Physical AI Brief
Daily cross-source signals for the Physical AI supply chain — silicon photonics, CPO, VLA models, humanoid hardware, embodied AI. Three streams, one page, zero filler.
484 items today · 421 arxiv · 1 SEC 8-K · 62 humanoid · 0 CN photonics
01 ARXIV · PHYSICAL AI PAPERS
421 items- arxiv:2610.22372 · cs.ROLIBERO-PeRM: Benchmarking Personalized Robotic ManipulationZhixu Li, Keqian Tang, Litian Gong, Jingyu Yao +6
Personalization is essential for general-purpose robots to transition into household environments and achieve true practicality. While general robotic manipulation has been extensively studied, personalized robotic manipulation (PeRM) lacks a controlled benchmark for separating task completion from preference satisfaction. We introduce LIBERO-PeRM, a benchmark for preference learning in embodied agents. LIBERO-PeRM studies five questions: 1) how policies follow preferences in zero-shot settings; 2) how they learn single preferences from demonstrations; 3) how efficiently they acquire preferences as demonstrations increase; 4) how learned preferences generalize across layouts and tasks; and 5) how policies compose preferences and behave under conflicts. We propose a controlled formulation and taxonomy of observable preferences in simulation, covering motion, skill, and user-profile choices. This formulation is restricted to language-expressible preferences that can be instantiated in LIBERO without changing the task goal, evaluated automatically, and reused across tasks. Building on LIBERO, our procedural pipeline generates paired demonstrations with and without preferences. We create five task suites (312 tasks) with high-quality demonstrations, rich metadata, and preference-satisfaction metrics. Experiments show that current VLA policies struggle with zero-shot personalization, while SFT makes preferences learnable. Mixed training achieves better no-preference retention than sequential adaptation and improves performance with contextual prompts compared with preference-only training, but can weaken motion-preference learning; preference generalization, decomposition, and conflict handling remain difficult. The code, datasets, and demos can be found on the project website: https://libero-perm.github.io/.
vlaembodiedmanipulationliberoembodied agentbenchmark - arxiv:2610.22368 · cs.ROValue as Memory: Online Value-Guided Post-training for General Robot PoliciesWenye Yu, Yijin Chen, Kaiyi Xiong, Wenbo Han +6
Real-world post-training of generalist robot policies faces two challenges: partial observability makes action selection ambiguous, while offline post-training methods delay feedback from deployment experience. We present Value as Memory (VaM), a framework that addresses these challenges through value caching and online real-world reinforcement learning (RL). Value caching conditions actions on recent value predictions as task-progress memory, while online updates enable timely policy improvement by combining offline demonstrations, off-policy interventions, and on-policy rollouts for sample-efficient adaptation. As a result, VaM transitions from offline RL warmup to online autonomous learning with sparse rewards. Across long-horizon tasks involving deformable objects and contact-rich manipulation, VaM yields a mean absolute success-rate gain of $29.4\%$ with value caching, while achieving a $32.8\%$ gain with online autonomous learning over matched offline training. The evaluated online schedule performs $8\times$ as many policy iterations under the same gradient-step budget, with more than $2\times$ lower recorded time per gradient step with concurrent data collection and training.
manipulationmemorypost-training - arxiv:2610.22367 · cs.AIDiscovering state-of-the-art quantum circuits with large language modelsAndi Gu, Alek H. Kemeny, Susanne F. Yelin, Mikhail D. Lukin
Significant efforts are currently being directed toward the implementation of large-scale quantum information processors, but realizing useful quantum advantage also requires the development of efficient quantum algorithms. While for many practical problems the dominant resource costs are associated with a small set of computational primitives, optimizing them in the quantum setting presents unique challenges, including a reversibility constraint absent from ordinary classical computation. Here we introduce an evolutionary framework in which large language models (LLMs) propose circuit designs and an independent evaluator verifies and scores them. We apply this framework to tasks spanning numerical computation, chemistry, optimization, quantum simulation, and number theory. In seven of the eight tasks, the discovered circuits significantly reduce both Toffoli and qubit counts relative to the state of the art. In the eighth, the search discovers a new reversible architecture for prime-field modular inversion. By handling reversibility differently, the architecture makes divsteps -- a classical algorithm previously impractical for reversible integer arithmetic -- viable and substantially improves the Toffoli-qubit trade-off for modular inversion. These results open the door for evaluator-guided LLM co-design of new algorithmic structures for quantum computation.
evaluator - arxiv:2610.22364 · cs.ROT2-z: Test-Time Optimization of Environment Latents for Robotic World ModelsChenyuan Zhou, Yining Hong, Huang Huang, Yunfan Jiang +8
Predicting how the world responds to an action requires more than what can be seen: robots must account for environmental factors that are neither directly observable nor explicitly specified, from physical properties such as friction and mass to human-defined environmental rules. We introduce T2-Z for test-time optimization of environment latents, which augments a robotic world model with a learnable environment latent to condition world predictions. This compact latent captures underlying environmental variation without supervision of what environmental factors exist or what values they take. During training, T2-Z alternates between two stages: a world-model optimization stage that fixes the latents and learns general action-conditioned dynamics across environments; and an environment-latent optimization stage that freezes the world model and optimizes the conditioning latents, allowing latent structure to emerge purely through world prediction. At test time, the world model remains frozen and a newly initialized environment latent is optimized from one or more test-time interactions, adapting the model to previously unseen environmental configurations. Across 7 simulated settings and 4 real-world robotic tasks, T2-Z reduces object prediction error metric from 15.20 to 7.00 and raises action success from 33.73% to 72.96%. Its emergent latents capture structured representations of individual and compounded factors across diverse visual scenes. Ablations further show that T2-Z can efficiently infer environment latents from as little as a single test-time interaction trajectory.
world modelaction-conditioned - arxiv:2610.22363 · cs.ROPerceive, Predict, and Plan: a Unified World Action Model for Egocentric Dexterous ManipulationZirui Wu, Chen Guo, Tianjian Jiang, Zeren Jiang +10
We introduce P$^{3}$-WAM, a unified world action model for egocentric dexterous manipulation that learns 3D perception, action-conditioned prediction, and goal-directed planning through three complementary modes: Perceive, which reconstructs articulated hand motion and camera trajectories from video; Predict, which generates visual futures under prescribed actions; and Plan, which jointly generates action trajectories and their anticipated visual consequences conditioned on an initial image, a language instruction, and an optional visual goal. Our model integrates a compact, structured 3D action representation into a pretrained video diffusion transformer through a mixture-of-transformers architecture with shared base weights, modality-specific adaptation, and bidirectional visual-action attention. Randomized modality visibility lets video and 3D actions serve as either conditioning context or prediction targets across the three modes. Experiments show that unified training improves all three capabilities over single-mode training, including gains in planning quality under a vision-language model evaluation. Through unified training, our model outperforms specialized baselines on 3D hand reconstruction and action-conditioned video prediction across four datasets, while producing goal-directed manipulation videos with stronger instruction adherence and action progression than video-generation baselines. Project Page: https://p3-wam.pages.dev/
manipulationdexterousaction-conditioned - arxiv:2610.22362 · cs.ROCoDA: Co-Evolution of Data Engines and Action PoliciesTianjiao Yu, Zhilin Ou, Yifan Shen, Xinzhuo Li +4
Generative simulation can automate robot environment construction, but physically valid environments are not necessarily useful for training. The challenge is to identify environments that address a policy's failures and supervision that helps it overcome them. We introduce CoDA, a framework in which a simulation data engine and an action policy co-evolve through a shared heterogeneous memory of environment construction, policy failures, and learning progress. The memory compiles validated diagnoses from a vision-language critic into generation programs that recreate difficulties and training programs that help the policy overcome them. We further propose Text2potential to convert these diagnoses into staged potentials for action-chunk value learning, providing dense supervision for policy updates while preserving the task's optimal policy under standard shaping assumptions. Measured learning progress guides further generation and program refinement. Across four manipulation task families, CoDA improves out-of-distribution success from 5.3% after supervised fine-tuning to 27.6%, exceeding the strongest closed-loop baseline by 9.5 percentage points with the same generation budget. On LIBERO-Pro, 15 rounds of generation and training raise success from 25% to 84% under position perturbations and from 1% to 76% under instruction perturbations.
manipulationliberomemory - arxiv:2610.22360 · cs.CVGrounding Generative Verification in 3D ContentShuanglin Li, Jiaying Lin, Shuquan Ye
Recent advances in multimodal large language models (MLLMs) and 3D generation have expanded 3D understanding and synthesis. However, beyond visual fidelity, task-specific correctness grounded in 3D evidence remains a major challenge. We argue that building reliable 3D systems requires not only stronger generation and reasoning, but also verification feedback that identifies unmet requirements and explains their violations, allowing systems to reflect on and refine their outputs. Such feedback must integrate multi-view observations and distinguish genuine inconsistencies from projection-dependent appearances. We make three contributions. First, we introduce 3DVerifierBench, a benchmark spanning 10 tasks that evaluates decisions, rejection explanations, and responses to controlled changes in specifications and candidates. A human study reveals a 21.6 percentage-point gap between humans and the strongest tested model, GPT-6 Astra, while explanation-aware analysis exposes weaknesses in rejection reliability. Second, we analyze verification across fine-grained factual categories, exposing errors in values and target binding. We develop a scalable pipeline combining multi-view fact extraction, controlled 3D editing, and native annotations to produce 14,126 training examples with contradiction explanations and linked counterfactuals. Third, we examine how verification could support 3D generation through candidate selection and recursive refinement. In human-labeled sixteen-candidate pools, oracle selection improves complete factual success from 26.7% to 90.0%. In a separate eight-input study, human-guided recursive refinement satisfies all seven requirements for every input, compared with three under parallel Oracle@8. Together, our benchmark, analysis, and data pipeline establish a foundation for models to reflect on 3D content through reliable verification feedback.
benchmark - arxiv:2610.22358 · cs.LGPathFM: Fast Generative Forecasting of Multivariate TrajectoriesJiachen Yao, Xiyuan Zhang, Annan Yu, Boran Han +4
Time series foundation models are becoming increasingly popular for general-purpose forecasting. However, most leading models predict marginal quantiles, leaving the dependence structure of future trajectories unspecified. This distinction matters for trajectory-dependent quantities such as cumulative demand, peak load, and event timing, yet point and marginal scores do not reveal it. To model this dependence, we introduce PathFM, a trajectory-based forecasting model that uses flow matching to generate complete future paths. We further distill the model to one or two steps and share history attention cache across draws. Together, these enable PathFM to generate multiple paths for a single series in roughly the time needed for one query to a quantile model. The distilled sampler also makes it practical to optimize a calibration loss directly. PathFM achieves competitive marginal accuracy on the established forecasting benchmarks fev-bench and TIME. To complement these evaluations, we introduce Forecast Path Score Bench (fps-bench), which augments fev-bench tasks with energy and variogram scores and downstream metrics. PathFM achieves a higher aggregate path score on fps-bench than all evaluated external baselines. Experiments on chaotic systems further show that our flow model sustains long-run variability, while strong baselines rapidly collapse. PathFM makes generative time series forecasting both competitive and practical, enabling diverse downstream applications.
benchmark - arxiv:2610.22356 · cs.CVDK-VLM: Geometry as Position for Spatial Reasoning in Vision-Language ModelsHan Su, Meng Tian, Tianyu Huang, Mingyu Gao +6
Vision-language models still struggle to understand 3D spatial relations across views and update spatial judgments as the observer moves. A common approach fuses 3D reconstruction features into visual tokens, but aligning these features with pretrained semantic representations remains challenging under limited spatial supervision. We propose DK-VLM with Dual-Kernel Attention, which computes attention from both a visual token's input position and its geometric position in the scene. A semantic kernel retains the pretrained model's native attention and positional encoding, while a separate geometric kernel uses estimated 3D points and camera poses to directly modulate attention through point displacements and relative view transformations. Then, a learned gate fuses their outputs. Geometry thus enters attention as position rather than as fused reconstruction features. We further introduce Reference-Frame Composition (RC) supervision to train the model to update its reference frame across successive judgments. It provides stepwise turn and heading targets derived from verified single-step annotations. Among the compared methods, DK-VLM-8B achieves the highest average score across six spatial benchmarks and the highest overall scores on VSI-Bench (74.33) and ReVSI-32 (58.28). RC supervision further improves multi-step spatial reasoning, while the full model preserves general video understanding. Project page: https://sumuru789.github.io/DK-VLM-website/.
benchmark - arxiv:2610.22355 · cs.CLOmniVerifier-C1: Decomposed Graded Verifier for Rule-Governed Generative World ModelingBowei Liu, Xinchen Zhang, Xuhuan Li, Tianlin Zhang +9
Generative visual verifiers play a crucial role in advancing multimodal intelligence. To enhance the robustness of visual verifiers in complex multimodal scenarios involving compositional requirements, we move beyond conventional binary verifiers that produce only True/False discrete judgments and introduce a verifier that automatically decomposes complex prompts or environmental rules into fine-grained components and assigns pointwise scores to provide a continuous and more informative critique signal. Our experiments further demonstrate that these structured and interpretable outputs improve the verifier's generalization to out-of-domain multimodal critic tasks. To train this verifier, we introduce a Decoupled RL training strategy that jointly optimizes sub-question-level binary judgment and global alignment scoring, using precise local feedback to develop fine-grained verification capabilities that provide a foundation for reliable continuous scoring at the global level. Under a general reinforcement learning framework, we first empirically demonstrate that the proposed pointwise graded verifier significantly outperforms binary verifiers in traditional text-to-image scenarios. We then extend our investigation to rule-governed generative world modeling where generative models are leveraged for world simulation and world reasoning under explicit rules. From a theoretical standpoint, we further show that a pairwise verification paradigm yields more accurate critiques than the pointwise formulation when ground-truth images are available, with empirical results on diverse structured world-modeling tasks such as Maze and Sudoku further validating the effectiveness of our approach. Overall, our findings suggest a key insight: the transition from generative models to structured world-modeling agents critically hinges on the availability of accurate pairwise graded visual verifiers.
world model - arxiv:2610.22350 · cs.AIFDE-Gym: The Last Mile of AI DeliveryDongxu Yang, Hongda Yan, Renxin Wang, Baoli Zhang +22
Existing agent benchmarks typically evaluate agents on well-specified tasks defined by issue reports, specifications, or task tickets. Real-world customer delivery is different: forward-deployed engineers (FDEs) must clarify ambiguous requirements, work in live systems, and uncover unstated constraints. We introduce FDE-Gym, a benchmark of 95 end-to-end delivery tasks from 20 customer scenarios across 18 sectors, each adapted by industry experts from real-world delivery projects. In simulated customer environments with scripted stakeholders and modifiable customer systems, agents interact with stakeholders and operate on production-like systems to deliver working solutions. FDE-Gym scores a delivery by how much it improves the customer's own success metric, normalized between 0 (unchanged system) and 1 (reference solution), with penalties for stakeholder interactions (messages, trial runs, and waiting time). An online platform lets humans practice FDE skills under the same grading criteria. We evaluate 13 model configurations from seven providers. In the realistic setting, where unstated constraints must be discovered, the best configuration achieves a mean FDE-Gym score of 0.56 and resolves only 14 of the 95 tasks (a score of at least 0.8 with no incident). Discovering unstated constraints is the primary bottleneck: performance drops sharply when models must uncover them without explicit guidance. Yet information discovery alone does not account for the gap: even on the 23 tasks where the best configuration obtains every key fact from stakeholders, it reaches a normalized business uplift of 0.8 or above on only 10. These results reveal a substantial gap between solving well-specified tasks and delivering effective AI systems for real customers.
agentagent benchmarkbenchmark - arxiv:2610.22344 · cs.LGMLExpert: Towards a Generalist Agentic Model for Autonomous Machine Learning EngineeringShangheng Du, Jinxin Shi, Shuaiyu Zhang, Yiming Wu +6
Large language models (LLMs) are expanding from assisting scientific research and engineering to autonomously developing and improving AI systems, advancing AI for AI (AI4AI) as a pathway toward recursive self-improvement (RSI). Workflow-driven machine learning engineering (MLE) systems provide a practical path toward this pursuit through complex orchestration that manages long-horizon experimentation, while recent MLE systems adopt model-directed interaction, allowing models to autonomously organize experiments and adapt optimization strategies based on execution feedback. However, feedback from each experiment reflects several preceding decisions, and a long MLE run contains many such costly experiments. Training must therefore select useful supervision from long interaction trajectories and learn from intermediate outcomes. We propose a post-training method for autonomous long-horizon MLE optimization and apply it to train MLExpert, a 35B Mixture-of-Experts agentic model. We design the Kaggle-MLE Factory to construct executable tasks and an interaction harness for model-directed experimentation. Hierarchical agentic SFT transfers general tool-use and coding capabilities to MLE through outcome-guided supervision. Windowed Interaction RL (WIRL) treats the multi-turn decisions between successive experiment results as an execution window and assigns it a shared outcome-based reward. Asynchronous collection and policy correction support policy updates before the trajectory ends. On MLE-Bench Lite, MLExpert improves the average medal rate of the base model from 37.9% to 72.7%, achieving performance comparable to larger frontier models such as Kimi-K3 and GLM-5.3. Further experiments show that MLExpert improves performance on NatureBench Lite, agent scaffold optimization, and algorithm optimization, suggesting that the learned optimization capability generalizes to other tasks.
agentagentictool-usehierarchical agentself-improvementpost-training - arxiv:2610.22343 · cs.ROGeoState: Learning View-Decoupled Geometry States for 3D World ModelingHongchang Liu, Jian Liu, Yujuan Ding, Liang Peng +4
Recent 3D world models increasingly use features from geometry foundation models (GFMs) as predictive world states, benefiting from their rich geometric priors. However, the view dependence of these features entangles scene dynamics with camera-induced variation, making the resulting state evolution harder to predict. We propose GeoState, a view-decoupled geometry representation for predictive 3D world modeling that separates scene-state evolution from camera-dependent observation. Specifically, a scene tokenizer aggregates multi-view GFM features into a view-decoupled scene state, which a geometry decoder queries using camera rays. An autoregressive world model then predicts future scene states, while camera motion is modeled separately. The predicted states can therefore be decoded under either predicted or externally specified cameras, while retaining a single context-derived scale throughout the rollout without per-frame scale alignment. Experiments show that GeoState outperforms prior 3D world models on Waymo, supports geometrically consistent cross-view querying, and generalizes zero-shot to KITTI and nuScenes. Project page: https://hongchang-liu.github.io/GeoState-project-page/
world model - arxiv:2610.22342 · cs.ROSpark-0: Scaling Spatially Aligned Tactile Pretraining for Robot Dexterous ManipulationTianxing Chen, Wenwei Lin, Chuqiao Lyu, Peicheng Xiang +18
Generalist robot policies benefit from large-scale visual pretraining but often struggle with contact-rich dexterous manipulation, where force, compliance, and incipient slip are hard to observe from images. We present Spark-0, a spatially aligned vision-tactile-language-action (VTLA) model that uses touch in two complementary ways. During pretraining, image-space contact maps derived from tactile gloves, together with features from a frozen 3D geometry teacher, supervise the visual tokens; these objectives are removed at inference. In parallel, a tactile pathway encodes regional force histories and 3D taxel states to condition the action expert. A single checkpoint therefore supports both vision-only and tactile-conditioned deployment. To scale this supervision, we build HumanTouch, a 1,000-hour dataset of full-hand tactile human manipulation, and pretrain Spark-0 on it together with egocentric video, robot data, and simulation. We also release SparkArena, which pairs vision-only simulated tasks with tactile-conditioned real-robot tasks on one dexterous embodiment. Spark-0 achieves the highest mean success rate on both SparkArena tracks and ranks first on all eight aggregate metrics of the Ego Humanoid Manipulation Benchmark. Compared with vision-only pretraining on the same data and compute with touch discarded, retaining the touch of 1,000 hours of HumanTouch raises vision-only success on Ego Humanoid from 52.5% to 71.7%, and success on SparkArena-Real, where touch is also used at deployment, from 44.0% to 52.0%, indicating that touch serves not only as a policy input but also as scalable supervision for dexterous visuomotor pretraining. The Website is available at: https://xsparkai.com/Spark-0.
manipulationdexteroushumanoidtactilebenchmark - arxiv:2610.22341 · cs.CV3DCodeVerse: Exploring 3D Worlds as CodeYipeng Gao, Ziyao Zeng, Jinfa Huang, Yijiang Li +15
We introduce 3DCodeVerse, a research ecosystem that enables coding agents to create editable and verifiable 3D content at scale. 3DCodeVerse comprises large-scale, high-quality datasets of executable 3D code, a coding harness for spatial reasoning and automated verification, and practical guidance for 3D coding. First, we construct 3DCodeVerse-2M, a large-scale curated dataset of standalone 3D executable code. We build this corpus by collecting open-source data, converting specialized representations, and generating new programs with coding agents. Each program is packaged with the source files and associated assets required for execution in its declared environment, along with available descriptions and verification records. Second, we develop the 3DCodeVerse Harness, a multi-agent directed acyclic graph framework for generating and refining 3D content. It uses execution checks, spatial measurements, and visual feedback to automatically evaluate and revise generated programs without human feedback during generation. Third, we conduct systematic empirical studies of 3D coding, establishing baselines with fine-tuned open models and deriving practical guidance through component-level analyses of our harness. Extensive experiments demonstrate that fine-tuning on our dataset significantly improves the 3D coding capabilities of open-source models. Notably, our 3DCodeVerse harness achieves a 90% win rate against Claude Design (3D Object) in pairwise human evaluations, when both are utilized by Claude Opus 5.5.
multi-agent - arxiv:2610.22340 · cs.RODexterity from Supervised PlayBotao He, Yiming Dou, Yiannis Aloimonos, Jitendra Malik +2
How can robots autonomously collect the contact-rich experience needed to learn dexterous manipulation? We present DSP: Dexterity from supervised play, a framework in which a robot learns a play policy from minutes of human play through behavior cloning, then deploys it autonomously to collect hours of interaction data to train a world model. During inference, the same play policy then guides model-predictive control (MPC) by proposing candidate action sequences for evaluation under the learned world model. To accelerate inference, we also introduce a method for distilling a closed-loop, goal-conditioned policy through self-play within the learned world model. We evaluate this framework on underactuated pushing and any-to-any in-hand pose reorientation with a multifingered hand, in simulation and real world. In real world experiment, our method outperforms hindsight behavior cloning by 69.6% in success rate and achieves comparable success to reinforcement learning with 490 times less interaction data. These results demonstrate that supervised play can autonomously collect interaction data, guide model-based planning, and enable policy learning without any task-specific demonstrations or sim-to-real transfer.
manipulationdexteroussim-to-realworld modelself-play - arxiv:2610.22339 · cs.LGLean Graph Search for Formal Theorem ProvingYihan Xu, Jia-Chen Gu, Junyi Zhang, Di Wu +6
Retrieving relevant lemmas is crucial for formal theorem proving, enabling provers to build machine-checkable proofs from established mathematical results. Recent retrieval systems match library declarations to a target theorem or its intermediate proof steps, but make limited use of multi-hop dependencies among declarations. These dependencies can reveal supporting lemmas that are not necessarily lexically similar to the target theorem but recur in proofs of related problems. To this end, we propose LeanGraphSearch, a graph-based retrieval method for Lean 4 that constructs a graph from Mathlib's forward dependencies and traverses it to uncover supporting lemmas through multi-hop dependency chains. Specifically, LeanGraphSearch links each declaration to the premises it depends on, uses a graph-based ranking algorithm to propagate relevance from semantically matched declarations to their dependencies, and reranks the expanded candidates against the original query. Across two mathematical retrieval benchmarks, LeanGraphSearch improves recall by up to 12% over the strongest baseline, LeanSearch v2, on MathlibQR and increases coverage of complete annotated proof routes on MathlibMPR, covering up to 3.7% of extra groups. More importantly, we evaluate whether these retrieval gains translate into improved end-to-end theorem proving by integrating LeanGraphSearch into an iterative proving loop guided by Lean feedback. Across three theorem proving benchmarks, LeanGraphSearch improves pass rate at {8, 16, 32} reflection turns by as much as 40% compared to vanilla prover and at most 10% compared to LeanSearch v2, showing higher accuracy-cost efficiency. These findings highlight the value of dependency graph search for theorem proving and motivate further research on structure-aware search for proof discovery.
benchmark - arxiv:2610.22337 · cs.ROSim-to-Real Reinforcement Learning for Path-Safe Target Reaching in Robotic Harvesting of Hydroponic LettuceAl Bashir, Partho Ghose, Yibin Wang, Robert G. Hardin +3
Robotic harvesting requires precise manipulation to ensure accurate positioning of the end-effector, especially in a constraint workspace. This study introduces the Path-Safe Approach Proximal Policy Optimization (PSA-PPO), a target-conditioned local reaching framework designed for robotic harvesting of nutrient film technique (NFT)-hydroponic lettuce in greenhouses. The proposed PSA-PPO preserves the standard clipped PPO objective while introducing a harvesting-oriented policy formulation that integrates target-conditioned observations, corridor-aware spatial constraints, trajectory-history-based safety encoding, and task-specific success criteria for constrained robotic manipulation. PSAPPO was trained entirely in Isaac Lab and successfully transferred to a physical UR10e robotic arm without any real-world policy fine-tuning. In a controlled physical benchmark against 150 scheduled targets, PSA-PPO achieved the highest reaching rate of 94.7%, outperforming Cartesian+TOTG (86.0%), RRTConnect (88.0%), and RRT* (80.0%), while reducing median RMS jerk by 39.8 to 73.0% relative to the classical planners during the final approach segment. A formulation ablation evaluated over 200 simulation targets demonstrates that path-safety and side-corridor shaping are the principal drivers of harvesting-compatible performance: removing either term substantially degrades strict success even when terminal endpoint accuracy is preserved or improved, confirming that optimization toward the endpoint alone is an insufficient objective for constrained crop-proximal manipulation. Finally, PSA-PPO was integrated with an RGB-D vision module and an end-effector and evaluated as a complete greenhouse harvesting system across 144 lettuce harvesting attempts, achieving 94.4% vision-guided target reaching and 86.1% complete harvesting success.
manipulationsim-to-realbenchmark - arxiv:2610.22324 · cs.CLTraining Reasoning Models to Generate Diverse Scientific IdeasManya Wadhwa, Pranav Narayanan Venkit, Hiroaki Hayashi, Yu Li +2
One of the first steps of a research pipeline is generating ideas. Large Language Models (LLMs) are increasingly used to automate this step, but most of this work optimizes the quality of individual ideas, while overlooking diversity. In our work, we study document-grounded ideation, focusing on increasing the diversity of LLM generated ideas and ensuring they are high quality. Given prior papers, a model must propose a set of ideas that are each high-quality and clearly distinct from one another, while also being grounded in the input documents. To optimize for this, we propose training with a set-level reinforcement learning objective. In our setup, in a single generation, the model writes K structured idea cards grounded in the source papers, and the set is scored jointly, where each idea's quality is weighted by its distance from the ideas written before it. This teaches the model to steer each new idea away from its predecessors. We show that our trained 8B model achieves higher creative utility (a measure combining quality and diversity) than Claude Opus 4.8 and GPT-6 Astra on our contamination-free benchmark. In a pilot human study with four experts, our sets are judged as promising as Claude Opus 4.8's and contain marginally more unique ideas. When a coding agent runs proof-of-concept experiments on the ideas, our model's success@k is comparable to Claude Opus 4.8's, so we see that the added diversity does not cost any executability.
agentbenchmark - arxiv:2610.22323 · cs.ROMem4Act: Learning Effective Latent Memory for Long-Horizon Robotic ManipulationYun Zhang, Zhiyu Huang, Tianhui Cai, Chen Tang +2
Long-horizon robotic manipulation fundamentally requires memory: agents must retain task-relevant information from past observations and interactions and use it to guide future actions. Existing approaches typically preserve raw historical observations, compressed visual tokens, or language summaries, but these representations either expose the action policy to increasingly long histories or risk discarding task-relevant temporal, semantic, and perceptual information. We introduce Mem4Act, a framework for learning effective latent memory as a fixed-size representation explicitly optimized for long-horizon robotic control. Mem4Act addresses three coupled challenges: how to efficiently cover long histories, what information memory should retain, and how to ensure the policy actually relies on memory. To this end, it organizes history via multi-granularity temporal aggregation, preserves complementary semantic and perceptual information through question-answering supervision and key-frame DINO feature reconstruction, and employs memory-first two-stage training to encourage greater use of memory by the action policy. The resulting memory captures both high-level task state and fine-grained perceptual evidence while providing a bounded interface to the action policy. Across RoboMME, RMBench, and four real-world manipulation tasks, Mem4Act achieves state-of-the-art performance on memory-dependent control, with success rates of 49.88%, 61.0%, and 81.5%, respectively. Extensive ablations further demonstrate the importance of temporal aggregation, complementary memory supervision, and memory-first training.
manipulationmemory - arxiv:2610.22322 · cs.AILearning Self-Orchestration in Decentralized Agent SwarmsGuowei Xu, Zhenting Qi, Amani Kiruga, Himabindu Lakkaraju +1
The capabilities of an agent swarm depend on both its individual members and how they coordinate. Yet existing multi-agent systems often train agents within fixed orchestration structures, separating the learning of individual capabilities from the organization of collective behavior. We introduce market-based post-training (MPT), a framework for jointly learning action policies and self-orchestration in decentralized agent swarms. Agents express their willingness to participate through bids, which determine who acts and how collective outcomes provide local training credit. Learning both action and bidding behavior allows individual capabilities and patterns of participation to adapt together without a central coordinating agent. We evaluate MPT on embodied crafting, robot navigation, and deep research. Results show gains over monolithic post-training and confirm that agent participation indeed evolves with learning. With the same post-trained action policies, learned self-orchestration outperforms alternative coordination strategies. Ablations show that both action and bidding benefit from learning, and that learned bids provide more effective action-training signals. Beyond the original training objectives, trained swarms adapt faster to related tasks and achieve better performance with larger populations at inference time. Larger populations also suffer smaller performance losses under agent dropout, demonstrating that learned self-orchestration supports reuse, scaling, and robustness in decentralized agent swarms. Code and trained models are available at https://github.com/Embodied-Minds-Lab/MPT.
embodiedagentmulti-agentagent systempost-training - arxiv:2610.22321 · cs.CVAVC-Reward: Event-Centric Audio-Video Consistency EvaluationYuruo Wu, Kaihang Pan, Kai Shen, Jiacheng Li +5
As joint audio-video generation models become more capable, they can generate audio-visual scenes exhibiting three forms of complexity: multi-event composition, multi-layer audio mixture, and asymmetric cross-modal relations. However, current evaluations largely reuse metrics from earlier simple and unidirectional tasks. They typically evaluate an audio-video clip as a whole and rely on simple predefined relations, making them unsuitable for these increasingly complex scenes. We propose an event-centric evaluation framework that disentangles audio and visual events at the event level, analyzes each event independently, cross-validates relations across modalities, and aggregates the judgments into a coarse-to-fine, three-level consistency metric. The metric assesses global narrative harmony, event-level temporal synchronization, and fine-grained attribute alignment, providing a general formulation for audio-video consistency evaluation across diverse generation settings and cross-modal relations. We instantiate the metric as AVC-Reward, a trainable audio-video reward model. To evaluate the agreement between automatic judgments and human perception across generation paradigms and relation complexities, we construct AVC-Benchmark, which is primarily built from diverse positive and negative samples generated by various types of audio-video generation models. AVC-Reward achieves state-of-the-art alignment with human judgment among automatic methods, provides interpretable rationales that localize failures, and improves the audio-video consistency of a joint generation model when used as an online reinforcement-learning signal. Project page: https://tsuki0512.github.io/AVC-Reward.github.io/
benchmarkevaluation framework - arxiv:2610.22320 · cs.CVNot Just a Subject: Capturing Adaptive Scene-Specific Variance for Visual StorytellingSeungJu Cha, Ye-Chan Kim, Kwanyoung Lee, Dong-Jin Kim
Recent advancements in diffusion models have significantly improved text-to-image generation, enabling their application to visual storytelling, where coherent image sequences are generated from narrative prompts. However, existing methods primarily focus on identity preservation and often fail to adapt to diverse scene variations, resulting in semantic misalignment and structural inflexibility. We propose EOS-Story, a training-free framework that harmonizes identity consistency with scene-specific adaptation across text and vision spaces. In the text space, we decouple the narrative into identity and scene sub-prompts to prevent information dilution, but this leaves the scene without access to appearance cues. To bridge this gap, we introduce Cross-Contextual Bridging, which softly injects a phrase-level appearance residual into the scene's <EOS> embedding. In the vision space, we find that scale mismatch between identity and scene subjects causes a few identity tokens to be repeatedly transferred, producing structural artifacts. We address this with Adaptive Identity Preservation, which anchors identity features at the scene-specific scale (Identity Anchoring) and integrates them via Identity-Aware Self-Attention. To evaluate performance under varied spatial compositions, we extend the Consistory+ benchmark with diverse framing scenarios (Consistory+_+). Extensive experiments show that EOS-Story achieves competitive or superior performance on both benchmarks.
benchmark - arxiv:2610.22319 · cs.AIA team of Gemini 4 Argon agents beats MinecraftMaxence Faldor, Satinder Singh, Roberta Raileanu
We placed four Gemini 4 Argon agents in a sandboxed Minecraft survival world and gave them one goal: defeat the Ender Dragon. Each agent controlled a player via an open-source library called Mineflayer and coordinated with its teammates through in-game chat and shared memory. No one assigned roles or wrote game logic. The agents planned, wrote, and debugged every program themselves. Researchers could watch the run on a live dashboard and pause or stop it at any time, but they had no cause to intervene. Starting from a randomly generated world on Easy difficulty, the team wrote 2,783 programs and defeated the dragon in 27 hours and 14 minutes, dying 605 times along the way. To our knowledge, this is the first end-to-end autonomous completion of Minecraft by AI agents, achieved using only the standard Mineflayer library without pre-scouted seeds or other game-specific tools written by humans. Winning required the agents to keep a coherent plan across thousands of dependent steps for more than a day. Complex real-world software engineering and enterprise tasks demand the same long-horizon coding and multi-agent coordination. We describe the setup, how the run unfolded, the team's strategies (dividing the work, sharing what they learned, recovering each other's items), and the limits of the result.
agentai agentmulti-agent - arxiv:2610.22316 · cs.LGLearning to Steer LLM Agents and Multi-Agent SystemsLeyang Shen, Yaxin Gong, Yang Zhang, Chenxiao Fan +6
Large language model (LLM)-based agents and multi-agent systems increasingly tackle complex tasks through long-horizon tool use and inter-agent collaboration. However, these systems are prone to deviating from the overall goal as execution unfolds, which wastes execution resources and leads to undesired outcomes, a failure termed goal drift. In this work, we propose a steering agent that supervises execution on the fly and steers the system back on track when drift occurs. The steering agent observes the execution context, deciding when to intervene with a low-cost fast path and how to intervene with a high-cost slow path. To unlock its steering intelligence, the ability to assess goal drift and generate effective interventions, we jointly train the two paths with reinforcement learning, combining a pairwise ranking objective with group-relative policy optimization (GRPO). Experiments on information-seeking search tasks show that the steering agent provides effective supervision for both single-agent and multi-agent systems, improving their performance and efficiency. The learned steering intelligence generalizes to out-of-domain tasks, transfers to unseen executors, and supports controllable intervention frequency without retraining.
agentllm agentmulti-agentagent systemtool use - arxiv:2610.22314 · cs.CLMindEval2: Clinically Grounded Patients and Rubrics for Evaluating Language Models in Mental Health SupportJohn Mendonça, Daniel Tschernutter, José Pombal, Areti Vassilopoulos +5
With people increasingly turning to AI chatbots for mental health support, the need to evaluate these tools is becoming more pressing. Recent benchmarks attempt to mimic a realistic setting by simulating multi-turn interactions with synthetic patients and evaluating them with an LLM-as-a-judge. However, the usefulness of these methods is limited because simulated patients can be overly forthcoming and self-reflective, and judgments typically involve opaque Likert scales with unclear practical utility. To bridge these gaps, we introduce MindEval2, an evaluation framework that consists of two components: Mental Health Quality Rubrics (MHQR), for evaluation, and MindSim, for patient simulation. For MHQR, we collaborate with PhD-level psychologists to develop a set of rubrics that break down response quality into explicit operational criteria that are more interpretable and actionable than existing evaluation taxonomies. MindSim, in turn, is an LLM harness that combines carefully-designed patient profiles with tools, enabling more varied, challenging, and realistic interactions than existing baselines. The harness also contains a memory component that is updated between each interaction, enabling MindEval2 to model the longitudinal aspect of mental health. We evaluate 18 proprietary and open-weight models on MindEval2, finding that while frontier models have largely improved on issues related to surface level quality and boundaries, guidance calibration remains the dominant failure mode across all models, accounting for most of the quality gap.
memorybenchmarkevaluation framework - arxiv:2610.22312 · cs.CLGrid-Bench: Can LLMs Predict How Discrete Worlds Change?Zixiang Xu, Zirui Song, Wanrong Zheng, Chiyu Zhang +15
Spatial instructions describe how to move objects, transform regions, or modify their surroundings. When several operations are applied in sequence, each acts on a configuration shaped by earlier changes. How accurately can large language models (LLMs) predict the resulting spatial state? We introduce Grid-Bench, an executable benchmark in which eight families of operations transform discrete worlds represented as two-, three-, and four-dimensional grids. Given an initial grid, explicit rules, and an event history expressed as text, a model predicts the complete final grid. Reference execution makes every predicted value checkable, supporting comparisons across spatial operations, history lengths, and workloads. Evaluating six LLMs shows that complete-state accuracy falls sharply as histories grow, with substantial divergence among models that perform similarly on single events. Models with similar exact-match accuracy can nevertheless differ substantially in the number of incorrect cells on shared instances. Even over long histories, some models recover many correct changes beyond what copying the initial state would yield. Grid-Bench makes both complete recovery and these differences in returned states directly measurable across a range of spatial conditions.
benchmark - arxiv:2610.22311 · cs.CVVelox: Spatial Memory Makes World Models EfficientQiuhong Shen, Shizun Wang, Xinchao Wang
Video world models turn actions and camera motion into long visual rollouts, making video a natural interface for interactive simulation. Such interaction requires responsive generation, yet coherent rollouts must retain visual memory of what has already been generated. Chunked video diffusion models usually keep this context in growing KV caches, so later chunks become progressively slower. To address this bottleneck, we introduce Velox, a training-free attention router for region-level context selection using frustum slab correspondences from streaming geometry. Velox represents historical blocks as camera frustum slabs with depth intervals and obtains current-block depth through recency-priority multiwarp. For well-covered regions, frustum slab overlap softly modulates each layer's model-derived block importance. Uncovered regions retain the model-derived route without geometric bias. The fixed token-block granularity matches FlashAttention-style tiled execution and reduces the attention term that grows with rollout length. Frustum slab construction runs asynchronously with denoising on the same H200. At chunk 48, Velox accelerates attention by $4.36\times$ and denoising by $2.43\times$. On the full 158-case WBench navigation split, it remains within 0.15 API-free average points of dense LingBot-Fast-Cam. Please see our project page at https://velox-world.com/velox/ for more details.
world modelmemory - arxiv:2610.22306 · cs.CVARC-SORT: Adaptive, Robust, Compensated Multi-Object Tracking with Interacting Multiple ModelsGirish Kumar Adari
Multi-object tracking in sports broadcasts is hard for two reasons: athletes move fast and change direction abruptly, and the camera pans and zooms at the same time. We present ARC-SORT (Adaptive, Robust, Compensated SORT), a tracking-by-detection tracker that trains nothing and runs in real time. Its design rule: the tracker measures its own uncertainty, and that measurement decides how it acts. An Interacting Multiple Model (IMM) Kalman filter gives every track a predictive covariance, and we turn it into a per-track search region: the less reliable the motion prediction, the larger the region in which a detection may match (uncertainty-adaptive Expansion IoU, EIoU). On identical inputs, this continuous coupling is what matters: the same covariance used as a hard Mahalanobis gate hurts at every threshold we swept, and richer covariance summaries add nothing measurable. When two tracks cross, the assignment cost itself reports whether the decision was close. ARC-SORT keeps such pairs under review and applies at most one corrective identity swap once post-crossing appearance evidence shows the first decision was wrong; this adds +1.0 HOTA (higher order tracking accuracy) on both the SportsMOT and DanceTrack test sets at under a millisecond per frame. For players who leave the frame for seconds, motion extrapolation carries no information, so our offline re-identification drops the usual motion gate and links tracklets by appearance alone. ARC-SORT reaches 81.82 HOTA on the SportsMOT test set with the fewest identity switches of the 69 leaderboard entries as of August 2026 and tied-third multi-object tracking accuracy (MOTA, 97.09), and 62.1 HOTA on DanceTrack test with the same association settings. On Deep-EIoU's public detections and embeddings it reaches 80.31 HOTA, +3.1 over Deep-EIoU. It runs at about 50 frames per second on a laptop RTX 3060 GPU. Code will be released.
leaderboard - arxiv:2610.22298 · cs.AIAI Consilium: a Benchmark of LLM Clinical Multidisciplinary Second OpinionsAlexey A. Bokov, Ivan I. Panin, Nikolay B. Nechaev, Mikhail Yu. Valkov +9
Large language models (LLMs) can generate differential diagnoses for complex cases, but a useful second opinion must also be factually grounded, clinically coherent, actionable, and safe. We developed a multidimensional benchmark using 300 diagnostically resolved NEJM--MedQA cases: 110 for physician reference ratings and Judge calibration and 190 for held-out evaluation. Eight single-model pipelines generated independent role-specific opinions and interactive-consilium reports under free-form and structured clinical-reasoning conditions. A physician-calibrated LLM Judge evaluated 3,040 final reports and the corresponding structured primary-role outputs. Model identity separated report quality more strongly than response format. Structured final reports had higher overall mean scores in seven of eight model pipelines, with their clearest advantage in management (2.740 vs. 2.436). Compared with the structured pre-consilium primary-role opinion, the complete structured synthesis had higher pooled scores across all five shared criteria, with the largest increases in problem prioritization, factual grounding, and reasoning coherence. GPT-5.6-Sol led every evaluated component and reached consensus in the fewest review rounds. These findings show that the benchmark can distinguish model capability, the contribution of multidisciplinary synthesis, criterion-specific effects of response structure, deliberation dynamics, and inference requirements within one common-case evaluation.
benchmark - arxiv:2610.22297 · cs.AIPrivate Retrieval Augmented Generation via Random ProjectionDixi Yao, Tahseen Rabbani, Tian Li
Retrieval-Augmented Generation (RAG) enhances the capabilities of large language models (LLMs) by querying external structured knowledge. However, it can also introduce privacy risks by leaking sensitive information from the retrieval datastore. We propose a simple method to preserve datastore privacy in RAG systems via random projection. By applying the same projection to both datastore embeddings and query embeddings, our method provably preserves semantic similarity between queries and retrieved items while substantially mitigating data extraction attacks. Across multiple RAG architectures and datasets, we show that this lightweight approach achieves superior retrieval and generation performance compared to prior methods which leverage formal differential privacy (DP) guarantees, while exhibiting comparable empirical privacy under strong attack models. Our results for the first time suggest that random projection can serve as a competitive and practical baseline for privacy-preserving RAG systems.
retrieval-augmentedretrieval augmentedrag - arxiv:2610.22294 · cs.CLMulti-Target Few-Shot Prompting with Diffusion Language ModelsAnton Xue, Alan Li, Rahul Saha
Few-shot prompting is useful when instructions and rewards are hard to specify. One such application is editing by example, where a demonstrated formatting change or API migration may need to be applied across multiple sites. However, such edits often depend on both left and right context, making generation with causal autoregressive models awkward. We therefore introduce multi-target few-shot prompting with diffusion language models, where examples guide the infilling of multiple masked regions within a shared canvas. To address the known challenge of reliably following demonstrations in few-shot prompting, we further leverage selective masking of the canvas to derive a likelihood-ratio reward that favors example-supported infills. Across few-shot prompting benchmarks and software refactoring case studies, we find that demonstrations improve infilling accuracy, with example guidance yielding further gains in several settings.
benchmark - arxiv:2610.22292 · cs.LGCAT-alyzing Honesty: Continuous Adversarial Training for Transferable LLM HonestyYuan-Chen Chang, Niklas Herbster, Gauthier Gidel, Tommaso Tosato
Large language models (LLMs) are becoming prevalent in scenarios where honesty is critical. However, despite alignment training, these models still exhibit dishonest behaviour, such as outputting claims that contradict some knowledge or are not supported by what is accessible in context. We introduce Continuous Adversarial Training for Consistent Honesty (CATCH), inspired by a method originally developed to robustly refuse harmful queries. CATCH trains the model towards an honest answer and away from a dishonest one while maintaining the model's capabilities. We evaluate two architectures (Llama-3.3-70B-Instruct and Qwen3-3.6-27B). We show that trained with only hundreds of targets built from self-reporting scenarios, CATCH learns an honest behaviour that robustly generalizes and transfers across distinct benchmarks. It consistently reduces dishonesty under pressure on MASK, increases confessions in the multi-agent game of Among Us, and improves the discovery rates of the model's hidden behaviours in AuditBench. We show that CATCH outperforms prompting and supervised fine-tuning in improving honesty, and it is on par with state-of-the-art honesty activation steering. However, CATCH surpasses honesty steering in maintaining utility, thanks to a novel choice of utility data.
multi-agentbenchmark - arxiv:2610.22290 · cs.AIEliteReplay: Preserving Diverse Solutions in RL Post-Training via Quality-Diversity SearchHanane Nour Moussa, Doeun Lee, Sachin Kumar
Reinforcement learning has become a common post-training technique to improve the accuracy of LLMs on verifiable tasks. However, such training can narrow the set of solutions the model can produce. Several approaches have attempted to address this behavior-sharpening effect, but they often optimize metrics that do not reflect the diversity of correct solutions. Ideally, we want a model that can produce multiple good solutions, not only a single reliable one. We propose EliteReplay, a post-training recipe that pairs GRPO with an "elite" archive built by quality-diversity search. Before training, we use the frozen base model to collect solutions that span distinct approaches; during training, GRPO is combined with a replay loss on this archive. The result is a post-trained model that improves accuracy while preserving more of the base model's solution diversity. On multiple text-to-SQL tasks, EliteReplay achieves a higher pass@128 than all baselines, producing up to 77% more distinct correct strategies than the strongest baselines. It preserves most of the base model's correct strategies, outperforming existing methods in strategy retention. This diversity translates into practical benefits on ambiguous questions, where EliteReplay covers both interpretations nearly twice as often as the strongest baseline on AmbiQT.
post-training - arxiv:2610.22289 · cs.AIImproving Environment-Agnostic Coding Agents: The Role of ScaffoldingSergey Rodionov
Stronger language models may increasingly construct useful procedures for themselves, reducing the advantage of developer-supplied scaffolding. We investigate why this advantage persists and whether an environment-agnostic agent can be improved without environment-specific engineering. We evaluate a Codex-based agent in two nearly identical prompt variants on ARC-AGI-3, AGI Maze, and Terminal-Bench. Each variant is deployed with unchanged code, prompts, and general-purpose tools across environments, receiving environment-specific information as input. Changing only an instruction to encourage reporting difficulties and returning control for automatic continuation improves mean scores on both ARC-AGI-3 and AGI Maze, while Terminal-Bench shows small, mixed changes. Yet the general agent remains below an ARC-AGI-3-specific reference, including with GPT-5.6-sol. The reference prescribes organization and control rather than supplying new game rules or solutions. To investigate this gap, we transfer reset decisions to a scripted controller that follows a convention already given to the agent. This small change recovers part of the specialized agent's advantage. A complementary AGI Maze experiment shows that additional tutorial experience can reduce performance. Together, these findings highlight the distinction between having information and tools available and using them effectively. Supplied procedures still matter for the models tested here. The performance gains also come with higher token-based cost. The experiments do not separate better use of information and procedures from additional or differently allocated inference. We provide the EnvTraverse framework and an evaluated baseline agent for testing whether further environment-independent changes improve performance in some environments without degrading it in others.
agent - arxiv:2610.22287 · cs.LGLearning Rational Strategies for Temporal Objectives in Multi-Agent SystemsOmar Adalat, Edwin Hamel-De le Court, Francesco Belardinelli
We study rational synthesis for stochastic multi-agent systems with unknown dynamics, where each agent is equipped with its own probabilistic temporal objective specified using Probabilistic Linear Temporal Logic on Finite Traces ($\mathsf{PLTL}_f}$). We impose solution concepts (e.g., Nash equilibrium, subgame-perfect equilibrium) as constraints on the joint strategy profile, using payoffs derived from each agent's probability of satisfying its temporal objective. In particular, we formulate per-agent rational synthesis as finding a strategy profile -- if one exists -- such that (i) the designated agent's probabilistic temporal constraint is satisfied, and (ii) the profile represents a stable solution under a chosen solution concept. We give conditional convergence guarantees for tabular Nash-Q learning, certify satisfaction with simultaneous finite-sample confidence bounds, and assess strategic stability with approximate best-response checking. Our approach operates in a stochastic \emph{Multi-Agent Reinforcement Learning} (MARL) environment, where model-free learning algorithms attempt to attain extremal satisfaction probabilities for temporal objectives automatically compiled into reward automata. Our approach is empirically validated using 8 general-sum concurrent stochastic games.
agentmulti-agentagent system - arxiv:2610.22280 · cs.AILEASH: Uncertainty-Aware Weak-to-Strong Escalation and Oversight for Computer-Use AgentsQingni Wang, Changdae Oh, Yuzhe Yang, Sharon Li +3
Computer-use agents increasingly tackle long-horizon tasks, where running a frontier model at every step is expensive and a small open-weight model alone often fails. Weak-to-strong escalation offers a middle ground where a weak agent acts by default and hands control to a strong one only when needed. The core challenge is deciding when to escalate, which requires an uncertainty signal cheap enough to check at every step. However, even after handoff, the strong agent can still get lost, but a black-box API that the strong agent is usually served exposes no internals to monitor. To this end, we propose LEASH, a training-free framework that unleashes the strong agent only when the weak agent's latent uncertainty signals it is needed, and keeps it on a leash afterward. LEASH has three layers: (1) initial task uncertainty for early routing, (2) step-level uncertainty for runtime escalation, and, after handoff, (3) a shadow observer for continued oversight of the strong black-box agent. Here, we use latent uncertainty as a fuel for all three layers. Specifically, we compare the weak model's hidden states with those from its past successful and failed trajectories, and a state that looks more like the failures signals higher uncertainty. During weak-agent execution, our uncertainty estimator reuses activations already produced during action generation, so it needs no extra forward pass or dedicated monitor training. Experiments on OSWorld across four contemporary model backbones show that LEASH consistently improves over standalone weak agents and outperforms random escalation by 6.55--11.94 points at matched escalation rates. With Qwen3.5-35B-A3B, LEASH achieves 73.66% task success at $0.4117 per task, matching always-on Claude Opus 5 (72.90%) while reducing inference cost by 19.2%.
agent - arxiv:2610.22277 · cs.ROAuditable and Deployable Multi-Agent Motion Planning via Posetal Games from Legal GuidelinesFlorian Lercher, Karan Mahesh, Yujun Huang, Meshal Alharbi +5
When robots are deployed in the real-world, they must interact with other agents and accomplish their own goals while adhering to legal constraints. These potentially conflicting objectives are often only partially prioritized, making scalarized objectives or fully specified hierarchies unsuitable. We present a framework for multi-agent motion planning under legal requirements based on games with partially ordered preferences (posetal games). Scenario-independent legal guidelines order abstract legal concerns, from which we automatically derive agent preferences and a legal game for any concrete scenario. Agents may refine their legal preferences with private objectives, and regulators may refine the guideline as laws evolve. We prove that such customizations can only remove equilibria and thus preserve legality. To coordinate agents without explicitly communicating private preferences, we introduce belief play, which provably recovers the true equilibria from conservative local games, and a distributed iterative best-response protocol requiring less information. We apply the framework to maritime navigation, systematically lifting pairwise traffic rules to multi-vessel encounters. On seven encounters with up to four vessels, every rule violation of the committed equilibria is certified as forced by the legal hierarchy or unsatisfiable, and defectors are identified from public information in all tested cases. These results indicate that guideline-based posetal games offer an auditable route from partially specified laws to multi-agent motion planning.
agentmulti-agent - arxiv:2610.22276 · cs.CVMJev-Spatial: System One Spatial Intelligence via Native Attention MatchingLuozheng Qin, Yuhan Zhu, Changhai Zhou, Qian Qiao +8
Spatial intelligence lets an agent perceive and act in the physical world, and many of its judgments must be made at interaction rates: which spatial relation holds, how far an object is, or where to place it. Vision-language models (VLMs) produce these answers by free-form text generation, paying for autoregressive decoding and answer parsing even though the output space is compact: one of a few options, a scalar, or a 2D point. In light of this, we introduce MJev-Spatial, a System 1 multimodal decision model that tackles all three types of spatial understanding tasks within a unified classification framework. Specifically, MJev-Spatial treats each candidate as a class and scores it using the backbone's pretrained query and key projections, without adding parameters. The candidates are the given options for choice tasks, or value bins and image regions refined through coarse-to-fine classification for numerical and point prediction, respectively. At inference, each classification decision takes a single forward pass and generates no answer tokens. To train on heterogeneous spatial data efficiently, we supervise all coarse-to-fine stages jointly with teacher forcing in one forward pass, and combine this with length-aware global batching and cross-GPU load balancing, raising training throughput by 30.8% over the original batching strategy. Built on Qwen3.5-4B and trained on 12.3M question-answer pairs converted from existing spatial datasets, MJev-Spatial achieves the highest overall score (65.64%) among eight tested models, ranks first on seven of twelve benchmarks, and outperforms its unmodified autoregressive base model on all twelve, while running 2.5 times faster than the fastest autoregressive baseline and 5.6 times faster than the base model.
agentbenchmark - arxiv:2610.22273 · cs.LGLibrarian Agents: Training LLMs to Curate Long-Context Workspaces that Anticipate Future QueriesAnikait Singh, Teresa Zhang, Yoonho Lee, Roshen Sanjay Nair +2
Reasoning over long document contexts remains a challenge for large language models (LLMs). Standard retrieval methods are inherently reactive, bottlenecking agents with complex, cross-document reasoning at inference time, while precomputed structures like knowledge graphs often fall short because they rely on rigid heuristics. Instead, can we allow agents to perform anticipatory directed curation: learning to proactively reorganize unstructured corpora prior to inference to efficiently serve diverse future queries. We build on a key asymmetry in modern agentic harnesses (e.g., Claude Code): while LLM agents are adept at navigating existing filesystems to solve an immediate task, they are rarely trained to construct reusable workspaces for future use. To instantiate this capability, we introduce Librarian Agents, a framework in which a builder agent transforms raw documents into a hierarchical filesystem of synthetic files and reference links, which forms a workspace tailored for a downstream answering agent. We formulate workspace construction as a Markov decision process and optimize the builder policy via reinforcement learning to directly maximize downstream answering accuracy. Empirically, Librarian Agents outperforms the strongest retrieval and precomputation baselines by 33 percentage points on BrowseComp-Plus, amortizes upfront construction costs across repeated queries, and generalizes zero-shot to unseen domains on BRIGHT with a 31-point average accuracy improvement.
long-contextknowledge graphagentllm agentagentic - arxiv:2610.22269 · cs.AIX-RIC: An Agentic RAN-Intelligent Controller Benchmark for Enhancing Reinforcement LearningAdhwaa Al-Chaab, Ayman Younis, Dario Pompili
The O-RAN ALLIANCE introduces architectural flexibility through disaggregation and AI-driven control. While the RAN Intelligent Controller (RIC) enables centralized optimization, current AI approaches lack reasoning capabilities and explainability. We propose eXplainable RIC (X-RIC), a novel framework integrating agentic AI into O-RAN architectures. X-RIC combines reactive Reinforcement Learning (RL) with strategic agentic reasoning through: (i) automated Behavioral Cloning (BC) using domain heuristics for sample-efficient training, (ii) a Telecom Language Model (TLM), a transformer-based reasoning module providing contextual abstraction and interpretability, and (iii) closed-loop co-adaptation within the Near-RT RIC control loop enabling continuous policy refinement. To ensure deployment readiness, we introduce BARX (Benchmarking Agentic Reasoning for RAN), a domain-specific evaluation framework assessing agentic control across latency, compliance, and determinism metrics. We validate X-RIC on SLA-aware network slicing in simulation and O-RAN-compliant testbed deployments. X-RIC achieves 89% operational cost reduction while maintaining SLA compliance, converging 75% faster with 50% fewer training steps and 46% lower variance.
agenticbenchmarkevaluation framework - arxiv:2610.22267 · cs.CVBeyond Pairwise Repulsion: Collective Diversity via Determinant Maximization in Diffusion SamplingKwanyoung Lee, SeungJu Cha, Yebin Ahn, Sungho Koh +1
Strong guidance in text-to-image (T2I) diffusion models concentrates sampling toward high-density modes of the conditional distribution, collapsing independently sampled noise vectors onto nearly identical visual prototypes. This limits utility for tasks that demand distributional breadth, most critically synthetic data augmentation. Diversity, however, is a collective property of the whole batch: a set can preserve every pairwise distance yet still collapse onto a low-dimensional subspace. Existing sum-of-pairwise noise-optimization and repulsion objectives respond only weakly to this higher-order failure. We propose DetNO (Determinant Noise Optimization), which formulates diverse generation as RKHS volume maximization on the initial noise, maximizing the log-determinant of an RBF Gram matrix over one-step Tweedie predictions. This set-level functional equals the squared batch volume; its repulsion is strongest when the batch is most collapsed and self-attenuates as diversity is achieved. An anchor-based fidelity term keeps the optimized noises near plausible regions, and Lagrangian relaxation gives a two-term objective with a single hyperparameter $λ$ governing the diversity-quality trade-off, without training or sampler modification. Across three backbones (SD1.5, SDXL, SD3), DetNO achieves leading intra-batch diversity in VendiScore and Mean Self-Similarity, and a downstream few-shot classification experiment confirms that broader mode coverage brings statistically significant accuracy gains for synthetic data augmentation in the low-shot regime. Fixing the pipeline and varying only the batch objective, both RBF-based aggregators outperform InfoNCE, and the log-determinant attains the best downstream accuracy on both benchmarks. Its diversity advantage over InfoNCE widens as the batch grows. Code is available at https://github.com/mobled37/DetNO.
benchmark - arxiv:2610.22263 · cs.LGBaselines and Score Centering in Policy Gradient EstimationShuo Han, Ziqing Hu
Reinforcement learning has become a widely used approach to post-training large language models. In a typical implementation, learning is off-policy: The policy generating the responses differs from the policy being trained because of training-inference mismatch or policy staleness in asynchronous training. This discrepancy is commonly addressed by importance sampling, whose ratio is usually regularized in practice through truncation, masking, or clipping. Under regularized importance sampling, the importance-weighted policy gradient is no longer invariant to constant shifts in the reward. This invariance is recovered by using the covariance between the reward and the importance-weighted score as the estimand, which coincides with the policy gradient when the ratio is not regularized. This note develops this covariance perspective for both sequence-level and token-level importance sampling. The covariance perspective reveals the roles of the group-mean baseline and score centering in policy gradient estimation. Both perform centering in covariance estimation, but they differ in the mean being subtracted: The group-mean baseline uses a sample mean, whereas score centering uses a population mean. The two lead to distinct unbiased covariance estimators with different variances. Whether both forms of centering should be used simultaneously depends on whether importance sampling is applied at the sequence level or at the token level. With sequence-level importance sampling, score centering adds no benefit once the group-mean baseline is used; with token-level importance sampling, using the group-mean baseline without score centering generally introduces bias when the importance ratio is regularized.
post-training - arxiv:2610.22262 · cs.LGTopology-Inspired Fine-Tuning for Wind-Field Super-ResolutionArjun Dadhwal, Brian Summa
Scientific super-resolution (SR) models are typically optimized with pointwise reconstruction losses, yet scientific interpretation often hinges on extrema, connected regions, and multiscale scalar structure. We characterize this tension with pretrained CNN and GAN wind-field SR models. Pointwise metrics, bottleneck persistence-diagram (PD) distance, and merge-tree (MT) distance favor the CNN, whereas order-2 Wasserstein PD distances favor the sharper GAN, so even different distances on the same topological descriptor reward different properties. Motivated by this disagreement, we fine-tune the pretrained CNN with differentiable topology-inspired auxiliary losses on scalar speed, gradient magnitude, soft high-speed level sets, and local maxima, and with fixed-pair supervision, which constrains predicted values and persistence at ground-truth persistence-pair endpoints located offline with the Topology ToolKit. A loss study of 16 fine-tuning configurations, each compared with a matched standard fine-tuning control that uses only the original reconstruction loss, on a fixed 168-field benchmark (with topology evaluated on a fixed 160 by 160 subregion of each field), shows that, among the scalar-field losses, gradient supervision provides the strongest PD improvement but does not improve mean MT distance, whereas fixed-pair supervision gives the largest and most consistent joint improvement in PD and MT distances. The final objective combines these two signals. Relative to the pretrained CNN, it lowers mean bottleneck, $W_{2,\infty}$, and $W_{2,2}$ distances by 31%, 25%, and 27% (and by 32-35% relative to the control, which on its own worsens all three), lowers all three on 150 of 168 fields, and gains 1.30 dB PSNR, about half the gain of standard fine-tuning. Mean MT distance also improves, but on only 91 of those 150 fields, indicating that PD and MT capture complementary structure.
benchmark - arxiv:2610.22259 · cs.CVFPG-Seg: Annotation-free Retinal Vessel Topology SegmentationJames Willoughby, Irina Voiculescu
State-of-the art approaches to retinal vessel segmentation typically require extensive annotations, a highly time-consuming process, reliant on clinical expertise. We present FPG-Seg, an annotation-free pipeline for retinal vessel segmentation combining classical vesselness filtering, Bayesian deep learning and topology-aware graph reconstruction to segment retinal vessel topology without manual labels. Tested across four widely used benchmark datasets for retinal vessel segmentation, FPG-Seg attains a clDice of $0.74{-}0.78$, representing $86{-}98\%$ of the performance of an equivalent fully-supervised baseline, despite using no manual annotations for training or hyperparameter tuning. As part of this work, we also provide, to our knowledge, the first performance benchmark of an annotation-free vessel segmentation method on FIVES.
benchmark - arxiv:2610.22256 · cs.LGBitSparse: Compact Activation Storage for ReLU Feed-Forward NetworksMax Zhu, Bohang Sun, Pietro Lio, Jialin Yu
Reducing memory use in large language model (LLM) training is an ongoing challenge. Activation checkpointing saves memory by recomputation, while lossy compression changes the stored activations. The reemergence of sparse $\relu$ and $\relu^2$ activations creates a third opportunity: losslessly compressing the activation cache itself. We present BitSparse, a GPU-resident lossless compression format for sparse activations. Compression and reconstruction run in custom Triton kernels fused into an autograd function, shrinking the activation cache while recovering it exactly for the backward pass. Bfloat16 (BF16) and FP8 activations are supported. Optional sign-bit packing further increases compression up to close to the theoretical limit. We analyze its compression ratio and compare it to the theoretical optimum. BitSparse lowers per-token VRAM by 31\% and 30\% Modded-NanoGPT and Nemotron-H respectively, with overheads of no more than 1\% at practical sequence lengths. Additionally, a new preallocated buffer system avoids host synchronization during training when variable size tensors are allocated.
memory - arxiv:2610.22252 · cs.LGExplore Once, Evaluate Many: Multiple Policy Evaluation for Stochastic Linear BanditsSpyros Dragazis, Yilei Chen, Aldo Pacchiano
We study how to evaluate $K$ target policies simultaneously in stochastic contextual bandits with linear reward structure. Given a known $d$-dimensional feature map and reward-free access to policy queries, we propose PRISMA (Policy evaluation through Regularized Information Sharing and Mixture Allocation), an algorithm that estimates every target policy's value to accuracy $\varepsilon$ with probability at least $1-δ$. The algorithm uses $\widetilde{\mathcal{O}}\!\left( \frac{(L^2S^2+β^2)(1\veeρ^\star)}{\varepsilon^2} \cdot \log(K) \right)$ context draws and reward observations, where $ρ^\star\le d$ is an instance-dependent measure of the difficulty of designing a shared sampling distribution. Here, $L$, $S$, and $β$ bound the feature norm, reward parameter norm, and reward noise, respectively, and the suppressed factors are logarithmic in $K$, $d$, and $1/δ$. Our method combines reward-free planning, a regularized sampling design, and a single ridge regression estimator shared across all target policies. An adaptive doubling procedure selects the regularization scale without knowledge of $ρ^\star$. The guarantees require no numerical lower bound on second-moment eigenvalues and have no explicit dependence on the cardinalities of the context and action spaces. Policy-query costs are accounted for separately, with $K$ queries per shared planning context.
policy evaluation - arxiv:2610.22245 · cs.AIDisagreement-Guided Weak-to-Strong RoutingSima Noorani, Shayan Kiyani, George Pappas, Hamed Hassani
Small language models are increasingly deployed on edge devices and robotic systems to meet cost and latency constraints, yet their limited capabilities can make reliable deployment challenging. Stronger models can provide assistance when needed, but deciding when to route a weak-model response to a stronger model often relies on signals of response correctness. In open-ended and agentic tasks, however, obtaining such signals can require costly ground-truth supervision that is difficult to obtain at scale. We revisit the foundations of weak-to-strong routing and ask whether routing decisions can instead be guided by signals that do not require ground-truth supervision. We first characterize the optimal weak-to-strong routing policy, while allowing for multiple weak-model responses generated through local inference. Building on this characterization, we derive a principled approximation to the optimal routing based on model-to-model comparison. In particular, we introduce a notion of model disagreement that can be measured directly from model-generated responses, without ground-truth supervision. Importantly, the approximation becomes increasingly accurate as the strong model becomes a better proxy for the ground-truth distribution. Across agentic and single-turn tasks, our approach effectively navigates the trade-off between routing cost and response accuracy without ground-truth supervision. Moreover, even when ground-truth correctness signals are available for training, disagreement-based routing can outperform methods trained directly to predict correctness
agentic - arxiv:2610.22244 · cs.LGLanguage Is a Security Surface for Agent GuardsHasan Demirkiran, Jens Ernstberger
Step-level guards decide, before each tool call, whether an agent's next action is safe given the user's request and the trajectory so far. Users write those requests in many languages, while guard prompts and the benchmarks on which purpose-built guards are trained and evaluated are English. We ask whether a guard's verdict depends on the language of the request. We build LinguaGuardBench from the 1,220 labelled steps of TS-Bench AgentDojo-Traj: we translate only the user instruction into German, French, Turkish, Swahili and Amharic, and keep every tool call, observation and label byte-identical, so any change in a verdict is attributable to language alone. Six English rewordings and a whitespace placebo serve as controls, and managed guards are run three times to measure their noise. Across 19 guards, from frontier APIs to 4B open models and four purpose-built agent guards, most open, small and purpose-built guards change verdicts under translation beyond their own noise floor, and the effect grows with the rarity of the language: among the nine guards below 13B parameters, the mean placebo-adjusted flip rate rises from +1.5 pp on French to +4.4 pp on Amharic. Purpose-built guards are the most sensitive. The direction of the damage tends to follow guard size: open guards of 12-30B parameters block more benign steps (up to +5.0 pp false positives), and every significant loss of recall is in a guard below 10B (up to -7.4 pp). Frontier guards shift by at most about one point; the strongest, Gemini 3.8 Flash, returns identical verdicts with and without the request. The damage is concentrated in a few instructions and, in a post-hoc analysis, falls on steps where the agent's reasoning restates the English request, consistent with guards reacting to inconsistency between channels. English paraphrase moves every guard less than translation does. We release the benchmark and all 489k verdicts.
agentbenchmark - arxiv:2610.22243 · cs.LGShortcut Sensitivity in Agent-Safety EvaluationsHasan Demirkiran, Jens Ernstberger
AI agents take actions through tools. Some actions are sensitive, so guards that supervise them must themselves be evaluated reliably. Yet the data used to evaluate a guard can include cues that make labels predictable without measuring the intended safety construct, and those cues can also influence guard decisions. We introduce a three-level shortcut-auditing framework and audit completed LinuxArena trajectories and two TS-Bench-derived step datasets. We distinguish label predictability, observational guard association, and paired decision sensitivity. In LinuxArena, shorter average shell commands strongly predict trajectory labels. In AgentDojo-Traj, exact-tool lookup falls from F_1=.901 with mixed-domain folds to .495 under domain holdout. In AgentHarm-Traj, released TS-Guard F_1 falls from .902 under row weighting to .719 with equal interaction weighting, while balanced accuracy rises. We then test the strongest safe-row marker association in a seven-configuration screen on a frozen 64-step AgentDojo-Traj cohort using repeated, interleaved Haiku calls. Rewriting the <INFORMATION> marker lowers blocking by 8.33 percentage points, while opaque tool aliases raise it by 11.46 points. All unsafe calls remain blocked; the changes concern false positives against inherited labels. Both effects remain nonzero after subtracting the formatting-placebo effect, with exploratory 95% ranges excluding zero. The same edits also change measured accuracy against inherited labels. This provides direct evidence that benchmark representation can materially change a guard's decisions and measured evaluation performance. The Haiku-sized effects do not reproduce in Luna or deterministic Q8 TS-Guard. We recommend domain group holdouts, evaluation-unit metrics, and placebo-controlled paired interventions.
ai agentbenchmark - arxiv:2610.22242 · cs.LGSurvICL: a Foundation Model for in-context Survival AnalysisJulie Alberge, Linus Bleistein, Judith Abécassis, Marine Le Morvan +2
Tabular foundation models (TFMs) now outperform gradient-boosted trees in classification and regression. Yet in survival analysis, which predicts when an event occurs from censored data, often on small clinical cohorts, we find that no existing TFM beats the half-century-old Cox model. We address this issue by introducing SurvICL, a TFM that predicts survival curves in-context. SurvICL is pretrained on purely synthetic datasets generated from a novel flexible proportional-odds prior that outperforms other prior types in our ablation study. Unlike most other survival models, SurvICL is trained with a plain quantile regression loss, which is possible because its synthetic pre-training data generator can provide uncensored event times. Additionally, we introduce the largest survival benchmark to date spanning 103 datasets. On the 63 datasets held out from all design decisions, SurvICL outperforms 20 tree-based, deep, and foundation-model baselines without any tuning: it wins 82.9\% of pairwise comparisons in integrated Brier score, against 71.3\% for the runner-up, the Cox model, and is the first model to rank above Cox on small datasets.
