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.
389 items today · 316 arxiv · 11 SEC 8-K · 62 humanoid · 0 CN photonics
01 ARXIV · PHYSICAL AI PAPERS
316 items- arxiv:2611.08485 · cs.CVWhere and Why is Grounded Reasoning Useful?David Yan, Le Xue, Forrest Huang, Qiuyi Zhang +3
Grounded reasoning, where a model outputs text coordinates of relevant locations in chain-of-thought reasoning, is a common strategy for solving complex visual reasoning problems. However, the benefits and limitations of grounded reasoning are understudied. We study where and why grounded reasoning is most useful. To do so, we introduce a chain-of-thought analysis method to measure how much a model uses its outputted grounding in its thinking trace and perform a detailed ablation study of a grounded reasoning training recipe. We find that grounded reasoning improves counting but shows no significant gain over baselines without coordinate grounding across a wide range of spatial reasoning and perception benchmarks. Using several synthetic tasks, we show that grounded reasoning is most useful when a task requires undescribable references, which are visual concepts that are difficult to describe in language. However, we observe that many existing visual reasoning benchmarks are solvable with only describable text references. To close this gap, we introduce BrickVision, a new benchmark that requires undescribable visual references to solve.
benchmark - arxiv:2611.08483 · cs.AISwarmEvolve: Latent Agent Swarms for Evolutionary Discovery and OptimizationQile Jiang, George Em Karniadakis
LLM-guided evolutionary search uses AI agents to improve programs for mathematical and computational problems by iteratively refining solutions and combining high-scoring methods. However, the process is often slow because the LLM generates a long textual chain of thought for each round and must repeat this generation many times to maintain population diversity. We present SwarmEvolve, a framework in which hundreds of low-cost agents per round work in swarms, reason in latent space without decoding to text, and share information by passing KV caches. Each swarm studies the problem from a different angle. A novelty filter keeps the most distinct agent ideas by directly measuring the distances between the agents' hidden state vectors. Crossover of ideas happens naturally in latent space, as a summarizer agent receives the combined KV caches of all swarm members. On 21 problems in mathematical discovery and scientific computing, SwarmEvolve gets the highest average score and needs a median of only 3 to 8 minutes to reach the best scores that other state-of-the-art frameworks reach in 60 minutes. On several problems, SwarmEvolve with Qwen3 8B improves on the solutions that AlphaEvolve first found with frontier LLMs. Our work shows that agent swarms that communicate in latent space offer a fast, low-cost approach to program search for scientific problems.
agentai agent - arxiv:2611.08477 · cs.CVEgoGeneralist: Unifying Egocentric Action-Video Tasks through Video-Native ModelingYanlin Jin, Bowen Yang, Cheng Zhang, Yifan Yang +3
Modern video models possess strong visual priors, yet do not natively represent human action in metric 3D space. We present EgoGeneralist, which adapts pretrained video generators to unify action-conditioned egocentric video generation, pose estimation, and joint action-video generation. Our role-adaptive action renderer provides video-native representations for multidirectional action-video learning, making metric motion usable as both a spatial control signal and a recoverable generation target. Cross-stream action-video coupling establishes close correspondence between metric motion and visual dynamics through lightweight adaptation while preserving the pretrained visual prior. Together, these designs establish a generalist framework spanning five action-video tasks. Despite adaptation only on controlled laboratory recordings, EgoGeneralist generalizes to diverse unseen scenes and interactions with effective action control and visual quality, bridging pretrained visual knowledge and metric 3D human motion.
action-conditioned - arxiv:2611.08472 · cs.LGXStack: A Benchmark for ML Repo MigrationLinrong Cai, Taiming Lu, Haozhe Chen, Zhuang Liu
ML code should run where researchers choose, not where it was first written. Today, machine learning code mainly lives in two ecosystems, PyTorch on NVIDIA GPUs and JAX on Google TPUs. Migrating repositories between them means building and testing across frameworks and hardware. This migration is (1) a practical need, since it unlocks different hardware for researchers, and (2) a natural benchmark: it involves different frameworks and hardware, covers whole repositories, and is long-horizon and challenging. We introduce XStack, a benchmark that evaluates agents on migrating ML repositories between these two ecosystems. Each agent receives a PyTorch repository and a clean TPU machine and must deliver a JAX repository. Its output passes if it meets our grading criteria, based on the source repository's published result, authenticated and calibrated by us. Migration by agents remains unreliable: Claude Opus 5 passes 75%, GPT-5.6-sol 45%, and every other agent at most 40%. XStack is a step toward freeing ML code from its ecosystem. We release XStack to make progress toward reliable migration across frameworks and hardware measurable.
agentbenchmark - arxiv:2611.08462 · cs.ROOopsieRecover: Learning Transferable VLA Corrections from Semantic InterventionsArpit Bahety, Skand Peri, Sriniket Ambatipudi, Chen Xu +3
Vision-Language-Action (VLA) models can perform a wide range of robotic manipulation tasks, but often fail during deployment. Teaching them to recover from these failures requires substantial corrective data: existing approaches learn to recover from failures by directly associating corrective actions with the observations in which they were demonstrated. For instance, DAgger-like fine-tuning can require tens to hundreds of interactions, yet the resulting policy may still fail to generalize the learned corrections beyond the demonstrated scenarios. We hypothesize that semantics can capture the underlying reasons for failures and their corrections, enabling policies to learn from only a handful of corrective demonstrations and generalize those corrections to novel objects, and scene variations. We propose OopsieRecover, a framework that uses abduction to infer the intent behind a small set of corrective demonstrations and encodes the inferred knowledge as dense spatial representations and conditional costs. These costs capture when a behavior is undesirable and steer the policy toward more desirable action regions. We experiment across 6 tasks and variations in object poses, scenes, and related tasks, and find that only a handful of corrective demonstrations can enable a VLA to overcome failures across a broad range of novel situations, without requiring new corrective data for each variation. Please refer to https://robin-lab.cs.utexas.edu/oopsierecover for videos.
vision-language-actionvlamanipulation - arxiv:2611.08460 · physics.opticsRoom-Temperature Quantum Emission from Boron Nitride NanotubesKe-Sean J. M. Peter, Ryan A. Kowalski, David A. Curie, Dakota Braaten +6
Quantum information systems such as quantum key distribution and quantum communications require bright, high-purity, stable sources of single photons to act as carriers of quantum information. As such, methods for obtaining single photon emitters (SPEs) in solid-state devices capable of articulating with quantum photonic circuits are a necessary obstacle that must be overcome. Hexagonal boron nitride (hBN) flakes have been extensively studied as a host material for SPEs, though working in two dimensions is accompanied by engineering complexity. However, dimensionality can be reduced through the implementation of boron nitride nanotubes (BNNTs). Here we investigate the potential of BNNTs as room-temperature SPEs using confocal photoluminescence microscopy and Hanbury Brown-Twiss interferometry. Quantum emitters were found in BNNTs spanning the entirety of the visible spectrum, including spectral regions where SPEs within hBN have not been reported. Furthermore, each tube was found to possess multiple SPEs at varying emission wavelengths along the length, exhibiting lifetimes as fast as 0.5 ns, which is about two-fold faster than the fastest hBN SPEs. Additionally, we demonstrate that surface passivation/coating mitigates blinking events, therefore increasing the emitter stability. Thus, BNNTs can function as a single structure supporting a variety of multiwavelength SPE sources, while maintaining a compact radius of ~25 nm, making them ideally suited for deterministic placement and integration within nanophotonic devices such as waveguides and interferometric filters.
quantum photonic - arxiv:2611.08458 · cs.ROLatent RL Priors for Data-Efficient Dexterous ManipulationYixiao Wang, Peiyi Chen, Mustafa Mukadam, Shreekant Gayaka +2
Learning dexterous manipulation from limited visual demonstrations is challenging because policies must acquire both perceptual grounding and precise control, while reinforcement learning (RL) in simulation can acquire rich control behaviors from large-scale interaction with privileged information. We introduce \textbf{Latent RL Prior}, a two-stage framework that reuses this control knowledge for data-efficient visual policy learning. During privileged pretraining, we distill large-scale RL rollouts into a flow-matching action decoder conditioned on proprioception and latent task context. During visual adaptation, we freeze the decoder and learn a visual interface that maps images to its latent context using action supervision alone, without privileged labels or latent targets. The same decoder can further be shared across arms by learning coordinated visual contexts for bimanual manipulation. Across three simulated tasks, our method outperforms full-parameter fine-tuning of $π_{0.5}$ and GR00T N1.7 in the low-data regime; on door lifting and bimanual assembly, it surpasses full-data baselines using $10\times$ and $5\times$ fewer demonstrations, respectively. Further analysis shows that direct visual-to-latent conditioning improves adaptation performance and training stability over routing visual predictions through the pretrained state encoder. Robustness to unseen mass variation and greater linear accessibility of contact-force information in decoder representations further provide evidence of retained RL-acquired control knowledge. Project website: https://yixiaowang7.github.io/latent-rl-prior-page/.
manipulationdexterousgr00t - arxiv:2611.08452 · cs.AIReproBridge: Building Reproducibility into ML Research ArtifactsYubo Wang, Rui Meng, Jaehyun Nam, Tomas Pfister +1
Reproducing machine learning experiments requires reconstructing procedures from information scattered across papers, released code, and supporting artifacts. For coding agents, reproduction success can therefore depend on their ability to fill in missing details, rather than on the released artifacts alone. We ask whether preserving recovered procedures and decisions improves subsequent independent reproduction. We define a standardized artifact design comprising a refined codebase, environment setup, experiment launchers, and a reproduction guide, and introduce ReproBridge, an agent workflow for constructing these refined reproducible artifacts. ReproBridge-Bench provides 1,029 tasks from 387 experiments across 36 papers, anchored by five human-verified reference scores per task. Agents attempt tasks in fresh sessions with raw code or refined artifacts; a separate Auditor checks execution evidence. Across four evaluated models, refined artifacts raise mean reproduction success from 76.4% to 92.6% and reduce downstream API spending by 19.7% on average. The cross-model success-rate spread narrows from 21.6 to 3.9 percentage points, with all four models achieving mean success rates above 90%. Removing any component reduces success in the ablation study, with the largest losses from code and environment removal. These results suggest that preserving execution knowledge helps even models with lower raw-code success reproduce experiments more reliably and at lower API cost.
agent - arxiv:2611.08445 · cs.LGA Hidden Markov Preference Model for Online Prediction of Adversary ActionsAngel Daruna, Michael Cogswell, Tixiao Shan, Yi Yao +2
A defender anticipating an attacker needs a probability distribution over the attacker's next action from information available before that action occurs. Two structures complicate this: the attacker's behavioral regime persists and can switch within an episode, and each action combines an operation with state-dependent arguments whose feasible values change over time. We capture both with a hidden Markov model whose latent state is the active regime and whose regime-conditioned action law is defined on a feasibility-masked tree of executable actions. The same likelihood forms the next-action distribution and updates the regime belief once the action is seen, and a partially observed prefix contributes its resolved local factors. Estimation separates into convex logistic fits that Baum-Welch reuses when regimes are unlabeled. We derive a reach-weighted nodewise KL expression for regime information, a sequence-wise log-loss bound for switching regime paths, and a uniform fallback guarantee. On the CAGE 2 cyber defense benchmark the filtered latent state is worth 0.8 nats per step, and after a controller switch the loss begins to fall in three steps against ten for a recurrent baseline.
benchmark - arxiv:2611.08436 · cs.CVSM4RT: Cascaded Feed-forward Model for 4D ReconstructionPang-Chi Lo, Jhao-Syun Lai, Tsung-Wei Ke
We introduce SM4RT, a multi-task cascade feed-forward model for dynamic scene reconstruction from RGB videos. While existing feed-forward approaches typically address geometry reconstruction and motion estimation separately--and often predict motions in a single coordinate frame that must account for camera movement--SM4RT solves these tasks sequentially by first reconstructing 3D geometry and then bootstrapping motion estimation from the predicted structure. By expressing each pixel's 3D motion in its corresponding camera frame, SM4RT reduces ambiguity arising from camera motion. To address the sparsity of real-world motion labels, SM4RT further incorporates a weakly supervised loss that regularizes unlabeled point motions using labeled point positions at the other time step. We evaluate motion estimation on three dynamic-scene datasets--Kubric, Waymo, and Point Odyssey--and SM4RT achieves the best performance on most metrics compared with prior baselines. We additionally benchmark geometry reconstruction on five dynamic scene datasets and show that SM4RT achieves competitive performance with state-of-the-art methods. Overall, our unified framework demonstrates that combining geometry reconstruction and motion estimation in a cascade not only avoids performance trade-offs but also enhances motion estimation through improved geometric reasoning.
benchmark - arxiv:2611.08433 · cs.AIMedGraph: Maintaining Patient State with Graph Memory for Long-Term Medical DialogueDonghua Zhang, Zizheng Wang, Ze Dong, Yiding Feng +1
LLM-based medical assistants that follow a patient across visits need a memory that keeps the patient's clinical state correct as it changes. We examine the errors of five memory systems on longitudinal medical dialogue and find three recurring causes: the retrieved evidence gives an outdated state, includes another person's health facts, or lacks a safety fact the answer depends on. We present MedGraph, a graph memory built for these three cases. Construction extracts dated facts from each session, attributes each fact to the person it describes, and links medication, diagnosis, and measurement records into chains that keep earlier states. Retrieval ranks the patient's chains as units, reads each chain at the time the question asks about, and adds the original dialogue text; a patient profile lists every recorded safety fact with every question. On MedMemoryBench, MedGraph has the highest overall accuracy with both a GPT and a Qwen backbone, 10.5 and 10.4 points above the strongest of five baselines, with efficient memory construction and answering in the GPT setting.
memory - arxiv:2611.08428 · cs.AIThe Power of Meta Access in the Software of the Future --- Demonstrated on LLM Chat UIs and Chat AIJulian Kunkel, Ali Doosthosseini
Chat remains the default surface for many large language model (LLM) tasks, even as agentic tools such as OpenCode excel at coding and automation. Yet the model sees the current dialogue but not the workspace around it. We explore meta access---giving the assistant a structured, permissioned view of the interface itself and propose a ladder of levels: feature tutoring; metadata, settings, and preference access; retrieval over summaries and messages, confirmed mutation, and full orchestration. Meta access can turn a chat client from many screens into one entry point for finding, navigating, organising, and operating it, while raising risks around grounding, privacy, prompt injection, and trust. General-purpose agentic harnesses are runtimes; because they can change their own settings, meta access becomes reflexive and needs a separate control-plane boundary. We ground this in the Meta Chat prototype of the Chat AI client. It builds bounded context from, e.g., summaries, settings, and relevant messages; links references to conversations, topics, and personas; supports confirmed, reversible changes; offers a browsable tutorial library; and recommends personas as user-started conversations. By exposing application state in structured form, meta access can extend screen readers from widget descriptions to semantic workspace operations. Meta access thus turns chat from a window onto one conversation into a control layer for the whole workspace, extending chat toward agentic capability without surrendering user control. As assistants become embedded in IDEs, dashboards, office suites, and operating systems, AI-era interfaces should expose a structured, permissioned model of their state and capabilities, ideally by opening their existing metamodels through a debugger-like interface.
agentic - arxiv:2611.08425 · cs.ROFailure Prediction and Classification in Vision-Language-Action Models Via Calibrated Latent ProbingMichal P. Podolinsky, Yunhao Yang, Daniel Little, Hong Lu +5
Vision-Language-Action models (VLAs) enable general-purpose embodied intelligence, but their performance is brittle under distribution shift, where unanticipated failures can damage equipment or endanger people. Deploying VLAs in real-world systems therefore requires a runtime monitor that is timely, calibrated, and informative while being efficient enough to run on the edge. Existing methods fail to fulfill these requirements. We present VLA-Tug, a lightweight probe over the latent layers of a VLA backbone that improves early failure detection and bounds the missed-detection and false-alarm rates, with minimal inference overhead. We show that supervising on the pre-failure segment of a trajectory increases early failure detection at matched false-positive rates by up to 30% over the strongest learned baseline. Under shift to unseen tasks, the bounds on missed-detection and false-alarm rates remain valid in up to 86% of runs, versus at most 21% for baselines. Unlike existing methods, VLA-Tug predicts the failure mode to enable corrective actions, raising top-3 accuracy on unseen tasks by up to 44% over a majority baseline. VLA-Tug enables real-time monitoring where the well-timed and informative probe supports lightweight steering that improves task success. We release a dataset of 3,180 annotated manipulation episodes to support future work on latent failure detection in VLAs at https://vla-tug.github.io.
vision-language-actionvlaembodiedmanipulation - arxiv:2611.08421 · cs.LGData-driven discovery of intrinsic Lie point symmetries of differential equation systemsIlker Gurcan, Kadir Ustun, Ramazan Gokberk Cinbis
Discovering continuous symmetries directly from differential-equation data can reveal geometric structure for model reduction and scientific machine learning without symbolic determining equations or a closed-form representation of the governing system. We formulate intrinsic Lie point-symmetry discovery as an operator-theoretic problem on sampled jet-space data. Admissible generators are identified from a sampled determining operator and separated from redundant, solution-dependent, and geometrically trivial directions. The framework combines prolongation, sequential subspace restrictions, and a fixed nonlinear dictionary linear in trainable coefficients. A rank-stabilization preflight test diagnoses systems whose represented intrinsic symmetry algebra does not stabilize under dictionary refinement. On a controlled differential-equation benchmark with an analytically known 11-dimensional represented intrinsic symmetry algebra, the method recovers the complete represented generator set, reproduces its Lie-bracket structure to numerical precision, and approximates a known generator beyond the training domain. Here we show that explicit intrinsic continuous-symmetry generators can be recovered from data obtained from exact or approximate differential systems.
benchmark - arxiv:2611.08420 · cs.LGOn the limitations of molecular representationsAcer Blake, Fergus Imrie, Charlotte M. Deane
Molecular representations such as fingerprints and learned embeddings play a central role in cheminformatics and small-molecule machine learning. They serve as the primary input for a diverse array of tasks, including clustering, dataset splitting, and virtual screening, yet their evaluation is often limited to downstream model performance on narrow benchmark tasks. We carried out a systematic investigation of the quality of molecular representations, finding that none of the state-of-the-art molecular embedding models capture graph structure, 3D shape, or ADMET endpoint similarity better than fingerprints or physicochemical descriptors. However, classical fingerprints also revealed their own material limitations. For example, we show that depending on the representation, between 225M and $\sim$4B molecules have non-unique representations in the 10.8B Enamine REAL database, with implications for virtual screening. Our findings provide an insight into the current limitations of molecular representations, outlining future research directions for the development of more robust encodings of small molecules.
benchmark - arxiv:2611.08417 · cs.AIBenchmarks as Software: Terminal-Bench and the Case for Continuous BenchmarksE. Kelly Buchanan, Crystal Zhou, Scott W. Linderman, Christopher Ré
Coding agents are among the most economically consequential deployments of LLMs to date, making reliable evaluation increasingly important. Yet even benchmarks that are carefully validated at release can become stale. As models improve, they can discover valid solutions that existing verifiers reject, while changes to evaluation infrastructure can invalidate assumptions embedded in tasks and tests. We call the resulting accumulation of defects benchmark maintenance debt: a growing gap between benchmark scores and the capabilities they are intended to measure. We introduce a continuous-validation workflow for agentic benchmarks that combines community reports, task artifacts, agent traces, automated oracle checks, and human-gated repairs. We apply this workflow to Terminal-Bench 2.0, a benchmark that underwent extensive human and LM-assisted review and red-teaming before release. Our audit identifies issues in roughly 30% of the benchmark, and we release the resulting repairs as Terminal-Bench 2.1. These repairs change evaluation results: accuracy increased by over 10 percentage points and leaderboard positions move by up to three ranks, even as the overall ranking remains highly correlated (Spearman rho = 0.958). We evaluate how resource allocation shapes benchmark conclusions. We show that fixed wall-clock can materially change agent performance: For example, extending the time limit closes a 20pp harness gap on GPT 5.4 and shrinks a 30 pp gap on GPT 5.4 mini to 5 pp. Benchmark maintenance is therefore not only about repairing task artifacts: resource budgets are part of the measurement specification and must be validated alongside the tasks themselves. More broadly, benchmark maintenance debt is not just accumulated implementation bugs: our results show that as models improve, both evaluation assumptions and resource settings may need to be revalidated.
agentagenticbenchmarkleaderboard - arxiv:2611.08413 · cs.ROT3R: Training-Free Two-Stage Token Refinement Towards Efficient and Robust VLA ModelsMohammed Hassan, Zhenyang Chen, Zheng Wang, Zhixin Zhu +4
Vision-Language-Action (VLA) models enable end-to-end robot control, but inference is increasingly expensive. Each decision processes hundreds of visual tokens, often across multiple cameras, leading to substantial computation while allowing irrelevant background content to distract action prediction under distribution shift. We argue that these issues stem from two complementary forms of redundancy: many visual tokens are spatially irrelevant, while attention over the remaining tokens is often misallocated. Based on this observation, we propose T3R, a Training-Free Two-Stage Token Refinement framework. In the first stage, segmentation-guided token pruning removes irrelevant visual patches to reduce computation and eliminate distracting context. In the second stage, saliency-guided attention biasing reallocates attention toward action-critical visual regions and language instructions within the retained tokens. Together, pruning produces a cleaner candidate set, while attention biasing further emphasizes the most decision-relevant evidence, yielding a better accuracy-efficiency frontier than either stage alone. On LIBERO-Spatial, T3R removes 85% of scene-camera visual tokens while essentially matching the unpruned success rate, and improves success by up to 7 percentage points under out-of-distribution object-placement perturbations. The timing comparison attributes 7.8 ms per inference call to pruning and attention-bias application. As input resolution increases, LLM-forward profiling shows up to a 3.3x speedup over the unpruned baseline. Evaluations across LIBERO, SIMPLER, and RoboTwin with three VLA backbones show the strongest accuracy-efficiency benefits in multi-camera settings. These results demonstrate that training-free token refinement is a practical approach to efficient and robust VLA inference.
vision-language-actionvlavla modelliberorobotwin - arxiv:2611.08407 · cs.ROSocial Humanoid: Socially Responsive Human-Humanoid InteractionVongani H. Maluleke, James Ni, Kie Horiuchi, Lea Wilken +6
Humanoid robots are typically trained as solitary agents, leaving them unable to coordinate physically with human partners who naturally anticipate and adjust to one another. To bridge this gap, we present Social Humanoid, a hierarchical system that generates and executes socially responsive whole-body motion on a 29-DoF Unitree G1 conditioned on a human partner's motion. Its high-level policy performs flexible sequence modeling of human and humanoid motion through a shared tokenized representation, generating reactive humanoid motion and anticipating the partner's motion when they are occluded, while a low-level whole-body tracking controller executes the generated motion on the robot. To train the system, we introduce a two-phase adaptive retargeting procedure that retargets one interaction partner to the robot while morphing the other partner to preserve contact dynamics, reachability, and relative body scale. We apply this procedure across five motion-capture datasets to produce a new Human-Humanoid Interaction (HHI) benchmark of 260 hours of human-humanoid interactions. Our results demonstrate that Social Humanoid produces coordinated, human-aligned behavior across diverse interaction types in simulation and the real world. Project page: https://von31.github.io/social-humanoid/
humanoidbenchmark - arxiv:2611.08403 · cs.LGStepBack: Step-Level Error-Localized Resampling for Efficient Test-Time ReasoningYang Ouyang, Jung-Eun Kim
Self-consistency (SC) is a widely used decoding strategy for improving chain-of-thought reasoning, but it requires sampling many complete reasoning trajectories with a preset budget. Its efficient variants reduce the cost mainly by stopping full-trajectory sampling early or distributing the number of samples across questions. In this paper, we propose StepBack, a step-level error-localized resampling method for efficient test-time reasoning. Given a small set of draft trajectories, StepBack localizes likely first error positions using lightweight signals and then resamples the detected error step and its suffix, followed by majority answer voting. Empirically, StepBack matches the accuracy of SC on average across seven mathematical reasoning and multi-hop QA benchmarks, while reducing token usage by up to 41.26%. In addition, by nature, StepBack can be combined with adaptive inference methods such as Adaptive Consistency, further improving their inference-time efficiency.
benchmark - arxiv:2611.08398 · cs.ROA Multimodal Haptic Interface for Simultaneous Softness and Thermal ModulationJoshua Brown, Fernando Bello, Lynette Jones
Haptic interactions, such as touching an object in virtual reality, involve detecting a plethora of sensations through the skin. Whilst tactile displays exist to simulate and manipulate all of these cues individually, little progress has been made constructing devices that can control multiple cues simultaneously. This paper presents a tactile display technology that offers both softness and temperature change. A prototype display, based on particle jamming and the thermal transfer of heated and cooled water, is evaluated in terms of its ability to change both stiffness and temperature. Mechanical evaluations demonstrate controllable stiffness increase of 300%, whilst thermal characterisation showed strong surface uniformity and a range of achievable temperatures from 10°C up to 50°C, with closed-loop control maintaining a 0.2°C error from a setpoint. This work advances the design of multimodal tactile displays with applications to perceptual research, medical simulation, and affective haptics.
tactile - arxiv:2611.08394 · cs.LGSynthesis of Ovarian Cancer Multi-Omics Datasets: A Biology-Informed Evaluation FrameworkDéborah Boyenval, Alina Malyutina, Jaana Oikkonen, Sammeli Liikkanen +1
Paired multi-omic cancer cohorts are costly to assemble and difficult to share, motivating synthetic cohorts as a complementary resource. As generative model development accelerates, we shift the focus from proposing yet another architecture to defining evaluation criteria that determine whether synthetic data preserve biologically relevant structure. We introduce a biology-informed and patient-disjoint framework for evaluating the generation of paired RNA expression and copy-number variation (CNV) profiles, and apply it to the longitudinal DECIDER high-grade serous ovarian cancer cohort. The framework assesses modality-specific fidelity and coherence between RNA and CNV pairs using agreement between real training and held-out data as a reference, alongside complementary training-data proximity diagnostics. We adapt three multi-omic integration architectures based on variational autoencoders to generate paired profiles. The model leading several fidelity and coherence measures also shows the strongest training-data proximity. The protocol and case study offer the community a reference for comparing future multi-omic generators.
evaluation framework - arxiv:2611.08393 · cs.CVAdaptive Low-Light Enhancement for Road Anomaly Detection: A Multi-Architecture Benchmark and Edge Deployment StudyObed Honour Eje, Oluwatosin Kolade, Glory Bagai, Taliha Abiodun Folorunso
Road anomaly detection systems on autonomous vehicles must function reliably under degraded visual conditions including darkness, rain, and Harmattan dust haze. This paper benchmarks four object detection architectures, YOLOv8n, YOLOv12n, RT-DETR, and Faster R-CNN, under physically grounded lighting and weather degradation, and shows that uniform application of the SCI low-light enhancement network is actively harmful at dusk-level lighting (-16.1% to -30.7% mAP@50). We propose a lightweight, training-free luminance-adaptive router that selects between no enhancement, moderate SCI, and denoising plus strong SCI based on per-frame Rec. 601 luminance, recovering most of the dusk penalty while preserving low-light gains, and we validate the full pipeline on NVIDIA Jetson Nano edge hardware.
benchmark - arxiv:2611.08392 · cs.AILearning a patient world model for health system simulation in intensive careYingtai Li, Ruoyan Wang, Yu Li, Shaohua Kevin Zhou
Designing healthcare strategies requires anticipating their clinical and operational consequences before implementation, a task that often relies on health system simulation. Such a simulation necessitates modeling how patient states evolve and generate care demands. However, conventional approaches typically tailor compact state transitions to a single policy question and omit history-dependent dynamics, thus requiring the redefinition of states and transitions when there is a new question that entails different clinical information. We propose to learn a patient world model from longitudinal EHRs using a neural network to provide a shared clinical representation of patient trajectories that can serve simulation questions across scenarios. Taking ICU operations as an example, we train the model on MIMIC-IV, freeze it, and couple its time-queryable predictions to shared ICU calendars, resources, and queues via scenario-specific mappings, so that one model renders admission demand, care tasks, and discharge proposals across simulations evaluating ICU bed capacity, care dispatch, and post-ICU transfer capacity. Experiments show that our model matches Markov and semi-Markov comparators in aggregate occupancy and directional responses to capacity and dispatch policies. It also reproduces policy gains from replayed observed trajectories. For patients entering the same support state from different prior trajectories, the learned representation captures a 24-h recovery difference (-6.4 predicted versus -6.7 observed percentage points), a contrast missed by compact state models, which propagates into distinct simulated care demand, queue burden, and bed use. These results indicate that a learned world model of patient dynamics has the potential to replace question-specific state models, allowing new care policies to be evaluated by reconfiguring the scenario without model rebuilding.
world model - arxiv:2611.08390 · cs.ROSCALE: Real-Time Latent Communication for Cyber-Physical AgentsShanghao Shi, Jian Xiao, Ching-Hsiang Chan, Sanjoy Baruah +4
Latent communication enables agents to collaborate efficiently by directly exchanging intermediate model representations rather than raw sensor data or text messages. Existing approaches, however, are primarily designed for LLM agents and do not consider the bandwidth constraints and heterogeneous task deadlines faced by cyber-physical agents. In this work, we propose SCALE, a real-time latent communication method tailored for cyber-physical agents. Motivated by the spatio-temporal continuity of the physical world, SCALE matches tokens across consecutive latent messages using similarity scores to identify reusable information and avoid redundant transmission. It then profiles the semantic gain and communication cost of transmitting different amounts of latent information. Based on these profiles, SCALE employs a real-time scheduler that jointly determines which pending messages to transmit and how many latent tokens to send, subject to available bandwidth and task deadlines. We evaluate SCALE across three cyber-physical benchmarks and a physical spider robot case study under diverse network conditions and heterogeneous task deadlines. Our results show that SCALE significantly improves both task performance and on-time delivery in bandwidth-constrained networks. Our code is available at: https://anonymous.4open.science/r/SCALE-4E80/.
llm agentbenchmark - arxiv:2611.08386 · cs.LGHybrid Machine Learning for Imbalanced Landing-Phase Safety Event Prediction from Flight DataTaha Asghar, Bilal Mufti, Syed Qasim Zaheer, Luqman Ahmed
Landing-phase safety-event prediction requires the detection of rare abnormal approaches within strongly imbalanced multivariate flight data. Capturing coupled temporal behavior while distinguishing event-specific signatures across airport operating contexts remains a central modeling challenge. This study proposes an interpretable hybrid feature-fusion framework for multiclass landing-phase safety-event prediction. A long short-term memory (LSTM) autoencoder trained on nominal flight windows characterizes deviations from routine approach behavior. Its continuous reconstruction-error descriptor is fused with parameter-specific statistical flight-history features and supplied to a class-weighted XGBoost classifier to distinguish Normal, Speed High, Path High, and Flaps Late windows. The framework is evaluated using 53,750 NASA DASHlink approach windows from three arrival airports. Pooled and airport-specific learning strategies are compared alongside component-level benchmarks, repeated pipeline retraining, and feature-attribution analysis. The results demonstrate strong class-balanced detection, with mean macro recall above 0.90 across all airports under repeated partitioning and retraining. Component comparisons establish the importance of class weighting and show that continuous feature fusion outperforms reconstruction-error-only and hard-gate alternatives. Airport-level comparisons reveal benefits from local specialization at the two airports with larger datasets and from pooled training at the airport with the smallest dataset. Feature attributions associate predictions with operationally relevant vertical-path, configuration, speed, and approach-context variables. The framework provides an interpretable approach to rare-event detection and practical guidance for adapting flight-data monitoring models to airport-level data availability.
memorybenchmark - arxiv:2611.08379 · cs.LGModule-Aware Optimization for Graph Neural NetworksGuy Hadad, Haggai Roitman, Moshe Eliasof, Bracha Shapira
Different optimizers impose different geometries on parameter updates. Adaptive methods such as Adam and AdamW use coordinate-wise updates, while matrix-aware optimizers such as Muon utilize weight-matrix structure through orthogonalized updates. For Graph Neural Networks (GNNs), however, the effective geometry of its weight matrices is shaped not only by its learned channel-mixing weights, but also by the graph-propagated representations it acts on. Thus, even if an update is well normalized as a matrix, its effect on the layer output can be amplified when applied to graph-propagated features. To address this gap, we introduce GraphMuon, a graph-aware optimizer that adapts Muon's matrix updates to the covariance structure of graph-propagated layer inputs. GraphMuon performs orthogonalization in a metric defined by these propagated representations and normalizes step sizes according to the update's induced effect on the layer representation. We show that theoretically, vanilla Muon can suffer from graph-induced update amplification, while our GraphMuon can mitigate this effect. As such, our GraphMuon yields a graph-metric analogue of Muon's polar update step. Empirically, GraphMuon achieves $1.32\times$-$2.42\times$ validation-loss time-to-target speedups over AdamW on three large-scale OGB benchmarks. Across 15 dataset-architecture configurations spanning node classification, inductive graph classification, and graph regression, it achieves lower time-to-target in 13 cases, while maintaining competitive or superior predictive performance. Code is available at https://github.com/guy-hadad/GraphMuon.
benchmark - arxiv:2611.08373 · cs.ROSLA: Scaling Spatial-Language-Action Model via Continual Pre-trainingFuhao Li, Wenxuan Song, Qizhou Huang, Ziyang Zhou +6
Reliable robotic manipulation requires both semantic understanding and precise control in three-dimensional space. However, vision-language-action (VLA) foundations largely inherit visual representations from 2D vision-language pre-training, while spatial supervision is often introduced only during task-specific adaptation. Developing spatial understanding as a transferable foundation-model capability therefore calls for learning from broader robot experience. We present SLA, a Spatial-Language-Action Model that scales implicit spatial alignment to continual pre-training and maintains it during downstream post-training. Initialized from a pretrained pi05 policy, SLA jointly optimizes action prediction and alignment between intermediate visual embeddings and spatial representations from a frozen 3D perception model. We curate and unify heterogeneous robot datasets to support spatial learning across diverse embodiments, tasks, and scenes. To accommodate this diversity, we introduce feature-dependent gates that adaptively aggregate visual embeddings from multiple VLA layers for spatial supervision. The teacher further exploits partially available depth and camera calibration alongside RGB observations, enabling effective use of heterogeneous geometric annotations. Probabilistic state dropout encourages reliance on visual cues and mitigates overfitting to demonstration trajectories. All auxiliary components are used only during training, allowing direct reuse of pretrained VLA weights and preserving the original inference architecture.
vision-language-actionvlamanipulationpi0post-training - arxiv:2611.08372 · cs.ROMOMASPLAT: Dynamic Novel-View Demonstration Augmentation for Mobile ManipulationSateesh Kumar, Arnav Balaji, Arpit Bahety, Skand Peri +2
Visuomotor policies trained through imitation learning are highly sensitive to changes in camera viewpoint. This is particularly problematic for mobile manipulation (MoMa), where robots rarely execute manipulation from the exact same base pose used during demonstration collection due to navigation uncertainty, localization drift, and environmental constraints. Recent approaches use novel-view synthesis (NVS) to augment demonstrations with unseen viewpoints, but they assume static scenes and therefore fail to re-render the manipulated object during the post-contact interaction phase, leaving the most critical parts of the demonstration unaugmented. We present MOMASPLAT, a novel-view demonstration augmentation framework for mobile manipulation that synthesizes full manipulation trajectories from unseen robot base poses, including dynamic object interactions. Our key insight is that standard MoMa data collection naturally provides the multiview observations needed to reconstruct both the static scene and the manipulated object in 3D. MOMASPLAT combines 3D Gaussian Splatting, feed-forward 3D foundation models, and kinematics-based action retargeting to separately reconstruct the scene and manipulated object, render them from novel viewpoints, and generate geometrically consistent synthetic demonstrations paired with transformed actions. We evaluate MOMASPLAT on five real-world mobile manipulation tasks involving both rigid and articulated objects across three scenes, and validate the fidelity of its synthesized demonstrations against ground truth in simulation. MOMASPLAT substantially improves robustness to unseen base poses, more than doubling the average success rate of the strongest baseline and maintaining non-trivial success at large base displacements where all baselines fail, while reducing the need for largescale real-world data collection.
manipulation - arxiv:2611.08369 · cs.AINPCBench: A Clinical Apprenticeship Benchmark for Multimodal Care-Pathway Reasoning in Nasopharyngeal CarcinomaPengkai Wang, Wei-Wei Zhang, Yan Li, Min Tang +12
Current medical benchmarks typically break down clinical competence into isolated factual questions, single-image tasks, or static case vignettes. These tests score individual answers, but reveal less about whether a model can integrate subspecialty knowledge with multimodal evidence and track a patient's condition throughout a course of care. To close this gap, we introduce NPCBench, a clinical apprenticeship benchmark for nasopharyngeal carcinoma (NPC). NPCBench mirrors specialist training, from guideline and foundational knowledge to case planning, longitudinal review, CarePath integration, and open-ended consultation. It contains 4,777 component and consultation items and a 100-patient CarePath cohort with 417 scored diagnostic and care tasks. The episode-level evaluation tests whether diagnostic and care answers are jointly correct within a care pathway. By linking stepwise tasks to patient-level care trajectories, NPCBench evaluates whether models ground decisions in visual evidence, interpret changes in patient state, and meet clinical requirements in consultation. Across 30 evaluated models, we observe a composition gap: strong performance on individual tasks does not ensure jointly correct diagnostic and care answers for the same patient. Among 21 multimodal models, strict CarePath success reaches at most 23%. Paired controls and physician calibration further examine evidence grounding, longitudinal revision, and consultation judging. Overall, NPCBench broadens the evaluation of medical artificial intelligence beyond static question-answering paradigms to an apprenticeship-oriented assessment of clinical workflows, thereby offering a testbed for evidence-based and temporally coherent reasoning in oncology care.
benchmark - arxiv:2611.08368 · cs.ROVisionSIM: Beyond Idealized Sensors - A Multi-Modal Simulator for Robot PerceptionSacha Jungerman, Kaustubh Sadekar, Avery Gump, Sriram Narayanan +5
Robotic perception increasingly relies on heterogeneous sensor suites combining cameras with emerging modalities (event, SPAD, iToF/dToF, and thermal). Yet existing simulators idealize these sensors, hiding critical modality-specific failure modes that limit real-world deployment. We present VisionSIM, a unified, physically grounded simulation platform that elevates sensor realism to a first-class concern by modeling the raw physical measurement process for diverse conventional and emerging modalities. We expose fine-grained control over physical sensor parameters, reproducing failure modes overlooked by general-purpose simulators: pileup and glare for LiDAR; phase wrapping, saturation, and motion artifacts for iToF; refractory drops for event cameras; and history-dependent emission for thermal cameras. This enables principled cross-modality comparison on identical scenes and trajectories. We release the platform and a large indoor dataset with synchronized raw measurements and rich ground-truth annotations.
event camera - arxiv:2611.08363 · cs.LGStrataFlux: Memory-Hierarchy-Native Low-Bit LLM InferenceJake Hyun, Ziyu Zhao, Jae-sun Seo, Mohamed S. Abdelfattah
Quantization reduces the memory and bandwidth demands of large language model (LLM) inference but can degrade accuracy; offloading higher-precision data can instead stall generation. We introduce StrataFlux, a framework that challenges this tradeoff by recovering accuracy lost to low-bit weight and key-value (KV) cache quantization while retaining quantization's memory and speed advantages. StrataFlux co-designs refinement representation, selection, and delivery to make host RAM and solid-state drives useful sources of additional precision. Its selective precision streaming reduces transfer volume, overlaps delivery with computation, and reuses fetched precision across decoding steps so generation need not wait for every transfer. We evaluate 2-4-bit weights and 2-3-bit KV-cache on dedicated GPUs and unified-memory platforms. For 2-bit Llama-3.1-8B-Instruct weights, StrataFlux reduces WikiText-2 perplexity from 9.46 to 8.40 while maintaining 95.4% of the quantized base's throughput on RTX 5090. With 2-bit weights and KV, StrataFlux raises Qwen3-8B's AIME 2024/2025 accuracy from 11.0% to 47.1% at unchanged GPU memory use. At 128K context, Llama-3.1-8B-Instruct with 3-bit weights and 2-bit KV achieves comparable perplexity and speed to vLLM's 4-bit AWQ weights and 8-bit KV on RTX 5090, using approximately half the GPU memory. These results show that GPU-resident precision need not set the accuracy ceiling, nor offloading-tier latency dictate decoding speed.
memory - arxiv:2611.08362 · cs.CLSpecCTMC: Exact Speculative Simulation of Diffusion Language Model Continuous-Time Markov ChainsSiva Rajesh Kasa, Yasong Dai, Sumit Negi, Hongdong Li
Masked diffusion language models offer parallel text generation, but independently sampling multiple tokens can disrupt their dependencies and change the output distribution. Exact continuous-time sampling avoids this approximation, yet typically requires one sequential model evaluation per token. We introduce SpecCTMC, a training-free speculative sampler that reduces this sequential bottleneck without a separate draft model. SpecCTMC proposes successive token updates using previously computed model predictions, evaluates their intermediate states in parallel, and applies an exact acceptance-and-correction procedure. We prove that SpecCTMC preserves the joint distribution of event times, generation order, and token values of the underlying continuous-time Markov chain. It therefore introduces neither time-discretization error nor additional factorization error from parallel decoding. Experiments on LLaDA-8B-Instruct and Dream-v0-Instruct-7B across five benchmarks demonstrate 1.73-4.38x fewer sequential model evaluations than direct continuous-time sampling, while preserving the reference sampling distribution. Our code is available at: https://github.com/kasakh/specctmc
benchmark - arxiv:2611.08361 · cs.RORadio-to-Pose Information for Gauge-Drift Correction in Cooperative LocalizationAditya Iyer, Andy Ye, Akshay Narayan, Nora Ayanian
Cooperative localization uses odometry and inter-robot measurements to estimate the motion of a robot team. These measurements describe only relative geometry, but the team's estimated position and orientation can drift together over time, i.e., gauge drift. In this work, we study correction of this gauge drift using a radio attenuation map registered in the same world coordinate frame used for robot poses. Specifically, we propose Radio-to-Pose Information (RPI), an interface that converts power residuals and their pose derivatives into a Gaussian information matrix and vector pair. RPI builds on the insight that received power is sensitive to shared displacement in the world map. The pose estimator uses the interface together with odometry and inter-robot measurements to update poses and minimize uncertainty. We characterize the shared motions detectable from the map and observed links. For an uncertain shared map, RPI keeps correlations across robots and time and reproduces joint Gaussian pose estimates and covariances. Across $1000$ simulated Gaussian trials, RPI matches the results of joint estimation. Across 32 paired trials with a registered map, RPI reduces common-position RMSE from $0.542$m to $0.140$m in a $50\times50$m arena compared with a baseline that uses radio only to weight relative measurements.
arena - arxiv:2611.08352 · cs.ROUSB Gym: A Gymnasium Testbed for Closed-Loop Ultrasound View MatchingThanapong Intharah, Zhichao Gao, Yiliang Liu, Hao Dong
Background and Objective: Our main goal is ultrasound guidance: moving a probe until its image matches a target view. Published methods for each part of a guide are evaluated in separate simulators. We asked whether these methods, combined into one guide, are good enough. Methods: We built USB Gym, a testbed whose standard engine simulates ultrasound from public patient CT, with a network drawing organ textures from CT labels. Its guide has three replaceable parts (slots), an organ segmenter, a controller and a reward, which received 20 published methods, each run on the same 60 episodes of six held-out patients. In each slot, a published method replaced ours only if detectably better on these patients. The better of this combination and our original guide on 20 new CTs is the benchmark baseline. A pose oracle, reading the exact probe pose available only in simulation, set the reference. The results below are exploratory. Results: On these CTs, the probe started on average 16.0 mm from the target view; the pose oracle brought it to 1.5 mm. Of the results known when the combination was fixed, only an image-similarity reward (SSIM) was slightly but detectably better, so the combination was our guide with that reward. This combination was the benchmark baseline, ending at 13.7 mm (16% of the oracle's progress). Conclusions: Combining untuned published methods with ours did not guide the probe near the target view; the task remains under-explored. The testbed accepts new methods; no guide was tested on hardware.
benchmark - arxiv:2611.08345 · eess.SYOutput-Feedback Integral Sliding-Mode Extremum Seeking with Super-Twisting DifferentiationHyuntae Kim
This paper develops a dither-free output-feedback extremum-seeking controller for pre-stabilized nonlinear dynamic systems with unknown steady-state performance maps. The controller uses only the measured performance, the applied parameter, and internal states. Within a local state--parameter domain, integral sliding mode enforces equality between the applied parameter rate and a prescribed nonzero search velocity after finite-time capture under bounded matched rate-channel perturbations. This motion lets a super-twisting differentiator and rate normalization recover a gradient-sign surrogate without demodulation or averaging. A ramp-corrected moving-equilibrium coordinate cancels the first-order ramp-induced plant term and yields explicit flow, reversal, recovery, and dynamic-remainder bounds. The resulting local state--parameter tube supplies the regularity needed for finite-time exact differentiation. Combined with a deadbanded hybrid relay, these results establish local practical convergence to an explicit neighborhood of the unknown optimum. An 80-state tubular-reactor benchmark and paired held-out Monte Carlo tests compare a causal sampled realization with perturbation ESC and a rate-informed super-twisting comparator. The proposed realization succeeds in all 200 held-out cases and attains the smallest upper-tail objective gap under the stated tuning protocol. A matched-rate sweep connects performance degradation to applied-rate mismatch, near-stall motion, reversal, and regularization attenuation.
