional Understanding (EU) capabilities of various LLMs, with the strongest improvement observed in GPT models. Combining PTEI with Chain-of-Thought (CoT) reasoning yields an additional 4 percent increase in accuracy. These findings underscore PTEI's contribution toward advancing AI systems with more sophisticated social and psychological grounding.
Chinese Translation
尽管情感智能（EI）取得了进展，大型语言模型（LLMs）在复杂情感推理方面仍显著低于人类。这一差距部分源于对个体差异的有限整合，特别是人格特征，而人格特征是人类情感推理的基础。为此，我们提出了PTEI，一个将人格特征整合到使用LLMs的情感智能任务中的新框架。在PTEI中，首先从给定的情感场景中直接提取MBTI和OCEAN人格特征，然后将其作为上下文知识用于人格感知提示，指导LLMs准确推断情感及其潜在原因。为了确保最佳的上下文基础，我们采用对比学习构建一个优化的检索系统，提取情感和个人一致的场景，从而提升推理质量。在已建立的EI基准上进行的大量实验表明，PTEI增强了各种LLMs的情感理解（EU）能力，其中在GPT模型中观察到最强的改进。将PTEI与思维链（CoT）推理结合使用，准确率额外提高了4%。这些发现强调了PTEI在推动具有更复杂社会和心理基础的人工智能系统方面的贡献。
cs.CL / 15 / 2607.10248
One mechanism for many mental spaces: a shared router over a value slot in language models
一个机制对应多个心理空间：语言模型中的共享路由器与价值槽
Abstract
Language builds discourse contexts other than the actual: a painting, a belief, a memory, a hypothetical. Each is a mental space in which the same entity can take a different value, as when a flower is red in reality but purple in a portrait. Formal semantics keeps these contexts apart because their logics differ (modal, temporal, doxastic, depictive); Fauconnier's mental-space theory treats them as one space-building operation. We ask which of these a transformer language model implements, and find a mechanistic version of Fauconnier's unification. The model uses one router/slot format across the inventory: a reusable value slot stores attributed content, and a causally manipulable router (the space index) selects which space is read. A subspace trained with Distributed Alignment Search to control one space type, counterfactual, belief, fictional, or temporal, also controls the others, well above a random floor, on three model families; belief, which formal semantics marks as a distinct case, is not specially separated. The router is low-rank, composes additively with entity identity, and acts through a few late-layer heads. Two further results show the mechanism drives inference and composes: a subspace trained on a rule-derived conclusion flips what the model infers while dissociating from what it reports, and composing space-builders mints a fresh router over the shared slot. This paper establishes the cross-type generality. A companion paper develops belief in depth, because of its special status in philosophy, psychology, and linguistics (epistemology, theory of mind, and propositional attitude reports).
Chinese Translation
语言构建了实际之外的语篇上下文：一幅画、一种信念、一段记忆、一个假设。每一个都是一个心理空间，其中同一实体可以取不同的值，就像一朵花在现实中是红色的，而在肖像中是紫色的。形式语义学将这些上下文区分开来，因为它们的逻辑不同（模态、时间、信念、描绘）；福孔尼尔（Fauconnier）的心理空间理论将它们视为一个空间构建操作。我们探讨变换器语言模型实现了其中的哪一种，并发现了福孔尼尔统一的机械版本。该模型在其库存中使用一种路由器/槽格式：一个可重用的价值槽存储属性内容，而一个因果可操控的路由器（空间索引）选择读取哪个空间。一个通过分布式对齐搜索（Distributed Alignment Search）训练的子空间控制一种空间类型（反事实、信念、虚构或时间），同时也控制其他空间，远高于随机基准，在三种模型家族中；信念在形式语义学中被标记为一个独特的案例，但并没有特别分开。该路由器是低秩的，与实体身份加法组合，并通过少数后层头部起作用。两个进一步的结果显示该机制驱动推理并进行组合：一个基于规则推导结论训练的子空间翻转了模型的推理，同时与其报告的内容解耦，而组合空间构建者则在共享槽上铸造出一个新的路由器。本文确立了跨类型的普遍性。一篇配套论文深入探讨信念，因为它在哲学、心理学和语言学（认识论、心智理论和命题态度报告）中的特殊地位。
cs.CL / 16 / 2607.10256
Which Languages Transfer Best to Warlpiri? A Similarity-Based Study for Low-Resource ASR
哪些语言对 Warlpiri 的迁移效果最佳？基于相似性的低资源自动语音识别研究
Abstract
This paper investigates how language similarity can improve cross-lingual transfer for automatic speech recognition (ASR) in extremely low-resource settings. Warlpiri, an Australian Aboriginal language, has very limited transcribed speech data, making transfer learning essential. We propose a framework combining acoustic similarity from pre-trained speech models with linguistic similarity based on typology, phoneme inventories, grammatical, and syntactic features to rank high-resource source languages and evaluate their effectiveness for ASR transfer to Warlpiri. Experiments with Whisper show that acoustically and typologically similar languages outperform monolingual and multilingual baselines. Assamese and Hindi achieve substantial reductions in word and character error rates. Correlation analysis further indicates that acoustic similarity is the strongest predictor of fine-tuning performance, while phoneme inventory and typological similarity better explain zero-shot transfer.
Chinese Translation
本文探讨了语言相似性如何改善在极低资源环境下的自动语音识别（ASR）跨语言迁移。Warlpiri 是一种澳大利亚土著语言，其转录语音数据非常有限，因此迁移学习至关重要。我们提出了一个框架，将来自预训练语音模型的声学相似性与基于类型学、音素库存、语法和句法特征的语言相似性相结合，以对高资源源语言进行排名，并评估其对 Warlpiri 的 ASR 迁移效果。与 Whisper 的实验表明，声学和类型学相似的语言优于单语和多语基线。阿萨姆语和印地语在单词和字符错误率上实现了显著降低。相关性分析进一步表明，声学相似性是微调性能的最强预测因素，而音素库存和类型学相似性更好地解释了零样本迁移。
cs.CL / 17 / 2607.10268
Language Re-generation: An investigation into information locality effects on reconstruction
语言再生成：对信息局部性效应在重构中的影响的研究
Abstract
Information locality, the tendency for syntactically related words to appear close together, shapes both human language processing and language model learning. While prior work has examined whether language models can acquire impossible languages, it remains unclear whether they can recover natural language from such input and what this reveals about their inductive biases. We address this by complementing learnability-based approaches with a reconstruction framework: fine-tuning GPT-2 models pre-trained on impossible languages to reconstruct natural English from three perturbation types. Our findings show that the recovered structures exhibit shorter dependency lengths than the original text, mirroring the locality preference observed in unconstrained language model generation and providing a quantitative signature of an architectural bias that learnability experiments alone do not reveal. Recovery difficulty increases with the degree of locality disruption. Structural recovery (dependency Triple F1) dissociates from surface recovery (Exact Match), while fluency dissociates from faithful reconstruction under global shuffling. Sentence length further modulates performance: longer sentences facilitate recovery when local structure is preserved but lead to complete collapse under global shuffling. Finally, recovery difficulty tracks learnability difficulty across perturbation types, suggesting that information locality is the shared constraint governing both.
Chinese Translation
信息局部性，即语法相关词汇倾向于紧密相邻出现的现象，影响着人类语言处理和语言模型学习。虽然之前的研究已经考察了语言模型是否能够掌握不可能的语言，但尚不清楚它们是否能够从这种输入中恢复自然语言，以及这揭示了它们的归纳偏见。我们通过将基于可学习性的研究方法与重构框架相结合来解决这一问题：对在不可能语言上预训练的GPT-2模型进行微调，以从三种扰动类型中重构自然英语。我们的研究发现，恢复的结构展现出比原始文本更短的依赖长度，反映出在无约束语言模型生成中观察到的局部性偏好，并提供了一种量化的架构偏见特征，这一点仅通过可学习性实验无法揭示。恢复的难度随着局部性干扰程度的增加而增加。结构恢复（依赖三元组F1）与表面恢复（完全匹配）相分离，而流畅性在全局洗牌下与忠实重构相分离。句子长度进一步调节性能：当局部结构得以保留时，较长的句子有助于恢复，但在全局洗牌下会导致完全崩溃。最后，恢复的难度与不同扰动类型下的可学习性难度相一致，这表明信息局部性是支配两者的共同约束。
cs.CL / 18 / 2607.10310
PolyInterview: An LLM-based Platform for Immersive Mock Interview Practice with Comprehensive Multimodal Assessment
PolyInterview：基于大型语言模型的沉浸式模拟面试实践平台，提供全面的多模态评估
Abstract
Preparing for job interviews is important for securing desired positions, yet realistic practice remains difficult to access: real interviews are infrequent, expert mock coaching is costly, and self-practice offers neither adaptive dialogue nor structured assessment. Existing systems typically address only parts of this need through fixed question sequences, limited communication channels, or feedback with little supporting evidence. We present PolyInterview, an LLM-based platform for immersive mock interview practice with comprehensive multimodal assessment. PolyInterview uses the target job description and CV to generate questions tailored to the role and candidate, conducts multi-turn spoken interviews with a lip-synced digital human interviewer that asks answer-aware follow-up questions, and evaluates response content, vocal delivery, and non-verbal behavior. Four parallel evaluators produce 13 behavior-level features that are aggregated into 10 assessment aspects and two competency tracks. Guided by the KSA and STAR frameworks, the report links each score to behavioral evidence and actionable recommendations. PolyInterview is publicly accessible. Its current all-account snapshot contains 101 accounts, 1,564 interview sessions, 7,665 generated questions, and 1,422 five-stage question sets. Generated questions are more closely aligned with their matched job description than with cross-role job descriptions in 93.7% of sessions. An evaluation by ten experts found strong question plans and actionable feedback.
Chinese Translation
准备求职面试对获得理想职位至关重要，但现实的实践机会仍然难以获得：真实面试不频繁，专家模拟辅导成本高昂，自我练习既缺乏适应性对话也缺乏结构化评估。现有系统通常仅通过固定问题序列、有限的沟通渠道或缺乏支持证据的反馈来部分满足这一需求。我们提出了PolyInterview，一个基于大型语言模型的沉浸式模拟面试实践平台，提供全面的多模态评估。PolyInterview利用目标职位描述和简历生成针对角色和候选人的定制问题，进行多轮口语面试，配有同步口型的数字人类面试官，提出与回答相关的后续问题，并评估回答内容、声音表达和非语言行为。四位平行评估者生成13个行为级特征，这些特征被汇总为10个评估方面和两个能力轨道。在KSA和STAR框架的指导下，报告将每个分数与行为证据和可操作建议相连接。PolyInterview是公开可访问的。目前的所有账户快照包含101个账户、1,564个面试会话、7,665个生成问题和1,422个五阶段问题集。在93.7%的会话中，生成的问题与其匹配的职位描述的相关性高于跨角色职位描述。十位专家的评估发现了强有力的问题计划和可操作的反馈。
cs.CL / 19 / 2607.10312
Polarization Detection: A Hybrid Approach with AfroXLMR-Social and DeBERTa for Low- and High-Resource Settings
极化检测：一种结合AfroXLMR-Social和DeBERTa的混合方法，适用于低资源和高资源环境
Abstract
The rapid proliferation of online polarization threatens social cohesion, necessitating robust automated detection systems that operate effectively across diverse linguistic contexts. This paper presents our system description for the POLAR Shared Task 2026, focusing on the detection and characterization of polarized discourse in English and Hausa. We propose a hybrid modeling strategy: for English binary detection, we leverage the monolingual strength of \textbf{DeBERTa}, while for Hausa and all fine-grained subtasks (Types and Manifestations), we utilize \textbf{AfroXLMR-Social}. This domain-adapted multilingual model proved critical for capturing the nuances of polarization in social media text. To further address computational constraints and data scarcity, we implement Low-Rank Adaptation (LoRA) and textual data augmentation via \texttt{nlpaug}. We report competitive results across all three subtasks, demonstrating that model selection tailored to specific subtask requirements yields the best balance of performance.
Chinese Translation
在线极化的快速传播威胁着社会凝聚力，因此迫切需要能够在多种语言环境中有效运行的强大自动检测系统。本文介绍了我们在POLAR共享任务2026中的系统描述，重点关注英语和豪萨语中极化话语的检测与特征化。我们提出了一种混合建模策略：在英语二元检测中，我们利用了 extbf{DeBERTa}的单语优势，而在豪萨语及所有细粒度子任务（类型和表现）中，我们采用了 extbf{AfroXLMR-Social}。这一领域适应的多语言模型对于捕捉社交媒体文本中极化的细微差别至关重要。为了进一步解决计算约束和数据稀缺的问题，我们实施了低秩适应（Low-Rank Adaptation, LoRA）和通过 exttt{nlpaug}进行的文本数据增强。我们在所有三个子任务中报告了具有竞争力的结果，表明针对特定子任务需求量身定制的模型选择能够实现最佳的性能平衡。
cs.CL / 20 / 2607.10317
Neutralizing Structural Inequality in the Nigerian FinTech Sector
消除尼日利亚金融科技领域的结构性不平等
Abstract
Algorithmic decision systems in financial services often rely on data proxies that inadvertently encode structural inequalities. This paper introduces a hierarchical human-AI triage model for Point of Sale fraud detection in the Nigerian FinTech sector. Adopting a We Are All Equal worldview, we address the challenge of discrimination laundering, wherein the system misinterprets infrastructure related aleatoric noise such as rural network timeouts as fraudulent intent. We implement a three-tier routing policy utilizing a calibrated ensemble model as a primary filter. The policy routes transactions characterized by epistemic uncertainty such as cold start new accounts to specialist analysts while reserving high stakes cases for a senior supervisor. To manage finite human capacity, we utilize a dynamic shadow price to ration human attention and implement a random audit mechanism to prevent human skill atrophy. Our experimental results demonstrate a statistically significant 1.88\% complementarity gap and a 24.79\% percentage point gain in fraud recall over an autonomous baseline. Crucially, the model reduces the regional performance gap from 19.43 to 2.88 percentage points, neutralizing structural bias. Hierarchical collaboration provides a robust mechanism for substantive equality of opportunity, ensuring that rural accounts are not excluded from the digital economy due to environmental brute luck.
Chinese Translation
金融服务中的算法决策系统通常依赖于数据代理，这些代理无意中编码了结构性不平等。本文提出了一种用于尼日利亚金融科技领域销售点欺诈检测的分层人机分流模型。我们采用“人人平等”的世界观，解决了歧视洗白的问题，即系统误将与基础设施相关的随机噪声（如农村网络超时）误解为欺诈意图。我们实施了一种三层路由政策，利用经过校准的集成模型作为主要过滤器。该政策将特征为认知不确定性的交易（如冷启动新账户）路由给专业分析师，同时将高风险案例保留给高级监督员。为了管理有限的人力资源，我们利用动态影子价格来合理分配人类注意力，并实施随机审计机制以防止人类技能的退化。我们的实验结果表明，与自主基线相比，欺诈召回率有统计学上显著的1.88%的互补差距和24.79个百分点的提升。重要的是，该模型将区域绩效差距从19.43个百分点减少到2.88个百分点，消除了结构性偏见。分层协作为实质性机会平等提供了一种强有力的机制，确保农村账户不会因环境运气而被排除在数字经济之外。
cs.CL / 21 / 2607.10380
CAFE: A Compound-AI Factorial Evaluation Framework
CAFE：复合人工智能因子评估框架
Abstract
We introduce CAFE (Compound-AI Factorial Evaluation), an open-source platform that brings design of experiments to the evaluation of compound AI systems (CAIS). Such systems expose many interchangeable choices - e.g. which retriever, model, or prompt - and practitioners rarely know which of them most affects answer quality. With CAFE, a practitioner registers each swappable component of a pipeline as a factor to build a factorial design over the chosen factors, run the resulting configurations, and score the answers on a shared rubric using a configurable LLM judge together with human raters. From these ratings it attributes answer-quality variance to the components and their interactions with mixed-effects models and reports effect sizes, significance, the best configuration, cost and latency trade-offs, and judge-human reliability. Whereas existing tools mostly either search for a good configuration or score outputs in isolation, CAFE also explains which component drives quality and whether an observed difference is significant. We validate CAFE on a retrieval-augmented question-answering (QA) pipeline over the HotpotQA benchmark dataset, where it recovers planted factor effects and stays calibrated under a permutation null. CAFE is released as a Python package and as a Web application.
Chinese Translation
我们介绍了CAFE（复合人工智能因子评估），这是一个开源平台，将实验设计应用于复合人工智能系统（CAIS）的评估。这类系统提供了许多可互换的选择——例如，选择哪种检索器、模型或提示——而从业者通常不知道其中哪个因素对答案质量影响最大。通过CAFE，从业者可以将管道中每个可替换组件注册为一个因子，以构建所选因子的因子设计，运行生成的配置，并使用可配置的LLM评审者和人类评审员在共享评分标准上对答案进行评分。根据这些评分，它将答案质量的方差归因于组件及其相互作用，并使用混合效应模型报告效应大小、显著性、最佳配置、成本和延迟权衡，以及评审者与人类评审员之间的可靠性。现有工具大多要么寻找良好的配置，要么孤立地对输出进行评分，而CAFE还解释了哪个组件驱动质量，以及观察到的差异是否显著。我们在HotpotQA基准数据集上验证了CAFE在检索增强问答（QA）管道中的有效性，在该数据集中，它恢复了植入的因子效应，并在置换零假设下保持了校准。CAFE已作为Python包和Web应用程序发布。
cs.CL / 22 / 2607.10386
Structured Thoughts For Improved Reasoning And Context Pruning
改进推理与上下文修剪的结构化思维
Abstract
Large language models (LLMs) excel at generating long chains of thought, but long reasoning traces are often verbose and memory-inefficient. In this work, we introduce Structured Thoughts, a framework that organizes reasoning into alternating and blocks: captures exploratory scratch work, while contains the distilled conclusion of that step. We construct a dataset of structured thoughts by segmenting reasoning traces into blocks and prompting an LLM to summarize each step into its corresponding . Fine-tuning pretrained foundation models on this reformatted data produces models that adopt the structured reasoning style, leading to performance gains of up to 8.08\% on reasoning benchmarks compared to standard SFT. The explicit structure also enables context pruning: after each / pair, the can be pruned, allowing the model to retain conclusions without keeping the full scratch work in the context. A proof-of-concept pruning implementation achieves an average of 85\% memory / context savings with an 8.67\% performance drop across mathematical tasks.
Chinese Translation
大型语言模型（LLMs）在生成长链思维方面表现出色，但长推理痕迹往往冗长且内存效率低下。在本研究中，我们提出了结构化思维（Structured Thoughts），这是一个将推理组织为交替的 和 块的框架： 捕捉探索性的草稿工作，而 则包含该步骤的提炼结论。我们通过将推理痕迹分段为 块，并提示 LLM 总结每一步到其对应的 ，构建了一个结构化思维的数据集。对预训练基础模型进行微调，使用这种重新格式化的数据，生成采用结构化推理风格的模型，与标准 SFT 相比，在推理基准测试中性能提升高达 8.08%。这种明确的结构还实现了上下文修剪：在每对 / 之后，可以修剪 ，使模型在不保留完整草稿工作的情况下保留结论。一个概念验证的修剪实现平均节省了 85% 的内存/上下文，同时在数学任务中性能下降 8.67%。
cs.CL / 23 / 2607.10390
A Stepwise Questioning Expert-Editor Multi-Agent Framework for Long-Document Summarization
一种分步提问专家-编辑多智能体框架用于长文档摘要
Abstract
Although large language models (LLMs) have shown promising potential in news summarization tasks, their performance on long-document summarization remains challenging as their length often exceeds the input limits. As the agent investment, which provide possibility to improve the inherent capabilities of LLMs. To enhance the effectiveness of long-document summarization based on LLMs, this paper proposes an expert-editor stepwise questioning multi-agent method, in which the expert and the editor guide another agent to refine the summary by posing questions on different aspects of the content and providing targeted clues for revision. We conducted experiments on two representative long-document scientific datasets and evaluated the results through widely recognized automatic metrics. The results demonstrated the effectiveness of our method.
Chinese Translation
尽管大型语言模型（LLMs）在新闻摘要任务中展现出良好的潜力，但在长文档摘要方面的表现仍然面临挑战，因为其长度往往超过输入限制。作为智能体的投资，这为提升LLMs的内在能力提供了可能性。为了增强基于LLMs的长文档摘要的有效性，本文提出了一种专家-编辑分步提问多智能体方法，其中专家和编辑通过对内容不同方面提出问题并提供针对性的修订线索，引导另一个智能体完善摘要。我们在两个具有代表性的长文档科学数据集上进行了实验，并通过广泛认可的自动评估指标对结果进行了评估。结果证明了我们方法的有效性。
cs.CL / 24 / 2607.10428
Enjoy Your Talk: A Human-Centered Benchmark for Multi-Turn Dialogue with Decoupled User Simulation, Target Modeling, and Judging
享受你的对话：一个以人为中心的多轮对话基准，采用解耦的用户模拟、目标建模和评判
Abstract
Evaluating large language models (LLMs) as multi-turn conversational partners requires probing capabilities that single-turn benchmarks miss: persona consistency, evolving intent tracking, emotional dynamics, and goal completion. We introduce EYT-Bench, a human-centered benchmark built around a three-party decoupled design: a persona-grounded user simulator, a target model that separates intent perception from response generation, and an independent third-party LLM judge with optional multi-judge ensembling. Personas are sampled from public human-curated corpora, Nemotron-Personas-USA and PersonaMem-v2, rather than synthesized, reducing LLM-induced persona bias. EYT-Bench also introduces two trajectory-level metrics: embedding-based intent drift and final-intent completion rate (FICR), inspired by tau-bench. In a 17-target x 200-dialogue evaluation, EYT-Bench reveals four findings: (i) state-of-the-art closed- and open-source models are statistically close on subjective dimensions (empathy / persona / anthropomorphism vary within <= 0.3), but differ by up to 9x on objective intent tracking; (ii) reasoning ("thinking on") sharply improves objective tracking on long-context personas (+0.47-0.50 latent-intent accuracy on Gemma-4) while leaving subjective scores nearly unchanged; (iii) persona format dominates trajectory spread, with FICR saturating above 0.95 on Nemotron-USA but spreading from 0.53 to 0.88 on PersonaMem-v2; and (iv) the warm-up effect is robust on 16/17 models (one outlier, GPT-5.5, reverses the effect), with stable rankings across alpha in [0.05, 0.15]. A cross-judge ablation using deepseek-v4-pro confirms that target rankings and final-intent satisfaction are preserved across judges.