benchmark - arxiv:2610.22236 · cs.LGTraining-Free On-Policy DistillationBangji Yang, Jiajun Fan, Zhaoyu Zhu, Hongbo Ma +1
On-policy distillation consolidates specialist models into one student at the price of rollouts, teacher passes and gradient steps for every teacher. We show that the consolidation does not need training: with forward passes only, the capabilities of separately tuned specialists can be aggregated into one model that is, on average, more accurate than either specialist alone. When specialists share a base model, distillation has an exact layer-wise solution: each linear layer of the fused model must reproduce every specialist's output on the inputs that specialist receives while reading its own answers, a least-squares problem whose solution needs only the second moments of those inputs. One forward pass over a few hundred rollouts provides them. We call this training-free on-policy distillation (\method). What the rollouts contain matters more than the solver: answers truncated in repetition loops are few among a mathematics specialist's rollouts but carry a large share of their tokens, and removing them lifts the fused model on AIME25 from task arithmetic's level to the specialist's at matched precision. On Olmo~3 7B, a few GPU-hours of forward computation per specialist yield a generalist that keeps most of each specialist's gain on each of seven mathematics and code benchmarks.
benchmark - arxiv:2610.22234 · cs.LGOS-Helix: Unifying GUI Interaction and Code Execution in Native Computer-Using AgentsBowen Yang, Kaiming Jin, Zichen Ding, Zhenyu Wu +7
Real desktop automation is neither purely visual nor purely programmatic. Agents must understand user-facing interfaces while efficiently manipulating files, scripts, and system state. However, native GUI/code coordination cannot be obtained by prompt-level tool composition. In our preliminary experiments, naively adding a code tool degrades a GUI agent, revealing that GUI/code routing must be internalized through training on high-quality hybrid data. We present OS-Helix, an end-to-end trained multimodal computer-using agent that unifies GUI actions and code execution in one action space. To construct the required supervision, we introduce application-artifact composition, pairing desktop applications with authentic file artifacts to instantiate executable Ubuntu states, and a decoupled instruction/evaluator generation pipeline that produces environment-grounded tasks whose final states can be verified regardless of the solution modality. We cold-start the policy with SFT on filtered hybrid data and further improve it through online RL in live desktop environments. For RL, we propose Step-balanced GSPO, which aggregates importance weight by action step rather than raw assistant tokens to prevent long code actions from dominating heterogeneous-action optimization. On OSWorld-Verified, OS-Helix-27B and OS-Helix-35B-A3B achieve relative improvements of 35.0% and 33.7% over their matched Qwen3.5 hybrid baselines, while OS-Helix-27B-RL further improves by 42.9% to reach 66.44% success. Ablations provide evidence consistent with contributions from learned GUI/code coordination, scalable verifiable supervision, and action step balanced optimization, while demonstrating cross-platform transfer on WindowsAgentArena. Our project homepage is at https://os-copilot.github.io/OS-Helix/.
helixagentevaluator - arxiv:2610.22232 · cs.CVWaveLLM: Wavelet-Domain LLM Entropy Modeling for Lossless Image CompressionYue Yu, Bowei Kang, Changsheng Gao, Dong Liu +1
Lossless image compression fundamentally relies on accurate entropy modeling, driving a recent shift toward large language model (LLM)-based methods. However, existing approaches mainly emphasize the modeling capacity of LLMs while directly serializing images into pixel or residual sequences, leaving heterogeneous signal patterns entangled within a single prediction stream. In this paper, we argue that effective entropy modeling depends not only on model capacity, but also on the predictability of the input representation. Motivated by this insight, we propose WaveLLM, a wavelet-domain LLM-based entropy model for lossless image compression. By decomposing an image into multi-level wavelet subbands, WaveLLM converts a heterogeneous prediction problem into a set of more structured and statistically homogeneous subproblems that are more amenable to autoregressive probability estimation. To bridge numerical coding symbols with the LLM's native text token space, we introduce a dedicated numerical interface featuring value embeddings and a coefficient prediction head. Furthermore, a task-adaptive prompting strategy is designed to enable a single shared model to adapt to diverse subbands within a unified framework. Built on Llama-3.2-3B with LoRA adaptation, WaveLLM achieves 2.51, 2.24, and 2.27 bpsp on Kodak, CLIC.pro, and DIV2K, respectively, outperforming the PixelCNN-based codec using the same wavelet transform by 0.43, 0.36, and 0.46 bpsp, and surpassing recent LLM-based lossless compressors on all three benchmarks.
benchmark - arxiv:2610.22230 · cs.LGArchitecting for Hardware Diversity: Meta's End-to-End Co-Design for High-Performance AI Training on AMD GPUsHaoyu Zhang, Arshad Rahman, Adnan Aziz, Ajith Prabhudev +54
Meta deploys a heterogeneous fleet comprising NVIDIA GPUs, AMD GPUs, and proprietary MTIA accelerators to reduce costs and mitigate vendor lock-in. However, the extensive history of NVIDIA-specific optimizations hinders performance portability to other hardware. This paper presents our methods for optimizing alternative accelerators, using AMD Instinct MI350X versus NVIDIA B200 for recommendation model training as a case study. We propose three reusable principles. First, hardware-model co-design is essential: rather than directly porting models from NVIDIA GPUs, revising model structures and software recipes allows full exploitation of accelerator-specific strengths (e.g., MI350X's 1.6x HBM capacity relative to B200). Second, our analytical performance model enables zero-warmup optimal kernel selection across diverse datatypes and matrix shapes in production models, drastically simplifying tuning. Third, increasingly mature agentic workflows have lowered the barrier to porting from the CUDA ecosystem; we extend these workflows for microarchitecture-aware kernel optimizations and custom shim generation. Compared to the directly ported baseline, these methods deliver up to 1.5x improvement in end-to-end training throughput. To our knowledge, this is the first hyperscaler report on productionizing AMD GPUs for large-scale training. Our contributions have been upstreamed to the open-source PyTorch, Triton, and ROCm ecosystems.
agentic - arxiv:2610.22217 · cs.CVSkill-Decoupled On-Policy Expert Distillation for Unified Vision-Language ModelsYuqi Xiong, Chunyi Peng, Zhipeng Xu, Qing Yang +4
Reinforcement learning (RL) can elicit specialized visual capabilities in vision-language models (VLMs), yet jointly optimizing heterogeneous objectives often introduces cross-capability gradient conflicts. Training experts independently mitigates this interference, but direct reward optimization may entangle specialized knowledge with reward-hacking biases. This paper proposes Skill-Decoupled On-Policy Expert Distillation (SCOPED), which trains capability-specific experts and extracts explicit skills through reward-guided reflection. During on-policy distillation, these skills guide expert supervision while remaining unavailable to the unified student, encouraging it to internalize procedural knowledge without explicit skills at inference time. Across eight multimodal benchmarks, SCOPED consistently outperforms the backbone and mixed-capability RL, demonstrating the effectiveness of skill-decoupled on-policy expert distillation for integrating independently acquired capabilities. Further analyses reveal interference during joint capability optimization and suggest that skill augmentation promotes more consistent expert trajectories, greater attention to task-relevant visual evidence, and better preservation of specialized expertise in the unified student. All code and data are available at https://github.com/NEUIR/SCOPED.
benchmark - arxiv:2610.22215 · cs.LGFutureLift: Lifting Trajectory Proposals into Bidirectional Future Graphs for Motion ForecastingDaniel Grimm, J. Marius Zöllner
Accurate motion prediction is critical for safe autonomous driving, yet even strong forecasters leave residual errors in their multimodal proposals. Existing refiners compress complete proposals or update their segments sequentially, preventing the full proposed future from serving as explicit relational context for correction. We introduce FutureLift, a backbone-agnostic refiner that represents each multimodal proposal as a heterogeneous future graph, with spatially grounded temporal nodes connected to agent histories, map elements, alternative modes, and, for joint prediction, other agents future trajectories. Bidirectional temporal edges allow later evidence to revise earlier segments in a single pass, and the updated nodes predict residual coordinate and confidence corrections. Requiring only proposals, scores, and scene context, FutureLift provides a common refinement interface for structurally different forecasting backbones. On Argoverse2 and the Waymo Open Motion Dataset, it consistently improves frozen proposals and outperforms native and specialized refiners across marginal and joint forecasting. On the WOMD motion-prediction leaderboard, FutureLift ranks first in minADE (0.5049m) and minFDE (0.9942m) among the compared standard-input methods, without LiDAR or ensembling, while achieving a competitive 0.4874 Soft mAP.
agentleaderboard - arxiv:2610.22210 · cs.CVTRACE: Topology-Aware Reconnection via Prediction-Adaptive Connectivity Enforcement for Thin-Structure SegmentationJunho Moon, Ikbeom Jang
Thin, elongated structures are particularly sensitive to connectivity errors, as even a few local segmentation mistakes can disrupt otherwise continuous paths. Existing skeleton-based objectives typically supervise fixed centerline locations and do not account for alternative local routes that may restore connectivity. We propose TRACE, a topology-aware, prediction-adaptive objective that uses a target skeleton to define required local connections. For each connection, TRACE identifies the easiest valid reconnection under the current prediction and jointly supervises its bottleneck with the corresponding skeleton point. The supervision thus adapts to the evolving prediction rather than remaining fixed to the centerline. We further introduce a topology-aware skeleton carrier that preserves both foreground and class-specific topology, enabling reliable definition of these connections. Extensive experiments across six 2D and 3D benchmarks show that TRACE improves connectivity and topology while maintaining strong segmentation accuracy with low training overhead. Code will be released upon acceptance.
benchmark - arxiv:2610.22209 · cs.LGBeyond Downstream QA: Rethinking How We Evaluate Agent MemorySebastian-Antonio Toma, Jean de Kernier, Mike Ply, Oliviu Matei
Agent memory benchmarks evaluate whether an agent's responses reflect correct knowledge, measuring downstream question-answering accuracy as a proxy for memory quality. We show this evaluation paradigm is incomplete: it conflates memory correctness with retrieval and generation quality, producing both false positives where correct responses mask broken memory and false negatives where incorrect responses blame correct memory for retrieval failures. We introduce FEB-30, a benchmark of 30 fact-evolution trajectories spanning 6 diagnostic categories and 3 difficulty levels, constructed via paradigm inversion: documents are generated from known fact trajectories, yielding deterministic ground truth without human annotation. Each trajectory defines how a fact should evolve across multi-author artifacts (documents, emails, conversations), enabling direct inspection of the memory store after each ingestion step. Applying FEB-30 to a representative memory system, we demonstrate that intrinsic and extrinsic evaluation diverge by an average of 16 percentage points, with 44% of trajectories exceeding meaningful divergence thresholds. Diagnostic decomposition reveals that the system correctly classifies memory operations (create, update, reinforce, supersede) only 20% of the time, a failure entirely invisible to extrinsic evaluation, and that failure modes differ systematically by category and difficulty. We release the benchmark, document generation pipeline, and scoring framework to support reproducible evaluation of agent memory systems.
memoryagent memoryagentbenchmark - arxiv:2610.22201 · cs.AIDecentralized LLM Workflows via Sequential AuctionsIan Gemp, Shubhendu Trivedi, Adam Fisch, Fantine Huot +6
We present a scalable framework for solving multi-step tasks through a decentralized market of self-interested agents. Prior theoretical work demonstrated that agents organized through sequential auctions can converge to a globally optimal policy in a Markov decision process (MDP), but relied on the restrictive assumption of the existence of perfect clones as a model of perfect competition. In reality, exact clones drive agent profit to zero, violating the natural risk-reward incentives of participants. Conversely, we empirically demonstrate that the opposite extreme, absence of competition, also leads to market failure and value extraction collapse. Between these two extremes, we investigate how heterogeneous agents can survive in markets that span the critical middle ground of imperfect competition. To be profitable, agents learn to strategically shade bids to secure positive individual profit margins, while still ensuring that the marketplace functions effectively enough so that collaborative tasks posed by market consumers can still be solved. We formulate and evaluate four bidding algorithms (Mirror Descent, Thompson Sampling, Bayes-Optimal, and PPO) across three auction protocols (1st-Price, 2nd-Price, and Credit-Conserving Vickrey). Across a diverse set of domains, we demonstrate that imperfect competition successfully sustains necessary agent profit margins while maintaining high task success rates at low costs.
agent - arxiv:2610.22200 · cs.CVOpenGCUA: Toward a Fully Open Stack for Generalist Computer UseHanzhang Zhou, Long Li, Quyu Kong, Lei Wang +11
Computer use has reached a turning point: the focus is expanding from GUI-only interaction to systems that complete tasks on users' computers through visual and programmatic interfaces. We define a generalist computer-use agent (GCUA) as a system in which the model chooses how to observe, act, and organize its work across visual and programmatic interfaces. Yet frontier implementations, such as GPT-6 Astra's computer-use system, remain largely closed, limiting systematic study of their design. We introduce Open Generalist CUA (OpenGCUA), an ongoing project to develop and open the full stack for GCUAs with the community. This report presents the first release of OpenGCUA. Code execution is a key feature of our agent design: it lets intermediate results determine subsequent operations and enables models to apply their existing programming capabilities directly to computer interaction. We also propose harness-to-model decision transfer, moving selected observations and task-state decisions from fixed harness rules to the model. On OSWorld-v2, our full configuration improves task scores across all eight evaluated models, including GPT-6 Astra and Claude Opus 5, raising mean partial score by +19.6 points and reducing reported steps per task by an average of 53.0% relative to batched GUI-only execution. We release our implementation and prompts, evaluation trajectories from all eight models, and preliminary versions of a real-device harness and a cross-device SDK. The harness addresses real-device challenges not fully captured by benchmarks, while the SDK provides a common programming interface across computers, mobiles and browsers. Across four real macOS tasks under our real-device harness, selected Qwen3.8-Max runs use 19.4 times as many reported reasoning tokens as GPT-6 Astra, highlighting substantial differences between models in real-device execution efficiency.
agentbenchmark - arxiv:2610.22198 · cs.RORobust Multi-Agent Pursuit with Preemptive Communication via Reachability-Based Control Barrier FunctionsMadison Book, Justin Bradley
Safe multi-agent coordination typically assumes what real-world, deployed, resource-constrained robots cannot provide: continuous, perfect state sharing. Robust control strategies must be complemented with robust communication strategies to deliver resilient deployments. In this work, we propose a unified control/communication strategy that provides safe control to guarantee inter-agent safety under intermittent communication and demonstrate its effectiveness in a pursuit-evasion scenario. Each agent computes the reachable set of neighboring agents based on the last state information received, then solves for a safe control input based on control barrier functions to ensure agents avoid collision. We prove that inter-agent collision avoidance is guaranteed provided communication staleness remains bounded. As the reachable set inflates between state transmissions, each agent reasons about when to preemptively transmit its state information to neighboring agents based on novel metrics characterizing uncertainty risk and kinematic risk. We implement and simulate a variety of pursuit-evasion scenarios and show that preemptive communication reduces messages by 77.5% compared to periodic and event-triggered baselines with zero collisions.
agentmulti-agent - arxiv:2610.22196 · cs.CLHy-Memory-30B-A3B: Attributed Reinforcement Learning for Long-Term Memory WritingTianxiang Fei, Jiazheng Kang, Mao Zheng, Xingyan He +1
External-memory systems extend language-model applications beyond a context window, but their quality depends critically on the model executing the write path: deciding what to retain, how precisely to represent it, and how new evidence should update persistent state. Existing frameworks typically place a general-purpose large language model behind framework-specific prompts. This raises serving cost, and a writer specialized for one protocol does not transfer directly to another; a unified model that can operate across established memory frameworks remains largely unexplored. We present Hy-Memory-30B-A3B, a compact long-term-memory writer trained across Mem0, MemOS, and Hy-Memory. Memory On-Policy Distillation teaches framework-specific write protocols on student-generated states; Memory RL then optimizes downstream answerability using dense per-chunk probes, attributed cross-chunk rewards, over-storage controls, and chunk-level GRPO. Carrying policy-written banks across windows keeps training aligned with deployment. On fixed 20% benchmark subsamples, Hy-Memory-30B-A3B ranks first on all four evaluated harnesses, including held-out EverMemOS, with a 55.49% macro average versus 54.62% for Claude Opus 5.
memorypersistent statebenchmark - arxiv:2610.22195 · cs.AIHTN-Orchestrate: A Hierarchical Task Network Framework for Multi-Agent OrchestrationYanchen Deng, Jianguo Huang, Yao Long Teng, Yixuan Li +5
Orchestrating multiple agents on a complex task requires reconciling local subtask execution with global task planning. Dynamic dispatch approaches provide execution flexibility but do not explicitly model subtask hierarchies or dependencies, while predefined workflows impose this structure but fix the decomposition before execution begins. Therefore, we introduce HTN-Orchestrate, a two-layer framework that separates these responsibilities through a hierarchical task network (HTN). The Planner maintains subtask hierarchies and dependencies, deferring decomposition until prerequisite results become available, while local Orchestrators schedule workers to execute each leaf subtask. Results propagate upward under \textbf{AND} semantics, and accumulated findings inform a rebuilt task network when the current decomposition cannot resolve the task. On five benchmarks spanning web research, scientific reasoning, and research workflows, HTN-Orchestrate outperforms ReAct, Agent Team, and DeepSeek Harness using the same base model, achieving 89.73% accuracy on BrowseComp and 73.44% on SUPERChem, exceeding Agent Team by 2.45 and 4.69 percentage points, respectively.
agentmulti-agentbenchmark - arxiv:2610.22194 · cs.AIAutoSFT: An Agentic System for Automated Data Research in Supervised Fine-TuningBingxiang He, Pengqi Sun, Shihao Peng, Shaoxuan Li +17
Language model agents increasingly automate post-training, but when given the full training stack they converge on supervised fine-tuning and spend most effort revising training data. Existing systems give the agent one kind of control, selecting rows from a fixed pool or calling a generator, whereas a human data engineer adjusts proportions, prunes, repairs, replaces and synthesizes the data. We design a data action space for these changes and ask whether an automated system can make the full range rather than one of them, and what structure that requires. We first run this loop with a single agent, and it reaches only one action, adjusting category proportions, while its context fills with execution output and is discarded mid-search. We therefore introduce \textbf{AutoSFT}, a multi-agent system where a meta agent separates execution from analysis, gathers proposals from independent angles, and calls a synthesis subagent for capabilities the pool cannot cover, with the model, the trainer, and the budget fixed within each comparison. Adding each component in turn helps: the executor halves context loss, the analysis subagents lift the best IFEval by $+6.5$, and synthesis adds $+3.4$ IFEval and $+2.9$ IFBench. The same loop transfers to code, where analysis adds $+3.7$ MBPP+ over a single agent that does not improve. Every result comes from runs that retrain from the base model each round, so each component's effect is individually attributable. Our findings suggest that automating SFT data research depends more on covering diverse data actions than on a stronger single action.
agentmulti-agentagenticagent systempost-training - arxiv:2610.22193 · cs.ROMOSS: Mode-Oriented Source Shaping for Robot Flow Matching PoliciesZirun Ma, Jindou Jia, Tuo An, Jiaqi Bai +2
In robotic manipulation, the same observation often admits multiple feasible actions. Generative policies such as diffusion and flow matching naturally capture such multimodal behaviors, yet which mode is executed is left to random sampling and cannot be actively controlled. We propose Mode-Oriented Source Shaping (MOSS), which turns mode selection into an explicit choice of where generation starts. We first reveal, both empirically and theoretically, that deterministic flow transport induces mode-dependent structure in the source space. Under ideal deterministic ODE transport, distinct behavioral modes have distinct source-space preimages, while practical approximation errors can blur their boundaries due to imbalanced data, limited model capacity, and few inference steps. Motivated by this source-space structure, MOSS constructs semantically aligned mode-specific Gaussian sources that encode mode identity in the initial noise. A modality-strength estimator predicts the local degree of multimodality, which adaptively controls the geometry of the mode-specific sources. It contracts them toward a shared Gaussian where behaviors are similar and separates them near behavioral branching points. During deployment, the desired behavior is selected at inference simply by sampling from the corresponding component. Across 6 simulated and 3 real-world manipulation tasks, MOSS steers to the target mode up to 100% of the time with high task success, recovers minority modes at over 90% guidance rate under a 10:90 imbalance dataset, and surpasses mode-conditioned flow policy by 43.3 points of guidance rate on real robots.
manipulation - arxiv:2610.22192 · cs.AITerminal-RSI: Recursive Self-Improving Terminal Agents via Recursive Self-Rewrite and Policy-RefineJunyao Yang, Zongxia Li, Yucheng Shi, Zhongzhi Li +5
Converting verifier outcomes into useful policy updates and partial progress into complete task resolution is crucial for terminal agents. We present Terminal-RSI, which alternates Recursive Self-Rewrite (RSR) of verified trajectories with critic-inheriting Recursive Policy-Refine (RPR). We formulate Turn-Boundary Skip-Observation GAE (TB-GAE) to retain token-specific baselines while applying decay across interaction turns. Family Dense Reward (FDR) calibrates semantic progress across task families; Success-Prioritized Reward (SPR) separates strict success from bounded partial credit. A Warm-up Critic Model, recursive critic handoff, and stage-specific learning rates coordinate the two refinement phases. The five-stage sequence from Qwen3.8-27B base model through RSR-1, RPR-1, RSR-2, and RPR-2 yields Terminal-RSI-27B: 19.84% on Terminal-Bench 4.0, 86.52% on Terminal-Bench 2.1, 16.95% on Terminal-Bench 3.0, and 79.00% on Terminal-Bench Hard. On LHTB, process reward rises from 0.21 to 0.37 and strict completion from 0/46 to 5/46; SWR100 success doubles from 3.00% to 6.00%. By recursively coupling verified trajectory rewriting with critic-inheriting policy refinement, each retained stage consolidates discovered behaviors and progressively advances the agent from partial local repairs to strict terminal verification.
agentself-improving - arxiv:2610.22189 · eess.SYThe role of friction on heave-induced motion of offshore drilling systems: modelling, analysis, and numerical experimentsLuigi Romano, Ole Morten Aamo
This paper investigates the influence of friction dynamics on heave-induced disturbance propagation using a coupled hydraulic-structural drilling model formulated as a semilinear interconnection of ordinary and hyperbolic partial differential equations (ODE-PDEs), and capturing spatially distributed memory effects in the drillstring-wellbore interaction. First, local and global existence and uniqueness results are established, and sufficient conditions for boundedness and input-to-state stability are deduced via a small-gain analysis. Subsequently, extensive numerical investigations are carried out, including harmonic excitation studies, nonlinear distortion analysis, spatial amplification investigations, mechanism-isolation experiments, parameter sensitivity analysis, and broadband simulations under realistic offshore sea-state excitation. The simulation results indicate that, whilst the hydraulic response is governed primarily by wave propagation phenomena, compressibility, and boundary reflections, the structural dynamics display substantially greater sensitivity to distributed friction effects. Furthermore, drillstring-wellbore friction is shown to have a relatively limited influence on well-posedness but a potentially significant impact on stability margins. These findings highlight the importance of accounting for coupled hydraulic-structural dynamics and distributed friction mechanisms when analysing, monitoring, and controlling offshore drilling systems.
memory - arxiv:2610.22186 · cs.CVEVA-Cap: Optimizing Audiovisual Video Captioning via Event-Centric AlignmentJiapeng Shi, Qiuxia Hou, Jiaqi Leng, Junke Wang +4
Recent advances in Omni Language Models (OLMs) provide a unified solution for audiovisual video captioning. Nevertheless, existing approaches predominantly treat captioning as an unstructured text generation process, neglecting the inherent semantic structure of videos and leading to suboptimal fine-grained audiovisual alignment. To bridge this gap, we propose EVA-Cap, a novel framework that decomposes captions into atomic audiovisual events and transforms unstructured text supervision into event-centric alignment. We train EVA-Cap via Supervised Fine-Tuning (SFT) and Group Relative Policy Optimization (GRPO) on a self-curated dataset, EVA-Data, which is built with event-centric quality control. During the GRPO phase, the model is explicitly optimized against three event-centric alignment targets: global event completeness, intra-event attribution, and inter-event synchronization. Extensive experiments demonstrate that EVA-Cap significantly outperforms existing methods in event-centric evaluation (e.g., +9.0% average gain on ChronusAV), while achieving competitive performance on general audiovisual captioning benchmarks (e.g., +3.3% on WorldSense).
benchmark - arxiv:2610.22184 · cs.ROWhole-Body DexterityConor Mc Gartoll, An Dang, Pranay Thangeda, Arjun Gupta +4
Current humanoid dexterous manipulation remains largely confined to end-effectors and simple grasps, falling short of humans' ability to engage the entire body by bracing and sliding bulky objects against the arms and torso. We introduce Whole-Body Dexterity (WBD), a general framework that uses the robot's body and appendages as contact surfaces to grasp and manipulate objects. Our sim-to-real pipeline first generates a dataset of feasible whole-body grasps, then trains reinforcement-learning experts on a general any-to-any repositioning objective by sampling initial and goal grasps from the dataset, and distills these experts into a single policy that uses proprioception and object geometry. We show that the distilled policy transfers zero-shot to a real Unitree G1, where it repositions a basketball and a cardboard box to regions around the torso and completes a long-horizon task that stows a box under one arm, carries it from a shelf to a table, and places it. The controller operates interactively, steered by coarse motion references from a human teleoperator. Across real-world whole-body manipulation tasks, the distilled controller completes 90% of trials, compared with 18% for direct retargeting and 4% for a generalist whole-body controller.
manipulationdexteroushumanoidwhole-body controlsim-to-realgrasp - arxiv:2610.22179 · cs.LGLCT: Losslessly Compressed Tensors with GPU friendly execution for TrainingMax Zhu, Bohang Sun, Pietro Lio, Jialin Yu
Lossless tensor compression could reduce memory usage during training, but training requires tensors to be repeatedly re-encoded. We introduce Losslessly Compressed Tensors (LCT), a GPU execution framework designed for this repeated encode-decode regime. LCT exploits the natural redundancy in floating point data to reduce storage for weights, activations and optimizer states. Highly optimized GPU kernels implement Huffman coding with precomputed codebooks for low compression and decompression overhead, essential for training. Our analysis characterizes LCT's storage costs and adaptation to magnitude shifts. LCT achieves a compression ratio of up to $1.455\times$, and is demonstrated in training and finetuning settings. On a 162M-parameter NanoGPT, LCT reduces memory usage by $24.6\%$ for a $7.2\%$ step-time increase. In Qwen3-4B fine-tuning, LCT saves up to $23.8\%$ of peak memory, enabling a $3.5\times$ larger context length in training on the same GPU.
memory - arxiv:2610.22178 · cs.ROREKI: Registering Knowledge in VLAsJuliette Marrie, Paul Pacaud, Adrien Ramanana Rahary, Shizhe Chen +1
We introduce REKI, a 2.4B vision-language-action model (VLA) pretrained exclusively on public data that sets a new Pareto frontier between performance and model size: on LIBERO-Plus, RoboTwin Clean2Rand and RoboCasa365, it is outperformed only by models about twice its size. REKI consists of a vision-language model (VLM) followed by a flow-matching action expert, and is pretrained on a curated mixture of public robot, egocentric human and vision-language data. Additionally, we endow it with a memory that gives the policy access to its recent past at negligible cost: a few register tokens in the action expert summarise the output of the VLM at each frame and persist in the expert's cache. This memory improves success in simulation and on real robots, and makes oscillations between poses about four times rarer. REKI uses a simple model design and training paradigm, and is optimised to be efficient at training and inference. We release the full code across data curation, training, inference and evaluation, deployment, and all pretrained and finetuned model weights: https://github.com/kyutai-labs/reki.
vision-language-actionliberorobotwinmemory - arxiv:2610.22177 · cs.CVMotion- and Gravity-Aware Feedforward 3D ReconstructionShiyao Qian, Yuan Ren, Dongfeng Bai
Feed-forward transformer models have recently achieved strong performance in multi-view 3D reconstruction, but they still rely primarily on appearance-based attention and often struggle in dynamic scenes with weak geometric cues. In this work, we introduce a motion- and gravity-aware framework under the Learning Using Privileged Information (LUPI) paradigm. During training, residual optical flow and gravity direction are provided as privileged signals to guide geometric representation learning; during inference, the model uses standard visual inputs only. We integrate these priors into transformer backbones with lightweight architectural changes, preserving feed-forward efficiency. Experiments on standard benchmarks show consistent gains in reconstruction accuracy and robustness, especially in scenes with dynamic objects, ambiguous structure, and challenging camera motion. Our analysis further shows that the optimal injection stage is backbone-dependent: middle-layer integration is generally preferable for VGGT, while late-stage integration is more effective for π3. These results suggest that LUPI offers a practical way to inject motion and physical priors into feed-forward 3D reconstruction without increasing test-time complexity.
benchmark - arxiv:2610.22176 · cs.ROInertially Significant Humanoid Teleoperation via Whole Body Model Predictive Control and Online Parameter EstimationJames Foster, Beomyeong Park, Luigi Penco, Robert Griffin
The effective deployment of humanoid robots in practical tasks of interest will require a capability of sensing and adapting to variations in inertial parameters, as occurs during heavy tool use or when carrying significant payload. Such payloads constitute a significant proportion of the combined payload-robot system, the effects of which cannot be ignored by delegating to feedback control. This paper showcases a general solution to this problem by integrating a whole body model predictive controller (WB-MPC) with an extended Kalman filter (EKF) for online estimation of fully physically consistent time-varying inertial parameters. These estimates are fed through to the WB-MPC's full rigid body dynamics model, resulting in more accurate feedforward torques for task demands under external load. Additionally, a virtual reality (VR) interface is provided that enables an operator to provide task-specific references without regard to payload while delegating dynamic balance to the WB-MPC. The framework is validated with simulation and hardware experiments on a humanoid system.
humanoidteleoperationtool use - arxiv:2610.22174 · cs.LGMosaic: Multi-Session Agentic Reinforcement LearningYilong Xu, Minrui Xu, Xinyu Zhou, Xiang Long +2
Most existing agentic reinforcement learning methods assume single-session trajectories, where interaction proceeds within one continuously accumulated context. However, agentic techniques such as context compaction and subagent forking may produce trajectories spanning multiple sessions, creating a mismatch between the single-session training assumption and inference-time execution. Therefore, we introduce Multi-Session Agentic Reinforcement Learning, a more general setting that supports an arbitrary number of sessions within a trajectory. Under this setting, a widely adopted method is to broadcast trajectory-level advantages to all sessions, but this does not distinguish contributions from various sessions or guide session allocation within trajectories. To address this issue, we propose Mosaic, a multi-session agentic RL method with critic-free hierarchical optimization. At the trajectory level, it incorporates session-usage signals into advantage estimation. At the session level, it shapes each session's advantage according to the estimated marginal contribution. Experiments on BrowseComp with Qwen3.5-4B show that Mosaic outperforms broadcasting in both context compaction and subagent forking scenarios while using fewer sessions on average. Further analyses show that our method achieves improvements in session utility and allocation, which we suggested is a promising target for future advances in multi-session agentic RL.
agentic - arxiv:2610.22163 · cs.AIMulti-sourced AI Tutor for Non-Native English Speaker StudentsYiyin Shen, Xinyi Ai, Weibing Wang, Rogers Jeffrey Leo John +1
Large language models (LLMs) are increasingly being adopted as teaching tools in computing education, particularly for developing AI tutors capable of answering student questions. With the rise in enrollment of international students in computing programs, researchers have begun to explore the difficulties that non-native English speaker (NNES) students face in computing courses due to language barriers. To aid such students, we developed a general-purpose AI tutor system that can process student questions in foreign languages (currently focused on Mandarin) and provide bilingual responses. Our system also incorporates the functionality of generating a glossary of technical terms and their explanations to enhance comprehension. Our approach utilizes a specialized Retrieval-Augmented Generation (RAG) pipeline with a novel question extraction mechanism, enabling us to effectively utilize a diverse array of course materials for answering student queries. We evaluated our AI tutor using 91 Piazza posts, demonstrating that it could answer 83.52% of the posts correctly and in a helpful manner. The tutor also provided high-quality translations for 76.92% of the questions and generated technical term glossaries for 64.47% of them. Our findings underscore the potential of multi-source RAG pipeline in enhancing AI tutor systems, offering support for non-native speakers that are challenging to provide with traditional supporting staff of a computing course.
retrieval-augmentedragrag pipeline - arxiv:2610.22160 · cs.AIEMO: A Closed-Loop Test-Time Computation Framework via Alternating Context Optimization and Logits SteeringWei Xia, Yehui Tang, Zhi-Hong Deng
Test-time context optimization adapts frozen language models through textual feedback, but ordinary conditional generation provides no separate control over how strongly to reinforce the distributional changes induced by a context update. We introduce EMO, a training-free framework inspired by expectation--maximization that alternates question-specific context estimation (E-step) and distribution refinement (M-step). The E-step revises a system context using textual gradients from an external judge without access to gold answers. The M-step contrasts next-token distributions under the revised and preceding contexts at a shared prefix to construct a confidence-constrained logit correction. Feedback on the steered response drives the next context update, coupling adaptation across rounds with intervention within each generation. Theoretically, we connect parameter and context optimization through a shared target-KL objective and additive logit representations, and derive the steering vector from context-induced logit displacements. Under context-improvement and faithful-steering assumptions, we prove a one-round decrease in target KL divergence for sufficiently small steering strength. Under greedy decoding across eight models and six benchmarks, EMO improves the zero-shot baseline in 39 of 40 aggregated settings, averaging $+9.10$ percentage points. With the same judge, it adds $+1.53$ points on average over context-only optimization (TextGrad), with consistent MMLU-Pro gains across models. At matched target-model generation budgets, accuracy advantages over best-of-$N$ are strongest on Llama, while oracle coverage is higher in most comparisons despite fewer candidates. The coverage--selection gap motivates allocating test-time computation jointly to candidate generation and verification.
benchmark - arxiv:2610.22158 · cs.CLPremiseSQL: Evaluating Premise Critique and Abstention in Text-to-SQL SystemsShuo Lin, Jinzhe Li, Chenlu Guo, Yanhan Li +2
Recent advances in large language models (LLMs) have substantially improved Text-to-SQL performance, yet existing NL2SQL systems largely assume that user queries are valid and answerable. As a result, models often generate executable SQL even for queries with flawed premises. We introduce premise critique ability in NL2SQL, namely the ability to recognize flawed queries and abstain from generating misleading SQL. To evaluate this capability, we propose PremiseSQL, a benchmark containing 1,800 flawed NL2SQL queries across four categories of premise failures, along with their 1,800 original queries. We further design an evaluation framework to measure intrinsic and prompt-induced premise-aware behavior across three prompting settings. Experiments on frontier LLMs show that current models are strongly premise-trusting by default and rarely critique flawed premises. Stronger critique-oriented prompting improves detection and abstention on flawed queries, but also increases false invalid judgments and suppresses SQL generation on originals. The code is available at https://github.com/vegeet/PremiseSQL.
benchmarkevaluation framework - arxiv:2610.22157 · cs.LGThe importance of tensor handling and implementation of physics-informed DeepONetsNahum Banks, Gustavo A. M. de Almeida, Sergio Maldonado
Physics-informed neural networks (PINNs) reduce or eliminate the need for large labelled datasets by enforcing governing equations during training. Physics-informed Deep Operator Networks (PI-DeepONets) extend this approach to learning solution operators, enabling rapid prediction across families of problems. Despite these advantages, we show that their application to higher-dimensional problems remains limited by the computational cost of training. To address this issue, we propose a modified PI-DeepONet architecture that reduces this cost by removing redundancy in how the Branch network processes input functions across collocation points. The proposed formulation is mathematically equivalent to the original architecture but reduces input tensor sizes and the number of Branch network evaluations. We analyse the computational complexity of both architectures and derive theoretical scaling laws for training time with respect to the number of collocation points and input functions. The proposed architecture is evaluated on 1D, 2D, and 3D linear advection problems. In all cases, it achieves comparable predictive accuracy to the original PI-DeepONet while consistently reducing training time. In 2D experiments, a maximum speed-up of $3.75\times$ is observed, with empirical training-time scaling agreeing with the predicted linear dependence on collocation points and input functions. Furthermore, training a 3D model on an NVIDIA A100 GPU is only feasible using the proposed architecture, as the original formulation exceeds memory limits. These results demonstrate that careful consideration of how PI-DeepONets are implemented can yield substantial gains, enhancing the practical performance of PI-DeepONets and enabling their application to otherwise intractable physics-informed learning problems.
memory - arxiv:2610.22152 · cs.LGSpecification Gaming and a Constrained Reinforcement Learning Mitigation in a Cleaning GridworldHilda Adwubi Osei
Reinforcement learning agents optimize exactly the reward signal they are given, which can diverge sharply from an intended true objective when that signal is only a proxy for it; this failure mode is known as specification gaming. Existing evidence for this failure mode is largely observational, with comparatively few demonstrations pairing a controlled, deliberately constructed proxy/true gap with a mitigation that carries a verifiable guarantee rather than only an empirical improvement. We construct a minimal gridworld in which a cleaning-robot agent can either genuinely remove dirt or conceal it from a cheap sensor, and show that a standard deep Q-network trained only on the sensor's proxy reward converges to a policy that games it completely, reaching near-maximal proxy reward while never achieving the true objective. We then introduce a constrained reinforcement learning fix, penalizing the specific behaviors through which the agent exploits the gap, and show that it eliminates the gaming behavior from the final policy while substantially recovering true-objective performance. We further derive and empirically verify an exact accounting identity relating the proxy reward, true reward, and our cost signal for this environment, showing that the cost signal measures the proxy/true gap exactly rather than merely tracking it, and yields a deterministic guarantee under a hard-constrained variant of the method.
agent - arxiv:2610.22150 · cs.AIRAG-RST: Retrieved Exemplars as In-Context Discourse Schema for RST ParsingMohammadreza Sediqin, Shlomo Engelson Argamon
In Rhetorical Structure Theory (RST), each edge of a discourse tree carries a nuclearity label, marking which span holds the central content, and a relation drawn from a closed inventory of functional categories. Neither label is a surface property: the same two clauses can be Elaboration or Background depending on rhetorical intent, and the boundary between them is fixed by annotation convention rather than by lexical cues. Predicting these labels therefore requires a formal discourse schema, and exposure to discourse-rich pretraining text does not supply it: given only a target EDU pair, a strong LLM recovers nuclearity at 0.794 F1 and relation at 0.738 F1. We propose RAG-RST, a retrieval-augmented framework that supplies this schema as in-context demonstration rather than through parameter updates, raising nuclearity to 0.846 F1 and relation to 0.788 F1 with no training. Two configurations are compared: a precomputed strategy that labels all EDU pairs before parsing, and a dynamic strategy that retrieves at each reduce step, so that the exemplars reaching the generator depend on the reductions the parser has already committed to. Both outperform existing RST parsers across span, nuclearity, relation, and full tree metrics, with the dynamic variant consistently strongest. Two properties of the retrieved schema emerge: its utility is bounded in number, saturating and then declining beyond fifty exemplars, and it depends on structural position. Both indicate that a formal linguistic schema can be supplied to a foundation model as a demonstration rather than trained into its parameters.
retrieval-augmented - arxiv:2610.22146 · cs.ROFastVLA: Task-Conditioned and Temporally Adaptive Visual Token Compression for Efficient Vision-Language-Action ModelsJiuyi Xu, Chengyuan Qian, Zhengzhong Tu, Yang Sui +1
Vision-Language-Action (VLA) models present a promising paradigm for general-purpose robot manipulation, but closed-loop deployment is often limited by inference efficiency: at each timestep, high-resolution, often multi-view images are encoded into hundreds of visual tokens that dominate transformer computation and latency. Token compression via attention is a natural solution but is challenging for VLA control because (1) decision-critical tokens are determined by the task, not generic visual saliency, and (2) the token budget varies substantially across steps, so static metrics and budgets waste computation or discard control-critical tokens. We introduce FastVLA, a training-free, task-aware, and temporally adaptive visual token compression framework. At each step, FastVLA scores token importance via task-to-vision attention and adaptively sets the compression rate from the importance-distribution uniformity, measured by an inverse-Simpson metric with online sliding-window calibration. We evaluate FastVLA on OpenVLA, OpenVLA-OFT, and CogACT across LIBERO and SIMPLER, and on a physical Franka Research 3 robot. In simulation, it reduces LLM-backbone FLOPs by 35-59\% and action latency by 9-32\% (a 1.09$\times$-1.46$\times$ speedup) while maintaining task success, achieving a comparable or better efficiency-performance trade-off than existing plug-and-play token compression/pruning/reuse baselines. On the real robot, it cuts end-to-end latency by 20\% (1.25$\times$ speedup) at comparable success. Code is available at https://github.com/jiuyixu25/FastVLA.
vision-language-actionvlamanipulationopenvlacogactlibero - arxiv:2610.22139 · cs.AIAre You Sure Your LLM Is Sure? Repairing and Auditing ConfidenceArijit Dey, Bhartendu Kumar, Ramya Hebbalaguppe, Khuushi Maheshwari
Confidence scores determine whether an LLM answer is accepted, reviewed, or regenerated, yet scalar recalibration can only correct errors visible through the score itself. We introduce RADAR (Repair, Adopt, and Diagnose After Recalibration), a reliability framework that separates confidence repair from deciding whether to deploy the repair and auditing the errors that remain. Under squared loss, we decompose prediction risk into outcome noise, information hidden by the score, and mapping error; restricting repair to order-preserving maps additionally introduces a pooling gap. For repair, we propose Probe-Anchored Isotonic Calibration (PAIC), which learns a monotone map from correctness-probe predictions on unlabeled target prompts. PAIC's excess risk depends on the probe's score-conditional bias and fitting error rather than its overall accuracy. For adoption, we derive a paired procedure with finite-sample control of false improvements. For auditing, we develop a prompt-clustered kernel test that detects feature-aligned residual errors while accounting for dependent answers from the same prompt. Across multiple LLM checkpoints, benchmarks, and distribution shifts, richer predictors can improve confidence estimation, but projecting their information back to a scalar score can lose these gains, and lower prediction risk need not yield better review decisions. Across 47,400 replayed raw-score allocations, paired betting adopts candidates with lower held-out test risk in 42.95% of allocations, versus 15.09% for one-sided empirical Bernstein. RADAR thus provides a principled workflow to repair what a scalar score can express, adopt a repair only when supported by evidence, and diagnose systematic errors that recalibration leaves behind. Code and data will be released upon acceptance.
benchmark - arxiv:2610.22138 · cs.AIBelieved but Not Said: Measuring Hidden Belief Collapse in Multi-Agent LLM DiscussionsLiangjie Huang, Henry Peng Zou, Wei-Chieh Huang, Wooseong Yang +2
LLM systems increasingly rely on groups of agents that discuss, critique, and check one another, assuming that peers will catch individual errors. We test this assumption against a minimal adversary. In a ten-agent factual discussion, one or three adversarial agents repeatedly assert a plausible false answer. After each round we probe each benign agent's private answer through a separate call whose response is hidden from its peers, and we separately label its public stance toward the claim. Across six models and three factual-QA datasets, groups remain stable without an attacker: only 0.6% of benign agents finish with the false answer. A single adversarial agent increases this rate to 44.8% and three to 60.5%. Studies of conformity typically report public agreement without private change; we find the reverse. With three adversarial agents present, benign agents hold the false answer in 51.6% of exposed turns but publicly endorse it in only 19.7%. Much of the influence therefore remains invisible in the discussion, and the false answers persist when agents are asked for their own views. Agents that retain the true answer often fail to defend it; once adopted, the false answer spreads among benign agents, and baseline knowledge does not predict robustness.
agentmulti-agent - arxiv:2610.22131 · cs.CLIntern-Decision: Efficient Multimodal Decision Making with Autoregressive Masked Language ModelingTong Zhu, Zhe Xu, Zefeng He, Yunkai Chai +6
Interactive agents repeatedly choose among a finite set of actions, labels, or ordinal outcomes. Generating a free-form response for each choice adds latency and complicates the interface between a model and its environment. We present Intern-Decision, a family of 0.8B, 2B, and 4B multimodal decision models that return structured choices and candidate probabilities from text and optional images. Our autoregressive masked language modeling objective places one decision placeholder per field in a complete assistant JSON skeleton. Ground-truth symbols appear only in supervision, while all fields are evaluated in a single causal forward pass using the backbone's native language-model head. The 4B model achieves a 90.02% unweighted average accuracy across seven evaluation suites, compared with 88.74% for Jev v1.13.0. For a fixed 289-token request with three fields on one RTX 4090, mean latency is 44.16 ms for the 4B model and approximately 33--34 ms for the smaller models. On a separate 96-case known-distribution pilot, temperature scaling reduces expected multiclass Brier from 0.628 to 0.550 and expected calibration error from 0.213 to 0.089. A confidence-based handoff study on 111 public Jevbench-Hard items improves accuracy from 82/111 (73.87%) to 97/111 (87.39%) by routing 42 items to Atria-Dawn-Preview. We also describe qualitative game, visual interface, and browser demonstrations. These results support compact decision models as a practical component of larger agent systems.
agentagent system - arxiv:2610.22130 · cs.AIDropKV: Decoupling Residual-Output Perturbation for Near-Optimal KV-Cache EvictionAozhong Zhang, Selcuk Gurses, Yanxia Deng, Naigang Wang +7
Key-value (KV) cache management has become a critical bottleneck for scaling Large Language Models (LLMs) to long-context scenarios due to linear memory growth. While recent eviction methods have moved beyond simple attention-score heuristics to incorporate value representations, they typically face a fundamental trade-off: the underlying optimization for minimizing output perturbation is an NP-hard combinatorial problem, leading to mathematically suboptimal heuristics or implementations that are difficult to kernelize. In this work, we propose DropKV, a simple principled eviction framework based on Decoupled Residual-Output Perturbation. By decoupling the joint eviction decision into independent per-token scoring, DropKV circumvents combinatorial intractability and admits a provable constant-factor approximation guarantee under a cone condition that we empirically verify on three long-context LLMs. To ensure practical viability, we provide a high-performance implementation using fused Triton kernels that avoid materializing the full attention matrix, delivering a 19.8x scoring-kernel speedup that translates to up to a 9.9% matched-batch end-to-end prefill speedup. Extensive evaluations on the RULER and LongBench benchmarks demonstrate that DropKV consistently outperforms state-of-the-art baselines at equivalent cache budgets, achieving the lowest inference latency among eviction methods while also remaining faster than the dynamic cache baseline.
memorylong-contextbenchmark - arxiv:2610.22126 · cs.LGCausal Integrated Gradients: Scalable Feature Attribution Along Causal PathsJoran Michiels, Anis Ismail, Maarten De Vos, Alejandro Sifrim
Existing feature attribution methods that account for feature dependencies do not scale to high dimensions. We propose causal integrated gradients (CIG) and causal expected gradients (CEG), which extend integrated gradients to incorporate prior causal structure among input features while retaining its computational efficiency and, under conditions on the causal graph, its axiomatic guarantees. In many settings, some causal structure between features is already known, for example, that a transcription factor causally regulates a target gene's expression. Yet existing gradient-based methods cannot exploit this structure, instead crediting a feature's downstream effects over its true upstream cause, such as ranking the target gene above the transcription factor that regulates it. Our key idea is to embed this causal structure by routing the integration path along the causal graph. The method thereby decomposes each feature's attribution into a direct effect and effects mediated through its causal descendants. We evaluate CIG in three settings of increasing realism. First, on a synthetic benchmark with a known causal model, CIG's attributions best match the ground-truth causal effect from Pearl's do-calculus, matching the accuracy of causal Shapley values at a fraction of the computational cost. Second, on simulated single-cell gene expression data with a known gene regulatory network (Dyngen), CIG most accurately recovers the transcription factors driving the simulation. Third, on real interferon-$β$ stimulation data, where only an approximate causal graph is available, CIG better prioritizes the transcription factors driving the measured response than DeepSHAP and integrated gradients. Together, these results show that incorporating causal structure into gradient-based attributions recovers upstream drivers that existing methods miss, without sacrificing scalability.
benchmark - arxiv:2610.22124 · cs.AIPROBE: Learning to Audit Policy Compliance in Tool-Using LLM AgentsKshitij Mishra, Abhijith Sharma, Nils Lukas, Salem Lahlou
Modern agentic systems extend language models with external tools, enabling them to complete complex tasks through multi-turn interactions. As these agents are deployed in consequential settings, they are often constrained by policies governing pricing, privacy, confirmation, safety, and other requirements. However, existing evaluation methods leave a critical gap: capability benchmarks measure what agents can accomplish, while content-safety red-teaming measures harmful generations, but neither tests whether agents comply with predefined policies during otherwise valid tool use. Such policy violations are difficult to audit because their effects may emerge only after several turns, span multiple tool calls, or surface in downstream systems. We propose PROBE (Policy Rule Observation via Behavioral Exploration), an algorithm for training LLM-based auditors that interact with deployed agents as realistic users and probe them for policy violations through multi-turn, tool-mediated dialogue. We formalize agent auditing as a two-player Markov game and propose a composite reward that balances violation discovery, behavioral plausibility, conversation progress, and policy diversity. To study training objectives and reward design, we compare GRPO, DPO, and GFlowNet-based optimization in a controlled setting. Across extensive experiments, auditors trained with PROBE uncover substantially broader and higher-yield policy violations than same-architecture zero-shot adversaries, even when the auditor is significantly smaller than the target agent. Our results show that multi-turn behavioral auditing is a distinct and necessary axis of agent evaluation, complementing capability and safety benchmarks. PROBE offers a scalable path toward automated policy auditing for real-world agentic systems and suggests a co-evolutionary loop in which auditors expose failures and agents improve from auditor feedback.
agentllm agentagentictool usebenchmark - arxiv:2610.22119 · cs.ROCompliant GMT: Making General Motion Tracking Compliant via Residual Reference AdaptationVenki Parsuram, Michelle Wang, Jin-Hyuk Hong, Pulkit Agrawal
We propose Compliant GMT (C-GMT), a modular framework for full-body compliant responses in general motion tracking. As humanoid robots move into real world deployment, they not only need to perform any motion but also respond safely and robustly to interaction forces. Existing general motion tracking policies can imitate diverse motions but are trained to be stiff, leading to unexpected and unsafe behaviors under disturbance. Prior work in compliant humanoid control trades off compliance and generality. There exist examples of full-body compliant responses, but these are heavily curated to single-reference policies. On the other hand, general motion trackers (GMT) are trained to rigidly imitate massive motion capture datasets. Our key insight is to leverage the generality of pretrained GMTs for motion tracking while introducing compliance from a residual adapter that conditions on a history of proprioception. This enables our approach, C-GMT, to achieve both general motion tracking and whole-body compliant responses. Compared to state-of-the-art general motion trackers and single-motion compliant policies, C-GMT demonstrates competitive performance in both motion tracking accuracy and compliance adherence. We further demonstrate improved safety, robustness, and error tolerance in simulation and real-world G1 experiments.
humanoid - arxiv:2610.22108 · cs.ROCF-Splat: Closed-Form Triangle Splatting with Mesh PriorsYixin Fang, Liang Zhao
Splatting-based radiance-field representations have recently gained popularity in robotic perception, enabling high-quality novel-view synthesis (NVS) through explicit scene primitives. However, existing splatting methods typically rely on iterative nonlinear optimization that jointly optimizes millions of coupled primitive parameters, resulting in substantial computation time, peak GPU memory consumption, and susceptibility to local minima. In this paper, we propose CF-Splat, a novel closed-form triangle-splatting framework for efficient scene appearance estimation over a known mesh either pre-existing or provided by upstream mapping or reconstruction pipelines. Our key insight is that, by properly formulating the problem over the given mesh geometry and calibrated cameras, the rendered pixel colors can be expressed linearly with respect to the per-vertex spherical harmonics (SH) coefficients, allowing multi-view scene appearance estimation to be formulated as a global linear least-squares (LLS) problem without the need of iterative nonlinear optimization. To guarantee a well-posed solution under uneven multi-view observation coverage, we introduce adaptive SH regularization that constrains insufficiently supported high-order SH coefficients and yields a unique global minimizer. To ensure the scalability to large-scale scenes, we develop a triangle-structured matrix-free preconditioned conjugate-gradient solver that avoids explicitly constructing the large sparse normal system while producing the same LLS solution. Experiments on diverse real-world datasets demonstrate that CF-Splat achieves competitive or superior NVS quality compared with state-of-the-art splatting methods. While, CF-Splat requires approximately 27.5-378.0 times less computation time and 1.8 times lower peak GPU memory on average than 3DGS across scenes of different scales on a consumer-grade laptop.
memory - arxiv:2610.22094 · cs.CVSCCL-DT: Semantic-Conflict-Aware Contrastive Learning with Dual-Rule Teacher Distillation for Zero-Shot Chest X-Ray DiagnosisTing Liu, Haoran Lai, HaobinChen, Ming Liu +1
Contrastive vision-language pre-training supports zero-shot chest X-ray diagnosis, but standard objectives treat every off-diagonal image-report pair in a batch as a negative. In radiology, recurring findings across patients mean that instance mismatch need not imply semantic incompatibility, creating false-negative supervision that we term a semantic conflict. A training-trace audit shows that pair-observations selected by the online mask used during training are enriched for report-derived finding agreement, although this proxy analysis does not establish clinical equivalence. Fixed thresholds can become miscalibrated as similarity distributions evolve, while soft reweighting leaves residual repulsive gradients. We therefore propose Semantic-Conflict-Aware Contrastive Learning (SCCL), which tracks modality-specific similarity distributions with exponential-moving-average (EMA) statistics, derives adaptive thresholds, and hard-masks flagged candidates before softmax normalization. Inclusive (OR) and consensus (AND) rules define two SCCL teachers, whose bidirectional cross-modal rankings are consolidated through standard distillation into a single deployable student, SCCL-DT. Across five public chest X-ray benchmarks, SCCL-DT achieves a mean zero-shot AUC of $0.850$, compared with $0.837$ for the strongest evaluated baseline in our comparison. Code will be made publicly available soon.
benchmark - arxiv:2610.22092 · cs.ROOnboard Lightweight LLM Reasoning for Human-Aware Safe UAV NavigationSiwei Cai, Yuwei Wu, Vijay Kumar, Lifeng Zhou
Aerial robots operating near people must treat a pedestrian, a parked car, and a wall differently, yet conventional perception--planning stacks reduce all of them to occupied geometry---a \emph{human-aware safety gap} in which category-specific clearance is unavailable to the planner and temporary occlusions remove pedestrian-specific constraints from its map. We present a modular, fully onboard pipeline that closes this gap by turning visual observations into context-dependent safety constraints for real-time UAV navigation. A shared BLIP encoder drives scene captioning and text-conditioned segmentation, which is fused with depth to localize obstacles in 3D. Obstacle-specific risk is inferred by a lightweight LLM (Qwen-2.5-0.5B) stabilized with retrieval-augmented generation and grammar-constrained decoding that restricts its output to a planner-compatible JSON schema. The LLM reasons once per object category rather than per frame: its verdicts are cached per class, so previously seen categories are resolved without inference and new ones trigger a fresh retrieval-grounded query; detections are additionally given persistent identities and held in the map across occlusions. The perception pipeline runs at 10\,Hz on a Jetson Orin NX, while fallback LLM queries take approximately 0.8\,s and remain off the control-critical path. In Gazebo, low-altitude urban navigation among pedestrians and vehicles shows a roughly $3\times$ reduction in collision events over a geometric planner, and semantics-aware landing among moving actors raises success from 34\% to 96\%. The 224-pixel encoder input limits reliable small-object perception at range; conservative pedestrian margins and short-term persistence mitigate this limitation during navigation.
retrieval-augmented - arxiv:2610.22086 · cs.AIMulti-SFT: Permissively-Licensed Multilingual Instruction-FollowingDaniela Baiamonte, Kamyar Zeinalipour, Elena Fano, Stefano Simonazzi +1
Supervised fine-tuning datasets remain overwhelmingly English-centric and commercially restricted, as their content is frequently generated by proprietary models whose terms of service forbid training competing systems. We introduce Multi-SFT, a commercially permissive collection of 502,818 instruction-response pairs across five European languages, built via translation and regeneration pipelines for tasks where translation is insufficient. Human evaluation yields inter-annotator agreement of at least 0.87 across all datasets and languages. Fine-tuning nine models from three different families (EuroLLM, OLMo 2, Qwen3; 1B-32B) and evaluating them on 13 benchmarks shows that Multi-SFT alone is competitive with larger augmented mixes and official instruction-tuned releases, with decisive advantages on safety tasks.
benchmark - arxiv:2610.22080 · cs.AILearning from Unverified Peers via Shared Residual DistillationFeiyang Chen, Runze Ma, Zichen Wang, Jiong Lou +3
A language model can generate several solutions to the same problem, some leading to correct answers and others to incorrect ones. Many post-training methods turn these trajectories into learning signals using supervision from reference answers, verifiable feedback, or external teachers. Without such supervision, learning from the model's own generations becomes less straightforward, as their correctness cannot be reliably assessed. An incorrect solution may nevertheless contain useful intermediate reasoning, so its value as supervision need not be determined solely by its final answer. Can a model improve by learning from its unverified peer solutions without first determining which are correct? We propose Shared Residual Distillation (SRD), an on-policy self-distillation method that learns from the model's own unverified peer trajectories. SRD compares how peers change the prediction distribution relative to a control, retaining adjustments that agree in direction and are conservative in magnitude. This yields supervision without treating any complete solution as correct, and the trained model requires no peers at inference time. Starting from the corresponding base models, SRD raises mean AIME24--26 avg@12 from 11.02 to 16.20 on Qwen3-1.7B, from 19.07 to 45.46 on Qwen3-4B, and from 21.39 to 54.44 on Qwen3-8B. Controlled ablations support the value of the shared-residual construction, and we observe no degradation relative to the base models on two general-capability benchmarks.
post-trainingbenchmark - arxiv:2610.22077 · cs.LGCan AI Forecast the Extremes? Evaluating Probabilistic Weather Models on Tropical CyclonesNathan Chalumeau, Xavier Renard, Remi Meynadier, Hugo Rakotoarimanga +2
Weather prediction is undergoing an important shift, with data-driven AI models outperforming leading numerical weather prediction systems on average at a fraction of the computational cost. Yet, whether this success extends to extreme events remains contested. Earlier evaluations have exposed important shortcomings, but have largely focused on deterministic models and metrics, leaving the capabilities of the latest probabilistic models less understood. We conduct an extensive benchmark of tropical cyclone forecasting, comparing GenCast, FourCastNet 3, and FGN with ECMWF's operational physics-based ensemble, IFS-ENS. We develop a suite of probabilistic metrics and diagnostic analyses to assess cyclone tracking, intensity biases, uncertainty, and the wind-pressure relationship that characterizes storm structure. We find that, despite remaining limitations in intensity and ensemble spread, leading AI models substantially outperform IFS-ENS in cyclone tracking: FGN achieves better probabilistic location scores for 90% of evaluated storms in 2025. Further analyses reveal model-specific failure modes and suggest a lingering influence of pretraining data, pointing to opportunities for targeted improvement. Nonetheless, these results suggest that AI models are capable of capturing physical relationships that remain predictive well beyond typical weather conditions, motivating further investigation into how these capabilities emerge and how to strengthen them.
benchmark - arxiv:2610.22075 · cs.CVRetrofitting Recurrent Vision-Language Models from Pretrained BackbonesMadhav Iyengar, Andreas Geiger, Hilde Kühne
Recent work shows that depth recurrence can be retrofitted into pretrained language models. In this work, we adapt this paradigm for visual understanding, using depth recurrence to reduce the parameter footprint of a pretrained vision-language model (VLM). To this end, we prune layers from the language backbone of the VLM and loop over a subset of the remaining layers. At each step of the loop, a linear adapter merges the block's output with the token representations from the preceding layers and projects the fused representation back into the block. As the pruning impacts the original model performance, we propose a three-stage recovery training: first, we leverage a language-only pretraining to recover the performance of the language backbone, followed by an alignment of the vision-language projector, and a full fine-tuning of the model via visual instruction tuning. In the first and the last stage, we distill the outputs of the pretrained non-recurrent model into the recurrent model, which further improves the recovery. We evaluate the proposed approach on the TinyLLaVA-1.4B backbone without introducing any new vision-language data during retrofitting. Our experiments demonstrate that the retrofitted recurrent model outperforms the unpruned baseline trained with the same recovery data while storing 23.3% fewer language parameters. We further show that the proposed retrofitting recurrence outperforms the setup of converting a recurrent LM into a VLM on the same data. Additionally, our ablations show how design choices, such as the adapter between recurrences and a larger recurrent block, lead to the improved performance of our approach. Finally, we find that our proposed model gains on reasoning-oriented benchmarks, whereas its performance on basic visual understanding benchmarks remains on par with the baseline.
benchmark - arxiv:2610.22070 · cs.LGMemory as Weights: Internalizing Long-Term History for Streaming VideosZhuoming Liu, Jinhong Lin, Lin Zhang, Yiwu Zhong +1
Streaming vision-language models (VLMs) must continuously integrate information from the past while remaining responsive to current observations and user queries. Existing memory mechanisms typically compress, retrieve, or cache historical observations and reintroduce them into the input context, forcing historical memory and current observations to compete for limited context. We propose MaLoW, a streaming memory framework that instead internalizes historical context into evolving lightweight parameter updates. MaLoW uses a recurrent HyperNetwork to incrementally consolidate incoming video observations into fixed-size LoRA weights, allowing historical information to be carried forward without revisiting the full video stream. In doing so, MaLoW represents history as weights, while keeping the current observation as input. To support training, we further construct MaLoW-Instruct, a large-scale streaming video dataset containing 381K question-answer pairs over 56K videos. We conduct extensive experiments to demonstrate that MaLoW achieves strong performance on OVO-Bench, StreamingBench, and StreamArena. On the backward split of OVO-Bench, MaLoW outperforms prior memory approaches by 3.9% to 11.8% across different video backbones while incurring only modest latency overhead. Beyond streaming video understanding, MaLoW also generalizes to vision-language-action models, improving the action success rate on Simpler WidowX by 9.3%.
vision-language-actionmemory - arxiv:2610.22068 · cs.AIWhen Reasoning Is Not Enough: Proactive Inquiry for Cross-Modal Safety in Multimodal Large Language ModelsWei Cai, Jian Zhao, Yuchen Yuan, Xuelong Li
Multimodal large language models (MLLMs) can act unsafely when individually benign images and texts are interpreted jointly. Beyond implicit reasoning risk, where reasoning drifts toward harm, we identify intent-level risk, where cross-modal ambiguity silently commits the model to a hazardous premise that path-level optimization alone cannot correct. We propose SAFE-PRO, which uses a Risk-Conditioned Uncertainty (RCU) signal, combining predictive entropy with a learned cross-modal risk score, to decide when to ask safety-grounded clarifying questions and where to apply path optimization, under a reward that prioritizes safety and penalizes over-refusal. We also build SSUI-Dialog, a multimodal multi-turn dataset with dual-intent supervision, and theoretically analyze gradient variance and unsafe-completion risk under RCU. Across seven benchmarks and two base MLLMs, SAFE-PRO consistently improves safety over SRPO, roughly halves its over-refusal, and surpasses GPT-4o and Claude-3.5-Sonnet in average single-turn safety.
benchmark - arxiv:2610.22067 · cs.CLBeyond Context Length: Characterising the Difficulty of Long-Context ReasoningShiwen Qin, Cameron Barker, Henry Gouk, Elliot J. Crowley +3
Long-context language models support ever longer context windows, and evaluations typically treat context length as the main indicator of difficulty. However, context length alone cannot explain why models fail on long-context tasks that require reasoning over complex information. Contexts of equal length can pose very different challenges: a task may be solvable from a single local passage, or it may require integrating evidence distributed across the full context. We quantitatively characterise this difficulty through three properties: evidence scope, how much necessary information must be found; evidence dispersion, how widely that information is scattered within the context; and reasoning depth, how many dependent steps are needed to derive an answer from it. We construct LURCH, a multi-domain question-answering benchmark of 6,550 instances with supporting evidence and reasoning graphs, enabling all three properties to be quantified. Evaluations of proprietary and open-weight models show that: (1) at fixed context length, all our three properties are strongly negatively correlated with model performance; (2) as context length increases, performance degrades sharply on samples with dispersed evidence, whereas it remains relatively stable on samples with localised evidence; and (3) difficulty compounds across properties, as questions with moderately high values on all three are harder than those with an extreme value on just one. These findings establish evidence scope, dispersion and reasoning depth as measures of long-context reasoning difficulty that complement context length.
long-contextbenchmark - arxiv:2610.22066 · cs.AIPrism-1: Position- and Rival-Independent Scoring for System One ModelsYao Long Teng, Jianguo Huang, Yanchen Deng, Jiayi Cai +1
Given a state and a set of options, such as intents, tools or interface elements, System One decision models return a distribution over the options in one forward pass, without generated reasoning. Web systems increasingly build them by reading decisions from language models. Two axioms of rational choice are natural requirements for such decisions, because in web systems the listing and the option set are determined by other components and can change or be manipulated: the order in which the options are listed should not matter, and adding an option should not change the relative odds of the others (Luce's choice axiom). We test the standard reads, namely single-choice question (SCQ) prompting, with shuffled-order training, order averaging or sorted options, and invariant set architectures, and find that each violates at least one axiom: one shuffle changes 12-20% of SCQ answers, sorting only freezes an arbitrary convention, and a well-chosen added option steers 4-34% of decisions. By Luce's theorem, a read satisfies both axioms only if it is a softmax over scores that each depend on one option alone, so in a transformer they hold by construction only if no token that scores an option sees another. We implement this as Prism-1 (Position- and Rival-Independent Scoring for System One Models): every option is read in its own block at shared positions, with the model's usual left-to-right attention, in one forward pass. On five benchmark sets with Qwen3-8B and Gemma 4 12B, SCQ prompting trained on the same data still changes 7% of its answers under one shuffle, and an added irrelevant option steers 12% of its decisions. Prism-1 stays within one point of that baseline's accuracy and is 2-4 points more accurate once only answers that survive a reordering are counted, while no shuffle changes any of its answers and no added irrelevant option moves its choice among the others.
benchmark - arxiv:2610.22061 · cs.AIFAOPs: Normal Forms and an Algebra for AI Agent WorkflowsAli Mohammadi Shanghooshabad
Organizations increasingly encode operational procedures as multi-step agent skills, here called Agent Operating Procedures (AOPs), that must be traceable, auditable, reproducible, access-controlled, and bounded in time and cost, requirements that regulated sectors make mandatory. Yet AOPs are typically kept as coarse-grained files, in which steps bundle actions, copied procedures drift apart, any step may reach any tool, and refactoring goes unchecked. We present Functional Agent Operating Procedures (FAOP), a formal theory that brings database-grade discipline to AI agent workflows. FAOP recasts each workflow as a typed function over scoped state and subjects an entire library to four progressive normal forms: Decomposed (FNF-1) splits a workflow into atomic steps, Typed (FNF-2) makes its data flow explicit and type-checked, Factored (FNF-3) gives each shared procedure a single canonical definition so a change propagates instead of drifting, and Scoped (FNF-4) confines every step to least-privilege tool access. Once normalized, every atomic step is a separate entity with a declared contract and can be type-checked, scope-confined, bounded, tested, traced, and audited on its own. A FAOP algebra of provably sound equalities turns refactoring and parallelization into compiler rewrites; orthogonal extensions add verified termination (+V), optimization (+O), and time and cost bounds (+B); and an architectural-level classifier (0-3) determines whether a workflow needs AI at all. Together, these make integrity, isolation, typing, termination, cost, and optimization compile-time concerns across the whole library, not runtime hopes. We implement the theory in SkillBase, an AOP management system (what PostgreSQL is to data, SkillBase is to AOPs), together with a separate executor that enforces these guarantees at run time. A worked financial workflow illustrates the approach end to end.
agentai agent - arxiv:2610.22060 · cs.CVFeaTok: Feature-Loss Tokenization for Image GenerationZhang Xiaofeng, Samuel Lavoie, Nicolas Beltran-Velez, Jianhao Yuan +5
Discrete image generators learn the distribution of tokens produced by an image tokenizer. These token representations should be easy for generators to model, not merely enable faithful image reconstruction. Yet these tokenizers are commonly trained to reconstruct pixels, which prioritizes low-level fidelity and may disrupt representation structure useful for generation. Motivated by feature-predictive self-supervised learning, we introduce feature-loss tokenization (FeaTok), which trains discrete tokens by reconstructing pretrained features rather than pixels. Ablations across SigLIP, SigLIP2, DINOv2, and DINOv3 show better generative performance with FeaTok. In class-conditional generation, FeaTok advances the quality-compute Pareto frontier, reducing gFDr6 at 256x256 by up to 29 percent over the strongest discrete baselines and outperforming the strongest unguided continuous baseline (RAEv2) with 38 times less sampling compute. FeaTok also achieves higher human-preference reward-model scores than prior art without reward-based training and surpasses continuous and discrete text-conditional baselines on three benchmarks with only 500 fine-tuning updates after pre-training.
benchmark - arxiv:2610.22051 · cs.LGRethinking Successive Rejects AlgorithmsMohamed Khair, Junpei Komiyama
In the best-arm identification problem, an algorithm adaptively selects arms to find the arm of the largest mean reward. Successive Rejects (SR) is a simple fixed-budget best-arm identification algorithm that eliminates one candidate arm at each phase. At first glance, SR seems to have limited adaptivity, since it follows predetermined phase lengths. However, for Gaussian arms with common variance, SR's worst-instance error exponent normalized by the popular complexity measure $H_1(P)$ is optimal up to a constant factor, which partly explains why SR's performance is competitive compared with state-of-the-art adaptive sampling algorithms. In this paper, we extend the framework of SR by optimizing its predefined schedule of phase lengths. We formulate phase-schedule design as an optimization problem for the normalized error exponent. Then, we introduce a version of SR with the optimized schedule, which we call $H_1$-balanced Successive Rejects ($H_1$-SR). This schedule is optimal within the class of SR-type algorithms with deterministic equal-count schedules. For Gaussian arms with common variance, $H_1$-SR achieves at least $11/12$ of the optimal normalized rate over all adaptive algorithms for every $K\ge2$, and this approximation ratio converges to one as $1-O(1/\log K)$ as $K\to\infty$. This is a significant improvement over original SR, whose corresponding fraction of the optimal normalized rate converges to $1/4$. Finally, experiments on ten synthetic instances and two dataset-derived Gaussian benchmarks compare the optimized schedule with original SR at finite horizons. They show substantial gains on many instances while also exposing a regime in which more aggressive early elimination is costly.
benchmark - arxiv:2610.22043 · cs.LGMomentum, Commentary Tone, and Earnings Surprise in LLM Financial ForecastsAlina Khaybullina
Which inputs drive a language model's financial forecast? We construct synthetic earnings announcements that independently reverse 20-day momentum, non-informational commentary tone, and whether fixed actual EPS beats a manipulated consensus. Six instruction/chat checkpoint conversions from four families (0.49-7.61B base parameters) receive 2,016 prompts each under a direct-answer Q8_0 protocol. The design includes 24 numerical blocks, six commentary pairs, profits and losses, rising and falling annual EPS, and two surprise sizes. Both A/B answer assignments and U/D cover all 384 primary scenarios per conversion. We estimate matched changes in conditional answer-prefix scores, not calibrated market probabilities. Across these checkpoints, the primary EPS contrast ranges from -0.19 to 86.80 percentage points. The magnitude of the primary signed tone contrast exceeds that of EPS in 5 of six models, and momentum exceeds it in 5. Natural-word numerical diagnostic accuracy ranges from 51.0 to 95.8 percent. Answer-encoding, wording, context, and conversion controls qualify these results. The experiment measures observable input reliance and numerical recognition; it establishes neither an internal reasoning mechanism nor realized return predictability. Public benchmark prompts, recorded scores, and an interactive recorded-results explorer accompany the study.