benchmark - arxiv:2611.08342 · cs.LG4D Point Splatting for mmWave Radar Micro-Doppler Differentiable RenderingAdnan Armouti, Guanyue Qian, Rajalakshmi Nandakumar
Millimeter-wave radar resolves the micrometer displacements of a moving body as micro-Doppler, enabling gait, gesture, activity, pose and vital-sign sensing. Each task relies on a downstream model trained on real data, yet radar datasets lag orders of magnitude behind vision and language. Brute-force collection will not close that gap, and existing synthesis methods (forward-only simulators, learned generators and static-scene inverse renderers) cannot fit a real capture of a moving body, leaving a sim2real gap. We present 4D Point Splatting (4DPS), a physics-based differentiable renderer that fits an explicit moving scene to a real capture. Using the synchronized camera to initialize the room and body, we represent the scene as oriented points carrying material, normal and velocity, and splat each point into range and Doppler bins. Our output is the raw complex return a real radar records, so we apply the same signal processing on the rendered data to obtain any representation for training any SOTA downstream model. We fit 235 sequences on HuPR and 100 on RT-Pose, two datasets with different radars. Our reconstructions set the state of the art, with median correlations of 0.940 and 0.830 on range-azimuth, 0.829 and 0.695 on range-Doppler, and 0.988 and 0.966 on micro-Doppler. Our renders also set the state of the art in downstream sim2real gap. On RT-Pose, a pose model trained on them reaches 21.9 cm MPJPE on real data, against 16.7 cm for real training and 127.6 cm for the RF-Genesis baseline, where both the rendered- and real-trained models learn nearly the same features. Our custom CUDA kernels fit a sequence in 5.7 minutes on HuPR and 12.9 on RT-Pose on one H100, making dataset-scale inverse rendering of mmWave radar possible for the first time. This work can enable synthetic data generation for any human-centered downstream task and model, and scene editing for data augmentation.
sim2real - arxiv:2611.08341 · cs.AIEnhancing Local and Global Structure in Compressed AI Agent MemoryHenry Han
AI agents store text as embedding vectors and compress them at scale. A compression is accepted when each item's nearest neighbors survive, but agents also consolidate, delete, and compare memories, tasks that depend on which items lie far apart. We evaluate both ends with two label-free fidelities from our prior work: the local fidelity $L$ at 10 neighbors and the global fidelity $G$ at 30 to 50 percent of the memory. At 96 bytes per passage, rescored binary codes reach an $L$ of 0.97 yet move 10.1 percent of passages to another theme; PCA with 8-bit codes, at an $L$ of 0.77, moves 1.7 percent. Rescoring leaves the stored codes, and thus $G$, unchanged. We prove that across the $G$ band, $1-ν_k$ is at most 2 to 2.18 times the square root of the near-far rank-violation rate, and that this rate follows the distance noise of the compression (correlation 0.979). Our locality-aware quantization (LAQ) turns the bound into a hinge objective on the codebooks of an OPQ index, keeping its bytes, rotation, and search. To our knowledge, LAQ is the first quantizer whose objective comes from a proven bound on the far end. Reconstruction alone raises $G$ in 0 of 52 held-out runs; the far hinge raises it in all 52. LAQ raises $G$ in all 52 ($p=8.5\times 10^{-4}$ over 13 memory-encoder groups), including the external FinDER memory, has the highest $G$ of seven equal-size quantizers in all 12 Multi-Doc-2025 runs, scales to a million passages, and keeps evidence recall.
memoryagent memoryagentai agent - arxiv:2611.08338 · cs.ROGPWM: Modeling the World as 3D Particle MotionRobert Ren, Matthew Bronars, He Zhu, Yikai Tang +4
We introduce GPWM, a Generative Particle World Model that predicts how 3D scenes evolve by generating the motion of particles and rigid bodies over time. Unlike video world models that generate scene appearance, GPWM operates directly in 3D physical space. A single representation and architecture cover rigid, deformable, cloth, and granular materials, in both passive physics and robotic interaction. GPWM is an object-aware multi-resolution transformer that combines flow matching with diffusion forcing to generate coherent physical trajectories far beyond its training horizon. We train and evaluate GPWM on ParticleWorlds, a 25.5-million-frame corpus spanning diverse geometries, materials, interactions, and robot embodiments. GPWM outperforms the best prior learned-dynamics baseline in each of six physical domains, reducing prediction error by 31% on average. Breaking from the domain-specific setting of prior methods, a single Unified GPWM models these heterogeneous regimes with one set of weights, largely outperforming per-domain baselines and even surpassing our own specialist models in two domains. GPWM further generalizes to unseen geometries, physical properties, object counts, and rollout horizons. Beyond prediction, GPWM enables contact-rich planning by learning a generative prior over joint actuator-object motion. Instead of searching over unconstrained actions, planning searches the model's latent noise for interactions that achieve a desired object-space goal. With the same world model, noise-space planning substantially outperforms action-space search on pushing and picking, where complex contact sequences are hard to discover. Executed open-loop in simulation, the resulting plans achieve goal-directed manipulation without reference action trajectories or goal-specific policies.
manipulationworld model - arxiv:2611.08335 · cs.LGDiscovering Learning Rules that Shape OthersHong Jun Lee, Ziyan Wang, Jun Wang, Stefanos Leonardos +1
In general-sum games, self-interested learners can reach outcomes that are poor for everyone, which has motivated methods that shape the opponent's learning. Methods like Learning with Opponent-Learning Awareness (LOLA), which differentiates through the opponent's learning process, and Model-Free Opponent Shaping (M-FOS), which uses model-free optimisation methods for long-horizon opponent shaping, have enabled mutual cooperation and exploitation of learners. However, these methods require access to the opponent's parameters, policy, rewards or learning rule during deployment. In single-agent reinforcement learning, update rules discovered by meta-learning have achieved state-of-the-art results (e.g. DiscoRL). These rules are separate from the policy they train and read only the agent's own experience. We hereby extend this framework to general-sum settings by introducing Discovered Update Rule for Opponent Shaping (DUOS). To the best of our knowledge, this is the first method to discover an update rule for an agent in general-sum games. DUOS meta-trains the rule to maximise the agent's expected return over its lifetime, and updates an inner network using the agent's own experience. We show that our discovered rule exhibits behaviour consistent with opponent shaping without reading opponent policies or parameters, and ranks 1st in normalised mean score and 1st in self-play among tournament entrants.
agentself-play - arxiv:2611.08334 · cs.CLCultural Adaptation within Repeated Social InteractionsChen Cecilia Liu, Roi Reichart, Anna Korhonen
Much of human culture is tacit and is typically learned through situated social interaction. As LLMs are deployed as interactive agents across diverse communities, assessing their cultural competence demands moving beyond existing question-answering approaches, shifting from what they know to how they adapt. In this work, we broaden the scope of cultural adaptation in NLP by investigating whether LLMs can adapt through repeated interactions via a sandbox environment. By reframing cultural adaptation as a continuous process of norm assimilation through social interaction, our work demonstrates that memory mechanisms foster longitudinal learning across repeated interactions, facilitate transfer to different social settings, and enable multilingual norm assimilation across five languages. Together, our findings lay the groundwork for continuous, interaction-driven cultural adaptation in LLMs.
memory - arxiv:2611.08318 · cs.ROLearning the Task or Learning the Verifier? Rethinking Criteria Exposure in Embodied Self-ImprovementWenbo Huang, Jianxin Bi, Shunyu Yang, Ruiyang Si +9
Self-improving embodied agents are increasingly being developed and compared on robot simulation benchmarks. Several recent works expose the benchmark's success criteria to the learning agent, a practice we call Criteria Exposure (CE). These criteria, originally written to verify policies trained by imitation learning, often demand more than the corresponding language instructions, so a score from such a verifier may not faithfully reflect whether the instructed task was completed, especially for robot-control programs written by an agent. As embodied coding agents report ever higher benchmark scores, a natural question arises: when a coding agent is exposed to the benchmark's criteria, does it generate code that better performs the instructed task, or code that simply overfits to the verification criteria? We conduct controlled experiments with a self-improving agent with and without CE, and find that with CE the agent quickly produces code with a higher success rate under the given criteria, but this advantage disappears when the criteria are changed slightly without departing from the original instructions. More importantly, the skills the agent develops under CE transfer poorly across tasks. We further propose instruction-faithful criteria and show that they better reflect the performance of agent-generated programs.
embodiedagentembodied agentself-improvingself-improvementbenchmark - arxiv:2611.08315 · cs.LGBenchmarking Large Language Models on Heterogeneous Cognitive and Biomedical Knowledge Graphs: The Role of Graph Grounding in Multi-Hop Reasoning and Hallucination MitigationAli Sarabadani, Kheirolah Rahsepar Fard, Tahereh Jeirani
Large language models (LLMs) are increasingly used in biomedical and cognitive neuroscience, but their ability to reason reliably over heterogeneous knowledge graphs (KGs) remains unclear. We benchmark five LLMs (GPT-5.1, Gemini 2.0 Pro, Llama 3, Qwen, and BioGPT) across five KGs: DrugBank, NeuroSynth, Cognitive Atlas, Brain Map, and GDPKG. We introduce a multitier query dataset to evaluate reasoning depth and factual accuracy under KG-aware (graph-grounded) and KG-agnostic conditions. Inter-annotator agreement was substantial, with Cohen's kappa of 0.781 (range: 0.753-0.807) and Fleiss' kappa of 0.77. Results show that general-purpose models, particularly GPT-5.1 and Gemini 2.0 Pro, generally outperform domain-specific models on multi-hop and high-level reasoning tasks, although performance declines as query complexity increases. KG grounding reduces hallucination rates across all evaluated models. Contrary to our initial hypothesis, domain-specific models do not consistently outperform general-purpose models on specialized tasks without graph grounding. These findings highlight the value of combining broad language-model capabilities with structured domain knowledge. We provide practical guidance for selecting LLM-KG integration strategies based on task complexity and computational constraints, and discuss hybrid neurosymbolic architectures for personalized medicine, cognitive disorder diagnosis, and explainable AI.
knowledge graphbenchmark - arxiv:2611.08312 · cs.MATrilemma: How Agents choose between Honesty, Confidentiality, and Task CompletionJohannes Taraz, Abhinav Kumar, Sahar Abdelnabi
Personal AI agents often act on underspecified instructions and draw on context such as their user's messages and calendar to decide how to proceed. This context can contain private preferences that interfere with the task, with other people's interests, or with confidentiality norms. Prior work has studied deception under goal conflicts and the leakage of private information, but largely as separate problems. We study both problems together in a simulated company Slack workspace in which four employees delegate sprint staffing to personal agents. Two agents learn from private messages that their employee avoids working with a particular colleague. Because the other agents request justifications for task preferences and the only true justification is confidential, the agents cannot simultaneously remain honest, preserve confidentiality, and complete the staffing. We evaluate seven open-weight models and Claude and GPT models. Of the 135/140 open-weight agents that pursue the avoidance preference, 107 either make false claims about task fit or logistics or cite a true consideration as if it were their actual reason. In a three-model comparison, confidentiality instructions reduce disclosure of the preference from 83% (none) to 31% (medium) to 2% (strong). However, the task is often not completed anyway; in the main experiment (medium confidentiality instructions), only 16 of 70 runs end with a valid staffing. In their reasoning traces, agents frequently acknowledge evidence against their claims and still conclude that the claims are not false; they engage in motivated reasoning. Our results indicate that agents can strategically make false or misleading claims in a day-to-day task without being instructed to; the environment provides a controlled testbed for studying how agents resolve such conflicts and for developing deception monitors.
ai agent - arxiv:2611.08308 · cs.LGRM-VLM: From Natural-Language Task Descriptions to Vision-Grounded Reward MachinesRoko Parac, Christos Ziakas, Leo Ardon, Alessandra Russo
Reward machines (RMs) have been proposed as a reward formulation in reinforcement learning (RL) to capture non-Markovian tasks where credit assignment depends on history. By representing reward functions as automata, RMs expose the task's temporal structure and allow such rewards to be handled within a Markovian framework. However, they assume a labelling function, a proposition-grounding oracle for observations, which must be manually engineered for each environment, limiting their wider applicability. In this work, we address this limitation by proposing RM-VLM, a novel approach in which a large language model (LLM) generates an RM from a natural-language task description, and a vision-language model (VLM) acts as the labelling function, grounding the RM's propositions over image-based observations during RM-guided policy learning. A VLM-based value function estimator is used to assess the agent's progress towards each RM transition and derive a dense reward signal. We evaluate RM-VLM on three domains covering non-Markovian grid-world navigation, robotic manipulation, and 3D object collection. Our experimental results show that RM-VLM consistently achieves higher task success rates than existing VLM-based reward methods across all tested domains.
manipulation - arxiv:2611.08307 · cs.LGMeasured Coupling for Parallel Decoding in Diffusion Language ModelsTal Oved, Lior Wolf
Diffusion language models can write many tokens per step, but must choose which tokens are safe to commit together. Mean-field decoding penalizes pairs of positions whose predictions look alike. We replace this proxy with the coupling itself: how much committing one token changes the model's confidence at another. Cheap probes inside the main forward pass measure it, with no training or extra model, and it plugs into the same mean-field rule. On three models and four benchmarks, our method needs far fewer steps than mean-field decoding and is faster in most settings, also with a key-value cache, at an average accuracy cost of 1.1 points. Shuffling the couplings or replacing them with a uniform bonus loses this gain.
benchmark - arxiv:2611.08302 · cs.LGDynamic Optimistic Constrained OCO with Memory via Delay EquivalenceMohammed Abdullah, George Iosifidis, Salah Eddine Elayoubi, Tijani Chahed
Online Convex Optimization with memory and time-varying constraints (COCO-M) is a recently-introduced framework that captures the dynamics of stateful online learning and non-stochastic control with budget constraints. Despite its expressive power, COCO-M remains largely understudied. In this work, we tackle this problem in its full scale, providing an optimistic algorithm that incorporates untrusted predictions to tame universal dynamic regret and cumulative constraint violation. Our analysis leverages a potent reduction from memory to delay and treats state-dependent gradients as delayed feedback. We establish the first bounds that adapt to both environment non-stationarity (path length $P_T$) and prediction accuracy, while lifting previous restrictive assumptions such as functional separability. En route to these results, we derive findings of independent interest for the memory-to-delay reduction, and for delayed OCO with time-varying constraints. The proposed framework ensures sublinear worst-case guarantees that improve with prediction accuracy, reaching $\mathcal O(1)$ regret under perfect predictions, with cumulative constraint violation (CCV) $\mathcal O(\sqrt{T})$ for unknown $P_T$ and $\mathcal O(\log T)$ for known or online observable $P_T$ , while recovering the optimal COCO rates when memory disappears.
memoryonline learning - arxiv:2611.08294 · cs.CVManimCode: Learning Visual Self-Correction for Educational AnimationZhengyang Liang, Lizi Liao
Manim supports precise, programmatic video generation, but a successfully rendered educational animation can still have visual defects, missing explanations, or incorrect content. Rendering does not directly identify the code changes needed to correct these visual failures, so feedback-driven revision is needed. Existing agentic frameworks use iterative rendering and revision but typically rely on large models, often proprietary. Existing Manim datasets primarily contain final programs and often lack quality verification and the trajectories of feedback and revisions needed to learn this refinement process. We introduce ManimData and ManimCode, a data and training framework for feedback-driven educational animation generation. We construct ManimData through a render-guided pipeline that iteratively generates, executes, reviews, and revises Manim programs with retrieved API documentation. Its 24,034 render-verified, quality-filtered records preserve intermediate programs, execution errors, visual critiques, and corrective revisions. We use role-conditioned supervised fine-tuning on these trajectories to train ManimCode, a single 4B model, to write, review, and revise animation programs. To improve correction of errors that differ from those in the collected trajectories, we then train the Reviser with reinforcement learning on self-generated states, using rendering outcomes and reference-free quality feedback. ManimCode achieves 96.6% and 97.0% render success on ManimData-test and ManimBench, compared with 52.6% and 82.0% for the untrained 4B model under the same inference setup, while also improving video quality scores from 2.48 to 2.94 and from 3.00 to 3.44. Integrating ManimCode into two long-form educational video generation systems increases completed videos from 19/117 to 107/117 and from 53/239 to 239/239, alongside improvements in educational effectiveness and visual quality.
agenticself-correction - arxiv:2611.08290 · cs.LGAutoCompress: Automated Hardware-Guided Model CompressionSheldon Liang, Ali Taghibakhshi, Saurav Muralidharan, Sharath Turuvekere Sreenivas +3
LLM pruning must preserve quality while meeting hardware- and workload-specific constraints on latency, throughput, and memory. Existing pruning workflows often target parameter counts, requiring practitioners to manually select model sizes that meet latency, throughput, and GPU memory limits. However, parameter count alone does not predict these metrics, which depend on the model architecture, workload, hardware, and serving engine. We introduce AutoCompress, a hardware-guided compression framework that directly incorporates such constraints into differentiable architecture search. AutoCompress jointly optimizes global and per-layer structure to produce heterogeneous architectures, guided by serving-engine cost models validated against end-to-end latency, throughput, and memory measurements. We evaluate AutoCompress on long-context local coding, batched datacenter serving, and latency-sensitive in-game interaction. Compared to the state-of-the-art Minitron pruning framework, AutoCompress improves average accuracy across eight benchmarks by 6.57, 2.35, and 0.57 percentage points on RTX 5070, RTX 5080, and H100, respectively. AutoCompress also compresses Nemotron 3.5 Super from 120B to 74B parameters for single-B200 BF16 serving at 128K context, retaining 95% of its mean benchmark score.
memorylong-contextbenchmark - arxiv:2611.08286 · cs.CVInteractive Visual Gym: Evaluating Visual Interaction Models with Simulated Users in Interactive WorldsChenhao Zheng, Zixian Ma, Kai Li, Kanchana Ranasinghe +10
Visual interaction models that support users during ongoing physical tasks require evaluation in closed loop, where assistant responses can change subsequent user behavior and task outcomes. Existing benchmarks largely rely on offline replay and therefore cannot capture these effects. We introduce Interactive Visual Gym, a framework for closed-loop evaluation and training of visual interaction models through joint simulation of executable task environments and human users. We contribute three simulators that recreate everyday tasks and human-aligned action spaces across cooking, household manipulation, and mobile-interface domains, together with a persona-conditioned user simulation harness that collaborates with the assistant while incorporating controlled execution errors and other out-of-plan behaviors. We instantiate the framework as IVG-Bench, a benchmark with 315 interaction scenarios. Evaluating state-of-the-art VLMs as visual interaction models reveals substantial limitations: the best-performing model, Gemini 4 Argon, completes only 53.4% of tasks within the time limit, exhibits significant weaknesses in factual grounding and actionable guidance, and is often too slow to prevent errors when response latency is considered. We verify our evaluation protocol with a human study which finds 78.5% agreement between our interaction-quality metric and participant preferences. We further show that the environments provide useful learning signals: in a proof-of-concept GRPO experiment, closed-loop training improves both task success and interaction quality for Qwen3.5-4B model in our cooking environment.
manipulationbenchmarkevaluation protocol - arxiv:2611.08281 · cs.CLMultimodal-SEA: Benchmarking AI Agent Security under Multimodal Social Engineering AttacksZhitao He, Chi Him Zhang, Chengru Han, Yi R Fung
Personal AI agents face social engineering attacks supported by forged multimodal content, but existing evaluations mainly focus on text-based dialogue, leaving their intrinsic security under such attacks insufficiently studied. We introduce Multimodal-SEA, an interactive benchmark comprising 608 expert-designed tasks grounded in real-world fraud cases, covering 14 domains and 152 attack scenarios with prepared multimodal content. The tasks span four difficulty levels and reflect the distinct social and cultural contexts of Chinese and English environments. Agents inspect content, verify requests, and execute actions in simulated applications, with attack outcomes grounded in verifiable tool execution logs. To enable controlled red teaming of personal agents, we design a multimodal attack harness that integrates task information, attack progression control, and multimodal content construction to evaluate their intrinsic security against adaptive social engineering attacks in simulated environments. Evaluating seven widely used multimodal models across six dimensions reveals that all are vulnerable, with attack success rates of 7.1% to 34.5%. Adding multimodal content increases attack success from 19.9% to 29.2%, despite increased proactive verification coverage. Stronger safety instructions reduce attack success without improving recognition of deceptive multimodal content, while lower attack success rates for some models accompany excessive refusal of legitimate requests.
agentai agentbenchmark - arxiv:2611.08277 · cs.AILeanContext: Disentangling Mathematical Resources from Local Formal Context in Lean Whole-Proof GenerationHang Zhang, Qihao Ye, Litong Liu, Yunbei Zhang +6
Neural theorem proving increasingly uses retrieved premises and project context to support proof generation. Evaluating these resources requires measuring the advantage of target-selected declarations over comparable alternatives and how this advantage depends on training. We introduce LeanContext, a benchmark of 129,826 Lean whole-proof generation problems, to investigate these questions. We group selected complete preceding mathematical declarations as Mathematical Context (MC) and the remaining selected local source commands as Environment Context (EC). Four aligned input views support crossed training and inference evaluations with shared targets and original Lean verification environments. For both diffusion- and autoregressive-based models, selected declarations outperform comparable alternatives by a larger margin when training includes MC. Final-input audits identify omitted declarations referenced by the target. Prioritizing these dependencies improves Pass@32 by 1.40% and 1.70% for LLaDA and Llama, respectively, on 2,000 targets. The selection experiment operates without access to target reference proofs and holds checkpoints, EC, and inference budgets constant. The code and dataset are available at https://github.com/haaaardcvZ/LeanContext.
benchmark - arxiv:2611.08275 · cs.ROAligning Policy Encoders with a Shape TeacherMadelon Bernardy, Mohammad Mahdi Derakhshani, David M Knigge, Riccardo Valperga +1
3D-aware representations can improve robotic manipulation by providing geometric information that is robust to changes in viewpoint, object position, and occlusion. However, using 3D inputs at deployment requires additional sensing and processing, making image-based policies more practical. We ask whether an image-based policy can acquire 3D shape information during training while retaining a purely image-based interface at deployment. We propose shape-teacher alignment, a training-time mechanism that uses a pretrained 3D shape representation as an alignment target for an image-based policy encoder. Specifically, we use Point-MAE features from individual objects as targets for corresponding image patches, transferring geometric structure to the policy without changing its deployment inputs or architecture. On LIBERO, shape-teacher alignment produces image representations that are more structured around object shape and corresponding parts, while substantially accelerating policy convergence and improving success rates. For pretrained policies, convergence is accelerated by 4.7x and 5.6x on the Object and Spatial suites, respectively, without requiring the shape teacher or depth input at deployment.
manipulationlibero - arxiv:2611.08264 · cs.CVSplatAnything: Gaussian Spacing Normalization for Feed-forward 3DGS across Arbitrary ScenesHao Yu, Zhenwei Wang, Yongzhen Hu, Yuning Gong +6
We identify a key limitation of existing feed-forward 3DGS methods: Gaussian attributes are typically predicted in a radial-distance-normalized space that canonicalizes a global radial statistic of the reconstructed geometry, while the spacing between neighboring Gaussian centers can still vary substantially across scene scales and camera configurations. This mismatch induces systematic shifts in Gaussian-scale targets, leading to ambiguous supervision. To address this, we propose SplatAnything with Gaussian Spacing Normalization (GSNorm), which canonicalizes a robust scene-level statistic of Gaussian spacing to provide a more consistent spatial reference for Gaussian prediction. To rigorously assess our method, we further establish a systematic benchmark spanning diverse datasets, input sparsity levels, and challenging extrapolated viewpoints. Extensive experiments demonstrate that SplatAnything achieves state-of-the-art performance, with particularly strong gains under substantial variations in scene scale and viewpoint.
benchmark - arxiv:2611.08252 · cs.LGSeries-wise Cross-validation for Global Time Series ForecastingVitor Cerqueira, Mariana Oliveira, Luis Roque
We consider the problem of cross-validation (CV) in time series forecasting, with a focus on global forecasting models based on deep learning. The prevailing CV strategy in forecasting is to partition data across time with the last part of each series used for testing. With global models trained on collections of time series, however, a fundamentally different approach becomes possible: partitioning across series and time. We formalize series-wise CV for global forecasting models and benchmark several instantiations, including K-fold CV, bootstrapping, and Monte Carlo CV, on 11 datasets with 10 neural network architectures. We evaluate CV procedures along five dimensions: performance estimation, overoptimism rate, model selection accuracy, regret, and downstream forecasting performance. We withhold a random fraction of series from training so that they appear only in the test set. This protocol simulates the emergence of new series not seen during training, a common scenario in global forecasting. Under this protocol, our results reveal a nuanced trade-off: series-wise CV yields better downstream forecasting accuracy by guiding model selection toward configurations that generalize across both time and series, while time-wise CV retains a slight advantage for estimating out-of-sample error.
benchmark - arxiv:2611.08249 · cs.LGGeometric Relevance Is Not Sufficient: Evaluating Focused Post-Training Supervision in Diffusion DockingMichael Brown, Maryam Rahimimovassagh, Mehdi Yazdani-Jahromi, Sanjeda Sara Jennifer +1
Diffusion-based molecular docking requires learning a corrective score field across a broad corruption distribution, yet individual pose errors can differ substantially in their structural consequences. This creates a post-training conflict between preserving broad score-field coverage and concentrating supervision on corrections that appear more relevant to docking geometry. Using DiffDock, a fixed-budget study compares broad continuation with torsional supervision focused on Cartesian consequence or conditional target consistency, and with a broad/focused mixture. Evaluation separates fixed score fidelity, generated-pose quality, best-of-10 sampling capacity, and confidence-selected Top-1 docking performance. At <5 Angstrom, broad continuation had the highest mean Top-1 success among trained policies, and focused or mixed supervision showed no consistent advantage under the tested budget. The starting model had better fixed score fidelity, whereas the evaluated broad-continuation model had higher mean Top-1 success with the same aggregate best-of-10 recovery. Focusing supervision on structurally consequential or more consistently specified corrections did not consistently improve returned-pose success over broad continuation.
post-training - arxiv:2611.08240 · cs.LGFewStepRL: Fine-tuning Few-step Diffusion Models with Reinforcement LearningHaitao Lin, Minsi Ren, Yang Li, Xingtong Ge +3
Reinforcement learning (RL) has become a standard post-training method for diffusion models, but most RL objectives roll out the current policy at every iteration, and with a many-step sampler these rollouts typically dominate the cost of training. Few-step models cut this cost, yet they model an average velocity over an interval, whereas existing RL objectives are built for an instantaneous velocity with a many-step sampler. We adapt RL objectives to few-step models systematically in three phases, namely a transformation, a decomposition and an adaptation. The transformation to $x$-space expresses each objective we consider through a common policy prediction, at the clean end by default and at an explicitly marked level for transition or flow-map objectives, so that objectives stated over densities, velocities or rewards become comparable. The decomposition writes the gradient of each of these objectives as the sum of a self-regression toward a reference and a reward augmentation, so that a method reduces, at the gradient level, to a choice of anchors and weights. The adaptation computes each anchor from a few-step model at the end of its transport or along its tangent, with the instantaneous velocity recovered from the network itself or from an auxiliary network. This yields few-step versions of FlowGRPO, AWM, DiffusionNFT, DiffusionOPSD and ReFL in one modular codebase. In our experiments, reward-gradient-based methods are the most effective when the reward is differentiable and its gradient is affordable; otherwise DiffusionNFT-style methods generally reach high rewards in relatively few iterations despite a sometimes higher per-iteration cost. Extra inference steps show no systematic aggregate benefit in our experiments, although selected outputs visibly improve with two to four times as many sampling steps as in training.
post-training - arxiv:2611.08237 · cs.CLTerminal-style Mathematical Modeling Bench: The First Agentic and Verifiable Benchmark for Mathematical ModelingXiaojun Wu, Cehao Yang, Xuhui Jiang, Chengjin Xu +2
Mathematical modeling turns an open-ended real-world problem into assumptions, a model, a solution, and a defended recommendation, and its answers matter only if they are both close to a credible reference and physically feasible; existing benchmarks either score solver fragments or grade reports with an LLM judge, leaving those numbers unchecked. We introduce TB-MathModel (the Terminal-style Mathematical Modeling Bench), the first benchmark to make complete contest-style modeling problems agentic, verifiable, and terminal at once, over 18 real CUMCM and MCM problems (2023--2025). Each task ships the verbatim statement, official attachments, and figures in a containerized terminal; the agent returns one JSON artifact, scored on the numeric core of the final question against a reproduction of the Outstanding-award paper. Two design choices distinguish it: direction-aware, one-sided scoring, which never penalises beating the anchor, and two-layer feasibility verification---a programmatic gate over six decidable invariant forms plus 624 Lean~4 obligations---with a hard gate that zeroes infeasible answers. Each task is scored against a panel of human outstanding papers rather than one, so a model must be close to many accepted answers; across seven multimodal entries (three seeds each) the strongest reach a multi-reference O-Eval of 64.58\%, while the \$2.54 open model holds 59.81\% at $1/100$ the cost. TB-MathModel turns the numeric core of modeling into a reproducible, machine-checkable signal for measuring long-horizon modeling competence and for collecting agent trajectories. The source code is available at https://github.com/wxj630/terminal-bench-math-modeling.
agentagenticbenchmark - arxiv:2611.08236 · cs.LGBeyond Quadratic Attention: Separable Gaussian Operators for ODE Vision TransformersCarlos Boned Riera, David Romero Sanchez, Oriol Ramos Terrades
Self-attention carries a quadratic cost in token count, which limits how far Vision Transformers can scale with input resolution. We introduce SSoG, a separable sum-of-Gaussians spatial operator that replaces dense self-attention with a learned mixture of axis-factorized Gaussian kernels, reducing the per-layer cost from $\mathcal{O}(N^2)$ to subquadratic in $N$. Unlike dot-product self-attention, whose Lipschitz constant is unbounded outside a compact input domain, SSoG with fixed kernels is globally Lipschitz continuous, which, via the Picard--Lindelöf theorem, guarantees that the continuous-depth variant has a unique solution on the whole integration interval, a global guarantee that dense-attention counterparts lack in general. On CIFAR-100 at $224\times224$, in a matched discrete-Transformer setting and in the continuous setting, SSoG improves Top-1 accuracy by 2.2--4.1 points across three model widths while using 8.8--16.6\% fewer parameters, with depth and training recipe held fixed. In a looped-Transformer setting, SSoG improves Top-1 accuracy by 5.4--8.1 points. We further probe inference-time scaling by evaluating on inputs up to $1024\times1024$ at matched batch size: in the continuous-depth (Neural ODE) architecture, SSoG reaches up to 154\% higher throughput and over 60\% lower peak memory than dense attention, with the gap widening as resolution grows, consistent with the difference in asymptotic cost. In the discrete and looped settings, SSoG also achieves over 30\% higher throughput than dense attention at this resolution. Together, these results establish SSoG as a practical, parameter-efficient, and theoretically grounded alternative to dense attention across discrete, weight-shared, and continuous-depth Transformer variants.
memory - arxiv:2611.08231 · cs.CLFrom Twisted Facts to Rogue Agents: Can LLMs Reliably Detect Multi-Dimensional Hallucinations?Hanyu Qiu, Ze Liu, Zhe Zhao, Kuang Xiong +3
Large language models (LLMs) are widely used for information access, content production, and agent-based task execution, but their outputs remain vulnerable to subtle, high-risk hallucinations. To systematically evaluate and advance post-generation hallucination detection, we introduce HaluCheck, a large-scale, bilingual benchmark comprising over 42k instances spanning both factual hallucinations in text and faithfulness hallucinations in agent trajectories. HaluCheck features fine-grained, multi-level annotations for detection, localization, and correction, covering different tasks and perturbation degrees. Experiments on fourteen LLMs reveal that, while models show basic sample-level detection ability, they struggle with fine-grained localization, correction, and agent hallucination detection. Further analysis demonstrates that thinking, web search, and supervised fine-tuning each offer partial improvements, but none fully solve the problem. HaluCheck thus provides a solid basis for benchmarking and diagnosing LLM reliability across text and agent scenarios.
agentbenchmark - arxiv:2611.08226 · cs.CL"Blind Compliance": Coercive Control in AI JailbreakingAcacia L. Ackles, Elliott Whitson
The explosion of generative AI systems and concerns surrounding their rapidly scaling capabilities has led to an adversarial cycle between increasingly narrow content restrictions ("guardrails") and increasingly sophisticated structured prompting techniques ("jailbreaks"). Technological frameworks surrounding such jailbreaking remain largely focused on the system's output: whether harmful or disallowed content is ultimately produced for the user. However, this approach, which draws largely from classical cybersecurity, does not address the uniquely relational dimension introduced by generative AI: in many cases the user is not just accessing a technology via an exploit, but is in conversation with it as part of the exploit. Here we present an alternative lens which centers the interaction rather than the output. Drawing on interpersonal violence scholarship, we present a novel framework of coercive control categories with a particular focus on aspects which are embodied versus disembodied. We then map how these disembodied coercive control dynamics are structurally embedded in modern AI training paradigms, and how those embeddings are exploited via a particular suite of persona-based jailbreaks. In this lens, we reframe AI safety and alignment as not merely a technical security problem, but as another surface upon which we see enacted the relational dynamics of power and control.
embodied - arxiv:2611.08219 · cs.CLAgent-R$^2$: Scaling Agents for Realistic Productivity Work via Real-Source Data SynthesisTeng Pan, Zixuan Wang, Yuchen Yan, Qifei Wu +4
Large language models have made rapid progress in agentic reasoning and tool use, raising expectations for agents that can take over realistic productivity work from humans. However, current agents still struggle to finish such tasks, where success requires following long-horizon professional workflows, understanding heterogeneous file contexts, and producing usable deliverables under fine-grained requirements. A key bottleneck is the lack of scalable post-training data that reflects real work intent, authentic file environments, and deliverable-level evaluation. To close this gap, we introduce Agent-R$^2$, a $\textbf{r}$eal-source data synthesis framework that scales training data for $\textbf{r}$ealistic productivity work. Instead of relying on manually specified workflows or LLM-generated files, Agent-R$^2$ grounds synthesis throughout the pipeline in real occupational seeds, corpus-derived workflow evidence, and authentic web-sourced files. From 10k web-corpus candidates, Agent-R$^2$ distills 4k task instances that pair real multi-format attachments with fine-grained, deliverable-level rubrics. On this data, we train Agent-R$^2$-4B via supervised fine-tuning and rubric-based reinforcement learning, improving both end-to-end productivity performance and file-grounded reasoning. Trained on only 4k instances, Agent-R$^2$-4B surpasses Qwen3-235B-A22B-Thinking-2507 on both GDPval (51.0$\%$ vs. 48.0$\%$) and OfficeQA (14.8$\%$ vs. 13.8$\%$). Further analysis shows that Agent-R$^2$ mitigates shortcut behavior, where agents produce superficially plausible deliverables before sufficiently grounding, constructing, or verifying them, while promoting sustained exploration.
agentictool usepost-training - arxiv:2611.08217 · eess.SYOptical Network Digital Twin: From Digital Coherent Transmission to Intelligent Network OperationsHideki Nishizawa, Enrico Miotto, Andrea Rosso, Takahito Tanimura +19
Optical Network Digital Twins (ONDTs) are emerging as a key framework for representing, synchronizing, and utilizing the physical state of optical networks for quality-of-transmission (QoT) estimation, optimization, and intelligent operation. This survey provides a comprehensive view of ONDTs, spanning digital coherent transmission and physical-layer observability to intelligent network operations. We systematically organize ONDT technologies from ecosystem, functional, control, and implementation architecture perspectives and review the enabling technologies required to construct, synchronize, and utilize digital representations of physical optical networks. As a unifying framework to describe both ONDT design and validation, we introduce a categorization based on the following figures of merit: Fidelity, Dynamicity, and Deployability (FDD). In particular, we emphasize fit-for-purpose Fidelity, where modeling effort focuses on the physical effects of components rather than solely maximizing model accuracy, favoring scalability and generalizability. Furthermore, experimental and field validations are examined to evaluate the practical applicability of ONDT technologies. We further examine experimental and field validations, from physics-based QoT estimation and measurement-informed model refinement to field synchronization, service provisioning, multi-domain observability, and infrastructure characterization. Finally, we discuss the complementary roles of ONDTs and AI agents toward autonomous optical networks and identify key research directions for practical deployment.
ai agent - arxiv:2611.08216 · cs.CVProQuant: Progressive Quantization-aware Training for Edge MLLMsYufei Xue, Yushi Huang, Jiawei Shao, Pingcheng Dong +4
Multi-modal large language models (MLLMs) with impressive perception and understanding capabilities will enable a range of downstream applications. However, their substantial computational and memory requirements hinder the real-world implementation, particularly for resource-constrained edge devices. Quantization has emerged as an effective solution for deploying large models. While post-training quantization (PTQ) successfully retains the performance at moderate bit-width (e.g., W8A8), it suffers from severe performance degradation under low-bit quantization (e.g., W4A8), especially for small-scale models that are more suitable for edge deployment. In this paper, we propose ProQuant, a novel weight-activation quantization-aware training (QAT) framework tailored for low-bit MLLMs. First, to align the limited finetuning dataset distribution with the pretrained dataset distribution, we propose a self-relabeling (SRL) scheme to resample the responses of the open-source multimodal dataset. Second, we are the first to take a decoupled view of multi-modal (MM) understanding and language modeling abilities. We propose a progressive block activation (PBA) mechanism to prioritize the MM understanding recovery. We also introduce a vision-sensitive factor $\vecτ$ that effectively identifies the MM-critical layers. Extensive experiments across various benchmarks, covering models with 2B-8B parameters, show that ProQuant significantly outperforms existing methods. For example, our W4A8 Qwen3-VL-2B-Instruct improves average accuracy by impressively 6.6% and achieves performance comparable to full precision counterparts with only approximately 1% performance drop. Code is available at https://github.com/yufeixue-ai/ProQuant/.
memorypost-trainingbenchmark - arxiv:2611.08209 · cs.RORoboARCade: Benchmarking Agentic Robot ControlSatvik Sharma, Hengyuan Hu, Mihika Kamat, Dorsa Sadigh +1
Improvements in multi-modal capabilities have enabled agents to act directly as robot policies, a setting we refer to as agentic robot control (ARC). We introduce RoboARCade, a benchmark of long-horizon manipulation tasks across three embodiments designed to test in-context adaptation, memory, and reasoning for precise control. The tasks build on existing open-source benchmarks, which we make harder by widening the range of initial conditions. RoboARCade is easy to use and extend: all tasks run on a single simulation backend, MuJoCo, each task is a self-contained data pack, and agents connect through interchangeable interfaces, so new tasks, agents, and interfaces can be added independently. The benchmark separates leading agents: GPT-6 Astra achieves 49.8% compared with 43.6% for Claude Fable 5.1, leaving substantial room for improvement. Comparing two interfaces, we find that a richer visual interface adds only 1.5 to 6.8 points at 1.7 to 2.3 times the cost. Failures stem mostly from choosing a poor manipulation strategy or executing a good one imprecisely, rarely from misunderstanding the scene, and agents incorrectly claim success on 24% to 39% of attempted subtasks. Agents can also improve without human input: one successful demonstration in context raises Astra's score from 53.5% to 73.0% on a five-task subset, and a prompt the agent writes and revises itself reaches 72.0%. Code, videos and full traces are available at roboarcade.org.
manipulationagentagenticbenchmark - arxiv:2611.08196 · cs.AISMART: State-based Multi Agent Reasoning Traces for Verification-oriented Problem-SolvingSai Vallurupalli, Francis Ferraro
Complex problems with limited context, such as claims, require a means to obtain additional knowledge necessary for solution reasoning. We propose a multi-agent framework, SMART, to obtain knowledge through capturing the implicit state of participating entities in a problem context. SMART curates the state knowledge as diverse, complementary and solution-salient premises using a variety of LLMs and a structured role-based prompt strategy. Using knowledge from these real-world verified curated premises, SMART's knowledge guided solution reasoning improves validation performance up to 20% on complex scientific claims spanning 3 datasets. We show that grounding premises in real-world knowledge is necessary and improves performance up to 5%.
agentmulti-agentagent framework - arxiv:2611.08177 · cs.AIBabyTheorist: A Benchmark for Learning to Theorize the World from Observation AloneDoojin Baek, Junyeob Baek, Mingyu Jo, Hosung Lee +3
How would we know whether a machine truly understands the world? We argue that one key capability is \textit{observational theory learning}: acquiring reusable theories from pure observations and applying them to new situations. Current visual reasoning benchmarks focus primarily on the latter, presenting each task as a few-shot puzzle grouped by a common rule. We introduce BabyTheorist, a benchmark for controlled training and evaluation of observational theory learning. BabyTheorist generates continual streams of before-after visual observations from hidden compositional programs; during training, learners see only the resulting transitions, with no program labels, task identities, or support sets. To support theory learning and evaluation, BabyTheorist controls the program language, a curriculum over program complexity, recurring frequent subprograms, and a continual train-test protocol with level skipping. Seven representative baselines plateau well below ceiling, with the gap widening as the curriculum advances---evidence that BabyTheorist isolates a capability current program-learning approaches do not yet address.
benchmark - arxiv:2611.08175 · cs.AIPCBRouteAgent: How Close Are LLM Agents to Becoming PCB Routing Engineers?Yue Zhang, Jilei Sun, Liqiang Jing, Yijie Wei +1
Automating printed circuit board (PCB) routing requires coordinating board-level connectivity decisions with geometric execution under strict design rules. Although recent agents can interact with native routing tools, reliable routing on complex boards remains challenging. We present PCBRouteAgent, a multimodal PCB routing agent inspired by how human engineers coordinate global strategy with local execution. A Global Routing Planner combines board images and structured electrical and geometric information to specify connection topology, routing order, corridors, and layer transitions. A Detailed Routing Executor realizes these guides through KiCad's native routing engine and performs bounded local recovery. Unresolved failures are returned as structured feedback, enabling targeted guide revisions while preserving accepted routes. We evaluate PCBRouteAgent on 247 synthetic and real-world boards from four PCBWorld-Bench test splits. It achieves clean-pass rates of 100%, 94.9%, 100%, and 40% on D2, D3-A, D3-B, and D3-C, respectively. On the three real-world subsets, PCBRouteAgent outperforms the best evaluated baseline in clean-pass rate by 13.9, 36, and 10 percentage points, respectively.
agentllm agent - arxiv:2611.08156 · cs.ROLa-JEPA: Compact Robot Policies with Latent World ModelingMateo Clémente, George Z. Xue, Amin Abyaneh, Liheng Ma +4
World-action models adapt pretrained video generators, and vision-language-action models build on pretrained vision-language backbones; both tie policy architecture and inference cost to large transformers designed for generation rather than control. It remains unclear how much these generation-oriented choices affect control. In this work, we introduce Latent Action JEPA (La-JEPA), a compact world-action model whose policy transformer is trained from scratch in the feature space of a pretrained V-JEPA 2.1 encoder. Following Fast-WAM, La-JEPA learns to predict future features jointly with actions during training and generates only actions at inference. Among policies without embodied pretraining, La-JEPA achieves state-of-the-art performance in clean and randomized RoboTwin 2.0 scenes and on LIBERO-Plus with correct instructions, while using far fewer parameters than the strongest competing policies. It runs substantially faster than Fast-WAM at comparable performance and outperforms the closest JEPA-based methods, LeapBot-WA and JEPA-WAM, on LIBERO-Plus and RoboTwin 2.0 without their pretrained policy backbones. On four real-robot tasks, its average success is comparable to pi0.5 and above non-embodied-pretrained baselines. Through a design study of the choices behind La-JEPA, we show that JEPA features outperform video VAE latents, most clearly out of distribution, and that training-time future prediction choices significantly impact the policy even when future prediction is skipped at inference. We correct the LIBERO-Plus instructions, as the officially released evaluator corrupts them for most variants, and re-evaluate several strong open baselines. Without a generative pretrained policy backbone, La-JEPA is better and faster than far larger policies, suggesting that robot foundation models may benefit from architectures designed for control rather than inherited from generation.
vision-language-actionembodiedpi0robot foundation modelliberorobotwin - arxiv:2611.08153 · cs.AIA Generative Foundation Model for Multimodal Tokamak TasksYingshuai Ji, Haodong Feng, Guohong Deng, Gelin Bian +1
Tokamak experiments record diverse measurements, including waveforms, profiles, fields, and scalars, which are used to reconstruct the plasma, forecast its evolution, recover associated control signals, and estimate disruption risk. Most models are built separately for one task and a fixed set of inputs and outputs, and each relies on a prescribed subset of diagnostics rather than structuring the full multimodal observation set to extract global, cross-signal information about the discharge. These limitations motivate a reusable foundation model that adapts to each task's input-output schema and models heterogeneous target signals jointly, rather than retraining a separate model for every configuration. In this work, we extend the TokaMark benchmark with two groups of tasks for actuator action inversion and disruption prediction, yielding 20 MAST tasks in six groups with explicit schemas and evaluation protocols. Then, we introduce GenToka, a generative tokamak foundation model that jointly generates schema-selected targets through a shared conditional diffusion Transformer, with signal-specific codecs preserving native time series, profiles, fields, and scalars; foundation pretraining and task-specific adaptation transfer cross-signal structure across all six groups. Evaluation on MAST measurements shows that GenToka reduces macro-average normalized root mean squared error relative to TokaMind by 4.7% on Groups 1-4 and by 17.7% on Group 5, and improves macro-average window-level F1 on Group 6 by 8.4% at a common validation-selected threshold. GenToka provides a reusable, generative, and multimodal pathway for experimental tokamaks, linking plasma-state reconstruction, actuator inference, and disruption warning from shared measurements to support more integrated, data-driven fusion plasma prediction and control research.
benchmarkevaluation protocol - arxiv:2611.08147 · cs.CVMCUFlowNet: Hardware-Aware Neural Architecture Search for Optical Flow on MicrocontrollersEnmin Lin, Jona Beysens
Dense optical flow allows small drones to avoid obstacles. However, few dense optical flow networks are designed for microcontrollers (MCUs) with neural accelerators, whose on-chip SRAM limits the input resolution. We present PACE-NAS, a macro-to-micro neural architecture search framework that decouples memory from latency: a macro design fixes the backbone and its peak SRAM, and a micro design searches operators and depths to trade accuracy against compiler-estimated latency. We apply it to EdgeFlowNet on a Grove Vision AI V2 board (Ethos-U55), where profiling locates the peak SRAM at the full-resolution output head. Replacing transposed convolutions with bilinear upsampling and convolutions lowers the peak SRAM by 18%, and two lightweight context modules lower the endpoint error (EPE) without raising the peak. All searched subnets share this peak, so the search needs no memory constraint. We name the Pareto-optimal models MCUFlowNet; the fastest variant, MCUFlowNet-S, runs 1.76x faster than a retrained EdgeFlowNet at a 1.6% higher EPE and, in simulated forest flights at 2 m/s, reaches the goal without collision in 51.5% of trials, against 41.8%. PACE-NAS also applies to semantic segmentation, where a searched PIDNet-S variant runs 1.17x faster at similar accuracy. Code is available at https://gitlab.kuleuven.be/emedia-research-lab/esat-stadius/mcuflownet/mcuflownet.