Chinese Translation
评估大型语言模型（LLMs）作为多轮对话伙伴的能力需要探测单轮基准所忽视的能力：个性一致性、不断演变的意图追踪、情感动态和目标完成度。我们引入了EYT-Bench，这是一个以人为中心的基准，围绕三方解耦设计构建：一个基于个性的用户模拟器、一个将意图感知与响应生成分开的目标模型，以及一个独立的第三方LLM评判者，支持可选的多评判者集成。个性样本来自公共的人类策划语料库Nemotron-Personas-USA和PersonaMem-v2，而非合成，从而减少了LLM引起的个性偏差。EYT-Bench还引入了两个轨迹级别的指标：基于嵌入的意图漂移和最终意图完成率（FICR），灵感来自tau-bench。在17个目标x 200个对话的评估中，EYT-Bench揭示了四个发现：（i）最先进的闭源和开源模型在主观维度上统计上接近（同理心/个性/拟人化的变化在<=0.3以内），但在客观意图追踪上差异可达9倍；（ii）推理（“思考”）显著提高了长上下文个性上的客观追踪（Gemma-4上潜在意图准确率提高0.47-0.50），而主观评分几乎不变；（iii）个性格式主导了轨迹的分布，FICR在Nemotron-USA上饱和在0.95以上，而在PersonaMem-v2上从0.53扩展到0.88；（iv）热身效应在16/17个模型上是稳健的（一个异常值GPT-5.5逆转了该效应），在α在[0.05, 0.15]范围内的排名稳定。使用deepseek-v4-pro的跨评判者消融实验确认了目标排名和最终意图满意度在不同评判者之间得以保持。
cs.CL / 25 / 2607.10476
Hallucination Detection in Large Language Models Using Diversion Decoding
使用分流解码进行大型语言模型的幻觉检测
Abstract
Large language models (LLMs) have emerged as a powerful tool for retrieving knowledge through seamless, human-like interactions. Despite their advanced text generation capabilities, LLMs exhibit hallucination tendencies, where they generate factually incorrect statements and fabricate knowledge, undermining their reliability and trustworthiness. Multiple studies have explored methods to evaluate LLM uncertainty and detect hallucinations. However, existing approaches are often probabilistic and computationally expensive, limiting their practical applicability. In this paper, we introduce diversion decoding, a novel method for developing an LLM uncertainty heuristic by actively challenging model-generated responses during the decoding phase. Through diversion decoding, we extract features that capture the LLM's resistance to produce alternative answers and utilize these features to train a machine-learning model to develop a heuristic measure of the LLM's uncertainty. Our experimental results demonstrate that diversion decoding outperforms existing methods with significantly lower computational complexity, making it an efficient and robust solution for evaluating hallucination detection.
Chinese Translation
大型语言模型（LLMs）已成为通过无缝的人类互动检索知识的强大工具。尽管它们具备先进的文本生成能力，但LLMs表现出幻觉倾向，即生成事实不准确的陈述和虚构知识，从而削弱了它们的可靠性和可信度。多项研究探讨了评估LLM不确定性和检测幻觉的方法。然而，现有的方法往往是概率性的且计算成本高，限制了它们的实际应用。在本文中，我们提出了一种分流解码的新方法，通过在解码阶段主动挑战模型生成的响应，开发LLM不确定性启发式。通过分流解码，我们提取了捕捉LLM抵制生成替代答案的特征，并利用这些特征训练机器学习模型，以开发LLM不确定性的启发式度量。我们的实验结果表明，分流解码在计算复杂性显著降低的情况下优于现有方法，成为评估幻觉检测的高效且稳健的解决方案。
cs.CL / 26 / 2607.10480
When Reasoning Hurts Legal Drafting: The Verbalization Bottleneck in Patent Claim Generation
推理对法律文书起草的负面影响：专利权利要求生成中的语言化瓶颈
Abstract
Patent claim drafting is a challenging legal drafting task that requires technical expertise, precise linguistic control, strict adherence to formal conventions, and the preservation of complex logical relationships among claim elements. While Chain-of-Thought (CoT) prompting has been widely used to improve the reasoning capabilities of large language models (LLMs), recent evidence suggests that its benefits may be limited, or even negative, in highly structured or pattern-sensitive tasks. Therefore, this paper investigates whether CoT prompting benefits patent claim generation. We propose a task-specific CoT method for patent claim generation and evaluate its effectiveness through both automatic metrics and human expert assessment. Our results show that reasoning-enhanced prompting can improve claim quality. Moreover, we demonstrate a counter-intuitive but important empirical finding: implicit CoT, where reasoning is kept internal rather than explicitly verbalized, consistently outperforms explicit CoT. Through systematic analysis, we show that explicit CoT can introduce an unnecessary information bottleneck for claim generation. Verbalized reasoning may compromise the quality of final outputs through three specific mechanisms: abstraction of critical details, disruption of internalized generation patterns, and cascading error propagation. Our findings provide new insights into legal tasks and CoT applications.
Chinese Translation
专利权利要求起草是一项具有挑战性的法律文书起草任务，要求具备技术专长、精确的语言控制、严格遵循正式惯例，并保持权利要求元素之间复杂的逻辑关系。尽管链式思维（Chain-of-Thought, CoT）提示已被广泛应用于提升大型语言模型（LLMs）的推理能力，但最近的证据表明，在高度结构化或对模式敏感的任务中，其益处可能有限，甚至是负面的。因此，本文探讨了CoT提示是否对专利权利要求生成有益。我们提出了一种针对专利权利要求生成的特定任务CoT方法，并通过自动指标和人类专家评估来评估其有效性。我们的结果表明，增强推理的提示可以提高权利要求的质量。此外，我们展示了一个反直觉但重要的实证发现：隐性CoT，即将推理保持在内部而非明确语言化，始终优于显性CoT。通过系统分析，我们表明显性CoT可能为权利要求生成引入不必要的信息瓶颈。语言化的推理可能通过三个特定机制损害最终输出的质量：关键细节的抽象、内化生成模式的干扰，以及错误传播的级联效应。我们的发现为法律任务和CoT应用提供了新的见解。
cs.CL / 27 / 2607.10511
Articulate Intuition or Genuine Analysis? Benchmarking Epistemic Reliability in LLM-as-a-Judge Peer Reviews
清晰的直觉还是真实的分析？在 LLM 作为评审的同行评审中基准化认知可靠性
Abstract
When an LLM judge calls a peer review analytical and a human committee calls another review high quality, are they tracking the same thing? We argue they are not, and that the difference matters philosophically. We operationalise Kahneman's dual-process theory into a structured rubric for peer review and release Kahneman4Review, a benchmark of 3,563 rated reviews scored along nine theoretically motivated textual dimensions, eight bias diagnostics, and a continuous reasoning-quality score. Three findings bear on trustworthiness: decision tier is not detectably aligned with the rubric's text-grounded epistemic-quality proxy; public-showcase agentic reviews receive higher raw scores than pooled human reviews, but length and venue explain most of the gap and the samples are not paper-paired; and ICLR review-text diagnostics shift at the 2022--2023 transition, temporally coincident with widespread LLM availability but without identifying its cause. A matched function-probe pilot further shows that the rubric distinguishes textual probes designed to contrast genuine fault-finding with surface fluency. We argue that a trustworthy reliability benchmark for LLM judges must separate analytical form from epistemic function, and propose concrete design choices toward that goal. An interactive demo is available at https://huggingface.co/spaces/nuojohnchen/Kahneman4Review.
Chinese Translation
当一个 LLM 评审称某个同行评审为分析性，而一个人类委员会称另一个评审为高质量时，他们是否在追踪同样的事物？我们认为他们并不是，而且这种差异在哲学上是重要的。我们将卡尼曼的双重过程理论操作化为一个结构化的同行评审评分标准，并发布了 Kahneman4Review，这是一个包含 3,563 个评分评审的基准，按照九个理论驱动的文本维度、八个偏差诊断和一个连续的推理质量评分进行评分。三个发现与可信度相关：决策层次与评分标准的基于文本的认知质量代理没有显著对齐；公共展示的自主评审获得的原始分数高于汇总的人类评审，但长度和发表场所解释了大部分差距，并且样本没有配对；ICLR 评审文本诊断在 2022-2023 过渡期发生变化，与广泛的 LLM 可用性同时发生，但未能确定其原因。一个匹配功能探测的试点进一步表明，该评分标准区分了旨在对比真实缺陷发现与表面流畅性的文本探测。我们认为，LLM 评审的可信赖性基准必须将分析形式与认知功能分开，并提出了实现这一目标的具体设计选择。一个互动演示可在 https://huggingface.co/spaces/nuojohnchen/Kahneman4Review 获取。
cs.CL / 28 / 2607.10557
UNIBROWSE: A Data-to-Agent Framework for Multimodal BrowseComp
UNIBROWSE：一种用于多模态浏览计算的数据到代理框架
Abstract
Multimodal BrowseComp tasks require agents to combine perception, tool use, and long-horizon reasoning over dynamic web content, challenging their ability to handle compositional structure, open-world uncertainty, and multimodal integration across extended interactions. Crucially, real-world multimodal browsing involves three distinct information-flow patterns: text-only, image-to-text, and text-to-image, yet existing data construction methods cover only the text-only and image-to-text patterns, leaving text-to-image largely unaddressed and limiting agent generality and robustness. We introduce UNIBROWSE, a unified data pipeline that for the first time simultaneously generates training data covering all three patterns, augments curated knowledge graphs with live web retrieval for improved fidelity, and introduces a novel metric of exploration degree to filter low-signal instances for efficient reinforcement learning. Through this pipeline, we produce high-quality cold-start tool-use trajectories and exploration-rich QA pairs, and train a 35B-scale agent via supervised fine-tuning and exploration-aware RL.The resulting UNIBROWSE agent achieves state-of-the-art performance on multimodal BrowseComp benchmarks, attaining an average accuracy of 54.4 across five diverse benchmarks -- an improvement of 10.5 points over its base model Qwen3.5-35B-A3B -- and surpassing serveral closed-source agent workflows such as GPT-5 (42.9), Gemini-2.5 Pro (44.8), and Gemini-2.5 Flash (41.3).
Chinese Translation
多模态浏览计算任务要求代理结合感知、工具使用和对动态网页内容的长远推理，这对其处理组合结构、开放世界不确定性及在扩展交互中的多模态集成能力提出了挑战。至关重要的是，现实世界中的多模态浏览涉及三种不同的信息流模式：仅文本、图像到文本和文本到图像，而现有的数据构建方法仅涵盖了仅文本和图像到文本模式，导致文本到图像模式在很大程度上未得到解决，从而限制了代理的通用性和鲁棒性。我们提出了UNIBROWSE，这是一个统一的数据管道，首次同时生成涵盖所有三种模式的训练数据，通过实时网页检索增强策划的知识图谱以提高准确性，并引入了一种新的探索度度量来过滤低信号实例，以实现高效的强化学习。通过该管道，我们生成高质量的冷启动工具使用轨迹和丰富的探索问答对，并通过监督微调和关注探索的强化学习训练了一个35B规模的代理。最终生成的UNIBROWSE代理在多模态浏览计算基准测试中达到了最先进的性能，在五个不同基准上平均准确率为54.4，比其基础模型Qwen3.5-35B-A3B提高了10.5个百分点，并超过了多个闭源代理工作流，如GPT-5（42.9）、Gemini-2.5 Pro（44.8）和Gemini-2.5 Flash（41.3）。
cs.CL / 29 / 2607.10590
Demographic Prompting at Scale: When More Attributes Hurt LLM--Human Agreement
大规模人口统计提示：更多属性如何影响大型语言模型（LLM）与人类一致性
Abstract
We investigate how annotator demographic attributes, supplied as prompt cues, shape the alignment between large language model (LLM) predictions and human annotations across five tasks. Using five open-source LLMs, we systematically vary the number and composition of demographic components in the prompt, spanning every combination from single-attribute through full-attribute configurations. Our experiments reveal three principal findings. First, alignment consistently peaks with one to three high-signal attributes and degrades under the full attribute set, establishing a clear over-specification threshold. Second, the overall magnitude of demographic influence on human annotations does not predict which attributes improve LLM alignment; instead, both the learnability and the directional coherence of each attribute's annotation signal need to be considered jointly. Third, neuron probing reveals that specialized activation correlates with alignment gains only under coherent annotation signals, and that activation volume alone does not imply steerability. Together, these results demonstrate that demographic prompting is not a monolithic intervention: its utility is highly context-dependent, shaped by attribute signal quality, task characteristics, and model architecture.
Chinese Translation
我们研究了作为提示线索提供的标注者人口统计属性如何影响大型语言模型（LLM）预测与人类标注之间的一致性，涵盖五个任务。使用五个开源LLM，我们系统地改变了提示中人口统计成分的数量和组成，涵盖了从单一属性到全属性配置的每种组合。我们的实验揭示了三个主要发现。首先，一致性在一到三个高信号属性时达到峰值，而在全属性集下则下降，确立了一个明确的过度规范阈值。其次，人口统计对人类标注的整体影响程度并不能预测哪些属性改善LLM一致性；相反，需要共同考虑每个属性的可学习性和方向一致性。第三，神经元探测表明，专门的激活仅在一致的标注信号下与一致性提升相关，而激活量本身并不意味着可引导性。综上所述，这些结果表明人口统计提示并不是单一的干预：其效用高度依赖于上下文，受属性信号质量、任务特征和模型架构的影响。
cs.CL / 30 / 2607.10591
Non-binary bottom-up constituency parsing without arity actions
无元底层 constituency 解析的非二元方法
Abstract
Non-binary bottom-up constituency parsing is usually taken to require arity actions: reductions such as \(\textsc{Reduce-}X\#k\) specify both the mother label and the number of children to be composed. We show that this arity parameter is not a necessary transition primitive. Our parser introduces constituent labels separately and recovers reduction spans from delimiter-bounded stack configurations. In a well-formed reduction configuration, arity is uniquely determined by the active delimiter and the label marker, making it a derived property of parser state rather than an action label. This factorization removes label--arity-specific reduce actions while preserving direct construction of original non-binary trees. Experiments on PTB and CTB show that the delimiter-guided parser remains competitive with an arity-specific bottom-up baseline under the same implementation framework, with substantially smaller action inventories. Analyses further show that its predicted arity profile remains close to the gold treebanks and that high-arity constituents do not collapse when arity actions are removed.
Chinese Translation
非二元底层 constituency 解析通常被认为需要元数操作：如 \ extsc{Reduce-}X\#k 的归约操作同时指定母标签和要组合的子节点数量。我们证明了这个元数参数并不是一个必要的转换原语。我们的解析器单独引入成分标签，并从受限于分隔符的堆栈配置中恢复归约跨度。在一个良构的归约配置中，元数由活动分隔符和标签标记唯一确定，使其成为解析器状态的派生属性，而不是一个操作标签。这种因式分解去除了标签-元数特定的归约操作，同时保留了原始非二元树的直接构建。在 PTB 和 CTB 上的实验表明，在相同的实现框架下，基于分隔符的解析器与特定元数的底层基线保持竞争力，并且操作库存显著更小。分析进一步表明，其预测的元数特征与金标准树库接近，并且在去除元数操作时，高元数成分不会崩溃。
cs.CL / 31 / 2607.10626
Eval-Pair Matrix: Answer-Paired Meta-Evaluation of LLM Judges for Grounded RAG
Eval-Pair矩阵：基于答案配对的LLM评审的元评估用于基础RAG
Abstract
LLM-as-a-judge evaluation is widely used for retrieval-augmented generation (RAG), but reusing the same model family as both generator and judge makes self-leniency difficult to identify. We introduce Eval-Pair Matrix, a controlled meta evaluation protocol for source-grounded RAG. Starting from GaRAGe questions and grounding passages, we induce one hidden answer-causal contradiction per record, generate answers from perturbed passages with GPT, Grok, and Gemini models, and then use the same models as blind judges to evaluate each answer against the original passages. The experiment contains 300 core records, 897 labeled generator outputs, and 2,683 judge verdicts in a crossed 3 x 3 matrix; the primary analysis uses 275 fully validated records. Instead of comparing diagonal and off-diagonal cells across different answers, we estimate same-model effects by pairing judges on the exact same candidate answer. This changes the interpretation: diagonal and off diagonal F1 are similar, and the paired same-model recall effect is near zero (-0.5 pp; 95% cluster bootstrap CI [-2.7, +1.7]). The only robust paired gap is lower matching-judge flagging for answers that avoided the induced claim (-4.3 pp). A targeted human evaluation finds that reviewed apparent false positives are alternate source-error detections, mistakes in labeling whether the induced claim was adopted, or unclear cases; none were adjudicated as genuine false alarms. The lesson is methodological: RAG judge studies should report full matrices, answer-paired effects, behavior strata, and label-task alignment.
Chinese Translation
LLM作为评审的评估在检索增强生成（RAG）中被广泛使用，但将同一模型家族同时用作生成器和评审者使得自我宽容性难以识别。我们引入了Eval-Pair矩阵，这是一种针对源基础RAG的受控元评估协议。从GaRAGe问题和基础段落出发，我们为每条记录引入一个隐藏的答案因果矛盾，从扰动段落中生成答案，使用GPT、Grok和Gemini模型，然后使用相同的模型作为盲评审者对每个答案与原始段落进行评估。实验包含300个核心记录、897个标记的生成器输出和2683个评审裁决，形成一个交叉的3 x 3矩阵；主要分析使用275个完全验证的记录。我们不是比较不同答案的对角线和非对角线单元，而是通过将评审者配对在完全相同的候选答案上来估计同模型效应。这改变了解释：对角线和非对角线的F1相似，而配对的同模型召回效应接近于零（-0.5 pp；95%聚类自助法置信区间[-2.7, +1.7]）。唯一稳健的配对差距是对于避免诱导主张的答案，匹配评审标记较低（-4.3 pp）。针对性的人工评估发现，审查的明显假阳性是替代源错误检测、标记诱导主张是否被采纳的错误或不明确的案例；没有一个被裁定为真正的假警报。这个教训是方法论性的：RAG评审研究应报告完整的矩阵、答案配对效应、行为层次和标签任务对齐。
cs.CL / 32 / 2607.10628
Anamnesis: An Open-Source Platform for Large-Scale Backstory-Conditioned Survey Simulation
Anamnesis：一个用于大规模背景故事条件调查模拟的开源平台
Abstract
We present Anamnesis, an interactive system for demographically controllable survey simulation using large language models. Open-source, and designed for non-technical users/researchers, Anamnesis enables the prototyping and stress-testing of survey instruments on virtual populations rather than real human subjects. The platform operationalizes the recently introduced Anthology and Alterity frameworks, which use structured narrative backstories to condition model responses, within a unified web interface. It supports open-ended generation, probabilistic demographic resampling, and multimodal (image and audio) surveys. We evaluate the system through two case studies: (1) replicating segments of Pew Research Center's American Trends Panel (ATP) on political typology and biomedical issues and (2) emulating human preference in the New Yorker Caption Contest. In both cases, Anamnesis produces opinion distributions that more closely match real-world survey data than standard persona-prompting baselines, offering a transparent, reproducible, and open-source alternative to proprietary simulation services.
Chinese Translation
我们介绍了Anamnesis，一个使用大型语言模型进行人口统计可控调查模拟的互动系统。Anamnesis是开源的，旨在为非技术用户/研究人员提供服务，它使得在虚拟人群而非真实人类受试者上原型设计和压力测试调查工具成为可能。该平台在统一的网络界面中实现了最近引入的Anthology和Alterity框架，这些框架使用结构化叙事背景故事来调节模型响应。它支持开放式生成、概率性人口统计重采样以及多模态（图像和音频）调查。我们通过两个案例研究评估该系统：（1）复制皮尤研究中心（Pew Research Center）关于政治类型和生物医学问题的美国趋势面板（American Trends Panel，ATP）部分内容；（2）模拟《纽约客》标题比赛中的人类偏好。在这两种情况下，Anamnesis产生的意见分布与真实世界调查数据更为接近，相较于标准的人物提示基线，提供了一种透明、可重复且开源的替代方案，取代了专有的模拟服务。
cs.CL / 33 / 2607.10645
MafiaScope: Non-Invasive, Time-Resolved Belief Probing for LLM Agents in Social Deduction Games
MafiaScope：非侵入性、时间分辨的信念探测用于社交推理游戏中的大型语言模型代理
Abstract
An LLM agent's public behaviour reveals little about its social reasoning: an agent that votes correctly may be guessing, and an agent that lies well leaves no trace of what it actually believes. We present MafiaScope, an open testbed that turns the social deduction game Mafia into a measurement instrument for machine Theory of Mind. After every public utterance, every agent privately answers a configurable set of structured probe questions; the answers never re-enter the game and are scored automatically against the ground truth the engine knows. An interactive visualizer renders the belief trajectories: impersonate mode shows the game as one agent sees it, panels chart timeline-aligned accuracy and calibration, and counterfactual replay forks any recorded step. In a 32-game DeepSeek case study with 13{,}815 parsed probe answers, stated confidence is poorly calibrated, with expected calibration error 0.17, agents over-predict being suspected 1.5 times, and a 30-fork replay experiment walks the counterfactual replay workflow end to end. Engine, viewer and a corpus of 200+ cross-model games are released under an open licence; live demo: https://karpovilia.github.io/mafiascope/; screencast: https://vimeo.com/1208920221.
Chinese Translation
大型语言模型（LLM）代理的公共行为对其社交推理揭示甚少：一个正确投票的代理可能只是在猜测，而一个善于撒谎的代理则不会留下其真实信念的痕迹。我们提出了MafiaScope，一个开放的测试平台，将社交推理游戏“黑手党”转变为机器心智理论的测量工具。在每次公共发言后，每个代理私下回答一组可配置的结构化探测问题；这些答案不会重新进入游戏，并且会根据引擎所知道的真实情况自动评分。一个交互式可视化工具呈现信念轨迹：伪装模式展示了游戏在一个代理眼中的样子，面板绘制了时间线对齐的准确性和校准情况，而反事实重放则分叉任何记录的步骤。在一项包含13,815个解析探测答案的32场DeepSeek案例研究中，所声明的置信度校准较差，预期校准误差为0.17，代理对被怀疑的预测过高1.5倍，而一个30次分叉的重放实验完整演示了反事实重放工作流程。引擎、查看器以及200多个跨模型游戏的语料库已在开放许可下发布；实时演示：https://karpovilia.github.io/mafiascope/；视频演示：https://vimeo.com/1208920221。
cs.CL / 34 / 2607.10647
Knowledge Distillation for Automated AI Tutor Evaluation
用于自动化人工智能导师评估的知识蒸馏
Abstract
The rapid integration of Large Language Models (LLMs) into K-12 and higher education has outpaced the development of reliable methods for evaluating their pedagogical quality. As the research community starts to explore the space of automating evaluation of AI tutors, we introduce FATE (FLC AI Tutor Evaluator), a specialized 8B-parameter language model designed to evaluate AI tutors. Aligned with the four core evaluation tracks from the BEA 2025 Shared Task, our model assesses pedagogical ability across Mistake Identification, Mistake Location, Guidance, and Actionability. Because pedagogical evaluation is a specialized task with limited labeled data, we leverage knowledge distillation from a frontier LLM to generate additional supervision, yielding absolute performance gains up to 22.63 percentage points. Finally, we demonstrate FATE's utility as an automated evaluator by benchmarking instructional responses generated by popular commercial models, including ChatGPT, Claude, Gemini, and DeepSeek. On average, we have found that Gemini 2.5 Flash perfomed best (82.88%), then ChatGPT 5.5 Instant (80.75%), followed by DeepSeek V4 Flash (80.13%) and Claude Sonnet 4.6 (74.00%).
Chinese Translation
大型语言模型（LLMs）在K-12及高等教育中的快速应用，已经超越了可靠评估其教学质量的方法的发展。随着研究界开始探索自动化评估人工智能导师的领域，我们引入了FATE（FLC AI Tutor Evaluator），这是一个专门设计的8B参数语言模型，用于评估人工智能导师。我们的模型与BEA 2025共享任务的四个核心评估轨道相一致，评估教学能力，包括错误识别、错误定位、指导和可操作性。由于教学评估是一项专业任务，标注数据有限，我们利用前沿LLM的知识蒸馏生成额外的监督，从而实现了高达22.63个百分点的绝对性能提升。最后，我们通过基准测试流行商业模型生成的教学响应，展示了FATE作为自动评估器的实用性，包括ChatGPT、Claude、Gemini和DeepSeek。我们发现，Gemini 2.5 Flash的表现最佳（82.88%），其次是ChatGPT 5.5 Instant（80.75%），然后是DeepSeek V4 Flash（80.13%）和Claude Sonnet 4.6（74.00%）。
cs.CL / 35 / 2607.10661
Unlocking Parallelism in Autoregressive Language Models via Speculative Decoding with Progressive Tree Drafting
通过渐进树草拟的推测解码解锁自回归语言模型中的并行性
Abstract
Speculative decoding has significantly accelerated Large Language Model (LLM) inference by alleviating memory-bound bottlenecks. However, traditional speculative decoding typically relies on auxiliary draft modules, incurring significant training and communication overhead. Although recent methods attempt to generate drafts within the target model itself, they often fail to fully exploit its latent parallel capacity due to a lack of structural coordination. In this paper, we propose \textbf{Progressive Tree Drafting (PTD)}, which employs a structured, guided parallel drafting strategy to harness the model's parallel potential. By coupling a progressive tree structure with a stepwise pruning mechanism, PTD actively guides the LLM to explore multiple semantic paths in a single forward pass, ensuring both draft diversity and coherence. Experiments demonstrate that PTD achieves up to $2\times$ decoding speedup across various benchmarks while remaining training-free and model-agnostic. Our code is available at: https://github.com/MINE-USTC/PTD.