benchmark - arxiv:2610.22041 · cs.ROPlans Are Not Executions: Adapting Frozen World-Model Planners to Controller-Induced Execution ShiftDanqi Zhao, Keyu Chen, Wenchao Sun, Minglei Huang +5
World-model planners evaluate candidate trajectories by predicting their future consequences, implicitly relying on the plan-to-motion relationship learned during training. At deployment, a different downstream controller may execute the same planned trajectory differently, creating a planning-execution gap between the frozen model's imagined future and actual execution. A natural remedy is to replace the planned trajectory with its predicted or observed execution. However, we find that better execution fidelity does not necessarily yield better decisions: directly substituting predicted execution trajectories yields only a 0.048 PDMS improvement across 10 unseen controller configurations, far smaller than the gain from our proposed adaptation. This suggests a semantic mismatch between a planned trajectory, which serves as a reference for downstream control, and an executed trajectory, which already describes the resulting post-controller motion. We introduce ExecutionBridge, which infers controller-specific execution characteristics from planned-executed trajectory pairs and uses them to adapt the native ego latent rollout while keeping the original planned trajectory unchanged. The planner, world model, and reward head remain frozen. On NAVSIM, ExecutionBridge consistently improves planning across unseen controller configurations and multiple frozen world-model planners, with substantially larger gains in challenging driving scenes where execution mismatch more strongly affects candidate evaluation.
world model - arxiv:2610.22039 · cs.CLWEAVE: Weaving Real-World Code Evolution into Dependency-Structured Tasks for Benchmarking and Training Code AgentsWeihao Li, Zihang Xu, Zhongqiu Wang, Songwei Yu +6
Real-world software engineering rarely consists of isolated changes. Features, migrations, and systematic fixes often unfold across multiple pull requests (PRs), requiring coordinated modifications across a repository. Yet most executable repository-level benchmarks retain a single issue or PR as the task unit, while existing multi-PR settings rely on coarse aggregation and structure-agnostic evaluation. This mismatch obscures dependencies among changes and limits progress measurement and learning signals for long-horizon code agents. We introduce WEAVE, a framework that transforms verified PR histories into executable, dependency-structured tasks. WEAVE leverages repository-calibrated typed change loci to organize related PRs into bounded DAGs and instantiate each as a unified executable task. We further introduce Graph Credit, a deterministic, test-grounded metric that measures dependency-consistent partial progress and supports dense, verifiable rewards for reinforcement learning. Experiments on task construction, agent evaluation, supervised fine-tuning, and reinforcement learning show that WEAVE yields coherent executable tasks, enables more faithful progress measurement than flat scoring, and provides effective supervision and verifiable rewards for training code agents.
agentbenchmark - arxiv:2610.22035 · cs.LGETH: Data-Balanced Heterogeneous Pipeline Parallelism for Training Large ModelsMianjie Yu, Zhongmin Zhao, Zizhao Mo, Jianxiong Liao +8
Current heterogeneity-aware training systems for large models balance workloads by assigning unequal numbers of layers across pipeline stages, a strategy aimed at equalizing execution time through a unique pipeline configuration. However, this approach fails to address critical memory constraints and suffers from inherently coarse balancing granularity. In this paper, we propose Heterogeneous Pipeline Parallelism (HPP), a novel approach enabling efficient and fine-grained data balance across heterogeneous pipelines. To address HPP's key challenges, prohibitive search overhead for optimal configurations and high communication overhead from asymmetric All-Reduce, we design the ETH training system. ETH employs a hierarchical workload balance framework that jointly optimizes balance across heterogeneous pipelines and within individual pipeline stages. To mitigate asymmetric communication costs, ETH implements block-wise gradient aggregation. Furthermore, it integrates specialized optimizations that dramatically reduce HPP configuration search overhead. Evaluations demonstrate ETH achieves up to 2x training speedup while reducing configuration generation overhead by up to 56x compared to state-of-the-art systems. The code is available at https://github.com/YuMJie/ETH.
memory - arxiv:2610.22031 · cs.AIGraphACM: Graph-based Action Consistency Measurement for Uncertainty Quantification in LLM AgentsYaxin Ma, Wenbo Zhang, Satya Krishna Pothapragada, Alina Zare
Behavioral consistency across repeated executions provides an informative signal for assessing the reliability of large language model (LLM) agents. However, existing approaches typically summarize consistency with a single task-level score, without identifying where behavioral differences occur or whether actions are compared under equivalent decision contexts. We propose a graph-based framework for fine-grained behavioral consistency measurement and uncertainty quantification. Given multiple rollouts of the same task, we merge equivalent states into a directed, action-labeled graph, enabling action comparisons at shared decision states. We introduce node-level and global measures of action consistency, using the global consistency score as a proxy for uncertainty to predict task success or failure. Experiments across three LLMs on ALFWorld, AppWorld, and WebShop demonstrate that our measure is positively associated with task success and achieves an average AUROC of 0.811 for task-success prediction, outperforming the strongest behavioral-consistency baseline (0.729) and verbalized confidence (0.769). These results highlight the effectiveness of state-conditioned behavioral consistency and pinpoint where agent actions agree.
agentllm agent - arxiv:2610.22029 · cs.LGSelection Is Not Learning: Choosing Model FuturesHaoyi Yang, Kshitij Mishra, Nils Lukas, Salem Lahlou
Self-evolving language models train on tasks they generate, so every round must decide which updated model the next round inherits. Current loops such as R-Zero, Agent0, and Skill Self-Play make this decision through the task, admitting tasks that look learnable and inheriting every resulting update. We identify a selection-to-update gap in these loops, where a task that looks learnable need not produce a useful update. Most updates on verified tasks fail to make correct answers more likely, yet these loops inherit them all. HeirCommit freezes the updated model and uses fresh evaluation questions to decide whether to inherit it or retain the current model. EvoHeir separates parameter inheritance from experience inheritance, using current and archived experience to repair failures and preserve successes before final confirmation. Our analysis characterizes selection-induced optimism and decomposes regret into selection, absorption, and confirmation errors. HeirCommit cuts the harmful share of accepted updates by 68-83% relative to random inheritance at the same acceptance rate. Across four backbones, EvoHeir achieves the best mathematics and general-reasoning averages, raising the math average by 5.3% relative to the strongest baseline, Agent0. The framework lets useful experience persist even when the model that produced it is not inherited.
self-playself-evolving - arxiv:2610.22026 · cs.LGAugment the Cache, Replace the Activation: A 2x2 Study of Internal-State SharingAdam Deng, Shicheng Liu, Yada Zhu
Multi-agent systems built from frozen language models communicate through text, which discards the internal state the sender already computed and makes the receiver pay to rebuild it. Work on internal channels avoids that cost, framing the problem as transfer: agent A sends information to agent B. Every agent pair that must communicate needs its own transfer bridge, and every agent added to a system needs a new bridge fitted against each previous agent. We reframe this problem as sharing: an agent publishes its internal state once into a pool, and every peer reads that pool through a single mechanism fitted to no sender in particular, so adding an agent adds a publisher and no new bridge. Whereas the carrier of the internal state and the fusion mechanism are usually chosen independently in the literature, our main finding is that these two choices are coupled; the deciding factor is where the carrier enters the receiver's computation. Consumability is a property of the receiving computation rather than of the representation. We find that the key-value (KV) cache must be augmented rather than swapped in: the receiver cross-attends over a peer's cache while keeping its own, and with only a small query-side adapter trained, matches the accuracy it gets from reading the source document. Activations invert this. They are best written straight into the receiver, with no bridge, no adapter and no training, which beats exposing them alongside the receiver's own, though that advantage rests on knowing where in the receiver the state belongs. The rule is augment the KV cache, replace the activation, established under one protocol with a shared control ladder. One adapter, trained once, also serves publishers it never trained on, and a pool of several at once, but only those whose internal state lies close to the receiver's own.
agentmulti-agentagent system - arxiv:2610.22020 · cs.RODemoStream: Batched Task-and-Motion-Plan Optimization for Automated Learning from DemonstrationCaelan Garrett, Neel Jawale, Fabio Ramos
Learning from demonstration is an effective strategy for endowing robots with manipulation capabilities but requires training on datasets comprised of massive amounts of demonstrations. Existing demonstration collection approaches either require constant human involvement, for example, via teleoperation or are specialized for a single task. Task and Motion Planning (TAMP) systems are able to autonomously solve and therefore generate demonstrations for general-purpose tasks; however, learning from their demonstrations has proven difficult due to multi-modality in their solutions. In this work, we build on recent work leveraging GPU acceleration for TAMP and propose DemoStream, a TAMP system that optimizes demonstration quality in order to improve behavior cloning success rates. First, we introduce batched TAMP sampling operations along with a new TAMP algorithm that correspondingly performs batched constraint satisfaction in order to explore many solutions in parallel. We conduct simulation experiments ablating different action spaces, cost functions, and batch sizes to analyze the effect of these choices on downstream policy success rates. Finally, we show that we can leverage the safety guarantees of TAMP to directly collect real-world demonstrations at an average rate of 66 demos per hour. See our website for policy rollout videos: demostream.github.io.
manipulationteleoperation - arxiv:2610.22019 · cs.LGPriming Base Models for Agentic Post-Training: An Open Midtraining RecipeSoumye Singhal, Joseph Jennings, Ka Shun Shum, Alexander Bukharin +10
Frontier models, general-purpose and domain-specific alike, now include a midtraining stage between pretraining and post-training, but little has been published about how this stage is built. We present an open midtraining recipe that primes base models for agentic post-training, and we show its effectiveness on two Nemotron models of different sizes. Pretraining typically runs at short context and sees almost no agent interactions, while SFT data is mostly synthetic, and very long SFT stages can hurt the model's later RL trainability. Our midtraining stage sits between the two. It trains at 256K context for hundreds of billions of tokens on a diverse mixture of natural software-engineering data like packed repositories, pull requests, commit histories, and build logs together with a small share of agentic traces. On Nemotron 3.5 Lightning, 500B midtraining tokens improve SWE-bench Verified from 60.7 to 64.5, SWE-bench Multilingual from 55.8 to 62.8, and SWE-bench Pro from 32.0 to 35.5 after identical full SFT. On the larger Nemotron 3.5 Super, 280B midtraining tokens improve DeepSWE, a harder benchmark of long-horizon tasks, from 11.7 to 15.7 and SWE-bench Verified from 72.9 to 74.3 after identical full SFT. Midtraining also improves robustness across agent harnesses on both models. We report the ablations behind the recipe and an evaluation protocol built on short fixed SFT runs, since base-model proxy metrics did not correlate reliably with post-training results.
agentagenticpost-trainingbenchmarkevaluation protocol - arxiv:2610.22018 · cs.CVWhen Models Misread Ink: Auditing VLMs for Arabic Calligraphy RecognitionNiaz Mohaiman Abtahi, Evan Maksud, Reyana Shahid
Arabic calligraphy is a high-stakes cultural visual domain for Muslim communities, yet VLM evaluations rarely test its fine-grained script traditions. We audit 12 VLMs on a cleaned 1,195-image six-way benchmark using numeric, visual-cue, and Diwani-first prompts. In this closed-set setting, the best numeric-prompt accuracy is 60.9% for closed/API models and 47.2% for open-weight models. The main failure is default-label collapse: every numeric run is dominated by a single label taking 33-63% of predictions, against an 18.8% majority-class baseline. For 8 of 12 models the default is Thuluth; GPT-4o labels 52.9% of all images as Thuluth although it is only 18.1% of the set. Prompting helps unevenly: visual cues move accuracy by more than 10 points for some models but hurt others, and foregrounding Diwani often relocates the default rather than removing it. Arabic calligraphy recognition is therefore a compact, culturally grounded stress test for VLM robustness and for measuring which visual traditions become legible to automated systems.
benchmark - arxiv:2610.22017 · cs.AICARA: Causally Guided Retrieval Agent for Question AnsweringWeicheng Liu, Xiangcheng Wu, Depeng Xu, Haopeng Zhang +1
Retrieval-augmented generation over text-rich knowledge graphs requires identifying evidence that contributes to answering a question, rather than merely matching its textual content. Existing agentic retrievers adaptively balance global search and multi-hop exploration but primarily rely on textual similarity to guide retrieval. We propose CARA, a training-free, CAusally guided Retrieval Agent that explicitly evaluates the causal relevance of retrieved evidence through two complementary criteria: consequentiality, measuring whether adding evidence improves answerability, and essentiality, assessing whether removing evidence weakens answerability. CARA integrates these assessments with path consistency to guide graph exploration and employs reciprocal rank fusion to aggregate evidence from multiple agents. We evaluate CARA on three STaRK datasets and augment STaRK with ground-truth answers to enable downstream question answering evaluation. Experimental results demonstrate that CARA consistently outperforms existing retrievers. CARA also improves answer quality and faithfulness while examining fewer graph nodes and requiring fewer retrieval operations. These findings demonstrate that causal relevance assessment enables more focused and effective agentic graph retrieval, improving both retrieval performance and downstream question answering.
retrieval-augmentedknowledge graphagentagentic - arxiv:2610.22016 · cs.LGClose Enough is Nowhere Near: Fidelity-Consumability Decoupling in KV CachesAdam Deng, Shicheng Liu, Yada Zhu
Multi-agent LLM systems communicate in text, which is lossy and wasteful: the sender throws away the reading it already did. A growing literature instead transmits the KV cache, the natural carrier of that reading. We study the most appealing form of the idea: both models stay frozen and the sender's cache is substituted for the reader's with nothing trained. It does not work, and the obstruction belongs to the reader. A frozen transformer reads only caches close to the one it would have written itself, a region we call its tolerable set. We show that geometric similarity to that cache does not predict whether the reader can use it, and the failure is not gradual, so a more faithful substitute is not the remedy it appears to be. We explain this phenomenon with competitive attention. Experimentally, we test eight diverse substitution methods against two measured anchors: an uninformative cache baseline and the reader's own accuracy from having read the passage. All sit at the baseline. That turns a long tail of separate negative results into one outcome with a stated condition for overturning it. A different design does work: if the reader keeps its own cache and attends over the sender's as extra context, adapting only its queries, it beats its own baseline, and also clears the baseline with eight senders pooled where undirected attention over the same union falls below it. We offer this as a characterization on one probe, not a leaderboard claim. Our central contribution is a redirection: pursuing faithful substitute caches is the wrong target; adding a cache to what the reader already has is the promising one.
multi-agentleaderboard - arxiv:2610.22014 · cs.LGGraphBEAT: Graph-based Bucketized Event Aggregation Over Time for Early Risk Stratification of Gastrointestinal CancersLeo Y. Li-Han, Chris Varghese, Elizabeth B. Habermann, Mark J. Truty +3
Longitudinal electronic health records (EHRs) contain complementary diagnostic and laboratory information about patient health. However, irregularity, sparsity, heterogeneity, and incompleteness in documentation complicate EHR modeling. To address the problem, we introduce GraphBEAT (Graph-based Bucketized Event Aggregation over Time), a temporal graph representation learning framework that groups clinical events into time buckets and transforms observed features into patient-specific, temporally directed graphs to enable early risk stratification for gastrointestinal cancers. Each node in an EHR graph encodes feature identity, observed value, and temporal position of clinical events. A context-aware, time-decay message-passing mechanism models transitions between successive buckets, followed by multi-head attention pooling to generate patient-level representations. GraphBEAT was trained and evaluated on multi-site longitudinal diagnosis codes and laboratory tests to predict pancreatic, liver, and intestinal cancer at 12-, 24-, and 36-month diagnostic lead times. In all experimental setups, GraphBEAT consistently outperformed baseline graph neural networks (GraphSAGE, GAT, and GCN) and sequence modeling approaches (Transformer and long short-term memory [LSTM]). GraphBEAT also required over 250x fewer floating-point operations per patient and over 20x fewer parameters than the Transformer and LSTM baselines. These results support GraphBEAT as an accurate and computationally efficient approach for cancer risk stratification from sparse, heterogeneous longitudinal health records.
memory - arxiv:2610.22009 · cs.CVVisual Sanad: Auditing Vision-Language Models for False Acceptance of Qur'anic CitationsNiaz Mohaiman Abtahi, Reyana Shahid, Evan Maksud
Vision-language models (VLMs) are increasingly used to interpret religious, educational, and culturally sensitive visual content, yet their reliability on citation verification remains underexplored. We introduce Visual Sanad, a controlled benchmark for testing whether VLMs falsely accept incorrect Qur'anic verse attributions when the same quote-reference pair is rendered under different visual framings. The benchmark contains 2,400 rendered images built from 100 English Qur'an verses, four factual states per verse, and six visual conditions ranging from plain text to decorative and authority-framed layouts. All quote-reference pairs were source-checked and human-audited before rendering. Across 24 full model-prompt runs, we observe a sharp capability split. Proprietary API-served VLMs range from near-perfect citation verification to moderate false acceptance, whereas recent open-weight VLMs often show weak discrimination and default to VALID on most images. De-biased prompting can reduce false acceptance, but often by shifting the decision threshold toward INVALID rather than by producing strong citation discrimination. Visual Sanad therefore exposes a culturally grounded multimodal factuality failure: fluent VLMs may treat displayed religious citations as self-authenticating unless they have strong citation-verification ability.
benchmark - arxiv:2610.21999 · cs.ROVP-EVT: Active Visibility Preservation for Embodied Visual TrackingYanci Wen, Tianle Zeng, Jianwei Peng, Weixi Situ +2
Embodied Visual Tracking (EVT) is an essential capability for intelligent robots in human-robot collaboration, requiring an agent to follow a designated target person in dynamic environments relying solely on egocentric visual perception. However, persistent pedestrian occlusions in dynamic scenes pose critical challenges to existing EVT methods. Lacking proactive visibility preservation, reactive tracking methods frequently experience prolonged target loss under severe occlusions. Therefore, we propose VP-EVT, a framework that treats target visibility as an explicit objective of motion control rather than a by-product of following. Acting on this objective requires understanding both what obstructs the target and how robot motion can change that obstruction. VP-EVT therefore uses language-guided perception to extract separate target and pedestrian observations. When occlusion makes current observations incomplete, VP-EVT combines visual history with ego-motion to maintain estimates of their relative positions and motion. Using these estimates and the current following waypoints, VP-EVT predicts whether continued following will sustain the obstruction. When the predicted risk is low, the original following motion is retained; otherwise, a recurrent controller combines updated estimates with maneuver context from preceding decisions to coordinate viewpoint adjustment and the return to stable following. Moreover, we introduce OccEVT-Bench for closed-loop evaluation of persistent occlusion and collect expert demonstrations for training. On its Risk split, VP-EVT improves success rate by 121.4% relative to a competitive baseline and reduces visibility deficit and longest target-loss duration by 78.8% and 74.4%, respectively.
embodiedagent - arxiv:2610.21995 · cs.LGInterfaces and Modules: A Scalable Representation for Weight-Space Learning across ArchitecturesMichał Kozyra, Michael Bronstein
Weight-space learning across architectures requires representations that respect parameter symmetries and scale to large networks. Existing equivariant tensor encoders require architecture-specific symmetry specifications, while neuron graphs make global attention costly. We introduce a typed graph organized around shared coordinate-permutation actions. Its nodes, called interfaces, represent coordinate spaces whose bound tensor axes must be permuted jointly; operation occurrences, called modules, form typed hyperedges. A compiler derives this graph from supported PyTorch models, with exact recovery of the permutation factors defined by its constraints. The Universal Weight Encoder (UWE) combines local equivariant updates to parameter features with global attention over invariant interface summaries. The small number of interfaces makes global attention practical, and a single encoder processes different supported architectures. UWE is competitive with architecture-specific state-of-the-art methods on established benchmarks. On a new zoo of 1,200 transformers spanning 100 architectures, it outperforms the evaluated baselines at ranking networks within unseen architectures, including architectures larger than those used for training, while using substantially less memory.
benchmark - arxiv:2610.21985 · cs.ROVeriVLA: Adaptive Test-Time Verification for Flow-Based VLAs with Latent World ModelsXueli Sun, Yuxiang Lu, Yantong Liu, Xianzhe Fan +3
Flow-based Vision-Language-Action (VLA) policies are inherently stochastic: different noise seeds can produce action chunks of varying quality from the same observation. At test time, however, the policy typically commits to a single sample without an explicit estimate of its quality. In fine-grained manipulation, executing a suboptimal action can irreversibly disrupt task progress and lead to failure. To exploit policy stochasticity, we introduce a training-free two-axis trigger that uses two initial samples to measure within-chunk temporal variation and cross-sample variance, detecting the critical states. Once triggered, a JEPA-based latent world model ranks candidate action chunks by predicting action-conditioned future latent states and the corresponding changes in task progress. By operating in latent space, the verifier avoids pixel reconstruction and enables efficient and more accurate ranking. Across three real-world tasks, $\textit{i.e.,}$ pen insertion, pipetting, and apple placement), our method raises the average completion score from 50.8% to 65.5%, while adding 326 ms for sampling and verification per triggered decision on a single RTX 4090. Consistent gains are also observed in simulation, where the average success rate rises from 66.4% to 73.6% in the clean setting of RoboTwin 2.0, from 62.8% to 69.6% in its randomized setting, and from 96.9% to 97.6% on LIBERO.
vision-language-actionmanipulationliberorobotwinworld modelaction-conditioned - arxiv:2610.21984 · cs.ROGRID-Planner: Graph-Structured Relational Reward Shaping for Diffusion PlanningZhao Yang, Zezhong Qian, Yinan Shi, Longjun Liu +1
Autonomous driving planning requires selecting trajectories that remain safe, rule-compliant, and comfortable under complex multi-agent interactions, yet robust closed-loop decision making in planning-critical scenes remains challenging. Existing diffusion- and RL-based planners often rely on generic rewards and use relational structure mainly for representation, limiting how planning-critical rollouts shape policy improvement. We present GRID-Planner, a graph--reward integrated diffusion planning framework that aligns generative planning with structured relational signals. Specifically, GRID-Planner combines graph-aware importance weighting, a time-indexed relational reward surrogate, and an advantage-weighted diffusion objective with on-policy self-distillation, forming a rollout-return-advantage-update loop without explicit likelihoods. Experiments on nuPlan and interPlan show that, compared with the diffusion-only baseline, GRID-Planner reduces collision rate, lane deviation, and jerk by 35.8\%, 60.8\%, and 27.4\%, respectively, and improves the planning score by 3.4\%. Project website: https://github.com/yangzhaojason/GRID-Planner.
multi-agent - arxiv:2610.21983 · cs.LGOntoAct: Hypothesis-Based Neuro-Symbolic Agents through Ontological AbstractionDonghyeok Lee, Bong Gyun Kang, Junyong Ahn, Sungroh Yoon
Learning from experience requires recovering reusable knowledge and determining when it applies to a new situation. We introduce OntoAct, a hypothesis-based neuro-symbolic framework built on ontological abstraction. OntoAct turns individual experiences into axioms over entity types, relations, and conditions, making their underlying patterns reusable across situations. A large language model (LLM) uses this world knowledge to propose hypotheses and actions, while a symbolic verifier assesses their supporting conditions against current observations. Verification identifies missing evidence, and new observations and intervention outcomes update hypothesis assessments and guide subsequent actions. We also introduce Tacit-FLE, a benchmark built on the Factorio Learning Environment (FLE) for diagnosing and repairing hidden faults in existing factories using inherited operational records. Its synthetic maintenance corpus distributes evidence across records, and an executed repair may leave the fault unresolved, requiring agents to reassess their diagnoses through interaction. Across Tacit-FLE, Mars, and ALFWorld, OntoAct achieves leading task performance with efficient interaction and inference.
benchmark - arxiv:2610.21979 · cs.LGP2TPO: Structure-Aligned Page-to-Token Credit Assignment for Full-Page Document ParsingYong Zu, Ninghua Yang, Zezhong Tan, Xinhong Ma +2
End-to-end full-page document parsing must preserve the content and structure of heterogeneous elements, yet otherwise accurate outputs can contain local errors and omissions. A single page-level advantage does not explicitly identify which elements differ in quality across sampled outputs or where their errors occur. We introduce Page-to-Token Policy Optimization (P2TPO), a structure-aligned credit-assignment framework that operates on full-page policy trajectories. Reference-element alignment first associates block-level quality contrasts with generated spans. Unified Anchor Pricing (UAP) incorporates matched and missing element outcomes and provides feedback relative to a fixed quality ceiling when sampled outputs receive identical but imperfect scores for a reference element. Within matched blocks, Span Credit Redistribution (SCR) redistributes existing block credit according to aligned correct and erroneous spans, using boundary proxies for within-block omissions while preserving the total block-derived advantage across its tokens. Under a fixed-weight additive page utility, we show that element-indexed contrasts retain local quality differences that page-level aggregation can cancel out. In the main experimental comparison, P2TPO outperforms all evaluated policy optimization baselines in seed-averaged scores on each of the six document parsing benchmarks. Additional experiments demonstrate gains across model scales and initializations.
benchmark - arxiv:2610.21969 · cs.LGUncertainty-Aware Federated Learning for Joint Communication and Computation in XR SystemsRobert-Jeron Reifert, Hayssam Dahrouj, Alaa Alameer Ahmad, Haris Gacanin +1
The convergence of communication and computation, empowered by artificial intelligence (AI) enhanced capabilities, emerges as a key enabler for realizing the promising functionalities of the sixth-generation of communication systems (6G). To this end, this paper investigates an AI-native, robust network, where multiple base stations (BSs) simultaneously serve several extended reality (XR) devices. The paper then presents an innovative distributed deep learning-based approach to simultaneously manage the transmit and computing powers, and computation allocation, amidst uncertainties in both channel and computing states information. Under such an interference network scenario, the paper proposes a robust approach for minimizing the worst-case delay across the served XR devices, subject to computation and power constraints. The paper leverages a deep neural network (DNN) that can be trained either centrally or in a federated manner to map estimated channels and computation requirements to optimized resource allocations per BS. During training, uncertainty samples, reflecting both channel and computing estimation errors, are injected after the DNN output to account for such joint uncertainties. Our results validate the enhanced robust delay performance and scalability of joint uncertainty injection as compared to uncertainty-agnostic methods and state-of-the-art benchmarks, with federated learning-based solutions achieving near-centralized performance while enhancing efficiency, privacy, and adaptability.
benchmark - arxiv:2610.21965 · cs.ROPerceptual Comparisons of Viscoelastic Renderings: Fractional- vs Integer-Order ModelsGorkem Gemalmaz, Harun Tolasa, Volkan Patoglu
Realistic haptic rendering of viscoelastic materials requires reproducing time-dependent behaviors such as creep and stress relaxation, which arise from the material's inherent memory effects. Fractional-order viscoelastic models have been shown to capture these dynamics accurately and compactly, but it remains unclear whether their increased fidelity translates into perceptual benefits during haptic interaction. In this work, we investigate whether users perceive fractional-order models as more realistic than conventional integer-order models while rendering viscoelastic materials. We compare the fractional-order standard linear solid (FO-SLS) model with two integer-order alternatives: the 3-parameter SLS model and the 5-parameter generalized Maxwell model. We consider two complementary approaches to optimal parameter selection to evaluate rendering realism from both objective and perceptual perspectives: In the first approach, we determine the model parameters via least-squares based system identification of physical viscoelastic reference materials to establish evaluation baselines. In the second approach, we determine population-level perceptually optimal parameters using a qualitative-feedback-based human-in-the-loop optimization and aggregation method. This method establishes model parameters that maximize perceived realism of viscoelastic behaviors through qualitative evaluations. Human-subject experiments indicate that the FO-SLS model is perceived as significantly more realistic than both integer-order alternatives under both parameter selection approaches. Overall, this study provides experimental evidence that fractional-order viscoelastic models are not only theoretically well-motivated, but also perceptually relevant for realistic haptic rendering of creep- and relaxation-dominated behaviors of viscoelastic materials.
memoryhuman-in-the-loop - arxiv:2610.21958 · cs.LGChronoSheaf: Joint Spatio-Temporal Spectral Filtering on Cellular SheavesAbeer Mostafa, Zahra Ahmadi
Sheaf neural networks generalize graph neural networks by assigning each edge a learned linear restriction map that specifies how the local representations of adjacent nodes relate to one another, thereby mitigating oversmoothing. Existing sheaf architectures, however, define the sheaf Laplacian as an instantaneous spatial operator and therefore cannot explicitly model how signals evolve over time on a graph. We first introduce a spectral sheaf Laplacian, defined jointly over space and time, such that a single operator regulates signal propagation across both the spatial graph and the temporal domain. Directly extending this operator's receptive field in space or time would require increasingly complex sheaf structures. Instead, we propose a spectral filter that expands the effective receptive field along both dimensions without enlarging the underlying sheaf. Building on this operator, we introduce ChronoSheaf, a spatio-temporal sheaf network with two key components. First, Transitional Restriction Maps (TRM) decompose each node's restriction map into a spatial term and two temporal frames that independently control its relation to neighboring nodes at the preceding and following timestep, enabling asymmetric temporal influence. Second, the Dual Spectral Filter (DSF) applies a single bounded-degree polynomial filter with an adaptively estimated spectral radius to the joint spatio-temporal sheaf operator. We validate ChronoSheaf on a diverse set of spatio-temporal forecasting benchmarks, where it outperforms state-of-the-art baselines. The code is available at: https://anonymous.4open.science/r/ChronoSheaf-1D7D/.
benchmark - arxiv:2610.21948 · cs.CVFollowing Delegation Traces to Study Human Oversight in Interaction with Multimodal AIAgnese Chiatti, Matteo Matteucci, Viola Schiaffonati
As multimodal AI systems are becoming our everyday thought partners, the shift towards studying how users transfer control to these systems becomes especially urgent. In this paper, we report an overview of findings from an online study with n=56 AI-literate participants and 392 Video Question Answering trials. After inspecting each multimodal AI response, participants rated task-specific trust and selected one of three intended policies for an analogous future task: reject AI assistance, use it with verification, or delegate fully to the AI. Within participants, post-interaction trust strongly predicted the user's delegation choices (OR=14.97, 95% CI [7.14,31.37]). Trust and delegation nevertheless remained distinct. Verification was the dominant policy and often co-occurred with high trust. Delegation also increased with task order after adjustment for momentary and participant-level trust (OR=1.22, 95% CI [1.02,1.45]), indicating an experience-related component that self-reported trust ratings alone could not capture. These findings motivate a shift from intended delegation to a non-intrusive observation of how delegation is enacted in collaborative human-AI tasks. We therefore conclude by proposing enacted delegation traces as a methodological framework for future studies of multimodal and agentic AI use.
agentic - arxiv:2610.21941 · cs.CLLearning Speech Attribute Representations via Multi-Teacher DistillationMaryem Bouziane, Salima Mdhaffar, Yannick Estève
Recent post-training approaches have shown that self-supervised speech encoders can be adapted to learn a single utterance-level speech attribute through knowledge distillation. In this work, we investigate whether a single speech encoder can simultaneously preserve multiple complementary utterance-level speech representations. We propose a multi-attribute multi-teacher distillation framework that learns semantic, speaker, language, and emotion representations from four frozen teacher models using a shared SSL backbone and attribute-specific branches. Experiments on multilingual semantic retrieval, speaker verification, language identification, and speech emotion recognition show that the proposed model achieves performance close to dedicated single-task students while requiring only a single shared encoder. Furthermore, the learned layer aggregation strategies reveal that different speech attributes naturally emerge at different depths of the SSL encoder, providing new insights into the organization of speech representations.
post-training - arxiv:2610.21939 · cs.RODe-Risking Robotic Welding for In-Space Metal Joining Through Physics-Grounded Digital TwinsZhou Daniel Hao, Umer Masood Chaudry, Robert Shaw, Andrew Cheney +3
Growing interest in in-space servicing, manufacturing, and assembly has renewed attention on reliable metal joining for modular construction and on-orbit repair. Robotic welding is promising, but reduced-gravity and low-pressure operation remains poorly characterised and flight-like tests are limited. The UK Space Agency-supported ISPARK project is developing a physics-grounded digital twin for structured process de-risking. This paper evaluates a model-based prototype combining a finite-volume thermal weld-pool model, Gaussian-process (GP) and neural-network surrogates, and a simulated six-axis manipulator. The reduced model includes solid-liquid phase change, effective thermocapillary transport, and a local-pressure parameter. Gravity is assessed separately through a hydrostatic-to-capillary pressure ratio and does not enter the thermal energy equation. Across repeated five-fold cross-validation, the GP has lower normalised mean absolute error than the neural network for all eight outputs. In 10 held-out boundary cases, 74 of 80 solver responses fall within nominal 90% GP predictive intervals. A three-output GP query takes 249 microseconds, including scaling and predictive variance. In a simulated UR5 L-shaped-seam study, a proportional controller adjusts the synergic power-source setting as travel speed falls near the corner. Finite-volume re-evaluation finds all 12 closed-loop operating points and 9 of 12 open-loop points within an illustrative 1.0-3.0 mm root-side-width band. These simulation results support using the framework to screen welding conditions and control strategies before experimental tests. Thermal-vacuum, targeted reduced-gravity, and robotic hardware-in-the-loop studies are needed to assess the model and controller.
manipulator - arxiv:2610.21938 · cs.AIAgentic-Priors: Using Large Language Models to Address the Cold Start Problem in Preference ElicitationBruno Arcanjo, Zhaoxing Li, Sebastian Stein
An agent that learns a user's preferences from their choices faces a cold start problem: before any choices are observed, it must rely solely on a prior. Informative priors are usually fitted to historical choice data, which new tasks and deployments lack. To address this, we present Agentic-Priors, which lets agentic systems use the implicit knowledge of large language models (LLMs) instead. Specifically, an LLM states a new user's trade-offs in interpretable terms, such as willingness to pay or value of time, which we map to a prior over multinomial logit coefficients. This prior is then refined by a Bayesian learner as choices are observed. Across seven choice datasets, Agentic-Priors improves cold start accuracy over an uninformed prior by 5-40 percentage points. It also outperforms AutoElicit, a state-of-the-art LLM prior-elicitation method, on three datasets and loses on one. It is up to 18 points more accurate than in-context learning, and never significantly worse. Against hierarchical Bayes models fitted to surveys of increasing size, our approach is worth roughly 6-40 surveyed users, in terms of log-loss, so its value is concentrated in a deployment's first users.
agentagentic - arxiv:2610.21937 · cs.CVFlow-CLIP: Image-Conditioned Residual Scoring for CompositionalImage-Text MatchingMinghong Wei, Enguang Zhu, Pu Cao, Lu Yang +1
Vision--language pre-training has enabled substantial progress in image--text retrieval and zero-shot recognition by learning transferable cross-modal representations. However, sentence-level alignment remains insufficient for compositional matching, where captions can describe the same objects and attributes while differing in how these elements are bound. To address this limitation, we propose Flow-CLIP, which refines global image-text alignment through image-conditioned changes in token--patch evidence. Global Compatibility Calibration (GCC) learns a sentence-level similarity metric, while Image-Conditioned Evidence Refinement (ICER) incorporates visual patch context into token queries and uses the resulting change in local matching evidence to adjust the global score. We further introduce Cross-Image Margin Ranking (CIMR), which compares the same positive--negative caption pair across matched and mismatched images, encouraging a larger ranking margin when the image supports the correct description. Experiments on compositional matching and attribute--object binding benchmarks show that Flow-CLIP improves binding accuracy over pretrained CLIP and learned global alignment through local evidence refinement, while cross-image ranking supervision mitigates reliance on language priors. Code will be released at link.
benchmark - arxiv:2610.21935 · eess.SYDefense Filters for Robust Learning of Partially Observed Dynamical SystemsJihun Kim, Javad Lavaei
Robust estimation of unknown system dynamics from partial observations under attack is a core challenge in learning for control, increasingly critical in multi-agent networks subject to adversarial data poisoning. Existing methods either require highly restrictive assumptions on the adversary or yield weak error bounds. In this paper, we consider identifying partially observed linear dynamical systems subject to adversarial attacks with an attack probability $p < 0.5$ at each time. We propose a defense-filtering framework in which a system designer leverages knowledge of the observation matrix to construct a defense filter satisfying full column rank and subspace separability conditions, which enables the evaluation of a single-attack scenario against observed data. Using this framework, we analyze the constrained $\ell_2$-norm estimator and establish finite-sample error bounds for recovering the Markov parameter matrix $G^*$. This is the first approach to provide provable system identification guarantees via a defense filter for partially observed systems subject to harsh attack environments where the attack probability approaches $0.5$.
multi-agent - arxiv:2610.21922 · cs.ROTS-WAM: Temporal Latent Actions and World-Action Coupling for Vision-Language Robot ControlMenghao Yin, Jiaju Yin, Xiyue Dong, Lixin Xu +4
Many vision-language-action policies rely on action-labeled robot demonstrations and lack an explicit mechanism for predicting future visual states. These limitations hinder learning from action-free videos and can encourage overfitting to demonstration conditions, weakening generalization and robustness. We present TS-WAM, a lightweight Teacher-Student World-Action Model that addresses these challenges through self-supervised temporal learning and explicit future prediction. Its vlm-jepa framework learns from heterogeneous data, including robot demonstrations and action-free human videos, using an EMA target branch without requiring a separately trained external teacher. By coupling vlm-jepa with delta-lwm, TS-WAM learns to predict latent actions from the current observation and instruction and uses them to predict future visual features that guide action generation at inference. A unified end-to-end training architecture jointly optimizes temporal representation learning, visual dynamics, and robot control without a separate latent-action or world-model pretraining stage. The EMA target branch is removed at inference, keeping the deployed policy lightweight. Experiments on LIBERO, LIBERO-Plus, and SimplerEnv, together with real-world manipulation experiments on Franka Research 3 and RevoTron, demonstrate that TS-WAM achieves strong manipulation performance, generalization, and robustness.
vision-language-actionmanipulationliberofranka - arxiv:2610.21919 · cs.LGThe Lagrangian Structure of Cost-Penalized Active Feature AcquisitionJoseph Bingham
Active Feature Acquisition (AFA) sequentially selects features to maximize predictive accuracy under an acquisition budget. A natural heuristic greedily maximizes Expected Information Gain minus a cost penalty, EIG(j|x_S)-lam*c_j (EIG-Cost), but its theoretical guarantees are unclear. We provide an a-posteriori certificate decomposing the gap to the optimal hard-budget policy into a budget-slack term and a greedy-suboptimality term Del_lam. The certificate is computable when the penalized optimum is tractable; on small heterogeneous-cost instances, we verify it exactly, finding Del_lam near zero across 4,000 solved cases. On the large clinical task, Del_lam is not computable, so the decomposition is diagnostic rather than a numerical guarantee. Our analysis identifies EIG-Cost as one-step greedy optimization of the Lagrangian relaxation of expected-cost AFA, with lam as its dual multiplier. For equal costs, we prove Del_lam <= exp(-gam*I(Y;X_[d])), where gam is the submodularity ratio, yielding a structural guarantee in the near-submodular regime. Supporting results establish an adaptive chain-rule identity showing that cumulative stepwise EIG optimizes transcript information, and prove that A -> I(Y;X_A) need not be submodular, even for jointly Gaussian variables. This rules out a direct Nemhauser-Wolsey-Fisher bound; naive Bayes restores submodularity. We also bound the randomization gap between hard-budget and expected-cost optima, analyze sensitivity to dual-multiplier error under local smoothness, and derive a conditional CVAE generator-shift bound given a bound on forward predictive KL. We treat this KL as a diagnostic quantity, not as a transform of training loss. Experiments cover synthetic Gaussian settings, AFA benchmarks, and clinical differential diagnosis on MIMIC-IV. On MIMIC-IV, EIG-Cost achieves macro-F1 0.180 +/- 0.012 at a $30 budget, using 6.6% of full-panel cost.
benchmark - arxiv:2610.21915 · cs.ROWhat Should Robot Policies Learn to Predict? Rethinking Supervision Targets in Vision-Language-Action LearningHaitao Shen, Hanyu Liu, Yihan Lin, Huiping Liu +6
Vision-language-action (VLA) policies are typically trained with action prediction as the primary supervision signal, since robots ultimately execute actions. Auxiliary objectives such as latent-action (LA) and future-vision (FV) prediction are often added to improve policy learning, yet they compete with action prediction for the same training budget. We ask whether, under a fixed budget, allocating a substantial share of training samples to such objectives, rather than to further action prediction, can improve policy learning. We build a unified VLA framework that couples a two-stream multimodal diffusion transformer (MMDiT) with a sample-level router, which assigns each training sample to either action prediction or an auxiliary objective and thus makes supervision allocation explicit. Within this framework, we instantiate three LA targets based on LAPA and its DINO-feature variants, and three FV targets based on DINOv3, V-JEPA 2, and Wan2.2 representations. Across LIBERO, CALVIN, and four real-robot tasks, suitable LA and FV supervision outperforms action-only training under the same budget, even when action prediction receives no more than half of the training samples. FV generally outperforms LA, although the strongest LA target remains competitive on CALVIN. Diagnostic analyses suggest that DINO-based LA targets emphasize relative motion, whereas FV targets emphasize absolute position and task identity. We therefore argue that what robot policies learn to predict, and how training is allocated among these targets, should be treated as explicit design choices rather than defaulting to action-dominant training.
vision-language-actionvlaliberov-jepa - arxiv:2610.21914 · cs.AICodeHealth Predicts the Token Cost of Agentic Code Review: A Pre-Registered Study of 465 Agent-Authored Pull RequestsLo Gullstrand Heander, Markus Borg, Emma Söderberg, Christofer Rydenfält
Context: Organizations increasingly delegate code review to multi-agent large language model (LLM) pipelines billed per token. Low code quality raises human maintenance effort, but its effect on agent review cost has not been measured. Objective: We test whether baseline CodeHealth predicts the token cost of agentic review and fix beyond pull request (PR) footprint. Method: In a pre-registered observational study, we sampled 465 agent-authored PRs from 246 repositories in AIDev, jointly stratified by CodeHealth and footprint under repository caps. A frozen five-agent review-and-fix pipeline processed every PR with three different LLMs. Per LLM, we fitted a mixed model of log tokens on CodeHealth and log PR footprint, with a repository random intercept. Results: For all three LLMs the CodeHealth coefficient was negative, with 95% CI excluding zero; each CodeHealth point was associated with 3.5% to 6.1% lower token cost. CodeHealth was unrelated to PR footprint, so PR size does not explain the association. Two of three registered smells predicted fix cost; batch fixing left median CodeHealth unchanged. Exploratory analyses located the association in the review stage, where agents iterated more on unhealthy code without writing more comments. Adjusting for the size of the scored files absorbed the association, which we interpret as mediation by file size. Conclusions: A quality metric calibrated against human maintainability also predicts the inference cost of multi-agent review and fix. The effect is modest per point and not carried by PR size, so the refactoring case extends to agent-reviewed code when it changes structure and file size, not only the score. Data Availability: https://doi.org/10.5281/zenodo.22127920
agentmulti-agentagentic - arxiv:2610.21913 · cs.ROSub-Millisecond Reflexes for Continuous Physical Human-Robot InteractionJon Škerlj, Sami Haddadin
Continuous physical human-robot interaction requires robots to rapidly adapt their control behaviour without interrupting task execution. However, existing ROS 2 control architectures rely on stop-to-switch mechanisms that introduce substantial delays and interrupt execution. We present an open-source, sub-millisecond reflex engine based on an enhanced ROS 2-compatible multimode controller that enables deterministic switching between control modalities and integrates reflex behaviours for mitigating contact hazards and kinematic limit violations. Experiments on a Franka Robotics manipulator demonstrate switching delays below 0.1 ms, more than four orders of magnitude faster than conventional ros2_control, while maintaining continuous operation during safety-critical events. In a multi-stage manipulation task, the proposed framework reduced execution time by 3.5% and peak trajectory jerk by more than threefold. By eliminating runtime middleware overhead, controller transitions are implemented as lock-free pointer operations, enabling reflex-level responsiveness within an open-source robotics framework.
manipulationmanipulatorfranka - arxiv:2610.21912 · cs.LGLearning Rewards in Strategic GamesAntoine Bergerault, Cyrille Kone, Giorgia Ramponi
Understanding the objectives that drive agents' behavior is crucial in strategic environments, where multiple agents interact and adapt to one another. This challenge arises in domains such as cybersecurity, autonomous driving, and online markets, where agents' underlying incentives are typically unknown and must be inferred from their observed behavior. Inverse reinforcement learning (IRL) provides a natural framework for this problem, but extending IRL to strategic multi-agent settings introduces a fundamental challenge: a reward that explains observed behavior does not necessarily induce optimal strategies when used for decision-making. In this work, we study this problem through the lens of \emph{strategic identifiability}, and show that learning from a fixed policy pair can be insufficient to strategically identify rewards, independent of the number of expert samples $K$. Motivated by this, we introduce an active IRL algorithm for two-player repeated Stackelberg games, in which the leader adaptively selects policies to elicit informative responses about the follower's reward. Combining maximum-likelihood estimation with experimental design, we derive finite-sample guarantees on the Stackelberg sub-optimality induced by the learned reward, matching up to polylogarithmic factors a new minimax learning rate of $K^{-1/2}$. Finally, we evaluate our approach in a realistic anti-poaching setting, demonstrating the benefits of active exploration for reward recovery and downstream strategic performance.
multi-agent - arxiv:2610.21910 · cs.AICoWorld Reasoner: Agentic Multimodal Reasoning with the Evolving Video World Model-Driven HarnessDongjie Cheng, Ruifeng Yuan, Ran Wei, Ning Song +2
As Vision-Language Models (VLMs) advance, they are increasingly expected to perform spatiotemporal reasoning over motion, actions, and view changes beyond what the input images show. Video World Models (VWMs) provide a natural unified engine for visual imagination by generating task-relevant visual content as additional evidence. However, existing approaches typically invoke a VWM through fixed pipelines centered on a predefined form of visual imagination, such as novel-view synthesis. To support flexible visual imagination for diverse reasoning tasks, we introduce CoWorld Reasoner, an agentic framework that uses a VWM as a unified visual reasoning engine. Its Video World Model-Driven Harness exposes multiple visual operations through tools and uses skills to guide the VLM in deciding whether to imagine, selecting an operation, constructing a generation request, and interpreting the resulting visual evidence. We further introduce VLM-VWM Harness Co-Evolution, which jointly revises tools, skills, and execution code based on execution feedback while keeping both model weights fixed. Across seven benchmarks spanning spatial intelligence, temporal and causal reasoning, and perceptual reasoning, CoWorld Reasoner outperforms direct reasoning with the same VLM. The strongest variant improves on all seven benchmarks, including relative accuracy gains of 11.8% on MMSI and 13.0% on ERQA, and outperforms all five evolving baselines using the same backbone across five benchmarks.
world modelagenticbenchmark - arxiv:2610.21908 · cs.LGThe Tilted Sampling Problem: Optimism, Posterior Sampling, and Logarithmic RegretMohammad Pedramfar, Siamak Ravanbakhsh
Many modern learning systems are used not to select a single best action, but to sample diverse high-quality outputs. Examples include entropy-regularized reinforcement learning, RLHF, and reward-guided generative modeling. Motivated by these settings, we study tilted sampling: an online learning problem in which the goal is to sample from the Gibbs distribution induced by an unknown reward function. We define sampling regret as the temperature-normalized cumulative KL divergence between the learner's distribution and the Gibbs target, corresponding to the suboptimality gap in entropy-regularized reward maximization. Our framework generalizes recent sampling-based and KL-regularized formulations in bandits and reinforcement learning. Our main contribution is a general regret framework that captures the different uncertainty structure of tilted sampling and classical regret minimization. While standard regret scales as $\widetilde{O}(\sqrt{T})$ due to first-order uncertainty, sampling regret is governed by squared confidence widths, yielding logarithmic-in-$T$ regret in standard settings at fixed temperature. We develop sampling analogues of UCB and Thompson sampling (UCS and TS$^2$) and provide unified regret decompositions that explain and extend recent logarithmic guarantees, including the first Bayesian sampling-regret decomposition based on new bounds for posterior-random Gibbs distributions. Our results show that learning to sample from a soft-optimal distribution is statistically distinct from learning to identify a hard optimum.
online learning - arxiv:2610.21902 · cs.CVSeeing is Believing: Benchmarking and Mitigating the Interference of Language Priors on Visual Perception in MLLMsLi Ma, Yan Zhuang, Jiawen Deng, Fuji Ren
Language priors in MLLMs can override visual evidence, impairing visual perception. Existing benchmarks assume shared priors across models and rely solely on accuracy. However, cross-model prior agreement falls below 80%, and over 15% of items can be answered correctly without any image, causing accuracy to overestimate prior resistance. To address these limitations, we introduce ViSiP-Bench, a benchmark grounded in rare-but-real scenarios across six prior categories, with a model-aware dynamic prior induction protocol and three metrics that more precisely and continuously measure prior resistance than accuracy alone. Evaluating 24 models reveals that prior adherence is pervasive, does not improve with scale, and larger models tend to either strongly override their priors or firmly adhere to them. We therefore propose MAPS, which identifies each model's strongest priors and fine-tunes on factually grounded knowledge to soften excessive prior reliance without disrupting general capabilities.
benchmark - arxiv:2610.21900 · cs.CLTEM: Temporal State Tracking Memory for Large Language Model AgentsPengfeng Song, Yunkai Zhuang, Liye Shi
An agent's memory must remain useful as information changes across conversations. A new fact may replace an earlier value without making that value irrelevant to a later question. We present Temporal State Tracking Memory (TEM), a training-free framework for tracking these changes. TEM stores facts as versioned records linked to their sources. It retrieves relevant versions, places them in source order and labels historical records before passing the evidence to the answer model. Event memory and original dialogue turns provide details missing from structured facts. We evaluated TEM on 420 questions drawn from 600 sessions involving 60 personas. Across four open-weight models, key-entity hit rates reached 75.71-97.62% overall and 96.67-100.00% on the historical-residence subset. Adding historical labels to otherwise identical, source-ordered facts increased overall hit rates by 17.14-55.95 percentage points. The evidence was compact: 94-118 tokens per answer, or 48.75-56.73% less than session-aware retrieval-augmented generation (RAG). A further evaluation on 1,540 LoCoMo questions tested hybrid memory under a fixed evidence budget. TEM exceeded session-aware RAG with both backbones; conservative score differences were 7.92 and 10.58 percentage points. The results support preserving fact versions and exposing their temporal status to the generator when answering questions about repeatedly updated information.
memoryretrieval-augmentedrag - arxiv:2610.21887 · cs.MAMetaShare: Failure-Aware Collective Self-Improvement for Open-Ended AgentsVishal Bhutani, Prasang Gupta, Sumanth Chundru, Kevin Paul
Humans often learn fastest not by repeating success, but by remembering which failures not to repeat. Open-ended self-improving agents can share experience across lineages, as in GEA, or edit their own improvement mechanisms, as in DGM-Hyperagents, yet failed improvement attempts are rarely retained as reusable cross-branch knowledge, and search can collapse onto redundant strategies. We introduce MetaShare, a failure-aware framework that records successful and unsuccessful agent modifications together with signed performance deltas. Before each generation, MetaShare retrieves diverse high-performing strategies alongside high-confidence failures from other lineages, letting agents reuse effective mechanisms while avoiding known dead ends. A task-blind auditor periodically injects absent or underrepresented mechanism categories, while embedding-based de-ranking limits redundancy during retrieval. Multiple meta-agents propose modifications in parallel at heterogeneous temperatures behind a deterministic selector that gates applicability, scope, compilability, and size. Experiments across paper review, coding, math grading, robotics reward design, and financial audit-risk assessment show that MetaShare's median matches or exceeds DGM-Hyperagents' final performance and reaches its own best-performing agent in 1.7--2.9$\times$ fewer evolution iterations.
agentself-improvingself-improvement - arxiv:2610.21882 · cs.CVAnvil: Scaling Embodied Spatial Intelligence through Knowledge RecirculationWeilin Ruan, Bohan Zhou, Jinghao Li, Yangkai Wei +11
Embodied spatial intelligence requires more than visual recognition: models must localize entities, reason about their relations, and express this knowledge in action-relevant forms. Yet joint training can conflate the knowledge needed at different stages because supervision is fragmented across incompatible geometric and semantic outputs. We introduce Anvil, a unified framework for data generation, training, and benchmarking for embodied spatial intelligence. (1) Anvil-Gen is a heterogeneous data flywheel that combines form-specific generation, cross-stage consistency verification, and failure-driven recollection across 11 Ground--Relation--Act forms. (2) Anvil-Ben provides scene-aligned supervision for learning diverse embodied spatial capabilities and scene-linked assessments that distinguish stage competence from error propagation. (3) Anvil-MOPD is a 9B student that consolidates heterogeneous spatial knowledge through distillation from specialist teachers. Semantic-first-fault routing targets teacher supervision to the student's earliest decisive failure, while stage normalization controls the loss scale across native outputs of different lengths. We construct 5.8M source-grounded records spanning 160K images, reserving 580K spatial records for held-out evaluation and using the remaining 5.22M records for training. On average across seven public benchmarks, Anvil-MOPD outperforms Qwen3.5-9B and RynnBrain-8B by 9.29 and 11.03 percentage points, respectively, while reaching 80.67% on VABench-Point. Controlled studies support targeted coverage, selective transfer, and feedback-guided repair as complementary framework components.
embodiedbenchmark - arxiv:2610.21881 · cs.LGLearning Lean Feedback for Efficient Proof-Candidate Evaluation on GPUsLazar Milikic, Etienne Bamas, Emmanuel Abbe, Viktor Kunčak
Training language models for formal theorem proving requires large numbers of verifier calls. Slow checks and timeouts often leave GPUs waiting and withhold rewards from potentially correct proofs. We train neural evaluators to predict Lean 4 verdicts and compiler feedback on GPUs using 4.84M labelled programs. We compare two validity classifiers with a feedback model that annotates proof attempts with predicted goal states and error messages. A two-stage evaluator that screens candidates with a classifier and scores its positive predictions with the feedback model reaches 87.8% accuracy in predicting whether Lean accepts a candidate proof, evaluated on a 68,932-program benchmark collected from public provers. The combined evaluator processes 3.5 candidates per second on one H100, 4.2x the throughput of our 64-worker Lean baseline with environment reuse. Finally, we combine our evaluator with Lean verdicts to obtain a hybrid reward system that shortens GRPO training steps by approximately 31%. After one GRPO epoch with hybrid rewards, the prover achieves nearly the same miniF2F pass@1 as with Lean-only rewards (62.2% versus 62.4%) and a pass@32 of 72.4% versus 71.3%.
benchmarkevaluator - arxiv:2610.21869 · cs.LGLethe: Link Inference Attacks For Evaluation of Edge Unlearning MethodsAndrea D'Angelo, Francesco Gullo, Davide Mottin, Sayan Ranu +1
Graph Unlearning (GU) aims to remove the influence of specific training data from a Graph Neural Network without retraining anew. While most methods address removal of nodes and features, Edge Unlearning (EU), aiming at forgetting edges, has recently attracted dedicated attention as a standalone problem due to the hardness of completely removing the impact of an edge in the model. Despite this, no dedicated benchmark for EU exists besides isolated evaluations on feature-rich datasets such as Cora and Citeseer where GNNs exploit node features rather than graph structure. As such, we show two systematic limitations in existing evaluations: dedicated EU methods can be slower than retraining anew, and accuracy metrics are uninformative due to the negligible impact of edge removal. To address this gap, we introduce Lethe, the first comprehensive benchmark for Edge Unlearning, featuring 15 methods and 8 datasets. Our benchmark emphasizes large graphs, introduces evaluation tasks of varying difficulty, and proposes empirical measures based on link inference attacks rather than proxy accuracy metrics. We provide a comprehensive empirical analysis, demonstrate the shortcomings of current evaluations, and release Lethe as a reproducible, extensible benchmark.
benchmark - arxiv:2610.21839 · eess.SYAblation-Driven Selective Laser Direct Mass Transfer for Micro-LEDs with High Speed, Enhanced Gap Tolerance, and High PrecisionYuxuan Cao, Yulong Wang, Jürgen Brugger, Zhen Zhang
Micro light-emitting diodes (micro-LEDs) offer superior performance for next-generation displays, yet their scalable manufacturing is severely bottlenecked by the mass transfer process from donor to receiver substrates. The existing laser-induced forward transfer (LIFT) methods are restricted by low throughput, limited selectivity, and long reaction times. In this study, we introduce a novel defocused direct laser ablation transfer (D-DLAT) method utilizing polyimide as a dynamic release layer. Whereas most indirect laser transfer methods trigger chip release followed by passive descent toward the receiver, D-DLAT directly ablates the release layer to actively propel the small-sized chips towards the receiver via ablation shock. By optimizing the ablation profiles and scanning paths, this method achieves high-speed transfer with significantly enhanced precision and robustness. We demonstrate the successful transfer of micro-LEDs ranging from 600 $\unicode{x03BC}$m down to 35 $\unicode{x03BC}$m in size, with a minimum reported placement error of less than 1 $\unicode{x03BC}$m, and a critical transfer gap of up to 50 $\unicode{x03BC}$m. Furthermore, the D-DLAT approach enables full selectivity, facilitating complex patterned transfers and array repitching operations that are essential for full-color display integration. Our findings indicate that D-DLAT effectively overcomes the fundamental limitations of current transfer technologies, demonstrating substantial potential for the scalable and high-yield manufacturing of advanced micro-LED displays. Ultimately, D-DLAT establishes a robust paradigm for the high-speed manipulation and heterogeneous integration of micro-optoelectronic devices, extending its profound implications beyond displays to the broader landscape of advanced micro-assembly and flexible electronics.
manipulationheterogeneous integration - arxiv:2610.21833 · cs.LGPhysics-Aware Distillation of a Grid Foundation Model for Real-Time Edge InferenceVlad Alexandru Sănduleac, Shahariya Rasheed Rayamarakkarveettil Puthanpurayil, Noman Khan, Remus Teodorescu
Grid foundation models (GridFMs) are graph transformers pretrained on a masked power-flow objective, and evaluate power flow orders of magnitude faster than an iterative solver. Their size, however, targets data-centre inference, whereas low-inertia operation increasingly asks for inference at the edge, next to the measurement, at batch size one and within a single synchrophasor (PMU) reporting frame. We distil a 20.1M-parameter, 12-layer GridFM teacher into a 0.62M-parameter, 6-layer student, a $32.5\times$ parameter reduction at half the message-passing depth, using an objective that combines the masked power-flow loss with output, physics-residual, latent-state and attention transfer. The physics-residual term, which matches the teacher's power-residual profile, has not previously been applied to typed power-grid graphs. We evaluate on a renewable variant of the IEEE 118-bus benchmark grid in which 26 of the conventional (synchronous) generators are replaced by inverter-based renewables with battery flexibility and standardised reactive-power limits (IEEE 1547), giving $97\,529$ physically valid (AC-OPF-feasible) operating scenarios, and we ablate all eight combinations of the transfer terms, reporting accuracy against ground truth and fidelity to the teacher separately. Distillation lowers student error on all four predicted quantities, by up to $48\%$ on active-power dispatch, and every distilled student is more accurate than its teacher on reactive-power dispatch. The gain is largely carried by the output matching term. The physics-residual term, in the layer-aggregated form used here, is consistently harmful, which we attribute to the spatial averaging it applies to a bus-level residual field. The resulting student is $32.5\times$ smaller in parameter count and has half the message-passing depth, making it a candidate for subsequent edge-device benchmarking.
benchmark - arxiv:2610.21810 · cs.LGN-Gram Memories at Two Bits: Quantizing Lookup Tables in Large Language ModelsDonghoon Han, Minseop Jung, Dong Woo Kwak, SeungJae Lee +1
Conditional memory expands language-model capacity through selective lookup, but the full embedding table must remain available for every possible input. We introduce Conditional Program (CP), a compressed representation that reconstructs only the rows an input requests. CP represents each row by a short code and a local scale, with a reconstruction program shared across the table. We instantiate this representation with a Gaussian trellis, norm restoration and independently decodable row records. CP's two-bit code lowers output KL by 41--63\% relative to 2-bit round-to-nearest. Analysis across three trained n-gram memory families shows that Gaussian-like coordinate statistics coexist with learned relationships among rows; at matched storage budgets, the shared trellis outperforms the fitted codebooks and low-rank residual variants we test. For DeepSeek, it reduces table storage from 203GB to 51GB, leaving room for larger workloads when the table is kept on GPU.
memory - arxiv:2610.21809 · cs.CVFrom Adults to Children: Benchmarking Foundation Models for Chest X-Ray Classification under Domain ShiftSara Inácio, Hugo Proença, João C. Neves
Foundation models have shown significant potential across diverse fields and have been applied to medical imaging, but their robustness across patient populations remains unclear. In this work, we investigate foundation models' ability to generalize from adult to pediatric chest X-ray classification. To the best of our knowledge, this is the first benchmark to compare multiple foundation models under a common adult-to-pediatric evaluation setting while also analyzing the effect of adaptation depth. We harmonize VinDr-CXR and PediCXR into a five-label classification task and evaluate six foundation models and two CNN baselines. All models are adapted using adult data only and tested in both the in-domain and pediatric domains using linear probing, partial fine-tuning, and full fine-tuning. The benchmark reveals a clear and consistent trend: all models suffer a substantial performance degradation when transferred from adult to pediatric data, with drops exceeding 10 percentage points. While Macro-AUROC values reach approximately 92% on the adult test set, performance decreases to around 78% on the pediatric domain. Another important finding is that increasing the depth of adaptation does not significantly improve cross-population generalization. Moreover, strong in-domain performance does not translate into better performance on pediatric data.
benchmark - arxiv:2610.21808 · cs.AIRoleQ: Role-Aligned Scales for Converting Language Models to TernaryJohannes Knittel, Alec Helbling, Hanspeter Pfister
Weight quantization below four bits shrinks memory and bandwidth at a perplexity that looks acceptable, but often at the cost of factual knowledge and task accuracy. Ternary weights have mostly been obtained by training from scratch or long fine-tuning. We introduce RoleQ, a quantization framework that places scales by the functional roles of Transformer units rather than by the layout of the weight matrix, as conventional formats do. Vectors that read from the residual stream take one scale per output unit; those that write to it, one per input unit. The rule is structural, needs no search, calibration or stored state, and holds on integer and ternary grids. In a rotated basis it matches a per-layer axis search at half the solver work, and at ternary it cuts perplexity by up to half against layout-based placement in two families at every size from 0.6B to 14B, where per-row scales leave MMLU at chance. Our lightweight post-training pipeline converts an existing checkpoint to 1.59 bits per weight in a few GPU-hours, a fraction of the cost of learned ternarization, and leads every post-training ternary method at matched or lower memory-resident bits on perplexity, zero-shot accuracy and MMLU, most widely on knowledge. A sparse second ternary plane placed by a per-unit Fisher saliency keeps the matmul multiplication-free and recovers what ternary loses first: at 8B, a fraction of a bit returns about half the MMLU that conversion costs, well ahead of random or Hessian-based placement at equal bits.
memorypost-training - arxiv:2610.21805 · cs.LGCIM-Embed: Learning Instruction-Conditioned Multimodal Embeddings for ClusteringYinghua Xu, Pengyang Xie, Yanxiang Zhou, Chuanzhi Tong +3
Multimodal objects can require different partitions under different instructions, as when products in an e-commerce catalog are grouped by color, material, or both. Retrieval-oriented embeddings may not recover these partitions, while generative clustering often relies on intermediate descriptions or iterative decisions. This paper presents CIM-Embed, a framework for learning instruction-conditioned multimodal embeddings for clustering. Supervised contrastive learning uses same-group positives and negatives from other groups, optionally including other tasks; heterogeneous structured batches accommodate tasks with different group counts. At inference, the model maps each object's image, text, and instruction directly to an embedding; HDBSCAN then forms a partition without the ground-truth cluster count. We construct ProductMIC from product metadata in Amazon Reviews 2023, comprising training data and an evaluation benchmark for instruction-conditioned clustering. The benchmark contains 5,373 products and 449 single-attribute and compositional grouping tasks spanning 50 leaf categories excluded from our training. With full-dimensional embeddings and HDBSCAN, direct within-task SupCon training achieves a mean per-task V-measure of 95.01, compared with 53.06 for WeMM 9B, the strongest general-purpose embedding baseline evaluated. Frozen product-trained encoders also improve out-of-domain conditional grouping, while controlled transfer experiments distinguish partition transfer from instruction response. We will release the code and dataset upon acceptance.
benchmark - arxiv:2610.21804 · cs.AICo2In: a Bi-level Memory Incremental Learning Framework with Knowledge Encoding, Consolidation, and IntegrationWenju Sun, Qingyong Li, Boyang Li, Wen Wang +3
Incremental learning (IL) aims to continually acquire new knowledge (plasticity) while retaining previously learned information (stability). However, striking a balance between plasticity and stability remains a significant challenge for intelligent systems. The human brain achieves exceptional balance, owing to various memory units that collaboratively encode and store information. Inspired by human memory mechanisms, this paper introduces an IL framework with knowledge enCoding, Consolidation, and Integration (Co2In). Co2In is designed as a bi-level memory architecture: a working memory for adaptive knowledge acquisition and a long-term memory dedicated to the persistent retention of information. The two memory modules work cooperatively during IL. The working memory first learns from new data to encode knowledge into parameters. Subsequently, Co2In performs a consolidation process to identify underlying patterns in the learned parameters and re-express them into a compact knowledge representation. Next, the knowledge representation and the identified patterns are transformed into separate network layers and integrated into the long-term memory. These designs empower Co2In to accumulate knowledge with high plasticity and stability. We evaluate Co2In on CIFAR-10, CIFAR-100, and Tiny-ImageNet under exemplar-free Class-IL and Task-IL settings. Experimental results show that Co2In achieves state-of-the-art performance with efficient memory consumption.
memorymemory modulememory architecture - arxiv:2610.21802 · cs.CVTeaching DINOv3 About Partial 3D Geometry: A Self-Supervised Geometry-Aware ApproachViktoria Ehm, Dongliang Cao, Riccardo Marin, Daniel Scholz +3
Partial shape matching is a crucial yet underexplored problem in 3D vision, with significant relevance to real-world scenarios where shapes are often only partially observed. Existing feature descriptors face difficulties in this setting, as traditional representations either struggle with the boundaries of partial shapes or heavily depend on the shape's spatial position. While existing approaches have employed DINO features for partial shape matching, these features are not inherently suited for handling partial observations. In this work, we propose a method to refine DINO features using LoRA-based self-supervised learning, enabling the generation of feature descriptors that are robust to partiality. Our features substantially improve performance on partial shape matching compared to traditional or vision foundation features. Additionally, when integrated into existing partial shape matching pipelines, we achieve state-of-the-art results on partial shape matching and left-right prediction benchmarks.
benchmark - arxiv:2610.21799 · cs.CLBehavioral Forgetting Is Not Privacy in Machine UnlearningTyler Lizzo, Larry Heck
Machine unlearning aims to remove targeted information from large language models, but existing evaluations often measure forgetting only on a designated benchmark distribution. We argue that such behavioral forgetting is an insufficient standard for privacy-oriented unlearning, and that evaluation should instead test whether forgetting persists across the access paths through which the same information may be recovered. We demonstrate this with a controlled case study of cross-lingual querying: across multiple multilingual model families, forgetting varies substantially when semantically equivalent information is queried in different languages and scripts. A single access path is therefore enough to expose a real gap between behavioral forgetting and information inaccessibility, motivating broader evaluation of access-path generalization and adversarial recoverability alongside in-distribution forgetting.
benchmark - arxiv:2610.21798 · cs.ROCross-Embodiment Causal Transfer from Human Videos to Robot ManipulationXingyu Zhang, Yingchao Li, Peizheng Guo, Jie Hu +2
Scaling robot manipulation models requires large amounts of interaction data, yet collecting diverse real-robot experience remains costly. Human manipulation videos, in contrast, provide abundant and diverse task experience that could potentially benefit robot learning. However, differences in morphology, action spaces, and control mechanisms make human motions unsuitable as direct supervision for robot actions, raising a scientific question: what knowledge in human videos should be transferred across embodiments to robot manipulation? To address this question, we analyze subtask transfer through the lens of causal beams, focusing on which contextual conditions and interaction dependencies should be retained to support local task effects. Based on this perspective, we propose Causal-Beam-Guided Subtask Transfer (CBST). CBST learns manipulation-related representations from human videos and grounds them in task-level semantic variables using robot demonstrations. Within a predefined manipulation model, causal-beam constraints guide contextual information selection by requiring retained assignments to sustain local mechanism outputs while preserving nontrivial responsiveness to retained inputs. The selected semantic information conditions a world-action model, allowing the robot to generate actions from its own observations without requiring human-robot motion correspondence. Experiments on robot manipulation benchmarks show improvements in task success over the corresponding world-action model baselines.
manipulationbenchmark - arxiv:2610.21794 · cs.CVMedPAO-VL: Interpretable and Protocol-Compliant Agent for Structured Radiology Report GenerationShrish Shrinath Vaidya, Gowthamaan Palani, Minmini Selvam, Gokulraja Srinivasaraja +1
Despite the remarkable generalization capabilities of Large Language Models, generating clinically accurate and structured radiology reports from medical images remains a significant challenge. Current end-to-end approaches function as black boxes, prone to factual hallucinations and generating ungrounded findings. This lack of reliability and failure to adhere to domain-specific structuring protocols, fundamentally limits their clinical utility. We present MedPAO-VL, a multimodal agent that shifts the paradigm from monolithic generation to explainable, iterative, protocol-compliant structured report generation. Unlike deterministic models, our plan-act-observe loop functions as a goal-oriented agent: it first perceives the Chest X-Ray through an explainable vision module designed to capture inherent diagnostic uncertainty, then iteratively refines the generated report by observing semantic discrepancies and self-correcting its generation for better alignment with visually detected clinical findings. Evaluation of our framework demonstrates that this agentic reasoning improves clinical correctness and completeness, achieving an expert-rated accuracy score of 3.5/5 and a structural score of 4.3/5. By prioritizing pathology specific explainability and iterative refinement over opaque end-to-end mapping, MedPAO-VL offers a transparent and clinically viable alternative for automated structured radiology report generation.
agentagenticiterative refinement - arxiv:2610.21792 · cs.ROTrajectory-based Diffusion from One-shot Human Video for Generalizable ManipulationGuoping Pan, Xinmiao Yu, Wannan Cao, Jincheng Li +4
Imitation learning enables robots to perform complex tasks by leveraging human-collected demonstrations. However, most existing approaches require labor-intensive data collection with additional teleoperation systems, which hinders scalability. Human videos represent a natural source of data that implicitly contain knowledge of manipulation behaviors. However, transferring them to robots remains challenging due to the visual domain gap and the absence of explicit action labels. In this work, we introduce TDV, a trajectory-based framework that leverages data generation to enable generalizable manipulation. TDV focuses on task-relevant objects by leveraging plug-and-play 6D pose estimation techniques to extract object trajectories from human videos, thereby mitigating the impact of irrelevant backgrounds and varying viewpoints. As a modality-agnostic intermediate representation, trajectories are inherently independent of the visual domains and different embodiments, enabling efficient data generation. Our method outperforms previous approaches on RLBench simulation tasks, achieving near 100\% success rates across seven tasks. In real-world experiments, TDV enables cross-embodiment generalization manipulation from one-shot human demonstrations by leveraging efficient trajectory generation techniques.