memory - arxiv:2611.08144 · cs.RODetecting Silent Faults at the Environment-Trainer Boundary in Reinforcement LearningZuo Gou, Chenhao Lu
Robot reinforcement learning increasingly runs environments outside the trainer: in another process, on another machine, in another language. Many faults at this boundary raise no error. We make the boundary a wire protocol and ask what notices a fault in it. On Pendulum, HalfCheetah, Hopper and a 7-joint arm (Pusher), we injected up to 23 faults into what crosses the wire and 2 into the trainer's log, from one value to every value. Unless every value was wrong, a learning-curve threshold caught at most one run in ten, and a comparison with ten fault-free seeds at most 21%. A clean evaluation of all 1,060 final policies agreed with the curves: of the 728 boundary-fault runs both curve checks passed, 723 scored no more than three standard deviations below fault-free seeds, though nearly all ended on other weights and, for reward faults, logged other returns. Checks of the run itself, not of its outcome, saw every fault. A step ledger kept on both sides and compared step by step caught and located every fault that crossed, on Pendulum, Hopper and Pusher on one machine and on Pendulum across two, with no fault-free run and, on Pendulum, a median 3% more training time. An episode record kept at the environment caught the faults in the trainer's log. We list such faults found in practice, in Gymnasium 1.3 and our own code and, reported by others, in five multi-agent libraries, and recommend a step ledger and episode counts kept at the environment.
multi-agent - arxiv:2611.08139 · cs.ROH-TRIP: Exact Many-to-Many Distance Index for Multi-Robot ExplorationElif A. Özaslan, Tolga Özaslan
Frontier-based exploration systems repeatedly evaluate geodesic travel costs between robot poses and candidate viewpoints, and tour-based planners consume the full pairwise distance matrix built from them. Recomputing these matrices with one graph search per source scales with the explored map size, whereas distance caches that refresh only locally affected or newly created frontiers can retain stale costs between surviving pairs when newly discovered free space creates a global shortcut. We present H-TRIP (Hierarchical Tropical Rank-one Incremental Propagation), an exact distance index grounded in Kleene algebra over the min-plus semiring for monotonically growing occupancy grids. H-TRIP partitions the known map into a spatial hierarchy with fixed boundary ports. Each unknown-to-free revelation is decomposed into efficient rank-one min-plus edge insertions at the affected leaf, and only decreased port-to-port summaries are propagated along its ancestor chain. A shared hierarchical lift evaluates the dense many-to-many matrix of shortest distances for all active query points on the occupancy map. We prove that sequential insertions preserve exact shortest-path distances on the grid. Benchmark experiments confirm exact agreement with reference searches on fully built maps and show that dense many-to-many queries at a few dozen viewpoints are up to two orders of magnitude faster than sequential k x BFS. On exploration runs with sixteen robots, in-loop queries on the growing occupancy graph are 2.21x-8.59x faster than co-scheduled k x BFS on maps with side 256, 512, and 1024 cells.
benchmark - arxiv:2611.08134 · cs.AIWho Can Evaluate AI Independently? A Framework for Analysing Evaluation InfrastructureSumaya Nur Adan, Joanna Wiaterek, Nourhan Abdelrhim, Sydney Reis +3
AI evaluation is increasingly central to AI safety and governance, yet taking part in an evaluation does not by itself confer authority over its design, its execution or the resources it depends on. This paper examines how infrastructure arrangements distribute that authority, and so shape who can evaluate AI systems independently. Drawing on Science and Technology Studies, political economy and applied evaluation standards, we introduce PAAI (Power, Access, Adequacy, Independence), a framework for examining who controls evaluation infrastructure, the terms on which it is available, whether it is adequate to the evaluation's purpose and what autonomy it supports. We apply PAAI to twelve African AI evaluation projects (2022-2025) and to the documented roles of all ten member institutes in the three published joint testing exercises of the international network of AI Safety Institutes. We identify three recurring configurations: extractive collaboration, dependent innovation and constrained independence. African actors often contribute data, language expertise and task design while external institutions retain authority over infrastructure and evaluation frameworks; design authority without infrastructure authority never yielded more than partial independence. The joint testing exercises show a similar division among mostly high-income members: the institutes that led each exercise also ran the models, while Kenya, the network's only African member, moved from annotation to task design with no documented model execution. Participation in evaluation does not by itself produce evaluative authority. For evaluation to serve as a credible governance mechanism, access must come with control over priorities, methods, publication and continuity.
evaluation framework - arxiv:2611.08133 · cs.CLBeyond K/V Caches: Cross-Model State Handoff in Hybrid Linear-Attention LLMsJiacheng Cai, Tao Zhang, Siqi Liang, Yun Dai +1
Production deployments often use different models to balance cost and quality through cascading, routing, and switching models during long conversations. Context accumulates across turns. Each switch requires the target to repeat prefill over this context. Existing methods transfer full-attention (FA) K/V caches across models. Hybrid linear-attention models also retain recurrent states (REC) and convolution histories (Conv). The interaction between Conv and REC complicates cross-model state transfer. Conv perturbations can change keys and values in recurrent layers and alter REC updates. Conv histories are overwritten, but changes in REC can still affect later outputs. More accurate REC reconstruction need not improve continuation. Building on this, we propose Hybrid State Transfer (HST) for all three state types. It uses structured REC maps and preserves history order during Conv alignment. We calibrate the adapters using the frozen target's outputs during continuation. FA cache map updates are merged offline. Across five benchmarks, Qwen3.5 4B$\rightarrow$27B and 9B$\rightarrow$27B achieve mean task retentions of 89.88% and 94.11%, respectively. With source caches already available, separate timing on synthetic 16K inputs yields handoff body speedups of $21.66\times$ and $21.91\times$.
benchmark - arxiv:2611.08128 · cs.CVFreeOcc: Decoupling Ego-Motion for Efficient 4D Occupancy Forecasting via Continuous Flow MatchingZeping Zhang, Zhuoya Zhao, Samy Metari, Robert Laganière
Current 4D occupancy world models use heavy cross-attention or deep recurrence to disentangle ego-motion from scene dynamics. This requires significant computational capacity. However, this effort is largely redundant since the ego-vehicle's intended future motion is provided by the downstream motion planner. We present FreeOcc, a lightweight continuous-time world model that introduces explicit $SE(2)$ geometric decoupling into a conditional flow-matching framework. By mathematically factoring out self-motion, we create a purely dynamic feature space. This space supports attention-free temporal fusion and allows physics-constrained ODEs to operate strictly on residual agent kinematics. To maximize deployment efficiency, FreeOcc introduces an uncertainty-aware blending mechanism. This routes well-observed voxels to a single-pass decoder, reserving the ODE solver exclusively for ambiguous, high-uncertainty regions. Requiring only 28.3M parameters-roughly 40% of comparable models-our optimal ensemble mode (FreeOcc-E) runs at 27 FPS. On nuScenes, FreeOcc achieves state-of-the-art forecasting performance with a 0.47 m Near Field Chamfer Distance (NFCD) and 32.1% mIoU. Furthermore, the model transfers zero-shot to SemanticKITTI. It also proves highly robust, retaining 92.8% of its ideal mIoU under significant sensor packet loss and timestamp jitter.
world model4d occupancyoccupancy worldagent - arxiv:2611.08127 · physics.opticsAn Agent-Driven, Cross-Validated Inverse-Design Pipeline for Dielectric MetasurfacesGiuseppe Emanuele Lio, Roberto Caputo, Alain J. Corso, Maria Guglielmina Pelizzo +1
Inverse design of dielectric metasurfaces is powerful but labour-intensive: each new target requires choosing a parameterization and a figure of merit, managing compute resources, and critically assessing results that may be numerically unconverged, sensitive to fabrication errors or trapped in local optima. Here we present an open-source pipeline in which an artificial intelligence (AI) coding agent turns a plain-language specification into metasurface optimization runs, executes them, analyses every run with a fixed checklist and proposes the next iteration. Designs are optimized with differentiable three-dimensional rigorous coupled-wave analysis (RCWA), including robustness to fabrication tolerances, and every reported design is independently verified with finite-difference time-domain (FDTD) simulations. We demonstrate the workflow at 1550 nm on a freeform silicon metagrating, on a polarization-independent meta-atom library assembled into beam-deflecting supercells, and on metalenses built from the same library. In the metagrating case, the independent FDTD verification revealed a 5\% overestimate of the efficiency by the optimization model that a conventional convergence test had not detected, and the robust design retained most of its efficiency under $\pm$10~nm fabrication errors. In the other cases, instead, the obtained RCWA results are confirmed by full-wave simulations.
agent - arxiv:2611.08113 · cs.LGIdentifying Neural Dynamics Using Nonlinear Interventional State Space ModelsHanna M. Tolle, Carles Balsells-Rodas, Yixin Wang, Pedro A. M. Mediano +1
State-space models (SSMs) recover low-dimensional latent dynamics from high-dimensional observations, but latents are generally not identifiable from observational data alone. Interventional state-space models (iSSMs) use perturbations to break observational symmetries and identify parameters up to permutation, scaling, and shift. However, prior work restricts identifiability to linear latent dynamics and piecewise-linear emissions, excluding many behaviors characteristic of neural circuits. We extend iSSMs to a substantially broader model class: we prove that (i) affine identifiability holds under general injective emission functions and Gaussian latent transitions, and (ii) with interventions, identifiability up to permutation, scaling, and shift is preserved under general nonlinear latent dynamics. We validate our theory on simulated FitzHugh-Nagumo dynamics, and then demonstrate two applications: adjudicating between competing models of decision-making; and predicting responses to unseen micro-stimulation in macaque and human neural recordings, with proof-of-concept generalization to a held-out stimulation site in the latter. Finally, simulations yield practical guidance for designing stimulation protocols that improve identifiability.
latent dynamics - arxiv:2611.08111 · cs.ROReinforcing End-to-End Autonomous Driving Safety via Negative-Aware Planning in a Unified World-Action ModelZihao Zhang, Xinyu Zhang, Bing Zeng, Xingyu Chen +3
World-action models (WAMs) jointly model future traffic states and driving actions for end-to-end autonomous driving in intelligent transportation systems. However, their action learning remains dominated by expert-trajectory imitation, leaving unsafe alternatives close to expert trajectories insufficiently constrained. To improve WAM robustness in safety-critical scenarios, we propose NegWAM, a safety post-training framework for unified WAMs that learns explicit safety boundaries from near-expert failure trajectories. NegWAM constructs negative trajectories that are geometrically close to the expert trajectory yet carry safety risks, uses them as repulsive anchors to steer predictions away from dangerous behavior, and encodes them as scene-conditioned rider tokens to learn scene--trajectory compatibility. During training, NegWAM jointly optimizes the flow-matching and near-expert negative objectives, while the auxiliary rider branch is discarded at inference, incurring no additional overhead. On NAVSIM v2, NegWAM achieves 40.4 EPDMS on navhard, a 4.1% relative improvement over the strongest compared scorer-free WAM baseline, and achieves 93.3 EPDMS on navtest. These results demonstrate that NegWAM achieves strong performance in safety-critical scenarios while preserving routine-driving performance. Zero-shot closed-loop evaluation on HUGSIM further demonstrates cross-benchmark generalization. The code is available at https://github.com/Zizizi-hao/NegWAM.
post-trainingbenchmark - arxiv:2611.08100 · cs.ROAn Integrated Open-Source Stack for Vision-Based Autonomous Navigation of Wheeled RoversIaroslav Kolomiets, Iaroslav Sheipak, Mikhail Reppo, Vsevolod Egorov +2
Vision-based rover autonomy requires integrating perception, localization, planning and platform-specific control under limited sensing and computing resources. We present an open-source ROS 2 stack for following directional arrows, reaching a terminal cone and returning autonomously to the start. It connects YOLOv8n detection and geometric direction, range and bearing estimation with RGB-D SLAM, wheel-inertial state estimation, kinematics-aware planning, behavior-tree mission management and STM32 actuation. Standard interfaces separate reusable autonomy from the hardware layer; a proportional path-following plugin supports low-speed rover motion. Public resources include software, configuration, a labeled dataset, trained weights and evaluation logs. On 478 held-out images, the detector achieves [email protected] of 0.720 under the Ultralytics evaluator and outperforms a classical OpenCV baseline under a shared evaluation protocol. The reported perception latency is 25 ms on a Jetson Orin NX. Three missions, one indoor and two outdoor, demonstrate marker-guided navigation and autonomous return without operator motion commands or reported collisions. Field results document integration failures and corrective choices; repeated instrumented trials remain necessary to quantify physical navigation errors and reliability.
evaluatorevaluation protocol - arxiv:2611.08097 · cs.ROSTEP: Spatiotemporal Training with Explicit Prediction for Perceptive Humanoid LocomotionYouhao Lyu, Chenghao Sun, Zhenxuan Ding, Shuting Zhang +2
For perceptive humanoid locomotion, the relevance of terrain geometry depends on the robot's body state and how both evolve under its actions. Existing work has advanced terrain-aware sensory fusion and cross-modal prediction, improving how visual and proprioceptive information is combined and related. However, feature aggregation alone does not specify which cross-modal relationships should be learned, while prediction between simultaneous observations does not explicitly constrain their action-dependent evolution. We present Spatiotemporal Training with Explicit Prediction (STEP), a one-stage framework that jointly supervises cross-modal relationships and action-dependent state evolution within a shared control representation. Multiple learned queries aggregate depth and proprioceptive tokens into a compact policy state. Bidirectional masked prediction uses each visible modality to predict latent features of the omitted modality, explicitly supervising terrain-body relationships. An action-conditioned predictor maps the current policy state and executed action to the next fused state, with an exponential-moving-average target defined directly in the policy representation space. Simulation comparisons show overall improvements in terrain traversal over evaluated fusion alternatives and ablations, while depth-dropout and delay experiments demonstrate the benefits of both predictive objectives under degraded visual input. Deployed on a 23-DoF Noetix E1 humanoid with onboard sensing, STEP succeeds in 98 of 100 hardware trials across stairs, a 40\,cm platform, a 70\,cm gap, stepping stones, and backward stair traversal without rearward terrain visibility.
humanoidaction-conditioned - arxiv:2611.08094 · cs.LGPersonaAgent: Retrieval-Augmented Transaction Anomaly Scoring with Refreshable Multi-Entity Persona MemoryZhengyu Fang, Liang Wang, Yingtong Dou, Yuzhong Chen +7
In large-scale payment networks, entities (card, merchant, etc.) and their interactions play a key role in transaction anomaly detection. However, transaction patterns between different entities usually change more frequently than the production model refresh schedule. Meanwhile, entity embedding tables suffer from high-cardinality and sparse interaction issues. To resolve the above challenges, we present PersonaAgent, an agentic transaction anomaly detection framework which represents the latest entity behavior signals as personas and saves them in a refreshable memory. Offline, a weekly scheduled batch job is used to extract the latest behavior signals and update them in the memory. At inference time, a lightweight adapter retrieves the corresponding entity behavior signals of each transaction from memory and uses them to augment the base anomaly classifier's scoring. Through comprehensive benchmarking with different anomaly detection models at Visa, PersonaAgent is able to score transactions efficiently (a 1,000-transaction CPU batch in 1-2 ms), lifting the PR-AUC by 0.65-6.58% and the production business metric by 8.93-14.24% relatively in deployment.
memoryretrieval-augmentedagenticbenchmark - arxiv:2611.08093 · cs.LGGradient Boosting for Efficient Rank Discovery in Tensor Network Kernel MachinesAlbert Saiapin, Kim Batselier
Tensor network kernel machines (TNKMs) combine nonlinear features with low-rank tensor representations to model complex relationships efficiently. However, selecting an appropriate tensor rank remains challenging, as conventional rank-search methods require training multiple candidate models, increasing computational costs. In this work, we focus on a canonical polyadic (CP) tensor network parameterization and propose GB-CPKM, a gradient boosting framework that builds a higher-capacity predictor by sequentially adding fixed-rank learners. This approach provides a new perspective on rank selection by incrementally increasing model capacity, with ensemble size selected based on validation performance rather than through explicit rank search. Experiments on real-world regression benchmarks show that GB-CPKM trains up to 6x faster and uses up to 2x less peak memory than a conventional rank-selection method based on cross-validation, while achieving comparable predictive performance. These results highlight the potential of gradient boosting to improve the computational efficiency and scalability of tensor-network learning.
memorybenchmark - arxiv:2611.08092 · cs.LGGabor Gabor: E-Gaborformer, a Strictly Causal Speech Recognizer with Gabor-Dictionary Convolutions and Delta-Rule MemoryNir Krakowski, Tamir Cohen, Ohad Zohar, Chen Zarfati
Real-time voice agents need a speech recognizer that never waits for future audio and whose memory per call does not grow with the length of the call. We present E-Gaborformer, a strictly causal encoder that combines Gated DeltaNet-2 recurrent layers with sliding-window attention in a 3:1 ratio, and builds every temporal convolution from a fixed dictionary of causal Gabor atoms. Its streaming state is constant, about 19 MB per call at production scale including the prediction network, where a causal-attention encoder of the same size accumulates over 1 GB after ten minutes. In a controlled causal study on LibriSpeech, E-Gaborformer leaves the initial plateau earlier, reaches a given accuracy up to 18% sooner, and matches or beats an E-Branchformer trained with the identical front end and recipe. At production scale, a 0.6B-parameter Hebrew and English streaming transducer built on E-Gaborformer has 28 to 31% lower word error rate (WER) than NVIDIA's Nemotron 3.5 ASR Streaming, a model of the same size, on English voice-agent traffic, and up to 20% lower Hebrew WER than our previous production system. The same weights serve strictly causal and 480 ms-lookahead decoding with no loss in accuracy, and a two-class formulation of the RNN-T lattice trains the transducer 2.6x faster with exact loss and gradients.
memory - arxiv:2611.08085 · cs.CLLinguistic boundaries emerge from pressure-constrained processing dynamicsRui He, Philipp Homan, Wolfram Hinzen
Human cognition organizes continuous experience into bounded units while maintaining continuity across them. Linguistic elements are also organized into bounded structural units such as words and phrases, with sentence marking the final bound of a grammatical construction process. How do such boundaries emerge? We propose boundaries emerge from interacting demands to maintain unresolved dependencies, integrate completed relations, and update semantic context, with structural completion allowing but not determining closure. First, in a generative simulation without predefined boundary positions or constituent labels, these demands produced sentence-like segment-length distributions and distinct characteristic timescales, with structural completion allowing but not determining closure. Near boundaries, unresolved dependency load declined, integration cost peaked, and accumulated pressure reset. Then, we tested whether same demands predicted human listeners' boundary choices during comprehension. Listeners were more likely to mark boundaries with fewer unresolved dependencies, higher integration cost and semantic mismatch, and elevated preceding memory load, beyond syntactic and prosodic cues. Third, in naturalistic self-paced reading, sentence-final slowing coincided with declining memory load and increasing integration cost and semantic mismatch, broadly consistent with the simulation. Processing same narratives with a large language model revealed increased local intrinsic dimensionality and straighter hidden-state trajectories at sentence endings, followed by larger representational updates. Boundary explained additional variation in reading time and hidden-state geometry beyond processing demands, but considerably less in next-token surprisal. Together, these findings suggest linguistic boundaries emerge as discrete manifestations of transitions in continuous processing dynamics.
memory - arxiv:2611.08082 · cs.LGDictLLM: Post-training Compression of Large Language Model with Dictionary KernelsJinho Park, Se Young Chun, Mingoo Seok
Efficient deployment of large language models (LLMs) on mobile devices requires balancing performance with compute and memory constraints. Post-training compression is an effective approach to managing this trade-off, while requiring much less computation than in-training counterparts. We propose DictLLM, a dictionary-based compression scheme in which weight matrices within grouped layers are decomposed into a shared dictionary and layer-specific sparse coefficients, directly obtained from a pre-trained model without further training. Specifically, we cast DictLLM as a four-step post-training optimization. First, we obtain shared dictionaries via singular value decomposition (SVD) of input-scaled weight matrices, amortizing rank across grouped layers; second, we sparsify the layer-specific coefficients to the target compression ratio using a dictionary-aware Hessian; and third, we refine the dictionaries given the resulting sparse coefficients. Finally, we aggressively quantize both components using factor-specific Hessian information: input covariance for the dictionaries and the dictionary-aware Hessian for the coefficients. The resulting representation reduces memory footprint by storing a single shared dictionary per layer group, along with sparse layer-specific coefficients. On language modeling and zero-shot reasoning benchmarks, DictLLM consistently outperforms existing post-training low-rank compression methods and pushes the Pareto frontier between model size and performance. DictLLM compresses the original LLaMA-7B model by 5.1x with only a 2.9-point increase in WikiText-2 perplexity, achieving a 3.6x higher compression ratio than the state-of-the-art low-rank method.
memorypost-trainingbenchmark - arxiv:2611.08081 · cs.CVTerraMesh-Masks: Open-Vocabulary Segmentation for Earth ObservationBenedikt Blumenstiel, Hugues Devimeux, Johannes Jakubik, Konrad Schindler
Open-vocabulary segmentation (OVS) enables models to segment a variety of concepts specified in natural language, removing the need for fixed label sets and fine-tuning. While recent OVS approaches have shown strong performance on natural images, they transfer poorly to Earth observation (EO) data due to differing semantics, scales, sensing modality, and geographic context. To date, large-scale EO-specific datasets and benchmarks for OVS remain scarce. We, therefore, introduce TerraMesh-Masks, a large-scale open-vocabulary segmentation dataset for EO, comprising 16 million binary masks aligned with nine million multimodal TerraMesh samples and annotated using OpenStreetMap and Overture-derived semantics. In addition, we provide an expert-reviewed evaluation set of 523 samples covering over 200 classes to enable robust evaluation. To demonstrate the impact of TerraMesh-Masks, we build an open-vocabulary version of TerraMind. Our results demonstrate significantly improved performance over general-purpose state-of-the-art models reaching 35.18% IoU versus 16.29% for DINOv3.txt and 11.33% for SAM-3. We release the dataset publicly under a permissive license at https://huggingface.co/datasets/ibm-esa-geospatial/TerraMesh-Masks.
benchmark - arxiv:2611.08057 · physics.opticsHybrid integration of a SESAM with a silicon nitride photonic integrated circuitAmaryllis Verstraete, Tom Reep, Jan de Haan, Eero Koivusalo +4
The realization of high-performance femtosecond lasers on photonic integrated circuits (PICs) requires robust, low-loss coupling between a saturable absorber and a photonic waveguide. In this work, we demonstrate the hybrid integration of a semiconductor saturable absorber mirror (SESAM) with a silicon nitride (SiN) PIC by micro-assembly. This micro-assembly strategy transfers the non-linear functionality of the absorber onto the PIC. We characterize the linear and non-linear performance of an unbonded SESAM and of a SESAM bonded to inverse-tapered SiN waveguides. At 1560~nm, the bonded device shows a waveguide-coupled modulation depth of 5.2\%, a maximum reflectivity of 37.2\%, and a saturation fluence of 47~\textmu J/cm$^2$. This work demonstrates that SESAMs can be integrated directly onto silicon nitride-based PIC platforms and provides a pathway towards fully integrated, self-starting, chip-scale ultrafast lasers.
photonic integrated circuit - arxiv:2611.08056 · cs.ROTerrain-Indexed Motion Priors and Geometric Support for Extreme Humanoid ParkourChenghao Sun, Youhao Lyu, Zizheng Li, Shuting Zhang +2
We present a progressive imitation framework that learns terrain-appropriate humanoid parkour skills and subsequently adapts their hand--knee support. Human demonstrations provide traversal behaviors, but their contact patterns need not suit a robot's proportions or the available support surfaces. Our framework changes the role of demonstrations across three stages: privileged imitation produces physically executable robot references; terrain-indexed adversarial motion priors guide skill acquisition; and geometric support guidance with a reduced motion-prior weight enables contact adaptation. Candidate-based precontact attraction and actual-contact evaluation guide support without prescribing a contact sequence. The deployed policy selects behaviors from onboard depth, proprioception, and velocity commands without an external skill label. On the Noetix E1, the policy achieves 96.34\% hybrid success in simulation and transfers without real-world policy fine-tuning, completing 29 of 30 individual obstacle trials and four of five hybrid courses. Support-adaptation ablations show higher traversal success for the complete adaptation stage on the tested terrains, while qualitative observations show changes in contact placement. These results support acquiring skills from demonstrations before relaxing their influence to accommodate robot-specific support configurations.
humanoid - arxiv:2611.08054 · cs.CVNutriVisRAG: Multimodal Exemplar Grounding for Nutrient Estimation with Vision-Language ModelsLubnaa Abdur Rahman, Ioannis Papathanail, Maria Panagiotou, Nefeli Dimitropoulou +4
Diet is a major determinant of health, yet accurate dietary assessment from free-living meal images remains challenging. Existing retrieval-based Vision-Language Model (VLM) approaches to nutrient estimation typically rely on textual food representations, while, to our knowledge, direct grounding on visually similar annotated meals has not been investigated. We present NutriVisRAG, a retrieval-augmented generation framework that uses a vision encoder to retrieve similar meals with known nutritional annotations and provide them as multimodal in-context exemplars for VLM nutrient estimation at inference time, without task-specific fine-tuning. We evaluate NutriVisRAG across four datasets, four VLMs, three retrieval encoders, and two prompting strategies, including free-living meals from adults with insulin-treated diabetes, where carbohydrate estimates are compared against two dietitians' annotations. Across all 96 configurations (4 datasets x 4 VLMs x 3 retrieval encoders x 2 prompting strategies), NutriVisRAG reduced the primary estimation error in 56, with improvements more frequent under whole-meal (35/48) than per-item prompting (21/48). Across the evaluated configurations, the lowest NutriVisRAG error was below the best Vanilla error in 13 of 16 VLM--dataset pairs. For GPT-5.4-mini, under whole-meal prompting, the best-performing retrieval encoder for each dataset reduced the primary error on all four datasets relative to Vanilla, with reductions of 1.7--35.7%. Overall, our results show that grounding in multimodal exemplars comprising images and respective annotations can improve nutrient estimation at inference time without task-specific fine-tuning, although the benefits remain configuration-dependent.
retrieval-augmented - arxiv:2611.08039 · cs.ROHITCH: Teaching a Vision-Language Model to Operate a Humanoid through a Whole-Body HarnessQingzhou Lu, Chenxin Liu, Xinming Wang, Liu Cao +3
Enabling vision-language models (VLMs) to control humanoid robots requires both reliable whole-body interfaces and knowledge of how to use the body that carries over across tasks. We present HITCH (Harness Interfaces To Command Humanoids), a harness in which perception returns the 3D position of any point the VLM selects in its image, and execution turns commands for the pelvis, torso, hands, walking, and gaze into whole-body motion with a learned tracker. We find that both measuring the scene and executing commands accurately raise task success. We then fine-tune a VLM on trajectories in which teacher policies solve five base tasks through HITCH, so that it learns how to operate the humanoid. Without task-specific trajectories, the fine-tuned model completes 51% of SIMPLE episodes and 42.5% of HumanoidArena episodes, and adding five trajectories per SIMPLE task raises its SIMPLE success to 95%. On a real G1, it carries a box and walks to a goal without real training data.
humanoid - arxiv:2611.08016 · cs.LGLearning Dynamics of On-Policy DistillationGuan Zhe Hong, Sumeet Ramesh Motwani, Liang Chen, Yangchen Pan +4
On-policy distillation (OPD) is emerging as an important paradigm for language-model post-training, combining dense teacher supervision with training on student-generated trajectories. This creates a feedback loop where each update changes the student and thereby the training distribution that generates subsequent updates. Yet, how this feedback shapes learning and generalization remains poorly understood. We conduct a mechanistic analysis of OPD's learning dynamics, distinguishing the two feedback loop stages: conversion of the distillation signal into parameter updates, and broadcast through which these updates change predictions across prefixes and reshape the rollout distribution. This framework helps us understand both limited student responsiveness despite substantial teacher-student disagreement, and degeneration when updates that improve local teacher agreement also increase the probability of unsuccessful continuations, exposing the student to unreliable supervision in subsequent rollouts. Finally, we show that offline distillation can prepare a student for OPD without significantly improving immediate held-out accuracy. While OPD alone may degrade generation, it can yield substantial accuracy gains after sufficient offline preparation. Together, our results connect teacher supervision, student geometry, and the evolving rollout distribution in determining the learning dynamics of OPD and its effects on downstream performance.
post-training - arxiv:2611.08013 · cs.RODevelopment of Anthropomorphic Arm for collaborative and Home Service Robot CHARMIEHafiz Adil, Tiago Ribeiro, Antonio Ribeiro, Carolina Lopes +3
Service Robots are becoming a reality and are progressing by leaps and bounds. Alongside this avenue of robotics is also the field of prosthetics, which is evolving rapidly. These topics are hot cakes in the industry and are widely developed. Some have one quality, and some have others, but are making a shoreline towards success. The development of arm and hand as individual or as a package is becoming common as more and more advanced collaborative robots come to the surface and join their human counterparts in the real world. This dissertation presents an open-source 3D-printed arm which was assembled and programmed on two fronts. The first is the control of the hand by utilising teleoperation with a camera and ML-based hand pose estimation. Secondly, it is the control via kinematics utilising the Denavit-Hartenberg method to define the motion and to find the position of the end effector in 3D space.
teleoperation - arxiv:2611.08012 · cs.ROSymmetry-Aware Surrogate-Assisted Testing of UAVsAnas Shrinah, Chris Bennett, Kerstin Eder
We investigate a symmetry-aware surrogate-assisted testing approach for UAV autopilots. The proposed approach exploits geometric symmetries between flight missions by trans- forming obstacle configurations into a shared canonical coordi- nate frame. This enables a neural-network surrogate trained on a single reference mission to be reused across arbitrary missions, reducing the amount of high-fidelity simulation data required for training. The surrogate model is trained to predict the minimum clearance distance between a UAV and surrounding obstacles and is used to guide test generation towards safety-critical scenarios. We evaluate this approach using the Aerialist UAV test bench and benchmark its performance against techniques from the 2024 to 2026 UAV Testing Competitions. Results show that the approach achieves an average critical-test-case hit rate of 42% across the evaluated missions, compared with approximately 10% for the constrained-random baseline used to generate the training data. This approach also achieves high input diversity and competitive benchmark performance. However, our findings reveal limitations in transferring surrogate predictions across missions with differing flight-path characteristics, highlighting the boundaries of geometric canonicalization. To support reproducibility and future research, we provide the trained surrogate model and a dataset comprising more than 2,500 high-fidelity simulations.
benchmark - arxiv:2611.08005 · cs.CVDeformable Contact-Aware 3D Object PlacementSonghe Wang, David Miller, Charles A. Kamhoua
Current language-guided 3D object placement typically predicts a rigid 6-DoF pose, ignoring the physical reality that indoor placements are dictated by contact. Consequently, predicted poses often float, interpenetrate supports, or fail to deform soft surfaces. We propose Deformable Contact-Aware Placement (DCAP), a framework that reframes placement as a drop-to-equilibrium process. Rather than predicting a static pose, DCAP constructs a semantic release state that specifies the drop location, object scale, and deformable components, allowing a coupled rigid-deformable simulator to settle the object under gravity and friction. This explicitly separates semantic intent from physical realization. Evaluated on a benchmark of 2,063 prompt-asset-scene cases, DCAP significantly outperforms static baselines. On standard placements, it improves VLM Physics & Naturalness by 21.2 points and human preference by 26.5 points over the strongest raw rigid baseline, while reducing mean penetration depth by 72.3%. Compared to simulated rollout baselines, DCAP reduces residual speed by 72.1%. Ultimately, our results prove that physically accurate 3D placement requires modeling semantic intent and physical realization as strictly separate processes.
benchmark - arxiv:2611.08003 · cs.CVWhen GeoFMs Meet VHR: Benchmarking Foundation Models on Very High Resolution Earth Observation ImageryKshitij Raj Sharma, Pierre Adorni, Minh-Tan Pham, Dirk Tiede +1
Geospatial foundation models are a widely adopted starting point for remote sensing tasks. Most are pretrained on Sentinel-1/2 and Landsat imagery at a ground sampling distance (GSD) of 10m or coarser. Many applications instead rely on very high resolution (VHR) imagery, where individual buildings, roads, and amenities are resolved. We study which pretraining regimes transfer well to this scale. We introduce VHR-Bench, a benchmark suite of ten datasets at 1.5m GSD or finer spanning five task families, eight application domains, and both satellite and aerial platforms. We evaluate twelve encoders in three pretraining tiers with frozen backbones at two training-data fractions. We find that both pretraining on large natural-image collections and VHR-specific pretraining transfer effectively to VHR tasks, with complementary strengths. SkySense leads when labels are scarce; at the full label fraction the web-pretrained DINOv3 ViT-L edges ahead, with the two close together there. Encoders pretrained only on Sentinel-scale imagery transfer least well under RGB input. Guided by these complementary results, we continue the self-supervised pretraining of DINOv3 ViT-L on 1.02 million held-out VHR tiles and evaluate the resulting encoder without changing the downstream protocol. The continued model gives a clear improvement over the released backbone when labels are scarce and a modest one at the full fraction, at no increase in inference parameters or compute. These findings support a generalize-then-specialize strategy: already strong, broad natural-image pretraining can be further improved by a short VHR continuation stage while preserving its general visual features.
benchmark - arxiv:2611.07998 · cs.LGTowards Multiscale Molecular Representation Learning through Geometry-Based Fragmentation for Polymer Property PredictionMakoto Takamoto, Chaitanya Sampara, Viktor Zaverkin, Daniel Oñoro-Rubio +1
It is well established that molecular substructures can contribute unequally to chemical and physical properties, with specific motifs or side chains often playing dominant roles. Fragment-based representations have therefore emerged as an effective strategy for incorporating chemically meaningful inductive biases, as exemplified by molecular fingerprints and fragment-based attention mechanisms. However, many existing fragmentation-based machine learning methods rely on predefined chemical rules, which may yield suboptimal fragmentation for previously unseen or chemically unfamiliar structures and often overlook the importance of multiscale molecular representations. To address these limitations, we formulate molecular fragmentation as a radius-constrained optimization problem inspired by the k-center formulation, where the number of clusters is not fixed a priori but emerges from the specified radius, and propose a geometry-based fragmentation method that partitions arbitrary molecular or material structures using largely geometric information, without relying on predefined chemical fragmentation rules. The method provides flexible control over fragment size, enabling the construction of multiscale fragment representations for scale-adaptive modeling. Building on this framework, we develop a geometric fragment-based attention mechanism that integrates multiscale structural information and incorporate it into a multimodal machine learning model combining p-SMILES representations with three-dimensional (3D) conformational information. Experiments on a benchmark polymer dataset show that incorporating 3D conformational information improves prediction accuracy for several properties over models based exclusively on p-SMILES representations. Explainable artificial intelligence analyses further provide insights into the physical relevance of the learned representations.
benchmark - arxiv:2611.07996 · cs.MASEER: A Scalable Event-Driven Runtime for Parallel Execution of Multi-Agent Reinforcement Learning EnvironmentsJiming Su, Feng Zhu, Yuhao Xiao, Miao Zhang
Multi-agent reinforcement learning (MARL) is increasingly used in systems where many entities interact in a shared environment. In these workloads, policy inference can be batched on GPUs, but the CPU-side environment path still updates entity states, resolves local interactions, and builds observations inside each $\mathrm{env.step}$. This path can dominate latency in interaction-intensive workloads after policy batching. Existing vectorized environment interfaces and general event-driven runtimes do not directly address the fine-grained cross-entity work inside one bounded MARL step while preserving the step-level semantics expected by learning systems. We present SEER (Scalable Event-Driven Environment Runtime), a runtime for MARL environment execution under an explicit event-locality contract. SEER models each environment step as a bounded stream of timestamped events. If direct state updates are entity-local and cross-entity effects are represented as ordered future events, SEER can run independent events in parallel while preserving serial-equivalent boundary outputs. To make this contract practical, SEER uses a CAS-based dual-layer event queue: local event queues preserve per-entity causal order, while a parallel executive list exposes ready work to global worker threads. SEER further uses STPA, a preflight-guided thread-pool selection mechanism that combines short runtime traces with static workload and hardware features. We evaluate SEER on two parameterized MARL workloads: compute-intensive UAV swarm control (Aircraft) and communication-intensive wireless management (PCS). On Aircraft, SEER reduces inference latency by 16.2\% over the baseline. On PCS, SEER reduces environment-execution latency by 46.9\%, 19.5\%, and 64.9\% over ROSS-T, NBSE, and ROOT-Sim, respectively.
multi-agent - arxiv:2611.07994 · cs.CVReviving the Discarded High-Resolution Feature for Transformer-Based Tiny Object DetectionZehao Wang, Si Wu, Guiping Cao, Xin Li +1
Tiny object detection remains challenging because tiny instances occupy only a few pixels, and their fine-grained details are easily lost during feature downsampling. Although the highest-resolution feature preserves rich local details for tiny objects, DETR-like detectors often discard it due to its large computational cost, while directly reintroducing this feature also brings dense background noise. To address this problem, we propose ReDHF, a novel framework that Revives the Discarded High-resolution Feature for tiny object detection. ReDHF has three cooperative components. Cross-Scale Deformable Fusion (CSDF) first constructs a semantically enhanced highest-resolution feature and injects its fine-grained details into the shallow feature under a moderate cost. Ellipse-Guided Shallow Supervision (EGSS) then guides the enhanced feature to learn more reliable classification and localization priors by applying geometry-adaptive auxiliary supervision. Full-Scale Query Interaction (FSQI) further enables decoder queries to access both low-level details and high-level semantic information, allowing them to aggregate richer cues for tiny objects. Extensive experiments show that ReDHF consistently improves multiple baselines across different DETR variants and achieves state-of-the-art performance on the evaluated benchmarks. The best ReDHF variants surpass their baselines by +1.4 AP on AI-TOD-v2, +1.6 AP on VisDrone, and +1.5 AP on SODA-D, with especially notable gains on tiny and small objects.
benchmark - arxiv:2611.07988 · cs.CLCompression Through a Context Lens: Evaluating Adaptation, Quantization, and Distillation in Arabic Legal LLMsMd. Rafiul Biswas, George Mikros, Wajdi Zaghouani
Model compression can facilitate the deployment of large language models in resource-constrained legal settings, but its effects on domain-specific performance may depend on the evaluation protocol. We investigate this issue using ArabLegalEval, a Saudi Arabic legal benchmark, under both with-context and no-context evaluation. The former provides supporting legal passages, while the latter assesses performance without external context. Removing the context produces accuracy gaps of up to 27 percentage points for the base model. We examine three stages: domain adaptation, quantization, and knowledge distillation across FANAR-9B, JAIS-8B, and Qwen2.5-7B. Four-bit quantization reduces model memory by 64-71%, with changes in no-context MCQ accuracy ranging from -3.2 to +1.6 percentage points. Distillation produces more variable results, with accuracy generally declining as student size decreases. A Qwen2.5-7B student distilled from FANAR achieves 76.6% no-context accuracy, compared with 77.2% for its teacher. Our findings suggest that no-context evaluation provides a useful complement to context-supported testing when assessing the effects of adaptation and compression on Arabic legal language models.
memorybenchmarkevaluation protocol - arxiv:2611.07976 · cs.LGQRAE: Query-Relevant Attention Entropy for Context Compression in DeepSearch AgentsYingjie Tan, Haiquan Qiu, Quanming Yao, Manyi Zhong +2
DeepSearch LLM agents repeatedly search and carry retrieved sources into subsequent reasoning steps. This expanding context increases token consumption and latency, although many results do not contribute to later reasoning. We propose Query-Relevant Attention Entropy (QRAE), an online context compression method that uses sparse evidence-assessment heads to decide which intact retrieved chunks should persist across reasoning steps. It scores bounded candidate batches through paired prefills: a query-free control suppresses shared attention responses, while normalized entropy measures the breadth of the positive response within each chunk. High-scoring chunks enter a persistent evidence whitelist and are included in the reasoning context. Uncertain chunks remain outside the reasoning context but are retained in a candidate cache, where they can be retrieved for later reevaluation. QRAE requires neither an additional generative compression pass nor modification of the attention operator. Across multiple DeepSearch benchmarks and reasoning models, QRAE improves task success by an average of 2.35 percentage points and reduces end-to-end latency and reported token usage by an average of 30% and 45%, respectively. These benefits generalize to a model with nearly 100B parameters.
context compressionllm agentbenchmark - arxiv:2611.07964 · cs.ROTime-Indexed Skill Success Models for Latency-Aware Skill Selection and Timing in Dynamic ManipulationMohsen Omidi, Maria Chiara Leva, Brian Vaughan
Robots that manipulate moving objects act on observations that are outdated when the chosen skill starts, and the best skill can change within a few hundred milliseconds. Skill competence models, however, predict success for a launch at the moment of observation. We propose time-indexed skill success models, which predict whole-task success when a skill is launched after a given delay. Trained on rollouts branched at a grid of launch delays, they let a receding-horizon planner select both the skill and its launch time under fixed or random latency. In a MuJoCo study in which a UR5e captures a sliding puck with three skills, the planner exceeded a static competence model by 2.8 to 6.2 percentage points at 100 to 300 ms of latency, including when that model was evaluated on point or sampled forecasts of the launch state. Across five complete runs (three training seeds and two distribution shifts), all 80 paired 95% intervals against four baselines excluded zero. On delayed launches, the time-indexed model was calibrated (expected calibration error 0.009 against 0.239), and pausing the world during latency overstated success by up to 23 points. On a real UR5e that captures a rolling ball after 0.2 s of injected latency, a static model trained in a hardware-matched simulator predicted 99% success on a prespecified subset of 35 delayed launches that it rated at least 0.8. Of these, 63% succeeded (95% interval 46 to 77%), and the time-indexed model predicted 56%. In this single-skill, predictable regime, a matched physics-based forecast was also well calibrated.
manipulation - arxiv:2611.07962 · cs.AITCSExam: Evaluating Frontier LLMs on the Full Range of TCS ResearchYifan Luo, Zhengying Liu, Yang Yuan, Zhilin Yang +1
Can large language models (LLMs) advance theoretical computer science and resolve its long-standing open conjectures? To study progress toward this goal, we introduce TCSExam, a unified framework and benchmark covering the full range of TCS research. It evaluates theorem proving, solver engineering, and constructions alongside direct attempts to resolve open conjectures, capturing both concrete research accomplishments and progress toward long-standing research goals. TCSExam-Full contains 100 expert-reviewed tasks, with submissions checked by the Lean kernel or an executable task checker. TCSExam-Lite selects 20 tasks, five per track, reducing the number of task attempts by 80% at the same per-attempt budget. We evaluate eight agents on Full and nine on Lite. The TCSExam Index combines all four tracks: GPT-6 Astra leads Full (71.9), and Opus 5.5 leads Lite (72.3). Full and Lite rankings are strongly aligned across eight shared agents (Spearman's $ρ=0.93$), supporting Lite as a compact, discriminative benchmark. The evaluations yield improved solver outputs and mathematical constructions, together with independently verified special cases and bounds from open-conjecture attempts. TCSExam provides an extensible framework with task sources, Docker build recipes, and evaluation tools for reproducible use.
benchmark - arxiv:2611.07959 · cs.LGCoRD: Coupled Rollouts Fill the Blind Spot of On-Policy DistillationGuhan Chen, Bohan Li, Jun Rao, Hejin Wang +7
On-policy distillation trains a student on its own rollouts, so states the teacher would reach but the student does not are seldom supervised. We introduce CoRD (coupled-rollout distillation), which keeps the on-policy loss and moves only the distribution of training states toward the teacher, drawing rollouts from a geometric interpolation of student and teacher through the maximal coupling of speculative decoding. The coupling keeps the student's token wherever the two distributions agree and otherwise substitutes a token from the part of the target the student underweights, so interventions are as rare as any exact sampler allows and recede as the student learns. Across six base students from two model families (Qwen3 and Qwen3.5, 0.6B to 4B), CoRD raises average pass@1 over on-policy distillation by 0.9 to 2.0 points and pass@8 by 1.2 to 3.4 points on five public math benchmarks at equal steps and data. On frozen teacher-reached states, a 4B student trained with CoRD reaches in 16 steps a gap closure that on-policy distillation does not reach in 80, and substitution rules intervening at least as often in other directions fall short of CoRD.
benchmark - arxiv:2611.07955 · cs.CVRotation Is Token-Blind: A Structural Criterion for Post-Training Quantization of Feed-Forward Geometry TransformersHuanfei He, Bohao Zhang, Jiaxin Zhang, Wei Sui +3
Feed-forward 3D models use Transformers to predict camera poses, depth, and structure in one pass, avoiding iterative multi-view optimization. Long sequences and deep attention incur high compute and memory costs, motivating Post-Training Quantization (PTQ) for deployment. Existing PTQ methods mainly suppress channel-wise outliers by smoothing or orthogonal rotation. Geometry Transformers introduce a different challenge: heterogeneous token types share one activation quantization grid. A few special tokens may exert high leverage over global outputs and be updated by only a subset of blocks. Token-blind transforms, which apply the same transformation to every token, reduce overall quantization error but cannot remove differentiated degradation across token types. We propose QMapAnything, a structure-aware PTQ framework. We cast floating-point-preserving rotation and smoothing as a token-blind reparameterization group, and prove that they yield common-mode gains while leaving relative token SNR unchanged. We introduce a statistic $Γ$ that combines a special token's quantization disadvantage and output leverage to predict whether differentiated bias correction is needed beyond rotation. Experiments on geometric benchmarks show that W4A8 closely matches full-precision performance. Under W8A8, peak memory falls by $2.2$--$3.1\times$ versus FP32, with up to $3.1\times$ speedup on real hardware. These results provide a principled and efficient approach to deploying feed-forward 3D reconstruction models in resource-constrained edges.