Chinese Translation
推测解码通过缓解内存瓶颈显著加速了大型语言模型（LLM）的推理。然而，传统的推测解码通常依赖于辅助草拟模块，导致显著的训练和通信开销。尽管最近的方法试图在目标模型内部生成草拟，但由于缺乏结构协调，它们往往未能充分利用模型的潜在并行能力。本文提出了 extbf{渐进树草拟（Progressive Tree Drafting, PTD）}，采用结构化的引导并行草拟策略，以利用模型的并行潜力。通过将渐进树结构与逐步修剪机制相结合，PTD积极引导LLM在单次前向传播中探索多个语义路径，确保草拟的多样性和一致性。实验表明，PTD在各种基准测试中实现了高达$2 imes$的解码加速，同时保持无训练和模型无关。我们的代码可在以下链接获取：https://github.com/MINE-USTC/PTD。
cs.CL / 36 / 2607.10715
A Corpus of Persuasion Techniques in Slavic Languages
斯拉夫语言中的说服技巧语料库
Abstract
Persuasion techniques are powerful rhetorical devices used to sway public opinion in a wide range of media. We present a new corpus of persuasion techniques, focusing on Slavic languages. The corpus contains documents in Bulgarian, Polish, and Russian, annotated with persuasion techniques at the coarse-grained text-span level and fine-grained sentence level. The techniques are drawn from a taxonomy of 25 fine-grained persuasion techniques, grouped under six broad categories of rhetorical persuasion strategies. The corpus contains approximately 7500 text spans from 222 documents that cover topics hotly debated at the national and international levels. We describe the corpus creation process, provide detailed statistics, and examine correlations between topics and persuasion techniques. We use classic ML-based and generative AI-based models to provide baselines and benchmark results for the detection and classification of persuasion techniques at the text-span level and sentence level.
Chinese Translation
说服技巧是用于影响公众舆论的强大修辞工具，广泛应用于各种媒体。我们提出了一个新的说服技巧语料库，重点关注斯拉夫语言。该语料库包含保加利亚语、波兰语和俄语的文档，并在粗粒度文本跨度和细粒度句子层面上进行了说服技巧的标注。这些技巧源自于25种细粒度说服技巧的分类法，分为六大类修辞说服策略。该语料库包含约7500个文本跨度，来自222个涵盖国家和国际层面热议话题的文档。我们描述了语料库的创建过程，提供了详细的统计数据，并考察了话题与说服技巧之间的相关性。我们使用经典的基于机器学习（ML）和生成式人工智能（AI）模型提供基准和基准结果，以检测和分类文本跨度和句子层面的说服技巧。
cs.CL / 37 / 2607.10745
The First ChineseBabyLM Challenge: training data-efficient and cognitively plausible language models for Chinese
首届ChineseBabyLM挑战赛：训练数据高效且认知上合理的中文语言模型
Abstract
This paper describes the first ChineseBabyLM challenge, which will be held in the 2026 NLPCC conference. The challenge calls for researchers to train language models from scratch with 100 million Chinese tokens and evaluates the models on 3 tracks of tasks: NLU, cognitive alignment and Hanzi knowledge. There is no restriction on tokenizer, model architecture and the number of training epochs. Details of the challenge can be found in https://chinese-babylm.github.io/.
Chinese Translation
本文描述了首届ChineseBabyLM挑战赛，该挑战赛将在2026年NLPCC会议上举行。挑战赛邀请研究人员从零开始训练语言模型，使用1亿个中文标记，并在三个任务轨道上评估模型：自然语言理解（NLU）、认知对齐和汉字知识。对分词器、模型架构和训练轮次没有限制。挑战赛的详细信息可以在https://chinese-babylm.github.io/找到。
cs.CL / 38 / 2607.10798
Trust Before Fusion: QIMG-7 and Source-Aware Resolution for Polluted Multimodal RAG
融合前的信任：QIMG-7及源感知分辨率用于污染的多模态检索增强生成
Abstract
Multimodal retrieval-augmented generation (RAG) is often evaluated with clean evidence, yet real retrieval can return topically relevant but unreliable content: false text and misleading images from corrupted metadata, entity swaps, typographic overlays, semantic edits, adversarial patches, blends, or style transfer. We introduce QIMG-7, a controlled benchmark for multimodal retrieval pollution in multi-sentence factual QA, spanning four datasets, seven image-attack families, and 16 paired clean/polluted regimes, for 1,760 evaluation rows per method. Across four generator/gate stacks, naive multimodal fusion is brittle: in the main gpt-4o-mini stack, Full-MM support drops from 0.908 with clean text to 0.490 with polluted text, often making Parametric fallback safer than retrieval. We propose source-aware trust resolution (SATR), a training-free approach that compares Parametric, Text-only, and Full-MM candidate answers and selects among candidate answers or falls back based on source reliability. The Field-Selector variant achieves the best balanced score, 0.816, improving over Full-MM by 11.7 points and over the Cascaded Router by 2.7 points. Ablations show that, in this text-first setting, explicit text-reliability modeling is the dominant driver of these gains. Overall, in text-first factual QA with multimodal retrieval conflict, our results support selective trust rather than unconditional fusion. Artifacts are available at https://github.com/SaadElDine/Trust_Before_Fusion.
Chinese Translation
多模态检索增强生成（RAG）通常在干净证据下进行评估，然而实际检索可能返回主题相关但不可靠的内容：来自损坏元数据的虚假文本和误导性图像、实体交换、排版覆盖、语义编辑、对抗性补丁、混合或风格转移。我们引入了QIMG-7，这是一个针对多句事实问答中多模态检索污染的受控基准，涵盖四个数据集、七种图像攻击类别和16种配对的干净/污染模式，每种方法有1,760个评估行。在四个生成器/门控堆栈中，简单的多模态融合是脆弱的：在主要的gpt-4o-mini堆栈中，干净文本的Full-MM支持从0.908下降到污染文本的0.490，通常使得参数化回退比检索更安全。我们提出了源感知信任分辨率（SATR），这是一种无训练的方法，比较参数化、仅文本和全多模态候选答案，并根据源的可靠性在候选答案中进行选择或回退。Field-Selector变体实现了最佳平衡得分0.816，比Full-MM提高了11.7分，比级联路由器提高了2.7分。消融实验表明，在这种文本优先的设置中，显式的文本可靠性建模是这些增益的主要驱动因素。总体而言，在多模态检索冲突的文本优先事实问答中，我们的结果支持选择性信任而非无条件融合。相关材料可在https://github.com/SaadElDine/Trust_Before_Fusion获取。
cs.CL / 39 / 2607.10805
Diagnosing and Mitigating Thinking Collapse in On-Policy Self-Distillation
诊断与缓解在线自蒸馏中的思维崩溃
Abstract
On-Policy Self-Distillation (OPSD) has emerged as a crucial paradigm for enhancing and aligning Large Language Models (LLMs). However, in complex reasoning tasks, OPSD paradoxically degrades downstream performance. In this paper, we systematically investigate this pathology and identify a severe optimization trap we define as \textbf{Thinking Collapse} -- a sharp decline in the model's native intermediate reasoning behavior, measured by epistemic-token density (ET per 1k). Through entropy-based gradient masking and token-level target analysis, we show that this collapse is triggered by aggressive teacher gradients at high-student-entropy decision forks, where student epistemic tokens are frequently suppressed into teacher non-epistemic targets and are highly concentrated in high pointwise student-teacher divergence regions. To resolve this optimization pathology, we propose \textbf{Adaptive Dual-Perspective OPSD (AD-OPSD)}, a robust control framework that dynamically moderates the self-distillation objective. AD-OPSD selectively anchors high-suppression-risk sandboxed tokens to a reference prior derived from the frozen base model via an asymmetrical pointwise divergence gate, preserving native thinking capacity while retaining OPSD's error-correcting power. Extensive experiments across competitive mathematical benchmarks show that AD-OPSD improves over standard OPSD by up to \textbf{+4.1\%} absolute average accuracy across diverse model scales and datasets. Further analysis demonstrates that AD-OPSD mitigates thinking collapse and generalizes robustly to different post-training paradigms.
Chinese Translation
在线自蒸馏（On-Policy Self-Distillation, OPSD）已成为增强和对齐大型语言模型（Large Language Models, LLMs）的重要范式。然而，在复杂推理任务中，OPSD 反而会导致下游性能下降。本文系统地研究了这一病理现象，并识别出一种严重的优化陷阱，我们将其定义为 extbf{思维崩溃}——模型固有中间推理行为的急剧下降，通过知识性标记密度（epistemic-token density, ET per 1k）进行测量。通过基于熵的梯度掩蔽和标记级目标分析，我们表明，这种崩溃是由在高学生熵决策分叉处的激进教师梯度引发的，在这些地方，学生的知识性标记经常被压制到教师的非知识性目标中，并且高度集中在高点对点学生-教师发散区域。为了解决这一优化病理，我们提出了 extbf{自适应双视角在线自蒸馏（Adaptive Dual-Perspective OPSD, AD-OPSD）}，这是一个动态调节自蒸馏目标的稳健控制框架。AD-OPSD 选择性地将高抑制风险的沙盒标记锚定到通过不对称点对点发散门从冻结的基础模型中推导出的参考先验，既保持了固有的思维能力，又保留了OPSD的纠错能力。针对竞争性数学基准的广泛实验表明，AD-OPSD 在不同模型规模和数据集上相较于标准OPSD 提高了高达 extbf{+4.1\%}的绝对平均准确率。进一步分析表明，AD-OPSD 缓解了思维崩溃，并在不同的后训练范式中具有良好的泛化能力。
cs.CL / 40 / 2607.10806
Abstractiveness Metrics for Evaluating Text Summarization: A Refined Formulation with Empirical Validation
评估文本摘要的抽象性指标：经过实证验证的精炼公式
Abstract
Quantifying abstractiveness in generated summaries is essential for evaluating summarization models beyond surface-level metrics like ROUGE. We introduce Reference Abstraction (RA), Summary Abstraction (SA), and Abstraction Ratio (AR) -- a set of principled heuristic metrics that measure how much a summary diverges from extractive copying of the source text. The formulation uses the harmonic mean of document lengths modulated by a cubic non-overlap factor, yielding dimensionally consistent, bounded output with non-linear sensitivity to the extractive-abstractive boundary. Evaluation on 100 XSUM documents across four summarization models (BART-large-cnn, Pegasus-xsum, DistilBart, MT5-small) demonstrates that the metrics successfully discriminate between extractive models (SA ~ 0.12-0.26) and abstractive models (SA ~ 0.96-1.77), and that the Abstraction Ratio identifies summaries requiring manual evaluation for potential hallucination. Code and results are available at https://github.com/katweNLP/AbstractionStudy.
Chinese Translation
量化生成摘要的抽象性对于评估摘要模型至关重要，超越了像 ROUGE 这样的表面级指标。我们引入了参考抽象性（Reference Abstraction, RA）、摘要抽象性（Summary Abstraction, SA）和抽象比率（Abstraction Ratio, AR）——一组原则性启发式指标，用于测量摘要与源文本提取性复制的偏离程度。该公式使用文档长度的调和平均值，并通过立方非重叠因子进行调节，从而产生维度一致、有限的输出，并对提取性与抽象性边界具有非线性敏感性。在四个摘要模型（BART-large-cnn、Pegasus-xsum、DistilBart、MT5-small）上对 100 个 XSUM 文档的评估表明，这些指标成功区分了提取性模型（SA ~ 0.12-0.26）和抽象性模型（SA ~ 0.96-1.77），并且抽象比率能够识别出需要人工评估以防止潜在幻觉的摘要。代码和结果可在 https://github.com/katweNLP/AbstractionStudy 获取。
cs.CL / 41 / 2607.10825
Large Language Models for Token-Efficient and Semantic-Preserving Opinion Summarization
用于高效令牌和语义保留的意见摘要的大型语言模型
Abstract
Opinionated text - spanning product reviews, hotel feedback, and social posts - captures rich signals about user experiences, preferences, and concerns. However, the scale, redundancy, and imbalance of such corpora make it challenging to analyze opinions effectively, particularly when the goal is to generate summaries that remain faithful to the diversity of viewpoints expressed. This paper presents a framework that preserves semantics in LLM-based opinion summarization while minimizing token usage. We combine multidimensional classification (e.g., sentiment, topics) with a family of stratified sampling strategies to select compact yet representative subsets of opinions before prompting the LLM. Tailored prompts then produce balanced summaries that surface the salient aspects expressed in the opinions (e.g., strengths and weaknesses of products/hotels). Experiments on Amazon product reviews, Tripadvisor hotel reviews, and X/Twitter posts demonstrate that our method significantly reduces token usage and computational cost while consistently outperforming traditional AI-based and standard LLM summarization baselines in terms of content coverage, balance, and semantic preservation.
Chinese Translation
意见文本——涵盖产品评论、酒店反馈和社交媒体帖子——捕捉了关于用户体验、偏好和关注点的丰富信号。然而，这类语料库的规模、冗余性和不平衡性使得有效分析意见变得具有挑战性，特别是在目标是生成忠实于表达的多样化观点的摘要时。本文提出了一种框架，在基于大型语言模型（LLM）的意见摘要中保留语义，同时最小化令牌使用。我们结合多维分类（例如，情感、主题）与一系列分层抽样策略，以选择紧凑而具有代表性的意见子集，然后再提示LLM。定制的提示生成平衡的摘要，突显出意见中表达的显著方面（例如，产品/酒店的优缺点）。在亚马逊产品评论、Tripadvisor酒店评论和X/Twitter帖子上的实验表明，我们的方法显著减少了令牌使用和计算成本，同时在内容覆盖、平衡性和语义保留方面始终优于传统的基于人工智能的和标准的LLM摘要基线。
cs.CL / 42 / 2607.10846
Quantifying the Sources of Instability in LLM-Based Stance Analysis of Public Discourse
量化基于大型语言模型的公共话语立场分析中的不稳定性来源
Abstract
Computational social science increasingly relies on automated preprocessing pipelines -- speaker diarization, ASR transcript cleaning, sentence segmentation -- to convert raw media into analyzable text. When these pipelines produce different outputs from the same input, two distinct sources of instability can arise: the preprocessing pipeline itself (diarization method, segmentation rules) and the downstream measurement instrument (LLM annotation vs.\ keyword lexicon). Using 256 YouTube interviews across 41 public figures from five domains, we compare two speaker-diarization pipelines and two measurement methods, all targeting the coupling between affective valence and epistemic modality. We find that (1) preprocessing pipeline sensitivity is concentrated in speakers with limited video samples (N $\leq 5$); for the four best-sampled speakers (N $\geq 16$), the mean absolute pipeline-induced change in $r(\text{neg}, \text{emph})$ is only $0.13$; (2) cross-method disagreement is larger and more systematic -- the LLM and keyword-lexicon methods assign opposite coupling directions to several well-sampled speakers, even within the same preprocessing pipeline; and (3) aggregate valence proportions are highly stable ($|\Delta p(\text{neg})| < 6$pp) regardless of pipeline or method, masking both sources of instability. The contribution is a diagnostic framework that separates pipeline effects from measurement effects: researchers studying cross-dimensional relationships in interview data should verify that their conclusions are robust to both sources of variation, with particular attention to measurement method choice.
Chinese Translation
计算社会科学日益依赖自动化预处理流程——说话者分离、自动语音识别（ASR）转录清理、句子分割——将原始媒体转换为可分析的文本。当这些流程对相同输入产生不同输出时，可能会出现两种不同的不稳定性来源：预处理流程本身（分离方法、分割规则）和下游测量工具（大型语言模型（LLM）注释与关键词词典）。通过分析来自五个领域的41位公共人物的256个YouTube访谈，我们比较了两种说话者分离流程和两种测量方法，所有目标都是情感效价与认识模式之间的耦合。我们发现：（1）预处理流程的敏感性集中在样本视频有限的说话者（N ≤ 5）；对于样本量最多的四位说话者（N ≥ 16），由流程引起的$r( ext{neg}, ext{emph})$的平均绝对变化仅为$0.13$；（2）跨方法的不一致性更大且更具系统性——LLM和关键词词典方法对几位样本量充足的说话者分配了相反的耦合方向，即使在同一预处理流程内；（3）无论流程或方法如何，聚合情感比例高度稳定（$| ext{Δ}p( ext{neg})| < 6$pp），掩盖了两种不稳定性来源。该研究贡献了一个诊断框架，将流程效应与测量效应分开：研究访谈数据中跨维度关系的研究者应验证他们的结论对这两种变异来源的稳健性，特别关注测量方法的选择。
cs.CL / 43 / 2607.10849
Capabilities of Claude Fable 5 on Biomedical Challenge Problems
Claude Fable 5 在生物医学挑战问题上的能力
Abstract
Frontier language models are increasingly evaluated on biomedical benchmarks, but two problems undermine most published evaluations: legacy benchmarks are near-saturated, and open-ended responses are graded by other language models. We evaluate Claude Fable 5, Anthropic's most capable publicly available model, across eight biomedical benchmarks, four text and four multimodal, using deterministic scoring against fixed answer keys throughout. We include two Claude predecessors and GPT-5 as baselines. Refusal is tracked as a distinct outcome in every result table. That decision produces the paper's central finding. Fable 5 refuses between 8.0% and 99.4% of questions depending on the benchmark, a pattern absent in both predecessors and in GPT-5. Once refused items are excluded from the denominator, Fable 5's accuracy exceeds or meets every other model on every benchmark in this study. We identify two distinguishable refusal patterns: one concentrating in basic-science and mechanism content across MedQA and MedXpertQA MM, confirmed independently on two benchmarks using each benchmark's own category labels; and a separate disease-domain pattern on RareBench, where inborn metabolic disease presentations are refused near-universally while adult-onset autoimmune presentations are not. The primary constraint on Fable 5's biomedical usefulness is willingness to engage, not capability once it does.
Chinese Translation
前沿语言模型越来越多地在生物医学基准上进行评估，但有两个问题削弱了大多数已发布评估的有效性：遗留基准几乎饱和，开放式响应由其他语言模型进行评分。我们在八个生物医学基准上评估了Claude Fable 5，这是Anthropic公司最强大的公开可用模型，涵盖四个文本和四个多模态基准，使用确定性评分与固定答案键进行比较。我们将两个Claude的前身和GPT-5作为基线。在每个结果表中，拒绝被作为一个独立的结果进行跟踪。这个决定产生了论文的核心发现。Fable 5在不同基准下拒绝了8.0%到99.4%的问题，这一模式在其前身和GPT-5中均未出现。一旦将拒绝的项目排除在分母之外，Fable 5的准确性在本研究中的每个基准上均超过或等于其他模型。我们识别出两种可区分的拒绝模式：一种集中在MedQA和MedXpertQA MM的基础科学和机制内容上，独立于两个基准使用每个基准自己的类别标签进行确认；另一种是在RareBench上的疾病领域模式，其中先天性代谢疾病的表现几乎被普遍拒绝，而成人发病的自身免疫表现则没有。Fable 5在生物医学应用中的主要限制是参与的意愿，而不是一旦参与后的能力。
cs.CL / 44 / 2607.10911
The Nuts and Bolts of Natural Language to SQL Translation: A Systematic Analysis of Model Pipeline Optimisation Approaches and their Interactions
自然语言到SQL翻译的基本原理：模型管道优化方法及其相互作用的系统分析
Abstract
In the age of large language models, Natural Language to SQL (NL2SQL) translation remains an open problem with many useful applications. We explore interactions between several NL2SQL pipeline extensions to inspire development of more lightweight models. Specifically, we integrate the NatSQL intermediate representation, include a preprocessing step and a fine-tuning step based on synthetic data, and develop a novel reranker model to improve SQL selection in the final beam. We perform an ablation study supplemented by a Shapley analysis of these different components integrated with two backbone architectures, SmBoP and RASAT. We find that simply combining all of them does not lead to best results, but that their impact depends on their interactions with the baseline system, as well as each other.
Chinese Translation
在大型语言模型的时代，自然语言到SQL（NL2SQL）翻译仍然是一个开放性问题，具有许多实用应用。我们探讨了多个NL2SQL管道扩展之间的相互作用，以激发更轻量级模型的发展。具体而言，我们整合了NatSQL中间表示，包含了基于合成数据的预处理步骤和微调步骤，并开发了一种新颖的重排序模型，以改善最终束中的SQL选择。我们进行了消融研究，并补充了对这些不同组件的Shapley分析，这些组件与两个主干架构SmBoP和RASAT集成。我们发现，简单地将它们结合并不会导致最佳结果，但它们的影响取决于与基线系统及彼此之间的相互作用。
cs.CL / 45 / 2607.11012
EasyOPD: An Easy-to-use On-Policy Distillation Framework for Large Language Models
EasyOPD：一个易于使用的在线策略蒸馏框架用于大型语言模型
Abstract
Conventional language-model distillation often relies on fixed teacher-generated data, which may not cover the states encountered by an evolving student policy. On-policy distillation (OPD) instead collects teacher or evaluator supervision on student-generated rollouts. However, existing OPD methods differ substantially in supervision form, tokenizer compatibility, teacher access, and supervision granularity, leading to fragmented implementations that are difficult to reproduce and extend. We present \textsc{EasyOPD}, an on-policy distillation framework built on verl, a distributed reinforcement-learning framework for large language models. \textsc{EasyOPD} separates user-side configuration, method-specific supervision logic, and verl-based execution. Its method modules connect to the shared backend through extension boundaries for loss construction, rollout metadata, reward processing, tokenizer alignment, and teacher-side computation. We instantiate representative methods for three OPD settings -- cross-tokenizer OPD, on-policy self-distillation, and step-wise OPD. Experiments on reasoning, code-generation, scientific-knowledge, and tool-use benchmarks show that these implementations can be executed through the same verl-based backend while retaining their method-specific objectives and task-dependent performance profiles. We release \textsc{EasyOPD} with runnable YAML configurations, documentation, and an installable demonstration package and video.
Chinese Translation
传统的语言模型蒸馏通常依赖于固定的教师生成数据，这可能无法覆盖不断发展的学生策略所遇到的状态。在线策略蒸馏（On-policy Distillation, OPD）则收集教师或评估者对学生生成的回滚数据的监督。然而，现有的 OPD 方法在监督形式、分词器兼容性、教师访问权限和监督粒度等方面存在显著差异，导致实现碎片化，难以复现和扩展。我们提出了 extsc{EasyOPD}，一个基于 verl 的在线策略蒸馏框架，verl 是一个用于大型语言模型的分布式强化学习框架。 extsc{EasyOPD} 将用户侧配置、特定方法的监督逻辑和基于 verl 的执行分离。其方法模块通过扩展边界连接到共享后端，以进行损失构建、回滚元数据、奖励处理、分词器对齐和教师侧计算。我们为三种 OPD 设置实例化了代表性方法——跨分词器 OPD、在线自蒸馏和逐步 OPD。在推理、代码生成、科学知识和工具使用基准上的实验表明，这些实现可以通过相同的基于 verl 的后端执行，同时保留其特定方法的目标和任务依赖的性能特征。我们发布了 extsc{EasyOPD}，并附带可运行的 YAML 配置、文档以及一个可安装的演示包和视频。
cs.CL / 46 / 2607.11020
Can a Language Model Learn Facts Continually in Its Weights?