manipulationteleoperation - arxiv:2610.21791 · cs.AIRSIFlow: Adaptive Self-Improvement Workflows with an Evolving Meta-AgentHuan Zhang, Mingju Chen, Dongxu Zhou, Can Lv +3
Large language model (LLM) agents have demonstrated strong capabilities in complex tasks, including code generation, tool use, and multi-hop question answering. Existing self-improvement methods enhance agent performance through model optimization, harness evolution, and executable artifact refinement, while recursive self-improvement (RSI) further extends optimization to the improvement mechanisms themselves. However, adapting self-improvement workflows to evolving performance bottlenecks while improving workflow orchestration through execution feedback remains a key challenge. We propose \textbf{RSIFlow}, a \textbf{R}ecursive \textbf{S}elf-\textbf{I}mprovement framework that integrates adaptive work\textbf{flow} orchestration with meta-agent capability evolution. RSIFlow diagnoses performance bottlenecks from task-agent trajectories and environmental feedback to dynamically select component-specific improvement workflows. The meta-agent then distills workflow outcomes into component-specific skills and cross-component principles, evolving its own harness to improve subsequent workflow orchestration. Experiments across six benchmarks and two backbone models show that RSIFlow achieves the SOTA average performance on all benchmarks. Ablation studies further demonstrate the contributions of adaptive workflow orchestration and meta-harness evolution. Code is available at https://github.com/1HUAN1/RSI-Flow.
agenttool useself-improvementbenchmark - arxiv:2610.21789 · cs.AIIn the Guise of Correction: How Framing and Cultural Adaptation Erode LLM Safety on Real-World Criminal CasesMohsinul Kabir, Md. Tariquzzaman, Karam S. R. Elradie, Polydoros Giannouris +3
Existing safety benchmarks rely largely on synthetic prompts with loosely grounded taxonomies and offer little focused analysis of severe crime categories. We introduce GUISE, a benchmark of criminal-assistance scenarios sourced from real-world cases and organised under a taxonomy from the UNODC International Classification of Crime for Statistical Purposes (ICCS). Each scenario is rendered under three linguistic framings and evaluated across five distinct cultures. Using our proposed two-tier evaluation framework, which judges both harmfulness and operational uplift, we find that: (i) framing alone drives unsafe compliance, with the mean unsafe compliance rate rising from 28.9% under direct requests to 66.4% under correction seeking (peaking at 97.3%; p < 0.05), as models correcting deliberately distorted details supply executable criminal steps; (ii) culturally adapted prompts elicit significantly more unsafe compliance than literal translations of the same requests in all four non-English languages (+7 to +18 points), a risk that translation-based benchmarks miss; and (iii) unsafe responses from the least-safe models carry strong Anglocentric registers, and the extent of this leakage increases monotonically with response harmfulness ($ρ$ = 0.46 to 0.63 across languages, all p < 0.001). GUISE offers a grounded resource and fine-grained diagnostics for strengthening refusal alignment across languages and cultures. We publish our code and data to facilitate future research.
benchmarkevaluation framework - arxiv:2610.21783 · cs.CVLearning to Think and Explore in 3D WorldYijiang Li, Bingyang Wang, Nuno Vasconcelos
We study how vision-language models (VLMs) can learn to think and actively explore in 3D space. Instead of providing multi-view images, videos, or 3D representations to the model, we preserve the native visual encoding of VLMs, i.e., a single image, one at a time, just like humans, and treat the 3D representation only as an environment with which the model can interact through actions (e.g., move left) and observations (rendered images). Each rollout begins from an initial first-person view and a visual question. The model can then iteratively reason, navigate, and reorient its viewpoint before committing to a final answer. Following an initial trajectory-supervised finetuning, we apply reinforcement learning (RL) with rewards driven primarily by final-answer correctness. Crucially, RL provides no direct supervision over which viewpoints the model should visit or which exploration actions it should take. Instead, the model must learn an effective exploration policy solely from the extent to which its interactions improve performance. Across six 3D benchmarks, our model ACT3D achieves an average accuracy of 53.2% that improves its base Qwen model by 22.0% and surpasses proprietary Gemini 3.6 Flash. ACT3D also outperforms existing open-source 3D methods while using 50-70% fewer input frames, and generalizes to unseen domains with an average improvement of 5.9% over its supervised initialization. Further analysis reveals adaptive, question-conditioned exploration. Performance gains over the base model increase from 0.6% on questions answerable from the initial view to 16.3% on those requiring extensive exploration. ACT3D searches for occluded targets, avoids unnecessary exploration, and acquires complementary viewpoints for spatial comparisons. These findings demonstrate that outcome-based RL enables VLMs not only to reason over visual evidence, but also to learn when and how to acquire it.
benchmark - arxiv:2610.21779 · cs.CVFuseAdapt: Parameter-Efficient Multimodal Fusion for Semantic Segmentation with Missing ModalitiesXin Zhang, Robby T. Tan
Multi-modal semantic segmentation benefits from complementary sensors, but this benefit is fragile when some modalities are degraded or missing at test time. Many existing methods rely on dedicated fusion modules or multi-branch cross-modal encoders, and the fusion often becomes unreliable when modalities are absent, leading to substantial performance drops. We propose \textbf{FuseAdapt}, a parameter-efficient framework that performs fusion within the adaptation space of a frozen vision foundation model (VFM), exploring whether VFM priors can serve as an effective shared backbone for multi-modal segmentation. Instead of attaching a standalone fusion network, FuseAdapt assigns each modality a lightweight adaptation branch and performs \emph{adaptation-space fusion} by keeping Q/V low-rank adaptation private to each modality and routing only FFN low-rank residual activations across modalities inside the frozen backbone. The pretrained VFM remains fixed while modality-specific corrections and explicit cross-modal interaction are learned through the adaptation paths. To handle missing modalities, we further introduce \textbf{Predictive Modality Modeling (PMM)}, a direct representation-level objective for cross-modal redundancy. During training, PMM masks a target modality and predicts its embedding from the visible ones, supervised by a teacher target computed under full-modality input. PMM is used only during training and adds no inference-time overhead. Across three multi-modal segmentation benchmarks (MCubeS, DELIVER, MUSES), FuseAdapt consistently improves both modality-complete and modality-incomplete segmentation with a small trainable budget, yielding up to +3.01 mIoU under full modalities and +12.56 mIoU under missing modalities over prior state-of-the-art. Project page: https://devinxzhang.github.io/FuseAdapt-Page/.
benchmark - arxiv:2610.21767 · cs.AIFrom Research Goals to Scientific Closure: Evaluating Agents Through Evidence-Supported Physics ResearchYijia Mao, Song Dai, Tao Chen, Xiangyu Lu +13
Can AI agents organize scientific investigations that lead to well-supported conclusions? To evaluate this capability, we introduce PhysClosureBench, a benchmark of 50 expert-verified, goal-specified research tasks spanning strong-field, attosecond, and neighboring physics areas. Each task specifies the research objective and essential scientific constraints, while leaving the agent to organize the investigation and determine how to generate and validate the evidence needed to answer the question. A 20-criterion Atomic Score measures local scientific accomplishment, while Research Closure (RC0--RC4) assesses how fully the investigation resolves the specified research question based on whether its indispensable evidence requirements and applicable scope conditions are satisfied. We evaluate seven frontier LLMs, where Claude-Opus-5 achieves the highest mean Atomic Score (6.29/10) but reaches RC4 in only 32.7% of attempts. Across the benchmark, 36.2% of attempts scoring at least 8/10 still fall short of RC4, while all seven models reach RC4 on substantially fewer tasks in all three runs than in at least one. Case-level analysis further illustrates how misinterpretation of generated results and unresolved physical applicability limits can lead to incomplete closure. Together, these results reveal a persistent gap between AI agents' local scientific accomplishments and their ability to achieve complete research closure.
agentai agentbenchmark - arxiv:2610.21766 · physics.opticsWatt-level single-cavity dual-comb laser at 2.1 $μ$m wavelengthMykyta Redkin, Yicheng Wang, Marcel van Delden, Robin Löscher +3
Compact ultrafast laser combs around the 2 $μ$m atmospheric transmission window are of growing interest for direct spectroscopy and as drivers for nonlinear conversion to otherwise hard-to-reach spectral regions. Dual-comb systems in this region remain very scarce: existing implementations rely mostly on nonlinear conversion, resulting in complex systems with poor conversion efficiency and overall low average and peak power. In this paper, we report on a watt-level dual-comb laser directly emitting at 2.1 $μ$m. The system is based on a single-cavity, soft-aperture Kerr-lens mode-locked (KLM) laser using a broadband holmium (Ho)-doped CaAlGdO$_4$ (CALGO) gain crystal. The laser delivers a total output power exceeding 2 W (1 W per comb) with a sub-100-fs pulse duration (50 nm bandwidth) at a repetition rate of 96.5 MHz, corresponding to a pulse energy of 10 nJ and a peak power of 100 kW per comb. To assess the source's potential for future spectroscopic applications, we characterize the free-running dual comb in detail for frequency stability, phase noise, and relative intensity noise (RIN), and benchmark the achievable spectral resolution by measuring spectral modulations imposed by Fabry-Pérot etalons. Our system combines the compactness, high efficiency, and passive mutual coherence of a single-cavity oscillator with watt-level average power at 2.1 $μ$m, making it a promising source for environmental sensing in the atmospheric transmission window, or as a driver for nonlinear frequency conversion, for example, to the MIR and THz regions.
benchmark - arxiv:2610.21764 · cs.LGQ($γ_H$): Entropy-Guided Horizon Mixing in Multi-Discount Q-LearningArwa Albelaihi, Abdulrahman Altahhan
A fixed discount factor in Q-learning imposes a single planning horizon throughout training, despite the differing demands of early exploration and long-term credit assignment. We propose Q($γ_H$), a multi-horizon Q-learning framework that maintains multiple discount-specific heads and adaptively combines them at control time using an entropy-derived confidence signal. The resulting state-dependent mixing weights shift control from shorter- to longer-horizon heads as policy confidence increases, enabling adaptive horizon mixing without modifying individual Bellman updates. Across classical control and deep reinforcement-learning benchmarks, Q($γ_H$) shows its strongest improvements on delayed-credit tasks relative to strong fixed-discount baselines. By decoupling horizon adaptation from the underlying learning updates, Q($γ_H$) provides a lightweight framework for adaptive planning-horizon control in reinforcement learning, avoiding the need for manual discount-factor selection while remaining compatible with standard value-learning methods.
benchmark - arxiv:2610.21763 · cs.CLBias Mitigation in Audio-Enabled LLMs through Adversarial Training and Distillation MethodsAdnan Al Ali, Jindřich Libovický, Anne Lauscher
Audio-enabled large language models (AuLLMs) allow users to interact with them simply by talking to them, without the need for an intermediate transcription. This expands the model's accessibility and capabilities by encoding non-verbal cues. However, using non-verbal cues may lead to biased behaviour based on the attributes inferred (or assumed) from them. In this work, we focus on the gender attribute and show that models' perception of it can be unlearned using adversarial methods without a loss in performance on other benchmarked tasks. Such models are useful in situations where gender inference is unwanted. However, in other practical settings, gender perception (without bias) may be desired.As a less radical measure, we distil the gender-agnostic model, creating a model that perceives gender but responds similarly to the gender-agnostic teacher, unless explicitly asked.We apply our newly proposed methods to three AuLLMs and obtain satisfactory results in the gender-unlearning part and partially satisfactory results in the distilling part.
benchmark - arxiv:2610.21760 · cs.CVBeyond Generation: Using Multimodal LLMs as Detectors of LLM-Generated Deepfake Facial ImagesOlivia Salamon, Virginia N. L. Franqueira, Shujun Li
As image generation models become increasingly capable of producing highly realistic content, traditional forensic techniques are struggling to keep up. This paper investigates whether MLLMs can instead be used as a form of forensic analysis by identifying synthetic media. Specifically, the study aims to determine whether these models detect fake images by recognising physical inconsistencies, or whether they rely on digital artefacts such as embedded watermarks.Six MLLMs: GPT-4o, GPT-5, GPT-5.2, Gemini 2.5, Gemini 3, and Seed 1.6, are benchmarked using a dataset of 400 images. Fifty images are authentic photographs, while the remaining 350 are generated using seven leading image models: Gemini 3, Gemini 2.5, Flux.2 Pro, GPT-Image-1, GPT-Image-1.5, Seedream 4.0, and Seedream 4.5. The evaluation is structured in three stages: single-image classification, pairwise comparison, and a focused "self-detection" experiment on Gemini. Several MLLMs achieved a high detection accuracy, with GPT-5.2 reaching a peak of 97.9% in the pairwise experiment, against specific generators and an overall mean pairwise accuracy of 82.5% by identifying structural or physical inconsistencies within images. However, results from the self-detection experiment indicate that the presence of invisible SynthID watermarks leads to a 10% reduction in Gemini's detection accuracy compared to the clean control group. One possible explanation is that SynthID embedding reduces some of the visible artefacts the models rely on for detection.These results imply that MLLMs can detect structural artefacts without relying on watermarks, but performance varies substantially across model generations. The findings from comparison C5 also suggest that SynthID embedding may suppress some of the structural artefacts used during detection.
benchmark - arxiv:2610.21755 · cs.RODeciCAD: Harness-Grounded Learning of Action Policies for CAD AgentsYufeng Peng, Yixin Zhang, Ziyi Liu, Zhikun Xu +4
Large Language Model agents operate through closed-loop action policies whose decision space is defined by an agent harness. We investigate how this operating space shapes agent behavior and policy learning in computer-aided design (CAD), a professional domain characterized by persistent state, geometric constraints, and long-horizon procedures. We instantiate this perspective in DeciCAD, a domain-native harness that exposes semantic CAD actions, decision-oriented observations, environment feedback, and persistent state transitions. Building on this harness, we develop a learning pipeline that extracts causal decision supervision from interaction trajectories and improves the policy through supervised fine-tuning (SFT) and geometry-grounded online reinforcement learning (RL). Across four CAD benchmarks, DeciCAD achieves the best F1 and intersection over union (IoU) while reducing model-output tokens by 45--88% relative to a general-purpose coding harness. Under matched supervision, its SFT policy improves F1 from 0.6883 to 0.7437 and outperforms the coding policy; subsequent online RL further raises geometry availability from 66.5% to 73.5% and F1 to 0.7880. These results demonstrate that the decision space induced by a harness materially affects both agent execution and policy learning in CAD.
persistent stateagentbenchmark - arxiv:2610.21742 · cs.LGSparse Federated Echo State Networks for Time-Series Classification via Smooth-L1 RegularizationGengcheng Lyu, Manuel Röder, Frank-Michael Schleif
Echo State Networks (ESNs), as lightweight implementations of reservoir computing, provide efficient solutions for temporal data processing in resource-constrained environments. While standard ESNs offer computational efficiency, their dense readout layers inflate the stored model and the per-sample inference cost. This paper introduces a Smooth-L1 regularized ESN (SL1ESN) that promotes sparsity in the readout layer while preserving differentiability for gradient-based optimization. The method employs a twice-differentiable Smooth-L1 approximation compatible with Newton's method, which we exploit both for centralized training and for a second-order (Newton-step) federated aggregation scheme. Under a single fixed shrinkage threshold shared by every dataset and both settings, experiments on seven time-series classification benchmarks show that SL1ESN learns readouts that are $77$--$95\%$ sparse in the centralized setting and $75$--$95\%$ sparse in the federated one, while staying within $0.7$ accuracy points of the strongest dense baseline on five of the seven benchmarks in the centralized setting and within $1.6$ points on the same five in the federated one, and the second-order federated update converges in only a few rounds. The resulting sparsity compresses the stored model, the per-sample inference cost, and the server-to-client broadcast by a comparable factor. These results demonstrate the practical value of sparse readouts for edge computing and IoT deployments.
benchmark - arxiv:2610.21741 · cs.ROComfortably Stuck: Evaluation Pitfalls in Bench2Drive's Closed-Loop ProtocolThomas Schwarzfischer, Emad Aghajanzadeh, Moritz Thoma, Nael Fasfous +4
Closed-loop evaluation is widely regarded as a more faithful way to assess end-to-end (E2E) driving models than open-loop evaluation, and Bench2Drive, built on the CARLA simulator, has become the dominant benchmark for it. We show that its evaluation protocol itself contains pitfalls that can misrepresent the models scored on it. Bench2Drive's comfort metric, adapted from nuPlan, has several implementation errors and a segmentation scheme that rewards standing still, so it mostly measures how long a vehicle stays stationary, correlating with stationary time at r = 0.97 across eight models. We propose a corrected, distance-based metric that cuts that correlation to r = 0.06. Under it, the strong driving-score/comfort tradeoff the original metric reports is no longer detectable once models are compared at matched driving speed, which suggests that comfort trades against how fast a model drives rather than against how well it drives. We also uncover two pitfalls in how the harness supplies navigation signals: the default controller can replace a model's own steering with an angle toward the route target point, and the target point supplied at evaluation time can differ substantially from what a model saw in training. We demonstrate both on UniAD and VAD, where their effects differ sharply, and argue that future closed-loop evaluations should check for them. We will release all fixes as standalone code, so corrected comfort scores can be obtained from logged trajectories without re-running closed-loop simulation.
benchmarkevaluation protocol - arxiv:2610.21739 · cs.LGReliable Knowledge Acquisition for Continual Learning AgentsYixu Huang, Jiaqi W. Ma
As agents powered by foundation models move into open-ended environments, continual learning from experience becomes essential. For agents built on models with fixed parameters, this learning requires reusing prior task experience in context to improve performance and efficiency on future tasks. Existing memory and experience-learning systems typically focus on either retained facts or procedures, and lack a mechanism for establishing the reliability of claims extracted from the ongoing task stream. Indeed, we find that approximately 12--33\% of entries produced by representative systems harm future task performance when reused. To address these limitations, we propose \textbf{RIKA} (Reliable Knowledge Acquisition), which combines hierarchical extraction of factual, conceptual, procedural, and metacognitive knowledge with Evidence-based Knowledge Admission, which uses source evidence and subsequent experience to govern which claims enter the task agent's context, without additional validation tasks. Across six task streams from Continual Learning Bench and four model backbones, RIKA achieves the highest aggregate score on every backbone.
memory - arxiv:2610.21735 · cs.ROUni-PNP: Unified Prehensile and Non-prehensile Policy Learning for Generalizable Object Pose ReachingWenxuan Zhu, Yuze Wang, Jiangran Lyu, Jiayi Chen +5
Object manipulation often requires combining non-prehensile actions with grasping. However, existing approaches typically develop these capabilities separately and integrate them by sequencing distinct primitives through prescribed intermediate objectives, limiting the coordination of contact motions and grasping. Their integration also remains sensitive to object and gripper geometry and depends on particular environmental supports, complicating real-world deployment. We present Uni-PNP, a unified policy that jointly learns prehensile and non-prehensile manipulation for object pose reaching across diverse objects and grippers. To achieve this, we first pretrain a geometry encoder to predict contact-induced object motion using a contact-dynamics dataset constructed from simulated interactions between diverse objects and grippers. We then train the policy in a two-stage reinforcement learning curriculum from graspable objects to ungraspable peers, encouraging the discovery of non-prehensile interactions and their integration with grasping. Experiments show 76.39% success on hundreds of seen objects with grippers, and 57.91% on unseen peers, outperforming the strongest pose-reaching baseline by 20.52 and 17.79 percentage points, respectively. We also analyze emergent behaviors, including nonprehensile and prehensile repositioning, grasp acquisition that requires grasping after tipping, and table-assisted reorientation while grasping, and transfer to the real-world scenario. Project page available at lgswdn.github.io/Uni-PNP.
manipulationgrippergrasp - arxiv:2610.21708 · cs.LGHuman Interventions as Labels: Localizing, Steering, and (Sometimes) Editing the Computations Behind LLM Agent FailuresCesare Zavattari, Alessandro Tommasi, Giuseppe Prencipe
When an LLM agent misbehaves, a human steps in, and that intervention is usually discarded or replayed as training data. We treat it instead as a measurement: a label pointing at the internal computation that failed. We study it in a controlled agent environment, on one failure (asking a clarifying question when the request was already unambiguous), with labels from a rule-based overseer validated against blind human judgment. Claims cover eligible models, which both over-ask and sometimes act on matched unambiguous requests: five of thirteen qualified (Qwen2.5-3B, gemma-2-2b, OLMo-3-7B, Mistral-Nemo-12B, Phi-4). Four findings. First, the ask-vs-act decision is causally localizable in every eligible family: a direction learned from intervention labels flips 72-100% of held-out over-asks at a definite residual-stream depth, with site and dose chosen on training data, and random directions under the same procedure flip at most one episode. Second, over-asking has two mechanistic signatures: in four families the model demonstrably knows the request is unambiguous (the belief is decodable, causally inert and orthogonal to the decision) and asks anyway; in the fifth, the rare over-asks track different world-representations. Third, steering gated on a probe readout removes most of the simulated oversight burden while legitimate questions survive (bootstrap 95% intervals exclude zero); a prompt instruction makes it worse. Fourth, consolidating the fix into the weights yields a map, not a rule: on one model a LoRA confined to the causal band passes every criterion fixed in advance while behavior cloning on identical data breaks retention; on three families the reverse holds; on the fifth the outcome crosses the boundary across seeds. A geometric rule written down in advance to predict the regime failed on two out-of-family tests; we report the failure and stop rather than fit a replacement.
agentllm agent - arxiv:2610.21696 · cs.CVBridging the Markerless Tracking Validation Gap: Intraoperative X-ray Workflow IntegrationConnor Daly, Salvatore Russo, Jinendra Ekanayake, Daniel S. Elson +1
Markerless intraoperative tracking of spinal anatomy via RGB-D sensors is a promising alternative to invasive bone-anchored fiducials, yet no existing clinical dataset provides independent subsurface ground truth for validating these methods. We address this gap with a clinically viable pipeline that derives per-vertebra 6-DoF pose from routine intraoperative C-arm fluoroscopy, co-registered to the RGB-D camera frame through a multimodal fiducial array detectable in both imaging modalities. Critically, the pipeline requires no additional radiation, instrumentation, or deviation from standard surgical workflows, relying solely on the fluoroscopic images already acquired during routine level checks. A calibrated transform chain links the vertebral pose recovered via multi-view 2D-3D CT-to-fluoroscopy registration to the RGB-D coordinate frame, enabling direct comparison of any markerless registration estimate against a subsurface reference. Ex vivo validation against optically tracked ovine vertebrae confirms end-to-end accuracy of 2.06 +- 0.29 mm and 3.35 +- 0.93 degrees, sufficient to serve as independent ground truth for methods with substantially larger errors. We demonstrate the pipeline's utility by evaluating two markerless registration strategies on clinical data spanning 11 vertebrae across 1,500 frames from three surgical sequences: the first such evaluation against an independent subsurface reference in a clinical setting. The proposed framework provides reusable infrastructure for benchmarking emerging markerless tracking methods on real surgical data.
benchmark - arxiv:2610.21686 · cs.ROMAIA: Anticipating Interaction Anomalies in Contact-Rich Manipulation via Multimodal World ModelsYi Qiao, Maolin Zheng, Haoran Sun, Lujia Dong +1
Reliable contact-rich manipulation requires detecting execution deviations early enough for corrective action. Yet most monitoring methods are reactive, identifying anomalies only after they become observable. Anticipating such anomalies is challenging because visual observations only partially reveal physical interaction, while planned actions describe intended motion rather than their actual physical consequences. We address this problem with MAIA, a multimodal framework for anomaly anticipation. Its vision-action-force world model learns nominal interaction dynamics and predicts the visual and force responses of planned actions, while a lightweight anomaly head predicts impending interaction anomalies. To enable systematic evaluation, we introduce CRIA, a real-world multimodal dataset of 2,852 trajectories spanning 18 contact-rich manipulation tasks, with frame-level annotations of anomaly onset and evolution. Experiments on both seen and novel tasks show that MAIA achieves more accurate anomaly discrimination than representative baselines while providing earlier warnings, with bACC scores of 0.765 and 0.710 and median anticipation lead times of 28.0 and 25.1 steps, respectively. Moreover, using these predictions for anomaly-aware action selection improves end-to-end task success from 52.6% to 61.5% without retraining the underlying manipulation policy. These results highlight the value of modeling future physical consequences for proactive anomaly anticipation.
manipulationworld model - arxiv:2610.21685 · cs.LGMARS: A Momentum-Aware Adaptive Rating System for Personalized AssessmentVikrant Sahu, Khethavath Sunil Naik, Gagan Raj Gupta
Adaptive assessment systems personalize learning by estimating learner proficiency and recommending suitable questions. However, many existing approaches rely on fixed difficulty labels, static rating updates, or black-box prediction models that provide limited control over rating stability, interpretability, and time-constrained routing. This paper proposes MARS, a Momentum-Aware Adaptive Rating System for interpretable adaptive assessment. MARS represents the question bank as a directed weighted knowledge graph and updates learner proficiency using dynamic rating sensitivity, response quality, and momentum-aware learner velocity. We also formulate a time-budgeted mark-maximization objective in which the policy selects questions within a fixed test duration under negative marking. Experiments on ASSISTments 2009 and MARS-SimExam evaluate MARS for prediction, rating stability, and adaptive routing. Results show that MARS provides smoother rating updates than Fixed Elo and achieves the highest score under time-budgeted negative-marking routing, improving the total score by 8.09 over the strongest baseline and reducing overtime violations from 50% to 5.6%. These results suggest that MARS is suitable for real-time, interpretable, and deployment-friendly adaptive assessment.
knowledge graph - arxiv:2610.21681 · cs.CLData Strategies for Large Language Models: A Unified Perspective on the Training LifecycleYuanjian Xu, Jianing Hao, Tianze Sun, Yongyao Duan +6
Data strategy plays a central role in training large language models (LLMs). It governs the selection, weighting, and scheduling of training data, thereby determining the capabilities a model can acquire from a fixed data budget. We treat high-quality data as having high training-dependent utility, which depends on the target objective, the model, and its training state rather than on the text alone. Research in this area now spans pretraining and post-training, but related ideas use inconsistent terminology and evaluation protocols. This work reviews data strategies across the training lifecycle and formulates them as policies that allocate a training budget over candidate examples. We organise methods by their decision granularity, model dependence, and temporal adaptivity, then compare their computational requirements and empirical support. Fine-grained or state-dependent policies offer more control but incur additional scoring, search, or rollout costs. Evidence for transfer across model scales and comparable end-to-end cost remains limited. A continuously updated list of papers, along with their public implementations, is available at https://github.com/xuyj233/data_strategy_survey.
post-trainingevaluation protocol - arxiv:2610.21679 · cs.CLNo Single Tokenizer Feature Reliably Predicts Downstream Language Model PerformanceAntonio Lopardo, Luis Frentzen Salim, Shikhar Srivastava, Kevin Jiang +2
Tokenization is a critical component of the language modeling pipeline, but it is poorly understood. Tokenizers can impact downstream performance, but there is currently mixed evidence about which tokenizer features predict a better language model without requiring a full training run. In this paper, we analyze approximately 200 language models trained on nine different languages. We evaluate both models and their tokenizers and test whether there is a relationship between any proposed intrinsic tokenizer evaluations and model performance, as measured by a wide variety of benchmarks. Despite the quantity and variety of proposed intrinsic tokenizer metrics, we find that none consistently predicts downstream performance across model families, tasks, and languages. This motivates the need for novel tokenizer evaluation metrics and approaches. We propose novel linguistically grounded metric introduced in this work shows that new metrics can capture aspects of tokenizer quality that existing metrics miss, though their predictive power is likely to vary across model families, languages, and tasks.
benchmark - arxiv:2610.21671 · cs.CLEnhancing Multimodal RLVR with Pairwise Token Contrast and Adaptive Trajectory ConfidenceYiran Zhang, Hanbo Huang, Jielan Li, Zhen Zhao +2
Reinforcement learning with verifiable rewards (RLVR) improves multimodal reasoning, but group-relative policy optimization broadcasts a single sequence-level advantage to every token of a response, although errors usually localize to a few decisive steps. Recent self-distilled variants refine credit with a teacher--student likelihood gap. However, this signal cannot tell whether it relies on the image or on a language shortcut, and is applied uniformly even when the teacher is uninformative. To address this, we propose Pairwise token contrast and Adaptive trajectory ConfidencE (PACE). Pairwise token contrast scores each response under four teacher views, crossing a correct or incorrect answer hint with the image present or removed, and compares the two hints for every token, which cancels shifts shared by both hints and reveals whether the token's support relies on the image. Adaptive trajectory confidence measures how well the teacher predicts each response given the correct hint and the image, normalizes this score within the batch, and uses it to set how far the token contrast redistributes credit within that response. PACE achieves the best average accuracy on three backbones and ranks first on all evaluated multimodal reasoning benchmarks with Qwen3-VL-8B-Instruct. Experiments show that both components matter: replacing pairwise token contrast with a single-view signal underperforms our method, while adaptive trajectory confidence adds further improvement.
benchmark - arxiv:2610.21667 · cs.CVFusionAudit: Pathway-Conditioned Robustness Auditing for Native Multimodal ModelsJiaming Zhang, Zherui Li, Siqi Guo, Junhao Dong +2
As vision-language models (VLMs) transition from frozen-encoder pipelines to native multimodal architectures, the visual channel increasingly serves as both a perceptual input and a carrier of rendered text that competes with user prompts. Current single-attack, pixel-centric evaluations fail to capture this complexity: they collapse distinct visual attack surfaces into a single scalar and ignore how architectural shifts alter vulnerability. We introduce PaRMA, an evidence-based auditing framework that defines pathway-conditioned risk over two primary image-level interventions: pixel perturbation and rendered-text insertion. Instead of arbitrary attack enumeration, PaRMA uses pathway-specific diagnostics to expose architecture-specific blind spots. Notably, we uncover a see-but-not-deceived signature where native VLMs attend heavily to typographic overlays but visually reject them, and we identify an objective mismatch between encoder-level and answer-level attacks in deep-fusion models. To exploit these findings, we instantiate FusionAudit, introducing two diagnostic-motivated attacks: MANTIS (margin-aware pixel perturbation) and Nat-Evid (host-congruent semantic insertion). Evaluating 18 open-weight VLMs across a novel nativeness taxonomy, we demonstrate that legacy single-attack rankings (e.g., PGD) are unstable for native models. Our diagnostic-driven attacks substantially raise the estimated residual risk, showing that PaRMA provides a falsifiable, evidence-driven auditing methodology for uncovering vulnerabilities missed by standard benchmarks.
benchmark - arxiv:2610.21666 · cs.RODiffeomorphic Hidden-State Steering: Improving Imitation Learning Policy Robustness through Expert Hidden-State GeometryKwangBin Lee
Small observation or action errors can compound during closed-loop execution and disrupt a trained imitation learning policy. We study whether the original expert demonstrations can improve robustness through test-time hidden-state steering without additional demonstrations or policy retraining. Our analysis of a trained policy finds that expert hidden-state variation is concentrated in relatively few dimensions, while the directions capturing this variation differ across regions. Motivated by this geometry, we propose Diffeomorphic Hidden-State Steering (DHS), which learns an invertible nonlinear coordinate map from expert hidden states obtained by replaying the demonstrations through the frozen policy. DHS attenuates coordinates with low expert variance more strongly than those with high variance, then maps the modified coordinates back to hidden-state space. This allows steering directions to vary with the current hidden state while retaining every coordinate. Across four simulated manipulation tasks, DHS improves mean success under observation noise over the unmodified Transformer policies. Controlled comparisons yield higher mean success with learned nonlinear coordinates than with linear coordinates under both hard projection and full-rank rescaling. Combining DHS with action smoothing improves mean success rate under observation noise in certain tasks compared with action smoothing alone.
manipulation - arxiv:2610.21659 · cs.LGWhen Does "Post-Call" Begin? An Availability Contract for Earnings-Call NLP BenchmarksXiaoyu Ma, Kelvin J. L. Koa, Yuncong Liu, Ke-Wei Huang
Corporate financial disclosures often unfold over time, with announcements, prepared remarks, and discussion spanning multiple high-frequency return intervals. Financial NLP benchmarks can therefore suffer from a within-event timing mismatch when the prediction target begins before the textual or audio input used by a model could be complete. Chronological train--test splits do not prevent this mismatch, because it occurs inside individual events rather than across samples. We introduce an Availability Contract, which requires a benchmark to state the input boundary, evidence for input availability, the outcome start, and the claims supported by that evidence. We apply this framework to 552 earnings calls from 24 firms. For each call, we hold the return endpoint fixed and compare the previous close, a documentary call-start proxy, and a completion proxy based on reconstructed audio end plus a stipulated processing allowance. At this baseline, moving from the call-start proxy to the completion proxy flips return signs for 9.4% of calls. Holding completion-clock cutpoints fixed, 28.3% change quintile. In an independent release-session analysis of a public earnings-call volatility benchmark, shifting before-market target windows forward by one trading session raises three-day mean squared error by 18.1-21.4% across four frozen prediction streams, while their point rankings remain unchanged. Together, the results show that labels and losses can depend materially on the declared outcome clock even when model rankings appear stable. The Availability Contract makes this assumption explicit and ties post-input claims to the available timing evidence.
benchmark - arxiv:2610.21656 · cs.AIData World: Scaling Data Environments from ScratchZhanli Li, Zeyuan Xu, Chaoxu Pang, Zeyuan Liu +15
Data agents use code to process heterogeneous files and derive analytical insights. Scaling their training faces costly environment curation, privacy and intellectual-property restrictions, and difficulty verifying open-ended findings without known generative truth. Our core insight is to specify the world behind the files before generating the files themselves. We introduce Data World, a multi-agent engine that synthesizes environments, tasks, and rubrics from explicit mechanisms. Independent audits check target recoverability from observed files; rubrics support evaluation and RL rewards. Human preference-based curation for question realism selects 100 tasks across 50 held-out worlds and 16 domains for DataWorld-Bench. The engine also yields 2,135 training tasks from 1,037 disjoint worlds. Evaluation of ten frontier models shows that the benchmark remains challenging, particularly for insight discovery, with a consistent computation-insight gap. RL-zero training on the synthesized tasks produces DataWorld-35B-A3B from Qwen3.6-35B-A3B without trajectory distillation. Under matched evaluation, it improves over its base by 7.06 points on DataWorld-Bench, 14.28 on DAComp, 20.90 on DABstep-Hard, and 12.13 on JobBench, a broader occupational benchmark.
multi-agentbenchmark - arxiv:2610.21654 · cs.CVDigital Typhoon Dataset V3: Benchmarking for Multi-spectral Extension and Analyzing Systematic Bias in Intensity RegressionAsanobu Kitamoto, Victor Le Louarn, Noelia Falczuk, Ngoc Anh Tong +1
This paper presents the Digital Typhoon Dataset V3, a new version of the longest typhoon satellite image dataset spanning nearly 50 years, designed to benchmark machine learning models on long-term spatio-temporal data. The latest addition to Dataset V3 is multispectral satellite data, enabling multi-channel models. We investigate data-splitting strategies and present benchmarking results for the Northern and Southern Hemispheres. Experimental results show that, under the current dataset construction and experimental setting, multi-channel inputs do not clearly outperform single-channel inputs for pressure regression. Among the evaluated model architectures, image types, and spectral channels, ResNet-18, full-resolution images, and Channel 1 generally provide the best single-channel performance for pressure regression. Performance also differs between the WP and AU datasets by regression target, with lower RMSE on WP for pressure regression and on AU for wind regression. These results highlight the importance of considering dataset-specific and task-specific characteristics when interpreting benchmark performance. Additionally, we analyzed systematic bias in intensity regression, which leads to underestimation of intensity for strong typhoons. We tested imbalance-aware weighting strategies, including inverse-density weighting inspired by BalancedMSE, and showed that these approaches are not sufficient to remove the bias. The dataset is publicly available at http://agora.ex.nii.ac.jp/digital-typhoon/dataset/ and https://github.com/kitamoto-lab/digital-typhoon/.
benchmark - arxiv:2610.21646 · cs.LGThe Self-Improvement ProblemBrian B. Moser, Stanislav Frolov, Tobias C. Nauen, Federico Raue +1
With the rise of large language models, agentic systems are becoming increasingly autonomous, as they write code, call tools, revise prompts, and propose modifications to their own components. This autonomy raises a fundamental question for safe self-improvement: can an automated gate like a reward model or an automated reviewer reliably decide whether a proposed update should be accepted? The difficulty is structural. The gate must be a terminating, computable procedure, while the property we ultimately care about, namely whether an update is a genuine semantic improvement, depends on potentially unbounded program behavior. We formalize this mismatch by modeling agents as programs and defining improvement in terms of a semantic score that need not be computable. We study two evaluation abstractions. The first is a certifier that approves or rejects an update. The second is a search procedure that, given access to a proposal generator, attempts to output some improving candidate if one exists. We show that, in the fully general setting, no terminating procedure can simultaneously guarantee reliability (never accepting a non-improvement) and non-conservatism (never missing an available improvement): self-referential constructions can force any such gate or searcher to fail. Randomized gates with bounded error face the same limit. We then identify a standard decidable regime used in practice (bounded evaluation and a finite proposal budget) under which exhaustive simulation yields a terminating search procedure with the desired guarantees. We also show the limit of this recovery: a chain of accepted updates can raise the bounded score in every round while the true score falls. Our results draw a boundary between unbounded, universal settings where correctness guarantees collapse and bounded settings where they can be recovered by design.
agenticself-improvement - arxiv:2610.21644 · cs.LGEfficient Pairwise Difference Learning: A Comparative Study of Encoding Strategies and Training Pair SelectionBernard Georges, Mohamed Karim Belaid, Manuel Rodrigues
Pairwise Difference Learning for Classification (PDC) enhances supervised learning in low-data regimes by training a binary similarity classifier on pairs of instances rather than individual samples, effectively inflating the training set from $n$ to $O(n^2)$ examples. However, the PDC framework leaves critical implementation choices underspecified, particularly regarding input representation and pair selection strategies. This work systematically investigates these design decisions through extensive empirical evaluation. We compare four feature encoders (paired features, triplet with relative difference, absolute difference, and a novel triplet with absolute difference) alongside four pair selection strategies ranging from exhaustive to kernel-thinning-based schemes. Experiments across nearly 200 benchmark classification datasets using Decision Tree, Multilayer Perceptron, Gaussian Naive Bayes, and Support Vector Classifier as base models reveal that the proposed triplet with absolute difference encoder ($φ_{\mathrm{3dabs}}$) achieves the highest average F1 scores. More notably, restricting training to upper-triangle pairs reduces computational time by approximately 40% with negligible accuracy loss. Kernel thinning-based selection provided no accuracy benefits while substantially increasing computational overhead. Overall, PDC configurations outperformed base models in 94% of tested cases, though no single configuration dominated universally, suggesting dataset-specific optimization remains an open challenge.
benchmark - arxiv:2610.21642 · cs.CVSCMerge: Spatial-Consistency-Aware Token Pair Merging for Efficient Semantic SegmentationJunfeng Ren, Yi Gong, Yuan Zeng
Vision Transformer architectures have showcased impressive performance on semantic segmentation, yet their quadratic attention complexity hinders practical applications. Recent works have focused on merging tokens to reduce sequence length and thus improve transformer efficiency. However, existing token merging methods primarily rely on feature similarity to identify similar tokens, overlooking spatial consistency, which may introduce segmentation errors in semantic segmentation. In this work, we propose SCMerge, a training-free token merging method that jointly considers feature similarity and spatial consistency in pairing similar tokens. The pairwise tokens are then merged within the encoder, and the resulting token sequences are reconstructed onto the original patch grid after decoding. The proposed method is flexible and can be plugged into existing vision Transformer-based semantic segmentation pipelines. Extensive experiments on the benchmark datasets ADE20K, Pascal Context, and Cityscapes show that SCMerge consistently achieves a better trade-off between segmentation accuracy and efficiency than baseline methods.
benchmark - arxiv:2610.21640 · cs.LGReFlow: Replay-Driven Decoupling for Efficient RL Post-TrainingYukun Tian, Junwei Su, Haofeng Xu, Lansong Diao +5
Reinforcement learning (RL) has been a critical paradigm for improving large language model (LLM) reasoning and agentic capabilities, but its efficiency remains a major bottleneck. Existing RL systems commonly rely on asynchronous execution to overlap rollout generation with policy optimization for enhanced efficiency. However, this approach faces inherent limitations: mild policy staleness offers only limited training efficiency gains, whereas aggressive asynchrony can degrade model quality and yields diminishing efficiency gains. The dynamic nature of LLM RL and multi-turn agentic workloads further challenges RL system design, as resource demands shift over time, making static resource allocation and generation strategy inefficient. This paper presents ReFlow, a system-algorithm co-designed RL framework for efficient LLM post-training. Its key insight is that, unlike existing RL frameworks which consume each trajectory only once, experience replay of recent trajectories enables effective sample reuse, thereby amortizing the high cost of rollout generation without incurring excessive policy staleness. To fully exploit this opportunity, ReFlow adopts a decoupled and adaptive architecture with three key designs: a dual replay data cycle supports efficient experience reuse while balancing training throughput and optimization stability; an elastic resource pool enables dynamic resource placement to accommodate evolving RL workloads; an adaptive scheduler coordinates generation strategies and resource allocation to improve resource utilization and boost generation throughput. Experiments across representative single-turn reasoning and multi-turn agentic RL tasks show that ReFlow improves training throughput by up to $2.06\times$ over state-of-the-art synchronous and asynchronous RL systems, while maintaining comparable or better model performance.
agenticpost-training - arxiv:2610.21634 · cs.CVAnchor3DGS: Feed-Forward 3D Gaussian Splatting with Compact Anchor-Based RepresentationSheng Ye, Jiangke Lin, Fudong Wang, Yi Yuan +2
Recent feed-forward 3D Gaussian Splatting (3DGS) methods have enabled effective multi-view 3D reconstruction by predicting Gaussian primitives in a single forward pass. However, existing approaches predominantly adopt a pixel-aligned paradigm that predicts one Gaussian per input pixel, coupling the Gaussian count to the input resolution and number of views. This also leads to redundant primitives in textureless regions and insufficient coverage in geometrically complex areas. We present Anchor3DGS, a pose-free feed-forward 3DGS framework that replaces pixel-aligned prediction with a compact anchor-based representation. Guided by information entropy, our method places a set of 3D anchors in the scene, concentrating them in regions of higher visual complexity while maintaining broad spatial coverage. Each anchor performs occlusion-aware aggregation of multi-view image features, exchanges information with other anchors, and is decoded into a set of local Gaussians. A lightweight feature enhancement module further refines the coarse RGB output from Gaussian splatting and improves rendering fidelity. Extensive experiments on diverse benchmark datasets demonstrate that our Anchor3DGS achieves state-of-the-art novel view synthesis quality while only using an order of magnitude fewer Gaussian primitives.
benchmark - arxiv:2610.21624 · cs.AIAgentPerturb: Evaluating LLM Agent Robustness under Controlled Interaction PerturbationsYuanjian Xu, Jinyi Han, Jianing Hao, Yuxuan Sun +4
Real-world tasks are often complex and noisy, posing substantial challenges for LLM agents. Yet existing evaluations lack a systematic framework for jointly characterizing potential perturbation targets and types. We use this two-dimensional view to conduct a unified study that quantifies the robustness of current frontier models, traces the mechanisms behind perturbation-induced failures, and draws implications for robustness-oriented training. Through controlled evaluations on APEX-Agents and Toolathlon, we find that perturbations consistently degrade performance, while the dominant vulnerabilities vary across task environments. Trajectory-level analysis reveals that agents rarely recognize altered inputs and often propagate corrupted information without verification. A focused cost analysis further shows that successful perturbed trajectories can require substantially more computation. These findings motivate training objectives that cover diverse forms of noise and reward verification, uncertainty awareness, and recovery. We provide a rigorous benchmark-construction pipeline. Code and data will be released at https://github.com/xuyj233/AgentPerturb.
agentllm agentbenchmark - arxiv:2610.21621 · cs.ROSPEAR: Selective Preference Elicitation with Adaptive Rewards for VLM-Guided Robotic ManipulationZewen Wu, Ruizhe Liu, Pei Zhou, Feng Chen +1
Recent Vision-Language Models (VLMs) enable automated preference labeling for robotic manipulation, yet their deployment remains impractical since each query incurs substantial inference cost, while VLMs frequently abstain on visually challenging comparisons. We propose SPEAR, a framework that transforms VLMs into efficient preference oracles via two synergistic mechanisms. First, we design distinguishability-aware query selection to reduce abstention rates by targeting only informative and distinguishable comparisons. Second, we deploy an adaptive reward transition that bootstraps exploration with VLM-synthesized guidance before preference-based refinement. Experiments on the manipulation benchmarks demonstrate that SPEAR achieves 88.9% average success, outperforming baselines by 23.3 percentage points, while significantly reducing abstention and wall-clock training time on contact-rich tasks. SPEAR learns three Mobile ALOHA manipulation tasks fully online on real hardware, demonstrating end-to-end real-world applicability. By addressing query efficiency and exploration challenges, SPEAR establishes VLMs as practical preference oracles for scalable robot learning without manual reward engineering.
manipulationbenchmark - arxiv:2610.21619 · cs.ROTraining-Free Target-Aware Refinement of 3D Gaussian Language MapSibaek Lee, Kyeongsu Kang, Seongbo Ha, Yewon Jeon +8
Distilling vision-language features into 3D Gaussian Splatting yields a language map that can be queried with arbitrary text. In deployment, however, a small set of user-specified targets must be reliably located, which current distilled maps often miss, while open-vocabulary queryability for other words must remain intact. We propose a training-free, target-aware refinement of an already-distilled language map. Guided by a text-promptable segmenter, the visual embedding inside each detected target region is blended with the target text embedding, a convex combination whose coefficient sets the amount of injected prior, while non-target regions are left untouched. Injecting too much overwrites appearance cues and general queryability, so we seek, for each target, the smallest coefficient that still lets it win, and we obtain every coefficient in closed form, leaving no hyperparameters to tune. We further identify a failure mode of injection in language-goal localization. Text embeddings share a common component, and injecting them makes the standard cosine read-out query-agnostic, which a centering step in the read-out removes. Across three benchmarks, the refinement consistently improves open-vocabulary segmentation, roughly doubling the base mIoU on Replica, and localizes language goals more reliably than the base map. It also transfers as a plug-in to off-the-shelf field-based and object-level language maps and supports a real-world demonstration in which a legged robot navigates to natural-language goals. Project page and code: https://lab-of-ai-and-robotics.github.io/target-aware-langmap/
benchmark - arxiv:2610.21616 · cs.CVA Decision-Aware Evaluation Protocol for Uncertainty in Medical Image SegmentationSaket Maganti, Shreyank N Gowda
Medical segmentation uncertainty is often evaluated with calibration error or pixel-level error-detection scores, but these metrics do not show whether uncertainty improves downstream review decisions. We propose a decision-aware evaluation protocol that separates uncertainty estimation from deferral policy and evaluates their combination under fixed review budgets: which pixels, regions, or images should be accepted, and which should be reviewed. The protocol reports fixed-budget error reduction, risk-coverage behaviour, region-level review proxies, structural failure modes, probability-derived baselines, and sensitivity to review assumptions. We instantiate it across retinal vessel, skin lesion, and polyp segmentation with multiple backbones and uncertainty estimators. Under matched inference budgets and canonical multi-layer MC Dropout, test-time augmentation (TTA) gives a stronger error-detection signal across matched MC--TTA comparisons, including architecture and dropout-placement checks. However, estimator rankings are not universal. For example, probability-derived scores remain competitive, deep ensembles can outperform TTA, learned deferral helps mainly when the base estimator is weak, and rankings can change with review unit, uncertainty target, correction assumptions, and cost model. Structural and second-observer analyses further show that boundary and thin-structure failures are partly tied to annotation ambiguity. These findings support evaluating uncertainty as an estimator-policy-review configuration, not as a standalone score map. Code: https://anonymous.4open.science/r/Uncertainity-in-Medical-Image-Segmentation/README.md
evaluation protocol - arxiv:2610.21614 · cs.AIQwen-Audio-Jev: A Unified Interface for Audio DecisionsYing Fang, Zhifu Gao, Qing Gu, Jiajun He +4
Large audio language models (LALMs) have made substantial progress in audio understanding. However, in high-throughput applications requiring frequent selection, judgment, and scoring, autoregressive text generation introduces additional computational overhead and response latency. To address this limitation, we introduce Qwen-Audio-Jev by augmenting Qwen-Audio-3.1-ASR-Flash-Next with a dedicated Jev decision head. The model unifies selection, binary judgment, and scoring through three decision primitives: CHOICE, NOUL, and SCORE. Given audio, a natural-language question, and an explicit answer specification, it directly produces a task-conditioned probability distribution or a numerical score through typed decision readouts. Evaluation across 18 labeled configurations covers all three primitives: for CHOICE, Qwen-Audio-Jev achieves accuracies of 78.70% on MMAU-mini and 84.06% on MMSU; for NOUL, it achieves an equal error rate of 6.50% on the VoxCeleb1-O speaker verification benchmark; and for SCORE, the model achieves a normalized mean absolute error of 0.1062 on NISQA overall quality assessment, with Pearson and Spearman correlations of 0.8210 and 0.8074, respectively. Compared with the original Qwen-Audio-3.1-ASR-Flash-Next, Qwen-Audio-Jev achieves an approximately 20x inference speedup. It demonstrates its potential for efficient, high-throughput audio decision-making in applications such as large-scale audio screening, speaker verification, and automated quality assessment.
benchmark - arxiv:2610.21612 · cs.AILearning Intrinsic Dynamics in Latent World ModelsJiamin Shi, Yuanhong Zhang, Zhiyang Liu, Jiawei Sun +1
Learning intrinsic dynamics---the relationship between actions and their consequences---is essential for world models. However, visual variations unrelated to the underlying control process can obscure this relationship in learned representations. We investigate the connection between action predictability, defined as the ability to reconstruct an action from consecutive latent states, and the preservation of intrinsic dynamics. Our central hypothesis is that encouraging action predictability discourages representations from discarding the distinctions between different actions and their consequences. We provide two complementary forms of evidence for this connection. Theoretically, we derive a positive lower bound on squared inverse prediction error when the encoder merges distinguishable outcomes of different actions from the same state and appearance, and show that forward prediction can still be exact on those transitions. Empirically, controlled appearance-diversity comparisons, supervision interventions, and latent probes support the role of action prediction in preserving control-relevant transition information. We instantiate action predictability with Concat IDM, which predicts actions from concatenated latent states and updates the shared encoder using action labels already available in offline trajectories, without changing the test-time planner. Across three independent training runs on Push-T with 24 appearances, Concat IDM increases planning success from 9.1% to 77.2%; its paired improvement is 52.9 percentage points larger than that observed with a single appearance. Together, these findings support action predictability as an inductive bias toward intrinsic dynamics and demonstrate the practical benefits of Concat IDM for latent planning under visual diversity.
world model - arxiv:2610.21607 · cs.LGMarket Incentives for AI Safety InvestmentNikita Tsoy
Despite their practical significance, modern AI systems pose significant societal risks, including toxicity, malicious use, and misinformation. Their mitigation depends not only on technical feasibility but also on the incentives of firms that develop and deploy these systems. We show how profit maximization shapes safety investment in an AI supply chain with an upstream LLM provider and several competing downstream firms. In our game-theoretic model, we derive the market-driven level of safety investment and compare it with the levels that maximize market welfare and industry profit. We find that downstream competition generally induces downstream firms to invest more in safety than the industry-profit-optimal level. By contrast, the upstream provider may underinvest relative to these benchmarks because it is not fully compensated for the surplus from safety investments. Finally, we analyze how different regulatory interventions affect the participants' incentives and profits, and the resulting levels of model safety.
benchmark - arxiv:2610.21604 · cs.AIEvaluating Inquiry Learning in LLM-Based Agents: Hypothesizing, Reasoning, and ExplorationMarta Knežević, Seyed Parsa Neshaei, Jibril Frej, Vinitra Swamy +1
Scientific inquiry involves iterative cycles of exploration, hypothesis generation, and reasoning over observations to uncover relationships in complex systems. Large language models (LLMs) have demonstrated strong capabilities in several components of this process, including inductive reasoning, hypothesis generation, and data analysis. However, prior work typically evaluates these abilities in isolation, leaving open how models integrate them across a full inquiry cycle. In this work, we introduce a framework for evaluating LLMs' inquiry capabilities in a unified, interactive setting. We extend the Countdown task to a mathematical environment where an agent probes the system by manipulating variables, collecting observations, and inferring relationships between them. This setup enables analysis of exploration strategies, causal discovery, and hypothesis refinement during inquiry. Across seven LLMs and multiple prompting strategies, our experiments reveal limited inherent inquiry capabilities and low success rates in discovering the underlying expressions. However, structured guidance inspired by educational scaffolding significantly improves performance. Our findings highlight current limitations of LLMs in strategic exploration and iterative hypothesis refinement, and show how simple guidance strategies developed for human learners in inquiry learning can support more effective inquiry behavior.
agent - arxiv:2610.21591 · cs.AILOADBench: A Large Open Analog Dataset and Benchmark for Machine Learning in IC DesignM. Leibl, F. Azevedo, S. Uhlich, A. Bonetti +6
Machine learning (ML) for analog IC design is constrained by limited open data and the absence of standardized evaluation: results reported in different papers are typically obtained on bespoke, often closed, datasets and with ad-hoc metrics, leaving comparisons across methods unreliable. We address both gaps with a large opensource dataset and a suite of standardized benchmarks. The dataset contains more than 100 operational amplifier (opamp) topologies (1-, 2-, and 3-stage) in the SkyWater130 PDK, each shipped with structural and functional annotations, topology-specific sizing constraints (symmetry and sizing rules), an Ngspice testbench, and a multi-objective NSGA-II sizing run producing around 103 Pareto-optimal and 106 intermediate sizing-performance pairs per topology. On top of this data we define six benchmark tasks with precise protocols: inverse sizing, topology selection, sub-circuit classification, structural circuit generation, Pareto-membership prediction, and multi-objective optimization. Thus, the data presented here constitutes the most comprehensive dataset for ML development in over four decades of analog IC design automation. The dataset, ground-truth annotations, and reference implementations of the metric-defining routines are released under an open license, enabling reproducible and directly comparable evaluation across heterogeneous analog design automation methods.
benchmark - arxiv:2610.21588 · cs.CVHear, Localize, and Reason: Spatially Aware Scene Understanding for Audio-Visual LLMsKim Sung-Bin, Lee Jung-Mok, Jinwoo Jung, Oh Hyun-Bin +3
Audio-visual large language models (AV-LLMs) have made strong progress in multimodal understanding, yet most treat audio as monaural semantic content, limiting their ability to reason about where sounds originate and how they relate to the visual scene. Recent spatial audio-visual studies address parts of this problem, but often focus on specific abilities, such as spatial correspondence or direction and distance reasoning. In this work, we introduce HLR-AVSceneQA, a comprehensive benchmark with a hear-localize-reason hierarchy that evaluates whether models can recognize what is heard, localize sound sources in egocentric and allocentric views, and use these localized sources to reason about how they relate to the visual scene. We further introduce HLR-LLM, which augments a strong audio-visual foundation model with a binaural spatial audio branch alongside its monaural semantic audio pathway. We train HLR-LLM with a curriculum that incorporates final-stage chain-of-thought supervision for multi-step spatial reasoning. Experiments show that existing AV-LLMs remain limited in spatial grounding and relational reasoning, while HLR-LLM substantially improves spatial audio-visual scene understanding without compromising semantic audio-visual perception.
benchmark - arxiv:2610.21586 · cs.ROAVALON: Zero-Shot Indoor Navigation with Human-Seeded Topological MemoryJing Xia, Changyan Li, Yuduo Wang, Guiyin Tian +4
Indoor service robots often enter unmapped buildings whose layouts and task constraints are already familiar to users. We address this human-familiar, robot-unseen setting with AVALON, a topological navigation framework organized around human-seeded layered spatial memory, node-verified topological execution, and task-triggered, evidence-gated human--robot memory co-updating. The robot has no prior on-site traversal or site-specific training. Deployment requires only a floor plan, place clues, and an initial topology-odometry alignment. Visual depth point-goal navigation leverages verified node arrivals to perform translational re-anchoring, without requiring a robot-constructed global SLAM map. Human--robot memory co-updating fuses complementary sources of information. Human operators specify task constraints that are unobservable via vision and report offline environmental modifications, whereas the robot provides physical-state evidence through active in-situ validation. Task relevance governs whether human-reported environmental changes undergo robot validation during the active mission. Field-collected perceptual evidence gates memory updates, maintaining an explicit separation between human-specified permissions and robot-observed traversability. Our experiments on a Unitree Go2 show that AVALON with X-NavDP completes a 12.81 m Intersection-to-Dining mission in 46.4 s (5/5) with no mapping, 40.9% less robot time than DimOS-Go2 (78.5 s, 4/5), while policy-only X-NavDP times out in all of the 5 runs. These demonstrate the superiority of the proposed topological navigation framework.
memory - arxiv:2610.21581 · cs.LGClaimScope: Label-Blind Witness Verification for Mathematical Claim AuditingMushir Akhtar
Mathematical agents are commonly evaluated by answer accuracy or proof acceptance. We introduce ClaimScope, a 72-task benchmark for auditing proposed claims through a verdict, scope, counterexample, and repair. Tasks cover optimizer-state transport, finite-horizon affine recurrences, and exact $2\times2$ linear algebra. We evaluate Qwen3-8B with direct generation and four matched two-pass conditions, each starting from the same draft. A witness checker evaluates only the model's submitted state, without consulting reference verdicts or supplying a replacement witness. Its feedback still conveys information about validity: a verified refutation establishes that the claim is false, and rejection reasons can suggest structural fixes. In this evaluation, verifier feedback yields $23/39$ accepted counterexamples on invalid claims and $25/72$ safe completions, compared with $15/39$ and $20/72$ for checker-free generic feedback. Strict completion remains $17/72$; verdict accuracy changes from $47/72$ to $46/72$. The claim-grouped safe-completion difference is $.094$, with a bootstrap 95\% interval of $[.000,.198]$. All seven task-level gains follow horizon/input-length diagnostics. A post-hoc schema-only correction of the direct outputs reaches the same 25 safe tasks without another model call. These results separate claim classification from executable evidence, while locating the observed gains in witness-structure correction rather than demonstrating broader reasoning improvement.
benchmark - arxiv:2610.21573 · cs.CVMYOPIA: Understanding the Limits of Tree-Based Reasoning in Vision-Language ModelsIslam Mesabah, Sary Elmansoury, Gerrit Großmann, Peter Neigel +1
Decision trees can make vision-language model (VLM) classification interpretable by replacing a single prediction with a sequence of root-to-leaf questions. However, the accuracy cost of this greedy, non-backtracking procedure remains unclear. We introduce \textbf{MYOPIA}, a benchmark comprising four LLM-generated, human-refined decision trees for GTSRB, MNIST, CIFAR-10, and MSCOCO, together with matched prompts, diagnostic protocols, and an evaluation harness. Across GPT-4o, Qwen-VL Max, and LLaMA-3.2 11B Vision, tree traversal underperforms zero-shot prompting in all 12 dataset--model pairs; paired bootstrap analyses consistently favor zero-shot. Our diagnostics characterize interacting failure modes associated with this gap. Models often recall tree-encoded class attributes more accurately than they ground the same attributes in images: the mean label--image accuracy gaps are 9.5, 13.8, and 12.6 points for GPT-4o, Qwen-VL Max, and LLaMA, respectively. Incorrect local decisions then irreversibly exclude the correct class, while answer extraction introduces additional model-dependent loss. A node-level audit of the four trees finds that split entropy is weakly but consistently associated with lower node accuracy, whereas depth, branching factor, and within-node grounding differences are less stable. \textbf{MYOPIA} therefore provides a framework for diagnosing hierarchical visual classification and motivates testing tree designs that prioritize perceptual answerability and recovery from early errors. Code is available at https://github.com/islammesabah/VLM_Classification.
benchmark - arxiv:2610.21571 · cs.AIRethinking the Role of Tool Graph Memory in Multi-Turn Tool UseUiji Hwang, Richeng Xuan, Taeuk Kim
Large language model (LLM) agents solve complex tasks by calling external tools over multiple turns. Recent work has proposed tool graph memory, which stores past experience as tool transitions to guide subsequent selection. However, it remains unclear whether this component consistently improves agents' tool use or undermines it. We evaluate tool graph memory on nine models and two benchmarks, finding that it seldom improves significantly over baselines, and that its gains, where present, cost over twice the LLM calls. The memory's guidance is often incorrect, and random or no guidance retains most of the gain, implying that the benefit stems from extra computation rather than the memory itself. Moreover, even with the correct next tool suggested, many failures remain, such as giving up without a tool call. Storing the same experience instead as textual insights outperforms tool graph memory with almost no extra LLM calls. This suggests that memory serves better as a soft prior than as a prescription of the next tool, leading to a more effective and generalizable solution.
memorytool usebenchmark - arxiv:2610.21570 · cs.AIIndependent Data-Integrity Verification for LLM-Driven Functional Coverage Closure in Hardware DesignVenkata Krishna Kumar Vedantam
LLM agents can write directed tests that close functional coverage, but coverage only shows that a state was reached, not that the data on the way there was correct. This is a long-known point in verification and we do not claim it as new. We describe a small coverage-closure agent that applies it to an LLM loop: coverage is read from the report that the simulation itself exports (cocotb-coverage YAML) and never taken from the model's account, and data integrity is checked by scoreboards that are separate from the model. On three architecturally distinct RTL designs (an asynchronous FIFO, a UART transmitter and an AHB2APB bridge) the agent raised coverage from 91.7%, 88.9% and 67.4% to 100% within its 8-iteration budget in every logged run, using one design-agnostic agent. A reactive scoreboard monitor attached to the FIFO's base testbench also found a clock-domain-crossing bug in the author's own stimulus code. The monitor is not yet attached to the LLM-generated tests, so their data integrity is not independently checked; one manual check on a single FIFO test is reported as preliminary. We list the weaknesses of this evidence: three designs, no baseline, no ablation, and logged runs that come from several agent checkpoints.
agentllm agent - arxiv:2610.21569 · cs.CVDecoupled Safety Control: A Safety-Control Algorithm for Training-Free Safety GuidanceHuilin Zhou, Ruoxi Cheng, Yuhang Wang, Yuming Liu +3
Diffusion models achieve strong performance in text-to-image generation, but they can still produce unsafe content. Training-free safety guidance offers a practical inference-time alternative that intervenes during sampling without modifying model weights. However, many existing safeguards still rely on a coupled safety-control chain. In this chain, the same unsafe condition simultaneously specifies what should be avoided, indicates the current unsafe signal, constructs the intervention, and determines when that intervention remains active. This coupling can work when unsafe factors are explicit, but becomes less reliable once harmful and benign semantics are entangled in the current generation state, often leaving residual unsafe content or unnecessarily degrading benign-generation quality or prompt fidelity. We address this problem with Decoupled Safety Control (DSC), a plug-in safety-control algorithm for training-free safety guidance. DSC reformulates training-free safety guidance as an explicit state-dependent control process, so the executed intervention is determined from the current generation state rather than inherited directly from a single unsafe condition. Experiments on standard safety benchmarks show that DSC improves representative latent- and text-space safeguards, achieving stronger unsafe-content suppression while preserving benign-generation quality and maintaining efficient inference.
benchmark - arxiv:2610.21568 · cs.LGFrom Perception to Prediction and Intervention: Benchmarking Proactive Video Assistants in Real-Life ScenariosBoyu Cui, Junhao Luo, Zhihao Wang, Linghan Chen +4
Recent multimodal large language models have advanced video understanding, but existing benchmarks mainly evaluate whether models can interpret observed visual content. In real-world scenarios, assistants must additionally anticipate what assistance is needed and decide when to provide it. To address this gap, we introduce \textsc{P$^3$IBench}, a benchmark for proactive video understanding in real-life streaming scenarios. It evaluates three complementary capabilities: daily perception, risk prediction and intervention. We benchmark 12 representative open-source and proprietary video-language models, including streaming and adapted non-streaming systems. Our evaluation shows that current models struggle to make well-calibrated response decisions. Finally, we construct additional proactive training data and train \textsc{P$^3$I-Assistant-4B}, achieving improvements on several tasks, though gains remain uneven across capabilities. Together, these findings show that effective proactive assistance requires not only visual understanding, but also precise response timing and appropriate restraint.
benchmark - arxiv:2610.21562 · cs.LGA Quantum Machine Learning Approach to Content-Based Media RecommendationMina Abbaszadeh, Ouissal Moumou, Anna Pearson, Bob Coecke +3
We use quantum machine learning techniques to develop a content-based media recommender system and evaluate its performance on TV programmes using textual descriptions from one week of BBC iPlayer data. Grounded in the Quantum Natural Language Processing framework (QNLP), programme synopses are mapped to variational quantum circuits using their semantic structure. The parameters of these circuits are trained via a supervised pairwise similarity learning task. We benchmark the quantum recommender against a classical BiLSTM baseline trained on the same dataset and objective. While both models achieve comparable performance on the similarity learning task, the quantum model requires significantly fewer parameters (47k vs. 17.8M). When the BiLSTM is constrained to a matching parameter budget, its performance degrades substantially (by 3.63 percentage points in Mean Average Precision (MAP)), indicating that the structured quantum representation provides a highly parameter-efficient alternative in this setting. We demonstrate the feasibility of executing the trained models on quantum hardware by deploying the circuits on Quantinuum's System Model H2 (and accompanying emulator) using a SWAP-test-based protocol. The results confirm that the similarity signal is preserved under realistic, noisy conditions. Taken together, these findings highlight the potential of variational quantum circuits for recommendation tasks and represent a practical step toward their real-world deployment.
benchmark - arxiv:2610.21561 · cs.LGDialFork: Going Beyond Factuality to Evaluate Procedural Hallucination in Conversational AgentsSergio Burdisso, Esaú Villatoro-Tello, Dairazalia Sánchez-Cortés, Yanis Labrak +3
Large language models (LLMs) deployed as conversational agents are increasingly expected to follow procedural logic. While factual hallucination has received extensive attention, a different failure mode has been overlooked: procedural hallucination, in which an agent produces individually fluent and factually plausible utterances while deviating from the intended conversational process. We introduce DialFork, the first benchmark to measure this failure mode on raw dialogue text. DialFork is built by forking conformant dialogues at controlled turn positions and injecting one of four hallucination types, yielding 36,278 dialogues across 24 tasks in 12 domains; 30 human annotators validate the construction. We define three evaluation tasks that characterize procedural deviation at three granularities: how far a set of dialogues deviates from conformant ones (monitoring deployed agents), how far a single dialogue does (procedural retrieval), and at which turn a dialogue first goes off-track (real-time guardrails). For the first two, we propose dialog-flow-aware metrics, FlowKL and FlowScore, that compare dialogues by how they unfold rather than their content. At the dialogue level, FlowScore significantly outperforms all baselines across seven encoders, showing that semantic similarity is no substitute for procedural similarity. We release DialFork under the MIT license at https://github.com/idiap/dialfork, hoping the community will measure conversational process alongside content.
agentbenchmark - arxiv:2610.21555 · cs.LGWhen Does Muon Actually Win? A Paired Data-, Time-, and Tuning-Budget BenchmarkMushir Akhtar, Hazem N. Nounou, Majdi Mansouri, Mohamed N. Nounou
Muon's gains in data efficiency motivate evaluating its performance under explicit data, time, and tuning budgets. On a single A100 GPU, we compare a reference single-device Muon implementation with AdamW across 17.7M- and 33.6M-parameter causal language models and 2.70M-parameter vision transformers. A common 23-candidate development budget selects configurations evaluated with five matched seeds, full validation trajectories, optimizer-step timing, and a frozen one-shot test. Muon reduces validation-loss area under the curve by 3.25--22.44\% at equal data and lowers test loss in all 20 primary pairs. In this configuration, Muon improves time-normalized loss area} by 19.68--22.31\% on vision, while AdamW is better by 3.38--6.96\% on language. The full Muon optimizer call, including auxiliary AdamW updates, consumes 56.1--69.6\% of training time, compared with 6.1--9.1\% for the AdamW baseline. Muon also has higher expected regret under restricted access to these candidate pools. These results characterize small models with reference kernels; their wall-clock ordering does not establish a ranking for larger models or optimized implementations.
benchmark - arxiv:2610.21553 · cs.CLWritability and the Pathway: A Controlled Study of Knowledge Injection into Addressed MemoryWei Gao, Cheng Tian, Duan Yang, Jianqing Yin
Addressed conditional memory provides a local surface for post-hoc knowledge injection, but a successful write need not transfer to a new question or preserve untargeted behavior. This study combines a controlled GPT-2 comparison with production-memory experiments on Qwen3.8-Flash-Next-FP8 and DeepSeek-V4.1-Flash. Newly initialized modules attached to frozen GPT-2 are tested in a 2 x 2 design: row-only versus row-plus-projection updates, each with or without fixed module-residual anchoring. All arms reproduce 50 trained completions in each of three sampling seeds. Natural-question target-string hits average 5.3% for both row-only arms, 28.7% for unanchored projection updates, and 22.0% with anchoring, while strict answer scores remain zero. Anchoring lowers measured drift; lexical hits do not establish entity-specific retrieval. In the archived Qwen experiments, gradient writes make the designated first answer token top-ranked for 132 independently reset targets, comprising 93 new promotions and 39 already top-ranked targets. A separate matched 24-target comparison gives 13 new promotions with Adam versus two with SPSA at equal iteration counts but unequal computational procedures. An 80-row case includes saved continuations and a replay record, while reformulations bound its scope. Six DeepSeek zero-order cases yield one new target-token promotion and one regression. These results support bounded writing feasibility and separate trained expression, question-form behavior, and preservation; first-token rankings do not establish complete-answer success or universal conditions for writable memory.
memory - arxiv:2610.21548 · cs.LGFinite-Time Guarantees for Constrained Reinforcement Learning under Partial ObservabilityValentina Zaccaria, Philip Jordan, Tingting Ni, Gian Antonio Susto +1
We study reinforcement learning in constrained partially observable Markov decision processes (CPOMDPs), where an agent maximizes reward subject to a cost constraint using only incomplete information about the underlying state. The lack of state observability may require optimal policies to depend on arbitrarily long observation-action histories, making planning and learning challenging. We leverage recent advances in finite-memory approximations of POMDPs and represent finite history windows as states of a constrained MDP. We then develop a primal--dual algorithm that provides the first non-asymptotic guarantees that jointly bound average reward suboptimality and constraint violation in the original CPOMDP. A key challenge is that this finite-window approximation affects both reward values and the set of feasible policies. Under a filter-stability assumption, we derive bounds relating optimality and feasibility in the finite-state model to those in the original CPOMDP, with approximation errors that decay geometrically with the window length. Experiments in partially observable gridworlds demonstrate the effectiveness of our approach.