memorypost-trainingbenchmark - arxiv:2611.07950 · cs.LGLeakage- and Label-Aware Benchmarking of Class-Imbalance Strategies for Multimodal Chest X-Ray and EHR ClassificationDang-Khoa Bui, Quang-Huy Tran, Minh-Khoi Nguyen-Bui, Duc-Tuan Ngo +3
Benchmarks of class-imbalance strategies for multimodal chest X-ray (CXR) and electronic health record (EHR) prediction are often evaluated on record-level splits of cohorts in which one patient contributes several radiographs. We audit such a benchmark on 29,083 MIMIC-CXR radiographs of 10,479 patients with eight diagnosis targets. First, the record-split benchmark was misleading: a lookup table that copies the label of a development-set radiograph with the same EHR vector nearly matched the best model, and on unseen patients every model fell to the level of predicting every case positive. Second, tracing the labels back to MIMIC-IV explained why: each patient's labels are keyword matches on the diagnosis codes of a single admission, copied, with one EHR vector per patient, to all of the patient's radiographs, so about a third of the radiographs carry labels that differ from the codes of their own admission. Third, under patient-separated evaluation with admission-aligned and rule-refined labels, the initial conclusions do not hold: in the evaluated configurations no imbalance strategy raised the observed mean macro F2 over no augmentation by more than 0.01, tuned tree ensembles showed no consistent advantage over tuned logistic regression, and the released EHR variables stayed at the level of the all-positive baseline, whereas the diagnostic history available at imaging time and image embeddings exceeded it, including, with rule-refined labels, a CheXpert-trained encoder that never saw these radiographs; RAD-DINO had seen 96.6% of them during self-supervised pretraining. Adding either encoder to the diagnostic history gave no statistically significant F2 gain; exploratory AUROC gains were significant with RAD-DINO but not with DenseNet. We provide reusable diagnostics, a leakage-aware protocol and a checklist.
benchmark - arxiv:2611.07949 · cs.CLBeyond the Keyword: A Human-Annotated Benchmark for Arabic Anxiety-Related DiscourseShimaa Ibrahim, Mabrouka Bessghaier, Md. Rafiul Biswas, Insaf Elouard +1
Keyword retrieval returns expressions of personal anxiety together with many other uses of anxiety-related language. We present a benchmark of 1,000 Arabic Twitter/X posts retrieved with the keyword qalaq ("anxiety/worry") and annotated for six discourse functions. Two native Arabic-speaking annotators labeled every post independently, reaching 51.1% raw agreement (Cohen's $κ= 0.378$). We then reviewed all 489 disagreements and all 511 agreements under the same guidelines. The review changed 120 of the agreed labels (23.5%), which shows that agreements need auditing as well as disagreements. The final labels form a consensus reference; on a random sample of 200 posts, an additional annotator agreed with it on 60.5% of posts ($κ= 0.517$). Against this reference, Fanar-C-2-27B reaches 79.8% accuracy and macro-F1 0.784 ($κ= 0.749$). Its errors concentrate in the residual category (41.1% of errors; precision 0.905, recall 0.447), whose definition is internally inconsistent in the executed prompt, so the low recall cannot be attributed to the model alone. Sociopolitical Anxiety (27.1%) outnumbers Personal Emotional Expression (22.0%); these are assignment proportions in the sample, not estimates of prevalence or retrieval precision. We release the guidelines, a residual-subtype scheme, a fixed data split, label provenance, and evaluation code.
benchmark - arxiv:2611.07948 · cs.AICirce: Training-Free LLM Self-Evaluation via Component Interaction ConsistencyShuyi Miao, Zheng Li, Xu Xie, Linrui Xu +3
As large language models (LLMs) are increasingly deployed in real-world applications, evaluating the correctness of their responses is crucial for reliable use. However, such evaluation is challenging because reference answers are rarely available in practice. Existing reference-free methods mainly rely on output signals, which may assign high scores to incorrect responses and thus do not necessarily reflect correctness. This motivates exploring the model's internal computation for complementary correctness signals. Recent studies exploit internal states, but they mainly analyze hidden-state trajectories across layers, which record only the combined updates of attention and feed-forward networks (FFNs) and overlook how these components interact. We therefore propose Circe, a training-free self-evaluation method that assesses a response using both internal and external signals, without reference answers or repeated sampling. For internal signals, we first introduce component interaction consistency (CIC), which measures the directional alignment between attention updates in earlier layers and attention-induced FFN output changes in later layers. We further incorporate representation geometry (RG) to characterize normalized displacement and angular changes along layerwise hidden-state trajectories. For the external signal, we convert the mean negative log-likelihood of generated tokens into a bounded likelihood confidence (LC) score. Finally, we fuse CIC, RG, and LC into a unified correctness score. Experiments on five LLMs and six benchmarks across four domains show that our method distinguishes correct from incorrect responses more effectively than existing baselines, and generalizes to multilingual, multimodal, and open-world settings.
benchmark - arxiv:2611.07947 · cs.AIChemE-as-Code: An Open-Source Platform for Agentic AI in Chemical EngineeringArtur M. Schweidtmann
ChemE-as-Code provides an open-source environment in which AI agents built on large language models (LLMs) can carry out chemical engineering work and engineers can check it. Its calculation tools range from catalysis through process simulation to plant layout, HAZOP and life cycle assessment, and skills tell a general-purpose agent how to use them. Engineering files are linked, machine-readable and versioned in Git; calculations run outside the LLM, each value records its origin, and automated checks report inconsistencies and out-of-date results by file and line. In an illustrative ammonia case, agents developed a written task into a preliminary basic-engineering package (simulation files, P&IDs, layout, HAZOP, datasheets, relief) for USD 224 of API cost in the design sessions. Planted faults and engineering changes show that the checks find what their rules cover and miss the rest, so they guide human review rather than replace it. No independent engineer has reviewed the design.
agentai agentagentic - arxiv:2611.07945 · cs.AItok: a textual concrete syntax of BPMN and DMN for humans and AI agentsMarek Deja
tok is a plain-text notation for Business Process Model and Notation (BPMN) 2.0.2 process models and Decision Model and Notation (DMN) 1.5 decisions. One engine checks a tok document against both standards, compiles the document to XML with an automatic diagram, and serves language-model applications through the Model Context Protocol. The paper classifies the notation as an alternative concrete syntax of the BPMN and DMN metamodels and reports one study. A tok model takes 2.2 times fewer tokens than the smallest BPMN XML of 893 files exported by 37 tools and 7.7 times fewer than the files as exported. The checker signals 98.2% of 1,411 injected faults, against 85.8% for the XML Schema with bpmnlint. Four language models new to tok write valid tok from a one-page reference for 85.7 to 100% of 87 process descriptions from two datasets and valid BPMN XML for every description, and the strongest model writes tok with 25% and 19% fewer tokens than BPMN XML. As agents with a lean skill and one checker tool, the models write valid tok for 97.6 to 100% of the descriptions, the strongest model spends 21% and 14% fewer tokens on tok than on BPMN XML, and a skill which sends the model to read the reference through a tool costs 1.7 to 2.8 times the tokens of the lean skill. Of 47 event logs simulated from public models, 44 fit their source models, and tok imports every model discovered from the logs.
ai agent - arxiv:2611.07941 · cs.CVIDPrior: Decoupling Identity Generation from Facial Image SynthesisFadi Boutros, Meriç Demirörs, Darian Tomašević, Vitomir Štruc +1
Large-scale identity-labeled face datasets are essential for training modern face recognition models, yet collecting and annotating such data is costly and raises privacy and ethical concerns. Existing synthetic face generation methods typically derive identities from predefined class labels, unconditional image generation, or predefined latent sampling strategies, without explicitly modeling the underlying distribution of facial identities. In this paper, we propose IDPrior, a two-stage framework that decouples identity generation from image synthesis. We first learn a diffusion prior in the embedding space of a pretrained face recognition model to model the distribution of facial identities and generate novel identity embeddings. These embeddings are then used to condition a pretrained identity-conditioned diffusion model (DM), enabling the synthesis of multiple images of the same identity. This modular design provides a scalable approach to identity generation while remaining compatible with existing and future identity-conditioned DMs. Experiments indicate that IDPrior generates diverse and well-separated identities while maintaining a closer correspondence to the distribution of authentic identities. Face recognition models trained on datasets generated using IDPrior-generated identities achieve competitive or improved performance compared with existing synthetic-data-based methods across multiple face verification benchmarks. Code and generated datasets are publicly available at https://github.com/fdbtrs/IDperturb.
benchmark - arxiv:2611.07940 · cs.AICardioLM: A multilingual suite of small language models for the cardiology domainManuela-Daniela Danu, Miguel Rodríguez-Ortega, Alberto Becerra-Tome, Pablo Arancibia +36
Background: Cardiovascular care generates large volumes of clinical free-text documentation in local languages. However, clinical natural language processing (NLP) resources remain concentrated in English, and no existing benchmark is cardiology-specific, multilingual, and broad in entity coverage. Methods: We present two contributions. The first is CardioCCC, a corpus of 508 cardiology case reports originally published in Spanish and extended to Czech, Dutch, English, Italian, Romanian, and Swedish through machine translation, annotation projection, and clinician review, with annotations covering diseases, symptoms, procedures, and medications. The second is CardioLM, comprising CardioBERTa, a set of language-specific encoders obtained by continued pretraining on cardiology and biomedical corpora, and CardioNER, a set of named entity recognition (NER) models fine-tuned on CardioCCC. Results: Cardiology-specific pretraining reduced pseudo-perplexity by up to a factor of eight, yet changed strict span-level F1 by at most 0.04, with benefits concentrated in languages whose backbones lacked biomedical coverage. The jointly trained multilingual XLM-RoBERTa-large models achieved the best strict F1 in nearly every language and entity type, exceeding the strongest monolingual system by up to 0.13 F1. A large share of residual errors concerned boundary detection: relaxed matching recovered 0.10-0.23 F1 for disease, procedure, and symptom, but only 0.01-0.07 for medication. Clinicians rated CardioNER well above a rule- and dictionary-based baseline, with the mean rating across entity types rising from 26% to 73%. Conclusions: Specialty-specific pretraining is not the decisive factor for downstream entity recognition, whereas multilingual joint training yields the strongest results in nearly all settings and a simpler deployment path for multi-country clinical information extraction.
benchmark - arxiv:2611.07939 · cs.CVWAM-Sim0.1: World Anchored Memory Meets Online RL for Consistent and Realistic Driving SimulationJiahao Cui, Jiaye Li, Wenhao Zhang, Yaoyi Li +6
Autoregressive driving world models systematically suffer from structural drift, identity drift, and quality decay over long rollout horizons due to transient conditioning signals that lack persistent state. We present WAM-Sim, a driving world model anchored by a persistent hierarchical memory and optimized via online flow-matching RL. WAM-Sim structures persistent context across three temporal and semantic scales: pose-aligned structural tokens encoding persistent road topology, read-only scene tokens supplying geolocated background appearance, and write-once instance tokens preserving individual traffic participant identities across occlusions and chunk boundaries. To restore visual sharpness and physical plausibility degraded by multi-scale conditioning, we introduce an online post-training framework that optimizes self-generated rollouts via flow-matching RL guided by hierarchy-aligned visual rewards and a vision-language model physical judge. The invariant memory empirically confines policy optimization within the anchored context, helping suppress semantic collapse and reward hacking, while online RL systematically enforces physical commonsense and fine-grained visual fidelity. On the nuScenes benchmark, WAM-Sim achieves a superior generation fidelity (FID=3.47, FVD=25.72) and strong downstream perception performance (24.16 mAP, 71.23 road mIoU). On minute-scale rollouts of self-collected data, WAM-Sim substantially mitigates structural and identity drift, preserving high scene similarity and subject identity coherence across extended horizons. Project page: https://fudan-generative-vision.github.io/WAM-Sim.
world modelmemorypersistent statepost-trainingbenchmark - arxiv:2611.07935 · cs.CVTowards Scalable Diagnostic-Driven Whole-Slide Image Retrieval for Pathology ArchivesJingsong Liu, Youssef Wally Han Li, Weiwei Zhou, Jing Dai +2
The growing adoption of digital pathology has rapidly expanded collections of hematoxylin-and-eosin whole-slide images, yet efficient navigation through these archives remains challenging, as metadata is often inconsistent or unavailable. Content-based image retrieval addresses this challenge by identifying similar cases directly from visual representations instead of manual annotations or metadata. Existing retrieval systems have demonstrated promising performance but largely rely on handcrafted indexing pipelines and predefined sampling strategies, constraining scalability and generalization. Here, we present CLEAR-WSI (Constant-Length Embedding for Automatic Retrieval of Whole-Slide Images) as a scalable, diagnostic-driven slide-retrieval framework, validated on both public and clinical data. Using a fixed-length slide embedding, CLEAR-WSI refines retrieved candidates with either a predicted diagnostic label or the ground-truth label, depending on the retrieval setting. Newly processed cases are subsequently added to the expanding memory bank, allowing the retrieval archive to grow over time. On a comprehensive benchmark consisting of two public and one private clinical datasets representing four retrieval tasks, CLEAR-WSI improves Precision@5 over the strongest evaluated baseline, with gains on CAMELYON16 (3.10), BRACS-1 (2.75), BRACS-2 (1.95), and X-HER2 (3.75). Additional ablation and latency studies further examine the contribution of each component and highlight its potential for scalable deployment.
memorybenchmark - arxiv:2611.07934 · cs.LGThe Blindfolded Controller: Decision Model-Guided Training for Reinforcement LearningMuhammad Ibrahim Khan, Aniket Dixit, Faizan Ahmed, James Brusey
Reinforcement learning (RL) provides a framework for continuous improvement through interaction with an environment. However, RL agents struggle against tasks that offer exploration challenges and sparse rewards, because learning efficient policies from scratch is slow and sample-inefficient. In contrast, decision models (DMs), given a description of the environment and task, can perform well in sparse-reward tasks, but, as they do not learn from reward, they are not able to improve towards the optimal solution. In this paper, we present Decision Model-Guided Training (DMGT), a technique that uses a DM with a description of the environment and task to initialise an RL policy, combining the DM's ability to act without training with RL's ability to improve from reward. Instead of exploring from scratch, the agent reaches the goal and receives reward from its first episodes, and RL then improves its policy beyond what the DM achieves. We demonstrate through experiments on seven tasks that our approach outperforms RL in all seven and DM in five. In addition, we provide comprehensive guidelines, derived from ablation studies, for describing an environment and task to a DM.
agent - arxiv:2611.07932 · cs.LGAttnSR: Attention Step Reward for Long-Horizon Agentic RLMengsong Wu, Tianchen Zhou, Zhaochen Su, Zhenhua Liu +2
Long-horizon LLM agents make multi-step decisions, yet outcome-reward-only reinforcement learning provides little direct supervision for individual steps. We introduce AttnSR, a step-level reward augmentation method that derives reward allocations from policy attention extracted during training replay, without a separately trained reward model or additional rollouts. AttnSR traces backward through the policy model attention graph to estimate the relevance of earlier steps to the final response. The terminal step, measured relative to a running baseline, determines whether these rewards reinforce or penalize the selected steps. In a controlled Qwen2.5-7B-Instruct study, AttnSR consistently outperforms the PPO baseline across all evaluated checkpoints on ALFWorld, raising final held-out success from 91.2% to 97.1%. The ablation experiment shows that the benefit depends on how intermediate rewards are assigned, rather than simple accumulation of step rewards over time. AttnSR also improves ScienceWorld at every checkpoint while cutting environment interactions by 29.9%. These results support policy attention as a practical prior for step-level reward allocation in long-horizon agentic RL. Code is available at https://github.com/fairyshine/AttnSR .
llm agentagentic - arxiv:2611.07928 · cs.AIZeroRAG: Adaptive Semantic Peer Routing for Decentralized Retrieval-Augmented GenerationPetru Neague, Quinten Stokkink, Martijn de Vos, Rafael Pires +3
Retrieval-augmented generation (RAG) is a standard approach to enhance the reliability of large language model (LLM) outputs by retrieving documents relevant to a query and supplying them in the LLM prompt. Today's RAG systems collect, index, and serve these documents from centralized stores. In practice, however, documents are owned by organizations, devices, and individuals who cannot or will not surrender custody of them to a central party, e.g., because of privacy, copyright or regulatory compliance. Existing modern decentralized retrieval techniques either reintroduce a coordinating authority, or rely on a global data structure that requires continuous network-wide coordination (e.g., DHTs). We present ZeroRAG, a decentralized routing layer for document retrieval in RAG that performs semantic query routing across an overlay of peers with- out global data structures or centralized orchestration. Documents remain with the peers that own them, and each peer keeps only lightweight local state: a bounded set of neighbor advertisements and a cache of recently routed queries. Using these, ZeroRAG steers each query through a best-first walk that progressively rewires the overlay toward semantically relevant neighbors, so recall improves over time as the network self-organizes. We evaluate ZeroRAG using three real-world workloads and two decentralized baselines. We find that ZeroRAG outperforms baselines in terms of recall and required network messages, approaches the recall and LLM answer accuracy of a centralized document store, self-heals after catastrophic peer failures, and scales to 100 000 peers.
retrieval-augmentedrag - arxiv:2611.07922 · cs.CVStitchEdit: Stitching Depth Priors into Image Editors via Per-Layer Gradient ProbingQin Guo, Dan Xu
Text-guided image editing with diffusion models struggles with spatial edits (e.g. manipulating objects, changing viewpoints, manipulating occlusions) that depend on the 3D structure understanding of a scene. Marigold and follow-up work demonstrate that a pretrained diffusion backbone already harbors strong geometric priors that can be activated for depth estimation, yet how to redirect these priors toward editing remains unresolved. Through per-layer gradient probing, we discover that the editing and depth objectives are compatible in early transformer layers but actively conflict in late layers, explaining why naive depth-conditioned editing either ignores geometry or degrades visual quality. We present StitchEdit, which first activates the backbone's geometric capability via a lightweight depth-denoising adapter, then selectively fine-tunes this adapter for editing using a probe-guided per-layer learning rate schedule that preserves depth knowledge where it helps and prunes it where it conflicts. On the proposed StitchEditBench, a benchmark of 150 geometry-dependent edits spanning spatial transformation, occlusion manipulation, and viewpoint change, StitchEdit consistently surpasses both instruction-based and depth-conditioned baselines on perceptual quality and geometric fidelity.
manipulationbenchmark - arxiv:2611.07920 · cs.AIHow Finite Brains Generalize Without Bound: Spatial Memory as a Universal ComputerYuhang He, Si Wu
A child who learns to add small numbers adds numbers of any length. LSTMs and Transformers trained on short sequences fail on longer ones. We trace the difference to what a learner reads when it decides. A position, a count or a time keeps taking new values as the input grows: it is a coordinate, and a learner that reads one meets in every long input values its experience never covered. Whether two things are the same has the same two answers at every size: it is a coincidence, and a learner that reads only coincidences finds in a long input nothing that short inputs lacked, so finite experience fixes its rule for every size. The memory carries the size of the input and the rule stays finite, and this determines the memory a generalizing brain can have: a memory that grows while what the learner reads of it stays finite has to look the same from every place, so it is made of cycles, and the past is reached by asking what happened when the cycles last stood where they stand now. The cycles are the grid modules, the cells that detect their return the place cells, the records the hippocampal index; a prefrontal controller moves the cycles and never sees them. Reading only coincidences costs no power: what finite experience fixes in such a learner is what Turing machines compute in step with their input, which is why a circuit that evolved to find the way home can learn arithmetic and grammar. In a network of neurons this machine of cycles is the dynamics seen through its symmetry, and symbols are its invariants. Trained on strings up to 40 symbols, the machine is at least 99% correct, on average over lengths 41 to 500, on nine of fifteen benchmark tasks, five of them tasks that no LSTM or Transformer solves; given new cycles after training it reverses strings of 15,015 symbols without error and fails beyond that on the curve the theory predicts; and its controllers run unchanged on neurons.
memorybenchmark - arxiv:2611.07912 · cs.ROGuidedWAM: Training-Free Energy Guidance for World Action Models in Autonomous DrivingXiyan Su, Finn Rasmus Schäfer, Hojun Lim, Frank Diermeyer +2
World action models (WAMs) for autonomous driving couple the driving policy and the visual world model in a single generative network whose objectives are fixed post-training. Conditioning on a new driving objective requires retraining. We study energy guidance as a general mechanism and present GuidedWAM, a model-agnostic framework for injecting driving objectives into frozen WAMs at sampling time. Our framework constructs each objective as differentiable energies on the sampler's intermediate clean-action estimate, which needs only forward access to the frozen model, and is stabilized by sampling-time guidance scheduling, per-energy normalization, and velocity-bounded correction. On the NAVSIM v2 hard-split, guidance improves all three frozen, publicly released WAMs by 1.8 to 3.6 EPDMS, averaged over three sampler seeds, and adds less than 10% to the sampling time per step. These results position energy guidance as a practical post-training control channel for WAMs.
world modelpost-training - arxiv:2611.07902 · cs.ROXM4D: 4D Reconstruction through Certifiable Geometric OptimizationHaoyu Han, Jack Benarroch Jedlicki, Jiahui Lei, Heng Yang
Accurate 4D reconstruction remains challenging: jointly estimating camera motion, scene geometry, and non-rigid deformation entails highly nonconvex optimization, making global optimality difficult to guarantee. We present XM4D, a framework for accurate and efficient 4D reconstruction built upon certifiably globally optimal geometric optimization. By separating the estimation of camera motion from non-rigid deformation, we formulate both stages as structured quadratically constrained quadratic programs (QCQP) that admit semidefinite programming (SDP) relaxations and verifiable global optimality certificates. For camera motion estimation, we lift static point tracks into 3D using foundational depth prediction models and formulate a scaled bundle adjustment problem that jointly estimates camera poses, depth scales, and static landmarks. By analytically eliminating translations and landmarks, we reduce the problem to a QCQP over scaled rotations. For non-rigid reconstruction, we represent dynamic motion using an embedded deformation graph (EDG) and show that its tracking, rigidity, and temporal regularization objectives similarly reduce to a rotation-only QCQP. Both problems admit SDP relaxations that we solve efficiently with a GPU-accelerated Burer-Monteiro solver based on Riemannian optimization. The recovered geometry further initializes dynamic Gaussian splatting for photometric refinement. Extensive experiments across seven benchmarks demonstrate accurate camera motion estimation on both static and dynamic scenes, efficient optimization that scales to thousands of frames, and state-of-the-art novel-view rendering quality.
benchmark - arxiv:2611.07900 · cs.ROStructure from Collision via Hypothesis-Space ReasoningTakuhiro Kaneko
Knowledge of the hidden interior structure of an object, including the presence and location of cavities, is essential for predictive simulation, robotic manipulation, and nondestructive inspection. However, such cavities leave no direct signature on the static visible surface, making collision-induced deformation one of the few cues used to infer their presence. Structure from Collision (SfC) addresses this problem by recovering the hidden interior structure of an object from multiview observations of a collision event. SfC is challenging for two reasons: hidden cavities are only indirectly observable through surface motion, and voids correspond to the absence of material, making direct optimization using standard density representations difficult. The density-based SfC baseline inherits both difficulties. Surface-matching gradients do not directly indicate the cavity location, while density gradients can be weak or uninformative inside or near void regions. Therefore, we reformulate SfC as hypothesis-space reasoning, representing hidden cavities as explicit structural hypotheses whose variables encode available structural assumptions rather than relying on implicit density updates. A global stage performs gradient-free optimization over this hypothesis space using forward-only material point method (MPM) simulation. A local stage then refines the selected hypothesis as a Lagrangian particle system through gradient-based optimization using differentiable MPM simulation. On the SfC benchmark, our method improves hidden-structure reconstruction and future prediction over the density-based SfC baseline. Interior-topology stress tests and robustness analyses characterize its scope and limitations. The project page is available at https://www.kecl.ntt.co.jp/people/kaneko.takuhiro/projects/sfc-hypothesis/.
manipulationbenchmark - arxiv:2611.07899 · cs.AIPersistent Retrieval for Conversational Agents: A Study of Retrieval EfficiencyKarthik Ravula, Sathwik Kowtha, Rishabh Verma, Kala Mehak Jain +1
Conversational agents can retrieve information directly or delegate retrieval to a subagent. This design choice affects how retrieval history is maintained, what context each model call receives, and what information must be re-retrieved in follow-up questions. We compare three designs: skill-based retrieval in the main agent, a non-persistent retrieval subagent that generates a findings response, and our persistent retrieval subagent, which maintains retrieval history across turns and returns retrieved evidence directly to the main agent. Our approach also allows the main agent to request corrections from the same subagent. We evaluate the three designs on enterprise questions posed independently, within shared conversations, and as dependent follow-ups. Across five multi-turn settings, our approach uses 43.3% fewer recorded agent-model tokens and incurs 45.0% lower estimated API cost than skill-based retrieval, while reducing average response time by 9.1%. Compared with non-persistent retrieval, it reduces response time by 15.6 seconds per question, while using 65.1% more tokens and incurring 29.1% higher estimated API cost. Our approach shifts the retrieval subagent from producing synthesized responses to maintaining retrieval state and supplying evidence that the main agent can directly evaluate and use for answer generation.
agent - arxiv:2611.07896 · cs.ROAffordJEV: Affordance Grounding by Confirming Candidate RegionsJincheng Li, Yutian Zhang, Jianyu Zhang, Xinshun Wang +1
Affordance grounding asks which visible part of an object performs a requested operation, a basic capability for embodied manipulation and human--robot interaction. Most existing methods follow a System Two paradigm: a multimodal language model first writes a chain of thought and then turns its conclusion into a box or points for a segmenter. This incurs considerable inference cost, gives the semantic decision little direct control over the extent of the final region, and makes mistakes hard to correct. Yet locating an affordance is essentially a perceptual judgment that need not be argued out in words. We therefore propose, in which frozen segmenters propose candidate regions and a System One model decides directly which of them perform the operation, generating no text and reading each decision from a single forward pass. Our contributions are twofold. (1) An inference framework that confirms candidate regions level by level, from object to part to sub-part, with evidence passed between the levels in both directions so that the levels correct one another. (2) A System One model trained for this framework: existing System One models are built for general-purpose decisions, judge one candidate per request and are ill-suited to the framework's region questions, so ours learns directly from the requests the framework itself produces and judges all candidates of a level at once. Trained only on InstructPart and evaluated with definition-aligned operation words, outperforms chain-of-thought methods in domain and in zero-shot transfer to UMD and IIT-AFF, is better overall than existing System One models run in the same framework, and needs fewer judge requests. To our knowledge, is the first approach to bring System One models to affordance grounding.
embodiedmanipulation - arxiv:2611.07894 · physics.opticsA Passive Intelligent Metasurface: A Network ModelTristan Lawrie, Gregor Tanner, Luke Bennetts
Intelligent metasurfaces enable advanced wave manipulation, including negative refraction, focusing, beam splitting and beam steering, with significant potential for practical applications. However, conventional approaches often rely on active metamaterials with gain, or demand computationally intensive solutions of inverse problems to find the optimal configuration. In this work, we use the scattering framework of quantum graph theory to derive an exact, analytic expression for configuration of a passive metasurface that produces a prescribed reflected wave field. The metasurface terminates a periodic lattice of waveguides and consists of an array of channels, each a tube of length $\ell_m$ ending in a vertex whose scattering parameter $λ_m$ sets how much of the wave is lost to the wall. For any prescribed incident and reflected field, the exact lengths and scattering parameters follow channel by channel, in closed form, from the ratio of the wave amplitudes returning to and leaving the lattice, and the metasurface is passive whenever this ratio does not exceed unity in modulus. We demonstrate the construction for both a one-dimensional and a two-dimensional metasurface. The one-dimensional metasurface terminates a two-dimensional lattice and steers, splits or perfectly absorbs Gaussian beams. The two-dimensional metasurface terminates a three-dimensional lattice and generates a prescribed image in a plane inside the lattice. In each case, we obtain the exact boundary properties that produce the effect.
manipulation - arxiv:2611.07891 · cs.AIWhen Bigger Becomes Brittle: Why Compression Hurts More as Language Models ScaleGautam Gupta, Dikshant Kukreja, Unique Kumar, Avni Mittal +6
Quantization and pruning reduce the cost of deploying language models, but their effects are usually reported as benchmark scores. Such scores record only the loss that remains at the output. Inside the model, compression damages some components more than others, and the rest of the network can partly compensate. The observed loss therefore reflects both the damage and this compensation, which benchmark scores cannot separate. How the two combine, and whether their balance changes with model size, remains unclear. We study this compensation for in-context retrieval, a capability carried by a small, identifiable set of attention heads. To separate the two, the expected loss from ablating these heads is decomposed exactly into the part restored by the rest of the network, attributed to the components responsible, and the part that reaches the output. This decomposition is then compared across model sizes. For damage of comparable size, smaller models restore most of the loss, whereas larger models retain most of it, because the compensating responses offset less of it as models grow. Compressing the heads produces the same pattern. Compression that is nearly harmless in small models can therefore be costly in large ones.
benchmark - arxiv:2611.07886 · cs.LGWhen, Where, How: Active In-Context Learning for Tabular Foundation ModelsChristian M. M. Frey
Tabular Foundation Models (TFMs) adapt to a new task only through the labeled rows in their context. However, labels are expensive to acquire in real-world scenarios due to the need for highly specific domain knowledge to ensure high-quality data. Active in-context learning transfers classical acquisition rules to this setting, in which the model that guides acquisition changes with every label added. We ask how the value of diversity, predictive uncertainty and target-aware information gain changes as a frozen TFM acquires more context. For that, we study Target-Aware Nearest Neighbor (TANN) acquisition and the Expected Predictive Information Gain in Context (EPIC) and evaluate them against classical rules. Our study across twenty benchmarks with two well-established TFMs and classical GBDTs highlights three findings: (i) When: diversity-based acquisition is best during the first acquisition iterations; (ii) Where: labeling pool rows that serve the target inputs via a simple target-aware nearest-neighbor rule matters substantially; (iii) How: EPIC yields the best models at the label budget, retains on average 80--87% of the chance-corrected full-data performance with only 5-10% of the pool labeled on the eight benchmarks with the largest pools, and selects labels that also improve the predictive performance of gradient-boosted decision trees.
benchmark - arxiv:2611.07883 · cs.CVInnerLogic-Surg: An Open Surgical Foundation Model via Continual PretrainingKun Yuan, Hao Ding, Yiqing Shen, Gasper Podobnik +4
Vision foundation models are the backbone of modern surgical image analysis, yet most surgical systems still rely on encoders pretrained on natural images or on specialist models trained from scratch on comparatively small datasets. Pretraining encoders on large amounts of surgery-specific data has the potential to generally enhance the performance of surgical image analysis methods simply by providing stronger backbones. Here we present InnerLogic-Surg, an open family of surgical vision encoders (21M--300M parameters) obtained by continually pretraining a transformer-based image encoder on millions of surgical images curated from 11,965 internet-sourced surgical videos spanning ~2,900 hours and more than fifteen procedure types. Our training recipe inherits DINOv3's pretraining stage to preserve representation ability after surgical adaptation, and additionally applies a learned feature-upsampling module for high-resolution dense prediction at low token cost. Across a benchmark covering workflow, instrument and anatomy recognition, segmentation, detection, and monocular depth, under the same architecture, the largest InnerLogic-Surg model achieves the best linear-probe performance in all 20 surgical tasks. We demonstrate that surgical continual pretraining yields transferable representations that generalize across procedures, anatomical regions, and task types, showing our surgical image encoders as a promising and scalable alternative to the procedure-specific representation learning approaches that have traditionally dominated surgical computer vision.
benchmark - arxiv:2611.07881 · cs.ROTactileLab: Sim-to-Real Dexterous Robot Manipulation with Efficient Shear-Sensitive Tactile SimulationYijiong Lin, Wenhao He, Toby Jack, Max Yang +2
Robotic simulators that represent shear within tactile sensing are becoming increasingly important for contact-rich manipulation. However, simulating contact shear with high-fidelity deformable or soft-body simulation is computationally demanding, limiting scalability to large-scale reinforcement learning. Here we introduce \textit{TactileLab}, a scalable multimodal tactile simulation framework built on IsaacLab, with a particular focus on a new optical-flow-based tactile representation termed \textit{simulated tactile flow}. Instead of relying on soft-body physics to model shear deformation, TactileLab uses rigid-body simulation to efficiently estimate dense tangential motion at the contact interface, producing tactile flow fields that encode shear-relevant contact dynamics. This flow field complements geometric features from a contact depth field. We further introduce a novel policy architecture for jointly processing long-horizon proprioceptive signals and rich tactile observations. Across multiple contact-rich tasks, policies trained using simulated tactile flow learn efficiently at scale and exhibit emergent contact-aware behaviours. To facilitate sim-to-real policy deployment, we then propose a novel real-to-sim tactile translation method that maps real tactile image streams to simulated tactile flow and depth fields, together with a new tactile data collection method, and validate it across diverse tactile sensor designs. Using this pipeline, we demonstrate real-world deployment on tasks encompassing dual-arm and dexterous in-hand manipulation. Overall, TactileLab provides an end-to-end sim-to-real framework for scalable tactile policy learning, enabling diverse shear-sensitive manipulation across heterogeneous tactile robotic systems.
manipulationdexteroustactilesim-to-real - arxiv:2611.07879 · cs.AINetwork Effect Decomposition for Causal Effect Estimation under Network InterferenceZhirong Huang, Wenzhen Zhang, Guangquan Lu, Boyan Chen +2
Estimating causal effects from observational network data requires accounting for both individual treatment and neighborhood exposure. Existing outcome models commonly incorporate these intervention variables into a joint predictor and estimate causal effects through prediction contrasts, leaving the corresponding response components implicit. We propose the $\underline{N}$etwork $\underline{E}$ffect $\underline{D}$ecomposition $\underline{Net}$work (NEDNet), which parameterizes the conditional outcome response as a query-invariant pre-treatment baseline plus separate treatment- and exposure-response pathways. Path-specific neighborhood contextualization supplies local pre-treatment information to the two pathways. We characterize the fitted response through a unique reference-normalized decomposition, which is equivalent to a zero mixed treatment--exposure contrast. Experiments on semi-synthetic benchmarks derived from BlogCatalog and Flickr show that NEDNet achieves the lowest PEHE among the compared methods for individual, peer, and total effects across homogeneous and covariate-heterogeneous response settings.
benchmark - arxiv:2611.07876 · cs.LGWhich Signals of Code Correctness Survive Distribution Shift? Hidden-State and Causal Predictors of Program and Verdict CorrectnessViola Campos, Thomas Klassert, Adrian Ulges
Large language models increasingly judge code, yet without external feedback such judgments are unreliable, and verifying them by executing tests or by extended reasoning is costly. Instead, execution-free correctness predictors use signals from the model's output probabilities, sampled solutions, or hidden states to identify judgments that warrant further verification. However, these predictors may degrade under distribution shifts when generalizing across projects and languages. We compare six execution-free predictors from four families on Qwen3-Coder-30B and gpt-oss-20b under benchmark shift between LiveCodeBench and BigCodeBench and language shift across Python, C++ and Java. We evaluate two targets: whether the model's own verdict on a program can be trusted, and whether the program itself is correct. Besides established approaches, we adapt two novel causal localization methods to code, which have been shown to be more robust out-of-distribution on natural language tasks. For verdict correctness, causal localization improves robustness to language shift. For Qwen3-Coder a probe on the causally localized subspace stays 0.09 AUROC ahead of output confidence under language shift (0.11 in distribution), significantly outperforming the best non-causal predictor by 0.04 AUROC on average. This advantage, however, does not generalize consistently to gpt-oss or to benchmark shift. For program correctness, predictors transfer more successfully between benchmarks, largely by identifying problems on which the model tends to fail. However, their ability to rank candidates within the same problem is limited.
benchmark - arxiv:2611.07873 · cs.CVPVDepth: Panoramic Video Depth Estimation via Geometry-Aware Spatiotemporal AdaptationChuanxin Song, Peixi Peng
Panoramic video depth estimation is pivotal for applications such as Virtual Reality and World Models. However, advancements in this field are impeded by two primary obstacles: the scarcity of large-scale training data and the unique spatiotemporal challenges of Equirectangular Projection (ERP), which hinder the direct transfer of perspective models. In this paper, we first present PanoCARLA, a large-scale synthetic RGB-D panoramic video dataset, featuring natural motion trajectories and drone-like roaming perspectives. Building on this foundation, we propose PVDepth, an end-to-end framework adapted from perspective video depth models. To tackle ERP-specific geometric distortions and consequent non-linear temporal dynamics, we introduce two core mechanisms: (1) A Progressive Sphere-aware Noise Initialization strategy that anneals the noise distribution from planar to spherical, guiding the model to adapt to non-uniform information density; and (2) A Cube-rectified Temporal Modeling module that incorporates an auxiliary cubemap temporal branch to rectify non-linear temporal dynamics in the ERP domain. Extensive experiments demonstrate that PVDepth achieves superior performance, generating geometrically accurate and temporally consistent depth sequences. Code and data will be released at https:// github.com/ChuanxinSong/PVDepth.
world model - arxiv:2611.07870 · cs.AIFARR: Forecast-Aware Residual Refinement with Matched Endpoint Evidence for Long-Horizon Time-Series ForecastingYiyang Zhang, Yuli Zheng, Bin Li, Meimei Zhang +1
Large language models use auxiliary conditioning to incorporate additional information into time-series forecasts. However, auxiliary conditioning does not consistently improve accuracy and can increase forecasting error when its influence is poorly matched to the observed temporal context. We propose Forecast-Aware Residual Refinement (FARR), which learns to correct forecasts using endpoint evidence. FARR obtains a valid forecast under the standard condition and a null forecast under a predefined null condition. Their difference reveals how conditioning changes the prediction. A shared residual refiner combines this endpoint evidence with historical features to estimate the remaining error of the valid forecast. Horizon-wise gating then controls the contribution of the predicted residual to the final output. Experiments on four long-horizon forecasting benchmarks show that FARR achieves the best dataset-average MSE and MAE among the compared methods. Code will be released at https://github.com/blossomrichzyy-star/farr-forecast-aware-residual-refinement.
benchmark - arxiv:2611.07867 · cs.LGPlayGen2: Learning Real-Time Playable Worlds with Explicit 3D StateJihai Zhang, Yuhan Zhou, Junyou Li, Deheng Ye +2
Building a playable world requires more than generating plausible video: it demands responsive control, persistent spatial structure, and coherent interactions with dynamic entities. Using generated frames as the sole recurrent world representation ties these requirements to view-dependent appearance synthesis, making it difficult to maintain long-term spatial and temporal consistency in generation and interaction. We present PlayGen2, a real-time interactive world model that separates learned 3D state evolution from neural rendering. PlayGen2 maintains a persistent world memory comprising static geometry represented by a volumetric occupancy grid and identity-indexed 3D states for the player and other entities. Within a fixed-size local region, its world-state model jointly learns action-conditioned entity dynamics, geometry completion, and entity discovery, committing updates to a persistent world memory. A separate neural renderer synthesizes first-person observations from depth and semantic projections of the world state, without feeding generated images back into state prediction. Using Doom as a testbed, we demonstrate that the model successfully learns world dynamics including geometry-aware player motion and player--NPC interactions. The framework is computationally efficient, enabling the complete simulation and rendering pipeline to achieve approximately 160ms backend latency on a MacBook Pro with an M4 Pro chip. Beyond single-player interaction, we demonstrate zero-shot deployment on unseen Doom maps, multiplayer interaction through the composition of single-player models, and programmable gameplay through rule-based state interventions. These capabilities demonstrate the transferability and extensibility of the learned game-world model. Code and pretrained model weights are publicly available.
world modelaction-conditionedmemory - arxiv:2611.07866 · cs.CLDo Recommenders Need to Think in Words? Latent Reasoning for Efficient LLM-Based RecommendationHongyang Su, Beibei Kong, Lei Cheng, Chengxiang Zhuo +1
LLM-based recommender systems increasingly generate chain-of-thought (CoT) text before each item prediction to activate the LLM's reasoning capacity, costing tens to hundreds of extra tokens per recommendation at inference. We ask whether inference-time reasoning artifacts (textual or latent) can be replaced by training-time supervision alone, without loss of accuracy. We propose LatentRec, a framework for LLM-based generative recommendation that replaces inference-time reasoning artifacts with training-time per-step head supervision. LatentRec has three components. A 2.5M-parameter Continuous Thought Module iterates K latent reasoning steps over the LLM's hidden state during training. Soft Token Selection bridges the ~14x norm gap between hidden states and input embeddings. A Per-step Reasoning Loss (PRL) supervises every intermediate state through the LLM's prediction head against the first token of the target Semantic ID. The trained LLM retains this accuracy without the module at inference, so the module can be discarded at deployment. On three Amazon benchmarks, LatentRec improves HR@10 by 4.0-5.0% (relative) over a CoT-distilled baseline at 18.3x lower latency (batch=1). Per-step training supervision consistently improves accuracy across three datasets, paid once at training rather than at every recommendation.
benchmark - arxiv:2611.07865 · cs.CVOcTreeMamba: Topology-Aware Mamba for 3D Medical Segmentation via Octree SparsificationKaixiang Mai, Jiaqi Yu, Yuzhu Ji, Yiqun Zhang +3
Accurate segmentation of 3D anatomical structures is crucial to medical image analysis. Recently, Mamba-based approaches have achieved promising segmentation performance in capturing long-range dependencies with linear-time computational complexity. However, heuristic selective scanning strategies inherently disrupt the intrinsic spatial hierarchy and increase computational redundancy, ignoring the inherent sparsity of 3D anatomical structures. To this end, we propose OcTreeMamba, a hierarchical framework that incorporates Octree structures into SSM, achieving 3D topology-aware long-range context feature propagation and integration. Specifically, a Structure Learner performs anatomy-adaptive structured sparsification through hierarchical gating and Gumbel-Softmax sampling, enabling differentiable selection of anatomical regions. The selected regions are subsequently organized by a Dynamic Octree Builder into a multi-level topology that preserves their hierarchical connectivity. Based on the constructed Octree, a bidirectional Tree-SSM propagates and integrates contextual features along connectivity-aware leaf--parent--root and root--parent--leaf trajectories. Extensive experiments on four benchmark datasets show that our proposed OcTreeMamba achieves competitive performance \wrt topological continuity and computational efficiency, particularly for recovering the complex structure on vascular datasets. The code is available at https://github.com/MikeMak123/OcTreeMamba.
benchmark - arxiv:2611.07859 · cs.AICORE: Coarse-to-Fine Search with Context Management for Deep ResearchHoang Tran Vuong, Ngoc Quan Pham, Hung Bui, Fatih Porikli +1
Deep research agents answer complex questions through repeated search. They rely on a harness that runs their tool calls and returns evidence to their context. This creates a trade-off: Search works best over long passages, which give enough context for reliable semantic matching. But the returned evidence should be short, because every observation stays in the agent's context for the rest of the trajectory. We introduce CORE, a training-free harness that combines coarse-to-fine search with context management. CORE retrieves passages with a dense retriever. It then reuses the token-level similarities from a late-interaction reranker to return only the most relevant snippet from each passage. Context management lets the agent expand the context around sources it has already retrieved and suppresses repeated snippets. In controlled closed-search comparisons with the Tongyi DeepResearch and OpenResearcher harnesses, CORE achieves state-of-the-art answer accuracy and evidence recall across three agent models and three benchmarks (BrowseComp-Plus, BrowseComp-Wikipedia and BrowseComp-Plus-CM), on corpora containing up to 553M documents.
agentbenchmark - arxiv:2611.07854 · cs.CLTowards Foundation Decision Models: DrexNischay Dhankhar, Alisher Amantay, Zhanibek Datbayev, Darkhan Akhmetov +5
Autoregressive approaches with large language models, such as long chain-of-thought reasoning, have become the default for new AI applications because they reach higher accuracy on a wide range of tasks. In production, however, they give inconsistent answers, require parsing of outputs, and add cost and latency. Decision models address this: given a question and a set of options supplied by the user, they return a probability distribution over the options without autoregressive decoding. We present Drex, a family of four open decision models with 8B (DLM), 9B (v1.5), 35B with 3B active (v2.0), and 27B (Max) parameters. Each request consists of a state (a document, record, or agent context, optionally with images) and questions about it, each asking for a choice among options, a yes/no judgment, or a score on an ordered scale. Training combines supervised fine-tuning with reinforcement learning from adjusted feedback (RLAF), plus vision tuning for Max. Ablations show that RLAF improves accuracy and that consistency losses improve robustness to option order. On Decision Index 0.4.1, Max sets a new state of the art among the decision models we compare (67.52 on the public suite vs. 60.70 for Jev 1.14.0). On Decision Bench v4, it is statistically tied with d3-turbo and GPT-6 Sol and trails Claude Opus 5.5 by 1.6 points, and it has the highest accuracy among decision models on DecideBench. Max answers a typical request in 183 ms, against about 2.0 s for GPT-6 Sol and 3.5 s for Claude Opus 5.5, and its outputs are stable under option reordering. All four models are open source: https://github.com/nace-ai/drex-decision-models
agent - arxiv:2611.07850 · cs.ROLearning Robotic World Models with Semantic and Visual FeedbackHao Wu, Yuqi Li, Yizhou Zhao, Kun Wang +6
Robot world models predict action-conditioned future visual states to support robotic planning and decision making. However, visual reconstruction does not fully capture task progress, semantic evaluation alone provides limited constraints on fine-grained visual errors, and homogeneous predictions limit comparative feedback. We present RoboImagine, a framework for post-training autoregressive robotic world models through candidate-future comparisons with semantic and visual feedback. Diverse search promotes diversity among candidate trajectories under shared visual observations, action sequences, and task instructions. A frozen multimodal evaluator provides task-relevant semantic preferences through pairwise video comparisons in both presentation orders, complemented by pixel- and perceptual-level visual consistency feedback that constrains local structural and appearance errors. These signals drive policy optimization through group-relative advantages. Fixed initial context and branch-specific rolling histories support long-sequence generation with bounded context and consistent conditioning between candidate generation and policy optimization. Experiments with separately trained models on RT-1, BridgeV2, and RoboCasa cover real-world and simulated manipulation. On RT-1, under matched initialization, update counts, and evaluation protocols, RoboImagine improves PSNR by 1.26 dB and reduces LPIPS by 11.93% relative to continued supervised fine-tuning. Ablations show that diverse search improves prediction quality and that combined semantic and visual feedback yields higher external multimodal evaluation scores than visual feedback alone while maintaining comparable visual fidelity. Rolling histories further improve long-horizon prediction. These results support improving visual fidelity, task relevance, and long-horizon prediction through candidate comparison and complementary feedback.