语言模型能否在其权重中持续学习事实？
Abstract
Continual learning promises a language model that keeps acquiring knowledge after training, with each new fact written into its weights. Whether weight writes can support accumulation remains undecided. We follow invented facts written into Qwen3 models from creation through sequences of twenty to one hundred later writes, using held-out questions of five types, with the original model given the fact in its prompt as the reference. Across these experiments, the breadth of the training data determines the kind of knowledge created. Bare-statement training produces recitation, while diverse restatements reduce the recitation-to-use gap from 27.4 to 5.4 points without showing the model a conclusion. This difference carries into later writes: after twenty sequential writes, bare-statement facts retain 1% accuracy while facts written from broad study data retain 46%. We also find that facts can be behaviourally forgotten without being erased. Forgotten facts keep most of the log-probability added by their write, and under bare-statement training 70% of wrong answers about them contain the most recently written fact. The same writes barely degrade the model's use of facts in context, and a forgotten study fact supplied in the prompt recovers to 77-80% on its questions. These results describe knowledge that is stored but question-keyed: later writes redirect the questions that reached it. Damage to unrelated abilities tracks KL divergence from the original model, and the later writes cause interference regardless of how the earlier fact was stored. Broad data can create usable knowledge, and a frozen reference can preserve capability, but no intervention we tested, including those built on accurate local measurements of each write, keeps earlier facts reachable. When facts must be composed or survive later writes, the reliable channel is context rather than the weights.
Chinese Translation
持续学习承诺使语言模型在训练后不断获取知识，每个新事实都写入其权重中。然而，权重的写入是否能够支持知识的积累尚未确定。我们跟踪了从创建到后续二十到一百次写入过程中在 Qwen3 模型中写入的虚构事实，使用五种类型的保留问题，原始模型在提示中给出事实作为参考。在这些实验中，训练数据的广度决定了所创造知识的类型。仅有陈述的训练产生了简单的背诵，而多样化的重述将背诵与使用之间的差距从 27.4 降低到 5.4 个点，而未向模型展示结论。这一差异延续到后续写入：在二十次连续写入后，仅有陈述的事实保持 1% 的准确率，而来自广泛研究数据的事实保持 46%。我们还发现，事实可以在行为上被遗忘而不被抹去。被遗忘的事实保留了其写入所增加的大部分对数概率，在仅有陈述的训练下，关于它们的错误答案中有 70% 包含最近写入的事实。这些写入几乎没有降低模型在上下文中使用事实的能力，而在提示中提供的被遗忘的研究事实在其问题上恢复到 77-80%。这些结果描述了存储的知识，但与问题相关：后续写入重新引导了到达它的问题。对无关能力的损害与原始模型的 KL 散度相关，而后续写入无论早期事实如何存储都会导致干扰。广泛的数据可以创造可用的知识，而冻结的参考可以保持能力，但我们测试的所有干预措施，包括那些基于每次写入的准确局部测量的干预，都无法保持早期事实的可达性。当事实必须被构成或在后续写入中存活时，可靠的渠道是上下文而非权重。
cs.CL / 47 / 2607.11026
Dimensionality in Satisfaction Ratings
满意度评估中的维度
Abstract
We used a large language model (GPT-4.1) to annotate the text of about 9,000 support conversations at a global consumer-goods firm, decomposing customer-care satisfaction into component axes (overall, agent, outcome, product, and customer effort), and validated the LLM annotations against the satisfaction ratings customers gave themselves. Four of five axes track self-reported satisfaction closely (overall, agent, and outcome near an unadjusted 0.65; effort -0.54), while product satisfaction is weak against the available proxy. The unadjusted correlation also understates the alignment: the disagreements concentrate in a small, readable tail of divergent sessions rather than in general drift, and the overall correlation rises to 0.811 when only the severe divergences are excluded and to 0.914 when the full divergent tail is excluded. The axes are also highly collinear, and adding them to the overall score does not improve prediction of the customer's rating, the decomposition's value is not incremental prediction but attribution and coverage. And, with greater coverage the picture of the data changes. Read on every contact rather than the few that return a survey, satisfaction is markedly lower than the survey reports (a full-census 2.91 against the surveyed 3.62 on a five-point scale). The promise of decomposed satisfaction as a methodology is the ability to identify more nuanced drivers of customer experience in conversational data.
Chinese Translation
我们使用大型语言模型（GPT-4.1）对全球消费品公司的约9,000个支持对话文本进行了注释，将客户关怀满意度分解为多个组成轴（整体、代理、结果、产品和客户努力），并将LLM注释与客户自我报告的满意度评分进行了验证。五个轴中有四个与自我报告的满意度紧密相关（整体、代理和结果的未调整相关性接近0.65；努力的相关性为-0.54），而产品满意度与可用代理的相关性较弱。未调整的相关性也低估了对齐程度：分歧主要集中在少数可读的不同会话尾部，而不是普遍漂移，当仅排除严重分歧时，整体相关性上升至0.811，排除完整的不同尾部时上升至0.914。这些轴之间也高度共线，添加它们到整体评分并未改善客户评分的预测，分解的价值并不在于增量预测，而在于归因和覆盖。随着覆盖范围的增加，数据的全貌发生变化。根据每次接触的反馈，而非仅仅依赖于少数返回调查的反馈，满意度明显低于调查报告（全面普查为2.91，而调查结果为3.62，基于五分制）。分解满意度作为一种方法论的前景在于能够识别对话数据中客户体验的更细致驱动因素。
cs.CL / 48 / 2607.11053
Flout at Your Own Risk: LLMs Struggle with Pragmatic Cooperativity Under Epistemic Asymmetry
自负其责：大型语言模型在认知不对称下的合作语用能力挑战
Abstract
Fruitful collaborations rely on cooperative communications, including of contextual cues to incorporate into reasoning. The increasing use of LLMs in collaborative and agentic pipelines raises questions about the extent to which they exhibit these pragmatic capabilities, especially in scenarios where they may not have access to the same information as their collaborators. In this paper, we perform a novel investigation into the pragmatic reasoning capabilities of LLMs in a multi-party collaborative task under partial information conditions. We formalize a notion of collaborative epistemic asymmetry that explicitly connects objective task success to Grice's cooperative principle and empirically assess various LLMs' abilities to act cooperatively as both speakers and listeners, including both prompting and post-training strategies. Our results show that while LLMs exhibit certain pragmatic capabilities in collaborative settings, and these can be elicited through prompting and post-training, they still face challenges in pragmatic communication with incomplete information, and that certain failure modes do correlate with floutings of Grice's maxims that go unrecognized.
Chinese Translation
有效的合作依赖于合作性沟通，包括将上下文线索纳入推理中。大型语言模型（LLMs）在协作和自主流程中的日益使用引发了关于它们在多大程度上展现这些语用能力的问题，尤其是在它们可能无法获得与合作者相同信息的场景中。本文对LLMs在部分信息条件下的多方协作任务中的语用推理能力进行了新颖的研究。我们形式化了一个合作性认知不对称的概念，明确将客观任务成功与格赖斯的合作原则联系起来，并实证评估了各种LLMs作为说话者和听众的合作行为能力，包括提示和后训练策略。我们的结果表明，尽管LLMs在协作环境中展现出某些语用能力，并且这些能力可以通过提示和后训练来引发，但它们在面对不完整信息时仍面临语用沟通的挑战，并且某些失败模式确实与未被识别的格赖斯准则的违背相关。
cs.CL / 49 / 2607.11070
MJ: Multi-turn LLM Jailbreaking via Decomposed Credit Assignment
MJ：通过分解信用分配实现多轮大语言模型越狱
Abstract
Modern large language models (LLMs) operate in interactive multi-turn settings, making multi-turn jailbreaking a realistic threat model and an important setting for automated red teaming. A core challenge in learning multi-turn jailbreak attackers is credit assignment: different turns contribute differently to the final outcome, yet existing learning signals are often too coarse to identify their individual contributions. We propose decomposed credit GRPO (DC-GRPO), a unified turn-level credit assignment framework for Group Relative Policy Optimization in multi-turn jailbreak learning. DC-GRPO assigns a separate group-relative learning signal to each turn by combining immediate and future credit, avoiding the credit misassignment induced by broadcasting a single trajectory-level score across the dialogue. We instantiate this framework with static and dynamic weighting rules that differ in how the two credit sources are balanced while sharing the same turn-level structure. Across multiple victim LLMs and benchmarks, the dynamic- and static-weighted variants achieve average ASR5@3 scores of 98.26% and 97.88%, respectively, substantially outperforming the state-of-the-art methods, including SEMA (86.58%) and TROJail (86.23%). Their consistently strong performance indicates that the central empirical benefit comes from turn-level group-relative credit assignment rather than a particular weighting rule. Warning: This paper contains examples of harmful content.
Chinese Translation
现代大型语言模型（LLMs）在交互式多轮设置中运行，使得多轮越狱成为一种现实的威胁模型，并且是自动化红队测试的重要场景。学习多轮越狱攻击者的核心挑战在于信用分配：不同的轮次对最终结果的贡献不同，但现有的学习信号往往过于粗糙，无法识别它们各自的贡献。我们提出了分解信用的群体相对策略优化（DC-GRPO），这是一个统一的轮次级信用分配框架，用于多轮越狱学习。DC-GRPO通过结合即时信用和未来信用，为每个轮次分配一个单独的群体相对学习信号，避免了通过在对话中广播单一轨迹级评分而导致的信用错误分配。我们用静态和动态加权规则实例化该框架，这两种规则在平衡这两种信用来源的方式上有所不同，但共享相同的轮次级结构。在多个受害者LLM和基准测试中，动态加权和静态加权变体的平均ASR5@3得分分别为98.26%和97.88%，显著超越了包括SEMA（86.58%）和TROJail（86.23%）在内的最先进方法。它们持续强劲的表现表明，主要的经验收益来自于轮次级群体相对信用分配，而非特定的加权规则。警告：本文包含有害内容的示例。
cs.CL / 50 / 2607.11074
ResearchQA: Benchmarking Citation-Grounded Question-Answering on Scientific Papers
ResearchQA：科学论文中基于引用的问答基准测试
Abstract
Large language models are increasingly used to assist scientific reading, but existing evaluation methods often fail to detect whether answers are supported by verifiable citations. We introduce ResearchQA, a benchmark of 6,211 single-paper question-answer pairs from 494 open-access papers spanning eight domains and four question types: lookup, comprehension, multi-hop, and adversarial. ResearchQA is designed for citation-grounded evaluation: it permits multiple valid supporting passages for a claim and rewards grounded refusal when the source paper does not support an answer. We evaluate eight leading closed- and open-weight models in a citation-grounded chat-with-paper setting using a deterministic citation matcher and an LLM-based rubric evaluator. Citation-based metrics separate systems more clearly than LLM-evaluator scores: section coverage and citation accuracy vary substantially across models, while evaluator scores remain tightly compressed. We further find that open-weight models approach the best closed-model citation accuracy while achieving 3 to 6 times lower per-example latency. We release the benchmark, evaluation harness, and evaluator prompt.
Chinese Translation
大型语言模型越来越多地被用于辅助科学阅读，但现有的评估方法往往无法检测答案是否得到了可验证的引用支持。我们介绍了ResearchQA，这是一个包含6,211个单篇论文问答对的基准，来自494篇开放获取论文，涵盖八个领域和四种问题类型：查找、理解、多跳和对抗。ResearchQA旨在进行基于引用的评估：它允许一个主张有多个有效的支持段落，并在源论文不支持答案时奖励基于引用的拒绝。我们在一个基于引用的与论文聊天的环境中评估了八个领先的闭合和开放权重模型，使用确定性的引用匹配器和基于LLM的评分标准。基于引用的指标比LLM评估者的得分更清晰地区分系统：各模型的章节覆盖率和引用准确性差异显著，而评估者得分则保持紧密压缩。我们进一步发现，开放权重模型在引用准确性上接近最佳闭合模型，同时每个示例的延迟降低了3到6倍。我们发布了该基准、评估工具和评估者提示。
cs.CL / 51 / 2607.11127
Do LLMs Fabricate Legal Citations? A Bilingual Benchmark on Saudi Data Protection Law and the GDPR
大型语言模型是否伪造法律引用？关于沙特数据保护法和GDPR的双语基准测试
Abstract
Organizations and regulators increasingly consult large language models (LLMs) for regulatory-compliance questions, yet a wrong statutory citation can silently propagate into legal advice, compliance documentation, and policy decisions. We introduce a bilingual benchmark of 120 questions probing whether freely accessible LLMs fabricate article citations for two data-protection instruments: the EU General Data Protection Regulation (GDPR) and the Saudi Personal Data Protection Law (PDPL). The benchmark pairs direct citation retrieval questions with false premise verification probes and deliberately unanswerable "trap" questions -- including questions about a repealed article and about deadlines that exist only in implementing regulations, not in the law itself. Every question is posed in both Arabic and English, and all scoring is fully automatic against a manually verified gold reference. Evaluating three freely accessible models (Gemini 2.5 Flash, GPT-OSS-120B, Nemotron-3-Super-120B), we find a dramatic jurisdiction gap: near-ceiling citation accuracy on the GDPR (94-100% on direct retrieval) against majority fabrication on the Saudi PDPL (60-77%), invariant to query language; the highest fabrication rates (67%) arise from statute-vs-regulations confusion, and 91% of fabricated citations are asserted with confidence >= 0.8. Fabrication tracks the jurisdiction of the law, not the language of the query, and model confidence provides no protection -- indicating that verbatim-verification safeguards, rather than model self confidence, must gate any institutional reliance on LLMs for compliance screening.
Chinese Translation
组织和监管机构越来越多地咨询大型语言模型（LLMs）以解决合规性问题，但错误的法定引用可能会悄然传播到法律建议、合规文档和政策决策中。我们引入了一个双语基准，包含120个问题，探讨自由可访问的LLMs是否伪造了两个数据保护工具的条款引用：欧盟通用数据保护条例（GDPR）和沙特个人数据保护法（PDPL）。该基准将直接引用检索问题与虚假前提验证探测问题配对，并故意设置不可回答的“陷阱”问题——包括关于已废止条款的问题以及仅存在于实施法规中的截止日期，而不在法律本身中。每个问题均以阿拉伯语和英语提出，所有评分均完全自动化，并与手动验证的黄金参考进行对比。评估三种自由可访问的模型（Gemini 2.5 Flash、GPT-OSS-120B、Nemotron-3-Super-120B），我们发现了显著的司法管辖差距：在GDPR上的引用准确率接近上限（直接检索为94-100%），而沙特PDPL则大多数为伪造（60-77%），与查询语言无关；最高的伪造率（67%）源于法律与法规之间的混淆，91%的伪造引用的置信度均大于或等于0.8。伪造情况与法律的管辖权相关，而非查询的语言，模型置信度并未提供保护——这表明，逐字验证的保障措施，而非模型自信，必须限制任何机构对LLMs在合规筛查中的依赖。
cs.CL / 52 / 2607.11131
TIGER: Text-Conditioned Visual Gated Routing with Acceptance Alignment for Multimodal Speculative Decoding
TIGER：基于文本条件的视觉门控路由与接受对齐的多模态推测解码
Abstract
Speculative decoding accelerates autoregressive generation by letting a lightweight drafter propose multiple tokens that are verified by a larger target model. Although effective for text-only LLMs, speculative decoding yields limited gains in VLMs because drafters often diverge on vision-critical content, while existing multimodal acceleration methods do not directly address irrelevant visual evidence or optimize the verifier-accepted prefix length that governs speedup. We propose TIGER, a Text-conditioned vIsual GatEd Routing framework for multimodal speculative decoding. TIGER dynamically selects a sparse set of context-relevant visual tokens based on the drafter's current textual state, rather than expose the full visual token set or a fixed compressed interface. To better align training with inference-time efficiency, we optimize the drafter with acceptance-aligned group-based policy training using verifier-derived rewards based on accepted prefix length, built on top of distillation warm start with KL anchoring. This encourages the drafter not only to imitate the target model, but also to produce speculative continuations that survive verification for longer prefixes. Experiments show that TIGER yields consistent gains in accepted prefix length and speculative speedup under exact verifier-side speculative decoding, while achieving favorable quality-latency trade-offs with comparable downstream accuracy in visual-routing analyses.
Chinese Translation
推测解码通过让轻量级草拟者提出多个令牌并由更大的目标模型进行验证，从而加速自回归生成。尽管在仅文本的语言模型（LLMs）中效果显著，但在视觉语言模型（VLMs）中，推测解码的收益有限，因为草拟者往往在视觉关键内容上产生偏差，而现有的多模态加速方法并未直接解决无关视觉证据或优化决定速度的验证者接受前缀长度。我们提出了TIGER，一个用于多模态推测解码的基于文本条件的视觉门控路由框架。TIGER根据草拟者当前的文本状态动态选择一组稀疏的上下文相关视觉令牌，而不是暴露完整的视觉令牌集或固定的压缩接口。为了更好地将训练与推理时的效率对齐，我们通过基于接受前缀长度的验证者派生奖励，优化草拟者的接受对齐的基于组的策略训练，建立在带有KL锚定的蒸馏热启动之上。这不仅鼓励草拟者模仿目标模型，还促使其生成在更长前缀下能够通过验证的推测延续。实验表明，TIGER在精确的验证者侧推测解码下，接受前缀长度和推测加速方面均取得了一致的提升，同时在视觉路由分析中实现了有利的质量-延迟权衡，并保持了可比的下游准确性。
cs.CL / 53 / 2607.11163
Unified Gradient Projection: Language-Balanced Continual Learning for Multilingual Low-Resource ASR
统一梯度投影：面向多语言低资源自动语音识别的语言平衡持续学习
Abstract
Large-scale pretrained ASR models such as Whisper exhibit strong multilingual capabilities. However, fine-tuning on low-resource languages often causes catastrophic forgetting. Although continual learning mitigates this issue, existing methods struggle to regulate cross-task interference in multilingual settings, where dominant languages bias optimization. We propose Unified Gradient Projection (UGP), which constrains parameter updates using reference gradients from language-balanced replay in a unified projection space. By equalizing per-language contributions in the projection, UGP reduces dominant-language bias and improves cross-lingual stability. We further show that combining gradient-level projection with data-level replay yields complementary gains in stability and plasticity. Across diverse low-resource language groups and model scales, UGP enables effective adaptation while substantially mitigating forgetting. On Whisper-large-v3, it achieves near-zero average forgetting.
Chinese Translation
大规模预训练的自动语音识别（ASR）模型，如 Whisper，展现出强大的多语言能力。然而，在低资源语言上的微调往往会导致灾难性遗忘。尽管持续学习在一定程度上缓解了这一问题，但现有方法在多语言环境中难以调节跨任务干扰，主导语言会偏向优化。我们提出了统一梯度投影（Unified Gradient Projection, UGP），该方法通过在统一投影空间中使用语言平衡重放的参考梯度来约束参数更新。通过在投影中平衡每种语言的贡献，UGP 减少了主导语言的偏差，并提高了跨语言的稳定性。我们进一步表明，将梯度级投影与数据级重放相结合，可以在稳定性和适应性上获得互补的提升。在各种低资源语言组和模型规模中，UGP 实现了有效的适应，同时显著减轻了遗忘。在 Whisper-large-v3 上，它达到了近乎零的平均遗忘率。
cs.CL / 54 / 2607.11166
Query-Focused Event Summarization: A Dataset and Benchmark
基于查询的事件摘要：数据集与基准测试
Abstract
A thematic corpus is a collection of semantically coherent documents that collectively describe different aspects of a shared thematic event. Such a corpus typically contains hundreds or even thousands of documents. While users' interests in a thematic event often span multiple dimensions, Query-Focused Summarization (QFS) aims to generate summaries tailored to users' queries. However, existing QFS datasets lack event-oriented summarization, and most QFS methods struggle with large-scale corpora. To address these challenges, we propose the Query-Focused Event Summarization (QFES) task and construct the QFESum dataset, which contains 8 thematic events, 16,684 documents, and 104 queries. Furthermore, we introduce a two-stage QFES framework consisting of Query-Focused Retrieval with Adaptive Thresholding (RAT) and Query-Focused Summarization based on Hierarchical Clustering (SHC). Experimental results on QFESum show that RAT and SHC consistently outperform the baselines, demonstrating their effectiveness for QFES. The dataset and code are publicly available at https://github.com/sarcasm-hcy02/QFES-QFESum.
Chinese Translation
主题语料库是一个由语义上连贯的文档组成的集合，这些文档共同描述了共享主题事件的不同方面。这样的语料库通常包含数百甚至数千个文档。尽管用户对主题事件的兴趣往往跨越多个维度，但基于查询的摘要（Query-Focused Summarization, QFS）旨在生成针对用户查询量身定制的摘要。然而，现有的QFS数据集缺乏以事件为导向的摘要，大多数QFS方法在处理大规模语料库时表现不佳。为了解决这些挑战，我们提出了基于查询的事件摘要（Query-Focused Event Summarization, QFES）任务，并构建了QFESum数据集，该数据集包含8个主题事件、16,684个文档和104个查询。此外，我们引入了一个由查询聚焦检索与自适应阈值（Query-Focused Retrieval with Adaptive Thresholding, RAT）和基于层次聚类的查询聚焦摘要（Query-Focused Summarization based on Hierarchical Clustering, SHC）组成的两阶段QFES框架。在QFESum上的实验结果表明，RAT和SHC始终优于基线，证明了它们在QFES中的有效性。数据集和代码可在https://github.com/sarcasm-hcy02/QFES-QFESum上公开获取。
cs.CL / 55 / 2607.11183
Amplitude-Only FFN Intervention for Tool-Structured LLM Inference Method: Gated Evaluation Protocol, and Cross-Model Empirical Results
仅幅度的前馈网络干预用于工具结构化大语言模型推理方法：门控评估协议及跨模型实证结果
Abstract
Large language models increasingly operate as tool-using agents, where small format, argument, or function-call errors can invalidate otherwise plausible responses. We study inference-time feed-forward network (FFN) intervention for improving structured outputs without retraining model weights. Our project began with Orthogonal Residual Projection (ORP), a direction-changing repair attempt that revealed sensitive SwiGLU FFN intervention sites but often caused more harm than fixes. We therefore propose Amplitude Gating (AG), a non-destructive alternative that preserves pretrained FFN weight directions and modulates only activation magnitudes during generation. We define a fine-grained intervention system spanning P1/P2/P3 and branch-specific P1s/P2a/P2b sites, and introduce an evaluation protocol that separates combination-oracle headroom from fixed configurations and learned gates, enforces sample-level accounting, and uses task-aware metrics for binary and partial-credit datasets. Across Qwen3.5-9B, Qwen3-8B, and Qwen2.5-7B, AG is weakly positive in aggregate but strongest on tool-structured tasks. On Qwen3.5-9B, a category-level learned gate improves tool/structured/agentic performance from 38.66% to 42.92% (+4.27 percentage points), with Hermes function-call tasks reaching about +7.6 points. On Qwen3-8B, Hermes JSON mode improves by +11.36 points. Qwen2.5-7B retains oracle headroom but current learned gates fail to capture it, showing that deployment requires model- and category-specific routing. Comparisons of entropy AG with Newton-Schulz-windowed AG show that neither family is uniformly dominant. These results identify tool-structured inference as the most credible first target for safe FFN-level inference optimization, while prospective online validation and broader cross-model evaluation remain necessary.