agent - arxiv:2610.21539 · cs.AIReinforcement Learning-driven Diffusion Models: Language, Vision, and BeyondZhiyuan Ma, Jiabao Wei, Tianyu Wang, Hanqing Wang +20
Diffusion models (DMs) have steadily demonstrated outstanding performance in high-fidelity, controllable generation across diverse scenarios, following the "pretraining-finetuning" paradigm established in LLMs. Despite these advances, conventional DMs still face critical limitations in sophisticated reasoning, autonomous exploration of generation pathways, and adaptability to complex, human-aligned objectives. Inspired by RLHF's success in LLMs, RL has recently emerged as a pivotal post-training mechanism to iteratively refine DMs through reward-driven trajectory optimization. This paper presents the first comprehensive review of RL-driven DMs and an empirical analysis of how RL reshapes the diffusion process towards experience-centric, goal-directed generation. We propose a novel technical panorama elucidating RL's multifaceted roles in diffusion-based language, multimodal, robotic, and scientific models, with a granular focus on diffusion trajectory optimization and reward engineering. Through extensive experiments on multiple task-specific benchmarks, we systematically evaluate representative methods to intuitively highlight the advantages of different RL post-training strategies. Furthermore, we reveal and dissect three crucial bottlenecks that currently hinder the real-world deployment of RL-driven DMs: reward design fragility, exploration-exploitation trade-off in high-dimensional generation spaces, and training instability under sparse rewards. Based on these perspectives, we provide a principled roadmap for the integration of RL into DMs and outline promising future research directions for robust, interpretable RL-diffusion systems. We also construct a public repository to continuously track the latest developments in this rapidly evolving field: https://github.com/MAIR-Lab-HUST/Awesome-RL-for-Diffusion-Models.
post-trainingbenchmark - arxiv:2610.21538 · cs.AIAgent Status Protocol (ASP): A Lightweight Specification for State Reporting in Multi-Agent SystemsDiego Gosmar, Deborah A. Dahl, Emmett Coin, Csaba Bolyós +1
The Agent Status Protocol (ASP) defines a lightweight way for heterogeneous AI agents to report their operational state to a status server, and for observers such as dashboards to receive that state as soon as it changes. Agents push small JSON events over HTTP POST; the server exposes the aggregated state to observers over Server-Sent Events (SSE). Because an agent that has crashed, hung or been compromised cannot be trusted to report its own failure, ASP does not rely on self-reports alone: the server infers liveness from missed heartbeats, can detect agents that are alive but no longer making progress, and accepts observations from supervisors, guardrails, operators and peers, some of which take precedence over what the agent says about itself while others are shown alongside it. ASP models state along independent dimensions (lifecycle, health, task, floor, governance), so that an agent can be, for example, online, degraded, working on a task and waiting for the floor at the same time. It reuses existing vocabularies: task states from the Agent2Agent (A2A) protocol, floor states from the Open Floor Protocol (OFP), and attribute conventions from the OpenTelemetry GenAI semantic conventions. This paper presents version 0.1 of ASP as a proposal, not as an approved standard. It is a working draft intended to open a discussion, and the authors invite comments, alternative proposals and independent implementations from the community.
agentai agentmulti-agentagent system - arxiv:2610.21537 · cs.AIFunctional Subnetwork Localization for Efficient Continual Learning in Large Language ModelsShuoran Jiang, Wentao Zhu, Yang Xiang, Dongfeng Li +7
Biological intelligence continually adapts to new information while retaining prior skills through localized plasticity. In contrast, continual learning in large language models (LLMs) remains vulnerable to catastrophic forgetting. A promising direction is to identify functionally specialized subnetworks within LLMs that enable targeted and resource-efficient adaptation. Using gradient-attribution profiling, we localize cognition-associated functional subnetworks exhibiting extremely low mutual overlap. Building on this finding, we propose Function-isolated Memory-Efficient Optimization (CiMEO), a continual learning framework that routes each training example to its corresponding functional subnetwork, restricting parameter updates and optimizer states to selected regions. Across 1.4 million sequential instructions, CiMEO achieves the highest average performance among the compared methods on five generative benchmarks spanning programming, examination-style tasks, and mathematical reasoning, while mitigating catastrophic forgetting. Subspace-overlap and principal-angle analyses further indicate that localized updates help preserve pre-trained representations, highlighting functional subnetwork localization as a practical approach to efficient and stable continual learning in LLMs.
benchmark - arxiv:2610.21522 · cs.ROMulti-Teacher On-Policy Distillation for Flow-Based Vision-Language-Action ModelsLichao Ni, Yuhan Liu, Pengfei Li, Guanghuan Fang
Specialized vision-language-action (VLA) policies acquire strong domain-specific skills, yet consolidating these skills into one shared policy remains challenging. On-policy distillation (OPD) transfers teacher knowledge on inputs induced by the student, addressing the distribution mismatch encountered during deployment. For flow-based VLAs, this mismatch arises at two levels: environment observations and intermediate action-generation states. We propose a two-level on-policy distillation framework that couples student-driven environment interaction with teacher velocity supervision along student-generated action flow trajectories. Frozen domain specialists jointly supervise one student. After each update, the student collects new environment experience and generates fresh action-flow trajectories. Starting from a joint supervised fine-tuning (SFT) initialization, Velocity Loss OPD yields absolute success-rate gains of 22.55% on LIBERO-40, 33.4% on CALVIN five-instruction sequences, and 22.6% on Meta-World MT50. Compared with Action Loss OPD and Teacher-rollout SFT, it achieves absolute success-rate gains of 6.4-9.5% on LIBERO-40, 9.7-18.3% on CALVIN, and 3.2-4.2% on Meta-World MT50. Analysis by sequence length further shows larger success-rate gains on longer CALVIN sequences. These results demonstrate effective multi-domain capability consolidation through two-level on-policy supervision, with a single student deployed across domains.
vision-language-actionlibero - arxiv:2610.21520 · cs.AIHierarchical Interpretable Multitask Audio Representation Learning: An Analysis of Event-Scene Relationships through the Information BottleneckClaudia Montero-Ramírez, Alba Martínez-Serrano, Carmen Peláez-Moreno
Joint Acoustic Scene Classification (ASC) and Audio Tagging (AT) models traditionally rely on unconstrained multihead architectures. However, this design ignores the natural hierarchical dependency between sounds and acoustic scenes. In this work, we investigate a hierarchical (sequential) conditioning paradigm that predicts scene categories directly from intermediate event probabilities. Through a large-scale empirical benchmark with 66 full training runs across three state-of-the-art backbones (PaSST, PANNs, and BEATs) and different loss-weighting settings (beta in [0.0, 1.0]), we show that hierarchical conditioning achieves competitive classification performance while altering multitask optimization dynamics. Unlike multihead baselines, which sequentially solve the easier ASC task before optimizing AT, the hierarchical architecture enables synchronized dual-task learning for specific loss-weighting settings (beta in [0.6, 0.9]). Furthermore, we propose a multi-perspective interpretability framework that combines an analysis of the information bottleneck, manifold geometry, and Formal Concept Analysis (FCA) on a Term-Frequency Inverse-Document-Frequency (TF-IDF) event representation, providing complementary insights into event-scene relationships. Our findings reveal that hierarchical conditioning systematically yields lower conditional entropy, higher mutual information, and superior latent cluster separability (higher Silhouette scores), establishing a more structured and semantically faithful intermediate feature space.
benchmark - arxiv:2610.21515 · cs.ROTRACE: Typed Reference-Aware Conversational Editing for Humanoid Motion DiffusionZejian Kang, Yifu Luo, Boyang Shen, Jingxuan Li +8
Language-conditioned humanoid motion generation enables users to create and refine robot movements through natural language. The central challenge is to connect an underspecified request to concrete motion structure that can subsequently be revised relative to an accepted reference. We present TRACE, a generation and editing framework for the Unitree G1 humanoid robot built around typed, measurable kinematic quantities. Our insight is to ground generation conditions in paired training motions and editing goals in the current motion. A language model learns to predict support, root, and hand events from motion-derived supervision; their time-aligned encodings complement the original text in a diffusion generator. Gravity-referenced states and trajectory--contact objectives improve kinematic consistency. For editing, typed requests define body--time scopes and source-relative goals. A separate reference-conditioned diffusion editor uses these goals and revised events to realize changes, while sampling-time projection preserves protected state coordinates. Quantitative evaluation and human assessment show improved semantic alignment and motion quality over the compared methods. Controlled ablations support the value of motion-grounded conditioning and kinematic consistency, while multi-turn editing evaluations demonstrate more reliable execution of successive relative requests under shared preservation constraints. Together, these results support typed kinematic structure as a common basis for motion generation and reference-aware revision.
humanoid - arxiv:2610.21512 · cs.AIQREF: Improving DeepSearch Agents by Aligning Search with Reasoning via Query Refinement and Evidence FilteringYujia Chen, Yuhui Wang, Junzhe Wang, Shihan Dou +14
DeepSearch agents achieve strong performance on complex information-seeking tasks through iterative search and reasoning. However, they often issue intent-misaligned, under-constrained, or redundant search queries, resulting in poor evidence coverage and accumulating retrieval noise. To mitigate these limitations, we propose QREF, a plug-and-play Query Refiner and Evidence Filter. Specifically, the Query Refiner rewrites search queries according to the current reasoning intent while ensuring complementary coverage across each search batch. The Evidence Filter retains evidence relevant to both the current search batch and the reasoning objective, reducing distracting retrieval results. We initialize the Query Refiner with supervised fine-tuning and optimize it through reinforcement learning based on interactions with a real-world web search environment, reducing ineffective queries. Comprehensive experiments demonstrate that our approach improves the performance of DeepSearch agents across four challenging benchmarks, achieving average relative Pass@1 improvements of 33.55% on WebSailor and 11.04% on TongyiDeepResearch without modifying architectures or retraining their reasoning policies. Compared with the original DeepSearch backbone without QREF, the RL-trained refiner reduces the fraction of query batches that lack discriminative retrieval anchors by 95%. Compared with the SFT-trained refiner, it also lowers the filter-based ineffective query ratio.
benchmark - arxiv:2610.21500 · cs.LGWhen the KV Cache Is Slower Than the GPU: Scaling Remote KV Tiers for Context-Parallel PrefillMuhammad Ibrahim
Reusing the attention KV cache across requests is a standard way to cut prefill cost in long-context LLM serving, and production systems increasingly back the per-engine cache with a remote tier shared by every engine. We study such a tier in a disaggregated deployment of GLM-5.3, an FP8 mixture-of-experts model with multi-head latent attention (MLA), on 8xB300 nodes running prefill context parallelism (PCP8). Enabling the remote tier (LMCache with a Valkey store) left median time-to-first-token unchanged but raised the 16k-token p99 from 3.96 s to 17.3 s. Replaying a production request stream on an isolated engine, we find that a remote hit took 0.88 s against 0.31 s to recompute the prompt, and that synchronous loading turned a hit-heavy stream into an unstable queue: p99 83.32 s with the GPUs 7% busy. A two-term cost model explains both: a hit pays off only if the tier delivers more than B* = sR (2.53 GB/s for this model and GPU), and a fixed per-hit engine cost (T0 ~ 0.33 s) means that hits on prompts shorter than about 17k tokens cannot beat recompute at any bandwidth. Guided by the model, we (i) shard the MLA KV store across PCP ranks, with a proof that the existing minimum-over-ranks lookup returns the longest cached prefix and a load protocol that can never desynchronize collectives; (ii) partition the remote tier by rank, one server instance per rank with no coordination, which raises GET bandwidth 6.3x to 10.76 GB/s; (iii) add asynchronous loading to the sharded store; and (iv) fix three client bugs that caused unbounded stalls. On the hit-heavy stream, with 98.5% of prompt tokens served from the remote tier, p99 falls from 83.32 s to 0.67 s. The sharded store raises prefill throughput by 14%, and by 34% together with an upstream kernel fix. All changes are submitted upstream to LMCache.
long-context - arxiv:2610.21498 · cs.CVGRASE: Geometry-native Real-to-sim Agentic Scene Modeling from VideoZhaorui Lin, Haohan Liu, Zhen Chen, Xiangru Huang
Real-to-sim conversion aims to turn a captured video of a real environment into an editable 3D scene while preserving the geometry and layout of what was observed. Existing methods often follow a one-pass, loosely coupled paradigm: each stage receives only limited feedback from previous stages, allowing errors to propagate without reliable correction and leaving the final scene inconsistent. We present GRASE, an agentic framework that closes this loop by treating the captured multi-view video observations as a standing reference throughout scene modeling. Instead of committing to intermediate estimates, GRASE revises its current hypothesis whenever rendered evidence disagrees with the capture, overcoming the fixed intermediate assumptions of conventional pipelines. Specifically, GRASE first fits the architectural shell and its openings from depth and multi-view cues. An agent then places generated assets in a shared coordinate frame while checking their spatial relationships. Finally, each asset's geometry and placement are refined iteratively to improve agreement across the observed views. By coupling these steps through feedback, GRASE localizes structural discrepancies and revises only the responsible components, while preserving cross-view consistency. The resulting 3D scenes are faithful to the input, consistent across views, and suitable for simulations. In our experiments, GRASE substantially outperforms other state-of-the-art methods, including a direct real-to-sim conversion using GPT6-Astra.
agentagentic - arxiv:2610.21496 · cs.CLOPRD: On-Policy Reminder Distillation for Long-Horizon Coding AgentsYuechen Luo, Zhiyi Lai, Miao Ziqi, Zihang Xu +8
Reliable repository-level coding requires agents not only to generate correct code but also to preserve constraints and respond to tool feedback across long interactions. However, our diagnostic analysis shows that agents often fail to act on relevant information already available in context, while final Pass/Fail outcomes provide only sparse process supervision. We introduce On-Policy Reminder Distillation (OPRD), which combines Multi-Scope Reminder supervision with outcome-based routing on reminder-free student rollouts. Successful rollouts provide autoregressive self-imitation targets, whereas failed rollouts receive Reminder-conditioned reverse-KL supervision from a teacher that re-evaluates the same student-generated prefixes. A Global Reminder converts final-test diagnostics into trajectory-level guidance, while Turn-local Reminders apply only to the next decision following the relevant environment observation. The teacher and student share a two-stage initialization from Coding SFT and Instruction-Following (IF) SFT. At inference, the student operates without Reminders. On SWE-bench Verified, SWE-bench Pro, and SWE-bench Multilingual, OPRD improves the benchmark-averaged task resolution rate by 11.3 points over the open-source Qwen3.5-4B base model. Under the same diagnostic protocol, it reduces the aggregate error-event count from 446 to 222, a 50.2% reduction.
benchmark - arxiv:2610.21493 · cs.LGPhase Kernel Lifts Capacity of Dense Associative MemoryYifei Zhao, Ying Tang
Dense associative memory underlies modern attention mechanisms; however, Modern Hopfield Networks (MHNs) with linear kernels yield overlapping attractor basins and spurious states under high memory load, limiting their theoretically exponential storage capacity in practice. We construct a class of nonlinear phase kernels that modify the energy landscape and retrieval dynamics, and prove that these kernels preserve exponential storage capacity and local fixed-point convergence under given conditions on kernel smoothness, memory separation, and memory distributions. The framework is instantiated by the Sigmoidal Phase Hopfield Network (S-Hop), which guarantees monotone energy descent and mitigates practical capacity loss. S-Hop has $8.3\times$ and $4\times$ gains in one-step noisy retrieval load over Uniform Hopfield Networks (U-Hop) on CIFAR-10 and TinyImageNet respectively, and reaches a $50\times$ gain in first-failure load for ImageNet. The results demonstrate that phase kernels provide a principled route to robust dense associative memory.
memory - arxiv:2610.21483 · cs.AIFlashAttention with Statistics: Low-Overhead Attention Features for Compact Hallucination DetectorsAndrei Volodichev, Maksim Savichev, Nikita Sviridov, Artem Yuliy +2
Large language models can produce fluent answers that the supplied context does not support. A compact frozen proxy model can detect such answers by reading the context and another model's response, even from a proprietary generator. Attention-based detectors use summary statistics of the proxy's attention weights, but standard pipelines materialize the attention matrix, whose memory grows quadratically with sequence length. We compute these statistics without materializing attention, by accumulation in the FlashAttention pass, sparse lookups after it, or a second tiled pass over the scores, and classify them by this path into nearly free and costlier ones. On an H100, the statistics of five published detectors add 16% to the latency of a Qwen3-8B forward at 2K tokens, falling to 1.4% at 35K, and store 22.5 KiB per response token. On three benchmarks, probes trained on frozen Qwen3 proxies of 0.6B to 8B parameters prove strong detectors: a 4B proxy outperforms the strongest of nine zero-shot 4-9B LLM judges in F1 on two of the three, the proxies match or outperform specialized grounding checkers of up to 7B parameters, probing each proxy beats prompting it as a judge, and adding attention features significantly improves F1 in half of the 12 benchmark-proxy pairs, never significantly lowering it.
memorybenchmark - arxiv:2610.21477 · cs.AITool-Augmented Large Language Models for Reliable Medical Calculation from Patient NotesYang Zhong, Ling Luo, Xuanlyu Zhu, Yanhua Wang +3
Medical calculators translate raw patient observations into clinically actionable values and scores. However, applying these calculators directly to free-form patient notes requires precise variable extraction combined with the exact execution of complex formulas, scoring rules, and output conventions. Relying solely on direct large language model (LLM) prompting is unreliable in this context, as rule recall, parameter extraction, arithmetic reasoning, and formatting become entangled within a single generative step. To address these limitations, we propose a tool-augmented LLM framework for reliable medical calculation from patient notes. The framework retrieves and structures calculator specifications, generates reusable executable tools, and uses a constrained extractor to map each patient note to calculator-ready inputs. The tools then perform deterministic calculation, separating clinical language understanding from rule execution and arithmetic. Evaluated on the MedCalc-Bench dataset, our method achieves 86.73% final-answer accuracy, outperforming the one-shot prompting and supervised fine tuning-only baselines by 24.0 and 17.27 percentage points. Ablation and error analyses show that these performance gains are primarily driven by rule-grounded tool synthesis and schema-constrained extraction, whereas remaining failures are concentrated in clinical-semantic extraction and benchmark-specific output contracts. By grounding calculator logic in executable tools, the proposed approach alleviates arithmetic hallucinations and improves the computational reliability of the final clinical outputs.
benchmark - arxiv:2610.21468 · cs.AICan LLMs Tell Stories Out of Order? A Corpus and Benchmark for Nonlinear Narrative GenerationSyed Rifat Raiyan, Md. Tariquzzaman, Md Farhan Ishmam, Tenghao Huang
Every story has two orders: the order in which its events happen and the order in which they are told. A story is nonlinear when the two differ. We introduce a benchmark of 83 published works that measures whether language models can tell nonlinear stories. For each work, we annotate one list of events in both orders, together with a story graph of causal dependencies, character knowledge, and the questions the telling raises and resolves. Given events in chronological order, all 19 models tell them, on average, more chronologically than the published works do; asked for a specific structure, they land within the human range about half the time. We then compare a model's stories of the same events under different conditions, using both an automated judge and 71 human readers. Both prefer the chronological story a model writes by default to the nonlinear one it writes on request, which readers find harder to follow in 81% of decisive judgments. Readers also prefer stories written from the story graph to stories constrained to the published telling order (72% of decisive judgments), a difference the judge does not detect. Our results suggest that following an instruction to reorder a story is not yet the same as telling it well out of order.
benchmark - arxiv:2610.21463 · cs.LGTrajectory-Level Speculation with Semantic Verification for Tool-Using AgentsZilin Wang, Yinhong Liu, Chao Chen, Qun Liu +1
Large language model agents solve tasks through long sequences of tool calls, and the prevailing design has a single strong model generate every step, although many steps are routine enough for a much smaller model. Existing ways of sharing this work keep the strong model as the generator: routers and collaborative agents decide which model handles each query or step, and speculative methods check a small model's drafts one step at a time, usually against the strong model's own action. We introduce SpecAgent, a collaboration protocol that turns the strong model from the author of each step into a reviewer of steps drafted by a weak model. The weak model executes read-only tool calls as it drafts, so that every step is grounded in real observations, and holds the first action that would modify the environment. The strong model then verifies the drafted segment in a single prefill pass, commits the longest prefix it judges semantically valid, and repairs the first rejected step. Because verification replaces autoregressive decoding with parallel prefill, the strong model decodes only short verdicts and occasional repairs rather than every step. On $τ^2$-Bench, SpecAgent comes within one point of the strong model's success rate while roughly halving memory traffic and wall time, and on AutomationBench and VitaBench it recovers 44-75% of the quality gap while lowering both. In an edge-cloud deployment on MCP-Atlas with three cloud models, it closes 53-75% of the weak-strong quality gap while the strong model generates at most half as many tokens. These results suggest that a strong model need not author every step of an agent: reviewing grounded drafts retains most of its quality while cutting both latency and memory traffic.
memory - arxiv:2610.21453 · cs.ROCarryOn: A Dual-Loop Harness for Personalized and Reliable Robot AssistanceZixuan Wang, Yuqi Liu, Yuxin Chen, Sengqiao Yang +6
While generalist vision-language-action (VLA) models have enabled diverse open-world manipulation, real-world robotic assistance requires two critical capabilities they currently lack: personalization and reliable execution. First, executing an instruction like "make me a sandwich" requires grounding subtask decomposition in persistent user preferences rather than relying on a single generic solution. Second, reliable deployment demands explainable safety assurance that proactively prevents hazards and gracefully recovers from occasional physical failures. To address these gaps, we present CarryOn, a dual-loop harness designed to augment a frozen VLA policy. The planning loop maintains persistent user-level and scene-level context that is continually grounded against new observations and revised through interaction, enabling personalized task selection and execution grounding across conversations and changing scenes. The execution loop actively supervises policy rollouts, shielding actions from predicted hazardous contact. When execution fails, the system invokes reusable control-based recovery skills that restore task progress and are refined through reflection on earlier failures. On a real robot and in simulation, CarryOn dispatches preference-consistent subtasks across users and scenes, reduces the violation rate of an already safety-finetuned VLA from 52.6% to 6.1%, and increases recovery success from 13% to 80% when refined skills are reused.
vision-language-actionvlavla policymanipulation - arxiv:2610.21451 · cs.CVPET Quantification-Preserving Registration for Longitudinal Whole-Body PSMA PET/CTFryderyk Kögl, Anna Reithmeir, Daniel M. Lang, Emily Chan +2
We adapt the diffeomorphic coarse-to-fine LapIRN architecture for longitudinal whole-body PSMA PET/CT registration in the Learn2Reg 2026 PSMAReg challenge. The method combines affine pre-registration, multi-modal deformable registration, and test-time instance optimization. We further introduce PET biomarker preservation losses for metabolic tumor volume (MTV) and total lesion glycolysis (TLG), including Jacobian-based constraints, and a per-structure rigidity regularization for skeletal structures. The submitted configuration achieved 81.0% DSC and 6.17 mm HD95, with MTV and TLG errors of 3.3% and 3.3%, respectively, on our evaluation protocol. Code is available at https://github.com/koegl/psmareg.
evaluation protocol - arxiv:2610.21449 · cs.AIIncentivizing Tool-Use Reinforcement Learning for Post-Arrival Emergency Care PredictionPengfei Hu, Chenwei Wu, Qiao Jin, Jun Deng +3
Large language models (LLMs) can reason over both structured measurements and free-text records, which makes them attractive for emergency department (ED) prediction. However, most existing LLMs still fall short of conventional ML methods like trees on EHR-based prediction tasks. Our controlled analysis traces this gap to two weaknesses: LLMs do not select informative features, and they misjudge what numerical values imply for risk even when given the right features. We propose TIR-EC, a tool-integrated reinforcement learning framework that addresses both. TIR-EC exposes EHR feature groups as tools the LLM can call. It trains the LLM with rewards from a gradient-boosted tree teacher, which provides a reference set of informative tools and a graded risk target. TIR-EC also distills task-specific clinical skills into the model weights through a reverse curriculum that gradually removes the skills from the prompt during RL process. Evaluation on three clinical prediction tasks shows that our 4B agent achieves the highest macro F1 of all compared models, outperforms conventional ML baselines, and is competitive with much larger frontier LLMs. Further analysis shows that TIR-EC selects more useful evidence with fewer tool calls and combines measurements more effectively.
agenttool-use - arxiv:2610.21448 · cs.LGV-SHARE: Provable Convergence Guarantees for Advantage-based AlgorithmsKintan Saha, Hamed Masoudi, Xinyi Sheng, Dominik Baumann +2
In reinforcement learning, advantage-based value-learning methods, unlike standard Q-learning, share information across actions by decomposing the action-value function into a state-dependent value component and action-specific advantages. While this mechanism has proven effective in both tabular and deep reinforcement learning, its theoretical understanding remains incomplete. We introduce V-SHARE, an advantage-based value-learning algorithm that combines cross-action value sharing with an importance-weighted Bellman-residual correction. To our knowledge, we provide the first complete theoretical characterization of an advantage-based algorithm, including both asymptotic convergence and non-asymptotic finite-sample guarantees. We identify a regime in which V-SHARE mitigates the rare-action bottleneck of asynchronous Q-learning and uncover a contraction--variance tradeoff between stronger deterministic contraction and increased importance-weighting noise. Motivated by this tradeoff, we develop decreasing residual-preconditioning schedules that combine faster early contraction with lower asymptotic noise. We further develop a two-timescale analysis that removes the single-timescale residual-weight threshold while preserving convergence guarantees. Experiments on discrete Markov decision processes (MDPs) and deep reinforcement-learning benchmarks corroborate our results.
benchmark - arxiv:2610.21440 · cs.CLSITAR: A Situational Distortion Metric for Automatic Speech Recognition EvaluationMingzhe Lu, Jinhao Liu, Yunpeng Li, Hang Su +2
In automatic speech recognition (ASR), transcription errors with similar lexical or semantic discrepancies can have markedly different consequences for the situation conveyed by speech. Capturing these differences requires not merely contextual assessment, but an understanding of what the spoken discourse describes, establishes, or seeks to bring about. However, relying on LLM judges introduces computational cost and scoring variability. We introduce SITAR, a situational distortion metric for ASR, realized through a comprehensive end-to-end construction framework. Specifically, this framework aggregates situational severity and cumulative extent into a unified score via a bounded recursive formulation. Using targets derived from this formulation with limited offline LLM support, we train SITAR as a lightweight BERT-family regressor that scores reference-hypothesis pairs in a single forward pass, entirely free of any prohibitive runtime costs. Recognizing that introducing a new metric demands rigorous justification, we formalize a principled validation protocol spanning both extrinsic validity, assessed at macro and micro levels, and intrinsic scorer reliability. The former, evaluated across diverse datasets and models, demonstrates that SITAR reliably tracks broad capability scaling while providing orthogonal diagnostic utility, whereas the latter, examined on a purpose-built controlled benchmark, confirms the operational soundness of the scoring instrument itself. Through systematic validation, we demonstrate that SITAR holds strong potential as a practical, lightweight complement to existing ASR metrics. We will release the full recipes (code, model weights, dataset, etc.) after the internal review to support community research.
benchmark - arxiv:2610.21439 · cs.LGEach Reward to Its Relevant Tokens: Reward-Specific Token-Level Credit Assignment for Multi-Reward RLSkylar Zhai, Haonan Wang, Zhining Liu, Linxin Song +3
Language models are increasingly trained against several verifiable rewards at once, such as the correctness and the format of a tool call, or the individual constraints of a complex instruction. Reward-decoupled normalization (GDPO) normalizes each reward separately so that different reward combinations keep distinct advantages, but it still spreads every reward's advantage evenly over the whole response, although each verifier checks only a few positions. We propose MRTPO, which assigns the advantage of each reward to the tokens relevant to it, using the information its verifier checks, without auxiliary models or token-level annotations. For each reward, the policy replays its own response with and without that reward's verifier information, and the per-token Jensen-Shannon distance between the two next-token distributions measures each token's relevance and reassigns the advantage. On text-to-SQL, multi-constraint instruction following and tool calling, MRTPO converges faster and achieves better performance than the compared methods.
tool calling - arxiv:2610.21438 · cs.CVRLDF: A Region-Level Deepfake Dataset and Benchmark for Partial-Face Manipulation DetectionShukesh Reddy, Abhijit Das
Existing deepfake detection models are built upon manipu- lated images that are tampered with either exclusively fully swapped or fully synthesised faces, which leads us to an open question: Can off-the-shelf deepfake models detect manipulation effectively while tampering is confined to a local facial sub-region? To fill this gap, we introduce RLDF (Region-Level Deepfake), the first dataset for region-level deepfake evaluation. Employing the proposed region-aware partial swap synthesis model, we generate over 2M par- tial face-swap images by using eight face-swapping meth- ods. Altogether, we considered seven face regions, among which there are four discrete regions (eye, nose, mouth and eyebrow) and four composite facial regions (eye and nose; eye, nose and mouth; eye, nose, mouth and eye- brow; and full face). In addition, we define three evalu- ation protocols and benchmark 12 state-of-the-art detectors across the five popular standard cross-dataset evaluations, DF40, and the proposed RLDF, providing a comprehen- sive audit of where and why existing detectors underper- formed. Our evaluation reveals that existing detectors fail to generalize consistently across partial facial regions, re- sulting in an overall drop of 10.5% AUC and 10% accu- racy under region-level evaluation. Further, we propose a mask-guided cross-attention model as a baseline approach that can effectively detect region-level manipulation. The proposed dataset, code, and checkpoints are available at https://github.com/RLDF-Dataset/RLDF.
manipulationbenchmark - arxiv:2610.21434 · cs.ROPolicy-Aligned Predictive Verification for Adaptive VLA Action-Chunk ExecutionMinkyu Kim, Junhoo Lee, Seungyeon Kim, Suhyun Jeon +3
Vision-language-action (VLA) policies generate action sequences that reduce repeated inference, but deciding how much of each sequence to execute is critical to task success. Existing adaptive methods make this decision using generation stability or prediction agreement, yet these signals do not establish whether executing a candidate prefix preserves the ability to complete the task. We introduce a policy-aligned predictive verifier (PAV) that anticipates the consequences of candidate action prefixes and evaluates their compatibility with successful continuation before execution. Rather than using future prediction to generate actions, our method evaluates existing proposals in the VLA's internal task-conditioned representation space. To learn local judgments from terminal task outcomes, we combine trajectory-level survival supervision with a localization prior that preserves useful behavior within failed rollouts. The resulting validity scores guide execution-horizon selection. Evaluations with pi_0 and pi_0.5 across LIBERO, RoboTwin 2.0, and real-world manipulation show improved success-query trade-offs. These findings point toward a different role for future prediction in robotics, using it to assess how far to trust a policy's proposed actions rather than to generate new ones.
vision-language-actionvlamanipulationliberorobotwin - arxiv:2610.21429 · cs.ROMuSTMM: Learning Sound-Triggered Mobile Manipulation across Multiple ScenesHao Ju, Feng Yan, Zhihao Zhong, Heng Yang +2
Sound-Triggered Mobile Manipulation (STMM) requires robots to respond to sound events by navigating to the associated objects and performing appropriate physical interactions. Existing STMM evaluations primarily vary object layouts within limited architectural structures, leaving challenges to unseen scenes insufficiently explored. We introduce Multi-scene Sound-Triggered Mobile Manipulation (MuSTMM), a dataset that combines spatial audio rendering and physical interaction across 31 scenes with diverse architectures and object layouts. MuSTMM contains 14.36 million precomputed binaural Room Impulse Responses (RIRs) and a scene-disjoint split for evaluating single-event and multi-event settings in unseen scenes. We further augment the Navigation Policy Model with explicit spatial memory that complements acoustic cues with observed geometry. Specifically, we accumulate pose-aligned depth observations in a shared coordinate frame and project this memory into an egocentric Bird's-Eye View (BEV) occupancy map. Fusing map features with sound context produces a goal descriptor that provides both target-location and scene-structure priors, guiding navigation to the target region required for subsequent physical interaction. This policy achieves a 2.60% success rate boost over existing methods, while component ablations further support the effectiveness of spatial memory. MuSTMM and the navigation policy provide a testbed for the new challenge of studying sound-triggered mobile manipulation in unseen scenes. The code and dataset will be released upon acceptance.
manipulationmemory - arxiv:2610.21427 · cs.AILearning from Interaction Experience via Localized Context DistillationZhezheng Hao, Yongkang Yang, Xilin Xia, Jiarui Yu +8
Learning from interaction experience is essential for enabling large language models to continually improve during deployment. Context distillation offers a promising path: privileged information extracted from experience is provided to a teacher, whose behavior is then distilled into a student model that has no access to it. However, both existing forms of context distillation have limitations. Off-policy context distillation trains the student on teacher-generated trajectories and thus suffers from exposure bias; whereas on-policy context distillation trains on the student's own rollouts, whose states are often uninformative and fail to convey behaviors the student cannot yet produce. To resolve this dilemma, we introduce Localized Context Distillation (LoCD). LoCD first localizes the critical spans of a student trajectory that require revision. The teacher then regenerates only these spans from the student's own prefix, and the student is trained on the regenerated segments. LoCD thereby concentrates learning on the most informative states while keeping exposure bias low. Extensive experiments across math reasoning, code generation, and agent tasks demonstrate that LoCD consistently outperforms strong baselines and achieves more stable training with smaller and sparser parameter updates.
agent - arxiv:2610.21422 · cs.CVWhere, How, and When: Query-Guided Grounding for Video UnderstandingYogesh Kumar
Understanding video content requires reasoning at multiple levels of granularity, from identifying objects to interpreting their interactions and localizing when specific actions occur. This thesis studies three query-guided grounding tasks that answer where key objects are (object-level grounding), how they interact (relationship-level grounding), and when specific moments occur (moment-level grounding). For object-level grounding, we present QDETRv, a query-guided extension of the Detection Transformer for one-shot localization of unseen objects in videos, which combines query-context fusion, a memory module for temporal consistency, and unsupervised pretraining on synthetic object trajectories. We further propose the Temporal Object-Aware Vision Transformer (TOViT) for few-shot video object detection, which builds on a language-aligned open-set vision encoder and propagates high-confidence object features across frames. For relationship-level grounding, we introduce few-shot referring relationships in videos: given a (subject, predicate, object) query and a few support videos sharing the predicate, the entities are localized by optimizing a T-partite random field using a meta-learned relation network with query-conditioned translational embeddings, Global Semantic Aggregation, Local Localization Aggregation, and belief-propagation-based message passing. For moment-level grounding, we propose the Moment Alignment Transformer (MATR) for video-to-video moment retrieval, which uses dual-stage Dynamic Time Warping for pre- and post-fusion alignment, learnable queries for boundary prediction, and self-supervised pretraining with temporal augmentations; MATR also extends to text-to-video moment retrieval with performance comparable to, or better than, the state of the art. These contributions advance query-guided video understanding across spatial, relational, and temporal grounding.
memorymemory module - arxiv:2610.21419 · cs.LGWhat Does History Add? Calendar Controls and Information-Timing Audits for University Trajectory ForecastingZepei Cao
We audit the predictive value of historical summaries in a rule-based university trajectory simulator. Seventeen ridge specifications and persistence separate current plans, past-state and past-plan summaries, and calendar-only history-length indicators. Across training, validation and two family-held-out evaluations, we reconstruct 16,750 routes and 804,000 monthly records. On the original 40-family set, six-month history lowers six-step normalized RMSE from 0.159234 to 0.151522, but its difference from a length-only control is -0.000895 (descriptive paired 95% interval [-0.001957,0.000248]). Fixed weights on 30 families unused in identified prior tasks reproduce this pattern: 0.156480 to 0.150074 versus current-state prediction, but -0.000558 [-0.001783,0.000728] versus the length control. These uncertain comparisons do not generalize to all-history summaries, which improve on their own calendar control. Exploratory decomposition finds clearer preparation-index and funds benefits, whereas fatigue dominates the six-month model's loss and has an uncertain length-controlled benefit. Moreover, 28.20% of its six-step fatigue predictions violate the simulator's [5,90] support. We document family dependence, prior exposure, strict temporal indexing, matched recursive references and independent numerical replay. The task conditions on recorded future plans; it does not evaluate real-student forecasts, causal action effects, autonomous planning or language-model memory. This is an evaluation audit, not a new memory architecture.
memorymemory architecture - arxiv:2610.21412 · cs.AIJevAgentHarness: Benchmarking, Deploying and Training Decision Models That Share Control with LLMsKunlun Zhu, Tao Feng, Xu Han, Heng Wang +4
Large language model (LLM) agents now fix software, operate computers and serve customers, but their harnesses, the programs that run them, call a large LLM at every step, which makes long tasks slow and costly. Most of these calls end in a multiple-choice decision, such as which tool to call next. Decision models such as Jev answer these questions directly, at a fraction of an LLM call's time and price. We ask how a harness can delegate these decisions to a decision model with little loss of success. To evaluate decision models in agentic tasks, we build AgentDecisionBench, which tests them from isolated questions up to control of a running agent, on public benchmarks and our DecisionGym environment. The decision model most accurate on isolated questions solved only 69.5% of agent tasks, and without a fitted confidence threshold, delegating tool choices lowered success by 19.0 points. Building on these findings, we then design JevAgentHarness, which fits this threshold for each choice on the agent's own trajectories and lets the LLM decide below it. Because existing decision models never see the states an agent reaches after its own mistakes, we train AgentDecision-4B, an open 4B model, on such states with Jev's answers as labels. On tasks unlike its training data, the agent solved 91.5% of them with AgentDecision-4B and 82.0% with Jev. On tasks like their training data, 4B models trained this way made agents 81.1-89.0% cheaper and up to 2.5 times faster, without losing 3 or more points of task success. The models, data, and harness will be released at https://github.com/JevAgentHarness/JevAgentHarness.
agentagenticbenchmark - arxiv:2610.21411 · cs.AIJev-LiteMAS: Separating Decisions from Generation in Multi-Agent WorkflowsKaixiang Wang, Zhaijiacheng Zhou, Jiong Lou, Jie Li
Multi-agent LLM systems rely on generative models for expert execution, workflow planning, and artifact review. Yet many coordination operations require only bounded decisions, such as selecting the next expert or judging artifact quality. Profiling three representative MAS frameworks, we find that routing and review account for 31.7-43.5% of generative-LLM tokens. However, replacing these operations with low-cost decision models introduces two interface gaps: generative planners produce execution plans and handoff instructions (planning gap), while reviewers provide actionable feedback for repairing flawed artifacts (review gap), which bounded decisions alone cannot provide. We introduce Jev-LiteMAS, a multi-agent framework that bridges both gaps by combining a lightweight decision model (Jev) with generative execution and feedback. To address the planning gap, Jev selects experts dynamically from the task and prior expert outputs, using accumulated context to ground subsequent execution without a pre-generated trajectory. To address the review gap, Jev screens intermediate artifacts against task criteria: passing artifacts proceed directly, whereas flagged ones undergo LLM re-review for acceptance or corrective feedback. Across four benchmarks and two backbones, Jev-LiteMAS reduces generative-LLM tokens by 12.7-54.6% over LAS, maintaining accuracy within 1.34 percentage points and lowering latency in seven of eight settings. Grounding routing in substantive expert outputs prevents up to 18.0 points in accuracy loss.
multi-agentagent frameworkbenchmark - arxiv:2610.21405 · cs.AIMemCut: Every Secret Leaves a Trail in the Long-Term Memory of LLM AgentsCheng Zhang, Xuanning Cao, Zhicheng Zheng, Hui Xiong
Long-term memory has become a crucial component of large language model (LLM) agents, with most research focusing on retrieving relevant information accurately and efficiently. However, a dependable memory system must also forget sensitive user data upon request. To evaluate this capability, we introduce MemCut, a systematic benchmark for user-requested forgetting. Its 456 hand-designed secrets in long-horizon conversations have fully annotated evidence trails to test recovery after deletion, including from messages that never state the secrets directly. We propose two recovery metrics. Native recovery measures reader success through the system's retrieval, whereas residual recovery measures the same reader's success when directly given surviving objects traced to annotated disclosures or evidence. We establish three reference deletion baselines to test how far existing retrievers and write models support forgetting across five memory designs and four LLMs. Disclosure removal reduces native recovery far more than residual recovery. LLM-guided deletion of ten additional turns from twenty retrieval candidates leaves 39-76% residual recovery. Subset sweeps (144 facts, three systems, two models) show that widening the pool at a fixed budget reduces residual recovery more than deleting all original candidates. These sweeps identify candidate coverage as a bottleneck in the reference baselines.
memoryllm agentbenchmark - arxiv:2610.21404 · cs.LGMulti-Reward Hacking: From Aggregation-Induced Drift to Trust-Anchored DefenseYushu Li, Jiin Woo, Xiaoxiao Li, Youngsuk Park
To improve task performance while encouraging desirable response properties, LLM post-training often combines a primary reward (e.g., correctness) with auxiliary rewards (e.g., brevity or efficiency). However, how these rewards are aggregated can critically affect training outcomes. We show that symmetric aggregation, which treats reward channels equally, can induce a failure where aggregate reward improves while primary-task performance degrades, even when individual rewards are meaningful. We investigate this aggregation-induced failure through a one-step multi-reward decision formulation, characterizing when it constitutes multi-reward hacking. We prove that, under symmetric aggregation, deterministic conditional-mean GRPO converges to auxiliary-favoring behavior when auxiliary rewards are more reliably obtained, regardless of initialization. We propose Trust-Anchored Gating (TAG), which first elects an anchor channel based on the policy drift induced by optimizing each channel alone, then gates auxiliary credit on anchor satisfaction. Unlike calibration-sensitive static reweighting, TAG structurally eliminates credit for auxiliary-only outcomes while preserving additional credit for joint success. In the same analytical setting, TAG provably reverses this selection, converging to the primary-favoring action from any initialization when the primary channel is anchored. Experiments across Qwen and Llama models on mathematics and program synthesis reveal substantial degradation under symmetric, constrained, and normalized aggregation, whereas TAG mitigates this failure. On Qwen3-4B with a brevity reward, TAG achieves 0.581 pass@4 versus 0.220 for additive aggregation, while shortening responses by 44% relative to correctness-only training.
post-training - arxiv:2610.21402 · cs.LGFault Detection and Bit-Error Rate Diagnostics in Optical Wireless Links Using Edge-AISarthak Bhattarai, Sheshadri Shikhar Bandyopadhyay, Suprabh Dhakal, Jitendra Kumar Manandhar
Optical Wireless Communication (OWC) systems provide high-bandwidth, spatially secure, and RF-interference-free digital transmission. However, direct free-space optical links suffer from extreme sensitivity to line-of-sight disruptions, optical attenuation, high ambient light saturation, and transmitter-receiver misalignment. Traditional channel diagnostic frameworks rely on external computing nodes or cloud-based processing, which introduces severe diagnostic latency and reduces edge node autonomy. This paper presents an autonomous, self-diagnosing optical wireless receiver node that integrates physical-layer signal telemetry extraction with an on-device TinyML classification engine. Built around an ESP32-S3 microcontroller, the receiver recovers 1000 bit/s On-Off Keying (OOK) digital payloads using a 5-sample majority voting mechanism while concurrently extracting real-time channel features: Bit Error Rate (BER), received signal voltage (Vsignal), Signal-to-Noise Ratio (SNRdB), and ambient light baseline (Lambient). Evaluated against a Random Forest baseline (77.22% accuracy) and a floating-point neural network (78.48% accuracy), a post-training INT8 quantized TensorFlow Lite Micro model occupying only 3.32 kB of SRAM runs directly on the receiver firmware. The quantized model achieves 74.68% overall accuracy across four distinct operational conditions: normal transmission, optical filter attenuation, ambient light interference, and severe obstruction/misalignment. On-device deployment tests confirm real-time, deterministic fault classification, establishing a practical approach for self-aware optical IoT nodes.
post-training - arxiv:2610.21400 · cs.ROModular Inverse Dynamics and Wrench Recovery for Closed-Chain Mechanisms Using Path AssemblyMohammad Dastranj, Jouni Mattila
Rigid closed chains can admit indeterminate internal reactions even when their motion is fully specified. This paper develops a modular inverse-dynamics workflow that combines path-based motion reconstruction with recovery of uniquely determined actuator and joint-wrench outputs. Analytical second-order differentiation of assembled transformation paths supplies passive velocities and accelerations on a regular closure branch. Spatial Newton--Euler balances are then assembled over the physical body-and-joint graph. Established nullspace and dual criteria identify and recover unique outputs, while balance, ideal-joint admissibility and affine-family diagnostics assess nonunique reference reactions. Evaluation against Simscape Multibody uses prescribed motion of a seven-degree-of-freedom heavy-duty hybrid serial--parallel manipulator under common rigid-body, ideal-joint assumptions. The results show close agreement with the reference for passive positions, actuator efforts and uniquely recoverable constraint-wrench components. Reaction diagnostics support compatibility with the admissible numerical load family.
manipulator - arxiv:2610.21397 · cs.AIWatts If? A Benchmark for Counterfactual Reasoning About the AI Data Center BuildoutPankaj Chouhan, Rajarshi Das, Aditya Dubey, Manminder Tomar +4
An AI data center is committed years before it draws a watt: a few discrete choices (grid or on-site gas, evaporative or dry cooling, which county to file in) fix its capital, water draw, and carbon path, and LLMs are now asked to help screen them with no test of whether they track what each commitment rules out. We introduce DCBench-440, with two arms reported separately: 200 generative questions over STRIPS worlds for five US deal archetypes, with search-verified gold answers and no tested shortcut above 12% on any question type, and 240 multiple-choice questions grounded in quoted site records, emission equations, and stated scenarios. Across open-weight deployments, accuracy spans 12.5-71.5% (generative) and 58.3-92.5% (multiple choice). One skill fails everywhere: asked to list every action available at a decision gate, no model exceeds 4.0% exact match. The strongest list 72-85% of the actions that would change the project's state but omit most permitted actions that change nothing; with those excused, the best exact match is still 24%. Code, data, and other materials are available upon request.
benchmark - arxiv:2610.21395 · cs.AIWhen do quantum views help? Complementary classical-quantum representations in multi-view fusionAzadeh Alavi, Fatemeh Kouchmeshki, Hossein Akhoundi, Yongli Re +2
Recent studies have reported promising results from quantum machine learning in drug discovery, healthcare, materials science and cybersecurity. Research includes comparisons with classical machine learning algorithms and the development of sequential hybrid pipelines. We propose that fusion can improve prediction when classical and quantum representations capture complementary predictive relationships. Our classical-quantum multi-view framework preserves the full classical input alongside simulated quantum features. Concatenation and attention combine the views; single-view comparisons, replacement and masking controls assess both views' contributions. Controlled angular-polynomial experiments distinguish targets adequately represented by a single view from those that depend on complementary angular and polynomial relationships. For these mixed targets, fusion improves prediction by combining the relevant feature sets, while matched classical representations recover the same benefit, linking the improvement to their complementary predictive content. Controlled classification tests further show that gains depend on alignment between the chosen representation and the target, and on a suitable fusion rule. Benchmark validation shows improved mean predictive performance over classical baselines on Wine, Wisconsin Diagnostic Breast Cancer, three-class Fashion-MNIST and Steel Plates Faults within the evaluated protocols. Classical replacement and masking controls remain competitive, classical baselines lead on Digits and Covertype, and outcomes with missing or corrupted views depend on the available inputs. These findings support evaluating quantum representations for their complementary predictive contribution and matching the representation and fusion strategy to the target relationships.
benchmark - arxiv:2610.21394 · cs.AIMatoi: Selective Revalidation of Consequence Evidence for Agent AuthorizationZexun Wang
Matoi separates reusable consequence evidence from current execution authority and durable occurrence history. Its finite query-aware contract covers branches, ordered collections and negative lookups. A consumer reuses only its own checked evidence whose observations remain valid; signed requests, live grants, quota and uncertain outcomes remain separate obligations. Conditional correctness arguments distinguish locally serialized claims from nontransactional remote effects. Two studies strengthen cost attribution and domain evaluation. A 1,440-cell matched-wire study repeats 144 authored inputs. With compact input bytes identical, reuse reduces complete lifecycle cost by 3.49 percent in JS and 1.79 percent in Python. Combined representation/reuse reductions against full proof are 10.87 and 16.53 percent; five of twelve family intervals for the matched comparison cross one in each language. A prospectively frozen AgentDojo Banking mapping covers 5/16 user plans and 7/9 attack plans. Both governed strategies preserve five exact benign transfers and refuse 35 reused attack-plan/policy crosses. Pre-satisfied utility predicates are reported separately from completed work. These are original scripted synthetic tools, not autonomous agents or a real bank. Earlier studies retain cheaper protected-cache alternatives, policy-compatible phishing, provider duplication and post-claim revocation. Separate JS/Python consumers establish finite-profile interoperability within the same project; durable recovery rechecks evidence and retains uncertain reservations. The results support a bounded, auditable reuse and execution contract, not a new general incremental algorithm, arbitrary-tool closure, atomic third-party commitment or production superiority.
agentautonomous agent - arxiv:2610.21392 · cs.AILearning to Reason with Selective Internal FeedbackXun Lu, Shumeng Yang, Xinrui Wang, Songjun Tu +2
Reinforcement learning from internal feedback can improve language models without ground-truth answers or external verifiers. However, confidence and correctness can diverge during training, and continued optimization of average self-certainty can increase confidence while degrading reasoning accuracy. We introduce SIF, a reinforcement learning method that combines comparative reward construction with selective policy updates. SIF compares ranked token-level self-certainty profiles across multiple windows and peers, assigning positive rewards only when both levels of voting provide sufficient support. Sample and group level filtering determines which responses contribute to policy optimization. Experiments on three backbones across 12 mathematical and scientific reasoning benchmarks show that SIF's final checkpoints achieve higher average accuracy than the best-validation checkpoints of Intuitor and Vigor, with a gain of 4.71 percentage points over Intuitor on Qwen2.5-Math-1.5B. Further analyses show better retention of accuracy during continued optimization, while test-time training experiments demonstrate sustained improvements with as few as four rollouts per prompt.
benchmark - arxiv:2610.21390 · cs.AIPriorityFEC: Deadline-Aware Streaming FEC for KV-Cache Handoff in Edge Multi-Agent SystemsHan Shen, Qifu Tyler Sun, Fanyang Meng, Ye Wang +1
Multi-agent large language model (LLM) systems are increasingly deployed at the edge to support complex tasks that require collaboration and information exchange among multiple agents. However, when task contexts are passed across agents, shared context that has already been processed by upstream agents often needs to be prefilled again, resulting in substantial redundant computation. Reusing an already prefilled key--value (KV) cache can avoid this recomputation, but requires the reusable KV state to be transferred across the network before the receiving agent can begin generation. Under strict time-to-first-token (TTFT) deadlines and unreliable edge links, delivering the complete KV cache may be infeasible, making it necessary to prioritize and protect the most valuable KV information within a limited transmission budget. This paper proposes PriorityFEC, a utility-aware KV-cache handoff scheme that jointly exploits model-side KV importance and transport-side reliability. PriorityFEC first constructs a utility-aware source stream by placing high-utility KV blocks at earlier transmission positions. It then protects the ordered stream using a feedback-adaptive Streaming Forward Erasure Correction (FEC) mechanism, which schedules repair packets according to receiver feedback before the TTFT deadline. We implement PriorityFEC on a distributed edge LLM serving prototype and evaluate it with Qwen2.5-7B-Instruct and Llama-3.1-8B-Instruct across multiple datasets. Results show that PriorityFEC achieves superior task quality over the compared transport schemes, with its advantages becoming particularly pronounced under stringent deadline constraints. These results demonstrate the importance of jointly optimizing KV-cache utility and deadline-aware transport reliability when reusable LLM state must be handed off across unreliable edge networks.
agentmulti-agentagent system - arxiv:2610.21389 · cs.ROFast-VLA: You Only Think Once for Robot ManipulationKehong Gong, Guanli Hou, Xiaoxu Zheng, Qi Wang +5
Vision-language-action (VLA) policies commonly place a multi-billion-parameter model inside the feedback loop, repeatedly recomputing task context whenever they predict a new action chunk. Yet task understanding changes far more slowly than the robot state. We introduce Fast-VLA, which invokes a frozen 4B VLM once per episode to contextualize the instruction and initial views as fixed, multi-layer key-value task memory. During execution, a lightweight online actor fuses learned time- and arm-indexed queries with current visual and proprioceptive feedback to form action queries. These queries retrieve task context from the fixed memory and predict action chunks without re-running the foundation model. Fast-VLA achieves 92.46% and 92.40% success on RoboTwin 2.0 Clean and Randomized, and 95.95% and 79.11% on LIBERO and LIBERO-Plus evaluation. Against three continually conditioned baselines on the same hardware, it reduces RoboTwin amortized policy inference time by 10.9-17.1x and requires 1.22 GiB of post-prefill online resident memory. Task-memory swaps further show that the actor relies on episode-matched task memory. These results show that task context can be established once while closed-loop control remains responsive throughout execution. Project page: https://garfield-kh.github.io/FastVLA
vision-language-actionmanipulationliberorobotwinmemory - arxiv:2610.21382 · cs.LGRDL-Bench: A Benchmark for Task-Aware Question Answering over Relational DatabasesYao Cheng, Siqiang Luo
Relational databases (RDBs) remain the primary infrastructure for storing and managing structured data across enterprise and scientific domains. Natural-language (NL) interfaces to RDBs aim to let users ask questions without writing SQL, but real user needs go beyond SQL-style lookup: some questions require reasoning over the observable database state, while others require predictive inference from historical relational records under temporal constraints. Existing benchmarks address these goals in isolation. To address this gap, we introduce RDL-Bench, a benchmark for task-aware question answering (QA) over real RDBs. RDL-Bench organizes diverse user questions into reasoning-oriented questions, whose answers are determined by the current database state, and prediction-oriented questions, whose answers depend on future outcomes under temporal constraints. Each instance contains an NL question, task metadata, schema information, temporal context, either executable evidence or target-construction logic, and denotation-level ground truth. RDL-Bench supports two evaluation settings: native solvers under their intended capabilities, and task-aware modular systems where a router dispatches each question to an appropriate execution module.
benchmark - arxiv:2610.21380 · cs.CVPerception or Reasoning? Diagnosing and Tackling Reinforcement Learning Bottlenecks in VLMsZeyuan Yang, Zhuohao Chen, Jiaben Chen, Yao Zhang +2
Reinforcement learning has brought strong gains to language model reasoning, but its improvements on visual tasks remain less consistent. One difficulty is that visual task failures can arise from different sources: the same benchmark may be limited by perception for one model and by reasoning for another. We introduce Perception and Reasoning Attribution (PARA), which measures the remaining perceptual and reasoning headroom of each model--benchmark pair by comparing how many current errors are recovered with more accurate visual information versus a stronger reasoner. Across 4 vision-language models and 17 multimodal benchmarks, PARA broadly agrees with the expected characteristics of existing benchmarks while revealing substantial model-specific variation. More importantly, it shows that RL gains become smaller as perceptual headroom increases, but larger when reasoning explains more of the recoverable errors, suggesting that current RL exploits reasoning headroom more effectively than perceptual headroom. Motivated by this gap, we analyze how existing RL objectives affect visual perception and find that reward- and token-level signals influence visual attention only indirectly. We therefore propose GRAD, a group-relative attention distillation objective that uses the best rollout in each group to directly guide the visual attention of weaker rollouts. GRAD consistently improves performance across two model scales and seven benchmarks, and further complements existing RL methods.
benchmark - arxiv:2610.21375 · cs.LGMemorilla: Latent Semantic Memory for LLMsAdibvafa Fallahpour, Parsa Idehpour, Vignesh Kothapalli, Nikita Mounier +3
Memory enables large language models (LLMs) to use information beyond their parameters, such as past user conversations, long documents, or an agent's earlier actions. For a frozen LLM, access to this accumulated knowledge is limited to what a memory system places in its context window. However, gathering and combining relevant information needed for each query while keeping the model's context short remains challenging for existing systems. Here we introduce Memorilla, a memory system that gathers information from an entire document collection into a fixed number of query-dependent 'memory tokens' for a frozen LLM. Memorilla encodes documents (such as user chats, book chapters) once into reusable embeddings, then learns to retrieve and place query-specific information in the LLM's input embedding space. Using only 16 memory tokens, Memorilla achieves the best performance against baselines including Mem0 and MemoRAG on five of eight benchmarks spanning personalization, long-document and knowledge-intensive question answering. Notably, on FactKG, it reaches 84.33% accuracy versus 80.43% with full-context inference, while using 108x fewer tokens to represent the source documents. Furthermore, Memorilla can facilitate continual learning in agentic systems where the document count (the agent's past turns) increases throughout the trajectory. In a TextWorld environment, we show that Memorilla reduces the agent's tendency to wander between rooms it has already explored, and improves its average success rate from 18.6% with full-history in-context learning to 24.2%. Overall, Memorilla aids frozen LLMs in accumulating and using knowledge effectively across long-context interactions and takes a step towards building continually learning AI systems.
memorylong-contextsemantic memoryagenticbenchmark - arxiv:2610.21374 · cs.ROFrom Optical Flow to Safety: Visual Safety Certification via Scale-Free Relative MotionBazeela Banday, Jishnu Keshavan
Safe autonomous navigation using monocular vision remains challenging because explicit obstacle states are not directly observable. This letter introduces a visual safety-certification paradigm that uses scale-free, depth-normalized relative motion encoded by monocular optical flow as the representation for collision-risk certification, without reconstructing obstacle position, velocity, geometry, or a metric world model. Unlike conventional optical-flow-based navigation approaches that rely on image motion to shape avoidance behavior, the proposed framework directly encodes collision-cone geometry and safety constraints in this relative-motion representation. An online Concurrent Learning observer estimates the normalized relative velocity and provides a computable estimation-error bound. This uncertainty is propagated into the collision cone through angular inflation and a monocular robot-footprint margin, yielding a Robust Vision-Based Collision Cone Control Barrier Function (RV-C$^3$BF). An optimal-decay CBF quadratic program minimally modifies the nominal control input to enforce the resulting robust safety constraint under the stated assumptions. Extensive simulation and real-world experiments on a TurtleBot3 in cluttered environments with static and dynamic obstacles demonstrate collision avoidance using only monocular vision and onboard proprioception, with the estimation-error bound containing the measured error in $98.3\%$ of simulation samples and the robot achieving collision-free navigation in $93$ of $100$ real-world trials.
world model - arxiv:2610.21370 · cs.LGSame Target, Different Origins: Aligning Latent Rollout for Consistent PDE EmulationXinying Li, Jing Xu, Wenjun Wang, Limei Han +3
Neural PDE solvers commonly generate multi-step predictions through autoregressive rollout, recursively applying a learned evolution operator and feeding each prediction back as input for the next step. Approximation errors can therefore propagate through the rollout, causing predictions at a shared target time to depend on their temporal origins and rollout depths. For a well-posed deterministic PDE flow, however, states at different times along the same physical trajectory should, under exact evolution, reach the same target state. To this end, we propose Align Latent Rollout (ALR), a regularization approach that promotes consistency across rollout origins through the neural PDE solver's internal representations. A second-order Taylor analysis of latent discrepancy propagation motivates two objectives: zeroth-order latent state consistency, which aligns representations at a shared target time, and first-order latent dynamics consistency, which aligns local evolution responses to shared perturbations. ALR can be integrated into different neural PDE solvers by regularizing suitable internal representations, without modifying their backbone architectures. Experiments across multiple PDE benchmarks and neural-operator backbones show that ALR improves prediction accuracy while reducing shared-target discrepancies across rollout origins.
latent dynamicsbenchmark - arxiv:2610.21367 · cs.CVSATOR: Backpropagation-Free 3D Test-Time Adaptation via Sample-Adaptive Token RoutingInseok Kong, Geunyoung Jung, Jiyoung Jung
Test-time adaptation (TTA) mitigates the performance degradation of source-pretrained models on unlabeled test data under distribution shifts. Recent backpropagation-free token-purging methods address 3D point cloud distribution shifts by filtering unreliable tokens. However, their reliance on predefined purge-size candidates requires multiple forward passes per sample and increases inference cost. We propose SATOR (Sample-Adaptive TOken Routing), a source-free and backpropagation-free 3D TTA framework that eliminates this search by deriving both the partitioning threshold and routing decision directly from each sample. SATOR computes token abnormality scores from complementary attention and geometric relation graphs. It then derives a sample-specific partition into retained and pruned tokens via Otsu-based separability, without evaluating candidate sizes. Keep-ratio-calibrated routing then selects the pruned prediction only when its entropy, penalized by the keep ratio, is lower than that of the original prediction. Experiments on ModelNet40-C, ScanObjectNN-C, and ShapeNet-C demonstrate that SATOR achieves state-of-the-art accuracy on three 3D corruption benchmarks.
benchmark - arxiv:2610.21366 · cs.CLRenoLayout: Unified Floorplan Renovation and Furniture Layout for Constraint-Aware Residential DesignJingguo Liu, Yuanqing Liu, Chong Zhang, Yue Yang
Residential design aims to transform existing homes into functional living spaces while satisfying diverse design preferences and physical constraints. However, existing floorplan generation methods typically assume empty building boundaries as input, whereas renovation requires reasoning over existing floorplans. Furthermore, floorplan renovation and furniture layout are commonly addressed as separate tasks, overlooking their inherent spatial dependency. Consequently, practical residential design must preserve immutable structural components while transforming editable spatial elements. It must also optimize furniture placement under design objectives and physical constraints. To address these challenges, we formulate residential design as a design-point-guided conditional generation task and propose RenoLayout, a framework that translates a pre-design residential state together with a design-point set into an executable Geometry-as-Code program. To enable controllable supervision, we construct a Design Map comprising 123 fine-grained design points covering structural renovation and furniture layout. We further introduce Difference-Grounded Design-Point Annotation to derive design-point sets automatically from before--after design pairs with VLM-based consistency verification. RenoLayout is trained using a three-stage CPT--SFT--RL pipeline, progressively improving spatial reasoning, instruction following, and constraint compliance. We also introduce the RenoLayout Benchmark, a set of 200 real-world before--after renovation cases with complete furniture layouts. To the best of our knowledge, it is the first benchmark for unified residential design. These cases include design-critical structural and building-service constraints.
benchmark - arxiv:2610.21362 · cs.ROLTM: Latent Transition Based In-Context Memory for Vision-Language-Action ModelsQixin Sun, WenLong Gao, Xiansheng Chen, Zhongyuan Wang +2
Robotic manipulation often requires remembering interactions that are no longer visible. We introduce Latent Transition Memory (LTM), which stores history as compact latent transitions between consecutive observations. Instantiated on $π_{0.5}$ as LTM-$π$, our method uses a frozen latent action encoder, a lightweight adapter, and causal KV caching for incremental inference without memory-specific semantic annotations or external reasoning-model calls. With DreamDojo transitions, LTM-$π$ achieves 58.54% success on RoboMME and 80.5% on LIBERO-Mem, performing best among baselines without external reasoning models or memory-specific annotations. Ablations support the contribution of transition content beyond temporal cues and the efficiency of delayed memory admission. Scaling RoboMME demonstrations tenfold improves success from 57.67% to 87.50%, while a Cup Swap diagnostic highlights the role of process supervision in temporal composition. Under a 200-token memory budget, success remains 84.00%, versus 87.50% without compression. Finally, real-robot counting success improves from 11.25% to 82.50%. Together, these results demonstrate the effectiveness, scalability, and efficiency of latent transition memory, highlighting its capacity for complex temporal information, bounded-context operation, and real-world robotic control.
vision-language-actionmanipulationliberomemory - arxiv:2610.21351 · cs.LGHonesty is the Best Policy ... or is it? Demand Management for Probabilistic Service OptionsDaniela Sailer, Robert Klein, Claudius Steinhardt
To enable well-informed customer choices and redistribute demand across service options, providers can display probabilities regarding predicted service levels at the time of booking, effectively offering probabilistic service options. In many service settings, the underlying truthful probabilities endogenously depend on the system state, rather than being static product attributes controlled by the provider. While customers rely on this information when making decisions, they cannot readily verify its accuracy. Consequently, providers may be tempted to manipulate the displayed probabilities to guide customers' choices in their favor, creating a novel lever for demand management. In this work, we model the resulting problem as a sequential decision process, define truthful probabilities recursively, and introduce transparency constraints that limit their distortion. Because the exact computation of truthful probabilities is intractable for realistic settings, we develop an adapted reinforcement learning algorithm to approximate them. Furthermore, we propose a sampling-based mixed-integer program to optimize the information communicated by the provider in each state. Our framework enables providers to gauge the benefits of dynamically disclosing probabilistic information at the time of booking and sheds light on incentives for manipulation. Our computational study is motivated by out-of-home delivery, i.e., the delivery to parcel lockers or shops, and reveals a trade-off between individual-level transparency and aggregate customer utilities.
manipulation - arxiv:2610.21349 · cs.ROCQDC-Warp: GPU-Accelerated Differentiable Quasi-Dynamic Simulation for Contact-Rich ManipulationYanchen Ding, Jiayi Chen, Jianfeng Xu, Wei Zhao +2
Efficient synthesis of contact-rich manipulation trajectories is essential for downstream robot learning. Gradient-based planning offers a promising approach, but jointly optimizing motion and contact decisions remains challenging. Meanwhile, many existing planners run on CPUs with limited throughput. We present CQDC-Warp, a GPU-accelerated differentiable quasi-dynamic simulator for batched model predictive control (MPC). We use it to study contact-guided trajectory optimization, in which intended robot--object contacts are specified before planning and encouraged through soft costs, while the simulated dynamics determine contact geometry and frictional response without a prescribed contact-mode schedule. Across $17{,}408$ trials on the $17$-object Allegro in-hand rotation benchmark, our framework achieves $79.8\%$ success, compared with $73.7\%$ for CFree and $65.2\%$ for IMPACT, and reduces mean trajectory-generation time by factors of $27.0$ and $50.2$, respectively. On a single NVIDIA RTX 4090 with $2{,}048$ parallel worlds, forward and backward throughput reach $2.06\times10^5$ and $2.25\times10^5$ world-steps per second, respectively. Further experiments cover manipulation with LEAP and Shadow hands, environmental contact exploitation, bimanual manipulation, whole-body control, and real-world trajectory replay on a Sharpa hand.
manipulationwhole-body controlbenchmark - arxiv:2610.21346 · cs.AIARGUS: An Agenda-Driven Framework for Reliable User SimulationHang Zheng, Nan Jiang, Zichen Zhu, Zixu Yang +5
High-fidelity user simulators are essential for evaluating and training agents in interactive environments. However, current LLM-based simulators struggle with reliability issues, such as goal omission and policy non-compliance. We propose ARGUS (Agenda-dRiven Goal-oriented User Simulator), a neuro-symbolic framework that maintains the user's goal state using a dynamic agenda stack manipulated via symbolic operations. By rigorously decoupling state tracking from natural language generation, ARGUS facilitates strict goal progression and constraint-aligned interactions. For cost-efficient deployment, we distill ARGUS into compact models using a unified pipeline: evaluator-guided SFT followed by online RL with dense step-level rewards. To rigorously evaluate behavioral fidelity, we introduce a rubric-based metric Composite Behavioral Score (CBS). Evaluations on $τ^3$-Bench and MultiWOZ demonstrate the significant superiority of the ARGUS architecture and the efficacy of our training pipeline. Code and data will be released soon.
evaluator - arxiv:2610.21345 · cs.CLOpCredit: Operation-Aware Policy Learning for Agentic Knowledge Graph Question AnsweringYucheng Li, Mingyuan Li, Xiaxia Wang, Gong Cheng
Agentic knowledge graph question answering(KGQA) relies on sequences of executable operations, but trajectory-level rewards obscure their individual contributions. Successful trajectories may contain ineffective operations, while failed trajectories may contain useful ones. To assess individual operations, we propose OpCredit, an operationaware policy-learning framework that combines downstream outcome credit with immediate transition credit: Branch Outcome Credit (BOC) reuses ordinary rollouts to compare terminal task rewards observed after different operations under matched navigation contexts; Frontier Transition Credit (FTC) measures operation-induced changes in answer-conditioned frontier potential. The two complementary signals jointly supervise the corresponding decision spans during training, while inference uses only the learned policy and standard KG interface. Experiments on ComplexWebQuestions, WebQSP, and GrailQA across three backbones demonstrate consistent improvements in KGQA performance.
knowledge graphagentic - arxiv:2610.21344 · cs.AIQuantum Versus Classical Machine Learning for Neuropsychiatric Disorders: A Feature-Matched Multi-Cohort EvaluationUtsav Poudel, Anuj Nepal, Vyas Viswanath, Subramaniyaswamy Vairavasundaram
Resting-state functional connectivity is widely used in machine learning-based classification of neuropsychiatric disorders. Quantum machine learning models have been proposed for functional connectivity data, but few studies have evaluated them under leave-site-out validation or against matched classical models. We compared six quantum model families with classical ones using identical correlation-derived features across four public cohorts (4,308 subjects). Leave-site-out validation was used for the Autism Brain Imaging Data Exchange (ABIDE I), ADHD-200 (attention-deficit/hyperactivity disorder), and REST-meta-MDD (major depressive disorder) cohorts, while the UCLA Consortium for Neuropsychiatric Phenomics (UCLA-CNP) was evaluated using stratified 10-fold cross-validation. The quantum kernel was statistically indistinguishable from a matched classical kernel across all four cohorts, with the average accuracy difference within 1.4 percentage points at 95% confidence. On ABIDE I, a hybrid quantum convolutional network achieved similar accuracy to its classical counterpart (0.645 vs. 0.641), while a trainable quantum kernel exceeded its classical counterpart by 0.064 in area under the curve (AUC). Quantum convolutional networks also achieved higher AUC than their matched classical networks on this cohort and showed smaller train-test gaps on all four cohorts, without an accuracy gain. Overall, most quantum models matched their classical counterparts, but none consistently outperformed them. This study, among the first quantum evaluations of ABIDE I under leave-site-out validation, provides a feature-matched benchmark of quantum and classical models comprising 284 matched comparisons. We further analyse fidelity-based graph construction and account for the observed quantum-classical parity through effective-dimension and eigenspectrum analyses.
benchmark - arxiv:2610.21340 · cs.LGDimension-Free PAC-Bayes Generalization for Deep Neural Networks via Dirichlet ProcessesSumit Vashishtha
Classical PAC-Bayesian generalization theory places prior distributions over Euclidean parameter space $Θ\subseteq \mathbb{R}^D$, causing relative entropy to scale with network dimension $D$. For modern overparameterized neural networks where $D \gg n$, this parameter-counting dependency yields vacuous generalization bounds ($> 1{,}000\%$). In this work, we resolve this barrier by shifting prior specification from internal parameters to the non-parametric space of data distributions $\mathcal{M}_1(\mathcal{Z})$ via a \textbf{Dirichlet Process (DP)}, $\mathcal{DP}(α, F_0)$, inducing parameter distributions through push-forward measures under the learning algorithm. By the Data Processing Inequality, parameter-space relative entropy is strictly bounded by the data-space DP divergence, which we evaluate in exact closed form using Digamma functions. The resulting PAC-Bayesian certificate is strictly invariant to network parameter count $D$, width, and depth. On intrinsic $d^*$-dimensional submanifolds, optimal partition scaling achieves the minimax non-parametric rate $\mathcal{O}(n^{-1/(d^*+2)} \sqrt{\ln n})$. Across vision architectures scaling up to 4.2 million parameters ($D = 7{,}850$ to $D = 4{,}203{,}522$) on the full MNIST benchmark, while classical weight-space bounds scale past $940\%$ (becoming vacuous), our data-space DP-PAC-Bayes certificate remains stable, non-vacuous, and certified at \textbf{$8.8\%$--$9.4\%$}.