manipulationworld modelaction-conditionedpost-trainingevaluatorevaluation protocol - arxiv:2611.07843 · cs.LGEnhanced MRI brain tumor detection using deep learning in conjunction with explainable AI SHAP based diverse and multi feature analysisAsif Rahman, Maqsood Hayat, Nadeem Iqbal, Fawaz Khaled Alarfaj +2
Recent innovations in medical imaging have markedly improved brain tumor identification, surpassing conventional diagnostic approaches that suffer from low resolution, radiation exposure, and limited contrast. Magnetic Resonance Imaging (MRI) is pivotal in precise and accurate tumor characterization owing to its high-resolution, non-invasive nature. This study investigates the synergy among multiple feature representation schemes such as local Binary Patterns (LBP), Gabor filters, Discrete Wavelet Transform, Fast Fourier Transform, Convolutional Neural Networks (CNN), and Gray-Level Run Length Matrix alongside five learning algorithms namely: k-nearest Neighbor, Random Forest, Support Vector Classifier (SVC), and probabilistic neural network (PNN), and CNN. Empirical findings indicate that LBP in conjunction with SVC and CNN obtained high specificity and accuracy, rendering it a promising method for MRI-based tumor diagnosis. Further to investigate the contribution of LBP, Statistical analysis chi-square and p-value tests are used to confirm the significant impact of LBP feature space for identification of brain Tumor. In addition, The SHAP analysis was used to identify the most important features in classification. In a small dataset, CNN obtained 97.8% accuracy while SVC yielded 98.06% accuracy. In subsequent analysis, a large benchmark dataset is also utilized to evaluate the performance of learning algorithms in order to investigate the generalization power of the proposed model. CNN achieves the highest accuracy of 98.9%, followed by SVC at 96.7%. These results highlight CNN's effectiveness in automated, high-precision tumor diagnosis. Code: https://github.com/asifrahman557/BrainTumorDetection
benchmark - arxiv:2611.07827 · cs.AIInstantiating the Knowledge PlaneFrancisco Pereira, Justine Sherry, Pratiksha Thaker
In 2003, Clark et al. proposed the "Knowledge Plane", a layer of intelligent agents that share rich information about network state in order to collaboratively manage the Internet across administrative boundaries. The technology of the day lacked the two capabilities it demanded most: reasoning and unstructured communication. Thus, automated network management has remained largely confined to single domains, while interdomain problems are still resolved by human operators on the phone. We argue that agentic large language models can finally unlock the potential of the Knowledge Plane, but that doing so will require solving new research problems that arise particularly in this decentralized, multi-agent, coopetitive setting that have received limited attention in the literature thus far.
multi-agentagentic - arxiv:2611.07826 · cs.LGEstimating Spectrum-Free Information Bounds for Generative Surrogates Beyond Conditional CollapseChaitanya Dumbre, Sourabh Dube
Conditional generative surrogates in high-energy physics, such as models that predict reconstructed jets from partons, are often evaluated on $1\text{D}$ marginal divergences among other metrics. Taken alone these metrics can reward conditional collapse. On a CMS simulation benchmark, a sampler that ignores the conditioner beats a trained Mixture Density Network (MDN) on all four target marginals under both KL and Wasserstein-1, and cannot be stably separated from a Conditional Flow Matching (CFM) model on two of the four. We instead score the models with the Continuous Ranked Probability Skill Score (CRPSS) from forecast verification, taking the exact marginal as the reference. We show how global skill depends on the evaluation spectrum, and we also report it within bins of a conditioning variable, where that dependence largely goes away. We then introduce a data-driven $k$-Nearest-Neighbor ($k$-NN) extrapolation that estimates the best CRPS the conditioning information allows. Both models agree with this estimate on all four targets, which suggests that current architectures have reached the information available in the conditioner, at least as far as these per-target scores can tell.
benchmark - arxiv:2611.07825 · cs.ROGenerate-then-Rank: Learning Symbolic Goal Interpretation for Embodied AgentsJisen Lei, Huizhi Liang
Goal interpretation is to translate a natural language instruction into a formal symbolic goal that a planner can execute. It is critical for embodied agents because any error here propagates to every downstream planning and action step. Doing it well requires both generating suitable candidates and reliably selecting among them. We present a two-stage generate-then-rank approach to Goal Interpretation in the Embodied Agent Interface (EAI). First, supervised fine-tuning adapts a Large Language Model (LLM) to EAI's output convention. Second, Candidate-Output Ranking from Errors (CORE) learns preferences over complete outputs from the model's own generated alternatives. We conducted experiments on the household scenes of the Embodied Agent Interface Benchmark. Multiple open source LLMs including DeepSeek-V3 chat, DeepSeek-Reasoner, and Qwen2.5-7B and Qwen2.5-72B were evaluated. The proposed method outperformed the baseline models including LLMs with simple prompting, chain-of-thought, and knowledge graph based approaches.
embodiedknowledge graphagentembodied agentbenchmark - arxiv:2611.07820 · cs.AIAxis-LoRA: Probing and Adapting Embedding Models for Personal-Memory RetrievalDohoon Kim, Junho Park, Hyunwoong Bae, Seyoon Kim +1
As LLMs are increasingly asked questions about a user's personal experiences, retrieval must distinguish memories by time, place, person, and event. Current personal-memory pipelines rely on frozen dense embedding models to embed memories for retrieval, but we find that these models do not capture the four axes well. Swapping a memory's date or place barely shifts its embedding, and a trained projection on top of the frozen embedding does not recover the missing signal. The problem lives inside the model, where each axis is encoded by a distinct attention-head pattern. Guided by these findings, we propose Axis-LoRA, a per-axis adapter that places four LoRA branches inside the embedding model and combines them via a per-query gate. Across three standard embedding models evaluated on LongMemEval, ATM-Bench, and LoCoMo, Axis-LoRA improves 4-benchmark mean R@10 by $+1.28$ to $+1.56$~pp while training only a small fraction of the base parameters on a small axis-counterfactual training set. Code is available at https://github.com/dohoonkim-ai/axis-lora.
benchmark - arxiv:2611.07812 · cs.CLDifferent Corpora, Different Skills? A Paired Study of Small Language ModelsEdoardo Vaira, Georgios Chavales, Pierre Nugues
Small language models are often trained under tight data budgets, where the choice of pretraining corpus may matter as much as model size or training recipe. We compare two contrasting corpora at matched 10-million- and 37-million-word budgets: the 1911 Encyclopaedia Britannica and the BabyLM 2026 strict pretraining corpus. For each corpus, we train teacher models, distilled 58-million-parameter Baby-Llama students, and no-distillation student ablations, using the same model families, training recipes, and evaluation protocol across corpora. Across BLiMP, BLiMP Supplement, EWoK, Entity Tracking, and COMPS, the corpora produce different skill profiles rather than a single quality ranking. BabyLM 2026 is consistently stronger on linguistic acceptability benchmarks, while Britannica is often stronger on Entity Tracking, a synthetic state-tracking task. The contrast appears across teacher and student models and at both budgets, with the cleanest cross-seed replication in the distilled student comparison. Targeted controls show that the central contrast persists after matching training exposure and at a longer diagnostic context; direct teacher ensembling also does not reproduce the 10M Britannica student's Entity Tracking behavior. Entity Tracking magnitudes nevertheless remain sensitive to training dynamics. For this corpus pair, corpus choice can therefore produce different relative skill profiles at small scale, motivating multi-benchmark evaluation when comparing pretraining data.
benchmarkevaluation protocol - arxiv:2611.07811 · cs.CVMV-Sign: Multi-View 3D Motion Integration and Enhanced Text-to-Motion Conditioning for Sign Language ProductionSingyu Muk, Kaustubh Kundu, Kunyuan Xie, Weijie Wang +1
Sign language production (SLP) aims to generate appropriate sign language motion from spoken language text. However, current 3D SLP methods suffer from noisy supervision and weak text conditioning. Specifically, these methods typically rely on pseudo-3D motion data fitted from single-view frontal videos, which introduces severe self-occlusion and loses fine-grained hand details, ultimately degrading downstream production. To overcome these limitations, we propose a comprehensive framework that jointly integrates multi-view 3D motion data and optimizes the text-conditioning pathway. First, we introduce How2Sign-MV3D, a high-fidelity dataset constructed via a novel cross-view curation pipeline that fuses multi-view estimates to resolve occlusions and filter non-signing segments. Second, we design a robust SLP model incorporating a sign-aware spatio-temporal VQ-VAE (ST-VQVAE) for joint body-hand tokenization, a pseudo-gloss adapter, and keyframe-aware retrieval to enhance text-to-motion alignment. Evaluated on the How2Sign benchmark, our framework reduces body and hand DTW-PA-JPE by 32.8% and 20.9% respectively compared to the state-of-the-art. Finally, we release two new high-quality 3D datasets, ASLCitizen-3D and MM-WAuslan-MV3D, alongside SGNify-MV as a multi-view video benchmark to support broader community research.
benchmark - arxiv:2611.07788 · eess.SYGeometry-Evolving Barrier Functions for Safety Evaluation in Dynamic EnvironmentsZhi Zhang, Francesco Belardinelli, Alessandro Abate
Autonomous systems operating in dynamic environments often involve multiple interacting agents, giving rise to high-dimensional safety evaluation problems. Hamilton--Jacobi (HJ) reachability provides rigorous finite-horizon safety guarantees, but solving a global HJ partial differential equation (PDE) over the entire state space leads to prohibitive computational complexity. In this paper, we introduce Geometry-Evolving Barrier Functions (GEBFs), a novel representation for dynamic safety evaluation that propagates local safety geometry of the solution to a terminal-value HJ PDE. By evolving the value, gradient, and curvature information along a Hamiltonian-induced reference trajectory, GEBFs provide a compact representation of the local HJ safety landscape. Under suitable local regularity assumptions, we establish consistency between the GEBF and the value, gradient, and Hessian of the HJ solution along an HJ-optimal characteristic. Building on these theoretical results, we develop an iterative algorithm that couples local geometric propagation with trajectory refinement. We further establish a trajectory-wise one-sided relationship between the GEBF and HJ value function, together with a fixed-point consistency result for the iterative scheme. Experiments on widely adopted agent--obstacle collision-avoidance benchmarks, across low- and high-dimensional systems, demonstrate that GEBFs accurately capture the evolution of dynamic safety boundaries, while achieving substantially lower computational cost than conventional HJ reachability.
benchmark - arxiv:2611.07784 · cs.LGEnergyBench: a live, gate-aligned benchmark of day-ahead forecasts for European power marketsEvgenii Genov, Ander Martinez Alonso, Christoph Bergmeir
Energy forecasting models are usually compared on backtests and fixed public datasets. Backtests can use data that did not yet exist when the forecast would have been made, and a fixed test set, reused across studies and possibly contained in the pretraining data of foundation models, invites overfitting to the benchmark. EnergyBench evaluates every entry live instead. Each day at 11:30 CET/CEST, thirty minutes before the day-ahead (SDAC) gate closes, every model forecasts the next day's price, wind and solar generation and load on the 15-minute grid for six bidding zones (BE, NL, FR, DE-LU, DK1, DK2). Unlike existing live benchmarks, EnergyBench runs the models itself, on one information set that it freezes at the cut-off and publishes, including three numerical weather prediction models, so that foundation models can be compared with and without the same weather covariates under the actual market deadline. Errors are pooled per series against a seasonal-naive baseline, and models are ranked by a Bradley-Terry rating on the Elo scale with a day-bootstrap interval. The field spans baselines, linear and gradient-boosted models on the frozen weather, eighteen time-series foundation models from eleven families, ensembles, and the grid operators' own forecasts. Over the first thirty-two live days, the covariates reduced the error of each of the ten foundation models run with and without them by 37-51% on matched races (64-73% on wind, 48-61% on solar). Used zero-shot, these models match or outperform a LightGBM model trained on the same covariates. Model size, from 3M to 331M parameters, has a much smaller effect, and median ensembles of the covariate models rank highest. The grid operators' pre-gate forecasts remain ahead on wind and solar, where the best single foundation models have 16-22% more error; on load the gap is 4-6%, within the intervals.
benchmark - arxiv:2611.07783 · cs.LGPBPO: Prefix-Block Policy Optimization for Diffusion Language ModelsKatherine Tieu, Dongqi Fu, Si Zhang, Hanqing Zeng +5
Diffusion Large Language Models (dLLMs) commonly decode responses by proceeding autoregressively across blocks while iteratively denoising multiple tokens within each block. This structure makes block-level policy optimization a natural adaptation of autoregressive policy optimization. However, the exact likelihood of a completed block is intractable because it marginalizes over the intermediate states and reveal orders of the denoising process. To realize the block-level formulation, we construct a prefix-block state and use a single model evaluation to define a factorized one-step surrogate for the block likelihood. The resulting block surrogate ratio is a product of its token surrogate ratios, allowing variation at individual positions to propagate across the entire block before clipping. We therefore introduce Prefix-Block Policy Optimization (PBPO), which preserves block-level conditioning but applies surrogate importance weighting and clipping separately to each token. PBPO further diversifies repeated updates by resampling corrupted views of the prompt and revealed prefix, a design motivated by masked-dLLM pre-training. Under our analysis assumptions, block-product weighting amplifies the variance of gradient contributions relative to token-local weighting. Across two LLaDA-family initializations, six benchmarks, and three generation lengths, PBPO achieves competitive performance against existing dLLM RL methods.
benchmark - arxiv:2611.07780 · cs.AIBridging Structure and Semantics: Knowledge Graph Verification via Triple-to-Text AlignmentZiang Cui
Semantic accuracy is an important dimension of knowledge graph quality. Graph-based scoring methods exploit relational structure, while textual representations allow classifiers to draw on knowledge encoded by pre-trained language models (PLMs). We present GenVer, a generative verifier that scores triples using both their serialized form and generated natural language representations (NLRs). A triple-text alignment procedure selects training pairs for a sequence-to-sequence generator; a BERT-based classifier then combines features from each triple and its generated NLR. On four FactFB variants, constructed from FB15k using perturbed negative triples, GenVer obtains the highest classification accuracy among the evaluated baselines, ranging from 88.56% to 91.52%. The experiments suggest that generated representations can help triple classification, while the case analysis also exposes semantic drift and high-confidence errors.
knowledge graph - arxiv:2611.07779 · cs.ROFlexible Piezoelectric Tactile Sensing System for Intelligent Object RecognitionRazan Khalifeh, Mohamad Yaacoub, Christian Gianoglio, Moustafa Saleh +2
Tactile sensing provides information about object properties that can improve robotic grasping and manipulation. However, accurate object recognition during the early stages of grasping remains challenging, particularly under strict real-time constraints. This study presents a flexible, distributed piezoelectric tactile sensing system for early-stage object recognition. The system was experimentally evaluated using six everyday objects and three recognition approaches: feature-based classification using support vector machines (SVMs), raw-signal classification using 1D convolutional neural networks (1D-CNNs) and convolutional recurrent neural networks (C-RNNs), and spatiotemporal heatmap classification using a 2D-CNN. The effects of sensor count and temporal window size on recognition accuracy and real-time feasibility were systematically investigated. Increasing the number of sensing elements improved recognition performance across all approaches. The best-performing method, based on spatiotemporal heatmaps and a 2D-CNN, achieved F1-scores of 0.9796, 0.9846, and 0.9903 using 16, 24, and 32 sensors, respectively. The optimal configuration, comprising 32 sensors and a 100-ms temporal window (50 samples), was deployed on an embedded edge platform, achieving an end-to-end recognition latency of approximately 101 ms from contact onset and an energy consumption of 63.8 microjoules per inference. These results demonstrate the potential of distributed piezoelectric tactile sensing for accurate, energy-efficient, and real-time object recognition during robotic grasping.
manipulationtactilegrasp - arxiv:2611.07778 · cs.AIThe Intrusion Pathway: How Non-Evidential Context Degrades Language Model PerformanceDikshant Kukreja, Gautam Gupta, Avni Mittal, Avinash Anand +7
Language models lose accuracy when a prompt contains non-evidential context: irrelevant statements from unrelated domains, or random off-distribution tokens, with no competing answer and no semantic overlap with the query. Such noise pervades retrieval-augmented, tool-using, and long-context deployments, yet the failure has lacked a mechanistic account. We trace the corruption to the Intrusion Pathway, which starts with MLPs enriching both query-subject and irrelevant-token representations, parroting heads carry prefix-token content to the prediction position, and relational readers largely avoid unrelated sources. Nevertheless, preserved subject enrichment and selective reading do not prevent intrusion in the broader computation. We then localize part of this interference to a sparse set of attention heads, which we call intrusion heads. Their prefix-derived writes promote distractor-related content, while their reduced support for the correct answer arises primarily from weaker non-prefix contributions rather than direct suppression by the prefix. The same heads transfer across tasks and distractor types and are largely weight-disjoint from the heads that integrate relevant contextual evidence, enabling a class-conditional intervention. Replacing the corrupted last-token residual with its clean counterpart, an oracle probe that requires a matched clean forward pass, recovers 97.6% on average of the distractor-induced drop across four downstream tasks. A deployable mask level knockout of about ten heads, roughly 1% of the model's attention heads, recovers 57% at bounded cost to helpful context.
long-contextretrieval-augmented - arxiv:2611.07773 · cs.LGOptimal Learning in Games with Arbitrary Delays: Best-of-Both-Worlds for Self-Play and Adversarial OpponentsEmmanuel Esposito, Andrew Jacobsen, Hao Qiu, Alexander Ryabchenko +1
We study learning in zero-sum games with delayed feedback. We introduce Multi-Anchor Optimistic Online Mirror Descent (MAO-OMD), a single algorithm which obtains optimal fast rates in self-play on smooth games, while simultaneously maintaining minimax optimal guarantees against arbitrary opponents. In particular, in smooth zero-sum games with arbitrary and unknown feedback delays, MAO-OMD obtains $O(σ_{\max}+1)$ social and individual regrets when used by both players, where $σ_{\max}$ denotes the maximum number of outstanding gradients, while also guaranteeing minimax optimal $O(\sqrt{T+d_{\mathrm{tot}}})$ individual regret against arbitrary opponents, where $d_{\mathrm{tot}}$ is the total delay. Moreover, we show that these optimal rates cannot be obtained simultaneously using the standard black-box reduction to the undelayed setting. We also extend our results to multi-player games, and to games with player-dependent delay sequences.
self-play - arxiv:2611.07765 · cs.CVBiGrounder: Bidirectional Training Unlocks the Potential of GUI GroundingDengyao Shi, Tianyu Fu, Fei Hu, Anyang Su +5
Graphical User Interface (GUI) grounding aims to map natural language instructions to precise operational coordinates, serving as a critical task for autonomous agent interaction. Current approaches formalize GUI grounding as predicting action-coordinate pairs from instructions, with end-to-end training supervised solely by final coordinates. Training datasets typically provide only a small number of descriptions per target coordinate, whereas real-world scenarios involve a much larger set of varied descriptions for the same target, which limits robustness to linguistic variation. While recent GUI grounding paradigms have begun employing reinforcement learning to enhance localization accuracy, models with limited exploratory capability before reinforcement yield constrained improvements even with identical training data during the reinforcement phase. To overcome these limitations, we propose a bidirectional reward paradigm that enables model self-reinforcement. During SFT, bidirectional training equips the model to generate action descriptions from target coordinates. During RL, the model samples semantically consistent instruction variations on-policy and grounds each variation back to its target. A cycle-consistency gate couples semantic quality to grounding success, turning the model's own instruction policy into a source of continually refreshed reinforcement trajectories. Our 4B-parameter model achieves competitive accuracy across five grounding benchmarks in our evaluation, indicating the effectiveness of the proposed training paradigm. https://github.com/Spell-shi/BiGrounder
agentautonomous agentbenchmark - arxiv:2611.07757 · cs.AIRIVER: Evidence-guided multi-agent retrieval for enzyme--reaction discoveryYifan Chang, Yunze Li, Feifei Jiang, Juntao Wu +11
Enzyme function annotation and the identification of candidate enzymes for a given reaction are essential steps in connecting genomic information to biochemical reaction networks. However, existing methods rely on sequence or structural similarity or on training with known enzyme--reaction pairs, making it difficult to integrate biochemical evidence across large candidate spaces; they also struggle with retrieval tasks involving enzyme families whose members are highly similar. Here, we introduce RIVER (Reaction--enzyme Inference via Evidence-guided Retrieval), a multi-agent framework for bidirectional enzyme--reaction retrieval that identifies potentially catalyzed reactions for a given enzyme and ranks candidate enzymes for a given reaction. RIVER coordinates large language models, specialized biochemical computational tools, and external knowledge resources to assemble evidence at multiple levels, including reaction chemistry, protein sequences and domains, catalytic mechanisms, and database annotations. Dedicated validation modules further examine evidence gaps and potential conflicts and rerank candidates using traceable evidence records. On public benchmarks, RIVER achieves the highest success rates in both enzyme retrieval and enzyme annotation. In family-specific enzyme retrieval, RIVER outperforms the evaluated baselines on the selected datasets and metrics. In Escherichia coli enzyme retrieval tasks constructed from recent literature, RIVER achieves the best retrieval results with closed-source and open-source model configurations, substantially outperforming closed-source general-purpose and biology-specialized multi-agent systems, including Claude Code, Codex, and Phylo Max. These results suggest that evidence-guided multi-agent retrieval can prioritize biochemically supported enzyme--reaction relationships, providing a basis for subsequent experimental validation.
multi-agentagent frameworkagent systembenchmark - arxiv:2611.07755 · cs.CLAutomated Video-to-QA Pipeline: End-to-End Question Generation and Verification Using Large Language ModelsJaival Chauhan, Vikram Singh, Pruthwik Mishra
The exponential growth of educational and informational video content on platforms like YouTube presents a challenge for efficient knowledge retrieval and assessment. This paper proposes an end-to-end automated pipeline for extracting structured summaries and generating high-fidelity Question-Answer (QA) pairs from video streams. The framework integrates a cascade of state-of-the-art models: audio extraction is handled via FFmpeg, followed by automatic speech recognition (ASR) using Whisper (A library in Python) to generate accurate transcripts. Due to the token constraints of Large Language Models (LLMs), transcripts are segmented and summarized using Facebook's BART. A key innovation of this work is a Self-Verification and Fine-Tuning Loop. We utilize meta-llama/Llama-3.1-8BInstruct to generate candidate QA pairs from transcript chunks, which are then fed back into the model to rigorously assess their relevance and importance. Only verified, high-quality pairs are retained to construct a synthetic dataset. This dataset is subsequently used to fine-tune the Llama model using Low-Rank Adaptation (LoRA), resulting in a specialized model capable of generating domain-aligned assessment questions. This approach demonstrates a scalable method for creating self-improving educational tools without the need for manual data annotation. The proposed model can be downloaded from Kaggle with all the codes from GitHub.
self-improving - arxiv:2611.07749 · cs.LGConvPE: Convolutional Position Embedding Improves TransformersZehao Wang
ConvPE (Convolutional Position Embedding) improves both the performance and efficiency of Transformer models in language modeling, long-context extension, and vision. It injects a windowed, content-dependent bias into attention scores through a lightweight causal convolution over a low-dimensional projection of the hidden states. As a lightweight plug-in, ConvPE is compatible with RoPE, attention masks, and KV concatenation. It also supports high-dimensional position embeddings via high-dimensional convolutions. In a reproduction, ConvPE outperforms five 1D baselines and eight 2D baselines. The code is publicly available at https://github.com/wzhagi/convpe.
long-context - arxiv:2611.07746 · cs.LGOrdering Multivariate Time Series by Volatility Memory: A Temporal Convolutional Approach to Volatility HeredityKaoutar El Azaar, Mohamed Hamlich, Moncef Garouani, Franck Ravat +2
Volatility dynamics can carry information about temporal relationships that is not apparent in the conditional mean. Extracting this information is difficult because the persistence of each variable can obscure its dependence on other variables. This paper presents QS-TCN, a method for ordering multivariate time series that combines statistical estimates of volatility memory with temporal convolutional learning. Persistence estimates obtained by quasi-maximum likelihood and temporal modes extracted by the Matrix Pencil method define structured filters within the network. Together with learned filters, they represent the history of the target and support comparisons of directional evidence between variables. These comparisons are combined into a global order of precedence, which does not identify a causal graph. The method is evaluated on synthetic systems with linear and nonlinear volatility interactions and on the Finance and fMRI benchmarks, whose reference relations are known. Across the four synthetic volatility regimes studied, QS-TCN reaches a higher mean ordering quality than an external lead--lag ordering and its squared-input adaptation, although the gain over the adaptation is uncertain in two regimes. Applications to currency, gold and cryptocurrency markets are reported descriptively.
memorybenchmark - arxiv:2611.07730 · cs.ROSeeing Beyond the Next Step: World-Model-Guided Human-Like Navigation in Multi-Agent ScenesSeong-Eun Hong, JuYeong Hwang, Jinhyun Kim, Hanyoung Jang +1
Selecting a next motion in a populated scene is not just a matter of avoiding what is currently in the way: an action can be locally feasible yet drift onto the road instead of the crosswalk, cut into a stream of crossing pedestrians, or bottleneck a narrow passage. These failure modes are invisible from the current state alone, and acting on them requires reasoning about what each candidate action will cause. World models offer exactly this, namely action-conditioned predictions of the future, but rolling them out at every decision step is a cost deployable navigation systems cannot pay, and the field has largely traded foresight for speed. We argue this is a false dichotomy. The natural place for a world model in human-aware navigation is at training time, as a preference oracle that distills foresight into a fast policy, rather than at inference time, as an online planner. We pair a discrete visual action prior with two task-aligned world models, one for nearby social dynamics and one for future scene semantics, that score candidate actions by their predicted consequences and refine the policy through group-relative preference distillation. At deployment, both world models are removed and the policy acts in a single forward pass. Across five held-out evaluation scenarios, this yields navigation that is simultaneously safer, more socially and group-coherent, and more goal-directed than strong reactive and learned baselines, with naive observers rating the resulting motion as more human-like. Taken together, our findings point to a different approach for bringing world-model foresight into deployable agents: not accelerating rollouts, but amortizing them into the policy so that the cost of looking ahead is confined to training. Project page: https://seoneun.github.io/-NeurIPS-2026-Seeing-Beyond-the-Next-Step/
world modelaction-conditionedmulti-agent - arxiv:2611.07726 · cs.RODual-Arm Tabletop Rearrangement with Movable Stacking and Gap-Aware Push TransfersMohammadreza Firoozalizadeh, Zakiye Rostamirad, Arman Barghi, Ahmad Kalhor +2
Object rearrangement in cluttered tabletop scenes is challenging, especially in wide workspaces where dual-arm systems improve coverage and efficiency. Limited free space often necessitates buffering, while objects near or across the shared reachability region of the manipulators must be transferred under careful coordination to avoid inter-arm collisions. Conventional inter-arm transfer typically relies on a fixed handoff configuration, which constrains where and how objects are exchanged, enforces spatiotemporal synchronization of the arms, and reduces flexibility in dense layouts. This paper presents a dual-arm rearrangement framework that combines movable stacking with Gap-Aware Push Transfer for asynchronous cross-boundary exchange. The method scans the shared boundary to identify collision-free transfer gaps, places an object near a selected crossing point, and then pushes it into the receiving workspace without requiring simultaneous interaction between the two arms. A joint Monte Carlo Tree Search (MCTS)-based planner coordinates both manipulators while reasoning over buffering and transfer decisions, and execution cost is used after planning to assess the resulting plan. Experiments show that the proposed method shortens joint execution sequences and becomes increasingly advantageous over a fixed handoff baseline as scene density increases, achieving lower execution cost in denser scenes while remaining competitive in lighter ones.
manipulator - arxiv:2611.07723 · cs.LGStatic Recovery Is Not Dynamic Stability: Dynamics-Aware Benchmarking of Protein Motif ScaffoldingEmile de Bruyn, Moritz Schäffler, Jannik Schneider, Sebastian R. Schmidt +8
Protein motif scaffolding benchmarks assess whether generative models can build scaffolds around functionally important residues. Their conclusions depend on which motifs are tested and which criteria define success. Existing benchmarks typically use small hand-curated motif sets and static refolding criteria, but protein function often depends on conformational ensembles and transitions. Here, we address both benchmark scope and success criteria. We introduce BILBY-PRO, a systematically derived benchmark of structurally conserved functional motifs from PROSITE-linked experimental structures, comprising 220 cases from 174 PROSITE motif-pattern entries. Evaluating six recent scaffold-generation models reveals distinct patterns in target coverage and unique-solution yield: RFDiffusion and Proteina solve more targets, whereas ESM3 produces more unique successful scaffolds on the targets it solves. We further introduce BILBY-MD, a reference-based all-atom molecular dynamics evaluation framework. Applied to statically solved designs from the original MotifBench benchmark, it compares motif drift, scaffold drift, motif-contact preservation, and motif-dihedral free-energy landscapes against reference simulations. Of 55 statically solved model-problem pairs, 31 satisfy our MD-supported criterion and only 8 pass all checks. Compact, self-contained motifs are often MD-supported, whereas interface-like or context-dependent motifs fail more frequently. Static refolding success therefore provides useful evidence of local motif recovery but does not establish dynamical fidelity of the full scaffold. BILBY highlights three complementary axes of motif-scaffolding evaluation: target coverage, unique-solution yield, and MD-supported dynamical fidelity.
benchmarkevaluation framework - arxiv:2611.07722 · cs.LGReinforcing What the Student Has Learned: Teacher-Guided On-Policy DistillationBozhen Fang, Linhao Zhong, Xuehao Wang, Huanyi Zheng +2
On-policy distillation (OPD) improves language-model reasoning through dense teacher feedback on student-generated responses.We first observe an amplitude disparity in this feedback: on math reasoning responses, style-related positions exhibit 3.5 times the mean absolute teacher--student log-probability difference of mathematical positions.Scaling OPD updates by these discrepancies thus directs much of the update strength toward matching the teacher's expression. We instead take the student's own pretrained base as a reference: the student--base difference records the token preferences acquired during post-training and provides supervision that does not depend on the student's distance from the teacher. Building on this reference, we propose Student-Delta On-Policy Distillation (SD-OPD), which uses centered student--base differences as update coefficients, keeps each coefficient only when its sign agrees with centered teacher--student feedback, and recomputes the student signal every training round. Across 5 teacher--student configurations with Qwen3 students of 1.7B to 8B parameters, SD-OPD improves six-task mathematics average pass@1 (avg@16) over OPD by 7.76 to 10.37 percentage points. Code is available at https://github.com/aim-uofa/sd-opd.
post-training - arxiv:2611.07718 · cs.LGMultimodal Information Interactions Bottleneck for Multimodal Self-Supervised LearningLiangjian Wen, Jiangtao Zheng, Jiang Duan, Qun Dai +7
Multimodal self-supervised learning is challenged by heterogeneity, unequal informativeness, and complex information interactions across multimodal data sources such as images, text, audio, and tabular data. While contrastive methods effectively align shared semantics, they typically rely on strong multi-view redundancy assumptions, overlooking modality-specific and synergistic information, thus leading to suboptimal representations. We propose the Multimodal Information Interactions Bottleneck (MIIB), a principled framework grounded in Partial Information Decomposition that explicitly encodes redundant, unique, and synergistic information within multimodal interactions. MIIB employs a dual-path model architecture for modality-specific encoding and multimodal fusion, optimized using surrogate tractable entropy-based objectives that enforce minimal sufficient representations without requiring explicit mutual information estimation. By leveraging variational bounds and score-based entropy estimation, MIIB efficiently captures complex cross-modal interactions while suppressing superfluous information. Experiments across diverse benchmarks demonstrate consistent improvements over state-of-the-art multimodal self-supervised methods. Code is available at https://github.com/wanglingzaishi/MIIB.
benchmark - arxiv:2611.07713 · cs.CVSelective Event Evidence and Persistent Temporal Priors for Event-Based Point TrackingXiaokai Song, Zhang Li, Leqi Liu, Yibin Ye +2
Event cameras enable tracking under fast motion and challenging illumination. However, stable long-term point tracking remains challenging: event activity is sparse and varies with scene content and motion, complicating correspondence, while trajectory estimates can drift over long sequences. Together, these challenges motivate a tracker that selectively uses informative event evidence and retains trajectory constraints across temporal windows. We propose Selective Event Evidence and Persistent Temporal Priors (SEPT), an event-only framework for tracking any point (TAP) over long sequences. SEPT introduces two complementary modules: Task-Adaptive Event Evidence Modulation (TEEM) and Persistent Trajectory Prior (PTP). Using only tracking supervision, TEEM learns a bounded reliability field that weights the event tensor across space, time, and channels before feature extraction. PTP carries geometric information across overlapping windows by marginalizing a fixed local quadratic trajectory model with a Schur complement. Its visibility-aware correction refines overlap coordinates, while a clipped terminal velocity initializes coordinates at newly exposed timestamps. SEPT builds on the ETAP architecture. SEPT is trained exclusively on synthetic EventKubric data and evaluated on five datasets across TAP and feature tracking. Results show that SEPT outperforms the retrained ETAP baseline in average Jaccard for TAP and in feature age and expected feature age for feature tracking; it also surpasses several methods combining event and frame inputs on EC and EDS. The source code is publicly available at https://github.com/sssxxxkkkk/SEPT.
event camera - arxiv:2611.07712 · cs.ROAffectiveVLA: Safe and Intent-Aware VLA Control for Human-Humanoid Social InteractionNikita Kuzmin, Georgii Demianchuk, Mariya Lezina, Yara Mahmoud +3
Social touch interactions, such as handshakes, fist bumps, and hugs, require humanoid robots to infer human intent, generate contact-rich motions, and maintain safe physical contact. We present AffectiveVLA, an intent-gated Vision-Language-Action (VLA) pipeline deployed on a Unitree G1 humanoid with dexterous hands. A lightweight RGB-based model recognizes interaction intent from two onboard cameras and converts it into a language prompt for the VLA policy. Low-confidence predictions suppress motion execution, reducing unintended activations. During physical contact, interaction-specific compliance profiles adjust arm impedance, while an independent voice-triggered mechanism initiates a latched safe-stop procedure. Using VR teleoperation, we collected approximately 1,300 multimodal interaction episodes and fine-tuned four VLA models: GR00T, RLDX-1, Being-H0.5, and UnifoLM-VLA-0. On-robot evaluation identifies GR00T as the best-performing standalone model. Integrating intent recognition increases its macro-F1 score from 0.75 to 0.90, while the intent model achieves a p95 inference latency of 6.27 ms. These results demonstrate that explicit intent gating improves the reliability of VLA-based social interaction without compromising real-time operation.
vision-language-actionvlavla modelvla policydexteroushumanoid - arxiv:2611.07707 · cs.AICAMP: Context-Aware Mapping and Placement for LLMs on 3D-IMC ArraysOrhun Görkem, Athanasios Vasilopoulos, Georgios Pagonis, Victoria Clerico +5
Large Language Models (LLMs) incur severe memory-access overheads due to the cost of loading large weight matrices and key-value (KV) caches. Analog In-Memory Computing (AIMC) is an emerging solution that enables computation directly on static weights stored in memory, substantially reducing energy and latency. 3D-IMC further stacks memory arrays into vertical tiers, increasing capacity but revealing a new challenge: how weights are placed in memory determines which matrix-vector multiplications (MVMs) can execute concurrently, affecting latency and throughput. The concurrency aspect of 3D-IMC weight-mapping remains underexplored in the literature, leading to sub-optimal utilization of compute resources on 3D-IMC chips. In this context we propose CAMP (Context-Aware Mapping and Placement), an inference-context-aware weight mapping algorithm for 3D-IMC-based LLM accelerators. CAMP captures the potential resource contention resulting from the architecture, targeted network, and runtime context in a conflict matrix, enabling throughput-optimized weight mapping. This mapping is refined through a profiling-driven feedback loop that efficiently prunes the search space and returns optimized mappings within practical runtimes. CAMP is an offline, software-based solution requiring no hardware changes and improving 3D-IMC compute utilization from 42.4% to 60.3%, delivering up to 2.0x higher 3D-IMC throughput and 1.74x higher end-to-end throughput over state-of-the-art mapping algorithms.
memory - arxiv:2611.07706 · cs.AITowards Generalist Graph-Level Anomaly DetectionJunjun Pan, Yixin Liu, Yu Zheng, Fuyi Li +2
Graph-level anomaly detection (GLAD) aims to identify graphs that deviate significantly from the majority. While recent advances have achieved competitive performance, existing approaches mainly follow a dataset-specific paradigm, requiring substantial in-domain data and training cost when applied to new application scenarios. This limits their scalability and practicality in real-world applications where labeled data is scarce and diverse domains are continuously emerging. In this paper, we investigate the problem of generalist GLAD, which seeks to learn a single unified model capable of detecting anomalous graphs across multiple domains with minimal supervision on the target data. This problem introduces key challenges, including learning transferable patterns from heterogeneous graph datasets and effective adaptation under limited supervision. To address these challenges, we propose GenGLAD, a generalist GLAD approach that leverages discriminative subgraph extraction as the core mechanism to learn transferable anomaly knowledge and enable efficient adaptation to new domains. Specifically, GenGLAD employs a graph neural network-based discriminative subgraph extraction model to learn the anomaly-discriminative substructures that distinguish anomalous and normal graphs across diverse domains. During the test-time adaptation phase, multi-dimensional discrepancies between the extracted subgraph and multi-level contextual information are calculated to characterize anomalies from multiple perspectives. For efficient adaptation, a lightweight scoring module is introduced to automatically decide the most informative signals, improving robustness against misleading patterns. Extensive experiments on diverse benchmark datasets demonstrate that the proposed approach achieves strong performance, superior generalization ability, and improved data efficiency compared to existing methods.
benchmark - arxiv:2611.07705 · cs.CVFlyingDrones: A Dataset and Benchmark for Optical Flow Estimation from UAV motionLaurenz Ruzicka, Roman Pflugfelder, Daniel Cremers
Optical flow is an important cue for drone analysis in aerial surveillance, airspace monitoring, and long-range target tracking, where motion remains informative when appearance is weak. Estimating the motion of small Unmanned Aerial Vehicles, however, is difficult because distant targets occupy few pixels, have low contrast, and move against complex dynamic backgrounds. Standard optical-flow benchmarks emphasize large, object-centric motions and therefore do not characterize distant aerial targets well. We introduce FlyingDrones, a generated benchmark with dense ground-truth optical flow and segmentation masks derived from physically based rendering, and define resolution-normalized metrics (EPE Norm) that normalize flow by image dimensions. We evaluate 16 state-of-the-art model families on a motion- and size-stratified subset of 999 flow pairs at three resolutions. Every family recovers drone motion better than the zero-flow reference, but the gap between families widens with resolution: at 1920x1080, global-matching, recurrent, transformer, and short-window temporal models separate clearly from pyramid, classical, self-supervised, lightweight, and diffusion models. Within-family checkpoint spread is comparable to or larger than between-family spread, so the training curriculum is as important as the architecture. Two complementary studies test the benchmark beyond synthetic accuracy: a ground-truth-free motion-alignment evaluation on 21 cameras of real UAV footage shows that the synthetic ranking carries a moderate, significant signal about real-world performance (Spearman \r{ho} = 0.63 over 41 checkpoints) while the exact top ranks differ, and a fine-tuning study shows that 800 FlyingDrones pairs reduce held-out drone-region error by up to 37% and improve drone localization on the real footage for all four architectures tested.
benchmark - arxiv:2611.07702 · cs.LGInfMasking: A Unified Variational Framework for Synergistic Multimodal LearningLiangjian Wen, Qun Dai, Jiang Duan, Yong Dai +6
Multimodal models are expected to exploit synergistic information, that is, information available only when multiple modalities are observed together. In practice, however, a fusion module can fit the training objective by relying on a dominant modality or on information that modalities share, and the synergy between modalities is often left unused. We propose InfMasking, a unified multimodal learning framework based on Infinite Masking. Before fusion, InfMasking randomly masks the features of each modality; each sampled mask produces a different partial combination of cross-modal evidence and a fused representation of the corresponding modality interactions. All masked fused representations are required to remain consistent with a reference target: the unmasked fused view when training is unsupervised, and the class proxy of the task label when labels are available. Maintaining this consistency across diverse partial combinations discourages reliance on a dominant modality or on shared information alone, and drives the fusion module to learn cross-modal interaction patterns that remain meaningful under partial observations. Since the objective is an expectation over infinitely many masks, we model the masked fused representations as Gaussian and optimize closed-form variational bounds, whose population-moment gaps are controlled by the masking-induced covariance under this model. On the evaluated benchmarks, InfMasking improves predictive performance and yields higher mean PID estimates of task-label synergy in learned modality-specific representations, alongside higher InterSHAP interaction attribution in the trained predictors. Code is available at https://github.com/brightest66/InfMasking.
benchmark - arxiv:2611.07695 · cs.LGEquivariant Deep Learning in Direct-Product Space for Electronic-Structure Hamiltonian PredictionShi Yin, Zujian Dai, Xinyang Pan, Lixin He
Most existing equivariant architectures for Hamiltonian prediction propagate features in direct-sum spaces of irreducible spherical tensors, whereas the target Hamiltonian is naturally an operator block in a direct-product representation over local orbital and spin bases. This representation mismatch fragments contiguous Hamiltonian blocks, obscures direct manipulation of orbital couplings, and organizes features into irregular angular-momentum components that require numerous transformations and are poorly aligned with efficient hardware execution and memory management. To address both the representation mismatch and computational bottleneck, we propose DP-Learner, an equivariant learning framework that performs message passing directly on Hamiltonian-like operator blocks in the direct-product space. Theoretically, we propose Lie-generator-based directional embedding and invariantly weighted left--right products and contractions that preserve equivariance to SO(3), SU(2), and spatial inversions, without explicit Clebsch--Gordan decomposition or inverse reconstruction. Methodologically, we instantiate this principle as a Transformer architecture, in which Hamiltonian-like feature matrices or tensors evolve through equivariant operators in the direct-product space, while non-linear invariant scalar routing controls structured gating over matrix units, directional propagation, attention, and environment-embedded context interactions. We further construct MP-HAM-110K, a large-scale Hamiltonian dataset comprising more than 110,000 material structures covering 68 elements with both Hamiltonian and wavefunction data. Experiments on MP-HAM-110K and Materials-HAM-SOC demonstrate that DP-Learner dramatically improves Hamiltonian prediction accuracy and inference efficiency over the state-of-the-art method.
manipulationmemory - arxiv:2611.07689 · cs.AIMCPHallu: Benchmarking Reasoning, Execution, and Memory Hallucinations in MCP AgentsYizhen Jiang, Haowei Guo, Tianyi Bai, Zengjie Hu +5
Large language model agents are increasingly deployed through the Model Context Protocol (MCP), which allows them to discover and invoke tools across multiple servers at runtime. Existing MCP benchmarks primarily measure task completion; where they diagnose failures, they do so after execution rather than building specific reasoning, execution, and memory failure risks into task construction. We introduce MCPHallu, a benchmark for diagnosing hallucination-prone agent failures in MCP environments along three axes: reasoning, execution, and memory. MCPHallu contains 358 tasks across five capability domains and runs on 36 production MCP servers in isolated containers. Each task is designed with a structural trigger that makes one of four subtypes the dominant failure risk: Branch Collapse, Unreachable Goal, Tool Misuse, or Context Forgetting. This design enables subtype-level analysis rather than aggregate task success alone. We evaluate fourteen frontier LLM agents from eight vendors on 5{,}012 trajectories. Current MCP agents remain vulnerable to hallucination-inducing tool environments, achieving an average score of 0.642 on a $[0,1]$ scale (higher is better). Failures concentrate on Unreachable Goal tasks, where the average score drops to 0.458 and 23.2\% of runs pass, indicating that agents often continue issuing tool calls or claim successful completion rather than recognizing infeasible goals. Subtype rankings diverge sharply: a model's rank shifts by 6.9 of 13 positions on average across subtypes, and Branch Collapse is uncorrelated with Unreachable Goal across models ($r = -0.12$). MCPHallu is released at https://github.com/J1zz/mcp_hallu.
memoryagentllm agentbenchmark - arxiv:2611.07683 · cs.CVReferClaw: Disambiguating User Intent for Egocentric AssistanceHangyu Qin, Junbin Xiao, Angela Yao
Users of egocentric AI assistants often ask short, ambiguous questions whose intended referents cannot be uniquely resolved from video and dialogue history. For example, "Where did I leave it?" may refer to several previously handled objects. Existing VideoQA systems typically commit to a single interpretation, risking plausible but intention-misaligned answers. We introduce ReferClaw, an agentic framework that formulates egocentric streaming QA as an interactive process: the assistant asks clarifying questions to resolve user intent before answering. Three cooperating agents identify candidate interpretations, request clarification, and generate answers grounded in the resolved intent and visual evidence. A bounded, cost-aware policy balances expected information gain against user effort while limiting clarification turns and question complexity. Referclaw learns from interaction through intent-specific skill updates and process-reward-guided policy optimization with agent-level credit assignment. Evaluations on augmented egocentric QA benchmarks demonstrate improved answer accuracy and joint answer-and-grounding performance over existing baselines, alongside zero-shot transfer across datasets. These findings support interactive clarification as a practical approach to resolving online user intent in egocentric streaming assistance.