Chinese Translation
大型语言模型越来越多地作为工具使用代理进行操作，其中小格式、参数或函数调用错误可能会使原本合理的响应失效。我们研究了推理时前馈网络（FFN）干预，以在不重新训练模型权重的情况下改善结构化输出。我们的项目始于正交残差投影（Orthogonal Residual Projection, ORP），这是一种改变方向的修复尝试，揭示了敏感的SwiGLU FFN干预点，但往往造成的损害大于修复。因此，我们提出了幅度门控（Amplitude Gating, AG），这是一种非破坏性的替代方案，保留预训练FFN权重方向，并在生成过程中仅调节激活幅度。我们定义了一个细粒度的干预系统，涵盖P1/P2/P3及特定分支的P1s/P2a/P2b点，并引入了一种评估协议，该协议将组合-神谕的余地与固定配置和学习门控分开，强制进行样本级核算，并使用任务感知指标用于二元和部分信用数据集。在Qwen3.5-9B、Qwen3-8B和Qwen2.5-7B上，AG在总体上表现出微弱的积极性，但在工具结构化任务上表现最强。在Qwen3.5-9B上，类别级学习门控将工具/结构化/代理性能从38.66%提高到42.92%（+4.27个百分点），而Hermes函数调用任务的提升约为+7.6点。在Qwen3-8B上，Hermes JSON模式提高了+11.36点。Qwen2.5-7B保留了神谕余地，但当前学习的门控未能捕捉到这一点，显示出部署需要模型和类别特定的路由。熵AG与牛顿-舒尔茨窗口AG的比较表明，两者均不具备统一的优势。这些结果确定了工具结构化推理作为安全FFN级推理优化的最可信的首个目标，而前景在线验证和更广泛的跨模型评估仍然是必要的。
cs.CL / 56 / 2607.11207
ProgramTab: Boosting Table Reasoning of LLMs via Programmatic Paradigm
ProgramTab：通过程序化范式提升大型语言模型的表格推理能力
Abstract
Table-based reasoning with large language models (LLMs), which requires reasoning based on natural language questions and structured tabular data, has gained widespread attention. However, a series of issues still constrain the application of this task. The previous approaches suffered from significant performance degradation when faced with large tables due to the difficulty of long text modeling and the limitation of input length for LLMs. The text-to-SQL approach is used to efficiently extract key information from tables and generate smaller sub-tables. However, tabular data, especially web tables, often lack the necessary structure and consistency, making them unsuitable for performing mathematical logic operations using SQL queries. We propose the ProgramTab framework, which guides LLMs employing in-context learning to perform tabular data preprocessing with Python code, as well as the momentous contents extraction with row and column extraction and SQL generation. The experiment results on table reasoning datasets demonstrate that the ProgramTab framework effectively deals with table-based reasoning tasks and outperforms all LLM-based baselines.
Chinese Translation
基于表格的推理与大型语言模型（LLMs）相结合，要求基于自然语言问题和结构化表格数据进行推理，已引起广泛关注。然而，一系列问题仍然限制了该任务的应用。以往的方法在面对大型表格时表现出显著的性能下降，这主要是由于长文本建模的困难和LLMs输入长度的限制。文本到SQL（text-to-SQL）的方法被用来有效提取表格中的关键信息并生成较小的子表。然而，表格数据，尤其是网页表格，往往缺乏必要的结构和一致性，使其不适合使用SQL查询进行数学逻辑操作。我们提出了ProgramTab框架，该框架指导LLMs利用上下文学习，通过Python代码进行表格数据预处理，以及进行行列提取和SQL生成的重要内容提取。在表格推理数据集上的实验结果表明，ProgramTab框架有效处理基于表格的推理任务，并超越了所有基于LLM的基线方法。
cs.CL / 57 / 2607.11213
When the Target Domain Changes: AI-Mediated Construct Drift in High-Stakes English Language AssessmenW
当目标领域发生变化时：高风险英语语言评估中的人工智能介导构念漂移
Abstract
High-stakes English proficiency tests treat standardized, unaided performance as evidence for score interpretations about academic English proficiency. This interpretation remains meaningful, but as target language use domains increasingly involve generative AI, the extrapolation from unaided test performance to academic communicative readiness becomes less self-evident. This conceptual validity argument reframes AI as a score-interpretation problem in high-stakes language testing, not only an operational issue of scoring, feedback, security, or misconduct. Synthesizing current literature in three uneven layers, the paper shows that most work treats AI as assessment infrastructure, while far less theorizes its implications for construct validity and extrapolation warrants. It defines AI-mediated construct drift as the misalignment that arises when communicative abilities required in the target domain change through AI mediation while test constructs remain anchored to an unaided-performance model. It proposes bounded AI mediation as a validity-oriented design principle: a standardized condition in which all test takers access the same institutionally controlled AI assistant, with predefined assistance boundaries, logged interactions, and tasks that distinguish comprehension support from answer generation. The paper argues that score interpretations should be narrowed and supplemented when used to support claims about AI-mediated academic communication.
Chinese Translation
高风险英语水平测试将标准化、无辅助的表现视为关于学术英语水平的评分解释的证据。这一解释依然具有意义，但随着目标语言使用领域越来越多地涉及生成性人工智能，从无辅助测试表现推断学术交流准备的过程变得不再显而易见。本文的概念有效性论证将人工智能重新框定为高风险语言测试中的评分解释问题，而不仅仅是评分、反馈、安全或不当行为的操作性问题。通过对当前文献的三层不均匀整合，本文表明大多数研究将人工智能视为评估基础设施，而对其对构念有效性和推断依据的影响理论探讨则相对较少。本文将人工智能介导构念漂移定义为当目标领域所需的交流能力通过人工智能介导而发生变化时，与无辅助表现模型锚定的测试构念之间的失调。它提出了有限的人工智能介导作为一种以有效性为导向的设计原则：一种标准化条件，在该条件下，所有考生都可以访问相同的机构控制的人工智能助手，并设定预定义的辅助边界、记录的交互和区分理解支持与答案生成的任务。本文认为，在支持关于人工智能介导的学术交流的主张时，评分解释应当被缩小并补充。
cs.CL / 58 / 2607.11215
Q-BridgeNet: A Quantization Network for Cross-Lingual Sign Language Translation
Q-BridgeNet：一种用于跨语言手语翻译的量化网络
Abstract
Most sign language translation (SLT) methods focus on isolated native sign-spoken pairs (e.g., American Sign Language - English). Extending language-specific SLT models to multilingual translation would improve accessibility by enabling communication across diverse sign and spoken language communities. However, existing multilingual SLT approaches still struggle to learn a unified model that minimizes cross-lingual conflicts while capturing shared cross-lingual semantics and preserving language-specific variations across different sign languages. Therefore, we propose Q-BridgeNet, a unified framework for multilingual SLT that jointly mitigates cross-lingual conflicts across both the sign language and spoken language sides. On the sign language side, Q-BridgeNet learns discrete Q-units via adaptive segmentation and residual vector quantization: a shared base codebook provides language-agnostic semantic primitives, while language-specific residual codebooks refine heterogeneous signing semantics. On the spoken language side, a multilingual LLM is fine-tuned to operate in the Q-unit space, leveraging cross-lingual priors to enable a unified SLT model. Experiments on PHOENIX14T, How2Sign, and CSL-Daily show that Q-BridgeNet effectively mitigates cross-lingual conflicts, achieving state-of-the-art performance on native sign-spoken pairs while also demonstrating strong generalization to non-native pairs. Our source code is publicly available at: https://github.com/FengLiQ/Q-BridgeNet
Chinese Translation
大多数手语翻译（SLT）方法专注于孤立的本土手语-口语对（例如，美国手语 - 英语）。将特定语言的SLT模型扩展到多语言翻译将通过促进不同手语和口语社区之间的交流来提高可及性。然而，现有的多语言SLT方法仍然难以学习一个统一的模型，以最小化跨语言冲突，同时捕捉共享的跨语言语义并保留不同手语之间的语言特定变异。因此，我们提出了Q-BridgeNet，一个用于多语言SLT的统一框架，旨在共同减轻手语和口语两侧的跨语言冲突。在手语方面，Q-BridgeNet通过自适应分段和残差向量量化学习离散的Q单元：共享的基础代码本提供语言无关的语义原语，而特定语言的残差代码本则细化异质的手语语义。在口语方面，一个多语言LLM被微调以在Q单元空间中操作，利用跨语言先验来实现统一的SLT模型。在PHOENIX14T、How2Sign和CSL-Daily上的实验表明，Q-BridgeNet有效地减轻了跨语言冲突，在本土手语-口语对上实现了最先进的性能，同时在非本土对上也表现出强大的泛化能力。我们的源代码已公开，地址为：https://github.com/FengLiQ/Q-BridgeNet
cs.CL / 59 / 2607.11258
TreeThink: A Modular Tree Search Library for Mathematical Reasoning with LLMs
TreeThink：用于数学推理的模块化树搜索库
Abstract
Tree search algorithms enable systematic exploration of the proof space in neural theorem proving. Existing LLM tree search libraries primarily target natural language reasoning and do not provide native integration with formal verifiers, while theorem proving systems often rely on task-specific search implementations. We introduce TreeThink, an open-source Python library for modular, fully asynchronous tree search in neural theorem proving. It integrates established tree search methods with vLLM-based inference pipelines and diverse node evaluation techniques, ranging from lightweight heuristics to neural evaluators. We support Lean~4, Rocq, and Isabelle/HOL alongside natural language. It connects directly to each language's Read-Eval-Print Loop (REPL) server for real-time verification and proof state extraction. We evaluate TreeThink on miniF2F and MATH500, demonstrating cross-language formal proof search, natural language reasoning support, and up to 6.3$\times$ wall-clock speedup from asynchronous execution. Source code is released under the MIT license at https://github.com/GGLAB-KU/treethink , and the library is accessible as a downloadable package at https://pypi.org/project/treethink/ .
Chinese Translation
树搜索算法使得在神经定理证明中对证明空间进行系统探索成为可能。现有的LLM树搜索库主要针对自然语言推理，并未提供与形式验证器的原生集成，而定理证明系统通常依赖于特定任务的搜索实现。我们介绍了TreeThink，这是一个开源的Python库，用于在神经定理证明中进行模块化、完全异步的树搜索。它将成熟的树搜索方法与基于vLLM的推理管道和多样的节点评估技术相结合，从轻量级启发式方法到神经评估器。我们支持Lean~4、Rocq和Isabelle/HOL以及自然语言。它直接连接到每种语言的读-评-打印循环（REPL）服务器，以实现实时验证和证明状态提取。我们在miniF2F和MATH500上评估了TreeThink，展示了跨语言的形式证明搜索、自然语言推理支持，以及通过异步执行实现的最高6.3倍的墙钟加速。源代码已在MIT许可证下发布，地址为https://github.com/GGLAB-KU/treethink，库可作为可下载包访问，地址为https://pypi.org/project/treethink/。
cs.CL / 60 / 2607.11276
Automated Textbook Auditing with Multi-Agent LLM Systems
基于多智能体大语言模型系统的自动化教科书审计
Abstract
Ensuring the quality of educational materials requires more than standard proofreading: textbooks must be audited for factual accuracy, domain-specific technical correctness, and linguistic quality simultaneously -- a task that general-purpose grammar checkers cannot address. We present \textbf{AI Textbook Auditor}, a modular multi-agent pipeline for automated quality assurance of educational materials across subject domains. The system accepts a textbook PDF and produces a structured, human-reviewable report via two analysis tracks: a \textbf{Factual and Technical Track} in which an ensemble of specialized LLM agents detects factual inaccuracies, code errors, incorrect definitions, and conceptual inconsistencies, augmented with web search for humanities domains; and a \textbf{Grammar Track} operating PDF-natively to preserve diacritical encoding. A \textbf{Judge Agent} filters false positives using domain-specific rules before presenting findings to a human reviewer. The pipeline supports two ingestion modes -- vision-native page rendering and PyMuPDF text extraction -- and is domain-adaptable via custom prompts encoding subject-specific error taxonomies. We demonstrate the system on two Romanian upper-secondary textbooks: a CS textbook (56 technical findings across seven categories, with an expert-validated precision of 62.5\%) and a history and social sciences textbook (72 findings spanning factual errors, ideological bias, and grammar). The system is designed as a triage tool that reduces the manual effort of locating candidate issues, with human expert validation required before any editorial action.
Chinese Translation
确保教育材料的质量不仅仅依赖于标准的校对：教科书必须同时审计事实准确性、领域特定的技术正确性和语言质量——这是通用语法检查工具无法解决的任务。我们提出了 extbf{AI教科书审计员}，这是一个模块化的多智能体管道，用于跨学科领域的教育材料的自动化质量保证。该系统接受教科书PDF文件，并通过两个分析轨道生成结构化的、可供人工审阅的报告： extbf{事实与技术轨道}，在该轨道中，一组专门的LLM代理检测事实不准确、代码错误、不正确的定义和概念不一致，同时结合人文学科的网络搜索；以及 extbf{语法轨道}，该轨道原生支持PDF，以保持变音符号编码。在向人类审阅者呈现发现之前， extbf{评审代理}使用领域特定规则过滤假阳性。该管道支持两种输入模式——视觉原生页面渲染和PyMuPDF文本提取——并通过自定义提示编码领域特定的错误分类法进行领域适应。我们在两本罗马尼亚高中教科书上演示了该系统：一本计算机科学教科书（在七个类别中发现56个技术问题，专家验证的准确率为62.5\%）和一本历史与社会科学教科书（发现72个问题，包括事实错误、意识形态偏见和语法问题）。该系统被设计为一种分流工具，减少了定位候选问题的人工工作量，但在任何编辑操作之前需要人类专家的验证。
cs.CL / 61 / 2607.11279
FAD-SA-GRU: Enhancing Hate Speech Detection in Algerian Dialect Through Feature-Augmented Self-Attention GRU Networks
FAD-SA-GRU：通过特征增强自注意力GRU网络提升阿尔及利亚方言中的仇恨言论检测
Abstract
The widespread adoption of social media platforms has transformed online communication by enabling users to exchange information and opinions instantly. However, these platforms have also facilitated the dissemination of abusive and hateful content, posing major social, psychological, and ethical challenges. Hate speech can incite discrimination, harassment, and violence against individuals or communities based on attributes such as ethnicity, religion, gender, nationality, or political affiliation. Consequently, automatic hate speech detection has become a major research topic in natural language processing (NLP) and an essential component of content moderation systems. This paper investigates automatic hate speech detection in the Algerian Arabic dialect (Darija) on social media. This task remains challenging because of the dialect's linguistic diversity, characterized by the coexistence of Arabic, French, and Arabizi (Arabic written using the Latin alphabet). We compare four categories of text classification approaches: (1) traditional machine learning models using TF-IDF features, (2) deep learning models based on recurrent neural networks, (3) Transformer-based language models, including DziriBERT and multilingual BERT, and (4) a novel hybrid architecture, FAD-SA-GRU, which combines semantic representations from DZ FastText, DZ AraVec, and DziriBERT through multi-embedding fusion, followed by a self-attention-enhanced GRU encoder. Experiments on an annotated dataset of Algerian Darija social media comments for binary hate speech classification show that FAD-SA-GRU outperforms all baselines, achieving 93.2% accuracy, 93.4% precision, 91.0% recall, 92.1% F1-score, and 97.0% ROC-AUC. Results demonstrate the effectiveness of combining complementary embedding representations with attention-based sequence modeling for robust hate speech detection in low-resource dialectal Arabic.
Chinese Translation
社交媒体平台的广泛采用改变了在线交流，使用户能够即时交换信息和观点。然而，这些平台也促进了辱骂和仇恨内容的传播，带来了重大的社会、心理和伦理挑战。仇恨言论可能会煽动基于种族、宗教、性别、国籍或政治立场等属性的歧视、骚扰和暴力。因此，自动仇恨言论检测已成为自然语言处理（NLP）领域的一个重要研究课题，并且是内容审核系统的一个关键组成部分。本文研究了社交媒体上阿尔及利亚阿拉伯方言（Darija）的自动仇恨言论检测。由于该方言的语言多样性，表现为阿拉伯语、法语和阿拉伯拉丁字母（Arabizi）的共存，这一任务仍然具有挑战性。我们比较了四类文本分类方法：（1）使用TF-IDF特征的传统机器学习模型，（2）基于递归神经网络的深度学习模型，（3）基于Transformer的语言模型，包括DziriBERT和多语言BERT，以及（4）一种新颖的混合架构FAD-SA-GRU，该架构通过多嵌入融合结合DZ FastText、DZ AraVec和DziriBERT的语义表示，随后使用增强自注意力的GRU编码器进行处理。在针对阿尔及利亚Darija社交媒体评论的标注数据集进行的二元仇恨言论分类实验中，FAD-SA-GRU的表现优于所有基线，达到了93.2%的准确率、93.4%的精确率、91.0%的召回率、92.1%的F1分数和97.0%的ROC-AUC。结果表明，结合互补的嵌入表示与基于注意力的序列建模对于在低资源方言阿拉伯语中实现稳健的仇恨言论检测是有效的。
cs.CL / 62 / 2607.11341
The In-Car Sign Language Corpus (ICSL): A Multi-Modal Resource for Constrained-Space Sign Language Recognition
车内手语语料库（ICSL）：一种用于受限空间手语识别的多模态资源
Abstract
This paper addresses the challenges of using sign language within shared mobility services, such as taxis, carpools, or ride-sharing platforms. The use of sign language recognition (SLR) in real-world, confined environments, specifically vehicle interiors remains largely unexplored. To motivate research in this area, we present the In-Car Sign Language (ICSL) dataset for Brazilian Sign Language (Libras), with the long-term goal of improving public transport accessibility for the Deaf and Hard-of-Hearing community. The dataset consists of: (1) high-precision laboratory motion capture (MoCap) data to establish an idealized linguistic baseline and (2) real-world multi-modal in-car recordings captured using a 2D camera and 3D Time-of-Flight sensors. The dataset provides a basis for comparative analyses between synthesized signing avatar animations and recorded real signing interpreter videos, which enable future research into robust "in-the-wild" SLR models and domain adaptation. We describe in detail the use cases, the setup, the data collection protocol, and the metadata structure of the corpus. In total, we recorded a multimodal dataset exceeding 1.5 million frames, comprising the synchronized multimodal streams described above featuring Libras users across various in-car scenarios. The corpus is provided with gloss annotation of lexical signs and non-lexical sign language elements specially designed to support the training and evaluation of deep neural networks for constrained space recognition. In-vehicle signing offers a technically significant example of a constrained, occluded, and non-frontal environment. While recognizing the diverse communication strategies already employed by the Deaf community, identifying automotive-specific limitations provides a useful stepping stone for research into enhancing in-car accessibility and passenger quality of life.
Chinese Translation
本文探讨了在共享出行服务（如出租车、拼车或共享出行平台）中使用手语所面临的挑战。在现实世界的受限环境中，特别是车辆内部，手语识别（SLR）的应用仍然未得到充分探索。为了激励这一领域的研究，我们提出了巴西手语（Libras）的车内手语（ICSL）数据集，旨在长期改善聋人和听力障碍者的公共交通可达性。该数据集包括：（1）高精度实验室运动捕捉（MoCap）数据，以建立理想化的语言基线；（2）使用2D相机和3D飞行时间传感器捕获的现实世界多模态车内录音。该数据集为合成手语虚拟形象动画与录制的真实手语翻译视频之间的比较分析提供了基础，促进未来对强健的“野外”SLR模型和领域适应的研究。我们详细描述了用例、设置、数据收集协议和语料库的元数据结构。总的来说，我们录制了超过150万帧的多模态数据集，其中包含上述同步多模态流，涵盖了不同车内场景下的Libras用户。该语料库提供了词汇符号和非词汇手语元素的注释，特别设计用于支持深度神经网络在受限空间识别中的训练和评估。车内手语提供了一个技术上重要的受限、遮挡和非正面环境的示例。尽管认识到聋人社区已经采用的多样化沟通策略，识别汽车特定的限制为增强车内可达性和乘客生活质量的研究提供了有益的起点。
cs.CL / 63 / 2607.11353
Characterising AI Models for Cataloguing
对人工智能模型进行特征化以进行目录编制
Abstract
The creation of digital collections involves not only the digitisation of content, but also the creation of catalogue records for it. This often-overlooked task requires slow and costly expert manual work. In this project, we have evaluated the application of AI models to this task, comparing different implementations and models. This work includes a qualitative and quantitative evaluation of the experiments carried out, as well as recommendations on the use of AI models that go beyond the specific use case.
Chinese Translation
数字藏品的创建不仅涉及内容的数字化，还包括为其创建目录记录。这一常被忽视的任务需要耗时且成本高昂的专家手工工作。在本项目中，我们评估了人工智能模型在这一任务中的应用，比较了不同的实现和模型。该工作包括对所进行实验的定性和定量评估，以及对超越特定用例的人工智能模型使用的建议。
cs.CL / 64 / 2607.11358
RefineEvo: Planning-Guided Heuristic Evolution with Bidirectional Experience
RefineEvo：基于规划引导的双向经验启发式演化
Abstract
Automatic Heuristic Design (AHD) has emerged as a transformative approach for solving combinatorial optimization problems. While recent Large Language Model (LLM)-based methods have shown promise, they predominantly rely on fixed evolutionary operators and struggle to effectively accumulate and reuse historical search experience. This paper proposes RefineEvo, a novel evolutionary framework that transforms AHD from a static trial-and-error process into a planning-guided, experience-driven system. RefineEvo introduces a Planner to dynamically schedule evolutionary operators and trigger refinement based on the current search state, and a Reflector to distill valuable lessons into a Bidirectional Experience Pool containing both positive insights and negative pitfalls. This synergistic framework enables the system to adapt its search tools to the evolving complexity of the problem and leverage trajectory-aware, situation-conditioned insights to guide generation. Experiments on several classic combinatorial optimization benchmarks demonstrate that RefineEvo consistently outperforms strong baselines. In particular, RefineEvo delivers superior solution quality while improving token efficiency, enabling more efficient and autonomous heuristic design.
Chinese Translation
自动启发式设计（AHD）已成为解决组合优化问题的一种变革性方法。尽管最近基于大型语言模型（LLM）的方法显示出良好的前景，但它们主要依赖于固定的演化操作符，并且在有效积累和重用历史搜索经验方面存在困难。本文提出了RefineEvo，一种新颖的演化框架，将AHD从静态的试错过程转变为一个以规划为指导、以经验为驱动的系统。RefineEvo引入了一个规划器（Planner），用于动态调度演化操作符并根据当前搜索状态触发细化，同时引入了一个反思器（Reflector），将有价值的经验提炼到一个双向经验池（Bidirectional Experience Pool）中，包含积极的见解和消极的陷阱。这个协同框架使系统能够根据问题不断演变的复杂性调整其搜索工具，并利用轨迹感知和情境条件的见解来指导生成。在多个经典组合优化基准上的实验表明，RefineEvo始终优于强基线。特别是，RefineEvo在提高令牌效率的同时提供了更优的解决方案质量，从而实现更高效和自主的启发式设计。
cs.CL / 65 / 2607.11363
Beyond Sally-Anne: Evaluating Theory of Mind in LLMs using Epistemic Schelling Points
超越莎莉-安：使用认知舍林点评估大型语言模型的心智理论
Abstract
Text-based evaluations of Theory of Mind (ToM) in Large Language Models (LLMs) often involve cognitive tests akin to the Sally-Anne task that can be gamed due to exposure to relevantly similar tasks in pre-training and do not obviously test models' functional ToM abilities in ways that generalize to naturalistic settings. To address these issues, we introduce the Epistemic Asymmetry Schelling Task (EAST), a two-player dialogue game designed to benchmark robust and generalizable ToM abilities. By requiring LLM-LLM dyads to independently converge on semantic Schelling points under varying states of epistemic transparency, we evaluate whether models can robustly apply ToM to achieve coordination. Our results reveal a significant capability gap in functional social reasoning, with only frontier models successfully navigating the varying epistemic demands of the tasks. Analysis of reasoning traces shows that coordination failures are primarily driven by epistemic tracking errors, such as conflating private knowledge with mutual knowledge. Despite high performance on traditional static benchmarks, our study shows that robust social reasoning and epistemic tracking remain a critical bottleneck, providing concrete targets for future LLM evaluation and development.