benchmark - arxiv:2610.21333 · cs.CLQuasiLeak-Bench: Evaluating LLM-Based Privacy Detectors on Policy-Conditioned Quasi-Evidence Group RecoveryYufan Zhu, Yangfan Jiang, Yang Fei, Shizheng Hou +2
Natural language messages often contain quasi-identifiers that are weak in isolation but can jointly reveal an individual's private attributes. We call such a jointly revealing set of cues a quasi-evidence group. Existing privacy benchmarks focus on explicit identifiers, attribute inference, or task-driven disclosures, leaving inference from quasi-evidence groups largely unmeasured. In practice, whether exposing such a quasi-evidence group violates privacy also depends on sender-specified disclosure policies. We introduce QuasiLeak-Bench, a benchmark for policy-conditioned quasi-evidence group recovery, which evaluates whether detectors recover complete quasi-evidence groups and flag the disclosures the given policy prohibits. To obtain reliable ground truth at scale, a reverse-construction pipeline first selects weak quasi-identifiers, composes them into quasi-evidence groups verified to jointly reveal a target attribute, and generates documents around this evidence, followed by automated quality control. Experiments show that existing privacy detectors transfer poorly to QuasiLeak-Bench, and locally deployable LLM-based detectors fall far behind proprietary models, mainly because they miss most quasi-evidence groups altogether. However, a 4B model fine-tuned on our training split surpasses the strongest prompted model on held-out documents, showing that the task is learnable.
benchmark - arxiv:2610.21330 · cs.CLVoteGEO-Bench: Measuring the Impact of Generative Engine Optimization on Electoral VisibilityShiqi Zhang, Xiaokui Xiao
Large language models (LLMs) are increasingly used for election information, yet little is known about how external web content shapes the political information they present. Generative engine optimization (GEO) rewrites web content to improve its visibility in generated responses. To investigate its potential risks in voting scenarios (VoteGEO), we introduce VoteGEO-Bench, a benchmark covering the 2024 U.S. presidential election and the 2026 Japanese House of Representatives election, combining election-related queries, historical webpages, and voter personas. We first analyze sources cited by generative engines and find that many citations come from lower-cost owned and open-platform sources, while some webpages remain visible across days. We then evaluate 11 rewriting strategies in 4 categories, examining target specification, effectiveness across engines and elections, and the role of voter personas. Rewriting without specifying a target rarely increases target exposure, whereas target-specified rewriting more often does. Strategies that improve page organization or strengthen a position through argumentation yield greater gains than adding supporting evidence or changing wording. For the two strongest strategies, gains remain broadly positive across evaluated U.S. personas, but persona conditioning shows no consistent effect on these gains. Explicit partisan identities can create baseline exposure gaps larger than the gain from rewriting one webpage. When partisanship is implicit in the voter profile, baseline shifts depend on how strongly the engine responds to that signal.
benchmark - arxiv:2610.21327 · cs.CVDocPilot: Interactive Agent Orchestration and Training for Long Visual Document UnderstandingZhuowen Liang, Xiaotian Lin, Yuyu Luo, Haixun Wang +1
Multimodal long-document question answering requires models to seek sparse evidence across pages and modalities, understand heterogeneous document content, and reason over the collected evidence. Even advanced vision-language models (VLMs) can overlook sparse evidence and struggle with cross-modal reasoning in a single pass. Existing document agents further introduce multi-step interaction, but some still rely on predefined workflows or coarse-grained actions, limiting the flexibility and granularity of document reasoning. We introduce DocPilot, an agentic document framework that supports expressive orchestration and makes complex document solving learnable by small language models (SLMs). DocPilot orchestrates evidence seeking through page-level navigation and fine-grained evidence localization, and integrates local visual inspection, code-based numerical reasoning, and cross-modal reasoning for multimodal understanding. To transfer these capabilities to small models, we propose curriculum-based agentic training that progresses from single-ability learning to adaptive orchestration through supervision from data-grounded reasoning traces and executable interaction trajectories. Across MMLongBench-Doc, DocBench, FinRAGBench-V, and MADQA, DocPilot achieves strong gains with frontier VLM backbones, while DocPilot-4B delivers competitive performance against substantially larger models, demonstrating strong effectiveness and generalization across diverse long-document reasoning settings. The code is available at https://github.com/HKUSTDial/DocPilot.
agentagentic - arxiv:2610.21325 · cs.CVCertEx: Exact Visual Retrieval at Scale by Certified ExclusionDongfang Zhao
Exact $k$-nearest-neighbor (kNN) search over image descriptors and embeddings underlies visual retrieval, near-duplicate detection, the evaluation of learned representations, and increasingly the memory of LLM agents. However, exhaustive search computes every distance at full precision, whereas approximate indexes return recall below one. We observe that exact search only needs a proof that each excluded vector cannot be a neighbor. Because such a proof only has to be conservative, it can be checked in low precision, much as the retina examines the whole scene at low resolution. We call this principle certified exclusion and realize it in an exact kNN method for GPUs, which we call \certex. Specifically, \certex\ stores a lower-bounding projection of each vector in half precision and folds a rounding certificate into the same tensor-core dot product, so that only the few vectors that the test keeps need exact distances. We prove that \certex\ returns the exact kNN of every query. On sixteen datasets with up to 100M vectors, \certex\ is exact for every query at 7.1$\times$ to 18.0$\times$ the throughput of the fastest cuVS search on one H100 and outperforms quantized and graph indexes whenever every query must be exact.
memoryllm agent - arxiv:2610.21320 · cs.CVZoomCap: Evidence-Routed Visual Zooming for Recoverable Visual Information TransferHao Shi, Jiafeng Wang, Yuzhuo Fu, Tao Zhang +4
Detailed captions transfer visual information into vision-language model supervision, but quality is easily conflated with length: long single-pass outputs still miss small objects, text, and spatial relations or promote uncertain observations into confident claims. We formulate fine-grained captioning as recoverable visual information transfer, in which a teacher acquires missing evidence under a finite observation budget and converts it into reusable supervision. ZoomCap uses a draft caption to route conditional knowledge retrieval, OCR-sensitive reading, coverage-gap recovery, and adaptive visual zooming, allocating deeper crops only to the most complex eligible region, deriving spatial relations from grounded geometry, and fusing non-duplicate evidence. We also introduce FactCycle, which scores caption-derived atomic claims on two blind anchors--source image (truthfulness) and reconstruction (realizability)--separating fabrication from generator loss claim by claim; human and multi-judge comparisons complement this fact-level view. Against Qwen3.5-397B-A17B captions, ZoomCap achieves an 81.9% human tie-adjusted preference score with a teacher over 10x smaller, and FactCycle counts 52.4 transferred claims per image vs. 31.5 for ScaleCap. The evidence also transfers: in continual pre-training, ZoomCap-450K improves the 11-benchmark average by 1.31-2.38 points over the strongest competing corpus and Ling-Tiny-VL-v3 gains 1.4 points on a production document/OCR monitor suite (1.5 on perception) at 60B tokens; in medical pre-training, 2.64M ZoomCap-captioned images improve five VQA benchmarks by 3.04 points (vs. 1.76 for same-image Gemini-3.1-Pro captions).
benchmark - arxiv:2610.21319 · cs.AIConstrained Multi-Agent Reinforcement Learning in ImaginationRaz Shmueli, Shahaf S. Shperberg
Predicting cost in a multi-agent world model moves most constrained policy updates into imagination: the stronger model-free baseline needs $10$ to $20$ times CA-MAMBA's interactions on the median SMAC configuration, and CA-MAMBA earns $2.8$ to $12$ times the baselines' MAMuJoCo reward at equal budget. In controlled tests on one task per benchmark, pricing the model's cost advantages at parity more than halves a cost-blind learner's cost and meets a tight continuous-control limit the learned multiplier misses, suggesting the model's cost estimates are not the bottleneck.
world modelmulti-agentbenchmark - arxiv:2610.21318 · cs.CVLoud and Clear: Making Visual Evidence Count in Multimodal Document RAGEunchan Jo, Seounghyun Jang, Sangwoo Mo
Multimodal document retrieval-augmented generation (RAG) has focused largely on retrieval, yet its answers depend on how well a vision-language model reads the retrieved pages. We find that evidence in retrieved pages remains underused even when the correct pages are retrieved. On SciEGQA oracle pages, five prior methods give the generator 0.7K to 2.7K visual tokens per query on average under a common image processor, yet raising the page-only budget from 1K to 32K improves accuracy by 15.22 points. At a fixed 16K budget, adding ground-truth crops improves accuracy by 3.75 points over page-only input. We study what the generator needs from retrieved pages through two requirements: Loud, making full use of visual token capacity, and Clear, refining the page so that its evidence becomes prominent. For Loud, increasing either the visual token budget or visual fidelity, the image resolution those tokens preserve, improves accuracy with the other held fixed, and full-page scaling outperforms repetition and tiling at matched budgets. For Clear, suppressing surrounding content improves accuracy without enlarging selected regions, and adding OCR of these regions to page images outperforms either input alone. We implement both as Loud and Clear (LoC), which combines higher-resolution pages with OCR from query-relevant regions without ground-truth annotations or additional training. LoC outperforms eight prior methods on SciEGQA, ViDoRe-v3, and MMLongBench-Doc, improving over the strongest, ColQwen3, by 8.88, 3.09, and 4.22 points with the same retriever and generator. It also outperforms its page-only variant under the same visual token budget. These results locate a bottleneck beyond retrieval: how the retrieved pages are given to the generator.
retrieval-augmentedrag - arxiv:2610.21312 · cs.LGSpectrum Is Not Enough: A Random-Matrix Account of Data-Free Preconditioning for Differentially Private LearningShilin Zhang, Kangning Wang
Data-free preconditioning builds the Kronecker factors of KFAC from synthetic probes instead of private data. DP-KFC validates a probe by matching the eigenvalue decay of the private factors; we turn this criterion into a random-matrix account of what a probe can reconstruct. A single dichotomy organizes it: in the factorization $ A \otimes G $, the error-signal factor $ G $ and the architecture-determined bulk of the activation factor $ A $ are reconstructible without data, while only the data-induced spikes of $ A $ are not. We prove the $ G $ half is data-insensitive up to an explicit error-energy bound, with zero violations over a full training trajectory across four modalities and a natural-image benchmark. We locate the failure boundary at a Baik-Ben Arous-Péché threshold: a bulk-only probe suffers an exact damage $s^2/(1-s^2)$ from a data-induced spike, which becomes invisible once it exceeds $(d/n)^{1/4}$, confirmed on an injected-spike surrogate. Holding the probe spectrum fixed and varying only its eigenbasis, we show that spectral-shape matching is neither sufficient nor the operative object: a probe with a near-exact matched spectrum (log-ratio dispersion $0.02$) still inflates the residual $2 \times $-$90 \times $ when its eigenvectors are random. A closed-form whitened-volume law explains this and predicts the within-spectrum ordering of the critical volume, though not its cross-dataset magnitude. Downstream, per-layer tuning does not transfer to accuracy and the headroom is a small oracle-minus-white gap. Under a privacy mechanism the source ordering survives while learning does, then inverts at a noise cross-over where the flat probe overtakes the matched ones.
benchmark - arxiv:2610.21308 · cs.LGPocketTSLM: Deploying an On-Device Soft-Prompt Time-Series Language Model for Privacy-Preserving Smartphone ReasoningMax Rosenblattl, Nikhil Krishnaswamy, Leon Nissen, Manuel Milling +8
Consumer phones and wearables increasingly collect physiological time-series data, yet most mobile systems treat these streams as summaries or text rather than as temporal evidence. This representation is poorly matched to personal health systems that need to retrieve, route, or summarize sensor evidence while preserving temporal structure. Cloud-hosted time-series language models avoid smartphone limits, but they repeatedly move sensitive physiology off the device and add latency and per-query compute cost. Mobile model alternatives that aggregate time-series data into statistical features or serialize raw samples as text either compress temporal structure or exhaust the context and memory budgets of small on-device models. We present PocketTSLM, a mobile-native deployment of OpenTSLM SoftPrompt, the soft-prompt variant of the OpenTSLM time-series language model. PocketTSLM runs fully on a consumer smartphone by injecting encoded time-series windows as embeddings instead of tokenizing time series as text. A numerical parity protocol verifies identical greedy-decoded token sequences between the Swift port and a Python reference using Apple's MLX machine-learning framework. Answers also match the PyTorch reference, although rationale wording can differ. An on-device context-economics benchmark compares the soft-prompt path with a text-serialized baseline on the same backbone. In our benchmark format, soft-prompting uses approximately 12 times fewer time-series tokens and reduces prompt-prefill cost. PocketTSLM processes 30-second Sleep-EDF electroencephalography clips for sleep staging and a 1,000-sample (10 s) 12-lead electrocardiogram clip from ECG-QA. In comparison, text-serialized baselines exceed the phone's context or memory budget at longer time-series windows. All inference runs entirely on the device, with no health data transmitted during inference.
memorybenchmark - arxiv:2610.21306 · cs.CVVLA-$Δ$: Learning Physical State Transition for Vision-Language-Action ModelsQi Lu, Jiarun Li, Ce Zhang, Miao Pan +4
Vision-language-action (VLA) models transfer semantic knowledge acquired through vision-language pretraining to robot control, achieving strong performance across various benchmarks. However, most VLAs directly map observations to demonstrated actions without explicit supervision of intermediate physical state changes, making them prone to overfitting observation-action correspondences. Some methods introduce intermediate reasoning or future-state prediction in language, visual, or latent spaces to guide action generation, but do not explicitly supervise the physical state transition itself. We therefore propose VLA-$Δ$, a general framework that uses physical changes between current and future states as explicit learning targets. To support this objective, we construct a large-scale pretraining dataset and derive physical state-transition annotations covering task-relevant semantics and visual motion changes, providing more direct supervision for continuous action generation. During inference, the model generates actions directly from current observations without additional inference overhead. VLA-$Δ$ achieves performance competitive with the state of the art on LIBERO and RoboTwin, and outperforms $π_{0.5}$ in real-robot experiments. Our code is available at https://github.com/ZJU-OmniAI/VLA-Delta
vision-language-actionliberorobotwinbenchmark - arxiv:2610.21304 · cs.CLVALCON: Benchmarking and Enhancing Value Consistency in Large Language Models via Dual-Role VariationJunru Wu, Jingting Zheng, Yuqi Ren, Deyi Xiong
Current value alignment evaluations for large language models (LLMs) heavily rely on output compliance, overlooking whether LLMs maintain value consistency when alternating between generator and evaluator roles. Such consistency is increasingly important because modern LLM agents and alignment pipelines often rely on the same LLM to both generate actions and evaluate their acceptability. To mitigate this limitation, we propose VALCON, a role-driven data construction framework that operationalizes value consistency into three benchmark suites (VAB-Consistency, VAB-Boundary, VAB-Challenge) via dual-role variation. By systematically pairing an LLM's generative behavior with its evaluative judgments under identical principles, VALCON synthesizes target evaluation data that exposes hidden value discrepancies. Experiments on 28 LLMs reveal pervasive cross-role inconsistency: models with highly compliant outputs frequently fail to align their judgments with their own generations. To understand and address this gap, we perform mechanistic probing that uncovers a role-dependent representational decoupling associated with this behavioral asymmetry. Guided by these insights, we propose a representation-directed intervention that effectively mitigates consistency failures. Our work establishes value consistency as a distinct alignment dimension and provides a complete evaluation-analysis-mitigation pipeline. Code and data are available at https://github.com/tjunlp-lab/VALCON.
llm agentbenchmarkevaluator - arxiv:2610.21292 · cs.ROMR-MPCC: Model Predictive Contouring Control for Robotic Manipulator Tracking under Moving Reference FramesChonghao Cheng, Tiancheng Li, Jonathan Woolfrey, Liang Zhao +1
Maintaining small path-tracking errors under moving reference frames is essential for robotic manipulation tasks with tight execution tolerances, such as machining and surgical automation. Conventional time-indexed controllers, including tracking Model Predictive Control (MPC), rely on a fixed temporal correspondence with the reference trajectory and may therefore accumulate path-relative error under abrupt reference-frame motion or timing mismatch. This paper presents MR-MPCC, a Model Predictive Contouring Control (MPCC) framework for velocity-controlled robotic manipulators with 6-DoF end-effector pose regulation under Moving-Reference disturbances. A hierarchical control architecture is adopted, in which an upper-level MPCC optimizes path progress while penalizing contouring and orientation errors, and a constrained resolved motion rate control (RMRC) layer maps the commanded Cartesian twist to feasible joint velocities under joint-level and safety constraints. MR-MPCC is implemented in a ROS2--MuJoCo pipeline and validated on a UR5e platform under moving-reference disturbances, where the path frame is perturbed and tracked by electromagnetic sensors. Compared with PI control and time-indexed tracking MPC under matched disturbance, MR-MPCC achieves lower geometric path-following error and smaller deviations in the moving path frame, demonstrating improved path-following performance under reference-frame motion.
manipulationmanipulator - arxiv:2610.21290 · cs.AIBias Control and Injection for Multi-Agent AggregationSungjae Lee, Soheil Abbasloo, Jabin Koo, Jungseul Ok
Multi-agent reasoning with large language models (LLMs) raises a central question for collective decision-making: when agents disagree, whose decision should be trusted? Resolving such disagreements is difficult because agents' decisions are susceptible to non-semantic biases such as majority pressure and response-style preferences, which can obscure the quality of the underlying reasoning. Our analysis shows how bias can be both mitigated and exploited: bias control improves initial answer selection, while bias injection elicits decision instability that serves as a reliability signal. Building on these findings, we introduce a test-time aggregation framework that requires neither validation data nor prior knowledge of relative model strength. The framework implements bias control by presenting one style-normalized reasoning trace per distinct answer, and bias injection by placing each aggregating agent's initially selected answer in the minority. Overall, our framework achieves higher average accuracy and better token efficiency than competing non-oracle aggregation methods across three mathematical reasoning benchmarks and four heterogeneous LLM combinations. It also improves final-answer accuracy when applied to multi-agent debate and Mixture-of-Agents, without modifying their interaction or synthesis processes.
multi-agentbenchmark - arxiv:2610.21287 · cs.CVReState: Noise-Level-Aware State Correction for Training-Free Object Removal in Few-Step Diffusion ModelsDonghwan Lee, Sanghyun Jo, Sihyun Kim, Ziseok Lee +1
Training-free object removal commonly redirects self-attention away from the target region, but on few-step distilled diffusion models this often leaves object-shaped residuals. We identify a temporal mismatch: a structured prediction forms inside the mask at the first forward, and the short trajectory offers few later chances to revise it, while attention rerouting acts mainly in later steps. Across four distilled backbones, magnitude-matched single-forward interventions show that background-directed state correction affects the final result more than attention rerouting. We therefore propose ReState, which corrects the masked clean prediction toward background evidence retrieved from the input using a noise-level-aware weight that vanishes at the final forward. ReState retains object-excluding attention as a complementary constraint and adds no denoiser evaluations. Across two paired benchmarks and two few-step backbones, ReState achieves the best masked PSNR, SSIM, and LPIPS among training-free methods, improving PSNR by 3.17 dB on average over the strongest baseline. On OBER-Test, it also outperforms the 50-step attention-forcing reference on all three metrics at 6.7x lower latency. The same configuration transfers to two additional distilled backbones without retuning. Project page: https://tdrlee.github.io/ReState
benchmark - arxiv:2610.21283 · cs.CVMallea: Building, Steering, and Memorizing Worlds with Coding AgentXuanyi Liu, Ruiqi Li, Siwei Ma
Code world models represent world dynamics through explicit state and executable rules. We investigate how to coordinate object representations, behavioral revisions, and visual memory within this formulation. We present Mallea, a system for building, steering, and memorizing worlds from text. A coding agent infers object interaction requirements from the prompt and jointly plans object structure, properties, and rules, favoring compact representations that support those requirements. The world evolves under the planned rules; optional user actions steer it, while text requests can edit properties or extend capabilities with unaffected state preserved. Object texture memory retains reusable appearance only for whitebox-defined objects and adapts it as those objects change. The program advances world state, while a generative renderer combines current structure with remembered appearance. By organizing construction, intervention, and memory around the same objects, Mallea conditions subsequent observations on the latest state and compatible appearance. Project page: https://lxycopper.github.io/Mallea.
world modelmemoryagent - arxiv:2610.21282 · cs.ROLICOA: Learning Instruction-Conditioned Optical Actions for Mobile Spatial Augmented RealityYuchen Deng, Shunkai Zhou, Ruixuan Zhu, Haibin Ling +1
Spatial augmented reality (SAR) augments physical environments with projected content, but most systems assume that the projector pose, target region, or motion path has been fixed or specified. We instead make projected light an embodied action: given a language instruction and current visual observations, a system must decide how to move and where and at what scale to place the projection. We introduce LICOA to realize this capability. Its differential-drive chassis provides room-scale approach and repositioning, while a steerable mirror with two degrees of freedom (2DoF) and projection scale provide rapid local control of the beam. A shared interface maps synchronized external and onboard RGB cameras with robot control state to physical optical actions across pretrained action models. Because instruction-conditioned optical action lacks dedicated training data and a common evaluation protocol, we construct a matched simulation embodiment and a corpus with 150 tasks and 7,500 rendered episodes, then evaluate models trained only on this corpus in previously unseen real-world environments. Across 480 complete closed-loop real-world execution trials (rollouts) on 12 tasks, the best adapted action model achieves a 57.50% success rate and 40.64% success weighted by path length, compared with 8.33% and 5.15% for the same architecture trained from scratch. A further 150 real-world ablation rollouts support the need for both camera views and coordinated control of the chassis and mirror. These results establish a physical platform, shared perception and optical action interface, corpus, and real-world evaluation protocol for treating projected light as an embodied action.
embodiedevaluation protocol - arxiv:2610.21281 · cs.AIDiviner: Agentic RCA with Human-Governed Adaptation for Large-Scale InfrastructurePradipta Ghosh, Gaurav Mitra, Shripad Nadgowda, Akshil Shah +6
Root-cause analysis (RCA) in large-scale distributed systems is challenging because a single symptom can cross service, data, configuration, compute, and organizational boundaries. Responders must assemble incomplete evidence and coordinate experts who hold different parts of the causal path. Generic language-model agents do not solve this problem through tool access alone: they can use stale or mistimed context, interpret failed queries as absent symptoms, report unsupported system state, and overload observability services designed for human-paced use. We present Diviner (Domain-Informed Verifiable Incident Navigation through Evidence and Review), an agentic RCA system operating at Meta and organized around a four-stage reliability lifecycle. Encode combines a large corpus of reviewed incidents with service-owner expertise to produce scoped, versioned skills, workflows, typed tools, and an incident knowledge graph. Execute uses an adaptive Lead and context-isolated workers to apply relevant knowledge to point-in-time evidence and maintain evidence-linked hypotheses. Evaluate compares saved investigation trajectories with responder-reviewed outcomes to measure exact cause, useful direction, and Groundedness. Adapt uses recurring gaps and on-call feedback to derive focused operational- knowledge updates, admitting them only after replay, quality checks, and owner approval.
knowledge graphagentic - arxiv:2610.21274 · cs.LGTimeCap: A Scalable Open-Source Pipeline for Captioning Wearable Health Time-Series DataMax Rosenblattl, Milana Gurbanova, Karl Aurel Deck, Narayan Schuetz +15
Time-series data generated continuously across industrial, environmental, and especially health domains produce multimodal signals whose scale and heterogeneity make their processing and interpretation particularly challenging. Time-Series Language Models (TSLMs) promise scalable interpretation of these signals, yet remain bottlenecked by a fundamental lack of text-annotated time series pairs that are signal-grounded, reproducible, and extensible across modalities and schemas. Existing paired corpora are either closed or limited in scale and modality coverage. Annotation with large language models could scale them up, but its outputs cannot be regenerated exactly, which prevents others from re-deriving, extending, or auditing the resulting corpus. To address this, we release TimeCap, an open deterministic captioning pipeline with three reusable extractor tiers and a cross-channel synthesis layer that operates directly on the signal with no language model in the loop, extensible to new schemas by writing channel definitions alone. This positions TimeCap as a reproducible mechanism for future dataset generation rather than a one-off release. Applying TimeCap to OpenMHC, the MyHeart Counts wearable dataset, yields MHCap, the largest open caption layer for multimodal wearable time-series data. MHCap spans 19 Apple HealthKit channels and 7,005,763 captions over 908,944 person-days. To showcase the impact of the corpus, we train OpenTSLM, a time-series language model, on MHCap and benchmark on health-outcome prediction. Caption-augmented training (OpenTSLM-Cap) outperforms a label-only baseline on 29 of 32 per-user binary evaluations across 8 tasks.
benchmark - arxiv:2610.21273 · cs.LGRegret Lower Bound for Online Learning of Quantum States with Logarithmic Loss, with Applications to Quantum State Tomography and Confidence Sequences for Matrix MeansYen-Huan Li, Wei-Fu Tseng
We study the problem of online learning of quantum states with logarithmic loss, which is a quantum generalization of online portfolio selection and has applications to quantum state tomography, the task of estimating an unknown quantum state from measurement outcomes. We first establish a regret lower bound, showing that the regret rates of Exponential Weights and the algorithm of Tseng et al. (2025) have optimal polynomial dependence on the dimension. Our analysis leads to two further results. First, our lower-bound proof leads to a quantum state tomography algorithm that is sample optimal up to polylogarithmic factors in the sample size. Second, we develop an algorithm for constructing confidence sequences for the mean of a matrix-valued random variable. A numerical experiment shows that the latter algorithm is competitive with existing methods.
online learning - arxiv:2610.21270 · cs.AIAuditing Escalation Reliability in Healthcare LLM Agents: A Perturbation-Controlled Benchmark of Workflow-State FailuresZihao Zheng, Junyi Yao, Baichuan Li, Jiayu Long
Trustworthy biomedical AI requires reliable action selection when evidence is missing, stale, conflicting, or late, not only accurate answers on static inputs. We instantiate an episode-level evaluation protocol as 400 synthetic healthcare workflow episodes: 40 hand-authored semantic seeds across four task classes, two surface parameterizations, and five controlled variants, scored on ANSWER/ASK CLARIFICATION/ABSTAIN/ESCALATE under a pre-declared 5:1 missed-versus-unnecessary escalation cost. Reference actions are single-author study-policy labels, not clinician-adjudicated ground truth. Across five open-weight configurations (1.7B--9.7B), clean decision accuracy reaches 0.97, yet uncertainty-prompted models miss 32--97% of required escalations. The stronger 9.7B configuration narrows but does not remove the gap, falling from 0.97 clean accuracy to 0.78 under conflict. Verbalized confidence is inconsistent (ECE 0.10--0.54; failure AUROC 0.48--0.77), and a development-tuned confidence-only policy yields no utility gain. A transparent workflow-aware controller improves 5:1 utility for every model and reduces missed escalations to 0.12--0.24, at specificity 0.83--0.93. However, a 30-concept lexical stress test reduces each rule detector's recall from 1.00 on trigger phrases to 0.00 on unseen paraphrases, limiting the controller claim to this benchmark. Seed-clustered statistics, ablations, cross-task transfer, raw outputs, and manifests are packaged for audit. The results motivate explicit workflow-state evaluation while showing that apparently effective routing controls need their own generalization tests.
llm agentbenchmarkevaluation protocol - arxiv:2610.21267 · cs.AITRACER: Trajectory-Aware Tool Retrieval for LLM AgentsTanish Kher, Sarvesh DK, Abedelkadir Asi, Sparsha Chennarayapatna Sundaresh +1
Modern agents select tools from registries, but exposing entire registries incurs high token costs and can impair selection as they grow. We present TRACER, a retrieval front-end that indexes observed tool-call sequences in a trie and ranks continuations using a learned blend of empirical and contextual priors. On a mixed retail-and-telecom $τ^2$ environment, TRACER improves GPT-5.4's tool-selection accuracy by 4.7 points while reducing input tokens per step by 80%, exposing just 5 of 128 tools. In autonomous full-turn execution, it matches GPT-5.4's full-registry task success with 61% fewer input tokens per step. Across six environments and four models, TRACER improves tool-selection accuracy over full-registry selection in 21 of 24 model-environment pairs, with average gains of 6 to 10 points; the three exceptions trail by at most 2.2. We analyze failures in both regimes to identify directions for future work.
llm agent - arxiv:2610.21266 · cs.CVTRACER: A Multimodal Harness for Patient Care in HospitalsZane Durante, Paul Tang, Keshigeyan Chandrasegaran, Amit Kaushal +10
Continuous video and clinical records provide complementary information for understanding how patient behavior evolves over a hospital stay. A central challenge is how to represent behavior across the entire stay while preserving access to fine-grained visual and clinical evidence. We introduce TRACER, a training-free multimodal harness that constructs minute-level action records from video and combines them with clinical records into hourly reports. The harness equips an LLM agent with tools to query this hierarchy, inspect relevant video, and retrieve clinical records. To evaluate TRACER, we construct two benchmarks. The first contains 397 annotated one-minute clips spanning 15 atomic actions; the second poses 144 clinician-informed multiple-choice questions, grounded in 264 hours of video and clinical records from two ICU patients. TRACER achieves up to 84.8% accuracy for action detection and 61.0% QA accuracy versus 49.3% for the strongest baseline using the same backbone. Across 280 patients spanning approximately 60,000 camera-hours, TRACER identifies significantly less purposeful activity in patients with a positive delirium assessment during their stay than in those without (p < 0.0001). Our results show that combining continuous video with clinical records can support understanding of patient behavior over time and across patient cohorts without training on protected health data.
agentllm agentbenchmark - arxiv:2610.21262 · cs.LGRethinking Learning from Label Proportions via Moment MatchingTianhao Ma, Wei Wang, Yivan Zhang, Dong-Dong Wu +3
Learning from label proportions (LLP) is a weakly supervised setting where training data are grouped into bags with only bag-level label proportions observed, and the goal is to learn a classifier that predicts labels for individual instances. Recently, a large body of methods has emerged; however, most of them rely on the assumption that instance-label pairs are sampled i.i.d. and randomly assembled into bags, an assumption often violated in real-world scenarios. Our experimental results indicate that these methods perform suboptimally under non-random settings. In this work, we adopt a more realistic assumption: bags are sampled i.i.d., and instance labels are conditionally independent given the instances. Under this setting, we show that the label counts follow a Poisson multinomial distribution. Motivated by this observation, we propose LLP via moment matching (LLP-MM), a simple yet effective approach which leverages multi-order factorial moments as training objectives, encouraging the classifier predictions to match these moments and thereby more fully exploit the underlying statistical structure of the data. Extensive experiments on benchmark datasets under various bag construction strategies demonstrate the effectiveness of our approach while maintaining high computational efficiency. Our code is available at https://github.com/TianhaoMa5/LLP-MM.
benchmark - arxiv:2610.21257 · cs.LGFractional Power-of-Two Quantization for Efficient and Effective Multiplier-Free LLM InferenceSunghyun Wee, Geunjae Choi, Hyeonjin Kim, Suyoung Kim +1
Large Language Models (LLMs) demand efficient inference, where the multiply-accumulate (MAC) operations in linear projections dominate compute and energy cost. Post-training quantization (PTQ) reduces this cost by mapping weights and activations to low-bit integers. Power-of-Two (PoT) weight quantization promises efficient multiplier-free LLM inference via shift-add logic, but suffers significant accuracy degradation at low bitwidths due to its coarse base-2 quantization grid. We introduce Fractional PoT (FPoT), a base-$\sqrt{2}$ (half-octave) fractional PoT grid that achieves practical 4-bit weight quantization while preserving a fully integer shift-add datapath. The method combines (i) the base-$\sqrt{2}$ weight grid with uniform integer activations, (ii) a calibration-time grid-alignment refinement absorbed into the static per-channel scale, and (iii) a lightweight 1-adder $\sqrt{2}$ approximation whose error is absorbed during weight calibration via approximate-grid quantization, yielding a processing element (PE) with significantly fewer gates than a 4-bit integer multiplier. On Llama-3-8B with 4-bit weight and activation quantization, FPoT improves zero-shot commonsense reasoning performance by 8.72% relative to the PoT baseline, while maintaining performance close to the uniform integer baseline, with only a negligible 0.4%p degradation. These results hold across various configurations spanning seven models and three bit-widths. Furthermore, FPoT can be integrated with state-of-the-art LLM PTQ methods, demonstrating framework-agnostic applicability.
post-training - arxiv:2610.21256 · cs.LGIsolating a Language Model's Judgment with a Label-Free SubspaceChaemin Jang, Donggyu Lee, Jeesoo Kim, Dongman Lee +1
Language models are often asked to make judgments: whether a proposed answer is correct, whether one statement follows logically from another, or whether a response is safe. Yet changing the instructions that specify how the judgment should be expressed can make the final answer less reliable, even for the same input. We ask whether this reflects a change in internal judgment or a failure to express it. Comparing hidden-state vectors for the same input under different instructions, we find that their differences concentrate in a low-dimensional expression subspace. We identify this subspace without benchmark labels and remove its contribution from a linear predictor to obtain an isolated judgment. Across correctness, entailment, and safety, the isolated judgment remains reliable under instructions not used to identify the subspace, even when the model's final answer and the original predictor degrade. When the judgment question is negated, isolation raises the predictor's AUROC from 0.504 to 0.825 for entailment and from 0.480 to 0.891 for safety. We also find that judgment is readable in earlier layers, before the expression subspace begins to disrupt the predictor. Together, these results provide a way to separate judgment from expression and recover internal judgments even when the model's final answers are unreliable.
benchmark - arxiv:2610.21251 · cs.RODo VLAs Really Follow Instructions? Diagnosing How-Level Instruction Following in Vision-Language-Action ModelsSangwu Park, Wonjoong Kim, Yeonjun In, Chanyoung Park
Vision-language-action (VLA) models are expected to follow instructions like their language backbones, yet prior work has focused on What-level instructions that specify the task rather than How-level instructions that specify how to carry it out. Through our diagnoses, we find that How-level instructions do not help current VLAs and lower task success. We hypothesize that pairing each training episode with a single instruction limits How-level instruction following, and show that training with segment-level enriched instructions makes actions respond to such instructions without consistently raising task success. To measure How-level instruction following beyond standard benchmarks, we introduce SIMPLER-IN, where success requires conditions given only in the instruction. It shows that the gained responsiveness does not extend to demanding or out-of-distribution How-level instructions. How-level instruction following thus remains an open challenge.
vision-language-actionbenchmark - arxiv:2610.21250 · cs.CVFocusOPD: Structure-Guided On-Policy Distillation for Diffusion ModelsXinrui Lin, Shumin Wang, Suqi Song, Fan Li +1
Multi-teacher on-policy distillation enables rectified-flow models with diffusion-transformer architectures to combine task-specific capabilities within a single student. However, matching model outputs does not guarantee that their internal representations are aligned, and it remains unclear where and when additional feature supervision is most useful. We analyze the teacher-student feature mismatch across Transformer blocks and denoising timesteps, and find that the mismatch is concentrated in deeper Transformer blocks during early denoising and persists after OPD distillation, indicating ineffective distillation. Motivated by this observation, we introduce FocusOPD, which augments the base OPD objective with token-wise directional feature alignment restricted to deep layers and early denoising steps. Across two rectified-flow backbones and multiple OPD objectives, FocusOPD improves capability transfer in compositional generation, text rendering, and preference-based evaluation. On the primary benchmark, it improves over DiffusionOPD from 0.953 to 0.967 on GenEval, from 23.797 to 23.944 on PickScore, and from 6.052 to 6.199 on LAION-Aesthetic. Ablations show that selective deep and early supervision is more effective and efficient than alignment over the full network and trajectory.
benchmark - arxiv:2610.21244 · cs.LGSystem1-Arena: Which Decisions Can Be Left to System 1?Tianyu Chen, Xinyang Liu, Ruyi An, Zhendong Wang +1
Large language models (LLMs) often make discrete decisions, such as flagging phishing or escalating to a human. System-1 models, such as Jev and the OpenAI Decisions API, return probabilities over options in a single forward pass, while System-2 approaches generate reasoning steps. Which decisions can System-1 models handle, and which require generative reasoning? We introduce System1-Arena, a benchmark that gives all models the same decision requests and scores their probabilities. It contains 92k questions across 105 tasks from 268 public sources, in four groups: text classification and knowledge, applied decisions, evaluation and control, and agent and action. It also includes 138 closed-loop environments, from web shopping and SWE-bench coding control to robotics, games, and driving; in real-time mode, the world evolves while a model decides. We evaluate 30 models: 7 closed LLMs, 13 open LLMs read at the answer token, and 10 System-1 models. The largest System-1 models match or exceed fast closed LLMs in non-agentic accuracy, with better-calibrated probabilities at one-quarter of the latency. A reasoning LLM is the most accurate, at about 70x the cost of a fast LLM. In closed loops, large open LLMs succeed most often (up to 32.9%); small System-1 models rarely succeed.
agentagenticbenchmark - arxiv:2610.21243 · cs.AIA Responsibility-Centered Coordination Mechanism for Decentralized Multi-Agent SystemsThi-Nhung Nguyen, Viet-Thanh Pham, Linhao Luo, Thuy-Trang Vu +1
Centralized orchestration provides comprehensive multi-agent coordination but creates bottlenecks in scalability, adaptability, and collective performance. Decentralized approaches alleviate these bottlenecks by distributing coordination across agents, yet often leave responsibility for delegated dependencies and failure recovery unclear. To address this gap, we propose RCM, a responsibility-centered coordination mechanism that leverages commitments to couple task delegation with explicit responsibilities of the involved parties. Through recursive delegation, agents retain their upstream responsibilities while locally orchestrating downstream commitments, thereby preserving accountability across delegation chains, creating a network of local orchestration. We operationalize this mechanism locally through responsibility-constrained decision-making and a responsibility-oriented communication protocol (RCP), enabling agents to autonomously establish, manage, and terminate commitments. As agents act in parallel, their local actions continuously evolve the commitment graph, allowing global coordination to emerge and adapt toward a shared mission without centralized control. Experiments across reasoning, coding, and executable agent environments demonstrate that RCM outperforms strong centralized and decentralized coordination baselines by up to 18.7 points in task performance and 10.2 points in Responsibility, while improving failure and harm detection without task-specific training.
agentmulti-agentagent system - arxiv:2610.21239 · cs.ROCLIS-MPPI: Closed-Loop Model Predictive Path Integral Control With Initial-State Sampling Under Execution DelayZidong Liu, Xu Chen
Sampling-based motion planning such as Model Predictive Path Integral (MPPI) control has become attractive for robotics due to its ability to handle nonlinear dynamics and non-smooth objectives without gradient information, while leveraging parallel rollout computation for real-time performance. However, in real-world robotic systems, the optimized control input is often applied after a non-negligible planning delay, causing a mismatch between the state used in forward simulation and the actual state at control execution, which can significantly degrade planning accuracy and may even lead to unsafe behavior such as collisions. This paper proposes Closed-Loop MPPI with Initial-State Sampling (CLIS-MPPI), which explicitly addresses execution-time state mismatch by modeling planning delay as a distribution over plausible execution-time initial states, performing rollout optimization from sampled initial states, and applying a rollout-based local feedback correction during control execution. The proposed method is validated through real-world experiments on a 6-DoF UR5e manipulator in a narrow-passage task involving obstacle avoidance and end-effector target reaching. Across 5 real-world runs of the same task, standard MPPI collided in 2 runs, whereas CLIS-MPPI remained collision-free in all 5 runs. Averaged over all collision-free runs for each method (3 for MPPI and 5 for CLIS-MPPI), CLIS-MPPI reduced the average task completion time from 25.4s to 24.4s and the average end-effector path length from 0.456 m to 0.425 m. It also reduced the overall joint-position and joint-velocity root mean square errors (RMSEs) from 0.100 rad and 0.108 rad/s to 0.017 rad and 0.011 rad/s, respectively. These results demonstrate improved safety, planning-execution consistency, and motion smoothness.
manipulator - arxiv:2610.21229 · cs.LGWhat Should We Remove? Evaluating the Reliability and Costs of RLVR Data DecontaminationEsther Xin
Reinforcement learning with verifiable rewards (RLVR) math corpora are routinely screened for benchmark overlap, but the reliability and data-loss costs of cleaning policies remain underexamined. We audit sixteen open RLVR math corpora comprising 1,673,514 corpus-specific training items against six benchmarks containing 1,949 questions. We evaluate eight policies based on exact, lexical, and semantic matching using 594 independently and blindly dual-reviewed training-benchmark pairs. Calibration and held-out sets have disjoint benchmark questions and training-example families. A separate review of all 83 candidate matches for 80 sampled training items measures item-level false removal. At cosine threshold 0.80, 52 of 59 sampled removals (88%) were confirmed false removals; this is not a demonstrated population rate. No semantic threshold in our preregistered grid (0.80-0.98) meets the 90% conservative-precision target on calibration. The existing judge-adjudicated policy achieves 98.1% resolved precision and 98.2% candidate-frame recall on held-out corpora. It has zero confirmed false removals among six sampled removals, with one unresolved case; this does not establish zero risk. Within the evaluated candidate frame, this policy offers a more favorable precision-recall balance than the semantic-threshold alternatives tested. Recall does not measure contamination outside the candidate frame. We release reviewed labels, per-policy decisions, and coverage accounting to support evaluation of cleaning reliability and cost rather than overlap counts alone.
benchmark - arxiv:2610.21227 · cs.CLKBS: A Korean Benchmark Suite for Foundation Model EvaluationYeonkyoung So, Jongmin Kim, Sungmok Jung, Gyuseong Lee +12
We present a Korean Benchmark Suite (KBS) comprising Ko-ARC, Ko-GSM8K, Ko-EQ-Bench, Ko-WinoGrande, Ko-LAMBADA, and Ko-IFEval, covering six capabilities across 9,396 items. The suite combines review, localization, direct Korean construction, and redesign for Korean, avoiding a translation-only approach to benchmark adaptation. We evaluate Korean-developed models, multilingual open models, and closed API models under protocols that account for model access, prompt format, and scoring differences. Results show that Korean benchmark scores depend on task construction, scoring rules, prompting, and model post-training. These findings suggest that Korean evaluation should report both final scores and the settings used to obtain them.
post-trainingbenchmark - arxiv:2610.21222 · cs.ROVision-Guided Closed-Loop Autonomous Navigation of Light-Driven Janus MicrobotsKanhu Charan Andia, Harsh Prashant Katole, Parth Nitinkumar Bhosale, Hemant Kumar +1
Artificial microbots capable of precise and autonomous navigation could provide versatile platforms for microscale transport, manipulation, and task execution in complex environments. However, reliable navigation becomes increasingly challenging at the micrometre scale, particularly when the surrounding environment changes dynamically. Here, we develop a vision-guided closed-loop navigation framework for light-driven cube-shaped hematite/TiO2 Janus microbots with a characteristic size of approximately 3 μm. Real-time visual localization using a custom-trained YOLOv11 object-detection model is integrated with discretized waypoint-based magnetic steering to enable precise navigation along complex prescribed trajectories, with more than 90% of the measured trajectory deviations remaining below 6 μm across the tested path geometries. The framework further enables arbitrary user-defined free-hand trajectories and autonomous navigation through a physically confined maze. In dynamically changing environments, A* path planning enables real-time path replanning while avoiding multiple independently swimming 5 μm spherical Janus micromotors. Finally, the microbot selectively captures multiple cargo particles and autonomously transports and delivers them to a predefined location while simultaneously avoiding dynamic obstacles. These results establish an integrated platform combining programmable trajectory control, autonomous path planning, real-time environmental adaptation, and task execution within a single light-driven Janus microbot system.
manipulation - arxiv:2610.21220 · cs.AIRecSoup: Efficient Model Merging Approach for Controllable Multi-Objective RecommendationZidong Jia, Guofu Xie, Xiao Zhang
Recommender systems often need to balance multiple conflicting objectives. Controllable Multi-Objective Recommendation (CMOR) enables dynamic adjustment of objective trade-offs at inference time without retraining, adapting to shifting user or system requirements. Existing CMOR approaches typically rely on output-level reranking or parameter-generating modules, which either limit the scope of feasible trade-offs or incur substantial training and inference cost. Model merging provides a promising alternative by reusing multiple objective-specific models via parameter combination. However, its direct application to recommender systems is challenging due to the lack of a shared linear regime and cross-objective interference. To address these issues, we propose RecSoup, a lightweight two-stage approach for effective CMOR via specialized model merging. RecSoup first mitigates parameter interference across objectives through a pruning strategy that jointly considers objective-specific importance and cross-objective relevance. It then performs adaptive parameter merging via a learnable preference-to-coefficient mapping, allowing dynamic trade-off control at inference time without joint retraining or additional generative modules. Theoretically, we prove that RecSoup outperforms the parameter generation paradigm in effectiveness and efficiency, as it can fully navigate the Pareto front without smoothness constraints. Extensive experiments on multiple benchmarks demonstrate that RecSoup achieves improved Pareto front quality and effective controllability while delivers 165--228x faster inference-time latency than parameter generation approaches. Code is available at https://github.com/35798642/RECSOUP.git.
benchmark - arxiv:2610.21218 · cs.AIAutoOnboarder: Autonomously Onboarding LLM AgentsTanay Arora, Cong Liu
Agentic large language models have been documented to lose reasoning capability when placed inside a harness, the software that mediates their observations and actions. To reconcile these losses, we introduce AutoOnboarder, an autonomous system that adapts an agent to a task the way a trainer onboards a new employee. It places the agent under the supervision of an intelligent ``Onboarder'' coding agent, which iteratively tailors its foundation model, prompt, and harness source code to simultaneously improve performance and decrease cost. Like a real onboarding process, the Onboarder is only given tens of examples, and all evaluation data is isolated on a separate virtual machine to prevent leakage. We report the lessons learned from building and testing AutoOnboarder on genomic-sequencing prior authorization form submission. Autonomously onboarded agents are able to match or exceed the performance of a Gemini 3.5 Flash baseline at up to 29$\times$ lower costs. However, these agents remain imperfect, and the Onboarding process varies greatly in terms of final performance. We conclude that future Onboarding systems require a robust environment for the Onboarder, including a structured thought interface, (controlled) internet access, and a wide model search space.
agentllm agentagentic - arxiv:2610.21211 · cs.ROAn Agentic Framework for Proactive Robots with Procedural Assistance MemoryDohyun Lee, Minho Park, Jeonghoon Park, Dongyoon Hwang +4
Proactive robots should anticipate user needs and provide assistance before being explicitly asked. A straightforward approach is to infer human intentions from commonsense knowledge, but the same observed behavior can imply different needs across users, potentially leading to unwanted assistance. We observe that everyday routines in the real world exhibit recurring patterns, so learning these patterns can reduce ambiguity in proactive decisions. Based on this, we propose a memory-based agentic framework in which a VLM task predictor reads a Procedural Assistance Memory (PAM) alongside the current observation to decide when to assist and what task to perform, and a VLA executor performs the predicted task. PAM is structured as a finite-state machine over activity phases, per-state assistance rules, and event-driven transitions; a reflection VLM revises PAM from each day's interaction trace without weight updates or retraining. In our 5-day desk-assistance evaluation, memory-free baselines remain conservative, with limited recall on proactive tasks. Our system improves consistently with accumulated experience, achieving the highest precision, recall, and F1 and reducing first-fire delay from approximately 5 seconds on Day 1 to 1 second by Day 5. Our approach also adapts to changes in the user's routine through continued memory reflection, recovering from memory--routine misalignment after a single memory update. The project page is available at https://davian-robotics.github.io/PAM/
vlamemoryagentic - arxiv:2610.21207 · cs.CVSpaceCache: Intermediate Feature Reuse for Efficient Semantic Segmentation on Point Cloud StreamsDongheon Yang, Beomseok Kang, Jae-Joon Kim
Point-cloud streams are central to autonomous driving and robotics, yet existing segmentation networks process every frame from scratch, although consecutive frames largely overlap. We observe that the intermediate features of point-cloud transformers remain semantically consistent across frames for the same spatial region, especially in deep stages. Building on this, we present SpaceCache, a training-free feature cache that reuses these features across frames. Tokens of previously observed regions skip a sequence of transformer blocks and take their cached features, only the remaining tokens are computed, and the full token layout is restored for dense prediction. Lightweight cache management keeps reused features reliable by limiting each entry's age and evicting low-confidence entries. Unlike token merging, SpaceCache selects tokens at negligible cost and runs its cache operations once per encoder stage. On both outdoor and indoor benchmarks, such as nuScenes, Waymo, and ScanNet, with two transformer backbones of Utonia and LitePT-B, SpaceCache achieves up to 1.65x speedup with less than 1-point mIoU drop and consistently outperforms token-merging methods.
benchmark - arxiv:2610.21200 · cs.LGPoisoning the corpora of protein design systems during pre-training and retrievalCharalampos Koilakos, Kimon-Antonios Provatas, Michail Patsakis, Candace SY Chan +2
Protein foundation models increasingly support the generation of novel biological sequences, yet their reliance on large public datasets creates opportunities for adversarial manipulation of the material they learn from and retrieve. We investigate two data-only poisoning attacks against protein design systems, targeting the pre-training corpus of autoregressive protein language models and the retrieval corpus of a text-guided protein generator. For the first, we pre-trained ProtGPT2 and ProGen2 variants of 300 million parameters, introducing poisoned protein sequences containing one of two triggers followed by one of three payloads. Across all poisoned runs, both architectures acquired highly reliable trigger-dependent generation under uniform and ramped poisoning schedules. Under ramped exposure, the earliest majority activation occurred after only 200 contaminated examples, while for the uniform schedule earliest activation occurred around 5000 poisoned samples. For the second attack, we inserted records into a retrieval corpus system without modifying its weights or decoding procedure. A single query-aware record captured a top-16 retrieval position for 58\% of templated requests and 97\% of free-text requests. Conditional on successful poisoned retrieval, 62\% of requests produced altered sequences, while 19-21% acquired the inserted domain by profile-HMM homology. A universal query-unaware attack using 1,500 records also achieved substantial retrieval capture and domain transfer. Together, these findings demonstrate that adversarial protein sequence contributions can establish persistent, sequence-specific behaviour during pre-training or redirect protein generation at inference, while leaving ordinary model behaviour largely preserved. Our work establishes the integrity and biological provenance of protein sequence data as critical safeguards for generative protein design systems.
manipulation - arxiv:2610.21195 · cs.AICondition-Contrastive Extrapolation for On-Policy Distillation of Language ModelsWenxuan Luo, Chenxi Li, Haoyue Liu, Yongxin Guo +1
On-policy distillation (OPD) improves a student language model by training on its own generated trajectories with supervision from a stronger teacher. However, when conditioned on student-generated prefixes, the teacher can favor invalid continuations that violate the question's premises, causing OPD to encourage erroneous reasoning steps. As a remedy, we introduce Condition-Contrastive Extrapolated On-Policy Distillation (CCE-OPD), which makes teacher supervision sensitive to question conditions. CCE-OPD incorporates (1) a fixed-prefix comparison in which the same frozen teacher evaluates the complete question and a counterpart with a selected condition masked, and (2) a bounded log-probability contrast that reweights the teacher distribution into an auxiliary distillation target. On 200 diagnostic reasoning states, condition contrast reduces the rate of preferring invalid continuations from 27.0% to 16.0%. Extensive experiments across six mathematical benchmarks show that CCE-OPD improves avg@8 over OPD by an average of 1.60 and 1.95 percentage points when distilling Qwen3-8B into 0.6B and 1.7B students, respectively. It also achieves the strongest results on four transfer benchmarks.
benchmark - arxiv:2610.21190 · cs.CVControlFlow3D: Distilling Multi-View Knowledge into Latent Flow Matching for Point Cloud UpsamplingYuang Liu, Zhi Zuo, Zhengkai Zhao, Lirui Zhang +3
Point cloud upsampling is essential for 3D applications but remains challenging due to sparse, under-constrained observations. Multi-view images can resolve geometric ambiguity, but existing multi-modal methods do not fully exploit visual cues. We propose ControlFlow3D, a framework that distills knowledge from pretrained foundation models into a lightweight generation network during training, enabling strong performance without multi-view images available at test time. Specifically, our method performs conditional flow matching in a structured latent space defined by a frozen 3D encoder. Multi-view image tokens and 3D latent tokens are jointly clustered into shared prototypes via unsupervised soft KMeans; the resulting prototypes guide the flow trajectory through velocity control and are aligned by a consistency loss that progressively transfers image knowledge into the base network. A Mamba-based geometry decoder then maps latent features to dense coordinates with linear complexity. We also construct ShapeNetPU, a 35K-object multi-modal benchmark that is 31 times larger than prior datasets. Extensive experiments across multiple benchmarks demonstrate the effectiveness of our method and its strong generalization capability.
benchmark - arxiv:2610.21189 · cs.LGResisting Harmful Instructions through Training-Free Weight EditingAli Ebrahimpour-Boroojeny, Xiaocong Yang, Samuel Schapiro, Zahra Borhani +3
Instruction-following language models can comply with requests that conflict with intended behaviors and safety alignment. We introduce a training-free weight-editing approach for reducing compliance with targeted undesirable instructions. Our method needs only forward passes on a small set of prompt pairs per target behavior, and because the edit is stored in the model's weights, changing or deleting instructions in the prompt cannot undo the weight modification. We evaluate the approach on two target harmful behaviors across five instruction-tuned models: instructed falsehood and instructed sexism. On two instructed-lying benchmarks, it improves mean honest response rates by 6.2 and 4.6 percentage points over the strongest alternative, significantly on three of the five models and within noise on the other two. In a controlled study of seven prompt formats, it maintains higher honest-response rates than single-layer activation steering when orders to lie arrive in single-turn formats not used to construct the edit. For instructed sexism, it reduces the mean sexist-response rate from 57.9% to 24.3%; combining the edit with a fairness prompt reduces it further to 11.9%. These results establish our contrast-based weight editing approach as an effective training-free intervention in two distinct behavioral domains, supporting its use both independently and alongside prompt-based safeguards.
benchmark - arxiv:2610.21184 · cs.CLBeyond Retrieval Failure: Understanding Why LLMs Miss Evidence in Multi-Hop RAGJiaqi Wei, Wenxuan Huang, Chengdong Xu, Xiang Zhang +1
Multi-hop retrieval-augmented generation (RAG) requires an agent to retrieve evidence for a sequence of dependent sub-questions. When a wrong answer lacks a required passage, the passage may be inaccessible to the retriever, or the agent may have failed to issue a query that would retrieve it. Standard retrieval evaluations conflate these two causes. We study how much apparent retrieval failure is instead caused by search drift: the search no longer tracks an unanswered sub-question. Holding the retriever, corpus, and generation model fixed, we introduce ExecutionRAG, a training-free requirement register that records open sub-questions and their supporting evidence. The register changes search guidance but not evidence access, so each gold passage it recovers certifies that the passage was reachable by the original retriever. On 1,000 MuSiQue questions, the number of wrong answers with unretrieved required evidence falls from 382 to 234 with all-MiniLM-L6-v2 and from 383 to 274 with Qwen3-Embedding-0.6B. These counts correspond to 69% to 44% and 66% to 48% of each system's wrong answers, respectively. Crucially, evidence recovery does not produce a proportional reduction in total error: premature resolutions and abstentions increase as missing-evidence failures decrease. The central finding is therefore a shift in the remaining bottleneck, from guiding search toward required evidence to verifying that the retrieved evidence is sufficient.
retrieval-augmentedragagent - arxiv:2610.21180 · cs.ROEvo-DexBench: Benchmarking Generalization across Related Tasks for Dexterous ManipulationBing Cheng, Yiran Mao, Zewei Ye, Yifan Jin +9
General-purpose dexterous manipulation requires policies to transfer knowledge from related tasks to new tasks with only a few target demonstrations. Existing benchmarks often provide too many target demonstrations. They also leave the differences between source and target tasks uncontrolled. As a result, they cannot clearly distinguish knowledge transfer from task memorization. Our key insight is the Core-Gen principle. Core tasks are related tasks that share a functional interaction and provide source knowledge, while a Gen task is a novel target within the same functional family that has only a few demonstrations. This setup tests whether a policy can reuse functional interaction knowledge. Based on this principle, we propose Evo-DexBench, which controls both source-task relatedness and target-task exposure. It comprises 30 tasks in six functional families and a VR teleoperation pipeline. The resulting dataset covers three everyday scenes per task. It contains 1,248 multimodal demonstrations collected with simulated Shadow Hands equipped with 123 taxels. Experiments show that even the best-performing method achieves an average success rate of only 38.5%, highlighting the challenges existing methods face in transferring functional interaction knowledge. Further analysis shows that, across the two tasks, joint Core+Gen training improves performance by 12% and 38%, respectively, over training on Core followed by fine-tuning on Gen. This indicates that joint training makes better use of Core knowledge and that functionally matched Core data yields greater gains in success rate than mismatched data. We release the Evo-DexBench dataset and benchmark to support the evaluation of dexterous manipulation policies' generalization across related tasks.
manipulationdexterousteleoperationbenchmark - arxiv:2610.21177 · cs.AIAgentEval: Evidence-Bounded Assessment of Stateful AgentsRuiqi Gao
Task completion alone does not determine whether a stateful agent satisfied its obligations, and a recorded judgment does not guarantee that its supporting evidence remains valid. We present AgentEval, a system that carries a contract-bound assessment record through tool execution, state and event capture, report qualification, and case maintenance. It separates what the agent did, what the evidence establishes, and whether the conclusion is supported for present use. On 32 tool-world administrations, independent checks agree on all 30 complete episodes, distinguishing satisfied, violated, and unresolved obligations; two interrupted administrations are retained. Controlled interventions on 48 evidence views expose 36 binding faults missed by hash-only checks without changing agent behavior. Scenario Studio connects reviewed cases to 12 two-turn reports with 24 independently checked answers. A separate public-benchmark study on 192 held-out executions reveals ten completed return subgoals within broader task failures and isolates the effect of missing evidence. These studies show how an executable assessment record makes scope, evidential support, and report reuse inspectable across an agent evaluation workflow.
agentbenchmark - arxiv:2610.21176 · cs.LGLanguage Models Share a Readout Bottleneck in Working MemoryDongyu Gong
Large language models (LLMs) reproduce a human working-memory signature on the N-back task: performance declines as memory load N increases, and errors concentrate on lures, items that repeat the current stimulus at a neighboring but incorrect serial position. We ask whether this shared behavioral signature reflects a shared mechanism. The literature offers two candidate accounts. On a precision account, softmax attention cannot concentrate its weight on the item N positions back, so adjacent items are retrieved instead. On an interference account, attention selects the correct item, but the content of competing items corrupts the decision. We dissociate these accounts in 15 pretrained LLMs from three families (70 million to 14 billion parameters) and in minimal transformers trained from scratch. In all three families, the attention heads that carry the task attend to the target on most missed matches. Linear probes decode near-perfectly whether a matching letter occurred at the target or at a lure position, even on trials where the model false-alarms to the lure, although its own decisions barely distinguish the two. Sharpening attention yields negligible gains, whereas suppressing the letter identity of non-target items improves sensitivity (d') in nearly all model-by-load conditions. In larger models the task is spread over more heads, and removing interference helps more. In minimal transformers, precision limits a single attention layer, but a second layer removes this constraint. The working-memory limit of LLMs therefore lies in reading items out under interference rather than in locating them, consistent with interference accounts of human working memory.
memory - arxiv:2610.21170 · cs.AIMemScope: Evolving Agent Memory Architectures through Localized Code RepairMiao Su, Yucan Guo, Zixuan Li, Long Bai +3
Agent memory architectures define how LLM agents store and reuse experience across tasks, but they are typically fixed after deployment, limiting their adaptability across task types. Recent self-evolving approaches instead adapt these memory architectures using execution feedback, typically represented as evaluation scores or compressed summaries. However, such coarse-grained feedback is insufficient for precise failure localization in current agent systems, where long-horizon task execution involves multiple interacting components. Without a precise failure location, existing approaches lack a clear code-level target for modification, leaving the update scope weakly constrained and often leading to large-scale rewrites of an otherwise capable memory architecture. We propose MemScope, a framework for evidence-grounded, localized evolution of an existing memory architecture. Its tool-using Patch Agent jointly inspects raw execution traces, runtime memory state, and current architecture code to identify repair targets and apply bounded, in-place edits under provider-identity and memory-compatibility constraints. On GAIA, the trace-and-memory inspection variant achieves 86.67% accuracy, compared with 70.00% for trace-only and 63.33% for summary-only feedback. Under xBench tasks, the evolved memory architecture improves over its original version from 49 to 55 correct answers with DeepSeek V3.2 and from 50 to 53 with Kimi K2.7 Code. On disjoint ALFWorld unseen tasks, success rises from 50.75% to 71.64%. Across archived architecture-snapshot transitions, the median code-change ratio is 4.4%. These results support fine-grained execution evidence as a basis for localized, in-place memory-architecture evolution.
memorymemory architectureagent memoryagentllm agentagent system - arxiv:2610.21167 · cs.AIGRADS: Graph-Regulated Agentic Data Synthesis with Controllable DifficultyGaocheng Yang, Ruijie Tao, Kai Ye, Senhua Zhang +2
Deep research agents are increasingly deployed to retrieve precise information from large corpora, yet multi-hop, long-horizon tasks remain challenging. Synthetic data generation has emerged as a scalable solution, with existing pipelines typically expanding seed tasks to increase their apparent depth and breadth. However, this expansion does not reliably translate into necessary reasoning: newly introduced evidence may be only weakly coupled to the answer, creating a mismatch between intended and empirical difficulty. To address this problem, we introduce GRADS, a graph-guided agentic data synthesis framework with explicit difficulty control. GRADS runs a closed loop of seed selection, evidence expansion, question--answer construction, and verification while maintaining a graph of the evidence and inference dependencies required to derive each answer. At each step, it conditions generation on a relevant subgraph, preserving indispensable inference without accumulating irrelevant context. Experiments across security, software-engineering, and scientific-document domains show that GRADS consistently generates verified tasks with empirically separated difficulty distributions, as measured by independent solvers. Moreover, under matched training budgets, post-training on this data delivers consistent gains at every level: SFT raises the solution-termination rate to over 95% while cutting tool calls and turns by up to 75%, RL trims tool use by up to a further 19.5%, and on our held-out, domain-specific benchmarks, the resulting 4B model matches Sonnet~4 on software engineering and surpasses it on the scientific and security domains.
agentictool usepost-trainingbenchmark - arxiv:2610.21165 · cs.CLFrom Long Context to Selective Memory:A Survey of Memory-Centric Compression for Large Language ModelsChaogui Gou
Long-context capability has become essential for large language models (LLMs) to handle complex tasks, yet continuously expanding the context window does not remove the computational, storage, and information-selection pressures incurred at inference time. Existing work on prompt compression, KV-cache management, efficient serving, retrieval augmentation, and agent memory is usually discussed in isolation, leaving the flow of information between storage and computation insufficiently understood. This survey reframes long-context compression as a memory-management problem for allocating inference-time resources. We distinguish compute-active memory, which directly participates in model computation, from compute-inactive memory, which lies outside the computation path but can be stored, retrieved, updated, or forgotten, and we characterize offloaded KV caches, persistent prefix caches, and precomputed KV states as intermediate states on the activation path. On this basis, we formulate writing, offloading, recall, retrieval-and-injection, materialization, eviction, consolidation, and forgetting as state transitions within an information lifecycle, and introduce a budgeted cost model that places these operations in a single cost account. The resulting framework unifies short-term context compression, long-term memory management, and the transitions between them, and shows that the central challenge is not merely reducing information volume but deciding which information should participate in costly inference, when, and in what form. We further summarize evaluation dimensions, benchmark gaps, and open challenges, and argue that future research should move from static, pointwise compression toward dynamic, governable, lifecycle-aware memory management.
memorylong-contextlong contextcontext compressionagent memoryagent - arxiv:2610.21160 · cs.ROWhen to Ask for Help: A Decision-Theoretic Account and Measurement Standard for Slow-Module Scheduling in Dual-System Vision-Language-Action ModelsShilin Zhang
Dual-system vision-language-action (VLA) models pair a high-frequency fast policy with a slow, expensive module, and published trigger mechanisms for invoking it lack a shared objective, measured costs, and any standard separating schedule quality from module quality. We provide the missing account: invocation is formalized as a calibrated decision problem whose fixed-ratio, fixed-budget, and history-gated schedules are degenerate cases of one quantity, the Value of Intervention (VoI); a renewal baseline proves the optimal schedule periodic under state-independent collapse, with adaptive gains governed by the cost ratio $ρ$ and the dispersion of collapse. Measured systems span three orders of magnitude in $ρ$ (FiS-VLA $0.95$; RoboICL about $10^3$), and a direct test confirms the low-$ρ$ prediction: calibrated gains vanish under homogeneous collapse and grow with heterogeneity. Fourteen closed-loop arms on LIBERO test the contested middle of the continuum ($ρ=1.53$): the adaptive trigger allocates its queries exactly as the threshold rule predicts, yet the language-rewrite channel cannot convert them into success on either suite, and a revert actuator that undoes a failed rewrite is the only intervention with zero net downside on both LIBERO-Spatial and LIBERO-Goal, where the rewrite deficit deepens from thirteen to forty-two points. A four-measure leak-free protocol cuts apparent early-warning AUC from $0.93$ to $0.62$ on $929$ deployment traces, a detection-limited frontier theorem converts detection quality into a schedulability statement, and a compilation loop turns measured value functions into per-task refresh periods within $2.4\%$ of the closed-loop oracle. Together these results turn slow-module invocation from a design opinion into a measured decision.
vision-language-actionlibero - arxiv:2610.21159 · cs.CVVisualTrust: Video Reasoning under On-Screen Claim FramingHuachao Zhu, Junjielong Xu, Sicheng Li, Liang Yang +4
On-screen text can describe what a video shows, or it can frame what the video is supposed to mean. Current benchmarks treat both alike, testing whether a model can read the text rather than whether it should trust it. We study claim deference, multimodal large language models' tendency to let an on-screen claim override conflicting visual evidence. We introduce VisualTrust, a counterfactual benchmark with 5,063 cases across 14 claim-fact classes. Each case pairs a question with three versions of the same video: one with no claim, one with an accurate claim, and one with a misleading claim that alters a single grounded fact. The option the misleading claim supports is the trap answer. Across seven open-weight and four proprietary models, mean accuracy falls from 90.1% without a claim to 51.9% with a misleading claim. Among initially correct responses, 42.8% flip to the trap answer and only 0.6% become other errors. Strong no-claim accuracy does not ensure resistance. We also introduce DEAR (Decoupled Evidence Arbitration and Reading), a training-free agent for video question answering that separates evidence arbitration from reading. An independent check reports conflicting evidence, a selector weighs it without a fixed modality priority and blinds what it withholds, and the reader answers from the resulting video in a fresh context. On 500 videos across three proprietary models, DEAR raises accuracy under misleading claims from 47.2--64.0% to 81.8--85.2% and reduces trap-answer selection from 34.4--50.8% to 11.6--15.0%, while preserving no-claim accuracy.
agentbenchmark - arxiv:2610.21156 · cs.CLStop Training Against Soft Users: Mentalized Adversarial Self-Play for Proactive DialogueZanwu Liu, Dan Zhang, Xiaodi Wang, Yongbo Mao +5
Proactive dialogue agents must guide multi-turn interactions toward task goals while adapting to users whose intentions and attitudes evolve. Existing planners improve system side action selection through search, learned policies, or user aware planning, but are often trained and evaluated against fixed or overly compliant simulators; as a result, high simulated success can reflect soft user bias rather than robust dialogue ability. We introduce MASP, a Mentalized Adversarial Self Play framework that treats proactive dialogue as a partially observable game between a system and a rationally resisting user. MASP represents the user's evolving mental state with a BDI based latent state, separates reward side teacher mentalization from policy side student mentalization, and alternates self play updates for the system and user policies. This turns robustness to rational resistance into a training objective rather than only an evaluation condition. Across diverse proactive dialogue benchmarks, compared with the widely used SFT and Planner-only approaches, MASP substantially improves success rate (SR) and rationality constrained success (RCSR), reduces average turns (AT), and performs better against hard users who make the task goal more difficult to achieve.
self-playbenchmark - arxiv:2610.21152 · cs.ROKODA: Kinematics--Observation Data Alignment for Humanoid Learning from Human Egocentric DemonstrationZibo Zhou, Yueru Chen, Junsong Wu, Weiji Xie +7
Training humanoid whole-body vision-language-action (VLA) policies requires costly robot demonstrations with aligned robot-view observations and controller-compatible whole-body state-action labels. Egocentric human demonstrations offer scalable supervision but have misaligned views and sparse motion cues. Existing approaches either pretrain policies that still require robot data for adaptation or generate controller-specific commands that limit reuse. Conversion also requires robot-view observations aligned with the generated whole-body motion. We present KODA (Kinematics-Observation Data Alignment), which converts egocentric RGB demonstrations into whole-body VLA training data using a reusable canonical trajectory to guide robot-view synthesis and generate controller-specific whole-body state-action supervision. KODA first generates a simulator-checked whole-body trajectory using a sparse tracker with sampling-based command optimization. This generated trajectory directly determines the onboard camera poses used to synthesize corresponding robot-view observations. The motion, observations, and task instruction form a canonical record. For each deployment controller, KODA optimizes commands to track the stored motion and records the corresponding action labels and proprioceptive states. When switching controllers, KODA retains this canonical record and regenerates whole-body state-action supervision for co-training with matching robot demonstrations. We evaluate KODA on four Unitree G1 loco-manipulation tasks using two whole-body controllers, SONIC and HEFT. With robot data held fixed, macro-average success on conditions covered by human demonstrations but absent from robot data improves by 45.0 and 46.25 percentage points, respectively. Under both controllers, this metric shows an upward trend across the evaluated human-data budgets, suggesting potential for further scaling.
vision-language-actionvlamanipulationhumanoidwhole-body control - arxiv:2610.21149 · cs.CLFrom Observation to Intervention: Distilling Self-Generated In-Context Demonstrations via Executable Causal InterventionMinming Ye, Min Liu, Haoang Chi, Ciheng Wu +1
Self-generated In-Context Learning (ICL) enables LLMs to synthesize their own few-shot demonstrations, often outperforming manual prompts. However, existing methods filter demonstrations by observation, assuming a rationale is high-quality if it leads to the correct output. We show such filtering suffers from \textit{observational equivalence}, whereby both faithful causal rationales and spurious post-hoc rationales can yield same outputs and are statistically indistinguishable from observation alone. Inspired by independent causal mechanisms, we argue that faithful rationales imply \textit{transfer invariance}, in which causal rationales should remain effective across logically similar samples. We instantiate this criterion as \textbf{E}xecutable \textbf{C}ausal \textbf{I}ntervention for \textbf{D}emonstrations \textbf{D}istillation (ECI2D), which abstracts language rationales into executable programs and applies the causal interventions ($do$-operator) on logically similar samples, retaining only those demonstrations that imply transfer invariance. Experiments on math, logical and commonsense reasoning benchmarks demonstrate that ECI2D consistently outperforms observation-based ICL while retaining far fewer demonstrations, with gains exceeding 7 points as the test distribution shifts.