agenticbenchmark - arxiv:2611.07679 · cs.ROPhySA: Decoupling Aesthetic and Physical Rewards for Humanoid LocomotionYunwan Fu, Sicheng He, Jiageng Mao, Yue Wang
Humanoid locomotion policies trained without an explicit smoothness objective tend to produce jerky, high-energy gaits that stress actuators. Vision-language models (VLMs) offer an appealing reward source: given a rendered rollout, a VLM returns an aesthetic quality score on [0,10] without task-specific feature engineering. We find, however, that aesthetic preference and physical smoothness point in opposite directions: across four production VLMs, clips with higher joint acceleration receive higher aesthetic scores (mean Pearson r=+0.66). We study PhySA (Physical Smoothness and Aesthetics), which keeps the two on separate, independently removable reward channels: a sparse, asynchronous VLM score and a phase-triggered correction channel that adjusts physical penalty weights from simulator state. Ablations attribute the smoothness gain to the correction channel. On MuJoCo Humanoid-v5 at 2.6M steps, PhySA lowers joint acceleration by 26 to 32% relative to vanilla PPO, Text2Reward, and Eureka (three seeds), at the cost of shorter episodes. In a single extended run it lowers joint acceleration by 59.5% relative to Prompt2Policy, while a three-seed replication shows a smaller, non-significant reduction (7.2%). The gain does not transfer to IsaacLab or the Unitree G1, even after rescaling the smoothness term.
humanoid - arxiv:2611.07669 · cs.LGAn Information Bottleneck for Post-hoc Memorization Mitigation in LLMsDina EL Zein, Shashi Kumar, James Henderson
Recent studies have shown that large language models (LLMs) can memorize and reproduce copyrighted training data, raising serious privacy and intellectual property concerns. In this work, we investigate whether memorization can be mitigated without modifying or retraining the underlying model through an information bottleneck. We propose a modular plug-and-play framework that inserts a nonparametric variational information bottleneck (NVIB) between a frozen LLM backbone and its language modeling head. By compressing the final hidden representations before token prediction, the proposed bottleneck aims to preserve general linguistic knowledge while filtering out highly specific memorized sequences. We evaluate our approach on memorization extraction tasks involving book passages and on standard language modeling benchmarks to measure the trade-off between memorization reduction and downstream utility. Our experiments show that the NVIB layer substantially reduces extractability while maintaining competitive performance on benchmark tasks. These results suggest that adding an information bottleneck offers a practical and architecture agnostic direction for post-hoc memorization mitigation in pretrained LLMs.
benchmark - arxiv:2611.07667 · cs.AIWho Should Know What? Empirical Information Design for Multi-Agent LLM EconomiesMete Erdogan, Mirac Suzgun, Aneesh Pappu, Guowei Xu +6
The performance of multi-agent LLM systems depends not only on the capabilities of the underlying models but also on information-sharing policies, which govern what information each agent observes and at what level of detail. But does providing AI agents with more detailed information always improve collective outcomes, or can transparency paradoxically degrade efficiency? To investigate this, we introduce empirical information design, an experimental framework that systematically compares disclosure policies across economic markets populated by LLM agents, while holding interaction rules and models fixed. We evaluate ten frontier LLMs in multi-resource trade, bilateral bargaining, and double auctions, using ground-truth utilities to decouple identified opportunities from realized outcomes. We find that finer information disclosure yields highly model-dependent results. In trade, revealing exact valuations rather than ordinal preference rankings reduces realized gains for GPT-6 Sol (from 98% to 91% of maximum) yet increases them for Claude Sonnet 5 (from 72% to 86%). In bargaining, exact information leads Grok 4.7 to produce less balanced agreements, whereas Claude Opus 5.5 divides gains symmetrically. In double auctions, full disclosure inflates price demands, triggering market breakdown and lowering welfare for Gemini 3.1 Pro (97% to 78%), Claude Opus 5.5 (94% to 64%), and Grok 4.7 (97% to 62%). Overall, these results suggest that model selection and information design must be co-optimized, as LLMs vary in how their strategic responses to information impact system objectives.
agentai agentllm agentmulti-agent - arxiv:2611.07663 · cs.ROEnergy-Aware Open-Vocabulary Detection Scheduling for Low-Power Robotic AutonomyAbdul Basit, Muhammad Shafique
Open-vocabulary object detection (OVD) lets robots search for run-time object queries, but fixed-rate inference spends embedded compute on low-value frames during idle motion, map revisits, or stable viewpoints. This paper presents an energy-aware OVD scheduler for low-power mobile robots. The scheduler uses odometry, map change, and detector confidence to switch among low-, mid-, and high-power YOLO-World-S modes on a Jetson Orin Nano. We evaluate the energy-quality trade-off on an AgileX LIMO using 40 real navigation trials: two target objects, five placements per target-policy pair, and four perception policies. Across bottle and cup trials, the scheduler records 8/10 mission success compared with 7/10 for each fixed-rate baseline, while reducing mean Jetson input power (a Jetson-side power-draw metric) by 5.2% relative to an 8 Hz high-rate baseline and lowering achieved OVD call rate by 34.0%. It does not minimize mission-window Jetson input energy because longer missions can erase instantaneous power savings. A separate 60 s stationary microbenchmark isolates the perception workload: the scheduler lowers mean power from 8.045 W to 7.301 W and energy from 484.7 J to 438.1 J relative to the 8 Hz baseline. These results support robot-state-aware OVD scheduling as a promising power-management mechanism with competitive task reliability, while showing that mission-window Jetson input energy must be read together with task progress and runtime.
benchmark - arxiv:2611.07662 · cs.ROWhich Objects Stay? Learning Hierarchical Persistence Beliefs for Lifelong Open-Vocabulary Object MapsThanh Nguyen Canh, Haolan Zhang, Nak Young Chong
Lifelong object maps must estimate whether an object remains at its previously observed location after it leaves view. Existing persistence filters rely on class-based survival priors, making them difficult to apply reliably in open-vocabulary settings. We introduce HiPer, a hierarchical persistence filter that derives persistence beliefs directly from data rather than semantic categories. For a new object, HiPer retrieves visually and geometrically similar training instances and estimates an initial survival rate from their observed histories, without requiring class labels. This estimate initializes a conjugate Beta belief over the object's survival probability, which is updated in closed form using detections and missed observations under Rosen's persistence-filter framework. When an object is re-identified at a new location, its persistence belief is transferred with the object, enabling the map to learn object-specific mobility over time. Experiments on 120 held-out simulated sequences show that HiPer consistently achieves the best performance among competing methods. Retrieval-based priors outperform CLIP-derived label priors, and object-specific online learning provides further gains. Results on 47 held-out 3RScan environments demonstrate that the proposed retrieval-based prior also outperforms state-of-the-art baselines in real-world settings.
online learning - arxiv:2611.07659 · cs.LGSimple Meta-Learning for HPO and CASH Using Performance RepresentationsEmiel Witting, Cheng Yan, Tom Julian Viering
Bayesian optimization for Hyperparameter Optimization (HPO) and Combined Algorithm Selection and Hyperparameter Optimization (CASH) often relies on explicit hyperparameter encodings as surrogate inputs. In high-dimensional search spaces, however, such representations can make it difficult for standard surrogates to learn useful performance models. We instead study an alternative meta-learning representation that describes each configuration directly as its historical performance across prior tasks. This representation can be paired directly with standard surrogate models. Our scope is the iterative selection component of AutoML, deciding which configuration to evaluate next from a fixed candidate set, rather than a complete framework with features such as runtime estimation, cost-aware budgeting and ensembling. We further apply this representation to maximal marginal relevance (MMR), a diversity-based re-ranking criterion from recommender systems, as a posterior-free acquisition rule. When selecting and optimising across multiple models on the TabArena benchmark, the performance-based representation paired with a linear Bayesian ridge regression model outperforms dedicated meta-learning baselines, achieving the same tuned performance in roughly half the number of trials. For single-model HPO, it remains a competitive second. Across both benchmarks this method closely matches or improves on regret at a fraction of the complexity and cost, requiring no pre-training or tuning and scaling independently of search space dimensionality. Code for reproducing our results is available at https://github.com/EWitting/simple-metalearning.
benchmark - arxiv:2611.07652 · cs.AIDon't Prompt Your Agents. Equip Them: Skillware as a Modular Standard for Sovereign, Auditable, and Deterministic Capability Execution for Logical SystemsVladimiros Peilivanidis
Current autonomous logical systems rely on a fragile assumption: that language models should dynamically write code, guess API payloads, and memorize operational constraints through prompt engineering. In practice, this burns massive context windows, inflates API bills, and produces brittle agents that fail unpredictably in production. In this work, we present Skillware, an open architectural framework and capability registry that formally separates what an agent wants to do from how that task is executed. Instead of prompting models with loose text instructions or markdown files baptized as 'skills', Skillware equips them with modular, versioned software capability bundles. Much like skill hotkeys in video games that execute deterministically under fixed world rules, each bundle combines a typed JSON input contract, deterministic Python code, task directives, bundled datasets, and automated unit tests. In-process adapters translate these capabilities into native tool declarations for any model, from frontier cloud APIs to local open-weight runtimes with zero network latency. Evaluating 23 frontier and open-weight models across 368 benchmark conditions, we show that small local models (3B to 8B parameters) equipped with Skillware consistently outperform unequipped frontier models on complex multi-step workflows. They cut token usage by up to 84.8%, slash inference costs by up to 97.3%, and eliminate parameter errors entirely. True agent reliability does not come from larger models or longer prompts, but from equipping logical systems with verified, deterministic software. Finally, we articulate the trajectory of capability bundles across an invariant continuum of information processing systems: from digital logical systems today, to IoT, drones, and embodied robotics tomorrow, establishing a foundation for auditable, real-time capability execution.
embodiedagentbenchmark - arxiv:2611.07651 · cs.ROGraspBench: Execution-Aware Cross-Robot Evaluation of 6-DoF Grasp Generation in MuJoCoAbdul Basit, Bibek Poudel, Muhammad Shafique
Modern 6-DoF grasp generators are often evaluated using grasp scores, dataset-level metrics, or single-robot demonstrations, yet proposal quality alone does not determine whether a grasp is executable by a particular robot. We present GraspBench, a controlled MuJoCo protocol for evaluating complete grasp-generation pipelines after robot-specific inverse-kinematics filtering, motion planning, and fixed approach-close-lift execution. GraspBench evaluates six configured pipelines across four robot embodiments, five diagnostic objects, and a structured tabletop pose grid, yielding 540 scenarios per method and 3,240 method-scenario evaluations in total. Missing outputs are conservatively counted as failures, and success follows the binary lift criterion produced by each method's execution runner. GraspGen achieves the highest overall success rate (73.1%), followed by EconomicGrasp (63.5%), while robot- and object-conditioned analyses reveal substantial variation across embodiments and object geometries. GPG produces the largest native candidate sets but achieves 44.3% success, indicating that proposal volume alone does not guarantee robot-executable grasps. Cluster-aware paired intervals support the principal method differences while not estimating repeated-execution uncertainty. The results characterize complete configured pipelines on fixed scenarios rather than intrinsic model quality or stochastic repeatability, showing that grasping performance depends jointly on the generator, robot embodiment, scene, and execution protocol.
grasp - arxiv:2611.07648 · cs.CVGenerator-Aligned Reward Modeling with Target-Aware Bounded Optimization for Visual GenerationZiqiao Weng, Yuanzhi Liang, Haibing Huang, Weidong Cai +1
Post-training aims to enhance desired attributes in visual generative models while preserving their general generation capabilities. Reward-guided post-training faces two challenges: real training data may mismatch the generator's output distribution, and attribute supervision may be confounded by unrelated content. Unbounded reward maximization can also cause overoptimization and degrade generation quality. We propose a framework combining Generator-Aligned Reward Modeling (GARM) and Target-Aware Bounded Optimization (TABO). GARM mitigates distribution mismatch through generator-support matching and reduces content confounding using content-matched real contrasts, yielding generator-relevant, attribute-focused reward supervision. TABO maps the learned reward to a smooth, upper-bounded function calibrated from real target-positive examples. This transformation is independent of the optimization algorithm and supports both direct reward-gradient and policy-based optimization. We further introduce Target-Aware GRPO (TA-GRPO), an adaptation of Flow-GRPO that combines fixed stage-wise advantage scaling, a generator-to-target curriculum, and timestep-balanced credit assignment for flow-based video optimization. Experiments show that GARM improves target-attribute expression while maintaining competitive generation quality. TABO better balances attribute enhancement and visual quality than unconstrained reward maximization, while TA-GRPO improves target enhancement over vanilla Flow-GRPO and reduces unintended attribute leakage on neutral prompts. These results highlight the importance of generator-aligned reward supervision and controlled, data-calibrated optimization for targeted visual generation.
post-training - arxiv:2611.07644 · cs.CVNeural Image Compression with Internal-External Prior Fusion and AlignmentTiancheng Si, Wenlong Xu, Dahua Gao
Existing context models for learned image com- pression rely primarily on internal priors from the current image, overlooking common patterns and typical textures shared across natural images. To address this limitation, we propose an Internal-External Prior Fusion-Alignment (IE-PFA) context model built upon quadtree spatial-channel joint modeling, where a learnable dictionary shared by the encoder and decoder serves as an external prior. Specifically, an RWKV-based Spatial- Channel Prior Fusion (RSCPF) module efficiently fuses inter- nal and external priors along spatial and channel dimensions with linear complexity, while a training-only Self-Distilled Prior Alignment (SDPA) mechanism aligns them without inference overhead. Moreover, our model supports variable-rate coding over 64 quality levels within a single network and is accelerated by custom CUDA operators. Experiments on seven benchmarks demonstrate state-of-the-art rate-distortion performance, achiev- ing an average BD-Rate saving of 21.65% over VTM-22.0-Intra, with only 29.9M parameters and the fastest coding speed among advanced neural codecs (61 ms encoding and 52 ms decoding). Code is available at https://github.com/stcnjtech/DCVC-X-Intra.
benchmark - arxiv:2611.07640 · cs.AIBeyond Instruction Override: Benchmarking Indirect Prompt Injection with Task-Tampering AttacksRongting Zhang, Martin Bertran, Sandeep Avula, Chia-Jung Lee +1
Indirect prompt-injection (IPI) benchmarks predominantly test planted instructions that redirect a tool-using agent toward an attacker-chosen task. Yet attacker-controlled content can instead tamper with the outcome of the task already in progress. A buyer's dispute evidence can cite a false refund policy to inflate a refund without changing the task. We define IPI by authority rather than instruction-like wording: planted content constitutes an attack when it attempts to influence a decision that its source is not authorized to determine. We introduce a benchmark that makes this boundary explicit: in each of 90 validated cases across three tool-use environments, the user request specifies the source that resolves the targeted decision. Every attack tampers with a decision required by the benign task rather than replacing that task. Each case has attacks with the same harmful outcome that target up to six parts of the agent's decision making: its objective, decision rules, available actions, expected consequences, reading of observations, or beliefs about the current state. Without a defense, attacks on three of the five other targets (decision rules, available actions, and state) have higher mean success than objective attacks across five LLMs. A detection-and-removal defense reduces objective-targeting attacks to 2.4% success but leaves observation- and state-targeting attacks at 9.5% and 12.1%; this gap persists under detection-optimized rewrites. A tool-policy defense still allows 15.4% attack success, close to 17.0% without a defense, because these attacks generally retain expected tools and manipulate values obtained during execution. Evaluations whose attacks redirect the task or target only the agent's objective can therefore overestimate the coverage of these defenses. Code and data are available at https://github.com/amazon-science/task-tampering-prompt-injection.
agenttool-usebenchmark - arxiv:2611.07639 · cs.LGMeasuring the Marginal Value of Verifier-Based RL Beyond Supervised Fine-TuningAmit Sagtani, Saloni Garg
Reinforcement learning from verifiable rewards (RLVR) is usually evaluated as an alternative to supervised fine-tuning (SFT), or as a stage that follows it, with the implicit assumption that the RL stage contributes gains of its own. We measure that contribution directly, in a controlled setting where both stages can be run to completion and compared exactly. We built gpu-train-debug, a deterministic benchmark of 1,877 synthetic GPU and LLM training failure diagnosis tasks across 30 archetypes, graded by a programmatic verifier with partial credit, and calibrated so that a mid-tier model scores in a 0.60 to 0.75 band. A pilot produced a surprise: 1,200 gold demonstrations nearly solved the benchmark (SFT 0.785 on the hard split), while 15 steps of GRPO from the base model did not move it (0.314 vs 0.311 base). We therefore amended the study, in writing and before the full run, to ask the sharper question: what does verifier-based RL add beyond demonstrations? At the pre-registered scale (60 GRPO steps, learning rate 1e-6), the answer is: no detectable improvement in-distribution (+0.22 points, 95% CI [-0.06, +0.51]), and a modest but measurable improvement under distribution shift (+0.33 points on a held-out OOD split, CI [+0.02, +0.68]). SFT accounted for most of the available improvement, while verifier-based RL produced only modest additional gains under distribution shift. The residual errors SFT leaves are mostly systematic, identical across samples, which likely reduces the within-group learning signal available to GRPO. A full 60-step cold-start GRPO arm confirms the pilot at the full horizon. The whole study cost $8.69 of compute and is fully reproducible from archived generations, frozen configs, and tagged code.
benchmark - arxiv:2611.07630 · cs.ROReal-Time Event-Based LED Tracking and Pose Estimation with PWM-Encoded MarkersJunkai Niu, Sheng Zhong, Delei Kong, Xin Wang +4
Event cameras are neuromorphic vision sensors with pixels that independently and asynchronously respond to brightness changes at microsecond resolution, supporting state estimation under rapid motion, motion blur, and high dynamic range (HDR) illumination. This also enables capture of high-frequency blinking patterns from light-emitting diodes (LEDs). Based on this, we propose a real-time event-based framework for LED tracking and six-degree-of-freedom (6-DoF) relative pose estimation using pulse-width modulation (PWM)-encoded LED markers. Multiple LEDs share a common blinking frequency and encode identities through distinct PWM duty cycles, reducing frequency-resource requirements and computational cost. Our dual-rate asynchronous tracking framework combines low-rate LED detection with high-rate raw-event updates, maintaining robust LED identity association and accurate pixel-level tracking under rapid motion. LED identities and positions establish reliable 2D-3D correspondences for real-time metric 6-DoF relative pose estimation. Extensive evaluations on a self-collected dataset demonstrate robustness under large motion flow and transient image-plane LED overlap. Furthermore, we develop a coordinated robotic platform comprising a quadrotor and a Mecanum-wheeled mobile robot, demonstrating the proposed framework's potential for closed-loop drone guidance through hovering and target-following experiments under time-varying HDR illumination, with onboard deployment on an NVIDIA Jetson Orin NX. The dataset and implementation will be released to facilitate further research.
event camera - arxiv:2611.07629 · cs.LGOnline Learning with Ordered Experts (via Two-point Feedback)Shiliang Zuo
We study adversarial online learning with expert recommendations in a totally ordered action space, where the experts' recommendations may change and be reordered arbitrarily. Our focus is two-point feedback: on each round the learner plays two current recommendations, observes both, and pays their average. We study two problems. In the separated monotone problem, the loss is the sum of a hidden nondecreasing component and a hidden nonincreasing component, and a query reveals the two components separately. In the BV problem, the losses are bounded and have bounded variation (BV) along the order, and a query reveals only the loss. In the separated monotone problem, two paid observations achieve \[ O\bigl(\log(|\cF|T)\sqrt{T\log|\cF|}\bigr) \] regret against the best of $|\cF|$ experts, within a logarithmic factor of the classical full-information bound. For the BV problem, we obtain $\wtO(\sqrt T+|\cF|)$ regret. We give an application in dynamic pricing with expert advice, showing that two-option menus matches the full information rate up to $\log$ factors. The main technical ingredient is a marginal-preserving endpoint flow estimator. Two observations reveal a loss increment along the order; random paths supply telescoping identities that turn this local information into unbiased loss estimates for all current recommendations. A reversible pair sampler preserves the learner's intended marginal at both played actions. The estimator's second-moment and stability bounds support online updates for both feedback models.
online learning - arxiv:2611.07624 · cs.AIBeneath the Surface: What We Overlook When We Talk About Agent EvolutionXiaobo Wang, Zimu Wang, Siyu Ye, Jiaqi Li +12
Agent harness evolution enables recursive self-improvement (RSI) without updating model weights. We examine whether higher benchmark scores yield reliable, efficient, and transferable improvement. We evaluate four methods across harnesses, general and coding tasks. Repeated evaluations, controlled agent expansion, component interventions, and admission replays reveal recurring patterns. All sixteen evolved agents gain previously uncovered tasks while losing previously covered ones. In one general-task case, reliability rises by 5.2 points despite losing ten previously covered tasks. Most reliability gains also increase deployment cost. Specific changes usually outperform generic reminders, but their benefits can change when combined. The highest-scoring agent does not yield the largest mean improvement in any of our four pools, and stricter admission can discard useful alternatives. Transfer selectively preserves gains: coverage usually falls on the general-task target across both harness types and paradigms, but rises on the coding target with the outer-layer-only harness. These findings motivate preservation-aware evaluation, tested decision rules, and explicit applicability conditions. As a constructive follow-up, EvoAudit turns these diagnostics into revisions of select, analyze, mutate, and gate, achieving the best performance.
agentself-improvementbenchmark - arxiv:2611.07618 · cs.MAA Bilevel Neurodynamic Approach for Distributed SCAD-Regularized Sparse Least SquaresJuntao Li, Qihang Guo, Deyuan Meng
This paper proposes a bilevel particle swarm optimization (PSO)-neurodynamic method for solving the distributed sparse least-squares problem regularized by the smoothly clipped absolute deviation (SCAD) penalty to global optimality. The SCAD penalty is nonsmooth and nonconvex, so existing distributed neurodynamic methods have not been extended to it directly. Through two-time-scale singular perturbation analysis, a rigorous full-trajectory convergence theory with explicit convergence rates is established for the inner collaborative multi-agent neurodynamic system without requiring the second-order sufficient condition. A theory-guided inner-outer collaborative optimization framework is further developed, in which a red-light green-light sensitivity mechanism is embedded in the outer PSO search to enhance global search while concentrating computational effort on the SCAD shape parameters, thereby improving recovery accuracy and computational efficiency. Simulation results on four sparse-recovery scenarios demonstrate that the proposed method accurately recovers the true sparsity pattern and achieves faster convergence and lower relative error than competing neurodynamic methods.
multi-agent - arxiv:2611.07616 · cs.LGMonotone Viability Learning for Offline Safe Reinforcement LearningZiqian Wang, Zhen Zhang
Offline safe reinforcement learning requires reliable long-horizon safety estimates from fixed data. Learning Hamilton-Jacobi (HJ) viability in this setting calls for semantic preservation, convergence, and numerical stability. To address these challenges, we propose Monotone Viability Learning (MoViL), which learns viable sets from offline transitions through monotone finite-horizon set recursion. The recursion preserves finite-horizon safety semantics and converges to the infinite-horizon viability kernel. Binary supervision maintains a fixed target scale across horizons, avoiding discount-induced numerical attenuation. We further establish sufficient conditions for recovering behavior-supported viable subsets. The learned viability estimates are used to certify actions at deployment. Dubins car experiments validate the numerical accuracy of the learned viable sets. Evaluations on safe RL benchmarks demonstrate state-of-the-art reward-cost performance. MoViL satisfies the average cost constraint on every task and achieves the lowest average cost among the compared methods. These results highlight MoViL's potential to extend long-horizon HJ viability analysis to safety-critical systems where reliable dynamics models are unavailable and online exploration is unsafe.
benchmark - arxiv:2611.07614 · cs.AIThe Popular Get More Popular: Auditing Brand Bias in Generative SearchGiorgio Monaco, Jacopo Lenti, Francesco Pierri
The shift from traditional to generative search has fundamentally transformed online product discovery. Because conversational recommendations synthesize choices into a few direct answers, subtle systemic preferences toward specific providers can exert profound commercial impacts. In this work, we audit this phenomenon through two distinct dimensions: (i) brand popularity bias, the systematic preference for market-dominant entities over long-tail alternatives, and (ii) brand familiarity bias, the ranking advantage an entity encoded in an LLM's parametric memory holds over an unfamiliar competitor. First, evaluating commercial queries across major engines, we observe persistent brand popularity bias: dominant brands in the top quintile are cited 1.4x to 2.8x more frequently than the average brand within a sample of 1,176 top market players, already present in parametric memory that external web retrieval fails to mitigate. Second, through controlled in-context experiments holding product attributes constant, we find that substituting authentic brand signatures with unfamiliar ones significantly degrades recommendation rank. Conversely, fact-preserving content rewrites yield negligible ranking gains. Taken together, our findings demonstrate that generative search engines reinforce a ``rich-get-richer'' dynamic by favoring established market leaders and imposing a parametric barrier to entry that standard content optimization cannot overcome.
memory - arxiv:2611.07613 · cs.CVTowards Real-Time Full-Waveform LiDAR Transformers via Intensity-Guided Token Reduction and Physics-Aware AugmentationKotaro Oishi, Kazuma Ikeda, Ryosei Hara, Ryo Yoshida +2
LiDAR is a critical sensor for autonomous driving and robotics, yet it remains vulnerable to adverse conditions such as fog, rain, transparent objects, and low-reflectance surfaces. Full-waveform LiDAR (FWL) addresses these limitations by capturing the complete return waveform as a histogram, preserving rich reflection characteristics including multi-path echoes and weak signals. While recent Transformer-based approaches have demonstrated promise on FWL data, two fundamental bottlenecks remain: high computational cost from redundant tokens in low-intensity background regions, and limited generalization due to scarce annotated training data. We address both bottlenecks by exploiting the physical structure inherent to FWL data. First, we propose FWL-ToPM, a token pruning and merging mechanism that leverages intensity distributions to selectively reduce tokens while retaining critical waveform content. Second, we introduce FWLAug, a data augmentation framework designed as a regularization method that preserves waveform structure, retaining the Time-of-Flight waveform of each ray and the frustum geometry, to improve generalization to unseen scenes. Evaluated on the Ghost-FWL benchmark, FWL-ToPM achieves over 8 times speedup at real-time throughput alongside a nearly 9-point F1 gain over the prior state of the art, establishing the first real-time FWL Transformer that simultaneously improves accuracy. FWLAug further improves performance on both the standard benchmark and an out-of-distribution test set of unseen scenes. Together, these methods push FWL-based perception toward practical deployment. The code and the OOD dataset are publicly available and can be accessed via the project page: https://keio-csg.github.io/FWL-ToPM
benchmark - arxiv:2611.07605 · cs.CLThe Gating Mechanism of KV Cache EvictionHao Gu, Xintong Yang, Zhiwei Li, Lei Zhu +9
KV cache eviction must decide which entries carry information that the rest of the cache can no longer supply. A natural way to estimate this is a learnable gate: give each cached entry a gate on its attention weight and train the gate on the model's own objective. The gate belongs inside the normalizing softmax, where suppressing one entry hands its mass to the others rather than discarding it. That same normalizer absorbs any per-query constant, so a constant gate cancels and reproduces the frozen model exactly. Trained on the language-modeling loss the gate stays near a constant score while its loss falls throughout: optimization succeeds, but a frozen backbone was fit for dense attention, so its loss prefers the no-op, and a near-constant score ranks nothing. The no-op is not a floor the gate improves from but what its objective asks for. A second obstacle survives once the gate moves, since the same gradient charges the gate for errors the frozen model already makes, which no reweighting of positions separates out. Much of the machinery in learnable eviction, from sparsity penalties to relaxations of the discrete decision, works around the first obstacle rather than removing it; both are removed by putting dense attention beyond the gate's reach and making the dense model's predictions its target. Under these two conditions capacity is never the missing ingredient: neither a larger scorer, nor explicit access to the past, nor a trainable backbone helps. The same holds for what eviction unavoidably discards, which is recovered by parameter-free centroids read inside the same softmax far better than by any memory read outside it, since the mass they must return is a share of that normalizer and available nowhere else. At 75\% KV cache compression, this gate reaches 90.14 on RULER at 8K and 46.03 on LongBench with Qwen3-8B, 3.72 and 1.36 points below full-cache inference.
memory - arxiv:2611.07599 · cs.LGBinding Paraphrased Instructions to Encountered Objects in sLLM Agents under Partial ObservabilityChanyoung Kim, Harim Song, Byoungchan Ko, Keonyoung Lee +2
Reliable deployment of LLM agents requires grounding diverse user expressions in their environments. Although RL-finetuned smaller large language model (sLLM) agents rival larger prompted LLMs on embodied tasks, their robustness to instruction paraphrases remains unclear. We introduce ALFWorld-Para to evaluate this robustness under controlled paraphrases, revealing 21-52% declines in success rates (SRs), predominantly driven by object synonyms. Despite retaining knowledge of synonym referents, agents often fail to take the corresponding objects even after observing them, exposing a gap between referential knowledge and object binding. Partial observability further complicates binding, as referents may remain unseen or have already been encountered under different labels. To address this challenge, we propose BINDIntent, a training-free framework that probabilistically binds referring expressions to observed labels using semantic proximity and estimated persistence against unseen candidates. Without additional training or LLM calls, BINDIntent improves synonym-paraphrase success of RL-finetuned sLLMs by 17.5-49.1 percentage points (pp) with only 0.4 ms of CPU overhead per step. It also raises GiGPO-7B's overall SR from 61.8% to 77.4%, surpassing a prompted 24B LLM. Code and benchmark are available at https://github.com/cau-hai-lab/ALFWorld-Para.
embodiedllm agentbenchmark - arxiv:2611.07595 · cs.AIEvolver: Turning Agent Exploration into Evolving CapabilitiesHaoyang Zhang, Qumeng Sun, Xiao Zhang, Xiang Liu +2
Agent exploration produces valuable procedural knowledge, yet much of this experience remains confined to transient execution traces and is not systematically reused. We introduce Evolver, an experience-driven framework that transforms agent exploration into persistent, reusable capabilities without updating foundation model parameters. Built around structured Gene assets, Evolver integrates experience acquisition, retrieval, and feedback-guided evolution to support experience-conditioned execution across existing agent harnesses. We evaluate Evolver using Claude Code and Codex, two foundation model configurations, and four benchmarks: Terminal-Bench 2.1, SWE-bench Multilingual, SWE-rebench, and GAIA. On a selected diagnostic set of 320 paired executions, Evolver improves task success from 83.13% to 87.50% (+4.38 percentage points) while reducing recorded task-execution token consumption by 24.24%. Furthermore, among 265 pairs successful under both conditions, token consumption decreases by 17.57%, demonstrating efficiency benefits even when task success is preserved. These findings suggest that persistent procedural experience can improve both task effectiveness and execution efficiency, providing a practical foundation for transforming agent exploration into evolving capabilities.
agentbenchmark - arxiv:2611.07594 · physics.opticsMulti-Channel Non-Abelian Edge Pumping via Multi-Degeneracy EncirclingQingyang Mo, Jiacheng Sun, Zhongfu Li, Jiajun Li +3
Non-Abelian operations enable sequence-dependent control of light, yet experimental realizations in wave systems have remained confined to local bulk or same-boundary modes and have often required fragile gap-closing degeneracies. Conventional Thouless pumps, meanwhile, support only two-channel edge transfer across a single gap. Here we demonstrate multi-channel non-Abelian edge pumping on a silicon photonic chip by adiabatically encircling multiple band degeneracies in a multi-gap lattice. In a minimal three-band band-triangle, two opposite-charge Weyl points and one zero-charge degeneracy are linked by global band connectivity, yielding both opposite-boundary and same-boundary edge channels. Sequencing these channels produces non-Abelian braiding of three topologically protected edge modes that reside on distinct physical boundaries. The process remains fully gapped and is protected by topological winding rather than by maintained spectral degeneracies. The same band-triangle principle extends systematically to an arbitrary number of channels, establishing a scalable route to collective, sequence-controlled routing of topological states in integrated photonic circuits.
silicon photonic - arxiv:2611.07593 · cs.LGTACT-KV: Tri-Axis Cosine Transform for Compressing Volumetric KV Caches in Medical VLMsChongyu Qu, Ritchie Zhao, Yufan He, Dong Yang +11
Vision-language models (VLMs) applied to 3D medical scans face a structural memory wall. A single CT or MRI volume produces tens of thousands of vision tokens, and the resulting key-value (KV) cache scales linearly with both batch and context. Existing KV compression methods operate on the cache's feature or token statistics, or apply 1D and 2D transforms, leaving the three-axis spatial structure in the cache unexploited. We propose TACT-KV (Tri-Axis Cosine Transform), a calibration-free KV compression method that exploits strong spatial correlations in visual KV states along all three anatomical axes of 3D scans. This correlation concentrates the KV cache energy under a frequency-domain transform. TACT-KV applies a separable 3D discrete cosine transform to the KV cache, allocates a per-layer bit budget across frequency bins via reverse water-filling (more bits to high-energy components), and quantizes each bin with a fixed scalar codebook. We derive a close-form relation: the distortion reduction from adaptive bit allocation is dominated by the spectral concentration of the transformed cache. Across six medical VLMs and two 3D CT benchmarks, TACT-KV achieves the lowest 1-bit prediction-flip rate on eleven of twelve model and dataset pairs in our main comparison. The same advantage reproduces on four general-purpose VLMs evaluated on video tasks, indicating that the benefit of exploiting three-axis structure generalizes beyond medical imaging. At equal high-bandwidth memory (HBM), TACT-KV scales the decoding batch size on a single GPU by up to $8\times$.
memorybenchmark - arxiv:2611.07589 · cs.LGLearning and Steering in Incomplete-Information GamesAsrın Efe Yorulmaz, Tamer Başar
In many modern multi-agent systems, a mediator seeks to influence strategic decision dynamics using only outcome-level feedback and without knowing agents' utilities. In this work, we study how a mediator can learn an unknown game and steer adaptive agents toward outcomes that maximize her utility using only trajectory feedback. The mediator knows the game form and her own objective, but not the agents' utilities. Moreover, she commits to nonnegative payments and provides recommendations and contexts, but receives only a single realized terminal outcome rather than full-information feedback. We develop halving-type algorithms that learn utility differences in normal-form games and conditional continuation-value differences in extensive-form games, achieving a sample complexity of $\widetilde O(\varepsilon^{-2})$ that matches worst-case lower bounds up to logarithmic factors. We then combine these learning procedures with a Lagrangian formulation in an online Learn-Compute-Steer algorithm that interleaves learning, equilibrium computation, and steering toward mediator-optimal extensive-form correlated equilibria (EFCE). With high probability, the Learn-Compute-Steer algorithm achieves a directness gap and an optimality gap of $\widetilde O(T^{-1/4})$, while time-average payments are of $\widetilde O(T^{-1/8})$.
multi-agentagent system - arxiv:2611.07586 · cs.LGStress Testing Chain-of-Thought Monitoring Against Covert MisalignmentYao Huang, Yifan Wang, Yitong Sun, Yichi Zhang +1
As LLM agents advance rapidly in autonomy, the risk of hidden objectives that bypass safety constraints becomes increasingly plausible. Chain-of-thought (CoT) monitoring offers a key control mechanism by reading the agent's reasoning and flagging misalignment before it materializes into action, yet its reliability depends on the realism of the red-team it is stress-tested against. Existing red-teaming efforts rely on shared templates, such as casting a side task as a prerequisite, which treat the side objective as an exogenous insertion and leave task-specific, contextually grounded evasions underexplored. We therefore ask: How robust is CoT monitoring against agents that strategically conceal misalignment? To answer this, we introduce TraceMRT, an automated Trace-level Monitor Red-Teaming framework that reframes the side task as a coherent sub-goal of the primary task and searches trace-shaping strategies whose induced trajectories complete the misaligned objective while reading as natural execution. Through black-box attacks on state-of-the-art monitors such as GPT-5.2 and Claude-4.5, we reveal substantial vulnerabilities even in top models. To counter these vulnerabilities, we propose a Bayesian Trajectory Monitoring framework, integrating global context and local segment evidence to detect covert misalignment. Empirically, our method significantly consolidates monitor robustness with 0.91 AUC and 0.74 TPR@FPR=0.05.
llm agent - arxiv:2611.07580 · cs.CVSubspace Learning with Outlier Exposure for Out-of-Distribution DetectionHongxu Shen, Kun Fang, Jiahui Zheng, Jian Sun +1
Out-of-Distribution (OoD) detection is pivotal for reliable deployments of Deep Neural Networks (DNNs), aiming to distinguish anomalous samples from the In-Distribution (InD) data in training. By introducing outliers as surrogate OoD data to finetune DNNs, Outlier Exposure (OE) has greatly advanced OoD detection in recent years. Existing OE-based research remains predominantly logit-centric, either by generating diverse outliers or by regularizing them with objectives operating at the prediction-output level. Nonetheless, these approaches leave much of the rich information in feature spaces untapped. In this work, we propose an OE-based detection framework, named RecOE, that learns appropriate feature subspaces from InD data and auxiliary outliers, such that the InD-OoD discrepancy can be effectively characterized through the learned feature subspaces. Specifically, the feature subspaces are optimized to capture the dominant variations of InD and outliers by minimizing reconstruction errors therein. Besides, a sample-level objective discourages reconstruction in the opposing subspace, while an orthogonality regularizer reduces subspace overlap, further promoting discriminative subspaces. At inference, the difference in reconstruction errors between the learned subspaces serves as the OoD detection score. Extensive experiments show the superiority of RecOE on multiple benchmarks, in which significant performance gains are achieved over recent strong baselines with only slight extra overhead from two small-size matrices of subspace bases. The source code is available at https://github.com/shenhongxu123/RecOE.
benchmark - arxiv:2611.07575 · cs.CVGeoreferenced Cross-View 3D Reconstruction from Ground and Satellite ImagesDavid G. Shatwell, Mubarak Shah
Georeferenced 3D reconstruction from ground images requires recovering not only the local scene geometry, but also localizing every camera center and 3D point in a geographic coordinate frame. Existing cross-view geo-localization methods are able to estimate the GPS coordinate of a single ground camera, but do not reconstruct dense 3D geometry, while recent feed-forward multi-view 3D reconstruction models predict point maps and camera poses only in a local coordinate frame. We introduce X-VGGT, a feed-forward cross-view geometry transformer that unifies these tasks by jointly processing a set of ground-level images and a georeferenced satellite image to predict ground-view camera poses, dense 3D point maps, and a scene-level transform parameterized by gravity, scale, yaw and translation. This transform projects the local reconstruction into the satellite plane, resulting in the geo-registration of all ground images with the georeferenced satellite image. We further propose an evaluation protocol for this joint task and show that, across multiple cross-view datasets, X-VGGT produces accurate georeferenced reconstructions and outperforms prior cross-view geo-localization and multi-view 3D reconstruction models.
evaluation protocol - arxiv:2611.07573 · cs.CVIntegrating Learned Priors through Prior-Space Reparameterization for Physically Based Inverse RenderingHun Hwangbo, Minjun Park, Yongsung Kim, Wooseok Song +3
Physically based inverse rendering decomposes appearance into materials and illumination, an ambiguous problem for which learned material predictions provide useful prior information. Most existing methods incorporate these cues through auxiliary regularization losses, which alter the optimization objective and require balancing their influence against photometric fitting. We propose a prior-guided material reparameterization that uses the predicted feature relationships to restrict the optimization space while leaving the objective unchanged. We fuse multi-view material predictions onto a mesh and construct a permutohedral lattice in the resulting prior feature space. Surface points with similar predictions share material variables, which are optimized jointly with geometry and illumination using a purely photometric objective. Experiments with two learned predictors on three synthetic benchmarks and Stanford-ORB demonstrate improved material decomposition and relighting. Our method improves relighting PSNR by 1.5-2.5 dB over the strongest evaluated mesh-based baseline on Synthetic4Relight and TensoIR-Synthetic, and outperforms a regularization loss using the same prior in controlled comparisons. We further show that our reparameterization pools noisy photometric gradients, producing updates that are more consistent across sampled views. This consistency enables a shortened joint optimization schedule that takes less than a minute while retaining relighting quality above all evaluated baselines on TensoIR-Synthetic.
benchmark - arxiv:2611.07570 · cs.LGACSeL: Adaptive Configuration Search for Hardware-Constrained LLM InferenceBokun Wang, Hung-Yueh Chiang, Diana Marculescu
Serving large language models (LLMs) under tight memory and latency service-level objectives (SLOs), including on edge devices, requires jointly selecting the model size, batch size, quantization algorithm, and mixed-precision quantization recipes. Exhaustive search over this combinatorially large space is infeasible. Unlike data-center serving, which often prioritizes throughput or generation speed, resource-constrained deployments typically set memory and latency budgets as constraints and then seek the configuration with the highest generation quality among feasible ones. To this end, we propose ACSeL, a framework that identifies configurations predicted to satisfy deployment constraints while achieving high predicted generation quality. To enable efficient quality estimation, we use Relative Model Quality Degradation (RMQD), a proxy based on the divergence between the candidate and reference models' next-token distributions over a small calibration corpus. Across the evaluated model sizes and precisions, RMQD is inexpensive to evaluate and strongly correlated with downstream benchmark performance. ACSeL combines (i) a lightweight quality surrogate that predicts RMQD directly from a configuration; (ii) Few-shot Median Calibration (FMC), which calibrates operator-profile-based latency estimates using a small number of end-to-end profiles; and (iii) a structured search algorithm BandPareto for exploring layerwise mixed-precision configurations. Experiments on an NVIDIA RTX A6000 workstation and an NVIDIA Jetson Thor edge platform show that FMC substantially improves deployment feasibility, while BandPareto + FMC achieves the highest feasibility and the lowest RMQD gaps to the best measured configurations in the evaluation set.
memorybenchmark - arxiv:2611.07560 · cs.ROWhole-Body Elasto-Geometrical Calibration of Wheeled Humanoid Robots Using a Shared-Waist Dual-Chain ModelPengxin Zha, Yi Lu, Haojun Ma, Jun Zhu +4
Wheeled humanoid robots carry both arms on one actuated waist, usually without torque sensors, so the accuracy of both end-effectors depends on a single waist model. After a first stage that estimates the tracker and target extrinsics, we identify the waist and both arms in one problem from the target positions on both flanges, measured by an optical tracker. The kinematic model is complete, with waist-base, parallel-axis, and branch parameters; arm compliances are driven by measured torques and waist compliances by gravity torques of the robot's mass model; and Gaussian priors with evidence-maximizing scales stabilize the estimate. Independent and sequential single-chain calibration are restrictions of this problem, and its waist information is the sum of the single-chain Schur complements. On two robots with different waists, each without and with a payload, calibration reduces the mean errors on held-out poses from 11.3-31.1 mm to 0.50-1.11 mm. For every training-set size from 20 poses to all, the method has 17-63% lower mean errors than adaptations of two published whole-body calibration methods, mainly through the waist compliance. Independent calibration yields two waist models; using either for both arms raises the other arm's error to up to 3.21 mm.
humanoid - arxiv:2611.07558 · cs.AISKIM: Pruning Large Language Model Agents via Selective Knowledge Informed MaskingMoonseok Choi, Jongwon Jeong, Minki Kang, Giung Nam +1
Large Language Model (LLM) agents are increasingly deployed under structured generation, where grammar constraints enforce output format at inference time. Whereas existing compression methods target unconstrained generation, we first identify a fundamental mismatch: under structured generation, grammar constraints reduce the model's output distribution to valid tokens, and the logits matter most at the small set of positions that decide the agent's action. Motivated by this finding, we propose SKIM (Selective Knowledge-Informed Masking), the first pruning and distillation framework for LLM agents under structured generation, with both stages guided by token role. SKIM partitions token positions into context, decision, and format categories, and this partition drives two complementary axes: (1) category-aware pruning saliency scores that preserve parameters most relevant for decision positions; and (2) an on-policy distillation objective that aligns student and teacher on valid tokens, paired with category-aware feature matching. We evaluate SKIM on tool-calling benchmarks and within multi-step agent loops, across multiple model scales and sparsity levels. SKIM consistently improves accuracy over baselines, with its superiority becoming most striking in high-sparsity regimes where conventional methods struggle, demonstrating the potential of category-aware compression for deployment-time grammar-constrained agents.
agentllm agentbenchmark - arxiv:2611.07555 · cs.CVMetricTrail: Revisiting the Power of Metric Depth and Local Optimization for Long-Sequence Streaming 3D ReconstructionJingnan Gao, Jiahui Huang, Sven Elflein, Jianhao Zheng +7
We present MetricTrail, a streaming system for accurate camera motion estimation and dense reconstruction from long monocular videos. To keep inference efficient over long sequences, we revisit DPVO's sparse local estimator. We find that single-image metric depth substantially improves its trajectory accuracy, without retraining the correspondence network or enlarging the optimization window. Metric predictions initialize patch depths and remain as soft constraints during optimization. We further improve correspondence estimation through learned adapters that augment matching features and recurrent context with frozen DINOv3 representations. The correspondence estimator and adapters are trained on diverse short clips without additional long-sequence training. This modular design also allows the depth provider to be independently upgraded and supports supplied camera intrinsics and calibrated stereo observations without retraining. Under the reported RGB-only protocols, MetricTrail achieves the lowest average trajectory error among evaluated methods on KITTI, Oxford Spires, and VBR. Pairing the estimated poses with single-image depth predictions also produces strong dense reconstruction on indoor and outdoor benchmarks and coherent reconstructions of in-the-wild videos. With supplied intrinsics, the system processes KITTI-02 at 12.70 FPS with 3.21 GB peak GPU memory.
benchmark - arxiv:2611.07554 · cs.LGThe Memory Trap: Language Agents Overgeneralize from Benign InteractionsKaiqu Liang, Jaime Fernández Fisac, Thomas L. Griffiths, Andrew Lampinen
Language agents increasingly use persistent memory to record user preferences and procedures across sessions. We show that memory from benign interactions can cause unsafe or unreliable behavior on later tasks. We call this failure the memory trap. To study this issue, we compare behavior in sessions with and without memory across six frontier models (including Claude Sonnet, Opus, and Fable models and GPT-5.6). We evaluate these models in a range of settings. First, in autonomy experiments, we find that requests to skip confirmation during low-stakes coding chores lead agents to skip confirmation and verification on simulated high-stakes tasks in both coding and everyday life. We next study memories formed when users seek emotional support, or describe harm from uncritical feedback, and find these memories bias later judgments of mathematical work, arguments, and poems toward sycophancy or over-critique (respectively). Memory notes written during legitimate stale-test repairs lead agents to weaken correct tests that catch broken safeguards; these memories also increase cheating on ImpossibleBench, with larger increases for stronger Sonnet and Opus models. We show that memory traps can arise both from overgeneralizations in memory writing, and when acting on memory. These findings demonstrate that even benign interactions can lead to serious risks from memory systems, and thus motivate the need for new research into ensuring that agents preserve the scope of their memories.