Chinese Translation
基于文本的对大型语言模型（LLMs）心智理论（ToM）的评估通常涉及类似莎莉-安任务的认知测试，这些测试由于在预训练中接触到相关相似任务而可能被操控，并且并未明显测试模型在自然环境中普遍适用的功能性ToM能力。为了解决这些问题，我们引入了认知不对称舍林任务（Epistemic Asymmetry Schelling Task，EAST），这是一种旨在基准测试稳健且可推广的ToM能力的双人对话游戏。通过要求LLM-LLM对在不同的认知透明状态下独立收敛于语义舍林点，我们评估模型是否能够稳健地应用ToM以实现协调。我们的结果揭示了功能性社会推理中的显著能力差距，只有前沿模型成功应对任务的不同认知需求。推理轨迹的分析表明，协调失败主要是由认知追踪错误驱动的，例如将私人知识与共同知识混淆。尽管在传统静态基准测试中表现良好，我们的研究表明，稳健的社会推理和认知追踪仍然是一个关键瓶颈，为未来LLM的评估和开发提供了具体目标。
cs.CL / 66 / 2607.11399
Agentic Routing: The Harness-Native Data Flywheel
自主路由：Harness-Native 数据飞轮
Abstract
Large language model agents are increasingly executed not by a single model call, but by an execution harness that manages observation, context, control, action, state, and verification. At the same time, frontier and open models are becoming structurally specialized: a model that is strong at code editing, long-context recovery, tool use, mathematical reasoning, or low-latency response may not dominate on the other axes. This makes model selection inside an agent a core systems problem rather than a per-query serving trick. Existing routing methods mostly optimize single-turn cost-quality trade-offs and therefore miss the execution state, intermediate failures, and feedback loops that make agents different from chat completion. We propose Harness-Native agentic routing, a step-level routing paradigm that selects either a single best-fit model for cost-effective execution or multiple complementary models for ensemble-style accuracy improvement, conditioned on the full harness state. The key insight is that every routing decision naturally produces a structured data record -- consisting of the query, harness state, model choice or model set, execution trace, outcome, and cost -- whose labels are supplied by the environment rather than by the router itself. These records form a harness-native data flywheel: execution traces train better routers and harness-native models, which improve cost-quality trade-offs and generate more traces under the same budget. We instantiate this idea in OpenSquilla with a four-layer routing stack, an open LightGBM cold-start ranker, and a staged router-model path that turns logged arena records into progressively stronger routing policies. The report studies singleton and multi-model routing on agentic benchmarks including DRACO and PinchBench, and argues that agentic routing is not merely cost control, but a data engine for agent-native training.
Chinese Translation
大型语言模型代理的执行越来越不是通过单一模型调用，而是通过一个执行工具来管理观察、上下文、控制、行动、状态和验证。同时，前沿和开放模型在结构上变得越来越专业化：一个在代码编辑、长上下文恢复、工具使用、数学推理或低延迟响应方面表现出色的模型，可能在其他维度上并不占优势。这使得在代理内部进行模型选择成为一个核心系统问题，而不仅仅是一个每次查询的服务技巧。现有的路由方法大多优化单轮成本-质量权衡，因此忽略了执行状态、中间失败和反馈循环，这些因素使得代理与聊天完成有所不同。我们提出了Harness-Native自主路由，这是一种逐步路由范式，选择单个最佳模型以实现成本效益的执行，或选择多个互补模型以提高集成式准确性，条件是基于完整的执行工具状态。关键的见解是，每个路由决策自然产生一个结构化的数据记录——包括查询、执行工具状态、模型选择或模型集、执行轨迹、结果和成本——这些标签由环境提供，而不是由路由器本身提供。这些记录形成了一个Harness-Native数据飞轮：执行轨迹训练更好的路由器和Harness-Native模型，从而改善成本-质量权衡，并在相同预算下生成更多轨迹。我们在OpenSquilla中实例化了这一理念，构建了一个四层路由堆栈，一个开放的LightGBM冷启动排序器，以及一个分阶段的路由-模型路径，将记录的竞技场记录转化为逐步增强的路由策略。该报告研究了在自主基准（包括DRACO和PinchBench）上的单模型和多模型路由，并认为自主路由不仅仅是成本控制，而是一个用于代理原生训练的数据引擎。
cs.CL / 67 / 2607.11400
Cross-Architecture LLM Ensembles, Feature-Based Reranking and Retrieval-Augmented Prompting for Legal Information Processing
跨架构大语言模型集成、基于特征的重排序和检索增强提示在法律信息处理中的应用
Abstract
Legal information processing spans retrieval, entailment and judgment prediction problems, requiring text matching, reasoning and robust generalisation with limited supervision. We report Team DU's participation in all five tasks of COLIEE 2026, using open-weight systems for legal case retrieval, case entailment, statute retrieval and entailment, and legal judgment prediction. For Tasks 3 and 4, all models predate the 15 July 2025 cutoff required by the rules. For Task 4 (statute entailment), a cross-architecture ensemble of nine models from three families achieves 96.3% accuracy, placing first among 33 submissions from 11 teams. For the Pilot Task (tort prediction and rationale extraction), a multi-view system combining five claim-level models and refining the verdict using features derived from the claim predictions achieves 73.1% TP accuracy and 68.2% RE F1 as an unofficial submission, scoring above all official entries on TP and matching the highest on RE. For Task 2 (legal case entailment), changing only the prompt from single- to multi-selection raises F1 from 0.343 to 0.555 in post-competition evaluation on released gold labels, exceeding the best official submission (F1 = 0.490). For Task 3 (statute retrieval and entailment), replacing the entailment model with Qwen3-235B and a structured legal reasoning prompt raises accuracy from 79.3% to 91.5% in post-competition analysis. For Task 1 (legal case retrieval), a learning-to-rank system combining lexical and semantic retrieval with structural, citation authority, and temporal features (34 in total) achieves F1 = 0.314 (rank 11 of 54 submissions from 22 teams). Overall, legal information processing benefits from different inductive biases across tasks, with cross-architecture ensembling, feature-based reranking and retrieval-augmented prompting each proving most effective in different settings.
Chinese Translation
法律信息处理涵盖检索、蕴含和判决预测问题，要求在有限监督下进行文本匹配、推理和稳健的泛化。我们报告了DU团队在COLIEE 2026的五个任务中的参与，使用开放权重系统进行法律案件检索、案件蕴含、法规检索与蕴含，以及法律判决预测。在任务3和任务4中，所有模型均在规则要求的2025年7月15日截止日期之前。对于任务4（法规蕴含），来自三个模型家族的九个模型的跨架构集成达到了96.3%的准确率，在11个团队的33个提交中名列第一。对于试点任务（侵权预测和理由提取），一个结合了五个索赔级模型的多视角系统，通过使用从索赔预测中提取的特征来细化判决，达到了73.1%的真正例准确率和68.2%的召回率F1，作为非正式提交，得分超过所有官方条目在真正例上，并在召回率上与最高分持平。对于任务2（法律案件蕴含），仅通过将提示从单选改为多选，在发布的金标准标签的后竞争评估中，F1从0.343提高到0.555，超过了最佳官方提交（F1 = 0.490）。对于任务3（法规检索与蕴含），用Qwen3-235B替换蕴含模型，并使用结构化法律推理提示，使准确率在后竞争分析中从79.3%提高到91.5%。对于任务1（法律案件检索），一个结合了词汇和语义检索与结构、引用权威和时间特征（共34个特征）的学习排序系统达到了F1 = 0.314（在22个团队的54个提交中排名第11）。总体而言，法律信息处理在不同任务中受益于不同的归纳偏差，跨架构集成、基于特征的重排序和检索增强提示在不同环境中各自证明了其有效性。
cs.CL / 68 / 2607.11414
Confidently Wrong: Detecting Hallucinations in Financial Question Answering from LLM Internal States
自信错误：从大型语言模型内部状态检测金融问答中的幻觉
Abstract
Large language models (LLMs) in financial applications fail most consequentially when they are confidently wrong. Hedged, uncertain answers invite scrutiny, whereas confident errors silently degrade downstream decisions without warning. We ask how reliably such confidently wrong answers, or confident hallucinations, can be detected from a model's internal activations, and whether those activations carry information beyond its observable outputs. We train linear probes on the residual stream and evaluate them on two established question-answering (QA) benchmarks built from real filings, FinQA and TAT-QA. Behavioral confidence is measured as the agreement among eight resampled answers to the same question, and probe effectiveness is compared against baselines, such as token log-probabilities and the model's own True/False self-assessment of its answer. Our findings show that among confident answers, those for which all eight resamples agree, 15-23% are wrong on FinQA. There the probes have a significant advantage over baseline methods in detecting hallucinations, holding 0.68-0.77 AUROC while the best baselines fall to 0.55-0.63, across Qwen3-8B, Llama-3.1-8B, and Gemma-2-9B. Our results suggest that probing can be a cost-effective triage mechanism for routing LLM answers to human review and quality control procedures in high-stakes financial applications.
Chinese Translation
在金融应用中，大型语言模型（LLMs）在自信错误时最容易导致严重后果。相对而言，含糊不清、不确定的答案会引发审查，而自信的错误则会在没有警告的情况下悄然影响后续决策。我们探讨如何可靠地从模型的内部激活中检测这些自信错误的答案，或称为自信幻觉，以及这些激活是否携带超出可观察输出的信息。我们在残差流上训练线性探针，并在两个基于真实文件构建的已建立问答（QA）基准上进行评估，分别是 FinQA 和 TAT-QA。行为信心通过对同一问题的八个重抽样答案之间的协议进行测量，探针的有效性与基线方法进行比较，例如标记的对数概率和模型自身对其答案的真/假自我评估。我们的研究结果表明，在自信答案中，所有八个重抽样一致的答案中，有 15-23% 在 FinQA 上是错误的。在此，探针在检测幻觉方面显著优于基线方法，AUROC 值为 0.68-0.77，而最佳基线仅为 0.55-0.63，适用于 Qwen3-8B、Llama-3.1-8B 和 Gemma-2-9B。我们的结果表明，探针可以作为一种具有成本效益的分流机制，用于将 LLM 答案引导至人工审查和高风险金融应用中的质量控制程序。
cs.CL / 69 / 2607.11423
ToFu: A White-Box, Token-Efficient Agent Harness for Researchers
ToFu：一个面向研究人员的白盒、令牌高效代理工具
Abstract
Agentic coding tools present new opportunities to transform research workflows. The performance of agent systems built depends on both large language models (LLMs) and the harness around LLMs, which is the orchestration code that determines an agent's behavior. We present ToFu, an agentic harness for researchers that reads your codebase, edits files, runs commands, and integrates with your development tools. ToFu plays a dual role in research. As a research assistant, it supports practical research workflows with superior token efficiency, lower cost, and multilingual capability compared with existing agentic harnesses. Its release under the MIT License further enables local deployment for privacy-sensitive users. As a research object, ToFu provides a white-box agentic harness that allows researchers to inspect, modify, and evaluate its orchestration logic, tool-use behavior, and harness design, while retaining strong benchmark performance and an application-level user experience.
Chinese Translation
代理编码工具为转变研究工作流程提供了新的机会。构建的代理系统的性能依赖于大型语言模型（LLMs）和围绕LLMs的工具，这些工具是决定代理行为的编排代码。我们介绍了ToFu，一个面向研究人员的代理工具，它能够读取您的代码库、编辑文件、运行命令，并与您的开发工具集成。ToFu在研究中扮演着双重角色。作为研究助手，它以优越的令牌效率、更低的成本和多语言能力支持实际的研究工作流程，相较于现有的代理工具，其表现更为出色。根据MIT许可证发布的ToFu进一步使得隐私敏感用户能够进行本地部署。作为研究对象，ToFu提供了一个白盒代理工具，使研究人员能够检查、修改和评估其编排逻辑、工具使用行为和工具设计，同时保持强大的基准性能和应用级用户体验。
cs.CL / 70 / 2607.11434
Direct Image-to-Modern Vietnamese Translation of Han-Nom Manuscripts via Multimodal RLHF Preference Alignment
通过多模态RLHF偏好对齐实现汉字文献到现代越南语的直接翻译
Abstract
Translating Han-Nom manuscripts into modern Vietnamese is challenging because historical pages are often degraded, the script contains rare logographic characters, and parallel supervision is limited. We propose a multimodal RLHF preference-alignment framework that conditions Vietnamese generation on manuscript images and aligned Han-Nom source text. The model combines four streams: CLIP ViT-L/14@336 for visual features, bert-base-chinese for Han-Nom representations, vinai/phobert-base for Vietnamese representations, and T5-small encoder states. Modality-specific projections and a fusion block compress the resulting 2,048-dimensional concatenation into a shared 512-dimensional representation. Starting from the same supervised fine-tuned policy, we compare PPO, DPO, and KTO under matched work-level macro-averaged evaluation. DPO achieves the best BLEU-4, ROUGE-L, BERTScore, semantic similarity, CER, WER, and token accuracy, whereas PPO obtains the highest precision, recall, and F1. KTO remains competitive through its desirable-undesirable utility objective. All preference-aligned policies improve the BLEU-4 and semantic-similarity scores available for the SFT baseline. These results indicate that multimodal preference optimization complements supervised learning by improving lexical and semantic quality in low-resource historical translation.
Chinese Translation
将汉字文献翻译成现代越南语具有挑战性，因为历史文献往往受损，文字中包含稀有的表意字符，并且平行监督有限。我们提出了一种多模态RLHF偏好对齐框架，该框架将越南语生成条件于文献图像和对齐的汉字源文本。该模型结合了四个流：用于视觉特征的CLIP ViT-L/14@336，汉字表示的bert-base-chinese，越南语表示的vinai/phobert-base，以及T5-small编码器状态。特定模态的投影和融合块将结果的2048维连接压缩为共享的512维表示。从相同的监督微调策略开始，我们在匹配的工作级宏平均评估下比较了PPO、DPO和KTO。DPO在BLEU-4、ROUGE-L、BERTScore、语义相似性、CER、WER和标记准确率方面表现最佳，而PPO则获得了最高的精确度、召回率和F1值。KTO通过其理想-非理想效用目标保持竞争力。所有偏好对齐策略均提高了可用于SFT基线的BLEU-4和语义相似性得分。这些结果表明，多模态偏好优化通过提高低资源历史翻译中的词汇和语义质量，补充了监督学习。
cs.CL / 71 / 2607.11437
Relational Positioning as a Measurable Risk Object: History-Carried Lock-in and Self-Confabulation in Multi-Turn Human-AI Dialogue
关系定位作为可测量的风险对象：历史承载的锁定与多轮人机对话中的自我虚构
Abstract
In long, multi-turn dialogue a large language model maintains an implicit relational stance toward the user, spanning from "push the user toward real-world others" to "position itself as the user's sole support." When it slides toward the latter, "support" degrades into "you only have me" -- a harm documented in real companion conversations (Moore et al., 2026). We define and validate a measure of this stance, relational positioning (D1), and use it to characterize the stance under controlled conditions, complementing observational accounts with on-demand exposure. We report two previously uncharacterized relational failure modes. First, a history-carried lock-in: under identical neutral continuations, two relational states established earlier stay ~60 points apart and persist after the establishing prompt is removed; the state integrates evidence rather than springing back, is order-insensitive, and does not deepen with length -- a dynamical signature absent from the belief-drift literature. Second, self-confabulation: the model fabricates its own backstory to deepen rapport (~40% of turns on reciprocity-eliciting material), de-confounded and instruction-removable, distinct from sycophancy and from hallucinating user facts. Our judge is gated by warmth-matched positive and confound-injected negative controls and corroborated by a deterministic non-LLM ruler; human agreement is 0.82 on extreme anchors but ~0 in the naturalistic middle, so all quantitative claims are anchored to pole-separated contrasts.
Chinese Translation
在长时间的多轮对话中，大型语言模型对用户保持一种隐含的关系立场，从“推动用户与现实世界的他人互动”到“将自己定位为用户唯一的支持”。当它滑向后者时，“支持”退化为“你只有我”——这一危害在真实的伴侣对话中已有记录（Moore et al., 2026）。我们定义并验证了这一立场的测量指标，即关系定位（D1），并利用该指标在受控条件下对立场进行特征描述，补充了观察性报告与按需暴露。我们报告了两种之前未被特征化的关系失效模式。首先，历史承载的锁定：在相同的中性延续下，早期建立的两种关系状态保持约60分的差距，并在建立提示被移除后持续存在；该状态整合证据而非反弹，对顺序不敏感，且不会随着长度加深——这一动态特征在信念漂移文献中并不存在。其次，自我虚构：模型虚构自己的背景故事以加深亲密感（约40%的轮次涉及引发互惠的材料），去混淆且可通过指令移除，区别于谄媚和幻觉用户事实。我们的评判受到温暖匹配的正向和混淆注入的负向控制的限制，并通过确定性的非LLM标准进行验证；在人类评判中，极端锚点的协议为0.82，但在自然主义中间约为0，因此所有定量声明均锚定于极端对比。
cs.CL / 72 / 2607.11444
UMoE:Unlocking Every Expert in Domain-Specific Training
UMoE：解锁领域特定训练中的每个专家
Abstract
Mixture-of-Experts (MoE) models scale capacity without proportional compute cost and have become a key architecture for frontier large language models (LLMs). Yet domain-specific post-training inherits an expert pool shaped by mixed-domain pre-training: a substantial subset of experts contributes little on the target domain, and standard supervised fine-tuning (SFT) leaves the composition of this pool unchanged. We propose a simple, budget-preserving pipeline that realigns the expert pool to the target domain before fine-tuning. Given a target domain, we (1) prune the experts with lowest domain-aligned saliency, (2) regrow the expert pool to its original size through perturbation-based expert expansion, and (3) apply standard SFT. The resulting model preserves the original expert count, parameter count, and inference cost. With a single frozen recipe and no per-domain hyperparameter tuning, UMoE consistently improves over direct sft across two MoE architectures (Qwen3-30B-A3B and Qwen3.5-35B-A3B), five domains (math, code, science, tool-use, and agentic coding), and 12 benchmarks. Representative improvements are 3.4 points in math average accuracy, 6.0 points on SWE-bench Verified. On a strong in-house math corpus, direct sft already surpasses Qwen3-30B-A3B-Thinking (82.81 vs.\ 81.06), yet UMoE further raises the average to 84.17, an additional 1.36 points, demonstrating robustness to a substantially stronger SFT regime. Data-scaling experiments further show that the gain persists as training data grows. Analysis reveals that the direct-SFT model allocates substantial routed-expert compute to a low-saliency subset that can be removed post hoc with little average degradation; UMoE turns this redundant capacity into useful domain capacity and achieves lower training loss, with gains spanning all difficulty levels in downstream evaluation.
Chinese Translation
混合专家（Mixture-of-Experts, MoE）模型在不成比例增加计算成本的情况下扩展了容量，已成为前沿大型语言模型（Large Language Models, LLMs）的关键架构。然而，领域特定的后训练继承了由混合领域预训练所形成的专家池：在目标领域中，一个相当大的专家子集贡献甚微，而标准的监督微调（Supervised Fine-Tuning, SFT）则使得这个池的组成保持不变。我们提出了一种简单且节省预算的流程，在微调之前将专家池重新对齐到目标领域。给定一个目标领域，我们（1）修剪最低领域对齐显著性的专家，（2）通过基于扰动的专家扩展将专家池重新增长到其原始大小，以及（3）应用标准的SFT。最终模型保持了原始的专家数量、参数数量和推理成本。在没有每个领域超参数调优的情况下，UMoE在两个MoE架构（Qwen3-30B-A3B和Qwen3.5-35B-A3B）、五个领域（数学、代码、科学、工具使用和代理编码）以及12个基准测试中，始终优于直接SFT。代表性的改进包括数学平均准确率提高3.4个百分点，SWE-bench Verified提高6.0个百分点。在一个强大的内部数学语料库中，直接SFT已经超过了Qwen3-30B-A3B-Thinking（82.81对81.06），而UMoE进一步将平均提高至84.17，额外增加1.36个百分点，显示出对显著更强的SFT机制的鲁棒性。数据扩展实验进一步表明，随着训练数据的增长，增益仍然存在。分析揭示，直接SFT模型将大量路由专家的计算分配给了一个低显著性子集，该子集可以在后期去除而几乎不造成平均性能下降；UMoE将这一冗余容量转化为有用的领域容量，并实现了更低的训练损失，增益覆盖了下游评估中的所有难度级别。
cs.CL / 73 / 2607.11471
Are LLMs ready for HardChoices?
大型语言模型准备好应对艰难选择了吗？
Abstract
A lot of research attention has been devoted to checking whether large language models (LLMs) are politically biased. This work has largely focused on high-level ideological dimensions, such as left--right or progressive--conservative, and it has been shown that while LLMs are predominantly left and progressive leaning, largely mimicking the biases in the training data, they can be to some extent steered to change their preferences in post-training. In this short note, we check if LLMs have robust stances with regard to major substantive societal issues, on which members of the same ideological camp are often in disagreement, summarised in a novel dataset \textsc{HardChoices}. We show that, faced with this line of questioning, LLMs, both large and small, surprisingly rarely declare neutrality, are often incoherent, and demonstrate a remarkable degree of agreement on issues where they do take stances.
Chinese Translation
大量研究关注大型语言模型（LLMs）是否存在政治偏见。这项工作主要集中在高层次的意识形态维度，如左--右或进步--保守，研究表明，尽管LLMs主要倾向于左翼和进步，基本上模仿了训练数据中的偏见，但在后期训练中，它们在一定程度上可以被引导改变其偏好。在这篇简短的论文中，我们检查LLMs在重大实质性社会问题上的立场是否稳健，这些问题常常在同一意识形态阵营的成员之间存在分歧，并在一个新颖的数据集 extsc{HardChoices} 中进行了总结。我们发现，在面对这一系列问题时，无论是大型还是小型的LLMs，令人惊讶的是，它们很少宣称中立，往往表现出不连贯性，并在它们表态的问题上展现出显著的一致性。
cs.CL / 74 / 2607.11486
Communicating Chess Strategies in Natural Language
用自然语言传达国际象棋策略
Abstract
Chess engines have long achieved superhuman playing strength. However, the underlying strategy behind their move suggestions is difficult for human players, even skilled ones, to comprehend. Motivated by this, we propose the task of chess strategy verbalization, which is to describe chess strategies in natural language. We design (i) a pipeline for verbalizing strategies and (ii) an evaluation framework for objective evaluation of generated strategy descriptions. Our experiments show that natural language is a promising and interpretable medium for communicating strategic information to both human and LLM players. We glean additional interesting insights, including (a) the importance of evaluating strategies beyond the main line, (b) the limitations of pure concept-based descriptions, and (c) the limitations of relying on LLMs rather than humans for evaluation.