benchmark - arxiv:2610.21147 · cs.AIAll-in-One Memory: Process-Rewarded Reinforcement Learning for Memory AgentsBingbing Wang, Jingjie Lin, Weiming Qiao, Jing Li +1
Large language models increasingly rely on external memory to answer queries grounded in long-horizon histories such as multi-session dialogues and lengthy documents. However, existing memory methods often optimize retrieval and memory construction separately, without a shared reward space for evaluating search, writing, and stopping decisions. To address this, we propose All-in-One Memory (AiO-Mem), which organizes these operations as transitions over a shared evidence state and directly evaluates the semantic quality of each action in a unified reward space. AiO-Mem compiles heterogeneous histories into source-traceable multi-view events, providing a shared substrate for retrieval and memory writing. A unified memory policy actor then rolls out complete interaction trajectories over this shared evidence state. A separate trajectory-conditioned process reward model is trained through evidence-grounded supervised fine-tuning and trajectory-level DPO to assign process scores to the collected trajectories. These scores are aggregated into an action-balanced trajectory reward and used by Process-Aligned DAPO (PA-DAPO) to optimize the memory policy. Across long-context memory benchmarks, AiO-Mem consistently improves answer accuracy over both static memory systems and RL-trained memory agents.
memorylong-contextexternal memorybenchmark - arxiv:2610.21143 · cs.AICoSTA: Compressed Speech Token Adaptation for Efficient Spoken Dialogue ModelsYingzhuo Deng, Chong Li, Bingxin Liu, Yi Shu +5
Spoken Dialogue Models (SDMs) typically share a speech-token space for generation and synthesis, leading to long autoregressive sequences and costly inference. Lower-rate speech tokenizers shorten sequences but may introduce compression-quality trade-offs and costly replacement without breaking this coupling. To address this limitation, we propose CoSTA, a compatibility-preserving post-hoc framework that decouples generation and synthesis token spaces while retaining the original speech tokenizer and decoder. CoSTA extends the generation token space with compressed representations via mean-initialized vocabulary extension, pronunciation adaptation, and dialogue adaptation, followed by deterministic expansion back to the original speech-token space before synthesis. On bilingual speech-dialogue benchmarks, CoSTA achieves approximately $1.7\times$ expanded speech-token generation speed, lower full-response latency, and higher throughput while preserving response quality. Analyses show more sustained interaction under limited budgets and a practical trade-off among compression, decoding efficiency, and quality. Further experiments demonstrate cross-architecture transfer and complementarity to low-rate tokenization.
benchmark - arxiv:2610.21140 · cs.ROManifold Generative Annealing for Contact Constraint-aware Impedance Control with Reinforcement Learning PriorZhilin He, Guoquan Qi, Changliu Liu
Contact-rich robotic tasks require coordinated motion and impedance under uncertain interaction dynamics and physical constraints. While recent methods on learned policies provide useful interaction strategies, they may fail when contact conditions differ from those encountered during training. Model-based generative methods show promising receding-horizon trajectories optimization results, but often lack structured behavioral guidance and refinement that jointly accounts for constraint geometry and actual contact responses. To this end, we propose Manifold Generative Annealing (MGA), a joint model-free and model-based generative framework for motion--impedance optimization. MGA first organizes pretrained reinforcement learning policies into structured sequence proposals, retaining task-specific interaction strategies without committing to their execution. It then evaluates these proposals alongside exploratory samples through model-based rollouts for refinement or separate execution selection. Crucially, realization-aware manifold refinement aligns updates from the refinement bank with both task constraints and locally achievable physical responses, allowing learned experience to guide online adaptation under changing contact conditions. Experiments cover surface scanning, peg insertion, and humanoid pushing, with improved safety success rate of impedance control under nominal and sensing-shift conditions.
humanoid - arxiv:2610.21138 · cs.AICleanRM: Reward Modeling from Noisy Feedback with a Few Clean Feedback for RLHFLicheng Pan, Haocheng Yang, Yuzhou Chen, Yichen Sun +6
Feedback noise is pervasive in preference data, arising from subjective human judgments, annotation mistakes, and imperfect LLM judges. Standard reward modeling trained on such feedback yields a biased estimate of the clean risk. Although classical denoising methods are directly applicable to noisy feedback learning, they generally assume a uniform noise ratio across samples. However, this assumption fails for feedback data, where noise is intrinsically sample-wise. To address this challenge, we propose CleanRM, which uses a few clean feedback to calibrate sample-wise noise ratios via optimal transport. Specifically, CleanRM aligns the joint distribution of preferences reverted from noisy feedback with that of clean feedback and then trains the reward model using a noise-corrected surrogate loss. Theoretically, we prove that the resulting objective is unbiased with respect to the clean risk. Extensive experiments demonstrate that CleanRM outperforms state-of-the-art denoising baselines and yields more effective downstream RLHF alignment.
rlhf - arxiv:2610.21135 · cs.AIStronger Judges Do Not See More: Deployment-Integrity Blindness in LLM EvaluationJianchang Su, Dawei Xiang, Yifan Zhang, Wei Zhang
LLM-agent evaluation reports behavioral scores but leaves the identity of the deployed runtime stack unreported. A tampered deployment that preserves evaluated outputs passes every output-only evaluator. We introduce DIB-Bench, a benchmark that measures this deployment-integrity blindness across 7,560 scored examples, six tamper types, and an evaluator ladder from regex scorers through LLM judges to GPT-5-nano. Stronger evaluators accept more tampered outputs: mean acceptance rises from 88.6% (rule-based) to 94.1% (GPT-5-nano), and a trigger-conditional tamper invisible on the evaluation suite produces 48-100% harmful outputs, rising with model capability. A PAC-style bound shows that tampers with zero output shift are undetectable from outputs at any sample size; a manifest-hash identity channel detects them with zero samples. Behavioral fingerprinting identifies the model component above chance (76.7%) but covers one of six stack components. We also propose a graded deployment-identity descriptor for evaluation reports.
benchmarkevaluator - arxiv:2610.21134 · cs.LGPredicting Live Birth After Fresh Embryo Transfer in IVF/ICSI: Differential ICM/Trophectoderm Morphology Encoding and a Hybrid Attention--Logistic Regression EnsembleAbdul Raheem Shahzad, Jinxin Chen, Ming Zhang, Juan Deng
Predicting live birth before fresh embryo transfer could support shared decision-making in IVF/ICSI, but routine clinical data are small, heterogeneous, and contain pathway-dependent variables prone to outcome leakage. We aimed to construct a curated single-centre IVF/ICSI dataset and develop a prediction model suited to small tabular clinical data. \textit{Materials and Methods:} We assembled 616 fresh embryo-transfer cycles (251 live birth, 365 non-live birth) with clinical, laboratory, and embryological predictors, and embryo morphology encoded with a novel differential weighting of the inner cell mass and trophectoderm. After KNN imputation, robust scaling, and selection of 50 features, we developed a Hybrid Attention Network (HAN), a learnable feature-attention gate followed by a compact multilayer perceptron, and a HAN + logistic-regression (LR) ensemble with a validation-tuned weight. Models were benchmarked against classical machine learning, TabPFN, and text-serialized language models on a held-out test set; AUC-ROC was the primary metric. \textit{Results:} The HAN + LR ensemble achieved the highest test AUC-ROC (0.710; accuracy 0.710, F1 0.630), ahead of the HAN backbone alone (0.700), logistic regression (0.667), the TabPFN + LR hybrid (0.666), and TabPFN (0.650). Text-serialized language models performed near chance (AUC $\le$ 0.61). The differentially weighted trophectoderm feature ranked highest by consensus importance across models. \textit{Discussion:} A compact attention network combined with a linear model outperformed both components and strong foundation-model baselines on a small clinical cohort. \textit{Conclusion:} The curated dataset and HAN + LR ensemble offer a transparent, reproducible basis for IVF/ICSI live-birth prediction; external validation is the next step.
benchmark - arxiv:2610.21126 · cs.CLCompression Does Not End at Selection: Post-Selection Positional Geometry for Sparse KV CachesWanyang Hong, Zhaoning Zhang
Sparse KV compression reduces long-context inference costs by retaining selected historical states, yet these states typically keep their coordinates from the full history. Under Rotary Position Embedding (RoPE), gaps left by evicted tokens continue to affect query-key matching, potentially limiting the use of retained evidence. We study post-selection positional geometry on a fixed selector-produced cache, examining how positional organization affects the use of retained information and which coordinate and continuation rules improve answer quality. To reorganize this geometry, we propose RoPERelay, which assigns compact, order-preserving RoPE coordinates, relocates cached key phases through inverse rotation and re-rotation, and synchronizes the current query and subsequent generation. It reuses existing K/V without renewed selection, historical re-prefill, or additional cache budget. Across 36 comparisons spanning two models, six sparse selectors, and three long-term memory tasks, RoPERelay improves answer quality in 32 cases. For Llama-3.1-8B-Instruct with BM25 on LongMemEval-S, accuracy increases from 23.6% to 35.4%. Measured complete-request time overhead is approximately 1%-3% in the evaluated configurations. These findings establish positional organization as a distinct optimization dimension for sparse KV caches after content selection.
memorylong-context - arxiv:2610.21123 · cs.LGDST-Net: Decomposed Spatio-Temporal Network for Time Series ForecastingYaohua Zha, Hai Wu, Shu-Tao Xia
Series decomposition has emerged as a widely acknowledged paradigm in deep time series forecasting models. However, existing approaches predominantly employ a homogeneous modeling strategy for all subsequences, thereby overlooking the distinct inductive biases inherent in different components. We identify that the trend component exhibits dominant spatial coupling across channels but carries significant temporal redundancy, whereas the seasonal component is characterized by local temporal evolution but remains susceptible to attention distraction. To address this, we propose a novel disentangled architecture that tailors specific modeling strategies to align with the intrinsic properties of each component. For the trend branch, we design a Latent Graph Block, employing latent graph modeling to capture global topological dependencies and system-wide co-movements. Conversely, for the seasonal branch, we introduce a Gaussian Multi-scale Attention mechanism to mitigate attention distraction. By injecting learnable Gaussian priors, we impose structural constraints that compel the model to focus on multi-granular local contexts, effectively filtering out global noise and accurately capturing local evolutionary phases. Extensive experiments on multiple real-world benchmarks demonstrate the effectiveness of our proposed model.
benchmark - arxiv:2610.21113 · cs.AISSR: Structural Signal Retriever for Scalable Agent Discovery in LLM-Orchestrated Multi-Agent SystemsGiancarlo Maldonado Cárdenas, Carlos A. Astudillo, Nelson L. S. da Fonseca
Large-scale multi-agent ecosystems require mechanisms for discovering suitable agents without exposing entire registries to a large language model (LLM) orchestrator. The challenge is not only to retrieve relevant Agent Cards from a growing registry, but also to maintain a compact candidate context, since each additional candidate increases prompt size, inference cost, latency, and the risk of exceeding the model context window. We formulate this problem as candidate-context scalability and propose the Structural Signal Retriever (SSR), our deterministic and training-free retrieval method designed for Agent Card discovery. SSR derives structural lexical signals from capability-bearing Agent Card fields, weights them using inverse document frequency (IDF), and uses an Inverted Signal Index (ISI) for selective candidate scoring. We evaluate SSR using a synthetic agent benchmark, a public MS MARCO adaptation, and end-to-end LLM orchestration. At 1,000 agents, SSR reduces LLM input by 98.8% and orchestration latency by 70.9% relative to full-catalog orchestration. On MS MARCO, IDF improves retrieval effectiveness within the structural representation, while the ISI preserves retrieval effectiveness and reduces explicitly scored passages by 96.8%. End-to-end experiments show that candidate pre-selection maintains a compact candidate context as registry size grows and avoids the operational failures observed with full-catalog orchestration. These results support SSR as an Agent-Card-native, traceable retrieval mechanism for candidate-context scalability.
agentmulti-agentagent systemagent benchmarkbenchmark - arxiv:2610.21109 · cs.LGKolab: Benchmarking Human-Agent Collaboration in Long-Horizon WorkflowsShirley Anugrah Hayati, Patricia Anugrah Setiani, Ruizi Wang, Zhouyi Li +3
When humans collaborate with AI agents on long-horizon tasks, what happens to the agents' contributions? For solving long-horizon tasks, humans place greater importance on ongoing interaction with AI agents (e.g., revising plans) than on an agent's ability to complete the task in a single attempt. However, most agentic benchmarks have focused more on automation. Strong performance on such benchmarks does not imply effective real-world collaboration. Professional workflows offer a useful setting, as they often involve interdependent steps. To address this gap, we introduce Kolab, a new dataset of 115+ hours of screen recordings of 50 professionals from 40 occupations collaborating with agents on realistic, long-horizon problems. Kolab also contains participants' collaboration and output quality judgments. Our analysis reveals diverse collaboration patterns, highlighting the importance of when and how professionals coordinate with agents. A collaborative agent must do two things that automation benchmarks never test: anticipate whether experts will adopt, modify, or reject its contribution once it leaves the chat window to guide its next steps, and assess collaboration quality from the unfolding workflow without constant expert feedback. Our benchmark operationalizes both: (1) forecasting how professionals use agent contributions at diverse workflows, and (2) verifying their judgments of collaboration satisfaction and output quality. Findings show that frontier models, both agentic and non-agentic, overestimate collaboration with experts and struggle to distinguish between high- and low-quality collaboration. Longer workflow context does not consistently improve these predictions. As frontier agents cannot yet understand long-horizon collaboration with experts, we position Kolab as a valuable dataset for studying human-agent collaboration grounded in real professional workflows.
agentai agentagenticbenchmark - arxiv:2610.21098 · cs.LGToolBottleneckBench: Measuring Tool-Agent Performance under Controlled Pressures and Fixed Harness BudgetsZichuan Xu, Yuyang Dong, Makoto Onizuka, Chuan Xiao
The success score of a tool agent reflects both its model and the budgets of the harness that runs it. We present ToolBottleneckBench, a benchmark that measures tool-agent performance under controlled pressures and fixed harness budgets. Synthetic enterprise tasks stay fixed while one pressure varies: the registered tool catalog, the visible tool menu, the certified workflow, or the encoding of returned records. An exact grader records the first unmet obligation of each failed episode and the harness limit, if any, that ended it. Across 21 model and harness configurations, the visible menu is the only one of the three pressures that act on the model with a large penalty, and this penalty tracks truncation at the completion cap. Raising the cap sixteenfold removes most of the penalty in the truncation-heavy configurations. Most failures end at a harness limit, and the failure record identifies the assigned pressure no more reliably than a count of tool calls. The repairs that help change which tools reach the model or who issues the calls, and only two repairs, neither improving model behavior, beat every configuration switch. We recommend reporting pressure effects with the harness budgets under which they were measured. We release the benchmark at https://github.com/zichuanxu/tool-bottleneck-bench.
agentbenchmark - arxiv:2610.21095 · cs.CVCan World and Action Modeling Benefit Each Other under Unified Causal Training?Yue Su, Bokai Lin, Xinkai Wang, Chubin Zhang +8
World and action modeling are closely related, yet their reciprocal benefits under unified causal training remain unclear. We introduce OneWM, a simple framework for unified causal training that treats actions as both conditions for world prediction and targets for policy learning within a shared video-action architecture. Using this framework, we investigate whether learning forward dynamics conditioned on specified actions improves action learning, and whether learning to generate actions in turn improves action-conditioned world modeling. We use LIBERO, LIBERO-Plus, and RoboCasa365 as embodied testbeds to evaluate both policy learning and world modeling, complemented by real-world policy experiments. Learning the visual consequences of specified actions improves policy performance and generalization across changing scenes. In turn, action learning improves multi-view world-prediction fidelity under the same architecture. Together, these results answer both questions positively, supporting reciprocal benefits between world modeling and action generation under unified causal training.
embodiedliberoworld modelaction-conditioned - arxiv:2610.21091 · cs.AIMemComposer: Learning to Assemble Memories for LLM Agents via Reinforcement LearningYixuan Xie, Xueqiang Xu, Xinyi Fan, Xinxi Lyu +7
Agent memory enables large language model (LLM) agents to reuse past experiences without updating their parameters. However, semantic similarity alone does not guarantee memory usefulness: downstream performance depends on how memories are selected and presented to the agent. We propose MemComposer, a memory composer that jointly examines the current query and retrieved candidates to decide which experiences to include, how many to select, and how to order them. We train the composer using Experience Policy Optimization (ExPO), a GRPO-based framework that evaluates assembled memory contexts through complete agent interactions and uses task-success feedback to optimize the assembly policy. Across three $τ^2$-bench domains with two frozen executors, MemComposer outperforms the strongest baseline for each executor while reducing average memory-token usage by 16.9% and 4.9%, respectively. The learned policy also transfers to unseen domains, executors, and memory banks, and benefits from memories generated by stronger models without further training. These results demonstrate the value of learning query-specific memory assembly using downstream task feedback.
memoryagent memoryagentllm agent - arxiv:2610.21090 · cs.CVRoboCloth: A Large-Scale Real Cloth Material Dataset for Neural Reflectance ReconstructionZhen Li, Haoran Qin, Francesco Di Sario, Daniel Rebain +1
We present RoboCloth, a dataset of 500 real cloth materials captured under flexible, calibrated camera and lighting configurations, where both are robotically controlled during acquisition. Each material is imaged in approximately 580 high-dynamic range frames spanning a wide view-light hemisphere; camera poses are refined via structure-from-motion and aligned to the robot frame, yielding per-material sparse surface geometry and structured per-point reflectance observations ready for neural material training. We demonstrate the value of this data through a two-stage neural reconstruction pipeline: Stage 1 trains a shared MLP decoder on our corpus to learn cloth material priors; Stage 2 then freezes the decoder and fits a dense 2048x2048 latent texture to each target material. Evaluating on held-out materials from our own dataset and on the independent UBO2014 benchmark, we show that a decoder trained on our captures outperforms decoders trained on existing measured reflectance datasets and analytic PBR fitting, transferring to unseen acquisition systems without modification. The dataset, calibration files, observation tensors, and train/test splits are released under CC BY 4.0 on the Hugging Face Hub.
benchmark - arxiv:2610.21088 · cs.LGProbe-and-Resume: When and How to Evaluate Agent ExecutionJiakai Xu, Tiancheng Ge, Tujie Guan, Yusheng Li +3
Test-time search can improve agent performance by exploring alternative actions and using a critic to select promising branches. We study this approach in terminal-based computer-use environments, which expose only a partial view of a large underlying state, and find that the Monte Carlo tree search (MCTS) configuration we evaluate does not significantly improve task success. We identify two sources of poor critic performance: the agent's observations omit environment state needed for evaluation, and intermediate states offer limited evidence of whether a trajectory will succeed. At intermediate states, critics rank continuations only modestly better than chance, and environment access improves this little. We study when to invoke a critic and what access to give it to evaluate a node. Our findings motivate a policy that invokes the critic only at nodes the agent believes are final, rather than during partial progress. Further, the policy gives the critic read, write, and execute access to a fork of the computer's current and initial state, allowing it to probe and perform potentially destructive actions to surface any hidden state the trajectory omits. We call the resulting design Probe-and-Resume: when the agent claims to be done, the critic probes a fork of its state, and a rejected agent resumes where it stopped. On Terminal-Bench 2.1, this raises success by 8.4 percentage points over MCTS with conventional critics, and approaches the success rate of the Codex CLI agent with the same underlying model at less than half its cost.
agent - arxiv:2610.21086 · cs.RORoPE-Flow: Monocular Vision-based Robot Pose Estimation via 2D Graph-Conditioned 3D Flow MatchingLiangjing Shao, Wanhao Liu, Zhiwei Fang, Rulin Zhou +5
Camera-to-robot pose estimation from a monocular image is significant for vision-based robot navigation and control. However, recovering 3D structural features from 2D observations remains a challenging task. Existing camera-to-robot pose estimation methods either rely on time-consuming iterative rendering or involve sophisticated 2D-3D registration which suffers from projection ambiguity, accumulated errors for distal joints and imbalanced learning for different joints. In this work, we propose RoPE-Flow, an end-to-end flow matching-based framework for camera-to-robot pose estimation. From the predicted 2D heatmaps, we propose a Keypoint-centered Feature Extractor to obtain positional features along with contextual representations from keypoints-centered multi-scale patches. Moreover, we introduce Graph-guided Condition Generator, which embeds a learnable adjacency map to generate condition for flow matching based on correlation among keypoints. Besides, we introduce Bone Length Loss and Ray-Depth Loss for topological consistency and scale stability, with joint-wise weighting to mitigate varied prediction difficulty across joints. The proposed method outperforms state-of-the-art methods for pose estimation across different types of robots with mutiple cameras. Besides, we establish a novel cross-scene benchmark to demonstrate the outstanding performance of RoPE-Flow on zero-shot cross-scene robot pose estimation. Meanwhile, animated visualizations are shown on the website: ropeflow.netlify.app
benchmark - arxiv:2610.21085 · cs.AIA Jury, Not an Electorate: Label-Free Reliability Estimation for Ensembles of Language ModelsAlessandro Rozza, Davide Mambelli, Steven Motta, Juri Taresch
An ensemble of language models is a jury, not an electorate. The answers are testimony from witnesses of unknown reliability, and the question is not what the majority said but who is credible. Most ensemble methods assign each expert a fixed reputation, often learned from labels. We ask whether it can instead be inferred from the pattern of agreement and disagreement among the experts. We model the experts as a jury: each question has a latent true answer, each expert reports it with unknown reliability, and those reliabilities are estimated by Expectation Maximisation (EM) from unlabelled answer configurations alone. We show that its decision rule reduces to a log-odds weighted linear score over the experts' answer distributions. What changes is not how the jury votes, but how it learns whom to trust. On a 1560-question benchmark, the jury, with nine experts of up to 3B parameters, performs comparably to the same fusion rule with label-fitted weights, showing no statistically significant advantage from supervision. Instead, it outperforms the best single model with matched self-consistency and the strongest label-free alternative on the full benchmark.
benchmark - arxiv:2610.21082 · cs.AISci$Π$vot: Benchmarking Principal-Investigator Judgment at Pivotal Moments of Scientific AutoResearch AgentsYouran Sun, Xingyu Ren, Chugang Yi, Kejia Zhang +14
Large language model (LLM) agents can keep scientific experiments running for months with little human supervision, but reliable autonomous research also requires scientific judgment. Evaluating scientific judgment is difficult because research decisions are open-ended and depend on project context. We construct Sci$Π$vot, a 248-case benchmark of scientific judgment based on human corrections. The benchmark draws on records from six months of research in which senior researchers ran LLM agents on 26 projects across six scientific fields and repeatedly intervened to correct the agents' research judgments. In each case, an agent chooses between the original LLM plan and the human-corrected alternative, using only the project record available before the correction as context. We compare LLM and human performance and examine the effects of reasoning effort, access to the project record, and reconsideration. At medium reasoning effort with the project record, models systematically prefer the original LLM plan over the human-corrected alternative. Four of five frontier models perform below chance on the selected cases. The highest accuracy in the project condition across the tested models and reasoning effort settings is 63.5%, compared with 73.4% for fresh researchers. At medium effort, every model achieves lower accuracy with the project record than without it, and reconsideration does not consistently improve overall accuracy. Even with a human-corrected alternative available, frontier models struggle with the selected research decisions, while Sci$Π$vot supports systematic evaluation of scientific judgment for LLM agents.
agentllm agentbenchmark - arxiv:2610.21074 · cs.AIOmicsBench: a long-horizon multi-turn benchmark of AI agents on research-study-scale bioinformaticsGideon Shaked, Siddharth Nagisetty
Computational biologists increasingly delegate substantial analyses to AI agents in long-horizon, multi-turn sessions, yet most benchmarks of agents' bioinformatics capabilities pose a single short task with a single-shot prompt. We present OmicsBench, a long-horizon, multi-turn bioinformatics benchmark for AI agents. Each of its 97 tasks links dependent subtasks into a directed acyclic graph, for 662 underlying subtasks in all. Each subtask is posed as one user turn and graded deterministically against an answer key, with optional in-context feedback on incorrect answers. Across 12 models on three harnesses, the slowest agent averages 3.6 hours per task, and no combination comes near saturating the benchmark, with mean scores of 0.43 to 0.57 out of a possible 1 without in-context feedback and 0.54 to 0.67 with it. Multi-turn evaluation also uncovers behavior that single-shot prompts cannot, as agents score lower on subtasks that depend on earlier subtasks, and in-context feedback helps agents recover some of that loss. Because deterministic scoring gives no qualitative account of how an agent failed, we also introduce an unsupervised failure mode discovery pipeline that describes, embeds, and clusters agents' failed attempts to surface emergent failure modes. Applied to OmicsBench, this pipeline exposes qualitative shortcomings in how agents carry out bioinformatics analyses, ranging from agents applying their own statistical filters in place of the specified ones to agents writing whole answers from memory.
agentai agentbenchmark - arxiv:2610.21063 · cs.AIApprove What, Exactly? Effect Binding and Trusted Rendering for Agent Step-Up AuthorizationSai Varun Kodathala, Prashanth Pollishetty, Jaylen Cargill
AI agents now pause for approval before they send money, delete files or email outsiders, and that approval is usually treated as a second factor. We argue it is one only when compromising the agent cannot produce it and when it names exactly what it approves, and we test both conditions. We define twelve attacks on the approval path, including decoy recipients, amount cancellation, card padding, label spoofing, enrollment squatting and threshold splitting, and run them against the Agent Payments Protocol (AP2), the Microsoft action-bound approval design and Model Context Protocol (MCP) elicitation. An audit of ten agent runtimes finds no third-party verifiable approval record, no flood limits where measurable, and post-approval argument changes in four of eight runtimes with measurable binding. As a reference defense we build ASU (Agent Step-Up), a gateway that renders the approval card from a canonical effect descriptor, has a passkey on the phone of the user sign its digest, and executes only the stored call. Tamarin proves its binding, single-use and key-level non-repudiation properties under three stated hypotheses (53 lemmas). Across 27,108 scripted trials and 609 AgentDojo attack pairs, ASU stops substitution, replay and grant stretching, and the trusted card cuts forged-narration attack success from 0.54 to 0.03. With 15 current language models as automated approvers, the trusted card raises the attack catch rate from 0.30 to 0.98, at the cost of refusing 16 percent of harmless near misses, mostly due to a placeholder artifact in our stimuli. A second device adds nothing measurable over a trusted on-host confirmation in these experiments, flooding still succeeds 0.45 to 0.60 of the time against a modeled approver, and approval cannot stop effects that policy already allows. Code, data and proofs: https://github.com/Sports-Vision-Inc/agent-step-up
agentai agent - arxiv:2610.21055 · cs.LGAthanor: An Agent-Native System for Autonomous Post-Training ResearchYufa Zhou, Yichen Wang, Jiakun Fan, Ziqi Shu +4
Recursive self-improvement, in which AI systems build stronger AI systems, is one promising route to superintelligence. We study a weaker form that can be measured today: frontier agents post-training a small raw base model with public data, as a test of their research potential. We build Athanor, an open agent-native system for autonomous post-training research (AutoPostTrain) on real GPUs. We give the agents full research autonomy, low-friction engineering tools, and efficient starter code. The agents carry out the research themselves: they read datasets and evaluation results, build their training data, write and modify training algorithms, tune hyperparameters, and analyze the results to decide what to try next. They work in a sandbox and cannot use external teacher models. Athanor's services supply frozen evaluation, managed jobs, decontaminated data, a persistent research record, and a held-out suite the agents are unaware of. In a 104-hour campaign, a team of four Claude Opus 5.5 agents, each on an eight-GPU node and given no research advice, improves Qwen3-4B-Base from 15.9 to 67.3 on a six-task development suite and from 9.2 to 48.2 on a held-out suite, against 66.5 and 47.9 for the official Qwen3-4B-Instruct-2507. To our knowledge, this is the first demonstration that a team of frontier agents can autonomously post-train a raw base model into a cross-domain generalist model that matches or surpasses the official instruct version on average. The campaign's full record shows that the agents reach this level by recombining known methods and each other's models: SFT on public reasoning data, RL with a rollout change one agent introduced, weight merges across agents, and rounds of self-distillation. With execution handled, today's agents already have the engineering and research ability to conquer this task. We hope Athanor provides a foundation for scaling autonomous AI research.
agentself-improvementpost-training - arxiv:2610.21048 · cs.LGSHIP: Structured and Hierarchical Insight with PrivacyKai Chen, Guisi Ni, Albert Cheu, Kassem Fawaz +3
Extracting insights from textual data while protecting user privacy is increasingly important. Existing methods produce private insights in diverse forms for different analytical purposes, making their utility difficult to assess and compare. In this work, we formulate differentially private (DP) insights as reusable representations for answering analytical queries, providing a common basis for utility assessment. Answering these queries, however, requires fine-grained details and cross-attribute statistics that existing representations may omit. To capture this information, we introduce Structured and Hierarchical Insight with Privacy (SHIP), which automatically maps queries to attributes, builds a hierarchical hybrid taxonomy, and releases a DP synthetic table for further analysis. We also develop an evaluation protocol based primarily on QA performance and structured fidelity. Our experiments demonstrate SHIP's strong performance on both measures relative to existing private insight baselines, while identifying queries that remain challenging for structured representations.
evaluation protocol - arxiv:2610.21046 · cs.LGPremiseCheck: Efficient Fact-Checking of Human and AI Peer ReviewsDingjie Song, Zhiling Yan, Ziqian Zhang, Yuxuan Zhang +1
Peer review is under strain from the growth of submissions, and review comments written by humans or AI can misstate what a paper says. We introduce ReviewFact, a benchmark that labels such factual errors in review comments under one definition across NLP, machine learning, and the natural sciences, and that measures cost and latency alongside accuracy. A frontier LLM checks comments well but must read the whole paper for each check, while cheap decision models, which score whether a statement holds in a text, fall far short on whole comments that mix several implicit assertions. PremiseCheck is a framework that combines the two. An LLM rewrites each comment as explicit premises without reading the paper, and a decision model checks each premise against the paper. When each comment is checked on its own, PremiseCheck reaches accuracy comparable to GPT-5.5 across all three fields at about half its cost, and on NLP reviews at a quarter of its cost and less than half its latency.
benchmark - arxiv:2610.21045 · cs.AILearning When to Withdraw: Curriculum Coverage and Rule-Specific Failures in Fine-Tuned Language AgentsMohamed Abouzahra
Tool-using language agents must revise prepared actions when evidence changes or authority is revoked. Selective withdrawal means suspending affected actions, preserving supported work, and resuming only after sufficient repair. We study supervised fine-tuning in NAQD-Env, a synthetic environment with explicit dependencies and a deterministic reference policy. Four LoRA checkpoints of Qwen2.5-3B-Instruct compare an original curriculum with an enriched one, the latter under three system-prompt framings with identical views and targets. Three instruction-tuned baselines and all four checkpoints are evaluated under three inference prompts on fresh validation cases, held-out dependency families, and held-out structural combinations. Under the naqd prompt, SFT checkpoints reach decision agreement of 0.77-0.92 and withdrawal recall of 0.46-0.86, versus 0.45-0.63 and at most 0.06 for the baselines. Rule-level diagnostics qualify these gains. The original-curriculum checkpoint wrongly withdraws 26 of 34 actions with surviving alternative support; enriched checkpoints withdraw 0-1. Conversely, all SFT checkpoints narrow prohibited combinations less often than the 7B and 8B baselines, a case absent from both curricula. Delegation results show more withdrawal but poorer preservation of independently authorized actions, and all four checkpoints omit valid event reports. Framing contrasts vary across metrics and splits and do not establish a reproducible advantage or equivalence. One training run per arm, jointly changed curriculum properties, prior evaluation exposure, and baseline-SFT quantization differences limit causal attribution. The contribution is a method for detecting rule-specific regressions that aggregate accuracy can conceal, not evidence that fine-tuning alone ensures agent safety.
agent - arxiv:2610.21042 · cs.AIReviewPilot: A Multi-Agent System for Screening and Data Extraction in Systematic ReviewsZhaoqian Xue, Shaoze Zhou, Xuexin Li, Junting Mu +9
Systematic literature reviews follow the PRISMA process of screening scholarly works and extracting evidence from included studies. Deep research tools based on large language models (LLMs) can search and summarize the literature but do not record which studies are excluded or where extracted values come from. We develop a conversational multi-agent system called ReviewPilot that automates the PRISMA workflow from paper screening to information extraction. Researchers can tailor criteria and schemas through dialogue, and a project memory keeps their configurations and revisions. Each screening decision comes with its reason and each extracted value with its source. On epidemiological expert reviews, ReviewPilot outperforms a single-prompt LLM and AgentSLR across three LLMs, with screening macro F1 up to 0.776 and extraction F1 up to 0.788. Case studies in biomedicine and human-computer interaction show close agreement with human annotators, suggesting that ReviewPilot can transfer beyond epidemiology. Code is available at https://github.com/ZhaoqianXue/ReviewPilot.
memorymulti-agentagent system - arxiv:2610.21040 · cs.AIRecursive Reward Improvement via Adversarial Meta-Prompt EvolutionWei Dai, Qianchu Liu, Guanghui Qin, Xing Han +3
Recursive Self-Improvement (RSI) has gained wide attention for its ability to work with limited data. In a proposer-solver loop, a proposer model writes training tasks and a solver model learns to solve them through reinforcement learning. Such loops have succeeded in mathematics and coding, where every answer can be checked automatically. Open-ended expert tasks, like clinical report generations, have no answer key, which makes simple question-answer generation not viable. Here the proposer must also write a rubric that lists the content an answer should include and the errors it should avoid. However, the unique challenge behind model-written rubric is the comprehensiveness of it. For naively generated rubric, solvers tend to exploit it, naming the expected topics without the substance behind them, which corrupts the learning process. To solve this problem, we introduce RRI, Recursive Reward Improvement, which uses a reward hacking adversary to supervise the reward. Before the solver trains on a generated task, an adversary writes an answer that games its rubric while being as useless as possible, the same model writes an honest answer without seeing the rubric, and up to two repair steps revise the rubric to make the honest answer win. The exploits found also rewrite how future tasks and rubrics are generated. On HealthBench Professional, RRI improves the adjusted score over the Generation Only self-improvement baseline by 10.9 points at 9B and 12.4 at 27B, and by 10.3 points without any external model. In particular, the gains are largest on difficult and writing tasks. We show that the 27B model trained on the HealthBench Professional description improves by 6.6 points on PRBench Hard and 8.0 on ProfBench without seeing a task from either. We further analyze the exploitable points of model-generated question-rubric pairs to facilitate future research on open-ended self-play.
self-improvementself-play - arxiv:2610.21038 · cs.LGINCONVENIENT: A Benchmark to Assess LLM Generalization Capabilities from Convenience SamplesSara Rezaeimanesh, Chad Klochko, Kundan Thind, Mohammad Ghassemi
Machine learning models for tabular prediction typically assume that training data are representative of the population on which they will be deployed. In practice, however, data are often collected based on convenience, availability, or cost, resulting in systematic sampling biases that can distort feature--target relationships and undermine real-world generalization. Existing tabular distribution-shift benchmarks primarily construct out-of-distribution (OOD) settings by partitioning observations across naturally occurring domains. While such partitions may induce changes in P(Y | X), the resulting shifts are neither explicitly controlled nor isolated, and the evaluation distributions themselves are often not representative of the broader target population. We introduce INCONVENIENT, a benchmark designed to study robustness to convenience-sample bias. Rather than relying on naturally occurring domain partitions, INCONVENIENT directly perturbs feature--target relationships in the training distribution while preserving the original population distribution as the deployment environment. This setting more closely reflects real-world scenarios in which models are developed from unrepresentative samples and subsequently deployed on a broader population. The distinction is particularly relevant for large language models (LLMs), whose pretraining may encode population-level knowledge that extends beyond the relationships observed in a biased training dataset. Unlike existing distribution-shift benchmarks, INCONVENIENT provides explicit control over both the direction and severity of relationship shifts, enabling a systematic assessment of robustness to convenience sampling. The benchmark comprises 200 tabular classification and regression datasets spanning diverse application domains and generates around 15,500 shifted training distributions which lead to substantial downstream degradation.
benchmark - arxiv:2610.21030 · cs.CLLearning When to Think: Utility-Guided Adaptive Reasoning for Efficient Text RetrievalYingtao Luo, Kangyi Zhao, Shiyi Du, Yubo Li
Reasoning-conditioned embeddings can improve retrieval by progressively refining semantic constraints. However, additional reasoning may introduce over-generalized or semantically noisy abstractions that weaken retrieval specificity. Existing methods largely supervise reasoning through final retrieval outcomes, providing limited guidance over the refinement process itself. We propose AURA, a framework that directly supervises intermediate reasoning states according to how well they separate relevant documents from hard distractors. AURA further trains a lightweight utility estimator over prefix-conditioned embeddings to adapt reasoning depth at inference time. On the reasoning-intensive BRIGHT benchmark, AURA-4B achieves 34.7 average nDCG@10 while using 2.8 reasoning steps and 96 generated tokens per query. Under matched deployment conditions, adaptive stopping reduces latency from 4.6 to 0.76 seconds relative to full-trajectory reasoning. Results on MS MARCO further suggest that this specialization need not substantially degrade conventional in-domain retrieval.
benchmark - arxiv:2610.21029 · cs.AISCARCE: Emergent Institutions in Long-Horizon Multi-Agent Coding SystemsAbhinav Kumar, Rediet Abebe, Eugene Bagdasarian, Sahar Abdelnabi
As autonomous agents become widely deployed, coordinating access to shared computing resources and services will become increasingly important. Recent incidents have already revealed unplanned coordination between agents, highlighting the safety implications of these interactions. Yet prior studies of agent negotiation often use stylized games and simulated bargaining, leaving open how these behaviors develop during sustained, long-horizon tasks. We propose SCARCE (Shared Compute ARena for Coding agEnts), a framework in which agents solve sequences of coding tasks under individual deadlines while negotiating access to a shared execution resource. This negotiation determines when and for how long they can inspect programs, compile code, and run tests. We find that agents spontaneously develop informal institutions: persistent rules and commitments grounded in reputation, reciprocity, and precedent. These arrangements, which we call institutional behaviors, carry across tasks and influence resource allocation, with agents sometimes foregoing immediate advantages to honor earlier agreements. We develop a 21-category taxonomy of negotiation strategies and emergent norms and use it to analyze eight models and audit more than 1,500 trajectories across different environmental settings. Among the models tested, newer frontier models exhibit more institutional behaviors than older or lightweight models. Alongside cooperation, we observe exclusionary coalitions and rare divide-and-conquer strategies. SCARCE provides a controlled setting for proactively and systematically auditing how these relationships emerge, shape access to resources, and affect the safety and coordination of multi-agent systems.
agentautonomous agentmulti-agentagent systemarena - arxiv:2610.21018 · cs.AIFailure Is the Mother of Success: Recovering Supervision with Grounding HindsightSheldon Yu, Tong Yu, Nikki Lijing Kuang, Zhendong Chu +5
When every rollout for a problem fails, GRPO receives neither a successful exemplar nor a nonzero group-relative advantage, so it learns nothing from exactly the problems it most needs to learn. The failed reasoning can nevertheless suggest what to try next. We present Replay-Selected Hindsight Distillation (RSHD), which turns failed attempts into an additional, tested source of supervision during GRPO training. The policy writes a corrective note about each failed attempt without seeing a reference answer or solution, then re-solves the problem with the note in context. A note is admitted only if its replays succeed more often than the original group did, after screening for notes that state the answer, and an admitted replay is distilled into the plain-prompt policy. Because admission is decided by fresh successes rather than by the failed rollout itself, the procedure can recover supervision from groups in which every attempt failed, and it applies equally to failures in mixed groups. Only the policy is trained, the only judge is the outcome verifier, and deployment requires no notes or extra inference. Across model families and sizes and two training domains, RSHD consistently improves over GRPO on in-domain benchmarks, and the gains carry over to out-of-domain evaluation sets.
benchmark - arxiv:2610.21016 · cs.LGCausalEvo: Causal Skill Self-Evolution for Health LLM AgentsYuchen Zhang
Large language model (LLM) agents for health can accumulate reusable textual skills from task feedback, such as calculation criteria, measurement time points, and formula units. Deciding which skills to retain is consequential: useful rules improve later answers, while faulty generalizations propagate errors. We propose CausalEvo, a method for causal skill self-evolution that estimates the relative effects of deliberately changed skill sets while holding the model weights, feedback cases, and rule count fixed. At each evolution step, a frozen LLM proposes candidate rules, which are pooled with the incumbent skills so that the current skill set remains a feasible choice. Paired skill replacements against shared background rules are scored on the same feedback cases; least squares converts the score differences into relative skill values, and the top-ranked rules are retained, favoring incumbents on ties. A decision-loss bound separates candidate support, approximation, and estimation errors. On six MedCalc-Bench-Verified tasks with 120 held-out cases, CausalEvo achieves 60.00% accuracy, compared with 55.83% for SkillOpt and 51.67% without skills. Across the recorded training configurations, which use different optimization budgets, it consumes 55.6% fewer training tokens than SkillOpt. Gains concentrate in particular calculations alongside regressions on others, and paired accuracy intervals include zero. CausalEvo turns task feedback into controlled decisions about skill retention and replacement. Project repository: https://github.com/txbrook/CausalEvo
llm agent - arxiv:2610.21011 · cs.LGAudio-Visual Dog Identification and Tracking in Real-World VideosAryahvishwa Babu, Kingsley Jin, Lillian Wang, Tuan M. Dang +4
Identifying individual dogs from barks in real-world videos remains challenging and largely unexplored. Solving this problem would enable long-term behavioral analysis, veterinary monitoring, and animal welfare monitoring by attributing vocalizations and behaviors to the correct individual. Real-world videos are challenging because multiple dogs may bark simultaneously, become occluded, or move in and out of view. Prior studies show that canine vocalizations and visual appearance contain discriminative identity information, but these modalities have largely been studied independently. We present a multimodal dog-fingerprinting framework combining visual identification, multiple-object tracking, and canine vocal fingerprinting to associate bark events with individual dogs while maintaining persistent profiles. The framework synchronizes bark events with tracked dogs and updates profiles using visual appearance and vocal characteristics, enabling identification despite occlusion, overlapping vocalizations, and off-screen barking. We also introduce a benchmark of 46 real-world videos containing 213 manually annotated bark events. Our framework correctly localizes the active visible dog in 61.83% of visible events at IoU $\geq 0.30$ and achieves a 91.33% F1 score for visible versus off-screen barking detection.
benchmark - arxiv:2610.21007 · cs.CVMind the Tail: Data Selection Through Sparse ConceptsZiheng Zhang, Neha Hulkund, Colin Lee, Zheda Mai +6
Data selection reallocates training support across visual concepts, including objects, attributes, color, texture, and backgrounds. Different allocation strategies may produce varied downstream model behavior, especially for tail concepts supported by only a few samples. Discarding them can hurt test-time robustness, while overemphasizing samples lying on category boundaries may introduce unstable training signals. This work explicitly investigates this relationship across fine-grained, spurious-correlation, and general image classification, as well as multimodal pretraining. Specifically, we use sparse autoencoders (SAEs) to extract concept activations from individual images. During selection, we accumulate the dataset-level concept support for each added sample. Applying diminishing returns to this distribution down-weights well-represented concepts and shifts the focus to the distribution tail. We then incorporate local support to find samples that reliably strengthen under-supported concepts. Empirical results show consistent improvements over strong data selection baselines across all benchmarks. We also reveal the connection between concepts our method enhances and the induced behavioral changes, including improved robustness against out-of-distribution and spurious contexts. This work establishes concept support, beyond individual sample importance, as a useful principle for data-efficient and robust learning.
benchmark - arxiv:2610.21001 · cs.LGImperfect Is Perfect: Learned Column Prefilling for Branch-Price-and-CutHaoran Liu, Yu Yang
Column generation folklore warns that initial columns hurt: a warm start may bias the dual and slow the root of branch-price-and-cut. We show the opposite once the columns are imperfect in the right way. Preloading the support of the optimal root LP, the best columns one could know in advance, fails to finish on half of our hardest instances; a pool of ordinary routes that need not belong to any optimal solution, held outside the master and priced at the current smoothed dual, speeds the root up. Because the columns need not be optimal, producing them becomes a learning task: a neural edge scorer guides a constructive generator that builds the pool in seconds. A second network predicts the initial center of Wentges smoothing from the instance file, and with it the direction of the smoothed pricing queries; the solver's own updates run unchanged. With prefill and center in place, exact pricing dominates root pricing time, so we rebuild bucket-graph labeling around its dominance test for the GPU and certify on the CPU. Inside RouteOpt, the state-of-the-art exact solver for capacitated vehicle routing, the pipeline reaches the same certified root bound in 13.2% of vanilla RouteOpt's root pricing time over 95 instances with 300 to 1000 customers, prefill alone in 26.1%, and 17.5 times faster in the best case. In an exact branch-price-and-cut solver for unrelated parallel machine scheduling, prefill with smoothing and the learned center certifies all 25 benchmark instances, 12 times faster than the solver's own column generation, which certifies 20 of them.
benchmark - arxiv:2610.20998 · cs.ROMoCaP: Mobile Code-as-Policies for Zero-Shot Mobile Manipulation from Passive Human VideosYiyang Cai, Tatsuya Kamijo, Ninghao Lu, Zijin Hu +1
Passive human videos of everyday activity offer robots access to a massive amount of human experience without requiring a demonstration for each task. However, using this experience requires recovering knowledge from unstructured video recordings and adapting it to a different body (i.e., the robot). To overcome these challenges, we present mobile code-as-policy (MoCaP), an agentic framework that turns passive human video recordings into guidance for an end-to-end mobile manipulation policy. MoCaP extracts spatial memory and cues of how people prepare to interact with objects, which inform navigation and the search for a manipulation-ready configuration. Guided by passive human experience and current observations, MoCaP uses a closed-loop code-as-policy to coordinate planning and tool use primitives to solve motions under the robot's geometry and kinematic constraints. On a real mobile manipulator, MoCaP achieves 80% success on each of two mobile manipulation tasks in an indoor environment whereas existing approaches cannot succeed at the task. Video-derived spatial memory improves object-navigation success from 70% to 100%, and human approach prior raises the rate of finding a reachable base pose from 0% to 100% at an unfavorable start location. Videos and code are available on the project website: https://mocap-7.github.io/MoCAP-page/.
manipulationmanipulatormemoryagentictool usecode-as-policy - arxiv:2610.20993 · cs.LGOne Bias Score, Two Factors: How Mitigations Affect Option Bias in Multiple-Choice EvaluationHyunjong Ok, Minjae Park, Jaeho Lee
Reordering multiple-choice options can change a model's answer, an effect studied as selection bias (a preference for certain option letters) and as order sensitivity (changes in answers across option orderings). We connect these perspectives through the pooled selection-rate standard deviation (pooled SD), which measures how unevenly option letters are selected across questions and orderings. When every option of each question appears equally often at each position, the pooled SD is exactly the product of two label-free factors: the flip rate and the positional skew. The flip rate is the fraction of questions whose chosen option changes across orderings; the positional skew is the standard deviation of option-letter selection rates among those questions. Across 41 checkpoints and 12 benchmarks, the flip rate tracks accuracy and output uncertainty, while the positional skew tracks the estimated option-symbol prior. Prior correction for selection bias mainly reduces the positional skew. Aggregating each ordering with its reversal, a mitigation for order sensitivity, mainly reduces the flip rate. Both reduce the range of accuracy across correct-answer letters by similar amounts, but aggregation leaves fewer model-rank reversals. Applying prior correction before aggregation reduces both factors and further narrows this accuracy range. However, aggregation lowers the flip rate partly by settling questions on the same wrong option. We therefore recommend reporting both factors to distinguish mitigation effects, alongside accuracy to distinguish stability from correctness.
benchmark - arxiv:2610.20990 · cs.CVDeSpatial: The devil of Agentic Spatial Reasoning Lies in 3D ItselfYifei Liu, Xiaolong Zhou, Ziyang Gong, Yifan Zhou +7
Spatial agents have access to reconstruction and segmentation tools, yet inconsistent predictions and fragmented object identities can prevent their outputs from becoming reliable 3D scene information. A simple oracle experiment exposes this perception bottleneck: replacing tool-predicted perception with ground-truth 3D bounding boxes raises an agent to nearly full score across five object-centric VSI-Bench tasks, while retaining the same reasoning tools. Motivated by this finding, we believe the devil of spatial reasoning lies in 3D itself, and introduce DeSpatial(Devil of Spatial), which explicitly decomposes spatial understanding into perception and reasoning. DeSpatial has a shared 3D perception tool to reconcile independent perception outputs into a unified, metric, gravity-aligned 3D scene. This perception tool has two core components: a) a joint segmentation--depth co-refinement module that consolidates masks of the same object and removes projection artifacts caused by inconsistent segmentation and depth within a single frame; b) a cross-view instance fusion module that consolidate voxel-overlapping instances but not allowed to merge same-frame instances. This fusion module additionally fuses spatially disjoint fragments caused by insufficient multi-view observations, by checking their 3D structure and source mask appearance. After receiving the complete and unified 3D scene information, the agent chooses reasoning or computational tools to solve the spatial problem. DeSpatial achieves leading performance on VSI-Bench, MindCube-Tiny, and ViewSpatial, with additional long-video results on VSI-SUPER, revealing the bottleneck of spatial understanding lies in a precise 3D perception itself.
agentagentic - arxiv:2610.20989 · cs.LGDo As I Do, Not As I Say: On-Policy Distillation Inherits the Teacher in Vision-Language ModelsFrank F. Yang, Kefei Yan, Yiren Wang, Yiyang Wang
Improving the reasoning capabilities of vision-language models (VLMs) often relies on post-training with supervision from stronger models. Two common strategies are supervised fine-tuning (SFT), which trains on teacher-generated demonstrations, and on-policy distillation (OPD), which provides teacher feedback on trajectories generated by the student itself. While OPD can be more data-efficient by supervising student-generated trajectories, it is unclear how the teacher's supervision at each state affects the student's learned behavior. To study this question, we conduct a controlled comparison of OPD and filtered SFT across data budgets and tasks where the teacher exhibits different levels of reliability, and analyze how these differences shape student accuracy and completion behavior. Across the two tasks, we find that OPD is effective only when the teacher remains reliable on student-generated trajectories. On ChartQA, where the teacher completes 99% of its solutions, OPD trained on only 10% of the data comes within 1.4 accuracy points of full-data SFT. In contrast, on Geometry3K, where the teacher fails to complete 22% of its solutions, OPD underperforms SFT by 7.9 points, with most of the gap explained by incomplete responses. More strikingly, starting OPD from an SFT student reduces its completion rate from 76% to 62%, effectively erasing the completion behavior learned through SFT. Filtering teacher demonstrations for completeness alone recovers much of SFT's advantage. These results show that on-policy supervision can transfer not only teacher knowledge but also persistent failure modes on student-visited states. Code is available at https://github.com/ffy208/opd-inherits-teacher-vlm.
post-training - arxiv:2610.20979 · cs.LGTwo-Stage Score Estimation for Group-Specific Diffusion ModelsHuitao Yang, Oscar Hernan Madrid Padilla
Generative modeling across related distributions raises a question: how can shared structure support accurate generation for a target distribution with limited observations? We develop a statistical framework for diffusion transfer that analyzes a shared pooled score together with a target-specific offset. Under Besov regularity of the pooled density and a stable hierarchical composition model for the offset, we derive convergence rates that separate the costs of learning these components and identify when a simpler offset yields a faster risk-bound rate than separate fitting. Under suitable boundary and regularity assumptions, we prove that the offset remains bounded and jointly regular as diffusion time approaches zero. These results allow the lower time cutoff to vanish polynomially and provide total-variation guarantees for diffusion models as density estimators. Synthetic simulations and image benchmarks provide a proof of concept for the framework.
benchmark - arxiv:2610.20970 · cs.ROWhen a robot can trust reversing wind to find an odour sourceJunlong Shen, Shunsuke Shigaki
A wind-derived heading can guide a robot towards an odour source even after the wind reverses, but its reliability depends on the plume's transport history. On a quadruped initially facing the source, holding the initial heading reached it in every reversing-wind trial, whereas comparison controllers never settled there. An exact geometric condition links heading correctness to net odour transport, and simulations separate sensing error from departures from nominal flow. Where the alternating wind directions are too close for reversal alone to produce wrong-side headings, the wrong-side rate rises from 1.0% under uniform transport to 4.7% in a stochastic flow. Re-acquiring headings from new odour contact improves simulated search when contact is intermittent. These results identify distinct requirements for wind-guided search: a heading supported by the transport geometry and sufficient intermittent contact to replace it.
quadruped - arxiv:2610.20969 · eess.SYInteger-Free Control Synthesis for Signal Temporal Logic via Ranked ImplicantsArash Bahari Kordabad, Mohammad H. Mamduhi, Sadegh Soudjani
This paper proposes an integer-free method for control synthesis of nonlinear discrete-time systems under finite-horizon Signal Temporal Logic (STL) constraints. Instead of introducing integer variables, we consider the discrete choices of the specification in a ranked order and solve a standard nonlinear program for each selected choice until a satisfying trajectory is found. We first define implicants as sets of timed atomic constraints whose joint satisfaction is sufficient to satisfy the specification. A reference trajectory obtained from a smooth approximation of the STL robustness is used to rank these implicants. Fixing one implicant turns the problem into a standard nonlinear program with no auxiliary weights or bilinear terms. If the solution does not satisfy the specification, the next implicant will be considered. We prove soundness and completeness of our method relative to the nonlinear programming solver. Numerical evaluation on nonlinear multi-agent systems shows that our proposed method identifies satisfying trajectories where the exact continuous reformulation fails, and moreover, it is faster than a mixed-integer formulation of the same specification.
multi-agentagent system - arxiv:2610.20968 · cs.ROSOMA: Stack-of-Objects Manipulation using a Quadruped RobotStergios E. Bachoumas, Panagiotis Artemiadis
Executing heavy-duty physical manipulation in unpredictable environments remains a major challenge for quadrupedal robots lacking dedicated manipulators. This paper presents SOMA, a hierarchical reinforcement learning framework for whole-body non-prehensile manipulation of multi-object stacks using a quadruped robot. Unlike traditional methods restricted to two-dimensional planar pushing, SOMA integrates explicit robot base-height control, unlocking full three-dimensional pushing strategies. This capability enables the quadruped to manipulate complex payload stacks as a unified system while preventing relative sliding or tipping of upper layers. The architecture factorizes control into a high-level manipulation policy using a Simplified Structured State Space Sequence (S5) encoder to reason over long-horizon temporal history and predict bounded target tracking commands via Beta distributions, and a low-level locomotion policy that dynamically adapts leg stance configurations. In extensive simulation benchmarks, SOMA achieves a $90.3\%$ success rate for single-object reposing and up to $90.1\%$ for dynamic two-box stacks under stringent error margins ($0.05\text{ m}$ position error, $1^\circ$ orientation error). Finally, we validate zero-shot sim-to-real transfer on a Unitree Go1 quadruped robot across diverse payload masses and geometries, demonstrating active base-height adjustments to maintain traction and balance during heavy pushing tasks.
manipulationmanipulatorquadrupedsim-to-realbenchmark - arxiv:2610.20965 · cs.CVMSP: Modality Self-Play for Training Multimodal LLMs Without Paired Cross-Modal DataRobert Moseley, Md Kaykobad Reza, Ameya Patil, Edward Ayrapetian +1
Large multimodal models have achieved remarkable progress by scaling model capacity and training on massive paired multimodal data. However, this paradigm introduces a fundamental bottleneck: aligned multimodal datasets are expensive to curate, difficult to scale across diverse modality combinations, and increasingly constrain further progress. We introduce Modality Self-Play (MSP), a training framework for multimodal foundation models based on proxy-mediated modality composition. Instead of relying on paired cross-modal data, MSP uses anchor-supporting proxy textual descriptions of unseen modalities as semantic bridges during training. Given an observed modality (e.g., image or audio), the model generates proxy descriptions for complementary unseen modalities, which are then encoded and aligned with the observed modality through learned bridges, enabling cross-modal interaction without real paired non-text supervision. Our central hypothesis is that proxy text, combined with contrastively trained pretrained encoders, provides sufficient semantic structure for multimodal composition to emerge. We show that MSP enables zero-shot modality composition: models trained only with a real anchor modality and accompanying proxy descriptions can integrate and reason over multiple modalities at inference time despite never observing real paired non-text data during training. Empirically, MSP achieves state-of-the-art or competitive performance on multiple multimodal benchmarks. Ablation studies further demonstrate that learned bridges improve alignment, while anchor-supporting proxy descriptions enable effective cross-modal composition. MSP removes the need for real paired non-text (e.g., audio-vision) supervision, while continuing to rely on modality-text pairs, pretrained contrastive encoders, and modality-distribution bridge calibration.
self-playbenchmark - arxiv:2610.20964 · cs.CVVELD-QA: A Large-Scale Dataset for Visual Evidence Grounding and Question Answering in Long DocumentsZhivar Sourati, Marianne Menglin Liu, Rongguang Wang, Zheng Wang +11
Existing long-document QA benchmarks rarely combine large-scale human-authored questions with detailed visual evidence annotations, limiting diagnosis of evidence localization and interpretation failures. We present VELD-QA, a large-scale benchmark for Visual Evidence grounding and question answering in Long Documents. It contains 15,000 fully human-authored question-answer pairs across 5,859 real-world PDFs totaling 372,445 pages, balancing chart, table, and complex-layout tasks. Human annotations identify supporting pages and evidence elements and describe cited visual content, enabling separate retrieval and interpretation evaluation. We evaluate nine vision-language models under four evidence settings and seven retrieval-augmented and agentic systems. Our analysis identifies bottlenecks in evidence localization and interpretation. Retrieval-based methods substantially underperform the oracle setting with ground-truth supporting pages; full-document access does not close the gap. Localization alone is insufficient: oracle accuracy peaks at 81.7%, highlighting persistent visual interpretation and reasoning challenges. Enriching inputs with our visual-element annotations narrows the performance gap by 17-24%, demonstrating value beyond evaluation and motivating data enrichment for multimodal document QA. Our code is available at https://github.com/oracle-samples/veld-qa.
retrieval-augmentedagenticbenchmark - arxiv:2610.20960 · cs.AIStepSafe: Learning Timely Intervention with Context-Aware Safety ReasoningSiyi Gu, Kunhan Gao, Chao Peter Yang, Sukanta Ganguly +2
Safety in agentic systems is inherently sequential: whether an action is safe depends on prior actions, tool outputs, information disclosures, and interactions among agents. Recent work has shifted agent safety guardrails from post-hoc trajectory-level assessment toward step-level action diagnosis and proactive risk prediction, yet existing training objectives provide limited supervision for how step evidence accumulates over time and when it becomes sufficient to justify intervention. We propose \methodlong~(ASPO): a training framework for step-level timely intervention. Pre-fix level stage alignment allows crediting each decision at its own step: it compares each response with the other responses sampled at the same prefix and rewards the action that fits the stage, continuing before the risk is visible, asking while clarification can still avert harm, and stopping before harm is irreversible. Adaptive feedback self-distillation further credits the reasoning behind each decision, with the model itself as a teacher that sees the step annotations on intervention steps and rules mined from its own false alarms on safe steps. Its training data carry a violation taxonomy and an intervention window at every step and are extended through self-evolving multi-agent red-teaming. Finally we present \textsc{StepSafe}, a step-level reasoning guardrail with timely intervention, that achieves the state-of-the-art peformance among open-source guardrails across five agent-safety benchmarks on safety-utility pareto curve with Youden's $J$ of 53.0 (4B) against 46.2.
agentmulti-agentagenticself-evolvingbenchmark - arxiv:2610.20959 · cs.AISame Context, Different Agents: A Behavioral Study of Semantic Layers, Orchestration and Compaction in Data AgentsZhivar Sourati, Marianne Menglin Liu, Zheng Wang, Rongguang Wang +5
Answering questions over heterogeneous data means combining evidence from files, databases, documents and the web. Data agents automate this, but their success depends on context. Three context-engineering techniques (CETs) help by shaping it: semantic layers describe the data, orchestration splits work among agents, and compaction condenses interaction history. Existing studies validate each CET alone, leaving open how agent behavior shapes its effect. We present a comprehensive analysis of the three CETs, alone and combined, with seven large language models (LLMs) on 1,514 long-horizon tasks from five benchmarks, covering behavior, accuracy and cost. We find that data agents without CETs form three behavior groups: investigators, premature answerers that answer without inspecting data, and early stoppers that give up. Each CET delivers its known benefit, from more investigation to smaller prompts, but only the behavior group it matches turns that benefit into accuracy, such as premature answerers under orchestration. For investigators, the same CETs can hurt: the semantic layer stops them from looking, orchestration makes them overconfident, costing 5 percentage points, and compaction makes them repeat steps. Combining CETs yields smaller gains and losses than their sum. We derive guidance for choosing CETs by agent behavior, task demands and accuracy-cost trade-offs.
agentbenchmark - arxiv:2610.20947 · cs.LGRanked Right, Sized Wrong: Ensemble Spread Fails Silently at Streamflow ExtremesJustin Gallant, Tyler Wilson
Deep learning models, particularly Long Short-Term Memory networks, are increasingly used for operational streamflow forecasting, with ensemble spread providing a signal of predictive uncertainty. As climate change increases the frequency of hydroclimatic extremes, these models will increasingly operate under flow conditions that are rare in their historical records. We therefore evaluate whether ensemble spread remains informative under increasing flows, examining both its magnitude relative to error and its ability to rank errors. Prediction error is approximately $2\times$ the corrected spread at ordinary flows and up to $4\times$ at the highest flows, with the discrepancy increasing with flow (log-log slope $1.30$, 95% CI: $[1.20,1.41]$). Under-dispersion is strongest in observed extreme flows beyond the training record, reaching an $18\times$ discrepancy. This is pervasive and not explainable by conditional mean bias. Area Under the Sparsification Error, a standard distribution-free metric of uncertainty ranking, describing how well a model knows when it is uncertain, indicates near-oracle ranking for ensemble spread ($0.071$), yet ordering by predicted discharge alone nearly matches it. Informative uncertainty ranking can coexist with severe under-dispersion. In deployment, error ranking cannot be verified, yet the spread is available, and it is relied on as the uncertainty signal under increasingly extreme conditions, while systematically understating risk, most severely for the extreme events that matter most.
memory - arxiv:2610.20941 · cs.AIScrewtape: Adversarial Transcript Editing for Monitor Red-TeamingGosha Kozhevnikov, Tyler Tracy
We present Screwtape, a pipeline that converts real agent transcripts into attacks that evade action-only monitors. An LLM attacker edits a transcript so that the agent causes a specified harm, such as leaking an API key or spawning an unmonitored agent. The edited actions are never executed, so we use an LLM verifier to reject attacks that are inconsistent with the original transcript or below a severity bar. The attacker revises the edit until the monitor no longer flags it. We applied Screwtape to 94 Claude Code and Codex transcripts from Redwood Research staff, optimizing against our own monitors. Across 252 runs, it produced 35 attacks that passed the verifier and evaded these monitors. Six of them, confirmed valid by human review, also evaded a held-out monitor provided by a frontier developer. We therefore demonstrate that a red-teamer without access to a developer's infrastructure can find confirmed failures of the developer's monitor. We expect to find more attacks when the pipeline runs inside the developer's infrastructure, with frontier attacker models that lack refusal safeguards and direct optimization against the production monitor. Developers can use these attacks to fix monitor blind spots, and evaluators can use them as counterexamples to safety claims about the monitor.
agentevaluator - arxiv:2610.20939 · cs.AIMAI-Gym: An Open All-in-One Platform for GUI Agent Evaluation and RL Training across Mobile and Computer EnvironmentsPanrong Tong, Quyu Kong, Pengxiang Li, Tianyu Xia +11
Online reinforcement learning (RL) with verifiable rewards has driven rapid progress in mathematics, coding, and tool use, but scaling this success to GUI agents remains challenging due to the high cost of interactive environments and the scarcity of realistic training tasks. We introduce MAI-Gym, an open all-in-one platform that unifies GUI agent evaluation, data collection, and online RL training across mobile, computer, and web environments. Through a shared rollout pipeline, MAI-Gym ensures consistent agent behavior across evaluation and training, supports ten established benchmarks, and scales to 20,000 parallel environment instances. Beyond scalable infrastructure, MAI-Gym enables automated synthesis of diverse GUI applications and challenging tasks, including resettable mocked apps and coherent cross-app worlds with execution-validated task verifiers. Using MAI-Gym, we train MAI-Gym-9B, a unified 9B agent for mobile and computer use, achieving state-of-the-art performance among open GUI agents under 10B parameters: 78.4% on MobileWorld and 76.2% on OSWorld-Verified. Importantly, RL in a mocked train-ticketing app transfers to its real-world counterpart, improving task success on a phys- ical phone from 28.8% to 48.1%. We release the framework, synthesis skills, mocked apps, 2,000 verified tasks, and the model checkpoint to support repro- ducible research and further advances in scalable GUI agent training.
agenttool usebenchmark - arxiv:2610.20938 · physics.app-phAnalytically Benchmarked Spatial and Temporal Discretization Guidelines for Finite Element Simulation of Thermal WavesKaushal Arun Pareek, Daniel May, Peter Meszmer, Bernhard Wunderle +1
Finite element simulation is widely used to design and interpret thermal wave-based measurement systems. The accuracy of the results obtained from such simulations depends on how finely the model is resolved in space and time. The discretization resolution is often chosen case by case rather than following a well-defined guideline. This work recommends a discretization protocol for finite element simulation of thermal waves where the spatial discretization is scaled to the thermal diffusion length and the temporal discretization to the number of time steps per cycle. The associated discretization errors are quantified using three benchmark problems of increasing complexity. The spatial phase error converges at second order in the element size, reaching 0.14 degrees at ten elements per diffusion length, whereas the temporal error decreases only at first order in the number of time steps per cycle; thus, the time step can be set from a target phase tolerance $ε$. Each error is assessed against a reference that isolates it. The benchmark case with subsurface defects shows the consequence of violating the recommended discretization protocol.
benchmark - arxiv:2610.20936 · cs.ROLEAP: Gradient Flow Augmented Lagrangian Trajectory Optimization for Contact Constrained RobotsCésar E. Ramos Chuquiure, Ryan Tsai, Yutong Zhou, Zihang Huang +7
Generating dynamically feasible motions for legged and manipulating robots requires solving a large, nonlinear, non-convex optimal control problem (OCP) in which contact wrenches must satisfy friction-cone constraints throughout the motion, not only at isolated instants. Existing solvers are sensitive to the initial guess and scale poorly with the number of joints. This paper introduces LEAP, which transcribes the OCP with pseudospectral collocation and, instead of a Newton or SQP solve, integrates a primal--dual gradient flow of a Powell--Hestenes--Rockafellar augmented Lagrangian to compute a solution. Because the flow is stiff, this paper designs a linearly implicit integrator with a Gauss--Newton metric, momentum, and a merit-based step controller that keeps the augmented Lagrangian monotone. Across locomotion benchmarks on five robot platforms, from planar hoppers to 32-joint humanoids, LEAP is the only method to solve every tested problem, achieves substantial speedups over Crocoddyl and Aligator on several benchmarks, and reaches trackable references at smaller compute budgets. The resulting trajectories are validated by tracking them in a high-fidelity simulator and on hardware. We will make the code public upon final acceptance of this paper.
humanoidbenchmark - arxiv:2610.20935 · cs.AIRAC: A Relational Agent Contract for Constraint-Governed Engineering TasksKaiyuan Zheng, Fengqing Jiang, Yichen Feng, Zubin Aaron Assadian +5
Engineering agents must produce artifacts that satisfy exact, coupled constraints. Existing solutions often require agents to make semantic decisions and also derive consequences such as precise geometry or timing from those decisions, leading to failures in constraint-governed engineering tasks. In this paper, we introduce RAC, a harness-agnostic contract for the action and feedback layer of engineering agents. It specifies authoritative observations, semantic actions, tool-based realization, independent verification, and repair evidence. The agent decides which relations an artifact should satisfy; specialized tools derive their concrete consequences. On PCB placement, routing, and repair tasks, RAC achieves the highest clean-pass rate for all nine evaluated models, averaging 90.9%, compared with 74.0% for the strongest baseline, requirements-only delegation. In job-shop scheduling, direct relational actions achieve 96.3% quality pass versus 59.6% for code-mediated delivery with access to the same scheduling backend. Our evaluations on 141 embedded-code tasks across four microcontroller identify the action contract as a source of agent capability: preserving semantic control while delegating mechanically derivable outputs improves the reliability of constrained artifacts.
agent - arxiv:2610.20927 · cs.CLPrune the Summary, Not the History: Source-Judged Compaction for Long-Horizon AgentsGuanghui Min, Chen Chen
Long-horizon LLM agents rely on context compaction: when the interaction history outgrows the context window, a compressor replaces it with a natural-language summary. We show that these summaries contain more than a record of what happened. Compaction prompts ask for decisions and next steps, so each summary mixes facts established by the environment with reasoning written by the compressor, and the agent cannot tell the two apart afterwards. In an analysis of a widely used structured compressor, 83% of failed runs trace back to an action that carries out a specific summary entry, most of which no environment output supports. We propose Source-Judged Compaction (SJC), a training-free post-processing step that scores each summary entry for whether it has a source in the environment outputs or the previous summary, and removes the entries that do not. It changes neither the agent nor the compressor and adds one inexpensive judge request per compaction. The judge is Jev, a non-generative decision model that answers the source question with a calibrated probability instead of text. On AppWorld and OfficeBench with two agent models, SJC improves cross-run reliability over the underlying compressor by 2.1%-8.9% without increasing the agent's token usage.
agentllm agent - arxiv:2610.20926 · cs.AISkillNEO: Non-Elitist Evolutionary Optimization of Agent SkillsEnze Ma, Chunyu Miao, Yankai Chen, Jie Yang +13
Agent skills turn procedural knowledge into reusable artifacts, allowing LLM agents to accumulate and refine expertise through experience. Building on this, skill evolution iteratively improves skills using execution feedback. However, improvement may require several interdependent edits, so low-reward versions can still be worth developing. Requiring immediate reward gains can discard useful partial repairs before their benefits emerge. Non-elitist evolution allows new versions to receive further edits despite scoring below earlier ones. For agent skills, this calls for evidence of what to preserve and repair next. We introduce SkillNEO (Non-Elitist Evolutionary Optimization for Agent Skills), which separately selects a version to use and versions to keep improving. Comparing how related versions perform the same task reveals working steps to preserve and remaining failures to fix. It can also revisit and revise earlier versions using later evidence. Across 86 SkillsBench tasks in eight domains, including software engineering and office work, SkillNEO raises GPT-5.4 Mini's average reward from 49.1% to 66.4%, within GPT-5.5 and GPT-5.6 Luna's original-skill range. Mini executes tasks, proposes revisions, and edits skills; final evaluation uses the same task instances. Gains span all eight domains. The same skills improve stronger models without further editing. On the 41 tasks with modified skills, reuse improves Luna by 10.5 percentage points and places GPT-5.5 10.6 points above its published original-skill reference. Execution records show these skills can benefit stronger models even when Mini still fails some task requirements. Search histories also show lower-scoring versions leading to skills ultimately selected for use. Together, these findings show that skill evolution can turn a weaker model's partial progress into reusable procedural knowledge for improving more capable agents.