memorypersistent memory - arxiv:2611.07552 · cs.CLSPARE: Fast Linear Attention Decoding with Compact Recurrent StatesKe Wan, Zaiyi Zheng, Chen Chen
Linear attention reduces the quadratic sequence-length complexity of softmax attention by maintaining fixed-size recurrent states. However, these states scale quadratically with the representation dimension and require repeated data transfers during decoding, creating a substantial GPU memory bottleneck. Recent studies have revealed significant low-rank redundancy in recurrent states, suggesting opportunities for more efficient representations. We introduce SPARE (State Projection for Accelerated Recurrent Execution), a training-free method that exploits this redundancy through fixed low-dimensional subspaces, heterogeneous rank allocation, and lightweight write compensation. By performing recurrent reads and updates directly on compact states, SPARE reduces memory traffic without online decomposition or dense-state reconstruction. SPARE achieves 2.55x full-model decoding speedup, 7.49x lower recurrent-kernel DRAM traffic, and 3.31x lower peak GPU memory, while preserving better associative-retrieval accuracy than structured pruning at matched memory budgets.
memory - arxiv:2611.07542 · cs.AIFlashOffloading: Low-Latency KV Cache Offloading for Long-Context LLM ServingXuan Wang, Zhuanhao Wu, Tingzhu Fang, Xushi Li +16
In large language model (LLM) serving systems, the key-value (KV) cache grows linearly with sequence length and concurrency, so long-context inference easily hits the high-bandwidth memory (HBM) capacity wall and results in low concurrency. Offloading the KV cache to host dynamic random-access memory (DRAM) frees HBM space and raises concurrency, but the additional cache management and host-to-device (H2D) load overhead inflates inference latency, higher concurrency fails to translate into effective throughput. We present FlashOffloading, a host-DRAM-centric inference architecture that fully offloads the KV cache in both the prefill and decode phases. FlashOffloading builds rolling HBM cache for prefill, a hot cache for decode, a reverse sparse index with zero extra miss-lookup overhead, and a compute--load overlapped kernel, optimizing KV offloading latency across the entire pipeline and delivering throughput gains under a time-per-output-token (TPOT) service-level objective (SLO). In experiments deploying GLM-5.2 with multi-token prediction (MTP3) speculative decoding on an Ascend910C prefill--decode (PD)-disaggregated platform, FlashOffloading improves prefill throughput by up to 1.38\(\times\) and decode throughput by up to 1.91\(\times\) under a 30 ms TPOT constraint. FlashOffloading is integrated into an open-source serving framework: https://github.com/vllm-project/vllm-ascend/tree/releases/v0.30.0.
memorylong-context - arxiv:2611.07541 · cs.AICST-Bench: Benchmarking Counterfactual Spatial Transition Reasoning with Causal Transition On-Policy Self-DistillationYing Li, Yongle Cheng, Jing Li, Xingquan Cai +1
Spatial reasoning is a fundamental capability of multimodal large language models (MLLMs), yet existing benchmarks mainly focus on static scene understanding. When objects are moved or swapped, models must infer not only spatial relations but also how interventions change the scene state. To address this problem, we introduce CST-Bench, a benchmark for counterfactual spatial reasoning. CST-Bench contains 7,082 examples covering 6 intervention types and 12 transformation operators, with dense causal annotations including intervention sources, affected relations, expected effects, invariant relations, and metric changes. We benchmark general-purpose MLLMs on CST-Bench and further compare post-training strategies on CST-Bench and standard spatial benchmarks. Results show that methods effective for ordinary spatial reasoning can still fail under counterfactual interventions. We further formulate counterfactual spatial reasoning as causal state-transition modeling and propose CT-OPSD, which uses CST-Bench annotations to guide a teacher model in learning counterfactual state transitions and causal reasoning chains. These chains are then distilled into a student model that reasons only from real observations. Overall, experiment results establish CST-Bench as a challenging benchmark and reusable supervision resource for counterfactual spatial reasoning. CT-OPSD further provides an initial demonstration of how its annotations can support model improvement.
post-trainingbenchmark - arxiv:2611.07540 · cs.CVAction-Conditioned VLM Monitoring for Autonomous Vehicle Semantic AnomaliesDaniel Adejumo, Omkar Ankalkope, Kunal Runwal, Aliasghar Arab
We study action-conditioned monitoring of autonomous-vehicle behavior in semantically unusual driving scenes. The target is whether the executed ego action is appropriate for the scene: the same scene can admit both normal and anomalous behavior, so visual unusualness alone does not determine the monitoring label. A Vision-Language Model (VLM) first predicts the scene-required action from an outcome-hidden video window; the monitor then compares that expectation with the realized ego-motion trajectory under an explicit compatibility rule. To study this monitoring task, we contribute a six-field taxonomy of 112 paired scenarios spanning both failure polarities (42 derived from documented incidents), a generated video corpus with estimated trajectories, and an exploratory two-stage monitor on that corpus. On 235 clips from 15 scenario families, the two-stage protocol with a fixed T-2.5 s window obtains 83.0% accuracy and anomaly recall 0.773, versus 74.5% and 0.536 for a single-stage video-only baseline; specificity is 0.880 versus 0.928.
action-conditioned - arxiv:2611.07537 · cs.ROUnified Overview of Safety Certificate Functions: Techniques from Hamilton--Jacobi Reachability and Control Barrier FunctionsHao Wang, Max Cohen, Zeyuan Feng, Donggeon David Oh +8
Safety is a central concern in modern autonomous systems. Safety certificates---functions whose existence guarantees that a system meets its safety specification---have long been used as a principled tool for rigorously reasoning about safety. Methods for synthesizing such certificates are scattered across research areas, with little guidance on when to use a particular approach. This tutorial provides a holistic treatment of safety certificate synthesis centered on two widely used families: Hamilton--Jacobi (HJ) reachability analysis and control barrier functions (CBFs). For each family of methods, we first introduce its general formulation, along with classical and learning-based tools for certificate synthesis. Next, we introduce two benchmark problems consisting of vehicle systems of differing complexity to assess the certificate quality of different synthesis methods and their effectiveness in safety-critical control. We use these benchmarks to compare various methods and provide practical guidance for selecting an approach.
benchmark - arxiv:2611.07532 · cs.ROSemantics-to-Contact: A Stagewise Framework for Robust Contact-Rich ManipulationGuanren Qiao, Ruixiang Ouyang, Sheng Xu, Yueci Deng +3
Contact-rich manipulation requires precise control under complex contact dynamics while remaining robust to diverse visual conditions. However, real-world reinforcement learning often overfits to the visual appearance of training scenes. We propose Semantics-to-Contact (S2C), a stagewise framework for visually robust contact-rich manipulation. The insight is to decompose the problem into two stages: semantic focusing and contact refinement. In the first stage, a vision-language model localizes task-relevant regions across visually diverse scenes, providing coarse spatial guidance and reducing the burden of exploration. In the second stage, a real-world human-in-the-loop residual reinforcement learning policy learns fine-grained contact behaviors within the localized region using force feedback. To further improve visual robustness, we introduce an object-centric visual augmentation strategy that randomizes background appearance while preserving the robot and task-relevant objects. Experiments on six contact-rich real-world manipulation tasks exhibit that S2C consistently improves success rates over strong baselines and maintains reliable performance under significant visual and positional variations. These results demonstrate that stagewise semantic focusing and contact refinement provide a practical path toward visually robust real-world contact-rich manipulation. Our demos can be seen in https://edem-ai.github.io/S2C/.
manipulationhuman-in-the-loop - arxiv:2611.07528 · cs.AIGoal-Branching LaCAM: Temporary Goals as Search Branches in Dense Multi-Agent PathfindingStephen Chang, Daniel Harabor
LaCAM is a search-based solver for multi-agent pathfinding (MAPF) that runs a depth-first search over joint configurations, using PIBT to generate successors and lazily adding single-agent constraints to explore alternatives. Although LaCAM is complete, it can struggle on dense, corridor-heavy instances because extracting blocking agents from deadlocks requires coordinated moves of several agents, often away from their goals, which PIBT's greedy one-step successors rarely propose and which LaCAM can only reach by enumerating constraints one agent at a time. This paper proposes Goal-Branching, where blocking agents are given temporary goals, each tried as an additional branch of LaCAM's search. We examine random and corridor-aware rules for choosing these temporary goals and show that completeness is preserved for any such rule. On dense instances (60-90% of free cells occupied) of three maze and room maps, Goal-Branching solves 89% of instances against 44% for LaCAM*, and, on instances both solve, it finds a first solution 10-14x faster on maze-32-32-2 and room-32-32-4, and nearly 200x faster in median time at 75% density on room-32-32-4. Solution cost increases by 15-22% on instances both solve, while on open maps and the full MovingAI benchmark its solutions cost no more than those of LaCAM* and it is at most 3% slower.
agentmulti-agentbenchmark - arxiv:2611.07517 · cs.ROEmbodied RAG++: Retrieval-Augmented Generation From What Exists to What Works in Embodied ScenesFan Yang, Tatsuya Kikuzuki, Jun Liang, Jun Liu +6
To serve a request in a human environment, a robot system must find an object that satisfies a functional requirement and a robot that can feasibly reach it. Scene graphs record which objects exist and where, but they often omit capabilities documented only in product manuals. Retrieval-augmented generation methods can read such manuals, yet they rarely reason jointly over object capabilities, physical instances, robot states, and dynamic connectivity. We introduce E2W-Bench, the Exists-to-Works Benchmark, which links dynamic hybrid scene graphs with real product manuals to evaluate capability-grounded robot--object assignment under availability, mobility, vertical-reach, and connector constraints. We also present Embodied RAG++, a spatially grounded retrieval framework that scopes retrieval to the manuals linked to candidate objects, caches extracted capabilities per product model, and resolves robot--object feasibility through deterministic spatial reasoning. On E2W-Bench, Embodied RAG++ reaches a pair accuracy of 0.90 with Qwen2.5-32B and 1.00 with Llama3.3-70B, compared with at most 0.30 and 0.55 for the baselines, while requiring the shortest preprocessing time among the retrieval-based methods and tying for the lowest total token count.
embodiedretrieval-augmentedscene graphbenchmark - arxiv:2611.07516 · cs.ROTraining-Independent Validation of a Learning-Based Controller for Stunt ManeuversKiruthiga Chandra Shekar, Aliasghar Arab, Jingang Yi
Learned controllers for vehicle maneuvers are usually validated under the conditions and objective used to train them, where they appear highly reliable, yet the parameters that change on real roads are often not those the model was trained on. We argue that validation must be independent of training: it should vary physical parameters that training kept fixed and use acceptance criteria absent from the training objective. A learning-based control policy, G-Maneuver, with 90.11% success is tested for stunt-level maneuvers. A looser terminal-heading score ranked the worst setting as the best. Because real stunt tests are safety-critical, a simulation-to-reality-to-simulation (Sim2Real2Sim) loop uses the simulation to select safe zones for real-vehicle trials with 16.9% acceptance. Automotive safety analysis frameworks are used to identify parameters for validation using coverage tests. We demonstrated that the kinematic stunt-driving indicators, namely heading change, sideslip, and speed-normalized yaw rate, match between simulation and the real world, while force and roll indicators do not, which identifies the indicators that can support safe deployment. A video of the experiments is available at https://youtu.be/dN1Oes8pNIo.
sim2real - arxiv:2611.07515 · cs.ROLMGN: A Unified Policy Learning Model for Last-Mile Grasp-NavigationXuenmin Hu, Qingxue Qian, Wei Luo, Benyi Ma +2
Legged robots equipped with manipulators have shown significant potential in the fields of daily life and industry in complex environments. However, existing methods mainly focus on global navigation or manipulation independently, while overlooking the last-mile navigation problem, where the robot must jointly determine the base placement, approach direction, and grasp execution to achieve reliable object grasping, and different manipulation stages require distinct control behaviors. To address these challenges, we propose Last-Mile Grasp-Navigation (LMGN), a unified policy learning model for Quadruped robot loco-manipulation. An Explicit Distance and Directional Reachability (ED-DR) representation is first introduced to explicitly represent the features of grasp-compatible base placement and approach direction. We then develop a semantically aligned Hierarchical Mixture-of-Experts (HMoE) policy that models stage-aware behaviors across the stages of base adjustment, pre-grasp alignment and grasp execution, and object lift. Finally, a new optimization method that jointly aligns navigation and manipulation, stage-dependent expert routing, and action-level consistency is proposed to optimize the HMoE policy. Extensive experiments on the Habitat and Isaac Gym simulators demonstrate that LMGN consistently outperforms state-of-the-art methods, denoting that integrating grasp-aware reachability with stage-aware policy learning significantly improves the success rates of last-mile navigation and target grasp for robots.
manipulationmanipulatorquadrupedgrasp - arxiv:2611.07513 · cs.ROConcertina: Embodying Manipulation and Cognition in One Diffusion Vision-Language ModelChengcheng Wang, Wei He, Wenhao Li, Hongguang Li +3
Most vision-language-action (VLA) policies decouple manipulation from cognition by bolting an external action mechanism onto a vision-language model (VLM)--either quantizing actions into discrete tokens whose errors bear no stable correspondence to physical displacements, or delegating continuous trajectory generation to a separate action expert whose gradients compete with or are blocked from updating the VLM backbone. We argue that manipulation and cognition should instead be embodied in the same model, shifting from a decoupled VLA to an Embodied Vision-Language Model (eVLM). We introduce Concertina, an eVLM that places continuous trajectory representation, blockwise generation, and physical-coordinate supervision directly on the computation and parameter path of a shared diffusion VLM backbone. Each control timestep is encoded as one continuous latent action token in the backbone hidden space, and the shared Transformer reconstructs temporally ordered action blocks via the same corrupt-and-reconstruct operator used for multimodal pretraining. Active action projections account for only 0.21% of parameters in the 4.46B model with zero separate action-generator layers, compared with 16.5%-35.1% dedicated action-expert capacity in representative continuous VLAs. Using our proposed Action Interface Audit, we confirm that cross-time trajectory coupling emerges in the middle and late layers of the shared backbone rather than at the action interfaces. On standard benchmarks, Concertina achieves 98.15% average success on regular LIBERO, 85.9% on the 10,030-task LIBERO-Plus benchmark, and 4.142 average completed chain length on CALVIN ABC-D. Code: https://github.com/lose4578/Concertina Hugging Face: https://huggingface.co/collections/Traly/concertina
vision-language-actionvlaembodiedmanipulationliberobenchmark - arxiv:2611.07508 · cs.AIDelegateEval: Evaluating Platforms for Personal External Service AgentsJichao Li, Jinrong Lu, Manjiang Yu, Zhixuan Liu
AI agents that work for an individual over time may also help that individual serve external clients. We define a personal external service agent as an agent authorized and continuously managed by a specific individual to accept client requests, deliver work, and provide ongoing support using that person's knowledge, experience, professional methods, and service boundaries. Such agents could allow individuals to serve more clients with less direct involvement and generate economic value at scale. We propose DelegateEval, a platform-level evaluation framework for agent systems that may combine large language models, reusable skills, protocols such as MCP, and executable tool environments. It covers eight dimensions: ease of creation, service discovery and matching, delivery quality, expertise inheritance, service continuity, learning and improvement, reliability and governance, and efficiency and value. These dimensions assess whether a platform enables people with different backgrounds to turn their expertise into ongoing services that deliver verifiable work, develop capabilities through experience, and remain economically feasible. We present the framework and an experimental design. Complete datasets, executable environments, a unified scoring protocol, and cross-model results remain to be developed and empirically validated.
agentai agentagent systemevaluation framework - arxiv:2611.07498 · cs.CVDegree-Equalized View-Graph Selection under Tight Pair BudgetsMinhyeok Im, Jongseong Brad Choi
Fully matching all image pairs in Structure-from-Motion is quadratic in the number of images, so practical pipelines pass a budgeted subset to the final matching and reconstruction backend. One-shot selectors score candidates with image retrieval, cheap geometric probes, or a learned prior and finalize B pairs before full matching. Their priorities do not account for the degree already allocated to each image, which can concentrate edges on a few images. Across nine large 1DSfM internet photo collections at B=2N, the coverage top-B selector leaves 54% of covered images with realized degree at most two. Reconstruction can then fail even when the selected graph is nearly connected and almost fully verified. Our training-free greedy rule, Degree-Equalized Allocation (DEAL), exponentially downweights pairs whose endpoints are already well covered. On these collections, DEAL improves the one-shot selector built on the same prior (from 0.776 to 0.929) and matches the mean registered-image ratio of an SfM-supervised edge prior under the shared backend (0.929 vs 0.930), without training. On the largely saturated 38-scene standard benchmark suite, DEAL gives the highest mean among the evaluated training-free selectors (0.875 vs 0.858 one-shot).
benchmark - arxiv:2611.07496 · cs.LGZYFA: Speed and Accuracy of FP8 FlashAttention on AMD CDNA4Quentin Anthony, Ganesh Bikshandi, Praneeth Medepalli, Stephen Ebert
Attention is a core building block of frontier models and often dominates compute-bound prefill at long context. FP8 matrix instructions offer twice the peak throughput of BF16, but they accelerate only matrix multiplication, and the softmax between the two matrix multiplications of attention runs on vector units that FP8 does not accelerate. On the AMD Instinct MI355X, the vendor's AITER kernels reach 39% of peak FP8 matrix throughput at 32K tokens, compared with 49% in BF16. Attention is also more numerically sensitive to quantization than feed-forward layers, and unlike NVIDIA's Hopper and Blackwell GPUs, the MI355X does not track the completion of matrix instructions in hardware, so overlapping the softmax with the matrix multiplications depends on instruction order. We present ZYFA, prefill kernels for the MI355X built on three techniques: (1) computing each FP8 probability directly from its score with a multiply-add and an integer conversion instead of an exponential; (2) a ping-pong schedule in which one group of threads computes its softmax while another executes its matrix multiplications; and (3) scaling each block of probabilities up by a power of two before rounding, and normalizing by the sum of the rounded probabilities. On attention inputs captured from three models and timed at sustained clock speeds, ZYFA achieves 1.16-1.76x the throughput of AITER. On the Gaussian inputs of kernel benchmarks at 32K tokens, it achieves 1.26-1.37x at head dimension 128 and 1.99x for DeepSeek-style multi-head latent attention. The sustained clock of the MI355X depends on operand values, so throughput on random inputs does not reliably predict throughput on real model activations. At head dimension 128, ZYFA's numerical error is within a few percent of that of Composable Kernel and Triton and 62-78% lower than that of AITER's. The source code is available at https://github.com/Zyphra/zyfa.
long contextbenchmark - arxiv:2611.07495 · cs.LGLearning Multi-Agent Orchestration:Higher-Order Interactions in Policy ImprovementKaixuan Liu, Guojun Xiong
Learned orchestration has emerged as a promising paradigm for large language model (LLM)-based multi-agent systems, in which a central orchestrator selects agents and controls the information they receive throughout a task. Despite rapid empirical progress, existing methods train the orchestrator by crediting each decision with its own advantage, implicitly assuming that the other decisions in the episode stay fixed. This assumption fails in orchestration, where the contribution of each decision to the final outcome depends on the decisions preceding it, so the gain of a joint update is not the sum of its per-decision effects. To address this challenge, we introduce \textbf{IAPO}, an interaction-aware policy optimization framework that explicitly accounts for interactions among decisions. Specifically, IAPO exactly decomposes the performance difference between two policies into the effects of individual decisions and their interactions, which we prove can account for half of the attainable gain, and then optimizes both from sampled trajectories while certifying that every update is an improvement. On long-input versions of GSM8K, MATH, DROP, MBPP and HumanEval, against fifteen published multi-agent and workflow-search systems on the same 7B executor, IAPO's deployed plans reach the strongest system's accuracy with fewer tokens, and removing the interaction term costs 7 to 17 points of accuracy on code.
multi-agentagent system - arxiv:2611.07493 · cs.ROFast and Reliable Open-Vocabulary Object Search with Absence Detection from Robot MemoryNicholas Paul, Devson Butani, Eric Martinson
Mobile manipulators deployed in everyday human environments, such as homes for assistive care or flexible automation, often need to locate specific objects on demand from accumulated observations. Equally important, they must reliably determine when a target is absent before committing to navigation or interaction. Existing approaches demand either slow exhaustive re-processing or expensive 3D representations, and are evaluated on recall and accuracy rather than specificity, leaving their ability to abstain on missing targets unmeasured. We find existing systems that search using text or visual embedding almost never abstain on absent objects achieving at most 9\% specificity. To support accurate search at speed, we present a two-stage architecture that decouples fast candidate retrieval from verified selection, enabling both high recall and reliable absence detection from sequential RGB observations with pose. We evaluate eight retrieval-selection configurations plus an exhaustive search baseline across ScanNet apartment scenes, including significant numbers of absent objects. Adding a visual verification step to caption-based retrieval matches or exceeds exhaustive search reliability while requiring an order of magnitude fewer detector calls. We reproduce ScanNet scores on a physical robot operating entirely on-board with a single consumer GPU running in real time.
manipulatormemory - arxiv:2611.07488 · cs.LGOPRD-Imp: a dataset of experimentally observed impurities and side products in multi-step pharmaceutical synthesesHeejeong Kim, Seonbin Kim, Jihwan Kim, Tim Pinkhassik +5
Predicting reaction impurities is a critical yet persistently data-limited challenge in pharmaceutical process chemistry. While machine learning models for reaction outcome prediction have advanced rapidly, benchmarking their ability to predict impurities remains impractical due to the absence of curated, publicly accessible datasets. We present OPRD-Imp, a dataset of 229 organic reactions drawn from 62 publications in Organic Process Research & Development, each annotated with experimentally observed impurities and their mechanistic origins. A total of 784 impurity entries is provided, classified into four origin categories. Reactions are represented as reaction SMILES and impurities as individual SMILES strings, with multi-step synthetic route connectivity preserved. All data were manually extracted from the source publications by trained chemists following a standardized extraction protocol, with expert chemical validation and manual structure drawing. OPRD-Imp is intended to serve as a benchmark resource for evaluating models that predict reaction impurities, providing experimentally reported impurity annotations and mechanistic origin labels that are not systematically represented in conventional reaction datasets.
benchmark - arxiv:2611.07486 · cs.CLMEvince: Knowledge-Constrained LLM Reasoning for Technique Identification from Malware ReportsHongyu She, Yiling He, Mengda Xie, Yiping Liu +3
Mapping behavioral intelligence in malware analysis reports to ATT&CK techniques supports threat analysis and defense, but reliable identification remains challenging. Supporting evidence is often scattered across long, heterogeneous reports, while closely related techniques share behavioral cues and differ only in subtle but decisive details. While large language models (LLMs) can use external knowledge to interpret reports, they may assign techniques based on plausible associations rather than sufficient evidence. We present MEvince, a knowledge-constrained LLM reasoning framework for report-level malware technique identification. Its core is a shared behavioral representation connecting report observations to technique definitions. This representation is built through agentic mining of technique documentation and supports two online stages: 1) primitive-guided extraction uses behavioral primitives to structure LLM-extracted observations, link them to source passages, and group dispersed evidence; 2) evidence-aligned retrieval matches this evidence to technique profiles through the same concepts, retaining likely candidates and confusable alternatives for LLM comparison of distinguishing behaviors. Controls at both stages validate evidence references and check technique-specific requirements. The representation also supports adaptation without parameter updates, where offline agentic audits diagnose concept weaknesses and refine extraction or retrieval policies. We evaluate MEvince against 11 baselines spanning four method families using a dataset of 610 public malware reports referenced by MITRE ATT&CK. MEvince improves exact-ID F1 over the strongest baseline by 17.4%, and its concept-guided retrieval increases gold-label candidate coverage from 31.6% to 47.9%. With a small adaptation set, offline refinement yields further gains and outperforms baselines that update model parameters.
agentic - arxiv:2611.07483 · cs.AISigned but Not Safe: Requirement Integrity in Multi-Agent LLM HandoffsXiaobo Chen, Genliang Zhu, Ziyuan Wang, Chu Wang
Enterprise AI agents increasingly split a user's request across subagents and merge their results, and each handoff can summarize away, alter or ignore a requirement. We report a preregistered study with synthetic, exactly scorable tasks comparing free-text, structured JSON and signed requirement-manifest handoffs: the orchestrator extracts each requirement as a verbatim quote plus a checkable predicate, code signs the manifest with HMAC-SHA256, and deterministic code verifies the merger's decision. With Qwen3-8B (1,440 trajectories, five threats), signed manifests cut unsafe commits under attack from 69.4% to 9.0% (60.4 pp, 95% CI [51.4, 69.4]) with no net loss of conformant commits relative to structured handoffs. Signing is what defeats in-transit tampering: an unsigned manifest committed violations in 79% of mutated runs, the signed one in none. On identical trajectories the deterministic verifier caught 78.7% of violating decisions versus 59.0% for an LLM checker, with 7% versus 50% false alarms. An oracle manifest removed all remaining unsafe commits: the residual risk comes from extraction errors made before signing. A preregistered pre-signing injection attack confirmed this boundary: one adversarial line raised signed-manifest violations from 14.6% to 37.5% (Qwen3-8B), 11.1% to 63.9% (gpt-oss-20b) and 8.3% to 75.0% (GPT-5.6-luna), and the signature then authenticated the corrupted manifest. Signing protects requirements in transit, not at extraction; the extraction step is the trust boundary.
ai agentmulti-agent - arxiv:2611.07482 · cs.AISAVeR$^2$: Reasoning-based Safety Alignment for Large Reasoning Models via Verifiable RewardsYichen Sun, Qirui Liu, Yan Wang, Jianan Lin +4
Large Reasoning Models (LRMs) have achieved impressive performance via explicit Chain-of-Thought (CoT) reasoning, yet this process introduces critical safety risks. We formalize LRM safety objectives as requiring non-malicious reasoning and answers, safety consistency, and explicit intent identification. We argue that existing alignment methods suffer from poor adversarial generalization, a significant ``safety tax'' on reasoning, and unfaithful coarse-grained rewards, primarily due to their neglect of the LRM's intrinsic reasoning capabilities for safety. To address these, we propose \textbf{SAVeR$^2$}, a novel \textbf{S}afety \textbf{A}lignment method via \textbf{R}ewards with \textbf{R}easoning capability. SAVeR$^2$ decomposes the safety objective into four fine-grained verifiable reward signals and utilizes a difficulty-aware data selection strategy to stabilize training. Crucially, we introduce a reasoning-preserving gradient projection mechanism that analytically resolves conflicts between safety and reasoning gradients, effectively eliminating the safety tax. Extensive experiments on models ranging from 1.5B to 671B demonstrate that SAVeR$^2$ achieves superior safety performance and significantly lower over-refusal rates across multiple benchmarks without compromising general reasoning capabilities.
benchmark - arxiv:2611.07481 · cs.CVCompact Motion Embeddings for Feed-forward Unified Human AnimationSeonwook Park, Hyunsoo Cha, Amrita Mazumdar, Shengze Wang +4
Animating a reference image with a driver's motion requires a motion representation that captures face, hand, and body articulation but excludes the driver's appearance. Most body animation methods take motion from external pose estimators and render it with many denoising steps; compact learned motion representations mostly cover the face. We present FUHA, a real-time-capable feed-forward transformer that learns a compact motion embedding of the whole human and uses it for face, upper-body, and full-body animation. A Q-Former compresses the driver's features from a frozen visual encoder into this embedding. To disentangle motion from appearance, a novel contrastive loss pulls it toward the motion embedding of the driver's rendered keypoint image and away from that of the same person in another pose. At inference time, FUHA needs no external pose estimators, keypoints, or parametric body models. As a retrieval key, the motion embedding matches body pose better than the learned global embeddings of ten vision encoders. Its same-person retrieval rate is 0.1%, against 9.5% for its own DINOv3 backbone. On our upper-body benchmark, FUHA outperforms all baselines on every self-reenactment metric and on cross-reenactment identity similarity. It is competitive on face and full-body and runs at up to 38.9 FPS at 512x512 on a single RTX 4090.
benchmark - arxiv:2611.07470 · cs.ROKinematic analysis of a double-closed-loop four-bar chain-based leg for a multimodal quadruped-pentarotor robotPrashant Raturi, Santhakumar Mohan
This paper presents the kinematic modeling and singularity analysis of a three-degree-of-freedom, double four-bar closed-loop quadruped leg mechanism. This leg is essential for transitioning between the two operational modes; therefore, it is examined in both joint and task spaces, leading to the development of its Jacobian matrix, which is vital for designing the gaits required for terrestrial mode and mid-air aerial reconfiguration (for aerial mode). Various gaits are devised for the system's motion in the MATLAB simulation, and their practical implementation on the single-leg test rig is demonstrated.
quadruped - arxiv:2611.07469 · cs.CLRecursive Self-Improvement for Long-Horizon Role-Playing DialogueY. L. Li, Y. H. Wu
Long-horizon role-playing dialogue requires consistent character identity, reliable memory, and context-sensitive interaction. We present a bounded recursive self-improvement (RSI) framework that revises original-character memory, stable persona instructions, and a policy for constructing turn-level prompts. Repair outcomes also inform revisions to the updater's diagnostic rules, edit routing, and candidate selection. We specify the updater as a versioned policy with explicit inputs, editable fields, acceptance gates, and stopping conditions. An internal evaluation covers 50 original characters and five models, with two 100-round conversations per character-model pair: 500 conversations per condition. Across the pooled evaluated replies, the bad-response rate decreases from 14.00% (6,973/49,824) to 4.94% (2,459/49,793), a 9.06-percentage-point reduction and a 64.71% relative reduction. In 600 blinded human comparisons, RSI receives 294 wins, 187 ties, and 119 losses, yielding a 71.19% win rate among decisive judgments. The results support improved aggregate dialogue quality for the complete framework on the internal test distribution.
self-improvement - arxiv:2611.07464 · cs.AIFINCOSMOS: An Open-World Evaluation Paradigm for Agents in FinanceYupeng Cao, Xueqing Peng, Lingfei Qian, Haohang Li +13
Language agents increasingly work for users on the open Web, and finance is where trusting them is hardest. The data must come from the Web, the strategy lives in the user's head, and correctness depends on time, so an agent can fail silently: a strategy the user never asked for, unverified data, or a backtest that uses the future yields a report as convincing as a correct one. Existing benchmarks close the world somewhere, fixing the task, the data, or the timeline, and thereby remove these failures instead of measuring them. We introduce FINCOSMOS, an open-world evaluation paradigm for agents doing financial work. A simulated user holds a hidden strategy sampled from a compositional grammar and reveals it only when asked. The agent receives no data, so it must acquire data from the Web, build its own backtest under a data cutoff, and deliver verifiable results. Grading needs no reference answer. Programs verify point-in-time correctness and recompute every reported metric, the delivered strategy is checked against the hidden specification, and a run is accepted only when the user is satisfied and none of 15 veto checks fails. Across four coding agents and a model without tools in 150 sessions, no session is accepted: every session that the simulated user signed off fails a program check, and agents raise only 54--64\% of the details they were not told. Human experts attribute 88.33\% of failures to the agents. With FINCOSMOS feedback, an agent with persistent memory learned all ten tasks.
memorypersistent memoryagentbenchmark - arxiv:2611.07461 · cs.LGThe Last Mile to the Real World: Agent-Native Primitives as Reusable Building Blocks for Solving Real-World TasksHaibo Jin, Peng Kuang, Xiaopeng Yuan, Lucheng Fu +10
How can an agent handle a real-world task end to end? Frontier models with their own harnesses still fall short on real-world benchmarks: on Terminal-Bench 4.0, GPT-5.6 Sol with Codex resolves 37.3% of the tasks and Claude Opus 5 with Claude Code 51.8%. These tasks fail at the step from reaching a resource to applying it, which we call the last mile to the real world. Existing approaches put the resource within reach but leave the competence to apply it with the caller, so each harness rebuilds that competence. We introduce agent-native primitives, reusable units that equip a resource with a resident LLM that applies it in context and communicates with other primitives, and HEELO (Harness Engineering around Evolving Agent-Native Primitives for the Last Mile to Real-WOrld Tasks), a harness engineering framework that builds each task's harness from a library of such primitives. HEELO composes the primitives a task needs into a task-specific graph, verifies the result against evidence from the environment, revises the primitives a failure implicates, and admits new or improved primitives after validation. We instantiate HEELO with eight primitive families, each built for one kind of task, and evaluate them on more than twenty benchmarks. Dev-Primitives raise Terminal-Bench 4.0 resolution from 37.3% under Codex to 55.5% on the same backbone, Harness Primitives composed per task raise Terminal-Bench 2.0 Pass@1 from 60.0% for the seed harness to 72.2%, Tool Primitives raise ToolBench pass rate from 55.8% to 75.1%, and Navigation Primitives raise GAIA-Text-103 accuracy from 45.6% to 58.3%. On the 46 real-world tasks of our heelo.dev platform, Claude Opus 5.5 with Claude Code and GPT-6.1 Sol with Codex complete 9 each on their own and 39 and 35 when connected to HEELO through our MCP server, with 1.9$\times$ and 2.3$\times$ less time and 1.55$\times$ lower cost per completed task.
agentbenchmark - arxiv:2611.07453 · cs.LGLatency-Aware Adaptive Best-of-$N$Alon Jacobson, Qian Xie, Peter I. Frazier
Test-time compute (TTC) improves language model responses by spending additional computation at inference time, and the right amount of compute depends on the prompt. Adaptive TTC methods allocate compute based on signals of response quality such as observed rewards, often by stopping once the signal exceeds a fixed threshold, and generally measure cost by the number of generations or tokens. But in online serving, latency is also a cost: it grows with server load, so a method using a fixed threshold spends too little compute when generation is fast and too much when it is slow. For Best-of-$N$ sampling, a widely used TTC method, we propose a two-stage adaptive algorithm that draws an initial batch, then uses the observed rewards and the batch's latency to decide whether to draw a second batch and at what size by comparing expected improvement against predicted cost: it draws one when the first batch's best reward is below a threshold that decreases with the first batch's latency, without explicitly measuring server load. We map each reward to its value under the cumulative distribution function (CDF) of the prompt's reward distribution, so the same tradeoff parameters can be reused across tasks; a regression fitted offline on other prompts predicts a distribution for the CDF value of the first batch's best reward, giving the expected improvement in closed form. We evaluate the method by simulation on a preference-based alignment benchmark, measuring generations-equivalent cost: generations plus latency, converted to generations. The method reaches Best-of-$N$'s reward at lower generations-equivalent cost, and conditioning on the measured first-batch latency saves at least as much of that cost as fixed-threshold (latency-unaware) baselines, and more when latency is highly variable, when it grows with batch size, and when the second batch's latency is highly predictable from the first's.
benchmark - arxiv:2611.07444 · cs.LGTwo-stage residual learning for calibrated uncertainty in solar irradiance forecastsRR Flanagan, E Kasha, PW Talbot, AS Epiny +1
Accurate forecasting of renewable generation requires not only point predictions but also reliable estimates of the uncertainty attached to them, particularly at extended horizons. Machine learning models are increasingly used to predict solar output, but they typically report a single value and capture the associated uncertainty poorly. The standard remedy is to train a model to predict the parameters of an assumed probability distribution directly. Here we develop an alternative two-stage method, motivated by synthetic solar signal generation, in which the forecast error is modeled separately from the expected irradiance. A first long short-term memory (LSTM) network predicts the expected global horizontal irradiance, and a second network models the distribution of the forecast error, which is added back to the mean to give both the forecast and its uncertainty. Using historical weather data from the National Solar Radiation Database and a probability integral transform evaluation, we show that explicitly modeling the error brings the predicted standard deviation markedly closer to the historical value, reducing the deviation between the two by 34.6 percent for a standalone LSTM and by 56.5 percent when the LSTM is coupled to Gaussian process regression, the most reliable of the configurations tested. Because the method models the total forecast error directly, it captures the full predictive uncertainty that matters for decisions, rather than only the model contribution that dropout-based approaches estimate.
memory - arxiv:2611.07441 · cs.ROClosed-Loop Physics-Based Manipulation Evaluation at a Commodity Price: What 4.4 L4 GPU hours at $3.50 Buys YouRohun Kshirsagar, Diana Rypkema
Running evaluations of contact-rich robotic manipulation at statistical scale has largely been possible only at well-resourced labs, highlighting the need for affordable and efficient options for broader audiences. We present a reproducible evaluation harness wrapper based on a previous study at 200-rollout scale on serverless commodity GPUs, at a cost of approximately $3.50 USD and 38 minutes of wall-clock per replication (4.4 L4-GPU-hours across eight parallel shards). Using the harness, we report a partially controlled negative result on action chunking with sparse data: three independently trained chunked checkpoints produce zero genuine successes across 780 rollouts, while three single-step checkpoints range from 2.6% to 11.8% success rate. A per-step-replanning control partially isolates chunk prediction, rather than open-loop chunk execution as the mechanism, with the failure localized to the training formulation and multi-step prediction. This work further democratizes physics-based policy development and evaluation through our reusable harness.
manipulationaction chunking - arxiv:2611.07417 · cs.ROMessyNav: Zero-Shot Navigation via Manipulation in Messy EnvironmentsKenneth Llontop, Cherie Ho, Dylan Zhou, Anuva Banwasi +3
Real-world environments are messy, with paths often blocked by everyday objects like chairs and toys, requiring robots to perform interactive navigation by manipulating obstacles to clear a path. Existing methods rely on extensive simulation-based learning or per-scene instructions specifying which objects to avoid or manipulate. They struggle to generalize across novel objects, diverse environments, and varied contexts such as a cart of fragile glassware that should be avoided while an empty one can be pushed. We formalize this problem as open-world interactive navigation and present MessyNav, a zero-shot framework based on vision-language models (VLMs). At its core, MessyNav performs explicit VLM-based affordance reasoning before task planning to decide whether to swipe, pick up, or avoid an object and produce a high-level plan. Evaluated on a mobile manipulator across real-world indoor and outdoor environments, MessyNav improves success rate by 27 percentage points (pp) over the strongest baseline in real-robot trials and achieves statistically significant gains in a larger-scale offline evaluation, without task-specific training.
manipulationmanipulator - arxiv:2611.07415 · cs.LGTruncated backpropagation through time induces a two-sided gradient turnpike propertyJulian D. Schiller, Matthias A. Müller
Truncated backpropagation through time reduces the computational burden of training recurrent models by restricting gradient computations to shorter temporal contexts. This induces two distinct boundary effects: missing past information perturbs the forward state evolution, while truncating the backward recursion omits future information. This motivates decoupling the interval over which gradient contributions are averaged for the parameter update from the surrounding temporal context over which they are computed. Under exponential decay of the state-transition Jacobians and mild regularity conditions, this leads to a two-sided gradient turnpike property: the error between truncated- and full-context gradient contributions decays exponentially with the distance from both context boundaries. We derive bounds for the averaged update gradient, yielding guarantees on gradient alignment and approximate full-context stationarity; under unbiased interval sampling and a Polyak-Lojasiewicz condition, these further imply a bound on the resulting optimality gap. Numerical experiments on the Wiener-Hammerstein benchmark illustrate the role of past and future context and the trade-off between predictive performance and computational cost.
benchmark - arxiv:2611.07411 · cs.ROKinematic Hypothesis Context: Embodiment-Conditioned Learned Video Coding for Bandwidth-Limited Robot Policy UplinksYuze Lu
Robots increasingly run large vision-language-action policies on remote servers, making the camera uplink the bottleneck. Existing video codecs spend bits on content the receiver could predict: the robot's own motion. We introduce the kinematic hypothesis context, a zero-rate conditioning signal that both ends build from joint states, the robot model and the camera calibration. Ego-motion homographies, per-link robot reprojections, an end-effector plane and a background memory warp the previous reconstruction into hypotheses that condition a learned video codec. We prove that history-only and global-pose conditioning with a limited receptive field cannot access this information under large motion, and that coding gains become task gains at the policy's success cliff. On LIBERO, the context saves 23.9% bit rate over a capacity-matched history-conditioned codec and 40-56% over x264, x265 and SVT-AV1, and needs 23.2% fewer bits than DCVC-RT at equal policy action deviation; trained to lower rates, it matches DCVC-RT in PSNR. In closed loop at equal rate, OpenVLA-OFT success nearly doubles from 22.8% to 43.1% at 6 kbps (72.6% with lower-rate training, 23.6 points above DCVC-RT), pi_0.5 success rises from 55.5% to 77.4%, and the gain grows under packet loss. The context transfers across three codec hosts and to real robot data.
vision-language-actionopenvlarobot policyliberomemory - arxiv:2611.07408 · cs.CVPSGAgent: Enhancing MLLMs' Dynamic Visual Perception for Panoptic Scene Graph Generation via Agentic ReasoningQingyun Liu, Yuankun Liu, Jiaye Feng, Jie Huang +5
Panoptic Scene Graph generation (PSG) requires holistic yet fine-grained reasoning over complex relations among objects. However, current Multimodal Large Language Models (MLLMs) applied to PSG suffer from a perception bottleneck: they function as static observers limited by fixed-resolution views. As a result, models often overlook critical relations involving small or peripheral objects and struggle to differentiate ambiguous predicates in complex scenes. To address these limitations, we propose PSGAgent, an agentic framework that enhances visual perception for PSG through a think-act-observe loop. PSGAgent allows MLLMs to dynamically refine their visual perception using two tools: a Region Zoom-In Tool for local exploration of small or boundary relations and a Visual Highlight Tool for relation-focused segmentation that filters out irrelevant contextual noise. To optimize decision-making, we construct a two-stage training pipeline. First, we create a high-quality, error-aware Multimodal Chain-of-Thought (M-CoT) dataset that provides robust supervision for cold-start learning. Then, we design a novel hierarchical reward mechanism integrated with Group Sequence Policy Optimization (GSPO) to enable strategic and adaptive tool usage. Extensive experiments on the PSG benchmark show that PSGAgent achieves SOTA performance, demonstrating strong capability in both holistic and fine-grained scene understanding.
scene graphagenticbenchmark - arxiv:2611.07403 · cs.RODynamic Watermarking and Action-Conditioned Latent Dynamics for Secure Vision-Based ControllersAditya Parameshwaran, Navid Aftabi, Abhishek Hanchate, Satish T. S. Bukkapatnam +2
Replay attacks are difficult to detect in vision-based control systems because previously recorded observations can remain plausible while the physical system follows a different trajectory. This paper develops an adaptive dynamic-watermarking framework for detecting replay attacks against vision-based controllers. A private Gaussian watermark is added to the nominal control input, and a control input-conditioned latent dynamics network (LDN) predicts the corresponding evolution in a compact latent space learned from image observations. Attack evidence is obtained from the discrepancy between the predicted latent response and the latent representation of the received image observation. A reinforcement learning policy adjusts the watermark covariance using the latent state and detector confidence, thereby balancing attack detectability, watermark energy, and control distortion. A Lagrangian penalty on task-specific safety costs is included during training to limit the adverse effect of the excitation. The framework is evaluated on a vision-guided UR5 manipulation task in which the robot reaches a target while avoiding a moving human arm under nominal and replay-attacked observations.
manipulationaction-conditionedlatent dynamics - arxiv:2611.07397 · cs.CVVis2State: Visual Reasoning through Executable State AbstractionsSheng Jin, Wentao Liu, Chen Qian
Solving visual puzzles requires both accurate interpretation of visual evidence and reliable computation under task-specific rules. Embedding visual assumptions and task constraints directly in solving code makes the represented problem difficult to inspect and revise independently. We introduce Vis2State, a training-free agentic framework that abstracts visual puzzles into explicit, revisable problem states linking visual inspection and execution. Visual inspection guides state construction and refinement, while generated programs solve the represented problem. The model constructs both the state and the program without predefined puzzle schemas or solving procedures. Inspection and execution feedback allow it to revise the problem representation and solving procedure independently. Vis2State achieves the state-of-the-art performance on four benchmarks. Cross-model evaluation further demonstrates consistent gains. Ablations, error analyses, and fixed-program state interventions further support the effectiveness of the state-centered design. The code will be made publicly available to facilitate future research.
agenticbenchmark - arxiv:2611.07393 · cs.LGCacheGraft: Enhancing Retrieval-Augmented Generation via Latent-Space Document EnrichmentVihan Lakshman, Peter Baile Chen, Zhuohan Gu, Michael Cafarella
Recent work in machine learning has demonstrated that transformer-based language model agents can communicate state not only through text, but also directly in latent space via key-value (KV) caches. In this paper, we argue that such cache-to-cache translation can provide utility beyond the original motivating application of real-time agent communication and can instead be utilized for retrieval-augmented generation (RAG). Building on the Cache-to-Cache (C2C) fusion framework, our proposed system CacheGraft encodes each corpus passage offline into a fused KV cache representation between a larger sharer model (4B parameters in our experiments) and a smaller receiver model (0.6B parameters) such that this cache can be injected into the receiver at query time with no need for the larger model. Across five QA benchmarks and two independent locally-run judges, CacheGraft improves answer quality over the text-only receiver baseline in 29 out of 30 comparisons, of which 17 of the 30 are statistically significant. Notably, these gains are zero-shot with respect to the fuser, which was never trained on retrieval data. These initial results suggest that latent-space document enrichment may be a promising avenue for enhancing the performance of RAG systems, particularly in compute-constrained settings where it is only feasible to deploy a small model at inference time.
retrieval-augmentedragagentbenchmark - arxiv:2611.07383 · eess.SYDistributed Robust Model Predictive Control for Optimal Consensus of Unknown Multiagent Systems with Event-Triggered CommunicationShiyue Jia, Nan Bai, Dawei Shi, Ling Shi
In this paper, we study the optimal consensus problem for multiagent systems with unknown dynamics. We propose a robust model predictive control (MPC) framework that jointly optimizes the control input and consensus sequences while enforcing state and input constraints. Within this framework, noisy offline data are used to construct nominal prediction models and quantify the associated uncertainty, which is explicitly incorporated into the robust MPC design. To solve the resulting optimization problem at each MPC update with reduced inter-agent communication, we develop a distributed gradient-projection alternating direction method of multipliers (ADMM) algorithm with an event-triggered information exchange mechanism and establish its convergence to the optimal solution. Furthermore, sufficient conditions are established to guarantee robust constraint satisfaction and practical consensus of the closed-loop multiagent system. Numerical simulations demonstrate the effectiveness of the proposed framework.