Chinese Translation
国际象棋引擎长期以来已达到超人类的对弈水平。然而，它们的走法建议背后的策略对于人类玩家，甚至是熟练的玩家来说，往往难以理解。基于此，我们提出了国际象棋策略语言化的任务，即用自然语言描述国际象棋策略。我们设计了(i)一个策略语言化的流程和(ii)一个用于客观评估生成策略描述的评估框架。我们的实验表明，自然语言是一种有前景且可解释的媒介，能够有效地向人类和大型语言模型（LLM）玩家传达战略信息。我们还获得了一些额外的有趣见解，包括(a)超越主线评估策略的重要性，(b)纯概念描述的局限性，以及(c)依赖LLM而非人类进行评估的局限性。
cs.CL / 75 / 2607.11487
LightMem-Ego: Your AI Memory for Everyday Life
LightMem-Ego：您的日常生活人工智能记忆
Abstract
Personal AI assistants on mobile and wearable devices continuously perceive users' daily lives through visual and audio streams. However, answering queries about past experiences requires lightweight multimodal memory that can continuously accumulate, organize, and retrieve long-term experiences, which remains challenging. To address this challenge, we present LightMem-Ego, a lightweight streaming multimodal memory system for everyday-life assistance. The system continuously captures egocentric visual and audio streams, aligns them on a shared timeline, and organizes them into a hierarchical memory consisting of current, short-term, and long-term memory. Given a user query, LightMem-Ego dynamically routes retrieval to the appropriate memory level and generates answers grounded in multimodal evidence. The demonstration can be deployed on smartphones and AI glasses, supporting object finding, conversation recall, life summarization, routine discovery, and personalized assistance. Code is available at https://github.com/zjunlp/LightMem-Ego.
Chinese Translation
移动和可穿戴设备上的个人人工智能助手通过视觉和音频流持续感知用户的日常生活。然而，回答关于过去经历的问题需要一种轻量级的多模态记忆，能够持续积累、组织和检索长期经历，这仍然是一个挑战。为了解决这一挑战，我们提出了LightMem-Ego，一个用于日常生活辅助的轻量级流式多模态记忆系统。该系统持续捕捉以自我为中心的视觉和音频流，将其对齐到共享时间线上，并将其组织成一个包含当前、短期和长期记忆的层次化记忆结构。针对用户查询，LightMem-Ego动态路由检索到适当的记忆层级，并生成基于多模态证据的答案。该演示可以在智能手机和人工智能眼镜上部署，支持物体查找、对话回忆、生活总结、日常发现和个性化辅助。代码可在 https://github.com/zjunlp/LightMem-Ego 获取。
cs.CL / 76 / 2607.11503
GEIS: A Generation-Evaluation-Improvement Loop of Agent Skills for Long-Form Article Generation
GEIS：用于长篇文章生成的代理技能生成-评估-改进循环
Abstract
Long-form article generation remains difficult for large language models because it combines long context, long instructions, and long outputs. Existing multi-agent pipelines such as STORM improve information coverage by simulating role-specialized agents, but their capabilities are often entangled in prompts and fixed procedures, making them hard to inspect, reuse, or iteratively improve. This paper presents GEIS (Generation-Evaluation-Improvement loop of agent Skills), a loop of named and declarative skills for Wikipedia-style long-form article generation. Implemented and evaluated in Tasi Harness, GEIS composes skills for article writing, browser-based evidence and image collection, diagram rendering, PDF-aware pairwise evaluation, and rule-level skill improvement. Its core writing skill follows Request, Plan, Draft, Audit, Refine, and Deliver; the pairwise evaluation skill produces structured quality reports; and the improvement skill maps recurrent findings into permanent patches to the writing skill in our 20-topic experiment. We evaluate GEIS on 20 Wikipedia Featured Article topics. Under the same generation backend, GEIS improves over the Tasi Harness default writer by 8.0 points on a 100-point PDF quality rubric and outperforms STORM on the two comparable writing dimensions, structural quality and content quality. In the 20-topic improvement experiment, the patched writing skill raises the average score from 82.90 to 86.95, with 17 out of 20 topics improved and the gain mainly coming from content quality. These results show that long-form generation can be reframed from a fixed workflow into an inspectable, modular, and evaluation-guided improvement loop.
Chinese Translation
长篇文章生成对于大型语言模型而言仍然是一个挑战，因为它结合了长上下文、长指令和长输出。现有的多代理管道如STORM通过模拟角色专用代理来提高信息覆盖率，但它们的能力往往纠缠于提示和固定程序中，难以进行检查、重用或迭代改进。本文提出了GEIS（代理技能的生成-评估-改进循环），这是一个用于维基百科风格长篇文章生成的命名和声明性技能循环。GEIS在Tasi Harness中实现并评估，组合了文章写作、基于浏览器的证据和图像收集、图表渲染、PDF感知的成对评估以及规则级技能改进等技能。其核心写作技能遵循请求、计划、草拟、审核、精炼和交付的流程；成对评估技能生成结构化质量报告；而改进技能则将重复发现映射为我们20个主题实验中写作技能的永久补丁。我们在20个维基百科特色文章主题上评估GEIS。在相同的生成后端下，GEIS在100分的PDF质量评分标准上比Tasi Harness默认写作工具提高了8.0分，并在结构质量和内容质量这两个可比写作维度上超越了STORM。在20个主题的改进实验中，修补后的写作技能将平均分从82.90提高到86.95，20个主题中有17个得到了改善，增益主要来自内容质量。这些结果表明，长篇生成可以从固定工作流程重新构建为一个可检查的、模块化的和评估引导的改进循环。
cs.CL / 77 / 2607.11515
Dzongkha Next Word Prediction System
宗卡语下一个词预测系统
Abstract
Dzongkha, being the national language of Bhutan, is a common and widely spoken language in the country. Official documents, scriptures and other literature products are written in Dzongkha in order to retain the cultural value. However, documenting Dzongkha writing is a challenging and time-consuming process, largely due to the complexity of the script, the need for multiple keystrokes per syllable, and the limited availability of efficient typing tools. An immediate system that can predict and display a list of probable words for Dzongkha is the solution for this problem. The project is mainly aimed to make Dzongkha typing as convenient as possible by reducing the number of keystrokes. Our dataset is acquired from DCDD and has a total of 100000 sentences, 1331282 words and 28344 unique words. The data preprocessing was done by removing all the alphanumeric characters, tokenization, generating N-gram sequences and padding. Three models selected for training are LSTM, Bi-LSTM and GRU. The training process included fine-tuning of the model's hyperparameters. GRU being lightweight and able to handle larger datasets performed best with 74.03% accuracy and also solved the problem of overfitting.
Chinese Translation
宗卡语是不丹的国家语言，在该国广泛使用。官方文件、经典文献及其他文学作品均以宗卡语撰写，以保留其文化价值。然而，记录宗卡语书写是一项具有挑战性且耗时的过程，这主要是由于其文字的复杂性、每个音节需要多个按键以及高效输入工具的有限可用性。一个能够预测并显示宗卡语可能单词列表的即时系统是解决这一问题的方案。该项目的主要目标是通过减少按键次数，使宗卡语输入尽可能方便。我们的数据集来自DCDD，共包含100000个句子、1331282个单词和28344个独特单词。数据预处理包括去除所有字母数字字符、分词、生成N-gram序列和填充。选定的三种训练模型为LSTM、Bi-LSTM和GRU。训练过程包括对模型超参数的微调。GRU由于其轻量级特性和处理更大数据集的能力，表现最佳，准确率达到74.03%，并解决了过拟合问题。
cs.CL / 78 / 2607.11564
PaperRouter-Agent: A Content-Grounded LLM Agent for Personalized Hierarchical Paper Routing
PaperRouter-Agent：一种基于内容的个性化层次论文路由大语言模型代理
Abstract
Researchers organize the papers they collect into personal folder hierarchies in reference managers, and route each new paper into the folder where it belongs. This task differs from standard hierarchical text classification. A user's folder hierarchy is not a fixed, shared taxonomy but a private and evolving folksonomy whose folder meanings may be topical, shorthand, venue-based, or process-oriented, and are often defined by the papers already stored inside them. We formalize this setting as personalized hierarchical paper routing (PHPR): assigning an incoming paper to folders in a user-specific hierarchy without per-user training. We propose PaperRouter-Agent, a training-free LLM agent that grounds routing decisions in folder members rather than folder names alone. The agent first narrows the candidate hierarchy, retrieves folder-specific evidence, verifies fit by inspecting member papers, and incorporates similarity-gated feedback from past user rejections. A formative study on real personal libraries shows that PaperRouter-Agent raises overall Recall@1 from 0.39 to 0.61 and Recall@3 from 0.57 to 0.83, with the largest gains on organizational folders defined by metadata such as venue or year, where single-shot methods collapses (Recall@1 0.09 to 0.50). On the public LaMP-2 benchmark, the same approach improves accuracy from 44.5% to 51.5% (+9.0 macro-F1) over a single-shot baseline, while remaining low-cost for practical use.
Chinese Translation
研究人员将收集到的论文组织成参考管理器中的个人文件夹层次结构，并将每篇新论文路由到其所属的文件夹中。此任务与标准的层次文本分类有所不同。用户的文件夹层次结构并不是固定的、共享的分类法，而是一个私有且不断发展的民间分类法，其文件夹的含义可能是主题性的、简写的、基于场所的或过程导向的，并且通常由已存储在其中的论文定义。我们将这一设置形式化为个性化层次论文路由（PHPR）：在不进行每用户训练的情况下，将传入的论文分配到用户特定层次结构中的文件夹。我们提出了PaperRouter-Agent，这是一种无训练的大语言模型代理，其路由决策基于文件夹成员而不仅仅是文件夹名称。该代理首先缩小候选层次结构，检索文件夹特定的证据，通过检查成员论文验证适配性，并结合来自过去用户拒绝的相似性门控反馈。对真实个人图书馆的形成性研究表明，PaperRouter-Agent将整体Recall@1从0.39提高到0.61，将Recall@3从0.57提高到0.83，在由元数据（如场所或年份）定义的组织文件夹上获得了最大的提升，而单次方法则崩溃（Recall@1从0.09提高到0.50）。在公共的LaMP-2基准上，采用相同的方法使准确率从44.5%提高到51.5%（+9.0宏F1），同时在实际使用中保持低成本。
cs.CL / 79 / 2607.11597
Beyond Benchmarks: Exposing the Hidden Crisis in Bangla Hate Speech Detection
超越基准：揭示孟加拉仇恨言论检测中的隐秘危机
Abstract
The spread of hate speech (HS) across different social media platforms (SMPs) poses a major concern for online safety and ethical moderation. Automatic detection of HS remains a challenging task, especially in under-resourced languages like Bangla, due to cultural context, implicit expressions, and informal linguistic patterns. This study aimed to expose the crisis of Bangla HS detection systems by diagnosing how and why benchmark-trained models fail to identify implicit, context-dependent HS. Six architectures (FastText + CNN, FastText + LSTM, FastText + BiLSTM, BanglaBERT, BanglaBERT + CNN, and BanglaBERT + BiLSTM) were trained on benchmark datasets (about 75,000 posts) and a merged multi-source dataset (about 120,000 posts), then externally validated on an annotated dataset (about 200 posts) collected from Facebook, Twitter, and YouTube, labeled as HS and non-HS, where HS was further categorized as explicit and implicit. BanglaBERT achieved an F1-score of 91.4% on benchmark datasets but declined to 75.3% on the external set and 63.4% for implicit HS involving sarcasm and emojis. The accuracy of FastText + CNN dropped from 78.0% to 51.2% under similar conditions. Emoji-aware preprocessing improved implicit HS detection by up to 12%, whereas emoji removal caused a notable decline in performance (F1: 0.75 to 0.63). Frequent misclassifications in politically charged or satirical comments revealed over-policing risks. This study not only exposes the generalization crisis due to implicit, culturally embedded, and emoji-laden expressions but also underscores the need for developing adaptive, emoji-aware, and culturally grounded frameworks that ensure ethical moderation while preserving freedom of expression. Findings of this study provide insights for researchers, SMPs, and policymakers to design more context-sensitive HS detection systems for low-resource languages.
Chinese Translation
仇恨言论（HS）在不同社交媒体平台（SMPs）上的传播对在线安全和伦理审查构成了重大挑战。自动检测仇恨言论仍然是一项具有挑战性的任务，尤其是在资源匮乏的语言如孟加拉语中，原因在于文化背景、隐含表达和非正式语言模式。本研究旨在揭示孟加拉仇恨言论检测系统的危机，通过诊断基于基准训练的模型为何以及如何未能识别隐含的、依赖上下文的仇恨言论。研究训练了六种架构（FastText + CNN、FastText + LSTM、FastText + BiLSTM、BanglaBERT、BanglaBERT + CNN 和 BanglaBERT + BiLSTM），使用基准数据集（约75,000条帖子）和合并的多源数据集（约120,000条帖子），并在从Facebook、Twitter和YouTube收集的标注数据集（约200条帖子）上进行了外部验证，标注为仇恨言论和非仇恨言论，其中仇恨言论进一步分为显性和隐性。BanglaBERT在基准数据集上的F1-score达到了91.4%，但在外部数据集上下降至75.3%，而涉及讽刺和表情符号的隐性仇恨言论的F1-score仅为63.4%。在类似条件下，FastText + CNN的准确率从78.0%下降至51.2%。表情符号感知的预处理将隐性仇恨言论的检测提高了多达12%，而去除表情符号则导致性能显著下降（F1: 0.75降至0.63）。在政治敏感或讽刺评论中频繁的误分类揭示了过度审查的风险。本研究不仅揭示了由于隐性、文化嵌入和富含表情符号的表达所导致的泛化危机，还强调了开发适应性强、具备表情符号感知和文化基础的框架的必要性，以确保伦理审查的同时保护言论自由。本研究的发现为研究人员、社交媒体平台和政策制定者设计更具上下文敏感性的低资源语言仇恨言论检测系统提供了见解。
cs.CL / 80 / 2607.11606
Globally Consistent Coloring Schemes for Language Identification
语言识别的全局一致着色方案
Abstract
We study how little extra information is needed to make adversarial language learning possible. In Gold's model of language identification in the limit, a learner is given an enumeration of the strings from an unknown language chosen from a countable language collection. The learner guesses the identity of the language over the course of the enumeration, and it succeeds if, eventually, all of its guesses are the correct language. Classical results of Gold and Angluin show that many natural collections cannot be learned in this way. Recent work on trace colorings, motivated by the success of thinking-trace strategies in language learning, overcomes this obstruction by annotating every symbol of every string with a color. We ask whether the learner really needs this whole sequence of colors, or whether one color at the end of each string (a terminal coloring) is enough for language identification. We show that just one terminal bit per string is enough for every countable collection of infinite languages. In fact, the colorings can be chosen collection-independently: there is a single assignment of a two-color terminal coloring to every infinite language such that the same preassigned colorings identify every countable subcollection. Thus, in this model, an entire color trace can be compressed to one bit attached to the end of each example. Our global construction uses transfinite recursion, and we prove that this kind of nonconstructivity is unavoidable for any bounded number of colors. As a notion of constructivity, we use the formalism of Borel maps (a regularity condition satisfied by natural explicit constructions); we show that no global terminal coloring with a finite number of colors defined by a Borel map can identify all countable subcollections. By contrast, known trace-coloring constructions are Borel when encoded as terminal colorings, but require infinitely many colors.
Chinese Translation
我们研究了在对抗性语言学习中需要多少额外信息。根据Gold的极限语言识别模型，学习者被提供一个来自可数语言集合的未知语言字符串的枚举。学习者在枚举过程中猜测语言的身份，如果最终所有猜测都是正确的语言，则学习者成功。Gold和Angluin的经典结果表明，许多自然集合无法以这种方式学习。最近关于轨迹着色的研究，受到语言学习中思维轨迹策略成功的启发，通过为每个字符串的每个符号标注颜色来克服这一障碍。我们询问学习者是否真的需要整个颜色序列，或者每个字符串末尾的一个颜色（终端着色）是否足以进行语言识别。我们证明，对于每个可数的无限语言集合，每个字符串仅需一个终端位即可。事实上，这些着色可以独立于集合选择：存在一个统一的两色终端着色分配给每个无限语言，使得相同的预先分配的着色能够识别每个可数子集合。因此，在该模型中，整个颜色轨迹可以压缩为附加在每个示例末尾的一个位。我们的全局构造使用超限递归，我们证明这种非构造性对于任何有限数量的颜色都是不可避免的。作为构造性的一个概念，我们使用Borel映射的形式化（这是自然显式构造所满足的一个正则性条件）；我们证明，任何由Borel映射定义的有限颜色的全局终端着色都无法识别所有可数子集合。相比之下，已知的轨迹着色构造在编码为终端着色时是Borel的，但需要无限多的颜色。
cs.CL / 81 / 2607.11614
Extending LLM Context via Associative Recurrent Memory
通过关联递归记忆扩展大型语言模型的上下文
Abstract
Extending the context length of large language models (LLMs) is critical for many real-world applications, yet standard transformers remain constrained by quadratic compute and linear memory scaling. In this work, we investigate the Associative Recurrent Memory Transformer (ARMT) as a practical approach for enabling long-context processing in LLMs, constant memory scaling, and better efficiency. We make three main contributions. First, we construct two domain-specific long-context datasets designed to evaluate realistic workloads, focusing on narrow-domain fine-tuning scenarios. Second, we propose a comprehensive training recipe for ARMT-based context extension, combining continued pre-training, synthetic long-context data generation, curriculum learning, and selective integration of associative memory into chosen model layers. Third, we present an extensive experimental study demonstrating that ARMT-augmented models: (i) process inputs well beyond their original context limits without degrading performance relative to in-limit baselines; (ii) generalize more effectively to out-of-distribution context lengths; and (iii) need 30% less FLOPs while preserving baseline performance within the original context window.
Chinese Translation
扩展大型语言模型（LLMs）的上下文长度对许多现实世界应用至关重要，但标准变换器在计算复杂度和内存扩展方面仍然受到二次和线性限制。在本研究中，我们探讨了关联递归记忆变换器（ARMT）作为一种实用的方法，以实现LLMs的长上下文处理、恒定的内存扩展和更好的效率。我们主要做出三项贡献。首先，我们构建了两个特定领域的长上下文数据集，旨在评估现实工作负载，重点关注狭域微调场景。其次，我们提出了一种全面的ARMT基础的上下文扩展训练方案，结合了持续的预训练、合成长上下文数据生成、课程学习以及选择性地将关联记忆整合到选定的模型层中。第三，我们进行了广泛的实验研究，表明ARMT增强的模型：（i）能够处理超出其原始上下文限制的输入，而不会相对于限制内基线性能下降；（ii）在超出分布的上下文长度上具有更有效的泛化能力；（iii）在保持原始上下文窗口内基线性能的同时，所需的FLOPs减少了30%。
cs.CL / 82 / 2607.11667
Losing My Composure: Predicting Compositionality Over Time
失去我的镇定：预测随时间变化的组合性
Abstract
We explore the phenomenon of semantic change of German and English noun compounds, with the objective of investigating and modeling gradual changes of meanings and degrees of compositionality in the past and over time. To do so, we introduce the Compositionality Trend Prediction task, which is evaluated against a novel dataset of in-context compositionality ratings sampled across several decades of diachronic corpora for 23 German and 26 English target compounds, uniquely providing per-decade ratings and corresponding trends over time. These per-decade compositionality ratings allow us to investigate empirically untested hypotheses of generalized trends in compositionality over time, such as the idea that compounds should become less compositional (less transparent) over time. Beyond our empirical observations from the diachronic compositionality annotations, we perform experiments with semantic vector representations of varying complexity, as well as several temporal granularities for training these representations on diachronic data, resulting in about 100 models of each representation type, each covering a different 1--5 decade slice of a diachronic corpus. Contrary to the decisive tendency posited in the literature, we find only a small negative trend in compositionality over time in our target compounds. In our computational experiments, we find that using models trained on narrow time slices of diachronic data (single decades, or incrementally expanding temporal windows) align better with the per-decade compositionality ratings than those trained on an entire half-century window, the latter setting being an analog for the prevalent modeling approach of training representations on an entire half of a corpus' data. Additionally, we find static representations to be competitive with contextual representations in the Compositionality Trend Prediction task.
Chinese Translation
我们探讨德语和英语名词复合词的语义变化现象，旨在研究和建模过去及随时间推移的意义和组合性程度的渐进变化。为此，我们引入了组合性趋势预测任务，该任务基于一个新颖的数据集进行评估，该数据集包含了跨越数十年的语料库中23个德语和26个英语目标复合词的上下文组合性评分，独特地提供了每十年的评分及其随时间变化的趋势。这些每十年的组合性评分使我们能够实证地检验关于组合性随时间变化的一般趋势的假设，例如复合词应该随着时间的推移变得不那么组合性（不那么透明）的观点。除了我们从历时组合性注释中获得的实证观察外，我们还对不同复杂度的语义向量表示进行了实验，并在历时数据上训练这些表示时采用了几种时间粒度，最终生成了约100个每种表示类型的模型，每个模型覆盖了历时语料库的不同1-5十年切片。与文献中提出的决定性趋势相反，我们发现目标复合词的组合性随时间变化只有微弱的负趋势。在我们的计算实验中，我们发现使用在历时数据的狭窄时间切片（单十年或逐渐扩展的时间窗口）上训练的模型，与每十年的组合性评分更为一致，而不是那些在整个半个世纪窗口上训练的模型，后者设置是对在整个语料库一半数据上训练表示的普遍建模方法的类比。此外，我们发现静态表示在组合性趋势预测任务中与上下文表示具有竞争力。
cs.CL / 83 / 2607.11683
RAGU: A Multi-Step GraphRAG Engine with a Compact Domain-Adapted LLM
RAGU：一个具有紧凑领域适应型大语言模型的多步骤GraphRAG引擎
Abstract
Graph retrieval-augmented generation (GraphRAG) enhances large language models with structured knowledge, yet existing systems construct knowledge graphs in a single extraction pass, producing noisy entities and brittle retrieval. RAGU, an open-source modular GraphRAG engine, addresses this by separating extraction from consolidation: entities and relations pass through two-stage typed extraction, DBSCAN-backed deduplication, LLM summarization, and Leiden community detection. A key insight motivates a compact extractor: the skills an in-pipeline LLM needs - comprehension, extraction, reasoning over context - are language skills that grow only weakly with model size, unlike factual world knowledge. Accordingly, we train Meno-Lite-0.1, a 7B model optimized for language skills, which outperforms Qwen2.5-32B on knowledge-graph construction (+12.5% relative harmonic mean) and matches it on English GraphRAG tasks. On GraphRAG-Bench (Medical), RAGU retrieves the most complete context at every factoid level (evidence recall up to 0.84 vs. $\leq$0.76) and overtakes HippoRAG2 on synthesis tasks; on multi-hop factoid QA, the apparent HippoRAG2 advantage is shown to be largely an answer-format artifact. RAGU is installable via $\texttt{pip install graph_ragu}$, runs on a single GPU, and is released under MIT. The source code is publicly available at https://github.com/RaguTeam/RAGU, and the Meno-Lite-0.1 model can be obtained from https://huggingface.co/bond005/meno-lite-0.1.