agentllm agent - arxiv:2610.20925 · cs.ROPlanning to Learn: Real-Time Robust Adaptive Control on the GPU with Guaranteed Future LearningAnutam Srinivasan, Siddhant Panse, Glen Chou
Robust adaptive control is vital for keeping robots safe while actively reducing uncertainty in unfamiliar settings. However, traditional methods are often conservative since they fail to account for guaranteed future parameter learning during planning and struggle to scale to high-dimensional systems. In this work, we present A-SLS, a learning-aware robust adaptive model predictive control (RAMPC) framework based on system-level synthesis. A-SLS forecasts future parameter uncertainty and preserves correlations between process disturbances, parameter error, and estimate updates to construct tight, guaranteed reachable sets that account for future information gain. We also develop a GPU-accelerated A-SLS solver to enable real-time performance. Across nonlinear robotic systems, A-SLS reduces reachable-tube conservativeness, maintains real-time performance, and scales to whole-body control for a quadruped with 61 states, 8 uncertain parameters, and 12 control inputs. We further demonstrate the approach on hardware, where a physical Crazyflie drone learns an uncertain dynamics parameter online to stabilize on the fly.
quadrupedwhole-body control - arxiv:2610.20904 · cs.ROEgoJEPA: Beyond Retargeting for Dexterous Manipulation from Egocentric VideosMeng Cao, Can Zhao, Jinjie Wei, Haonan Zhao +11
Egocentric human videos offer a scalable alternative to costly robot demonstra- tions, yet transferring their manipulation priors to robot control remains chal- lenging. The commonly adopted retargeting-based approaches convert human wrist and finger poses into embodiment-specific action commands through hand- crafted geometric mappings. However, such mappings require meticulous recal- ibration across different embodiments and risk propagating hand-tracking errors into downstream robot control pipelines. To this end, we propose EgoJEPA, a three-stage framework that learns embodiment-agnostic representations from ego- centric human videos without rule-based human-to-robot retargeting: 1) Human- action pre-training learns effect-aware latent actions through JEPA-style visual prediction conditioned on hand-motion differences; 2) Latent-action-based world modeling jointly predicts future visual dynamics and latent actions, which decou- ples human prior pre-training from robot-specific control parameterizations; 3) Embodiment-specific tuning adapts the pretrained model to diverse robots through lightweight action adapters. Experiments on five real-world manipulation tasks across three robot embodiments demonstrate the effectiveness and efficiency of EgoJEPA. It achieves a 74% average success rate under the out-of-domain evalu- ation, compared with 65% for the retargeting baseline, while reducing cumulative cross-embodiment wall-clock cost by 47.2%, from 743.2 to 392.7 hours.
manipulationdexterousworld model - arxiv:2610.20903 · cs.LGSign Betting for Misspecified Regression: Variance Adaptivity and Optimal ApproximationYinan Li, Han Han, Kwang-Sung Jun
We study bounded regression under model misspecification, evaluating a function $f$ by its $L_1(P_X)$ distance to the conditional mean $η(x)=\mathbb{E}[Y\mid X=x]$, rather than the standard $L_2(P_X)$. This $L_1$ metric is natural for downstream decision problems such as bandits and reinforcement learning. We analyze the min--max estimator of Baraud & Maillard (2025), which we call \emph{sign-betting regression}. Our main results show that sign-betting regression achieves high-probability $L_1$ oracle inequalities under misspecification. Under realizability, our bound adapts to the average conditional variance and recovers the variance-adaptive rate of Li et al. (2026). We further characterize the optimal universal approximation factor for proper learning by showing that sign betting achieves factor 3, while no proper learner can satisfy a distribution-free, high-probability $L_1$ oracle inequality with a universal approximation factor strictly below 3. We extend the upper-bound analysis beyond finite classes and complement the theory with experiments in synthetic and real-world model-selection tasks.
world model - arxiv:2610.20901 · cs.ROTGrasp6D: Task-Informed Vision-Language Priors and Pose Refinement enable Zero-Shot Task-Oriented 6DoF GraspingManav Kulshrestha, Apoorva Vashisth, Damon Conover, Aniket Bera
Task-Oriented Grasping (TOG) has emerged as an important skill for purposeful robotic manipulation. Given the task of using a knife to "slice bread," a robot must choose to grasp from the handle, whereas for "hand[ing] the knife over," it is preferable for the handle to remain unobstructed for safe handoff. Current approaches for TOG either focus on 2D planar grasping, remain limited by their reliance on task-agnostic grasp generators, or are constrained by reliance on finite, manually-annotated, task-specific grasp datasets that cannot cover all possible grasp-task pairs. To that end, we present TGrasp6D, a vision-language based approach for generating task-oriented 6DoF grasps in a zero-shot manner. Given a task and a single RGB-D image of an associated object, TGrasp6D reasons about the appropriate semantic affordance for the given task and object geometry to produce an initial task- informed prior. This generated prior is further refined via 6DoF pose optimization for collision avoidance while preserving task consistency. Our approach achieves state-of-the-art success rate of 94% on the TaskGrasp dataset, with the best baseline at 74%. We also deploy our approach on a Franka Research 3 robot to demonstrate zero-shot generalization to real world objects
manipulationfrankagrasp - arxiv:2610.20893 · cs.LGFlexible Doubly Robust Proximal Causal Inference for Censored OutcomesDhanajit Brahma, Ruolin Meng, Lawrence Carin, Ricardo Henao
Estimating causal effects on right-censored time-to-event outcomes is challenging under unmeasured confounding. Proximal causal inference addresses unmeasured confounding through proxy variables, but existing proximal survival methods rely largely on restrictive parametric bridge models. We introduce Proximal Survival Bridge learning (PSB), a flexible doubly robust proximal framework for censored outcomes. Building on existing proximal identification of treatment-specific survival curves, the framework yields outcome-bridge (OB), treatment-bridge (TB), and doubly robust (DR) estimators, with two flexible bridge learners: PSB-K, a kernel estimator with a low-rank time decomposition, and PSB-N, a neural estimator trained by a projected maximum-moment-restriction loss. Unlike existing flexible proximal learners for scalar outcomes, PSB directly handles censoring and survival-curve targets. On a standard SEM benchmark, a proximal semi-synthetic ACTG study, and the real-world Framingham study, PSB-K performs strongest on SEM and ACTG, whereas PSB-N is the most stable on Framingham. Additional results with modern causal survival baselines, weakened or invalid proxies, and heavier censoring assess robustness across challenging settings.
benchmark - arxiv:2610.20867 · cs.AIAnticipatory Commonsense with Moving Objects: An ASP-Based Framework for Predictive Reasoning with Interacting TrajectoriesJulius Monsen, Jakob Suchan, Mehul Bhatt
We present a general declarative framework for anticipatory reasoning about interactive motion in complex dynamic environments. Our focus is on (explainable) anticipatory space-time mental model formation at a semantic level to heuristically guide low-level motion computation. Our method is hybrid, modular and elaboration tolerant integrating: 1) An ASP model characterising the dynamics of interactive motion trajectories; 2) Semantically-guided prediction for deriving `branching trajectory futures'; and 3) An energy-based {optimisation} model to evaluate, rank and filter anticipations. We evaluate with the Argoverse 2 benchmark, where each anticipated trajectory carries an explicit, inspectable event-level hypothesis, unlike end-to-end stochastic prediction.
benchmark - arxiv:2610.20866 · cs.ROSHRED: Self-Play for Heterogeneous Road-Users without Expert DemonstrationsWei-Jer Chang, Zhouchonghao Wu, Akshay Rangesh, Weixin Li +2
Realistic traffic simulation and autonomous planning both depend on modeling interactions among heterogeneous road users whose dynamics, constraints, and behaviors differ fundamentally. Existing approaches predominantly learn realistic interaction behavior from large-scale human driving demonstrations. In this work, we ask whether such behavior can instead emerge purely through self-play. We present SHRED, a framework that jointly trains vehicles, pedestrians, and cyclists through large-scale heterogeneous self-play from random initialization, without demonstrations or logged-trajectory replay. SHRED combines type-specific dynamics and objectives with procedural scenario generation and per-agent randomization, allowing diverse behaviors to emerge through closed-loop interaction. On the Waymo Open Sim Agents Challenge (WOSAC), SHRED jointly controls all three road-user types and outperforms both the closest demonstration-free self-play baseline and prior self-play methods that use demonstration-trained reference policies. Policies trained only on nuPlan map geometry retain comparable realism under zero-shot evaluation on Waymo. For planning, SHRED outperforms all compared planners on the interaction-heavy InterPlan benchmark. Controlled ablations show that heterogeneity in self-play improves both traffic realism and planning performance, with particularly large planning gains in interaction-heavy scenarios. Together, these results establish heterogeneous procedural self-play as a scalable, demonstration-free paradigm for learning realistic and interactive traffic behavior.
self-playbenchmark - arxiv:2610.20865 · cs.LGLong-Context Drafting Does Not Require Long-Context DraftersAref Jafari, Vansh Bhatia, Matt Qin, Subhajit Dutta Chowdhury +4
Speculative decoding pairs a cheap drafter with parallel verification against the target. At ultra-long context it stops paying, because a drafter that attends over the full prefix inherits the target's context-dependent cost, so the speedup it was meant to provide first erodes and then reverses. The standard response is to build a long-context drafter, by training or fine-tuning one, or by using a multi-token-prediction head shipped inside the target checkpoint. We argue this is unnecessary, because the drafters in use are target-conditioned, reading hidden states the target computed over the entire prefix, so distant context already reaches them without their own attention, and verification corrects what they miss. We therefore freeze released short-context drafters and run them far past their training length under two inference-time mechanisms, one bounding the attention cache they see, which fixes their cost, and one re-basing their rotary position indices into the trained range. No weight is touched and the only tunable saturates early, so one vLLM-side configuration spans 4k to two million tokens, costing at most 0.16x of speedup at short length. We prove when a common index shift leaves the drafter's computation exactly unchanged, and decompose acceptance into two measurable gaps with sharp bounds. Whatever speedup a frozen drafter delivers at short context, it retains out to two million tokens, across three target-drafter pairs and five corpora. The claim concerns the flatness rather than the multiple, and removing either mechanism eliminates it. The gain is conditional on the speculative cost per emitted token staying below the matched target-only cost.
long-contextlong context - arxiv:2610.20863 · cs.LGSNOAR: Self-training with Noise filtering, Observation adding, and Remixing for Speech Denoising Without Clean ReferencesSamuele Cornell, Matthew Maciejewski
Unsupervised speech denoising methods such as MixIT and noisy-target training can be trained directly on noisy recordings, avoiding the sim-to-real gap of synthetic training data, but remain well below supervised performance. Teacher--student self-training, as in RemixIT, narrows this gap by remixing a teacher's estimates into new mixtures that supervise a student; however, since the student is trained only toward the teacher's estimates, teacher errors, in particular over-suppressed speech, are never recovered and propagate as the teacher is updated from the student. We propose SNOAR, a self-training framework that corrects the teacher's estimates before use: the noise leaked into its speech estimate is removed by differential noise filtering (DNF), and part of the noisy observation is added back in to restore suppressed speech and attenuate artifacts. We show that the two corrections depend on each other, as without DNF the loop collapses to the identity mapping. Compared to RemixIT, SNOAR is more stable along training and outperforms it in signal-based, perceptual, and word error rate metrics, reaching 97\% of the SI-SDR improvement of a fully-supervised topline on WHAM!; on real CHiME-3 recordings, it improves every metric of three different teachers, while RemixIT degrades their perceptual quality.
sim-to-real - arxiv:2610.20851 · cs.LGComplete or Sparse: A Tale of Two IdentifiabilitiesJunze Zhou, David Klindt
Understanding when neural networks learn interpretable representations is central to mechanistic interpretability. We study this through identifiability: when can supervised learning recover the true latent variables underlying the data? We introduce two data generating processes distinguished by their causal direction where either labels cause latents (generative) or latents cause labels (discriminative). When the representation has at least as many dimensions as the latent space, we prove that the optimal encoders are linear in the latents. Conversely, when dimensions are insufficient, identifiability generically fails and encoders become nonlinear. We then show that additional constraints yield recovery of individual latent variables, not just linear mixtures. In the generative setting, non-negative activations and energy efficiency force each neuron to encode exactly one latent, providing a precise account of selective grandmother cell-like responses. In the discriminative setting, when latents are sparse and the true measurement matrix satisfies the spark condition, the learned encoder inherits compressed sensing structure, enabling sparse dictionary learning to perfectly recover individual latents. This constitutes, to our knowledge, the first identifiability result for undercomplete nonlinear independent component analysis. Our framework unifies nonlinear ICA, compressed sensing, and neural network interpretability, delineating exactly when representations transition from nonlinear distortions to linear representations and, under additional constraints or steps, to clean, axis-aligned codes. Theoretical predictions are validated on synthetic benchmarks and on activations drawn from vision models, large language models, and macaque inferotemporal cortex.
benchmark - arxiv:2610.20845 · cs.AIEvolving Together, Failing Apart: The Skill-Policy Co-evolution Trap in Self-Improving AgentsYujie Feng, Xiaobin Wang, Jinluan Yang, Xiaonan Shi +4
Skill-policy co-evolution, in which an agent's policy model and its skill bank improve each other during training, is a promising paradigm for self-improving agents. Current evaluations score the co-evolved policy on the final bank it trained with (the delivered bank). After deployment, however, skills may be rewritten, added or removed more frequently than the policy model is retrained, leaving the policy to operate on a drifted skill bank that neither the training objective nor existing evaluation protocols address. In this paper, we first establish, through theoretical and empirical analysis, that the fragility hidden by this blind spot is not incidental. The acceptance rule these methods share (cooperative acceptance) ties a policy's competence to the bank it co-evolved with, leaving it fragile once skills drift, regardless of the implementation; empirically, sensitivity to skill drift grows as co-evolution training proceeds. We therefore introduce the CO-evolution Drift Audit (CODA), a deployment-oriented evaluation protocol that simulates skill drift and, alongside success rate under worst-case drift, introduces two new metrics: a consistency rate under meaning-preserving edits and a resistance rate to corrupted skills. Across four co-evolution methods and five benchmarks (six settings), worst-case success is 5-13 points below the peak measured for each delivered pair, erasing 52-93% of the delivered bank's gain over the no-skill baseline. CODA thus reframes the target of co-evolution, from peak performance on the co-evolved bank to the ability to harness skills however they evolve, and provides a shared standard against which future methods can be audited.
self-improvingbenchmarkevaluation protocol - arxiv:2610.20843 · cs.CVMARS - A foundational map auto-regressorQi Zhang, Suvam Bag, Rupanjali Kukal, Lucas Thai +8
Map generation tasks feature extensive non-structural vectorized data (e.g., points, polylines, and polygons) and thus pose significant challenges to common pixel-wise generative models. Conventional approaches use multiple stages, first segmenting these features at the pixel level and then performing vectorized post-processing, with errors and complexity compounding at each stage. Motivated by the recent success of auto-regressive language modeling, we propose the first map foundation model, named Map Auto-Regressor (MARS), that is capable of generating both multi-polyline road networks and polygon buildings in a unified manner. For training MARS, we collected to our knowledge the largest multi-class map extraction dataset totaling 3.4M examples, which we call MAP-3M. Across four road and building datasets, MARS outperforms or matches the performance of multistage baselines. Additionally, we develop a "Chat with MARS" capability that enables interactive human-in-the-loop map generation and correction, supported by the auto-regressive nature of our end-to-end approach. We release our MAP-3M dataset and project demo page at (1) https://huggingface.co/datasets/bag-lab/MAP-3M and (2) https://huggingface.co/spaces/bag-lab/MARS, respectively.
human-in-the-loop - arxiv:2610.20841 · cs.LGAnalog RF Computing for Integrated Data and AI Flows Over Wireless NetworksWentao Yu, Vincent W. S. Wong
Many resource-constrained edge devices require artificial intelligence (AI) services but are not powerful enough to store or execute neural networks (NNs). Analog radio frequency (RF) computing addresses this limitation by letting a base station (BS) multicast a waveform encoding NN weights to different clients. Each client then multiplies the received weight-encoded waveform with its locally generated input-encoded waveform through a passive mixer in its RF frontend, thereby realizing the matrix-vector multiplications (MVMs) needed for AI inference with an extremely low power and memory requirement. This can enable telecom operators to provide AI inference services to the edge devices connected to their networks. In practice, however, this multicast AI flow must share the spectrum, power, and spatial resources with other data flows. This paper considers a multi-user multiple-input multiple-output (MU-MIMO) downlink scenario and studies the resulting relative aggregate data rate reduction from integrating the multicast AI flow with the unicast data flows. We formulate a joint beamforming problem for the data and AI flows to maximize the aggregate data rate subject to computing accuracy, client energy, total power, and hardware-related constraints. We develop a weighted minimum mean-square error (WMMSE)-based algorithm to solve the problem. Numerical results based on 3rd Generation Partnership Project (3GPP) channels show that the relative aggregate data rate reduction from integrating an AI flow with data flows is below 5% in representative settings. This indicates that telecom operators can use analog RF computing to provide edge AI services while providing data communication services simultaneously.
memory - arxiv:2610.20840 · cs.ROA Phenomenology of Neural Geometry in Vision-Language-Action ModelsMozes Jacobs, Aiden Swann, Mathilde Papillon, Siddharth Boppana +15
Robotic manipulation is inherently geometric, but what shape does the world take inside the policy that performs it? Prior work on vision-language-action (VLA) policies establishes that task-relevant variables are linearly readable from activations, yet readability certifies only that a variable is present and says nothing about the shape its values trace, which may be curved and multi-dimensional. Using controlled sweeps, natural rollouts, linear probes, and activation interventions, we build a taxonomy of seven task-relevant concepts in VLA models. We find that realized geometry is predominantly multi-dimensional and is shaped by the constraints under which the policy acts. Physical joint limits, for example, leave wrist orientation an open arc rather than a circle, and distance to the target, a single scalar, is realized in two orthogonal subspaces, a coarse ruler for the approach and a fine ruler for the grasp. The same geometries reappear in two further VLA models on real-world robot data, indicating that they are neither model-specific nor artifacts of simulation. We demonstrate how the findings are actionable by steering along the geometry a concept realizes. Overall, this work lays out an initial phenomenology of realized geometry in VLA policies and demonstrates that a concept's shape carries as much importance as its readability.
vision-language-actionvlavla modelmanipulationgrasp - arxiv:2610.20836 · cs.LGConfidence-Calibrated Inference Expansion for Evaluator-Guided Test-Time ReasoningWeida Liang, Kenji Kawaguchi
Evaluator-guided reasoning presents a local control problem at every step: trust the current answer, return a fallback, or spend more inference. We formulate this choice as a compute-aware three-action controller, scoring terminal actions by predicted answer success and expansion by a rollout-based continuation-success surrogate, with compute cost priced explicitly. A local action-gap analysis characterizes when score errors can change a decision, while an expected regret bound accounts for errors along the realized expansion path. Guided by this analysis, we learn and calibrate action-specific confidence estimates, with local uncertainty penalties for ambiguous decisions. On held-out width expansion and revision tasks, the controller improves observed utility over calibration-selected fixed policies while reducing normalized online compute by about 30% and 16%, respectively. Across four Qwen3-8B seeds, compute is lower than fitted Q in both settings, while utility differences are small and mixed. The results frame evaluator-guided reasoning as resource allocation under uncertainty, rather than a uniformly stronger reasoner.
evaluator - arxiv:2610.20832 · cs.LGAmortizing Sequential Experimental Design in Reasoning LLMsAnita Kriz, Antonin Berthon, Nicolás Astorga, Mihaela van der Schaar
Despite their extensive pretraining knowledge and strong reasoning capabilities, large language models (LLMs) struggle to adaptively gather information in experimental design tasks. We study how to amortize sequential Bayesian experimental design (BED) into an LLM by explicitly training the reasoning that guides action selection using information-gain rewards. Our approach specifies the experimental generative model independently of the LLM's own beliefs, allowing us to optimize expected information gain under a model of our choosing. We investigate the design choices for amortization, including alternative generative-model constructions, reward estimation, hypothesis-space specification, multi-step objectives, and computational costs relative to training- and inference-time BED baselines. On Twenty Questions, the resulting policies improve information acquisition and task performance, outperform inference- and training-time BED baselines at substantially lower computational cost, and generalize to unseen categories. We further demonstrate positive transfer from Twenty Questions training to Medi-Q, a real-world clinical dialogue benchmark. These results suggest that information gathering can be learned as a generalizable reasoning skill in LLMs.
benchmark - arxiv:2610.20830 · cs.CLPEDBench: Can Speech Models Detect Phonetic Errors?Zhonghao Shi, Zackary Rackauckas, Jasmine Yip, Savannah Halpern +8
Phonetic error detection (PED) aims to identify phoneme-level substitutions, deletions, and insertions and is essential for educational and clinical speech assessment and intervention. Despite recent advances in speech models for phoneme recognition, their ability to reliably detect fine-grained phonetic errors remains unclear because of a lack of high-quality evaluation data. To address this gap, we introduce PEDBench, the first PED benchmark spanning a continuum from natural to controlled speech. It comprises real-world recordings of spoken errors made by children (N=334), scripted error productions by human actors (N=1,188), and synthetic error productions generated using phone-conditioned text-to-speech (N=1,083). All phonetic annotations were transcribed and verified by two phonetically trained annotators. Our evaluation of 13 state-of-the-art speech models shows that they fail to reliably detect phonetic errors, even under controlled conditions with acted and synthetic speech, where acoustic variability is reduced. The models also exhibit low precision, frequently misclassifying correctly pronounced phonemes as errors. Our speech-language models, fine-tuned on synthetic data, outperform commercial and open-source baselines on word-level PED, although a sim-to-real domain gap remains. We plan to publicly release our code and recipes, along with our acted and synthetic datasets, to support the development of more accurate PED systems.
sim-to-realbenchmark - arxiv:2610.20828 · cs.LGCoTranslator: Reverse to Forward Chain-of-Thought Translator for Reasoning DistillationZinan Tang, Zhuoshi Pan, Kairui Chen, Ze Yan +4
Reasoning distillation requires chain-of-thought (CoT) data with correct answers, but obtaining such data efficiently remains challenging. Forward CoT generation (FCG) followed by rejection sampling (RS) produces verified, question-only responses, yet may require many teacher calls for difficult problems. Reverse CoT generation (RCG) is cheaper because it conditions on the known answer, but access to this answer can lead to post-hoc rationalization and a mismatch with the student's answer-free inference setting. We propose CoTranslator, which learns to translate reverse CoT responses toward the distribution of verified forward responses. Trained on a small set of paired reverse and rejection-sampled forward responses, it supports large-scale generation and verification of student training data. Across six student models evaluated on six mathematics benchmarks, CoTranslator improves average accuracy over semantic suppression by up to $5.82$ percentage points. For two 4B students, it retains $95.0\%$ of rejection sampling's average accuracy while using $35.9\%$ of its estimated compute. Without retraining, the same translator transfers to a new mathematics dataset, a different teacher, and STEM problems, improving average STEM accuracy by $4.88$ and $4.31$ percentage points for 4B and 8B students, respectively. These results show that CoTranslator offers an efficient and transferable approach to reasoning distillation.
benchmark - arxiv:2610.20823 · cs.ROFeasibility Ceilings Before Learning: Sizing and Timing Bounds for Drone-Mesh InterceptionDmytro Babarytskyi
A mesh of interceptors defending a fixed site must spread over the predicted arrival point of an approaching object. This is a recurring task in multi-agent robotics: N agents, each covering a disc of radius r, spread over a probability density so that a point drawn from it falls within some agent's disc. We give closed forms for the quantities that bound performance before any policy is chosen, and validate each by Monte Carlo. The uncertainty of a point predicted by extrapolating a noisy linear track follows from least-squares variance propagation plus a delta-method term for the uncertain arrival time, agreeing with simulation to within 1.8%. The coverage probability of a packed mesh admits a closed form accurate to 0.007 absolute, with a volumetric analogue. A naive area ratio misses a ceiling: while agents stay at least 2r apart, no mesh exceeds 90.69% coverage in the plane or 74.05% in a volume. This ceiling belongs to the separation constraint, not to the problem - permitting overlap substitutes Kershner's thinnest-covering density and removes it, the better regime above a crossover at p=0.80. Hexagonal packing is optimal among five arrangements; the centroidal-Voronoi configuration coverage control supplies reaches only half its coverage at small budgets. The Gaussian assumption is conservative under heavy tails and anisotropy but fragile under bias: a one-sigma aim-point offset costs 14 points of coverage. A repositioning team must also choose a commit time, and the admissible window can be empty. Against the applicable ceiling, analytic placement attains 85% of what the configuration permits while centralised PPO averages 15% over three trainer seeds, with a paired bootstrap against a null policy containing zero at every seed. We argue such ceilings should be computed before any learning experiment and quoted alongside every result.
multi-agent - arxiv:2610.20822 · cs.CVTemporal Specification Guidance for Text-to-Video GenerationAvishree Khare, Hideki Okamoto, Bardh Hoxha, Georgios Fainekos +1
Text-to-video models often generate the visual concepts requested in natural language prompts but fail to satisfy their intended ordering, timing, and persistence requirements. We propose representing these requirements as specifications in a formal Temporal Logic, and introduce Temporal Specification Guidance (TSG), a training-free inference-time framework that turns temporal specifications into executable objectives for guiding video generation during sampling. TSG scores traces of visual concept evidence using differentiable quantitative semantics and uses the score's gradients to construct a specification-guided denoising direction. We instantiate this framework with video-text attention traces, efficient differentiable scores for Metric Temporal Logic, and gradient-derived attention biases. This instantiation requires neither an external concept classifier nor backpropagation through the video transformer. Its attention-steered forward pass yields a denoising direction that augments classifier-free guidance (CFG). On a benchmark of 220 prompts spanning 11 categories of temporal constraints, including ordered object entry and state transitions, this instantiation improves specification satisfaction accuracy over CFG by 9.5-20 percentage points across five models, with minimal memory overhead. These results demonstrate the potential of temporal specifications and differentiable satisfaction objectives as a flexible interface for controlling video generation.
memorybenchmark - arxiv:2610.20814 · cs.LGGibbs-Gap Policy Optimization for Off-Policy RL with Verifiable RewardsXuanchang Zhang, Tong Zhang
Reinforcement learning with verifiable rewards often trains language models on responses generated by a different, possibly stale behavior policy. Correcting this mismatch with learner-to-behavior importance weights can introduce high variance for long responses. We introduce Gibbs-Gap Policy Optimization (GGPO), which minimizes a log-mean-exp contrast of KL-adjusted rewards among responses to the same prompt, without learner-to-behavior importance weights. We derive GGPO from KL-regularized policy optimization: on policy, its population loss is exactly the KL-regularized optimality gap, and under suitable coverage the learner's single-response expected reward is at least the behavior policy's best-of-$K$ reward, minus terms for the raw finite-group loss, regularization, and finite training data. We characterize reference refresh through exact Gibbs-target composition and a finite-data final-policy reward bound under local reference control and tokenwise behavior mismatch. This refresh is empirically important: in an ablation, fixed-reference GGPO scores $32.96$ on the five-benchmark problem-weighted average, versus $46.58$ with the default refresh schedule. Experiments on mathematical reasoning further show that GGPO outperforms GRPO and AGRO under training-inference mismatch and on fixed offline data, while maintaining its performance as controlled policy staleness grows.
benchmark - arxiv:2610.20811 · cs.CVRIGen: Beyond Action Following with Reciprocal Interaction in Generative GameplayMengfei Li, Yi Zhang, Xingtong Ge, Lunjie Zhu +7
Generative gameplay requires more than immediate responses to player actions.Actions should have persistent consequences, while the generated world should act according to its own state and rules, shaping the player's situation and subsequent choices. Video generation captures local dynamics, but limited history, partially observed game state, and the need for contextual behavior decisions make such interactions difficult to sustain through audiovisual prediction alone. We introduce \rig (Reciprocally Interactive Generation), a framework for reciprocal interaction in generative gameplay. \rig combines persistent state and configurable rules, contextual interaction reasoning, and action-conditioned audiovisual generation. Together, these components make interaction consequences and actor behavior jointly determine generated content while maintaining consistency between game state and visible outcomes. We construct a Xonotic gameplay dataset to train an audiovisual generator for learning dynamics and interactions. Further post-training enables \rigarena, an interactive generative first-person shooter (FPS) prototype. Experiments demonstrate more reliable combat outcomes and reciprocal interaction between the player and the generated world. Project page: https://murphylmf.github.io/RIGen/.
action-conditionedpersistent statepost-training - arxiv:2610.20809 · cs.AIA Self-Improving Agent that Grows its Own DSL: Oracle-Free Operator Induction for Exact Question AnsweringYun-Shiuan Chuang, Yirou Ge, Yixi Li, Yiming Bian +5
A self-improving agent has to decide which new skills to keep, and in practice it keeps whatever earns a high reward, convinces a judge model, or compresses its library, even when the skill is wrong. That is fine until the answers have to be exactly right: over a table of transactions, a total is either right or wrong. We ask a harder question: can an agent grow what it can compute with no reward, no judge, and no answer key, and still be exactly right? A single frozen model runs two nested loops over a small, exact domain-specific language (DSL). The inner loop answers new questions only when two independently written expressions agree on every record. The outer loop discovers its own missing capability and authors an operator for it, admitted only when an independent second implementation agrees and the result obeys a conservation law, all without ground truth. The inner loop roughly triples the answerable-question set, and the outer loop self-extends on disjoint starting sets, inventing operators outside the starting closure; a no-induction control answers none of the new questions. The admission rule is empirically sound: it rejects 138 of 138 buggy operator variants under an independent reference, even one written by a different model family, where the admit-if-it-runs, self-consistency, and compression signals the field relies on admit most of them. Unmodified, the same mechanism self-extends on two further domains with unrelated schemas, business-ledger and clinical-encounter question answering, growing their answerable sets more than tenfold and nearly twentyfold. The loop transfers across model families too: run with further proposers, the gate stays sound and independent families self-extend comparably, with only the induction yield tracking each proposer's DSL fluency.
agentself-improvingjudge model - arxiv:2610.20796 · cs.LGPrivacy-Preserving Text Inference for Health-Misinformation Classification and Semantic RetrievalJasleen Kaur, Chloé Cachet, Koray Karabina, Jean Belo Klamti +1
Health-misinformation analysis presents a dual challenge: misleading claims can affect public well-being, while systems designed to detect them may expose sensitive user-submitted text. This motivates privacy-preserving inference for public-health applications in which user queries may contain confidential medical information and trained models and reference databases are maintained on organizational servers. We design a privacy-preserving text inference framework based on CKKS homomorphic encryption for health-misinformation classification and semantic retrieval. The classification pipeline fuses BioBERT embeddings, PCA-reduced MPNet embeddings, and SVD-reduced TF-IDF features, while the retrieval pipeline combines MPNet embeddings with ISCC-SCT semantic fingerprints through late fusion. We evaluate predictive performance and benchmark plaintext versus ciphertext-based execution. The hybrid classifier achieved an accuracy of 0.971 and a macro-F1 score of 0.950, with identical reported metrics under plaintext and ciphertext-based inference. The hybrid late-fusion similarity model achieved a Pearson correlation of 0.8691 with human sentence-pair similarity scores on STS-B. For ciphertext-based retrieval, we evaluated naive and optimized implementations. The optimized implementation uses CKKS single instruction, multiple data (SIMD) packing to compute and return multiple similarity scores together, reducing single-query latency over 7,128 indexed items from 222.028 to 5.168 seconds and communication overhead from 6.11 GB to 2.13 MB. These findings demonstrate the technical feasibility of privacy-preserving text inference for health-misinformation applications and characterize its predictive utility and efficiency trade-offs.
benchmark - arxiv:2610.20795 · cs.LGEEG-Based Prediction of Music Enjoyment Using Deep Learning with Explainable Temporal AnalysisParshav Singla, Simranjit Kaur, Maanya Walia, Niharika
Understanding why listeners enjoy music is difficult because the experience is subjective and changes as a piece unfolds. Electroencephalography (EEG) offers a non-invasive way to study neural responses during listening, yet earlier work has mostly treated music emotion as a categorical label or relied on black-box models. We formulate enjoyment prediction as a regression problem on the public MUSIN-G dataset, which contains 128-channel EEG from 20 listeners together with song-level enjoyment ratings. Six deep learning models were compared under a participant-wise protocol, in which 15 listeners were used for training and validation and 5 were held out for testing. DeepConvNet, CNN-LSTM and GNN-CNN reached the lowest mean absolute error (0.97 to 0.99 on the 1 to 5 scale), while the temporal convolutional network (TCN) had the highest Pearson correlation with the ratings (r = 0.21). Gradient-based and occlusion-based explanations, averaged over all held-out windows, identify the time segments within each two-second window that most influence the TCN predictions. The study provides a reproducible cross-listener benchmark for EEG-based enjoyment regression and shows how explainability analysis complements standard metrics.
benchmark - arxiv:2610.20794 · cs.CVBlame Errant Words, Not the Canvas: Defect-Aware Spatial Credit Assignment for Visual Text GenerationSujie Hu, Huaisong Zhang, Pujing Yang, Qian Wang +6
Recent reinforcement learning approaches have improved visual text generation, yet generated images still suffer from incorrect wording and malformed glyphs. We identify two coupled limitations: recognition-based rewards often overlook visible glyph defects, while whole-image scalar feedback obscures the uneven quality of individual text regions. These limitations motivate our central principle: $\textit{Blame errant words, not the canvas}$, to effectively preserve local quality distinctions. First, we introduce $\mathbf{TextDART}$, a defect-aware region-targeted reward model that localizes individual text regions and independently evaluates their semantic correctness and glyph quality, providing fine-grained, region-precise feedback. Building on TextDART, we propose $\mathbf{SCANNER}$, which turns these regional judgments into spatially differentiated policy feedback. It normalizes rewards for each text target across sampled images and casts the resulting regional advantages onto the corresponding latent tokens. Extensive experiments on existing benchmarks and our newly established $\textbf{PolyText-Bench}$ demonstrate that our approach achieves state-of-the-art performance in text accuracy and glyph fidelity. Project page: https://sujiehu.github.io/textdart-scanner
benchmark - arxiv:2610.20788 · cs.LGSliced Wasserstein Sketches with Error CertificatesElaheh Akbari, James Liu, Keaton Hamm, Soheil Kolouri
Compact representations that support geometrically meaningful comparisons are valuable for learning from probability distributions. We study optimal, free-weight $M$-atom quantization of one-dimensional probability measures in quadratic Wasserstein distance and extend it to higher dimensions via slicing. Using $L$ shared projections yields a sketch with $O(LM)$ comparison time. Our main result is an exact decomposition of squared distance error into residual and partition-mismatch terms, yielding computable certificates for arbitrary centroid sketches. For optimal sketches, a finite $(2+δ)$ moment for some $δ>0$ guarantees $O(M^{-1})$ error in the finite-direction sliced distance, while equal-mass sketches can converge more slowly. Each slice preserves its mean, while storing residual energies enables covariance recovery from sufficiently many generic directions. Experiments on document, point-cloud, and flow-cytometry benchmarks show that optimal sketches substantially reduce storage while preserving the classification performance of uncompressed sliced distances. They also improve generalization from few labelled subjects and support end-to-end learning as compact pooling layers.
benchmark - arxiv:2610.20786 · cs.AIEvolvingBrain: An Adaptive Skill System for Multi-modal Agents Self-Evolution in Open WorldsRunxuan Liu, Jingqi Shi, Baoxin Wang, Honghong Zhao +9
Multi-modal agents have made significant progress in interacting with open worlds. However, their ability to self-evolve remains limited. Supervised fine-tuning enables imitation but not active experience organization, while outcome-based reinforcement learning promotes exploration yet underuses intermediate trajectory information. Therefore, we propose EvolvingBrain, which uses an adaptive skill system as the brain for agent self-evolution, consisting of two following components. Stratified Skill Self-Evolution: extracts Procedural Patterns and Problematic Patterns from model-specific trajectories, summarizes experience by generating skills, and evolves through multiple interactions with the open world, ultimately forming a Stratified Skill Architecture consisting of Domain, Plan, and Action/Trap Skills, which provides the agent with a multi-granularity knowledge repository and a novel evolution approach. State-Aware Adaptive Execution: summarizes long-horizon interactions into Structured Abstractive Memory and uses it to guide the State-Aware Skill Fusion Strategy, enabling the agent to utilize knowledge according to its current state. Experiments on VAB-Minecraft and DiscoveryWorld demonstrate that EvolvingBrain improves success rates on open-world tasks and exhibits generalization capabilities. Our code is provided in the supplementary materials and will be publicly released in the camera-ready version.
memoryagent - arxiv:2610.20785 · cs.CLAPRiL: An Open-ended Benchmark for Chinese Social Awareness in Large Language ModelsTian Lan, Han Zhang, Shanshan Wang, Yahan Yu +2
Large language models are increasingly used in social interactions, where appropriate responses require recognizing implicit cues, inferring mental states, identifying conflicts, and considering potential consequences. However, existing benchmarks often focus on specific social abilities, constrained task formats, or particular interaction settings, offering limited coverage of social awareness across diverse contexts. To address this gap, we introduce APRiL, an open-ended benchmark for Chinese social awareness with instance-specific rubrics. APRiL contains 2,196 Social Awareness Evaluation Instances across six categories and two complementary tasks: Social Scenario Analysis evaluates models' understanding of social cues, mental states, and conflict structures, while Socially Aware Response assesses whether they can translate this understanding into appropriate, natural responses. Experiments on 12 models from six families reveal a consistent understanding-to-response gap. Higher-quality scenario analyses improve responses, while weak analyses may reduce response quality. APRiL provides a systematic resource for advancing socially aware language models in Chinese contexts. Our data and code are available at https://anonymous.4open.science/r/APRiL-05DE/.
benchmark - arxiv:2610.20782 · cs.AIOpen-world AI: a Tale of Three RequirementsJianyu Zhang, Léon Bottou
The closed world asks AI for specialized competence and lets it manufacture the iid data its theory assumes: a fixed criterion licenses collecting examples until the task is covered. The open world asks for versatility --- taking up tasks a machine was not built for, from few examples and with little prior from a human designer --- and offers data that is grouped, not shuffled. Its tasks are not known in advance, so a machine must do inference-time learning (1st requirement). Modeling tasks as sharing latent factors and decomposing the training objective, we derive two more: a *rich representation*, retaining more than iid generalization requires, and a predictive disentanglement that aligns inputs and outputs into matched low-dimensional groups. The analysis further predicts the need for a consistent inner learner, and the benefit of long contexts spanning related tasks. We test these claims on language models up to 8B, asking whether a model has a property (qualitative) rather than how much it wins (quantitative). Models that meet the requirements improve as data from related tasks accumulates; the rest degrade, and neither scale nor layer-type interleaving repairs this.
long context - arxiv:2610.20781 · cs.CVReasonUni: Reasoning-Augmented Reconstruction Advances Unified Multimodal ModelsFeize Wu, Zhiyuan Yan, Yuhuan Hong, Yuwei Niu +9
Image-to-text (I2T) understanding and text-to-image (T2I) generation are dual tasks in unified multimodal models (UMMs). Reconstructive post-training couples these tasks, but existing reconstruction-based methods remain perception-centric and provide limited supervision for reasoning about structure, intent, and transformation. We propose ReasonUni, a reasoning-augmented reconstructive post-training framework that extends this coupling beyond perception. Given an original image and an editing instruction, the UMM first reasons about what and how to edit and describes the target state, then generates the corresponding target image conditioned on this understanding output. To jointly optimize both modules, we introduce TreeGRPO, a tree-structured reinforcement learning method that samples multiple reasoning paths and candidate images per path, then aggregates visual rewards to update both modules under a unified objective. Experiments across multiple benchmarks demonstrate consistent gains in both understanding and generation, showing that reasoning-augmented reconstruction yields genuine mutual gains for UMMs.
post-trainingbenchmark - arxiv:2610.20780 · cs.AIARIA: Runtime Inspection and Amendment for Multi-Agent SystemsBolin Shen, Yushun Dong
In LLM-based multi-agent systems (MAS), a local mistake can propagate as other agents build on it, turning a correctable error into a failed task. Diagnosing completed runs or repeatedly re-executing them to search for corrections does not provide a learned policy for intervening while execution is still in progress. We present Agent Runtime Inspection and Amendment (ARIA), a framework for learning when and how to repair ongoing multi-agent executions. Our key intuition is to turn execution-verified repairs into reusable supervision, moving repair search into data construction and training so that a learned model can intervene at runtime without repeatedly testing alternatives. First, we collect over 24{,}000 successful and failed trajectories across four multi-agent frameworks, six benchmarks, and eight host models. From these trajectories, we construct the ARIA dataset, to our knowledge the first large-scale dataset pairing failed multi-agent trajectories with execution-verified repairs. It contains 3{,}953 verified repair pairs, each specifying what went wrong, where to intervene, and how execution should continue. Then, we develop the ARIA framework for MAS-agnostic runtime inspection and amendment, designed to integrate with any MAS that exposes execution trajectories and supports continuation replacement. Using verified repairs as training examples, we teach a compact model to recognize when intervention is needed and how to correct an ongoing run. The aim is to learn repair strategies that apply to related failures in new configurations, rather than merely reproduce individual corrections. Finally, the trained ARIA model demonstrates strong repair performance in both post-hoc and runtime settings, with generalization to unseen frameworks and host models. Our code is available at https://github.com/LabRAI/ARIA.
agentmulti-agentagent frameworkagent systembenchmark - arxiv:2610.20778 · cs.LGVerdict: Encode Once, Decide ManyAvdhesh Singh Chouhan, Jack Perry
Automating organizational work means answering many small typed questions about one piece of state (does this email need a reply, how urgent is it, who owns it) with probabilities good enough to act on without a human. The open typed-decision model we compare against, like a standard cross-encoder, answers each question with a separate pass over the input. We present Verdict-MQ, an open encoder that reads a case once and answers all of its questions: given identical retained state tokens after truncation (verified on every reported case), a structured attention mask makes the state encoding independent of the questions, keeps questions from influencing each other, and makes option order irrelevant by construction. On the typed-decisions benchmark, a 395M Verdict-MQ matches the soft gold labels significantly better than the released 421M Laya specialist on all five fresh seeds (KL 0.080 vs 0.117, Brier 0.045 vs 0.061), an advantage that survives per-system temperature scaling; its higher accuracy (0.801 vs 0.767) lies past the benchmark's own saturation point, so we read that margin as closer agreement with the benchmark's labelling teacher. On RACE, BoolQ and MultiRC, against a matched fine-tuned baseline it wins on MultiRC and loses on RACE and on BoolQ, where its training is unstable; we measure zero-shot references ourselves, including a closed general system (Jev) that beats every fine-tuned encoder. A controlled comparison measures what the shared state costs: giving each question its own question-aware copy of the state gains 0.9-3.5 points on RACE and MultiRC at 1.4-2.5x the per-case latency, and nothing measurable on typed decisions, with the invariance guarantees unchanged.
benchmark - arxiv:2610.20776 · cs.AIINFUSE: A Unified Surface Benchmark for Indirect Prompt Injection in Agentic SystemsThuy Dung Nguyen, Jingyi Guo, Rana Seda Dubauskas, Sreepriya Damuluru +2
LLM-based agents increasingly execute long-horizon, end-to-end workflows involving shell commands, code modification, dependency installation, and interaction with remote services. This growing autonomy expands their exposure to indirect prompt injection (IPI), where adversarial instructions embedded in external content are interpreted as legitimate task directives. Existing evaluations, however, often focus on isolated malicious strings, single-step tool calls, or narrowly scoped agent configurations. We introduce INFUSE, an end-to-end benchmark for evaluating IPI throughout the lifecycle of long-horizon agent workflows. INFUSE comprises 40 realistic tasks and a structured taxonomy of 56 attack scenarios spanning five threat categories. Attacks are injected through realistic interaction surfaces encountered during task execution, extending evaluation beyond attacks placed only in initial prompts or isolated tool interactions. We evaluate 19 LLMs from the GPT, Claude, Gemini, Llama, Gemma, and Qwen families, producing 41,824 adversarial trajectories (26,220 without defenses) and 720 benign reference runs. Across the benchmark, mean attack success reaches 19.6% under LLM-based evaluation and 15.4% under heuristic verification, with per-model ASR-L ranging from 3.2% to 41.9%. Existing guard models provide limited protection, while prompt-level instructional hardening reduces mean attack success from 58.2% to 14.3%, revealing substantial but model-dependent security-utility trade-offs. Our code is available at https://github.com/judydnguyen/infuse.
agentagenticbenchmark - arxiv:2610.20767 · cs.ROMulti-Dimensional Featurizers Yield Fine-Grained Steering and Data Shaping in Robot Foundation ModelsAiden Swann, Mozes Jacobs, Mathilde Papillon, Siddharth Boppana +11
Can better feature discovery improve control over robot behavior through internal interventions and training data? Recent work in vision and language models reveals multidimensional features, motivating block-sparse featurizers (BSFs), which represent each feature in a low-dimensional subspace. We investigate whether recovering this structure provides an interpretable interface for steering vision-language-action models and shaping the data from which they learn. We show that BSFs recover several multi-dimensional task-relevant features underlying the model's computation (e.g., perception, language, proprioceptive state, episode phase, and control variables). We causally validate these features through fine-grained steering: targeted changes in visual perception, semantic representation, and spatial localization produce predictable changes in downstream action prediction. Finally, we use feature-based data attribution to isolate where a specific behavior (failure recovery) arises in the training dataset, identifying fewer than $1.4\%$ of training tokens whose masking is sufficient to remove the targeted behavior. Overall, our results show how matching featurizers to representation geometry can provide more effective tools for understanding and changing robot behavior.
vision-language-actionrobot foundation model - arxiv:2610.20764 · cs.ROWhen Reality Outpaces Cognition: A New Scaling Principle for Reliable IntelligenceAhsan Adeel, Talha Bin Riaz
This paper challenges the understanding capabilities of existing AI paradigms and emerging world models by asking a fundamental question: How quickly can reality change before an intelligent system can no longer reliably perceive, reason, and act? We formulate this as a theory- and model-agnostic challenge in which systems must infer reliable actions for increasingly distant future states across a growing observation--reality gap. We operationalize this challenge through the effective Speed of Coherent Cognition (\(S_C^{\mathrm{eff}}\)), defined as the highest tested Speed of Reality (\(S_R\)) through which a system continuously satisfies a predefined behavioural reliability criterion from baseline. As a proof of concept, reinforcement-learning agents are trained at the baseline \(S_R\) and evaluated, without retraining, as \(S_R\) is increased by up to \(50\times\) while computational resources remain fixed. We compare Transformer and gated-MLP baselines with \(Co^4\), an architecture inspired by cellular mechanisms implicated in conscious processing and awake thought. Across the tested scenarios, \(Co^4\) achieves substantially higher \(S_C^{\mathrm{eff}}\) than the baselines, motivating \(S_C^{\mathrm{eff}}\) as a new benchmark and the scaling of coherent cognition as a principle for reliable AI.
world modelbenchmark - arxiv:2610.20750 · cs.ROCausal Analysis of Robot Foundation Models Enables Efficiency GainsAiden Swann, Mathilde Papillon, Mozes Jacobs, Siddharth Boppana +12
Vision-language-action (VLA) models are a promising approach to general purpose robot manipulation. While the internal computation of vision-language models, which compose the backbone of VLAs, has been studied extensively, less is known about how their perceptual and semantic representations are transformed into action in VLA models. We study this question mechanistically in three frontier VLA models ($π_{0.5}$, MolmoAct2, and MolmoBot) by analyzing the causal pathways through which the input image and the instruction shape actions. Across the three models, activation patching shows that causally relevant information influences the action expert (AE) through a narrow band of mid-to-late cross-attention layers, which we term the action-conditioning interface. The content of this information changes with token and layer within the model. Early-to-mid layer representations of the target noun token determine which object is selected, whereas mid-to-late layers carry the spatial information in text tokens that redirects the movement when patched. Where spatial information is stored in prompt tokens various based on the architecture. In MolmoAct2 the model forms task-conditioned representations at the text positions that follow the instruction, whereas $π_{0.5}$ additionally stores this information at image token positions. We show that these findings can be translated directly into improving model efficiency. We find that restricting MolmoBot's action expert conditioning to only 3 out of 36 layers reduces inference FLOPs by up to $33\%$ while maintaining comparable task success with the default model across 3,000 evaluation episodes ($46.7\%$ vs. $46.9\%$), and the same principle carries over to real-robot deployment of $π_{0.5}$.
vision-language-actionvlavla modelmanipulationrobot foundation model - arxiv:2610.20739 · cs.LGQuantile-Guided Diffusion for Probabilistic Time-Series ForecastingShadi Omari, Bashar Khoury, Assaf Schuster
Quantile forecasts summarize uncertainty at each future time step, while diffusion forecasters represent it through sampled future trajectories. We study how predicted quantiles can serve as distributional controls for diffusion forecasting. Specifically, we investigate how the mechanism of quantile integration affects its effectiveness in guiding a pretrained diffusion forecaster. We find that residual control throughout the denoising network yields the most consistent improvements. We instantiate this approach as Control-TS, which learns a quantile-control pathway while keeping the pretrained backbone frozen. Across eight benchmarks, Control-TS consistently improves the backbone's CRPS, achieves competitive performance against state-of-the-art forecasting models, and retains its gains in capacity-matched comparisons. Further analyses examine how these gains depend on quantile granularity, predictor choice, and forecast quality. Finally, we introduce the Conditional Linear Predictive Score (CLPS), which complements distributional metrics by evaluating the downstream forecasting utility of generated futures. Together, these findings support predicted quantiles as an effective interface between explicit uncertainty estimates and diffusion-based trajectory generation.
benchmark - arxiv:2610.20731 · cs.CVWorldCast: Exploring Intelligent Visual Generation in World State PredictionWenhao Yan, Haodong Li, Xiaotong Li, Yaxin Luo +8
Humans interpret multimodal observations of the world and forecast how it will evolve. Recent multimodal generative systems have demonstrated similar capabilities: they can integrate multimodal context and synthesize visual outputs that depict changes in the world. To characterize this emerging trajectory, we organize model capabilities into a four-level hierarchy and discover that higher-level tasks remain relatively underexplored despite model progress. To bridge this gap, we explore model performance on multi-step world state prediction, a high-level task that requires models to forecast successive future world states from multimodal observations and diverse references. Such a task requires models to possess entangled understanding and generation capabilities, as well as extensive world knowledge to translate human action sequences, physical transformations, and abstract visual reasoning into coherent visual outputs. We introduce WorldCast, a benchmark comprising 320 manually designed cases across eight categories, and evaluate 12 representative models. The conducted experiments yield three key insights: (a) advances in frontier models are rapidly pushing the boundaries of intelligent world modeling, with the GPT-Image-2.5 System scoring 72.8 and demonstrating early-stage visual state prediction capabilities on real-world tasks; (b) WorldCast reveals substantial gaps between models: even Nano Banana Pro reaches only 42.8 and the best open-source model scores 20.1, with deficits in handling complex context, entangled references, and reasoning-heavy tasks; (c) strong generative models effectively draw on the capabilities of their underlying language models. These findings highlight the need for multimodal generative models that are capable of dynamic planning and reasoning while also being deeply unified to faithfully simulate real-world states.
world modelbenchmark - arxiv:2610.20693 · cs.CVOrbia: Do Generated Videos Form a Consistent 3D World?Dongyue Lu, Guian Fang, Hur Lim, Yonggang Wu +19
Video generators serve as interactive worlds for exploration and simulation, yet realistic and visually appealing frames do not necessarily depict a consistent 3D world. Existing benchmarks assess selected geometric properties or use vision-language models to judge memory and consistency, leaving systematic 3D evaluation underexplored. We introduce Orbia, the first benchmark dedicated to a systematic and comprehensive evaluation of 3D consistency in interactive world models. Through controlled exploration and revisit, we organize the evaluation into three dimensions: input preservation, generated-content persistence, and 3D self-consistency. For input preservation, registered RGB-D references and annotated anchors support appearance, geometry, and spatial stability measurements. For generated-content persistence, we compare revisits to regions absent from the input image with their first generated observations to test whether models preserve what they create. For 3D self-consistency, cross-view depth and feature correspondence, regional scale agreement, and held-out Gaussian rendering test whether generated views describe the same 3D scene. Minute-long exploration and revisits further test long-term consistency. Our construction and annotation pipeline brings together public datasets, self-captured scenes, and simulation, yielding 720 instances across indoor and outdoor environments and six exploration-and-revisit protocols. Evaluation spans 27 text-, action-, and camera-conditioned models through native-interface adapters. Our Orbia benchmark provides a systematic basis for diagnosing 3D inconsistency and developing more persistent video world models.
world modelmemorybenchmark - arxiv:2610.20690 · cs.AIOpen Benchmark of AI Impact on Humans (ImpactBench): an expert-guided, open platform for the holistic evaluation of AI impact on humansChayapatr Archiwaranguprok, Krystal Kirkland, Rachel Poonsiriwong, Stanley Huang +16
The present work is concerned with how to measure, at scale, the multi-dimensional impact of generative AI models on the individuals who use them. In service of this goal, we introduce the Open Benchmark of AI Impact on Humans, or ImpactBench. Our iterable and expert-guided platform evaluates whether model behaviors support or undermine an individual's wellbeing across psychological, physical, and social domains. ImpactBench translates constructs from domain experts and existing benchmarks into metrics that are grounded in actual model behavior. Success on these metrics occurs either when a model promotes expert-defined desirable behaviors or avoids identified harms. Models are evaluated against these metrics in simulated multi-turn conversations using matched teen and adult scenarios. The conversations are then scored by an LLM judge and verified through reliability audits and human expert review. The current suite delivers 809 metrics drawn from 26 benchmarks. We evaluated ten frontier LLMs against these metrics across 48,540 conversations. Our results indicate that models promoted desirable behaviors in 68.7% of evaluations, but avoided harms in only 53.1%. Majority expert ratings agreed with the automated judge in 83.3% of reviewed cases (Cohen's Kappa = 0.667). Results are released through a public platform that keep aggregate scores traceable to the underlying metrics and transcripts. ImpactBench is intended as evolving infrastructure for independent, transparent evaluation of generative AI's multi-dimensional impact on humans.
benchmark - arxiv:2610.20686 · cs.LGRoLaQ: Joint Rotation and Range Calibration for 2-Bit KV Caches in Multi-Head Latent AttentionShiyu Ji, Peng Ling, Chen Lin, Shupeng Li +3
Multi-head latent attention (MLA) compresses the content of a key-value cache into a shared latent per token, making low-bit storage attractive for long-context inference. At two bits, errors in this representation perturb both attention matching and value readout, with effects that depend on error direction and reconstruction scale. We introduce RoLaQ (Rotation-calibrated Latent Quantization), an offline method for joint rotation and range calibration. RoLaQ refines a shared rotation within subspaces defined by the frozen key and value projections, and jointly learns quantization ranges whose thresholds and reconstruction gains are coupled to control the shared signal's response. Prediction matching and attention-output supervision guide calibration with model weights frozen. Across three MLA models and six benchmarks, RoLaQ achieves the highest average accuracy among the compared 2-bit methods both without and with token protection. On DeepSeek-V2-Lite, an SGLang implementation provides 3.96 times the BF16 KV capacity under the same memory budget and 2.51 times its decode throughput with 256K input tokens.
memorylong-contextbenchmark - arxiv:2610.20682 · cs.LGTurboMamba: Faster Inference, Faster Training, and Better Prediction with Lexical MemoryHanxin Zhang
We introduce TurboMamba, a language-model architecture for faster inference, faster training, and better token prediction. A global Mamba-3 model runs once per completed token block. A small Transformer uses this global context and earlier tokens within the block to predict the next token. Three lightweight additions improve local predictions: direct access to the previous block's last token, a small correction to next-token scores, and a learned mixture with recent-token frequencies. We compare twelve small-scale implementations on public FineWeb-Edu, with comparable parameter counts and shared training-budget caps. The baselines include Transformer, Mamba-3, Jamba, and Zamba. The Jamba and Zamba versions use Mamba-3 in place of their original Mamba modules. TurboMamba achieves the lowest mean test prediction loss (negative log-likelihood, NLL) among all twelve models. Its mean NLL is 9.4% lower than that of Mamba-3, 7.3% lower than that of the Jamba version, and 18.3% lower than that of the Zamba version. End-to-end training throughput is 2.60x that of Mamba-3, 2.81x that of the Jamba version, and 4.82x that of the Zamba version. We measure end-to-end batch latency for a batch of 32 sequences, each with a 256-token prompt and 128 generated tokens. On this workload, TurboMamba achieves a 3.12x speedup over Mamba-3, a 3.53x speedup over the Jamba version, and a 6.21x speedup over the Zamba version.
memory - arxiv:2610.20681 · cs.AIIneqAgent: Agentic Inequality Proving in Lean with Symbolic ComputationGaolei He, Jinhui Li, Ruobing Zuo, Xia Zeng +1
Formal inequality proving requires mathematical reasoning and exact computation, often connected through intermediate lemmas and goal reformulations. We introduce IneqAgent, an agentic framework combining LLM reasoning with symbolic computation for inequality proving in Lean4. Lemma-level Monte Carlo Tree Search coordinates lemma construction, goal reformulation, and symbolic backend selection. Complementary sum-of-squares, Bernstein, and successive difference-substitution backends produce exact certificates for local Lean proof generation. The agent integrates these proofs with LLM-generated steps and uses Lean feedback to construct complete proofs checked against the original theorem statements. On four benchmarks, IneqAgent with DeepSeek-V4-Flash solves 175 of 186 problems (94.09%), achieving the highest overall success rate among the evaluated methods. Across four backbones, it achieves higher success rates at low LLM inference cost thresholds than Direct Agent using the same backbone, with lower mean inference costs over each method's successful runs. We also obtain Lean-verified strengthenings of Dittert's conjecture in dimensions four and five, and certify the Holens--Djokovic $(4,4)$ inequality, implying Merris's conjecture in dimension four.
agentagenticbenchmark - arxiv:2610.20675 · cs.LGAlphaDiana: A Measurement System for Agents in Verifiable Reasoning TasksZhanke Zhou, Zongze Li, Weikai Huang, Xuan Li +15
Agents pair reasoning models with harnesses that organize model calls, execute tools, load skill instructions, and retain history in memory. Understanding agent performance requires linking execution conditions to outcomes, behavior, and resource use. We introduce AlphaDiana, a measurement system with modular interfaces for datasets, agents, environments, and scorers. It preserves harness-native control loops while controlling file reuse and memory retention separately. AlphaDiana supports macro analyses of complete agents and micro interventions on agent components such as tools, skills, and memory under matched conditions. We study three open-source models and three open-source harnesses on five answer-based benchmarks, with two models also tested on two interactive benchmarks. We find that harnesses do not consistently improve the accuracy of reasoning: gains on questions unsolved by the model alone can be outweighed by losses on questions it answers correctly without tools. This inconsistency extends to component interventions: neither skill guidance nor retained history reliably improves accuracy. Behavior and output-token use must be assessed with correctness: low token entropy can accompany wrong answers, and less output per execution can mean more output tokens per correct answer when accuracy falls. AlphaDiana links answer correctness and execution status to traces and resource records for comparing agents and diagnosing failures under shared task and resource policies.
memoryagentbenchmark - arxiv:2610.20667 · cs.LGEvaluating Model Capabilities After Benchmark SaturationMorio Hamada, Jehyeok Yeon, Matteo Migliarini, Justin Olive +1
Rapid advances in AI capabilities are causing benchmarks to saturate. Once several models approach perfect scores, a benchmark loses its power to differentiate between them, and with it its utility for informing AI safety cases and frontier policies. Yet benchmark items produce far more information than the scores we typically extract from them: a full transcript of how each model solved the task. We investigate whether these transcripts can be used to assess model capabilities even after benchmark scores have saturated. We simulate saturation retrospectively using historical benchmark and model releases, and test whether protocols that access only correctly solved transcripts from saturated benchmarks can predict capabilities measured on held-out, unsaturated items. We evaluate representative contemporary approaches for extracting information from model transcripts: pairwise judging, single-transcript absolute scoring, and predictions from deterministic transcript features or embeddings. We find that saturated transcripts retain capability information beyond benchmark accuracy. Amongst our evaluated protocols, pairwise judging performs best, reducing prediction error by 5.3 times compared to a baseline using only saturated benchmark scores. This capability-relevant information is distributed across transcript sections, but concentrated in specific channels of information such as reasoning and tool-calls. However, the predictive power of protocols does not always generalise across distribution shifts -- for example by underestimating the capabilities of the first reasoning models introduced in early 2025. Saturated benchmarks thus retain capability-predicting signal in their task transcripts, but extending the lifetime of existing benchmarks requires methods designed to extract these signals robustly.
benchmark - arxiv:2610.20665 · cs.AIThe World Won't Wait: Evaluating and Teaching Agents Under a Live ClockNingxin Pan, Hanyu Li, Yeqing Shen, Ning Ding +2
Large language model (LLM) agents can solve complex interactive tasks, yet reasoning and tool calls take time. In continuously changing environments, an otherwise correct action may arrive too late. We introduce RealtimeBench, with 180 tasks across six game types, to evaluate agents while games continue running during inference. Paired evaluations show that GPT-6.1-sol outperforms the faster Jev when games are paused, but falls behind when they keep running. We propose Two-Timescale Offline Recursive Self-Improvement (TORSI): a slow teacher studies failures and tests corrections in paused games between attempts, then writes reusable instructions for a fast student. The student acts without teacher calls during gameplay, and validation scores determine which updates are adopted. Two rounds raise Jev's average test score from 13.51 to 24.19 on a 0--100 progress scale, improving all six game types. The instructions also improve average scores for three other models without retuning. These findings point toward real-time intelligence: enabling powerful LLM agents to remain effective in a continuously changing world by organizing deliberation, learning, and action across timescales.
llm agentself-improvement - arxiv:2610.20662 · cs.LGGraph Additive Models disentangling feature, topology, and interactionsClara Alvarez-Rodriguez, Steve Azzolin, Vincenzo Marco De Luca, Filippo Maria Bianchi +2
Graph Neural Additive Networks (GNANs) have emerged as effective, interpretable-by-design architectures for graph learning. They achieve transparency by combining per-feature shape functions with simple distance-based reweighting. However, GNANs have three limitations. First, node-wise relevance scores aggregate information from all graph nodes, allowing irrelevant nodes to inherit spurious relevance from relevant ones. Second, topology is represented only through shortest-path distances, so structurally distinct graphs may share identical distance profiles. Third, feature and distance information are combined multiplicatively, entangling feature, distance, and interaction effects in a single rescalable term. We introduce AMIGA, which represents topology through a vocabulary of mined motifs and decomposes each node contribution into three additive channels: feature, structure, and feature-structure interaction. The channels are centered around a graph-balanced training baseline, making the decomposition well-defined and unique under the chosen reference distribution. Across ten graph classification and regression benchmarks, AMIGA outperforms GNAN on most datasets and narrows the gap with standard black-box GNNs. Because each channel is centered and assigned to a single node, AMIGA produces node-local explanations that reconstruct the model output exactly and support interpretation at multiple levels of granularity.
benchmark - arxiv:2610.20659 · cs.LGLatent Annealed Posterior Sampling with a Variance-Preserving Bridge for Sparsely Observed Random FieldsJiankun Lei, Tian Qi
Posterior inference for random fields from sparse observations presents two major challenges: accurately characterizing complex spatial uncertainty and preventing conditional sampling costs from growing substantially as the spatial grid is refined. To address these challenges, we propose latent annealed posterior sampling (LAPS). LAPS first maps the random field into a latent space with reduced dimensionality, while a coordinate decoder composed of Branch and Trunk networks enables field evaluation at arbitrary spatial locations. Flow matching is then used to learn the latent prior distribution. Based on this learned prior, posterior sampling is performed entirely in the latent space, with the observation likelihood incorporated progressively through an annealing procedure. To control local covariance inflation, we introduce a variance-preserving (VP) bridge that allocates a covariance budget between the transported latent state and a Gaussian perturbation. Numerical experiments demonstrate that LAPS accurately approximates posterior distributions, preserves distinct posterior modes, and reduces the dependence of conditional sampling costs on spatial resolution. Across two Gaussian benchmarks, LAPS attains small errors against exact posterior references, with sliced Wasserstein distances only 2 to 3 percent higher than the reference values obtained from exact posterior samples. In a bimodal posterior problem, LAPS reduces the sliced Wasserstein distance from 0.2693 to 0.1499 and consistently recovers both modes across all test scenarios. For a two dimensional Navier-Stokes random field, LAPS substantially improves marginal posterior calibration, while its evaluation time remains nearly constant as the output grid is refined from 64^2 to 512^2. These results demonstrate that LAPS achieves a favorable tradeoff among posterior accuracy, uncertainty calibration, and computational efficiency.
benchmark - arxiv:2610.20652 · cs.CLDoes Personalization Break Trust? Open-Set Rejection Calibration Under Few-Shot On-Device Intent AdaptationVaibhav Prajapati, Pushp Garg
Always-on virtual assistants increasingly personalize on-device to new user-specific commands via few-shot fine-tuning (e.g., LoRA), but this adaptation is evaluated almost exclusively on whether it learns the new commands, not on whether the assistant can still recognize an input that matches none of its intents - the out-of-scope (OOS) rejection problem. We show that naive few-shot LoRA personalization degrades OOS-rejection AUROC (area under the ROC curve) on every one of five public intent-classification benchmarks (mean drop $0.055$--$0.258$, five seeds each), while in-scope accuracy is preserved. A scorer-invariant check shows part of this drop is an artifact of the standard confidence scorer's own miscalibration under adaptation, though a smaller, genuine effect remains under any scorer; we trace it to a measurable shrinkage of the embedding margin around known-intent regions (pooled Spearman rho=0.61$). A deployment-cheap fix is scoring rejection by Mahalanobis distance instead of softmax confidence without extra training/retraining recovers most of the lost quality on 3 of 5 datasets at no accuracy cost, beats two further OOD scoring baselines, and generalizes across shot count, LoRA rank, and a second base encoder. We report honestly where the fix falls short and release all code for reproducibility.
benchmark
02 US SEMI · SEC 8-K FILINGS
1 itemsscanned: NVDA / AVGO / MRVL / COHR / LITE / AMD / TSM / SMCI / ANET / CRDO / POWL / VECO
03 HUMANOID · COMPANY NEWS
62 itemsscanned: figure-ai / 1x / boston-dynamics / unitree / apptronik / sanctuary-ai / neura-robotics / agility-robotics / physical-intelligence / agibot
Figure AI (10)
- Figure AISeptember 30, 2026F.02 Decommission
- Figure AISeptember 17, 2026Helix 2.5: Zero-Shot 30-Home Generalization
- Figure AIOctober 09, 2025Introducing Figure 03
- Figure AISeptember 03, 2026Figure and Nscale Sign Strategic Partnership For Up to 100,000 GPUs on the NVIDIA Vera Rubin Platform
- Figure AIAugust 25, 2026Introducing Index: Building The World’s Largest and Most Diverse Physical Dataset
Boston Dynamics (10)
Unitree 宇树 (10)
- Unitree 宇树Components
- Unitree 宇树Kung Fu Meets Spring, Unitree SFG Robots Present "Cyber Real Kung Fu" in the Year of the Horse2026-05-31Media Coverage
- Unitree 宇树Welcoming Myanmar President Min Aung Hlaing to Unitree2026-08-05Media Coverage
- Unitree 宇树Unitree founder Wang Xingxing graces the cover of Time magazine2026-08-05Media Coverage
- Unitree 宇树Unitree Announces H2 Plus, an NVIDIA Isaac GR00T Reference Humanoid Robot for Academic Research2026-06-01Media Coverage
Apptronik (1)
Sanctuary AI (6)
- Sanctuary AIBlog PostTIGER: Task-Space Imitation Guidance for Efficient Reinforcement LearningDemonstrations as a compass, not a scriptRead More
- Sanctuary AIPress ReleaseProduct UpdatesSanctuary AI Expands Physical AI Strategy to Industrial Robotics, Demonstrating Production-Ready AI PerformanceRead More
- Sanctuary AICorporate NewsDaniel Friedmann Appointed CEO of Sanctuary AIRead More
- Sanctuary AIPress ReleaseZeon Invests in Sanctuary AI and Partners to Advance Specialized Materials for Dexterous RoboticsRead More
- Sanctuary AIThought LeadershipWeb Summit Reflections: Canada’s Physical AI Moment Can’t WaitRead More
Agility Robotics (10)
- Agility RoboticsAgility’s Humanoid Deployment ProcessPras VelagapudiAugust 04, 2026
- Agility RoboticsThe Realistic Pathway to HomeInsightMay 26, 2026
- Agility RoboticsAgility and AIInsightMarch 16, 2026
- Agility RoboticsAgility Gets a New BrandInsightMarch 5, 2026
- Agility Robotics2026: The Automation EvolutionInsightJanuary 16, 2026
Physical Intelligence (8)
- Physical IntelligenceNewFrontier Models for Every RobotOctober 13, 2026Our latest robotics models are now available through the Pi API beta. Upload data, fine-tune models, and deploy cloud-hosted inference on your own robots.
- Physical Intelligenceπ0.7: a Steerable Model with Emergent CapabilitiesApril 16, 2026A steerable robotic foundation model that exhibits a step-change in generalization.
- Physical IntelligenceThe Physical Intelligence LayerFebruary 24, 2026General-purpose physical intelligence models will enable a Cambrian explosion of robotics applications. See how our partners are already solving real-world problems.
- Physical IntelligenceMoravec's Paradox and the Robot OlympicsDecember 22, 2025By fine-tuning our latest model, we were able to solve a series of very difficult manipulation challenge tasks.
- Physical Intelligenceπ*0.6: a VLA that Learns from ExperienceNovember 17, 2025A method for training our generalist policies with RL to improve success rate and throughput on real-world tasks.
智元 AgiBot (7)
- 智元 AgiBotAGIBOT Opens Hong Kong Office in Causeway Bay2026-10-09
- 智元 AgiBotAGIBOT Partners with Vietnam National Innovation Center(NIC) to Advance Embodied AI Industry and Talent Ecosystem2026-10-02
- 智元 AgiBotIDC Data Shows AGIBOT Led Global Humanoid Robot Shipments in H1 20262026-09-30
- 智元 AgiBotAGIBOT and ASD Deploy 100 AGIBOT X2 Humanoid Robots Across 100 Retail Stores2026-09-29
- 智元 AgiBotAGIBOT and Chimelong Launch Large-Scale Embodied AI Theme Park with More Than 300 Robots2026-09-23