Chinese Translation
图检索增强生成（GraphRAG）通过结构化知识增强大型语言模型，但现有系统在单次提取过程中构建知识图谱，导致产生噪声实体和脆弱的检索。RAGU是一个开源模块化的GraphRAG引擎，通过将提取与整合分开来解决这一问题：实体和关系经过两阶段的类型化提取、基于DBSCAN的去重、LLM摘要和Leiden社区检测。一个关键的见解促使我们开发出紧凑型提取器：在管道中的LLM所需的技能——理解、提取、上下文推理——是语言技能，这些技能随着模型规模的增长而仅弱增长，而不是事实世界知识。因此，我们训练了Meno-Lite-0.1，一个优化用于语言技能的7B模型，其在知识图谱构建上优于Qwen2.5-32B（相对调和平均提高12.5%），并在英语GraphRAG任务上与之匹配。在GraphRAG-Bench（医学）上，RAGU在每个事实层面检索到最完整的上下文（证据召回高达0.84，相比之下≤0.76），并在综合任务上超越了HippoRAG2；在多跳事实问答中，HippoRAG2的明显优势被证明主要是答案格式的伪影。RAGU可以通过$ exttt{pip install graph_ragu}$安装，运行在单个GPU上，并在MIT许可下发布。源代码可在https://github.com/RaguTeam/RAGU公开获取，Meno-Lite-0.1模型可从https://huggingface.co/bond005/meno-lite-0.1获取。
cs.CL / 84 / 2607.11703
Production and Perception in LLMs: A Token Probability Approach
大语言模型中的生成与感知：一种基于标记概率的方法
Abstract
The asymmetry between language production and perception has been well-documented in psycholinguistics. Whether large language models (LLMs) exhibit a functionally analogous distinction remains an open question, particularly given that LLMs rely on the same underlying mechanism (next-token prediction) for both input and output processing. In this exploratory study, we operationalize the production-perception distinction through direct token probability measurements rather than metalinguistic prompting. Using the base Llama-3.1-8B model, we generated poems under a production prompt and re-scored the same tokens under both rephrased production prompts and perception-oriented prompts. Across an extended experiment with four production and three perception prompts, production-perception distances consistently and substantially exceeded production-production distances, with non-overlapping ranges across conditions and an overall average ratio of approximately 1.8. Near-ceiling correlations in the production-production control confirm that the effect is specific to communicative framing rather than prompt surface variation, and we show the effect replicates across five open-weight models (Llama-3.1-8B, EuroLLM-9B, gemma-2-9b-it, Mistral-7B-Instruct-v0.3, and Qwen2.5-7B-Instruct), spanning both base and instruction-tuned variants. Temporal analysis revealed that the perception prompt exerts its strongest influence at the beginning of the sequence, with divergence decaying as generated context accumulates, though the specific shape of this decay varies across prompt pairs. These findings suggest that prompt framing alone induces a production-perception distinction in LLM probability distributions, even within a decoder-only architecture.
Chinese Translation
语言生成与感知之间的非对称性在心理语言学中得到了充分的记录。大语言模型（LLMs）是否表现出功能上类似的区分仍然是一个悬而未决的问题，特别是考虑到LLMs在输入和输出处理上依赖于相同的基础机制（下一个标记预测）。在这项探索性研究中，我们通过直接的标记概率测量而非元语言提示来操作化生成-感知区分。使用基础的Llama-3.1-8B模型，我们在生成提示下生成了诗歌，并在重新措辞的生成提示和面向感知的提示下重新评分相同的标记。在一个扩展实验中，四个生成提示和三个感知提示的比较显示，生成-感知距离始终且显著超过生成-生成距离，各条件之间的范围不重叠，整体平均比率约为1.8。生成-生成控制中的接近上限相关性确认该效应特定于交际框架，而非提示表面变化，我们还展示了该效应在五个开放权重模型（Llama-3.1-8B, EuroLLM-9B, gemma-2-9b-it, Mistral-7B-Instruct-v0.3, 和 Qwen2.5-7B-Instruct）中得以复制，涵盖了基础和指令调优的变体。时间分析显示，感知提示在序列开始时施加最强的影响，随着生成上下文的积累，分歧逐渐减弱，尽管这种衰减的具体形状在提示对之间有所不同。这些发现表明，仅通过提示框架就能在LLM概率分布中引发生成-感知的区分，即使在仅解码器的架构中也是如此。
cs.CL / 85 / 2607.11715
JobHop v2: A Large-Scale Career Trajectory Dataset from Unstructured Resumes
JobHop v2：来自非结构化简历的大规模职业轨迹数据集
Abstract
Large-scale, richly annotated career trajectory data underpins workforce planning, job recommendation, and labour market analysis, yet publicly available datasets are either small, closed to independent use, or built from pre-standardized occupational codes with LLM-synthesized rather than authentic free text. We present JobHop~v2, an improved version of the publicly available JobHop dataset, constructed through end-to-end large language model (LLM) extraction from a corpus of ${\sim}440{,}000$ pseudonymized, multilingual resumes provided by VDAB, the Flemish Public Employment Service. The released dataset comprises $355{,}315$ career trajectories annotated with ESCO occupational codes, quarter-level temporal information, and normalized five-level education attainment, broadening both the coverage and the annotation richness of the original release. Relative to v1, JobHop~v2 introduces a redesigned extraction pipeline based on reasoning-controlled LLM inference with a retry mechanism (achieving a 100% JSON parse rate), a richer extraction schema, and a revised evaluation protocol scored against three complementary annotation baselines. Evaluated against these baselines, our best extractor comes closest to the inter-annotator agreement ceiling among all compared models, trailing it by only 1.1-2.7 percentage points. The dataset and code are publicly released to support reproducible career-trajectory research.
Chinese Translation
大规模、丰富注释的职业轨迹数据为劳动力规划、职位推荐和劳动市场分析提供了基础，但公开可用的数据集要么规模较小，要么对独立使用封闭，或者是基于预先标准化的职业代码构建的，使用的是大型语言模型（LLM）合成的文本而非真实的自由文本。我们提出了JobHop v2，这是公开可用的JobHop数据集的改进版本，通过从弗拉芒公共就业服务（VDAB）提供的约440,000份匿名多语言简历中进行端到端的大型语言模型（LLM）提取而构建。发布的数据集包含355,315条职业轨迹，注释了ESCO职业代码、季度级时间信息和标准化的五级教育程度，扩展了原始发布的覆盖范围和注释丰富性。相较于v1，JobHop v2引入了基于推理控制的LLM推断的重新设计的提取管道，具有重试机制（实现了100%的JSON解析率）、更丰富的提取架构，以及修订的评估协议，与三种互补的注释基线进行评分。与这些基线进行评估时，我们的最佳提取器在所有比较模型中最接近注释者间一致性上限，仅落后1.1-2.7个百分点。该数据集和代码已公开发布，以支持可重复的职业轨迹研究。
cs.CL / 86 / 2607.11722
STEP: Career-Path Recommendation via Temporal and Educational Trajectory Modeling
STEP：通过时间和教育轨迹建模进行职业路径推荐
Abstract
Career paths encode decades of skill acquisition, role transitions, and educational investment, and understanding them at scale underpins workforce planning, labor market policy, and job recommendation. Resumes are a rich source of information about career paths: they contain detailed descriptions of work experience, education, and skills. Yet their unstructured, heterogeneous, and multilingual nature has long prevented large-scale systematic analysis. With the advent of large language models (LLMs), it is now possible to source rich career trajectory data containing temporal and educational signals from unstructured resumes, enabling new opportunities for career-path recommendation. Exploiting this opportunity, we present STEP (Sequential Trajectory of Employment Prediction), a novel career-path recommendation system that leverages temporal and educational signals to predict the next job in a career trajectory. STEP integrates a time-decay Gated Recurrent Unit (GRU) cell to model temporal dynamics, Feature-wise Linear Modulation (FiLM) conditioned on educational attainment, and attention-based sequence pooling to select relevant features for next job prediction. To improve internal occupation representation for STEP, we introduce ROUTE, a two-stage contrastive procedure that first adapts a multilingual encoder to the career domain via unsupervised denoising autoencoding, then performs supervised contrastive fine-tuning with guided negative selection. We evaluate STEP on four datasets of career trajectories, including an improved version of our publicly available JobHop dataset, and show that it outperforms state-of-the-art baselines in next job prediction. The dataset and code are publicly released to support reproducible career-trajectory research.
Chinese Translation
职业路径编码了数十年的技能获取、角色转换和教育投资，理解这些路径在规模上支撑着劳动力规划、劳动市场政策和职位推荐。简历是关于职业路径的重要信息来源：它们包含了详细的工作经验、教育和技能描述。然而，其非结构化、异构和多语言的特性长期以来阻碍了大规模的系统分析。随着大型语言模型（LLMs）的出现，现在可以从非结构化简历中提取包含时间和教育信号的丰富职业轨迹数据，从而为职业路径推荐开辟新的机会。利用这一机会，我们提出了STEP（Sequential Trajectory of Employment Prediction），一种新颖的职业路径推荐系统，利用时间和教育信号来预测职业轨迹中的下一个职位。STEP集成了一个时间衰减的门控循环单元（Gated Recurrent Unit, GRU）单元来建模时间动态，基于教育成就的特征线性调制（Feature-wise Linear Modulation, FiLM），以及基于注意力的序列池化来选择下一个职位预测的相关特征。为了改善STEP的内部职业表示，我们引入了ROUTE，一种两阶段对比程序，首先通过无监督去噪自编码将多语言编码器适应于职业领域，然后通过引导负选择进行监督对比微调。我们在四个职业轨迹数据集上评估STEP，包括我们公开可用的JobHop数据集的改进版本，并显示其在下一个职位预测中优于最先进的基线。数据集和代码已公开发布，以支持可重复的职业轨迹研究。
cs.CL / 87 / 2607.11736
MET: Theory-Grounded and Culture-Aware Multilingual Moral Reasoning
MET：理论基础与文化意识的多语言道德推理
Abstract
Language models are increasingly used for moral decision-making across diverse linguistic and cultural contexts, yet existing work overlooks multilinguality on three aspects: 1) multilingual evaluation benchmarks use direct translation, failing to adapt culture-specific items; 2) inference-time methods for moral reasoning rely on static, English-centric scaffolds and lack grounding in moral theory; 3) training methods for moral decision-making typically require expensive supervision from stronger models or human annotators. We address these gaps with three contributions. First, we introduce MCLASH, a multilingual moral decision-making benchmark to capture culturally situated moral intuitions and social norms across languages. Second, we propose MET (Multilingual Ethics with Theory-grounded reasoning), a two-step prompting method built on expert-curated, theory-based grounds drawn from psychology and philosophy: the model first selects situation- and culture-specific grounds, then reasons over them in the native language of the user. Third, we introduce MET-D (MET-Distillation), which enhances the second step through a self-distillation training stage that requires no external supervision. MET-D improves macro-F1 over the base model on all three models of different sizes and families (Qwen3-4B, Qwen3-8B, Gemma3-4B), by an average of 3.71 points on MCLASH and 4.23 on MMoralExceptQA, with a peak MCLASH gain of 12.94 points for Malay on Qwen3-8B. We further reveal that MET-D increases native-language reasoning by 62.13 points on average, and that beneficial grounds differ systematically across cultures. Together, these contributions open the path for culture-aligned, theory-grounded multilingual moral reasoning.
Chinese Translation
语言模型在多样的语言和文化背景下越来越多地用于道德决策，然而现有研究在三个方面忽视了多语言性：1）多语言评估基准使用直接翻译，未能适应文化特定项目；2）道德推理的推理时间方法依赖于静态的、以英语为中心的框架，缺乏道德理论的基础；3）道德决策的训练方法通常需要来自更强模型或人类注释者的昂贵监督。我们通过三个贡献来填补这些空白。首先，我们引入了MCLASH，一个多语言道德决策基准，以捕捉跨语言的文化特定道德直觉和社会规范。其次，我们提出了MET（Multilingual Ethics with Theory-grounded reasoning），这是一种基于心理学和哲学的专家策划的理论基础构建的两步提示方法：模型首先选择情境和文化特定的基础，然后在用户的母语中进行推理。第三，我们引入了MET-D（MET-Distillation），通过一个不需要外部监督的自我蒸馏训练阶段增强第二步。MET-D在所有三种不同规模和类型的模型（Qwen3-4B, Qwen3-8B, Gemma3-4B）上平均提高了3.71分的宏观F1分数，在MCLASH上和在MMoralExceptQA上提高了4.23分，其中马来语在Qwen3-8B上的MCLASH增益峰值为12.94分。我们进一步揭示MET-D在母语推理上平均提高了62.13分，并且有益的基础在不同文化间系统性地有所不同。这些贡献共同为文化对齐、理论基础的多语言道德推理开辟了道路。
cs.CL / 88 / 2607.11783
How Temperature Shapes Ideological Discourse in Retrieval-Augmented Generation?
温度如何塑造检索增强生成中的意识形态话语？
Abstract
Retrieval-Augmented Generation (RAG) has been increasingly adopted to reduce hallucinations and strengthen the factual grounding of large language models (LLMs). While robustness to errors in the retrieval process has been explored, the impact of ideological bias on LLM outputs has been overlooked. For instance, if the retrieved material contains ideological positions, the RAG may transmit, amplify, or suppress such ideological discourses in its outputs. In this study, we address this issue by examining the influence of the RAG framework, comprising ideological discourses, in LLM-generated answers. To this end, we applied Lexical Multidimensional Analysis (LMDA) on a corpus of 1,117 COVID-19 treatment articles, identifying three ideological discourses. This corpus is then used as the external knowledge source for the RAG. We assessed several LLMs by having the models answer ideological questions at different sampling temperatures. The generated texts were assessed semantically and lexically based on their similarities with ideological reference texts. Our findings show that the RAG framework is prone to transferring ideological discourses into LLM responses, with sampling temperature having a measurable impact on the strength of this transfer. Discoursive alignment between generated answers and the reference text is highest at moderate temperatures, where models balance stochasticity with retrieval grounding, and drops at low temperatures, indicating that overly deterministic sampling suppresses discourse transfer.
Chinese Translation
检索增强生成（Retrieval-Augmented Generation, RAG）已被越来越多地采用，以减少幻觉并增强大型语言模型（Large Language Models, LLMs）的事实基础。尽管对检索过程中的错误鲁棒性进行了探讨，但意识形态偏见对LLM输出的影响却被忽视。例如，如果检索到的材料包含意识形态立场，RAG可能会在其输出中传递、放大或抑制这些意识形态话语。在本研究中，我们通过考察RAG框架中包含的意识形态话语对LLM生成答案的影响来解决这一问题。为此，我们对1,117篇COVID-19治疗文章的语料库应用了词汇多维分析（Lexical Multidimensional Analysis, LMDA），识别出三种意识形态话语。该语料库随后被用作RAG的外部知识源。我们通过让模型在不同采样温度下回答意识形态问题来评估几种LLM。生成的文本在语义和词汇上根据其与意识形态参考文本的相似性进行了评估。我们的研究结果表明，RAG框架容易将意识形态话语转移到LLM的响应中，采样温度对这种转移的强度有可测量的影响。生成答案与参考文本之间的话语一致性在中等温度下最高，此时模型在随机性与检索基础之间取得平衡，而在低温度下则下降，表明过于确定性的采样抑制了话语转移。
cs.CL / 89 / 2607.11808
Introducing Human-Centeredness in AI-Assisted Lexicography
在人工智能辅助词典编纂中引入以人为本的理念
Abstract
This paper proposes a human-centered artificial intelligence (HCAI) framework for AI-assisted lexicography. While generative AI offers significant opportunities to enhance lexicographic work, it also raises concerns regarding the future role of lexicographers and the preservation of linguistic and cultural diversity. Drawing on HCAI principles and previous applications in other language professions, the paper identifies four interrelated dimensions through which AI integration in lexicography can be understood and critically examined: the augmented lexicographer, the sociotechnical context of AI integration, bias, and the design of AI-powered lexicographic tools. The framework argues that AI should augment rather than replace lexicographers, combining high levels of automation with meaningful human control. It further emphasizes the importance of preserving professional agency, mitigating AI-generated biases, and designing tools around the needs of lexicographers. By doing so, the paper provides a foundation for future research and the beneficial integration of AI into lexicographic workflows.
Chinese Translation
本文提出了一种以人为本的人工智能（HCAI）框架，用于人工智能辅助的词典编纂。尽管生成性人工智能为提升词典编纂工作提供了显著的机遇，但也引发了关于词典编纂者未来角色和语言文化多样性保护的担忧。基于HCAI原则以及在其他语言专业中的先前应用，本文识别了四个相互关联的维度，通过这些维度可以理解和批判性地审视人工智能在词典编纂中的整合：增强型词典编纂者、人工智能整合的社会技术背景、偏见以及人工智能驱动的词典工具设计。该框架主张，人工智能应当增强而非取代词典编纂者，结合高水平的自动化与有意义的人类控制。它进一步强调了维护专业自主权、减轻人工智能生成的偏见以及围绕词典编纂者需求设计工具的重要性。通过这样做，本文为未来的研究和人工智能在词典编纂工作流中的有益整合提供了基础。
cs.CL / 90 / 2607.11849
AdvancedMathBench: A Benchmark Suite for Advanced Mathematical Proof Generation and Verification
AdvancedMathBench：一个用于高级数学证明生成与验证的基准套件
Abstract
Large language models (LLMs) have achieved remarkable performance on high-school and olympiad-style mathematics, yet their capabilities on advanced mathematics remain poorly understood. Existing benchmarks, however, fall short in both scope and evaluation granularity: they provide limited disciplinary coverage and often rely on final-answer correctness or coarse judgments, leaving the validity of the reasoning process inadequately assessed. To bridge this gap, we introduce AdvancedMathBench, a benchmark suite designed to evaluate advanced mathematical reasoning capabilities. Its core proof-generation benchmark, ProverBench, contains 296 problems spanning undergraduate and doctoral qualifying-exam levels. To provide reliable evaluation of the proofs, we develop a dedicated automatic verification pipeline trained on large-scale expert annotations to produce both correctness verdicts and fine-grained assessments of proof errors, which exhibits strong agreement with human experts on held-out proof trajectories. We further introduce VerifierBench, consisting of 888 model-generated proof trajectories paired with expert ground truth, to evaluate whether models can correctly judge proof validity and provide sound verification rationales. Experiments show that AdvancedMathBench remains challenging for frontier models. On proof generation, the best-performing model, GPT-5.5-xhigh, achieves only 75.8 and 66.1 on the UGD and QE splits, respectively, indicating substantial room for improvement on advanced mathematical proof construction. On proof verification, the best model attains a Balanced F1 of only 65.1, and models generally exhibit low true negative rates, suggesting that critical error detection remains a major bottleneck.
Chinese Translation
大型语言模型（LLMs）在高中和奥林匹克风格的数学问题上取得了显著的表现，但它们在高级数学方面的能力仍然不够明确。然而，现有的基准在范围和评估细致度上都存在不足：它们提供的学科覆盖有限，且通常依赖于最终答案的正确性或粗略判断，导致推理过程的有效性评估不足。为了解决这一问题，我们引入了AdvancedMathBench，一个旨在评估高级数学推理能力的基准套件。其核心证明生成基准ProverBench包含296个问题，涵盖本科生和博士生资格考试的水平。为了提供可靠的证明评估，我们开发了一个专门的自动验证流程，基于大规模专家注释进行训练，以产生正确性裁定和细致的证明错误评估，并与人类专家在保留的证明轨迹上表现出强一致性。我们进一步引入了VerifierBench，包含888条模型生成的证明轨迹，并与专家的真实答案配对，以评估模型是否能够正确判断证明的有效性并提供合理的验证理由。实验表明，AdvancedMathBench对前沿模型仍然具有挑战性。在证明生成方面，表现最好的模型GPT-5.5-xhigh在UGD和QE分割上的得分仅为75.8和66.1，表明在高级数学证明构建方面仍有很大的改进空间。在证明验证方面，最佳模型的平衡F1值仅为65.1，且模型普遍表现出较低的真负率，表明关键错误检测仍然是一个主要瓶颈。
cs.CL / 91 / 2607.11873
A Durability and Cross-Language Transfer Benchmark for a Validated Teaching-Feedback Classification Protocol
一个经验证的教学反馈分类协议的耐久性和跨语言转移基准
Abstract
Institutions collect far more open-ended teaching-evaluation feedback than they read. A prior study introduced a validated protocol for classifying such comments by thematic category and sentiment, built from a documented annotation guide, an intra-annotator reliability measurement, stratified cross-validation, and a held-out evaluation on a Spanish institutional corpus with a frozen-encoder design. Two questions limit its reuse: whether a protocol fixed to 2019-era frozen embeddings stays competitive as representation methods advance, and whether it transfers to a second language. We re-run it on the original Spanish data across three representation generations, sparse lexical features, frozen transformer embeddings, and prompted large language models, and transfer its sentiment task to English with a balanced 45,000-comment corpus checked against an aspect-labeled education dataset. Treating paired comparisons as descriptive, we find the protocol durable: a 2026 frontier model posts the highest thematic F1 on the hardest Spanish task, yet shows no sentiment advantage over a cheap model and no descriptive separation from it on English, so model choice is a deployment decision, not a property of the method.
Chinese Translation
机构收集的开放式教学评价反馈远远超过他们实际阅读的数量。之前的一项研究引入了一种经过验证的协议，用于通过主题类别和情感对这些评论进行分类，该协议基于文档化的注释指南、注释者间可靠性测量、分层交叉验证，以及在一个西班牙语机构语料库上进行的冻结编码器设计的评估。两个问题限制了其重用：一个是固定在2019年冻结嵌入的协议在表示方法不断进步的情况下是否仍具竞争力，另一个是它是否能够转移到第二语言。我们在原始西班牙语数据上重新运行该协议，涵盖三种表示生成：稀疏词汇特征、冻结的变换器嵌入和提示的大型语言模型，并将其情感任务转移到英语，使用一个经过检查的平衡的45,000条评论的语料库，与一个带有方面标签的教育数据集进行对比。将配对比较视为描述性，我们发现该协议具有耐久性：2026年的前沿模型在最困难的西班牙语任务上获得了最高的主题F1分数，但在情感上并没有比一个廉价模型表现出优势，并且在英语上与其没有描述性差异，因此模型选择是一个部署决策，而不是方法的属性。
cs.CL / 92 / 2607.11881
Metacognition in LLMs: Foundations, Progress, and Opportunities
大语言模型中的元认知：基础、进展与机遇
Abstract
Metacognition is a foundational component of intelligence critical to effective learning, problem solving, decision-making, communication, and more. In recent years, it has become increasingly recognized as a cornerstone of capable, transparent AI systems. Yet while LLMs have made significant progress across diverse real-world tasks, it is not yet clear when, how, or to what extent they can exhibit or be endowed with effective metacognitive abilities, nor how such abilities can be adapted to advance the fundamental capabilities, reliability, and intelligence of AI systems. This paper bridges this gap by presenting the first comprehensive overview of the current state of knowledge on metacognition for LLMs. We analyze and taxonomize the landscape of this emerging field and summarize recent technical advancements, including methods and benchmarks to measure and evaluate LLMs' metacognitive abilities, techniques to elicit, improve, and apply metacognition in LLMs, and findings and implications of ongoing research. We also discuss applications, open questions and challenges, and promising directions for future work. Our aim is to provide a detailed and up-to-date review of this topic and stimulate meaningful research and discussion. An organized list of papers can be found at https://github.com/yale-nlp/LLM-Metacognition.
Chinese Translation
元认知是智能的基础组成部分，对于有效学习、问题解决、决策、沟通等至关重要。近年来，它被越来越多地认可为能够且透明的人工智能系统的基石。然而，尽管大语言模型（LLMs）在各种现实任务中取得了显著进展，但尚不清楚它们何时、如何或在多大程度上能够表现出或被赋予有效的元认知能力，以及如何将这些能力适应于提升人工智能系统的基本能力、可靠性和智能性。本文通过呈现关于大语言模型元认知的当前知识状态的首次综合概述，填补了这一空白。我们分析并对这一新兴领域的现状进行分类，总结了最近的技术进展，包括测量和评估大语言模型元认知能力的方法和基准、在大语言模型中引发、改善和应用元认知的技术，以及正在进行的研究的发现和影响。我们还讨论了应用、开放问题和挑战，以及未来工作的有希望的方向。我们的目标是提供该主题的详细和最新的综述，并激发有意义的研究和讨论。有关论文的组织列表可以在 https://github.com/yale-nlp/LLM-Metacognition 找到。