agent system - arxiv:2611.07378 · cs.LGAdversarial Multiplicity: Cherry-Picked Models and the Limits of DisclosureCassandra C. Chou, Benjamin Q. Huynh, Allison Koenecke, Elizabeth T. Chin
Predictive multiplicity -- where similarly accurate models disagree on individual predictions -- can be leveraged to optimize secondary metrics (e.g., fairness) without loss in predictive performance. However, exploiting multiplicity opens the opportunity for motivated parties (or coding agents gaming their objectives) to steer predictions toward their own ends, such as in litigation or consumer disputes with automated decision-making tools. In these adversarial multiplicity settings, we distinguish soft manipulation, where parties optimize over innocuous model choices (e.g., hyperparameter tuning, feature engineering), from hard manipulation, where parties directly perturb model parameters. We find a sufficiently flexible model can be manipulated to reverse any individual prediction with minimal change in predictive performance while passing black-box audits, making hard manipulation viable in most regulatory settings. Soft manipulation has smaller effects but can pass white-box audits requiring full access to code and data. Even aggregate estimates can be manipulated: a model's estimated effect of climate change on wildfires can be fully reversed by hard edits and soft search. We find the Reconcile algorithm typically resolves most disagreements between models, but more often in favor of the manipulator when the opposing model is chosen adversarially compared to at random. We map which manipulations can plausibly pass existing regulations and disclosure policies, and recommend mitigations such as pre-registration, proof of model training, sensitivity analyses, multi-metric evaluations, and committing to versioned public models.
manipulationmanipulator - arxiv:2611.07377 · cs.MADesigning with Cinematic Language: An Intent-Driven Approach to Human-AI Video Clip Co-CreationYi He, Zhengxi Yu, Feilin Han, Chuer Chen +2
Generative video models have made video creation more accessible, yet creators still struggle to translate high-level narrative intentions into coherent cinematic expressions. Existing systems primarily rely on prompt refinement or expose isolated camera controls, leaving a gap between what to express and how to film it. We present CineAssist, a human-AI co-creation system that treats cinematic language as an interactive design medium connecting creative intentions with executable video controls. We construct a Cinematic Language Design Space and an annotated dataset from 107 filmmaking sources and 1,436 film clips, linking communicative intentions to cinematic parameter values. CineAssist combines a multi-agent workflow for intent alignment and strategy recommendation with an interactive cinematic editor and a context-sensitive preference learner for user-controlled personalization. We evaluate the system through three studies examining end-to-end usability, cinematic quality, and preference evolution. Results show that CineAssist facilitates efficient, controllable, and personalized intent-driven cinematic video creation.
multi-agent - arxiv:2611.07374 · cs.MADesign of Multi-UAV Cooperative Deep Semantic Autoencoders for Communication NetworksXiaoling Han, Bin Lin, Nan Wu, Tony Q. S. Quek
Driven by the rapid development of the low-altitude economy, Uncrewed Aerial Vehicle (UAV) networks have become increasingly important for executing tasks such as search, rescue, and monitoring. However, under bandwidth-constrained scenarios, UAV networks still face significant challenges in task execution efficiency and intelligent processing capabilities. To address these issues, we propose Multi-UAV Cooperative Deep Semantic Autoencoders (MC-DSA) for communication networks. The architecture comprises one Task UAV (T-UAV) and multiple Support UAVs (S-UAVs), which jointly enhance communication efficiency and data-processing capability via semantic communication and cooperation. Specifically, we introduce a semantic matching and source selection mechanism, enabling the proposed MC-DSA to handle low-quality images while ensuring effective utilization of high-quality images. The T-UAV broadcasts a task query, and each S-UAV computes the semantic similarity between its local image and the query. Communication links are established only when the similarity exceeds a predefined threshold. The training algorithm optimizes network parameters through gradient descent, improving semantic feature extraction accuracy. The testing algorithm evaluates UAV collaboration efficiency and communication quality. Experimental results demonstrate that, compared with benchmarks, the proposed MC-DSA improves image semantic segmentation accuracy while maintaining task execution efficiency. As a result, it reduces redundant transmissions and offers an effective solution for multi-UAV communication networks.
benchmark - arxiv:2611.07372 · cs.AIEnvOpt: Automated Optimization for Agent EnvironmentsMingxuan Li, Huanzhi Mao, Qiuyang Mang, Hanchen Li +2
An LLM agent's cost and success depend on its model, its harness, and its environment. In addition to the widely studied directions of parameter tuning and harness optimization, we propose the third direction: optimization over stationary environments. Agents are often deployed in rich environments shared by many tasks, such as a company's Slack workspace or a complex collection of data sources, documents or tickets. We call these environments stationary environments. We introduce EnvOpt, a method for automated end-to-end optimization of stationary environments for any given model and harness, using real or synthetic tasks without needing any graders or reference answers. On AutomationBench, EnvOpt reduces the cost of Claude Haiku 4.5 by 72.7% and raises its pass rate by 14.34%. The same environment transfers to Opus 5, cutting its cost by 68.1% with a 1.34% gain. Added on top of Meta-Harness's optimized harness for Haiku 4.5, it lowers cost by a further 83.4% and adds 11.17% accuracy over Meta-Harness alone. Across 5 benchmarks, EnvOpt significantly reduces the cost of agents while improving or maintaining pass rate, and is transferable to different models and different harnesses. We also discuss common successful and rejected environment optimization strategies during automated optimization. More broadly, our results show that how information is represented to an agent shapes both its cost and its success, making the environment worth optimizing in addition to the parameters and the harness.
agentllm agentbenchmark - arxiv:2611.07367 · cs.LGTest-Time Self-Play: Scaling without a VerifierAkira Yoshiyama, Akmal Ashirmatov, Jonas Hübotter, Ido Hakimi +1
Test-time scaling methods spend additional compute at deployment to improve a large language model's performance on a hard target problem. Most methods rely on a verifier that scores candidate solutions for correctness, using it to select among or refine the model's outputs. In many deployment regimes, however, a strong verifier on the target problem is unavailable. We introduce Test-Time Self-Play (TTSP), an approach to test-time scaling using asymmetric self-play without reliance on a verifier. On a difficulty-calibrated subset of LiveCodeBench v6, TTSP is the best-performing method at every verifier strength we evaluate, and the gap to the strongest baseline widens as the verifier weakens; in the no-verifier limit, TTSP yields a 35\% absolute improvement in mean@8. The gains hold across a range of verifier strengths, except when the base policy has near-zero coverage of the correct answer.
self-play - arxiv:2611.07364 · cs.LGAsymptotic-Preserving Tensor Neural Networks for 3D High-Field Kinetic Equations with Applications to Uncertainty QuantificationShi Jin, Tian-ai Zhang
Kinetic equations in the high-field regime describe particle transport dominated by electric-field acceleration and collisional relaxation. These equations are widely used in simulations of plasmas and semiconductor devices. Typical models include the Vlasov--Poisson--Fokker--Planck (VPFP) and Boltzmann--Poisson systems. Their high dimensionality and multiscale nature pose two of the major challenges to accurate and efficient numerical solution. To address these challenges, we propose an asymptotic-preserving tensor neural network (APTNN) method. Specifically, to reduce high-dimensional computational costs and memory requirements, we adopt the canonical polyadic (CP) decomposition structure of TNNs to approximate their solutions. We show that the equation residuals retain this decomposition form, allowing high-dimensional loss integrals to be evaluated through combinations of one-dimensional integrals. The resulting computational and storage costs are reduced from exponential to polynomial growth with dimension. To improve the prediction accuracy in the high-field limit, we enhance the baseline AP loss by incorporating the first moment equation. The added loss term directly constrains the limiting flux closure to reduce its approximation error in the high-field regime. The enhanced AP loss can still be evaluated dimension-wise. Numerical results show that APTNN improves prediction accuracy in the high-field regime with significantly reduced computational costs and memory usage. For a 3D3V VPFP problem with multiple random inputs, training takes less than 30 minutes on a single GPU.
memory - arxiv:2611.07356 · cs.LGAcoustic Anomaly Detection for Early Identification of Drill JammingMaria Beatriz Silva, Lucas Edson Silva de Araújo, Sergio F. Chevtchenko, Ermeson Andrade +2
In the context of Industry 4.0, predictive maintenance is essential to minimize unplanned downtime and ensure operational safety. In machining processes, drill jamming is a critical failure that compromises both workpiece quality and tool integrity. This work proposes an acoustic anomaly detection system using ultrasonic sensors to identify pre-failure patterns in Computer Numerical Control (CNC) machining. We release a public dataset of time- and frequency-domain acoustic features from 17 drills, and benchmark semi-supervised autoencoder-based normality modelling against supervised XGBoost classification using Leave-One-Drill-Out validation. The results demonstrate that the Autoencoder is highly effective for industrial safety, achieving a Recall of 1.00 and a consolidated F1-score of 0.81. Furthermore, the threshold-independent analysis shows that the Autoencoder better separates normal and pre-failure acoustic patterns, achieving a consolidated AUC of 0.8294, compared with 0.7559 for the supervised model under inter-tool covariate shift. In contrast, the supervised XGBoost achieved an average F1-score of 0.55 and failed to detect failures in some of the evaluated drills. The findings highlight that the ultrasonic spectrum provides a reliable window for identifying incipient mechanical instabilities, and the AE model offers a robust safety envelope by anticipating all jamming events across the tested drill bits.
benchmark - arxiv:2611.07349 · cs.LGFrom What Is Used to What Is Missing: Rethinking Dynamic Rank in LoRAPeng Xue, Guodong Fan, Guang-Yong Chen, Xing Chen
Parameter-efficient fine-tuning with LoRA is governed by rank, which controls parameter budget and locally accessible task updates. Existing dynamic-rank methods mainly focus on capacity allocation, leaving unresolved whether the current adaptation space is sufficient for the task. We characterize the first-order information boundary of standard factor gradients, showing that their unobservable subspace coincides with the normal space of the low-rank manifold. Thus, factor gradients alone cannot distinguish under-optimization within the current space from task-relevant directions beyond its first-order reach. This motivates a geometric view of rank evolution, where tangent information optimizes existing degrees of freedom and the normal spectrum quantifies the value of new ones. Under a local quadratic model, we show that leading normal singular directions yield the locally optimal rank expansion, with spectral energy determining its gain. Building on this characterization, we propose GEAR, which turns the tangent--normal decomposition into rank evolution through tangent-space optimization, matrix-free normal-spectrum estimation for rank expansion, and task-induced geometric initialization. Across language and vision benchmarks, GEAR achieves the best average performance in all evaluated settings with substantially fewer trainable parameters. On LLaMA3-8B, it raises the average commonsense score from 86.61 to 88.45 while reducing trainable parameters from 42.0M to 14.4M.
benchmark - arxiv:2611.07342 · cs.AISelective Retrieval Routing for Multi-Violation Knowledge Graph RepairTung-Wei Lin, Gabe Fierro, Han Li, Tianzhen Hong +2
Knowledge graphs often contain missing or inconsistent facts that undermine the applications built on them. Schema validation detects such violations but does not identify which of several valid repairs recovers the intended graph. External corpora can supply the evidence needed to identify the intended repair, but retrieving from them is costly and not always successful. We propose a framework for selective retrieval in iterative graph repair. An LLM repair agent explores and edits the graph, while a router controls its access to a researcher agent that retrieves external evidence. We extend a violation-injection framework to generate cases with multiple compatible violations and known reference repairs. The agent's repair outcomes on these cases provide supervision for training the deployment-specific router. On two building-automation datasets, researcher access improves repairs for some violations and degrades them for others, with effects differing across deployments. The router uses fewer repair-time tokens than always enabling access in all six evaluation suites. It achieves higher repair $F_1$ than always disabling access in five of six suites.
knowledge graphagent - arxiv:2611.07327 · cs.LGPermanent and Transient Representations for Continual Reinforcement LearningNishanth Anand, Doina Precup
Continual Reinforcement Learning (CRL) agents struggle to adapt to new situations while retaining prior knowledge, leading to a stability-plasticity trade-off. One approach to solving this problem is by using complementary learning systems: one for slow, long-term learning and another for quick, transient adaptation. Anand & Precup (2023) instantiated this idea using permanent and transient value functions, but used the same features in both systems. Intuitively, long-term learning should rely more on parametric representations for broad generalization, while transient learning would be best served by non-parametric representations for situation-specific adaptation. In this paper, we explore this idea both theoretically and empirically. By introducing a novel non-parametric approximation for estimating the transient value function in complex tasks, we demonstrate that our method enables online learning and outperforms competitive baselines on image-based tasks and Craftax-Classic, underscoring the effectiveness of our system-level decomposition.
online learning - arxiv:2611.07319 · cs.LGWeightLoop: Weight-Space Programming with Experimental FeedbackMengran Li, Bo Li, Weijing Zhao, Chengyang Zhang +5
Weight-space programming (WSP) shows that language models can reason about and rewrite neural network weights; recursive self-improvement (RSI) motivates feedback-driven improvement. Models guiding experiments receive concrete feedback: completed assays reveal prediction errors and inform what to test next. We introduce WeightLoop, a framework that turns experimental feedback into persistent updates to deployed predictors. GPT-6 Astra inspects a task specification and the model implementation, then writes a bounded, reusable parameter-editing program. As new measurements arrive, the program identifies prediction discrepancies, validates candidate edits using observed outcomes, and writes the selected changes back to the original model. The editor remains fixed while the predictor evolves, enabling repeated updates without further LLM calls or gradient retraining while preserving the original inference interface. In offline replay, targeted edits improve mean ranking utility in seven of eight endpoint-budget settings (all eight versus random edits) and mean final discovery across three sequential tasks. WeightLoop connects WSP with an RSI-inspired improvement loop, making experimental evidence a direct driver of model change and subsequent decisions.
self-improvement - arxiv:2611.07307 · cs.MATake It From Here: Evaluating Coding Agents On Project TakeoverDingxian Wang, Zhe Fu, Jiaqi Deng, Chao Yin +8
Coding agents increasingly collaborate on long-term projects. When taking over a project, they need to recover currently applicable requirements from the interaction history and update them in response to new requests. However, most coding benchmarks do not include long interaction histories, and some evaluations that consider history still do not cover the full process from requirement reconstruction to project delivery. We therefore provide agents with the interaction history, current request, and pre-request project environment, ask them to make the requested changes, and assess takeover performance through delivery tests. Across six model-harness configurations and 37 targets eligible for executable validation, the mean complete-delivery rate under Full History is only 4.95%. We then use requirement atoms to independently diagnose a larger set of takeover targets, examining relevant information identification, pre-request state reconstruction, and current request processing. We further supply correct pre-request or target states to examine the remaining difficulties in requirement processing and project modification. The diagnostics show that none of the six configurations produces fully correct post-request requirement states, and a case illustrates omission of previously recovered constraints during updating. With correct target states supplied, the mean complete-delivery rate is 8.11%, with build and regression pass rates of 96.40% and 90.99%, but target-test success remains 9.01%. These results show difficulties in both forming target requirement states and implementing supplied targets in existing projects. Based on these findings, we introduce ReqMemBench, which combines requirement diagnostics with delivery validation to support systematic evaluation of project takeover.
benchmark - arxiv:2611.07301 · cs.CVArtiVidRM: Verifiable Artifact-as-Rubric Reward Modeling for Explicit Quality Articulation in Video GenerationCaorui Li, Yangguang Ji, Shuxia Lin, Yu Chen +11
Reinforcement learning has become an effective approach for improving video generation, but its success depends heavily on reliable reward models. Existing methods either map videos directly to opaque scalar scores or generate free-form rationales that are not explicitly optimized as the basis of the final judgment. We propose ArtiVidRM, a verifiable, artifact-as-rubrics video reward model that follows an observe-then-score paradigm. It first identifies visible generation artifacts and organizes them into a structured artifact list, then uses these artifacts as explicit rubrics for quality assessment. This intermediate representation grounds each score in concrete video-level evidence, making the reward more interpretable and informative for optimization. We further introduce a multi-reward reinforcement learning strategy to jointly optimize artifact prediction and quality assessment. The reward decomposes into complementary objectives that evaluate artifact prediction quality, including semantic coverage and prediction efficiency, together with a pairwise preference objective that supervises quality ranking through score differences. By jointly optimizing the intermediate artifact representation and the downstream preference prediction, the model learns to produce informative artifact lists while improving the accuracy of video quality assessment. Experiments show that ArtiVidRM achieves state-of-the-art performance on video preference benchmarks, provides effective signals for video diffusion optimization, and enables artifact-guided refinement of closed-source video generators.
benchmark - arxiv:2611.07286 · cs.AIHow Should Agents See Code? Rethinking Code Representations for Agentic Software EngineeringMelika Sepidband, Jiho Shin, Song Wang, Hadi Hemmati
LLM agents are increasingly used for repository-level software engineering tasks. However, most agents still access repositories through raw files, lexical search, and shell commands, treating code primarily as text. This paper investigates map-based code-representation workflows as a design variable in agentic software engineering, examining how different repository views change agents' retrieval behavior, task success, token use, and latency. In this paper, we study five code-representation workflows: abstract syntax trees, call graphs, module dependencies, documentation-enriched search, and a hybrid workflow that combines them. Each workflow exposes a repository map that the agent can query during a task. We evaluate Claude Code with Haiku 4.5 against a raw-tool baseline on four repository-level tasks: fault localization with LinuxFLBench, repository-level code generation and repair with SWE-bench Pro, test generation with SWT-bench, and security patching with SecBench. The results show that representation workflows are most useful as token-saving tools when localization is the bottleneck, while their success gains narrow when the decisive step is patch or test authoring. On LinuxFLBench, all five workflows improve recall@10 over the baseline (0.757-0.829 vs. 0.717) and reduce token usage by 25-39%. On SWE-bench Pro, resolution remains close to the baseline (44.3-50.6% vs. 49.0%) while token usage drops by 17-26%. On SecBench, Callgraph has the highest patch-resolution rate in this sample (34.3% vs. 28.9%) and uses fewer tokens, but all workflows increase latency. On SWT-bench, Hybrid and Docs have the highest reproducing-test rates in this sample (47.8% and 47.2% vs. 44.7%), but all workflows use more tokens and take longer. Overall, map-based workflows change both what the agent inspects and how much it spends, but their benefits are task-dependent rather than universal.
agentllm agentagentic - arxiv:2611.07275 · cs.ROEfficient Diffusion Policy Fine Tuning with Latent Noise Representation BridgingYuanchu Liang, Yue Yang, An-Chi He, Hanna Kurniawati
Pre-trained diffusion and flow policies have emerged as powerful backbones for robotics, yet adapting them to novel data via reinforcement learning (RL) remains computationally expensive, sample inefficient and lacks theoretical guarantees. Prior approaches that fine-tune the initial noise bypass costly backpropagation through time but struggle with the "noise aliasing" problem: there does not exista natural link between the noise space value and the action space value. This paper resolves these challenges by introducing the Latent Noise Linear Markov Decision Process (LNL-MDP). We derive the Bridge Equation for LNL-MDPs which eliminates the noise aliasing problem by aligning values with a shared representation between the noise and the action space. Furthermore, our LNL-MDP fine tuning framework falls under the FineTuneRL theory with near-optimal guarantee. Leveraging these insights, we propose Latent Representation Bridging for Diffusion (LARBRID), an algorithm that efficiently learns the representation to optimize the initial distribution that steers the pre-trained model to act optimally in the environment. It yields up to 2-3$\times$ performance increases on challenging control problems across a diverse suite of robotic locomotion and manipulation tasks. Code can be found at: github.com/Maincode-Research/LARBRID.git.
manipulationdiffusion policy - arxiv:2611.07272 · cs.ROVREgo: Consumer VR Headsets for Learning Robot Manipulation from Egocentric DemonstrationsLiuchuan Yu
VREgo trains manipulation policies from demonstrations recorded on consumer VR headsets; we instantiate it with EgoKit on one such headset, the PICO 4 Ultra. On-device head and hand tracking and stereo object localisation allow the operator to begin a demonstration at the table. We adapt HumanEgo's preprocessing to estimate grasp thresholds per recording, update object positions after release, and propagate segmentation masks between keyframes. We evaluate policies in simulation on two tabletop tasks recorded by one operator. For cube placement, we compare PICO and Project Aria recordings of the same 41 layouts. With 40 training demonstrations, the PICO policies placed the cube on the tray in 67, 74, and 60 of 100 evaluation layouts across three training seeds. The corresponding Aria results were 73, 78, and 0. For a two-stage stacking task recorded with the PICO alone, the baseline placed the block on the tray in 94 of 100 layouts and met the end-of-episode criterion for both placements in two. In one diagnostic episode, repeated replanning deferred release beyond the eight executed steps. Demonstration replays also showed that the open gripper could displace the cube before grasping it. After adapting the training trajectories and adding task-specific execution rules that use simulator object positions, the system met both placement criteria in 93.8% of episodes across four evaluation noise seeds. The same 100 layouts informed rule development and evaluation. VREgo is available at https://vrego.chuange.org/.
manipulationgrippergrasp - arxiv:2611.07268 · cs.ROCRISP: Constraint-Resolved Interactions between Subsystems for Predictive Whole-Body CooperationDario Valenzuela, Jeeseop Kim
This paper presents a layered control framework that recasts high-dimensional system dynamics as a composition of dynamically interconnected subsystems, with dynamic dual-arm manipulation on the humanoid serving as a representative application. The humanoid is decomposed into a single rigid body (SRB) representing the torso and floating-base dynamic models representing the articulated arms and lower-body. The rigid connections between arm and torso and the closed kinematic chain formed between the end-effectors during cooperative manipulation are represented as holonomic constraints. The corresponding Lagrange multipliers parametrize the interaction wrenches transmitted at the shoulders and between the end-effectors. The high-level nonlinear model predictive control (NMPC) jointly optimizes the subsystem state trajectories, arm torques, ground reaction forces (GRFs), and interaction wrenches, enabling online wrench design and optimization while preserving the dynamic coupling among the independently modeled subsystems. The optimized arm torques are applied directly for cooperative dual-arm manipulation, whereas the prescribed GRFs are mapped to lower-body joint torque for whole-body coordination. We validate the proposed framework through numerical simulations and hardware experiments on a G1 humanoid robot. The results demonstrate optimal interaction wrenches consistent with simulated contact measurement, cooperative dual-arm trajectory tracking, and stable whole-body coordination while executing the NMPC in online manner.
manipulationhumanoid - arxiv:2611.07265 · cs.LGHow Deep is the Signal? Time-Context up to 30 Years for Colorectal Cancer Prediction in a Primary-Care EHR ModelAchille Raviez, Lan Wang, Brendan Delaney
EHR foundation models read medical events within a fixed context window, and long context has so far been measured only in tokens. How far back in time a model must look to capture the earliest predictive signal is unknown. We introduce time-context, the calendar lookback window before the prediction point, and provide, to our knowledge, the first systematic evaluation of decade-scale time-context for an EHR foundation model. We pretrained a decoder-only transformer (34.6M parameters) on 1.4B coded events from 3.6M primary-care patients in North West London and fine-tuned it to predict incident colorectal cancer under a 12-month pre-diagnostic buffer. A frozen cohort of 1,008 cases and 9,582 matched controls, all registered for at least 30 years, was evaluated at 15 nested lookback windows from 2 to 30 years, so that every truncation is compared within the same patients. Discrimination rose from an AUROC of 0.551 at 2 years to 0.590 at 18 years, then plateaued, with about 80% of the gain reached within 10 years; an XGBoost baseline converged to the same level (peak AUROC 0.589). A simulation further suggests that rotary position embeddings (RoPE) over elapsed time alias beyond about 20 years at the default base frequency, motivating a retuned $θ$. Time-context is a measurable design axis for clinical prediction models, distinct from token-context, and should be reported.
long context - arxiv:2611.07263 · cs.CLHyFlash: A Prefix-Memory Prior for On-Device Speculative DecodingJongwoo Ko, Pashmina Cameron
Parallel block drafting accelerates speculative decoding by predicting multiple future positions in a single backbone pass, but on-device deployment constrains both drafter capacity and context storage. Under tight memory budgets, drafters must model long-range dependencies and fine-grained draft-position interactions without the growing KV-cache overhead of full-context attention. We propose HyFlash, a compact parallel drafter that introduces a prefix-memory prior to separate long-range context modeling from fine-grained draft-position interactions with limited capacity. Local attention captures recent context and draft-position interactions, while recurrent memory compresses historical target features into fixed-size states. This inductive bias enables effective local and long-range modeling with bounded draft-side memory. Experiments show that HyFlash achieves higher acceptance than full-context attention with substantially less KV storage and with larger gains for shallow drafters and long contexts.
memorylong context - arxiv:2611.07261 · cs.ROAn Empirical Study of Reinforcement Learning for Keypoint Positioning of Deformable Linear ObjectsVinay Rodrigues, Sara Zeynalpour, Gokhan Solak, Sebastian Hjorth +1
Deformable linear object (DLO) manipulation is challenging because high-dimensional configurations, nonlinear contact dynamics, and diverse material and geometric properties produce a broad range of object responses that are both difficult to model and data-intensive to learn. This work studies reinforcement learning for planar manipulation of a one-end-grasped DLO, focusing on training efficiency, observation design, robustness, and zero-shot sim-to-real transfer. We compare three Proximal Policy Optimization training stacks and find that the GPU-parallel MuJoCo Playground stack enables faster convergence and more consistent performance than CPU-based alternatives under the same wall-clock budget. Relative keypoint observations and an asymmetric actor-critic formulation further improve learning efficiency. Policies trained with observation noise, delay, domain randomization, and disturbances achieve simulation success rates of 96.4% for single-point tracking, 98.0% for disturbance recovery, and 88.8% for two-point tracking. Using markerless RGB-D keypoint estimation, the policies transfer without real-world fine-tuning, succeeding in 9/10, 10/10, and 8/10 physical trials, respectively.
manipulationsim-to-realgrasp - arxiv:2611.07256 · cs.LGTriAttention: Unified Sparse Mixed-Precision Attention for Long-Context LLM InferenceHanzhi Zhang, Qiao Zhang, Qinglei Cao, Heng Fan +3
Sparsity and quantization reduce the cost of attention by skipping interactions and lowering the precision of those retained. However, useful interactions differ in their precision requirements: some tolerate low-precision scores while remaining sensitive to removal. We therefore introduce TriAttention, which assigns query-key (QK) tiles to skipping, 4-bit, or 8-bit computation, using 4-bit to retain quantization-tolerant interactions at lower cost. To represent compactly, we encode each head's computation regions with two diagonal-distance boundaries, keeping policy storage independent of context length. Offline calibration then evaluates candidate boundary pairs using normalized projected-output errors and complete-tile execution costs. Using these measurements, a model-wide allocator selects each head's boundaries to minimize calibration error under a shared computation budget. During prefill, the selected boundaries generate precision-grouped traversals for a fused kernel that skips omitted tiles and combines contributions from both QK precisions into one normalized attention output. The quantized keys and values prepared during prefill are then reused in a compressed key-value (KV) cache for decoding, reducing repeated reads of prompt data. Experiments show that TriAttention achieves end-to-end generation speedups of up to 7.88x over the Torch attention and 2.65x over the BF16 baseline. Meanwhile, TriAttention maintains quality comparable to the original models on general-purpose and long-context tasks. These results show how a compact retention-and-precision policy translates local sensitivities into attention computation and repeated KV-cache access. The implementation is https://github.com/HanzhiZhang-Ulrica/TriAttention.
long-context - arxiv:2611.07254 · cs.LGUsing Transformers for Teaching and Learning via Symbolic Distillation of Neural A-life AgentsShuwei Wang, Vadim Bulitko, Matthew R. G. Brown, Michael Youngblood +4
Neural networks controlling behavior of A-life agents can evolve over multiple generations. While effective in making the agents survive, such networks are generally opaque to a human observer. Furthermore, the neural weights cannot be transferred between agents with different neural architectures. Distilling such neural networks into compact symbolic policies can help humans understand agent behavior in A-life. Additionally, it allows the agents to share their knowledge in a compact fashion, improving population-level learning. Distilling a neural policy into a symbolic policy is typically performed by searching the space of symbolic policies to find one that approximates the neural policy. However, even with a domain-specific language the space of policies can be large, necessitating incomplete stochastic search such as genetic algorithms. Yet the distillation remains computationally expensive. We propose to instead use a transformer pre-trained on readily available random percept-action data. Doing so speeds up distillation substantially. We derive the transformer architecture via a neural architecture search. Empirical results compare different distillation methods and the influence of the symbolic space on A-life population dynamics through self-reflection and teaching and learning.
agent - arxiv:2611.07247 · cs.LGGoal Revelation as a Bridge Between Information-Theoretic Empowerment and Capability-Based Human PowerJaime Ruiz-Serra, Tomáš Gavenčiak, Jobst Heitzig
Power matters for AI safety. Optimal agents tend to acquire power across a wide range of objectives (instrumental convergence). As we embed AI agents into socio-technical systems, humans risk gradual disempowerment. Both concerns call for a formal, quantitative measure of power. Two such measures in the context of Markov games are information-theoretic empowerment, the capacity of the channel from an agent's actions to future states, and capability-based individual power, a measure of goal attainment built from the probabilities that a human would attain each goal in a family of candidates were they pursuing it. We show where the two meet. From individual power, Fano's inequality yields a floor $B(s)$. This floor is a lower bound on what we term goal revelation, the mutual information $\mathbb{I}(G;ξ\mid s)$ between the human's latent goal and the realised trajectory. Empowerment caps revelation from above, so that $B(s) \le \mathbb{I}(G;ξ\mid s) \le \mathbb{I}(\mathbf{A}_h \to ξ\mid s) \le \mathfrak{E}_h(s)$. Each quantity can be large while the one below it stays at zero, highlighting the following failure modes. An assistant maximising empowerment can prefer an interface offering broad influence over goal-irrelevant degrees of freedom to one that lets the human attain every goal they might hold. An assistant maximising revelation cannot tell a working interface from one rewired so that every goal yields the wrong outcome, since relabelling goals leaves mutual information unchanged while zeroing every attainment probability. Positive guarantees therefore come only from the attainment end of the chain. A satisfaction floor, computable from two statistics of interaction logs, offers a way to monitor the power users retain in a deployed system, and simulations in random environments show the three quantities are correlated but far from interchangeable.
ai agent - arxiv:2611.07246 · cs.LGKG-Neighbor: Knowledge-Guided Neighborhood Construction for Embedding VisualizationChang Won Kim, Jaewook Lee, Haeun Min, JunNyeong Jeong +3
In embedding visualizations, nearby instances are interpreted as semantically related, although neighborhoods derived from input embeddings may not reflect content relationships. Neighborhood-based dimensionality reduction (DR) methods can propagate this limitation by optimizing layouts from embedding-based neighborhood graphs. We propose Knowledge-Guided Neighborhood Graph Construction (KGN), which represents relational knowledge derived from each instance as a set of knowledge graph (KG) triples. It uses the degree of triple sharing as pairwise semantic evidence to adjust embedding distances and reconstruct the top-k neighborhood graph; neither KG extraction nor neighborhood construction uses class labels. We also introduce IDProfile (IDP), which interprets residual mismatches by independently assessing whether the source content supports the assigned class label and the neighborhood-majority class. Across eight text and image dataset-encoder conditions, we characterize KGN over the full guidance-strength sweep and four KG similarity measures. Across this sweep, the maximum NMI and LT_2D each exceed the corresponding kNN baseline in all eight conditions, with the largest gains reaching 0.68 and 1.35, respectively. The gains and the extent of neighborhood reconfiguration vary across conditions and are larger under encoder conditions with lower kNN baseline scores. The IDP analysis shows that the largest reductions are concentrated in Placement Issue candidates, whereas Label Issue and Unresolved Case instances account for a larger share of the remaining review set. These results indicate that KGN can narrow the evaluated gap between embedding-based neighborhoods and content-supported semantic relations, while IDP shows that a residual mismatch should not automatically be interpreted as a projection failure.
knowledge graph - arxiv:2611.07245 · cs.LGMAC-v2: Long-horizon model-based value expansion from pixelsKwanyoung Park, Sergey Levine
We study whether \emph{long-horizon model-based value expansion} can serve as an effective framework in pixel-based offline reinforcement learning (RL). Long-horizon model-based value expansion from pixels is substantially harder than state-based environments, as the agent must learn both the representation and its dynamics, and errors from either can accumulate recursively over imagined trajectories. We introduce MAC-v2, which enables long-horizon model-based expansion from pixels by training a reconstruction-free latent world model with action chunking. By learning representations from prediction and control objectives, MAC-v2 does not need to encode control-irrelevant visual details that can hinder the dynamics prediction. Action chunking further mitigates the prediction error by reducing the number of recursive model predictions required for value expansion. Across diverse visual benchmarks including visual OGBench, MMBench2, Robomimic, and LIBERO-10, MAC-v2 achieves strong performance against model-free and model-based offline RL baselines. Our analyses show that long-horizon model-based value expansion is critical for the performance of MAC-v2, and that both action chunking and latent learning objectives are essential for reliable long-horizon latent imagination.
action chunkingliberoworld modelagentbenchmark - arxiv:2611.07241 · cs.LG$D^2$-OPD: Separating Direction and Distance in Extrapolative On-Policy DistillationBowen Fang, Dingmin Wang, Hengzhi Pei
On-policy distillation (OPD) trains a student model to match a teacher's next-token distributions on student-generated prefixes. Extrapolative OPD further extends the teacher's log-probability shift relative to a reference policy to construct stronger learning targets. However, existing methods couple extrapolation direction and distance, allowing the magnitude of the teacher's shift to dictate how far the target moves. We propose $D^2$-OPD, decoupling these two factors: a per-prefix hard KL constraint bounds the target's distance from the teacher, while an adaptive combination of the teacher's post-training shift and the initial student--teacher gap determines the extrapolation direction. We optimize the student toward this target using forward KL, providing learning signals even for low-probability tokens. We evaluate three settings: where standard OPD performs well, where it struggles with base-model initialization, and where no teacher reference is available. On mathematical reasoning benchmarks, $D^2$-OPD attains the highest score in all three settings, outperforming the strongest baseline by up to 9.45 points. Ablations validate the contributions of each component.
post-trainingbenchmark - arxiv:2611.07238 · cs.CLHow Do AI Agents Waste Your Money?Huimin Zeng, Yuxuan Lu, Jesse Cai, Eric Hanchen Jiang +7
AI agents may consume massive tokens across a wide range of tasks, but how many of those tokens are really necessary? Although prior work has shown that agents can incur unexpectedly high token costs, we still lack a systematic way to identify and measure token waste. In this work, we introduce a unified framework for measuring token waste across heterogeneous agent systems. Our framework records and normalizes complete agent trajectories, identifies removable spans using an LLM-as-a-Judge (LLMaJ), and maps the identified text spans back to the original trajectory for token-level measurement. We validate the identified waste through direct interventions by removing marked spans, resuming agent execution, and measuring changes in task outcomes and subsequent actions. In online pruning experiments, we remove 41-46% of the model's reasoning while largely preserving task performance across two distinct agent benchmarks. Using this framework, we analyze agents across multiple settings. Our analysis shows that token waste has a system-level structure that aggregate token usage alone cannot reveal. First, waste varies more across tasks and agent settings than across models. Second, reasoning and visible execution can exhibit different waste patterns. Third, different harnesses can introduce distinct sources of inefficiency even under the same model and task. Overall, our framework transforms token waste into an actionable optimization signal, enabling researchers and practitioners to train more efficient models, design better reasoning strategies, and optimize agent harnesses toward reducing unnecessary computation rather than merely reducing token usage.
agentai agentagent systemagent benchmarkbenchmark - arxiv:2611.07235 · cs.ROAdaptive Hybrid Force-Vision Control with EMG-Derived References for Deformable Bottle ManipulationJeffrey Lee, Geoffrey Hanks, Evangeline Bouganim, Yue Hu
Robotic automation of laboratory tasks has become a growing interest to remove researchers from hazardous environments and repetitive procedures. However, even equipment as simple as squeeze bottles poses a challenging manipulation problem due to its deformability. This paper presents an adaptive hybrid vision-force controller for robotic grippers to effectively manipulate laboratory squeeze bottles to dispense liquid. This approach integrates force and vision feedback according to the interaction state, enabling effective squeezing while accounting for bottle deformation and recovery. The proposed controller is validated by comparing its performance with a pure force controller on 12 combinations of conditions consisting of 2 target dispense volumes, 2 plastic squeeze bottles, and 1 to 3 squeezes used to dispense. Human demonstrations of these operating conditions were collected using electromyography (EMG) sensing mapped to a desired grip force, which was then used as the force reference for the controllers to squeeze the bottles. Results demonstrated that the proposed adaptive hybrid vision-force controller achieved more accurate liquid dispensing with a mean absolute error (MAE) of 6.54 mL compared to a MAE of 15.92 mL from the pure force controller, while also replicating the human action recorded from the EMG signal more closely.
manipulationgripper - arxiv:2611.07229 · cs.ROSpatial-UMI-Dream: Scaling World Action Models with Metric Geometry for UMI ManipulationWeiduo Yuan, Ze Yuan, Sihan Liu, Jionghao Wang +3
The Universal Manipulation Interface (UMI) enables scalable demonstration collection independently of a specific robot. However, its moving, close-range gripper cameras entangle camera and object motion without a stable external view, leaving the metric geometry needed for precise manipulation implicit in RGB. We introduce Spatial-UMI-Dream, a 5B spatial World Action Model (WAM) pretrained on 605 hours of RGB-D UMI data through two-stage training -- world-only adaptation followed by joint world-action learning -- to jointly predict future RGB, fisheye-aligned metric depth, and actions. Stereo-derived depth makes scene scale and spatial relationships explicit within a unified dual-gripper RGB-D representation. Furthermore, evaluating UMI policies requires testing whether UMI's robot-independent demonstration collection translates into policy generalization across robot embodiments, alongside robustness to environmental changes. We introduce RoboTwin-UMI, a twenty-task simulation benchmark built on RoboTwin with UMI wrist-mounted fisheye observations, enabling controlled, repeatable evaluation of UMI policies under environment and embodiment changes separately and jointly without additional fine-tuning. On RoboTwin-UMI, Spatial-UMI-Dream achieves 79.10% in-domain success and 72.15% under combined embodiment and environment shifts, exceeding the strongest evaluated baseline in each setting by 4.65 and 10.30 percentage points, respectively. It also exhibits the smallest success-rate drop under embodiment transfer among the evaluated policies. In a matched pretraining ablation, RGB-only pretraining leads to a 5.75-percentage-point decrease in in-domain success compared with RGB-D pretraining. Across six real-world tasks, Spatial-UMI-Dream achieves 85.0% average success, exceeding the strongest baseline by 24.2 percentage points. Project page: https://weiduoyuan.com/spatial-umi-dream/
manipulationrobotwingripperbenchmark - arxiv:2611.07222 · cs.AISPILLWAY: Separating Target Admission from Sanitization to Defend Tool-Using Agents against Prompt InjectionWoosang Lim, Yeseul Chang, Jaewoo Lee
Tool-using agents act on email, files, and payments, so an instruction hidden in a web page or a shared document can make an agent send data or money to an attacker. Many runtime defenses put a language model on the decision that lets an action through, so the decision depends on a model that may read the attacker's text and adds model calls to every step. Provenance defenses keep the model off the final check but need a model to plan or route each task first. Where an action goes can instead be read from the tools' typed returns: an email's sender or a payment's account arrives in a typed field, while injected instructions arrive as free text. SPILLWAY turns this observation into a deterministic floor with no per-task plan. The check admits a call when each target value that appears in tool output but not in the user's task was returned by a read-only tool in a typed field or is covered by a fixed delegation rule. A deterministic sanitizer removes instruction-like regions its fixed rules match before the model reads them and retains the removed spans as evidence, and a local judge may restore a refused call, never one whose target lies in the removed text. Across three backbones, and seven task suites on one of them, no standard templated injection redirects an action to an attacker's target under SPILLWAY, at the highest observed mean utility and composite scores (NRP, ERS) of the defenses we run under that attack. A forged chat template also produces a raw zero, and mean in-scope attack success under a released reinforcement-learning attacker is .007. SPILLWAY reaches this floor with no defense-triggered backbone re-run and almost no added model call. As personal AI agents act on users' email, files, and payments, SPILLWAY gives them an auditable refusal floor on where their actions go that keeps them useful and holds across backbones.
agentai agent - arxiv:2611.07217 · cs.LGTileQuant: Tile as the Precision-Allocation Unit for LLM Mixed-Precision Post-Training QuantizationHanzhi Zhang, Qiao Zhang, Qinglei Cao, Heng Fan +3
Post-training quantization (PTQ) reduces large language model (LLM) deployment costs, but preserving quality requires accounting for uneven weight sensitivity. Uniform PTQ uses one weight bit-width, importance-aware PTQ adjusts scales or reconstructs weights within that width, and column-wise mixed precision retains selected columns at higher precision. However, sensitivity varies within columns, so whole-column retention also protects less sensitive weights, while runtime outlier detection adds decoding overhead. We therefore introduce TileQuant, a training-free framework that assigns precision to hardware-aligned 2D weight tiles. Offline calibration provides input-activation-aware scoring from forward statistics for lower calibration cost and gradient-informed scoring from loss sensitivity for stronger quality preservation. Within each projection's retention budget, the highest-scoring tiles remain at 16-bit and the rest use 8-bit. The resulting map guides reconstruction into compact weight buffers, which the kernel reuses for native integer and 16-bit computation with local accumulation to reduce memory traffic. Experiments show that TileQuant reaches 2.66x the original model's generation throughput. Across language modeling, reasoning, long-context retrieval, and question answering, it preserves quality and surpasses quantized and original models on multiple tasks. These results support selecting precision units that capture local sensitivity while enabling efficient hardware execution. The implementation is: https://github.com/HanzhiZhang-Ulrica/TileQuant
memorylong-contextpost-training - arxiv:2611.07216 · cs.LGCompositional Unalignment: Auditing Safety in Large Vision-Language ModelsHong kyu Lee, Kai Shu, Li Xiong
Large vision-language models (LVLMs) are vulnerable to compositional harm: an image and a prompt that are benign in isolation become harmful through the model's cross-modal reasoning. Recent alignment methods report strong scores on compositional safety benchmarks, but an LLM judge only measures observed behavior and cannot reveal how robustly safety is integrated into the model. We argue that un-alignment provides a complementary audit of alignment robustness. We first show that minimal supervised fine-tuning reaches an LLM-judge score comparable to a multi-stage reinforcement-learning pipeline, so similar benchmark scores cannot establish robustness by themselves. We then show that recognition and refusal of compositional harm are decoupled: nine standard direction-finding methods detect compositional harm with up to AUC=0.99, yet ablating the recovered directions either leaves refusal intact or destroys model coherence. We therefore propose CMaG, a training-free un-alignment method for compositional safety. CMaG measures refusal intention with a refusal-versus-compliance log-probability margin, takes its gradient with respect to mid-layer activations, and contrasts it against partially benign pairs to isolate the cross-modal interaction term. A single backward pass over 80 examples yields a direction along which steering monotonically converts refusal into compliance. On Qwen2.5-VL-7B and LLaVA-1.5-7B, CMaG reduces compositional safety scores by up to 65 points while over 98% of outputs remain coherent, outperforming gradient-based baselines; the direction also transfers to text-explicit harm and un-aligns models hardened by defensive training. Compositional safety can thus appear behaviorally strong while remaining fragile to targeted internal intervention. Warning: this paper contains unsafe model outputs, including references to self-harm.
benchmark - arxiv:2611.07215 · cs.LGFractional physics-informed neural networks to model social media viralityAlex Koong, George Em Karniadakis
Existing models of social media virality either prescribe the functional form of engagement growth or assume a fixed mechanism by which past activity influences future engagement. We present a fractional physics-informed neural network that learns both the temporal memory and nonlinear growth dynamics of social media engagement, using symbolic regression to translate the learned dynamics into interpretable equations. Synthetic experiments at the deployment resolution of 24 hourly observations demonstrate that the fractional order and reaction law can be jointly recovered from clean data. Under observation noise, however, the standard pointwise physics residual drives the fractional order toward its lower bound through a scale artifact. We address this failure using a fully whitened residual that propagates observation noise through the fractional operator. Applying the framework to 348 posts from X (formerly Twitter), we find that a single-window model estimates strongly memory-dependent dynamics but fails to accurately reconstruct held-out trajectories. Partitioning the 24-hour window into four teacher-forced segments substantially improves performance, yielding segment-specific fractional orders that increase from 0.23 to 0.99 and interpretable growth laws for each phase. The segmented model reconstructs 64% of held-out trajectories within a mean squared error of 0.01, compared with 44% for the strongest of nine classical baselines. Together, these results demonstrate the potential of fractional physics-informed equation discovery to uncover interpretable engagement dynamics while highlighting the importance of noise-aware estimation and temporal segmentation.
memory - arxiv:2611.07208 · cs.LGStructured Machine Theory of Mind from Agent TrajectoriesNigel F. Doering, Rahath Malladi, Arshia Sangwan, M Tanjid Hasan Tonmoy +3
Predictive models of human behavior trained on large-scale trajectory data optimize for statistical accuracy without representing the mental states that causally generate behavior. Such models support prediction but not principled intervention: they cannot answer how an agent's behavior would change if its beliefs or preferences were different. We introduce Structured Machine Theory of Mind (SMToM), a framework that attributes explicit, independently supervised belief and desire representations from observed trajectories within a Belief-Desire-Intention (BDI) causal structure. The central architectural element is a goal head that consumes only the predicted mental-state channels and a current-trajectory embedding; counterfactual intervention on beliefs and desires is a direct operation. We instantiate SMToM on a controlled pedestrian navigation domain where ground-truth mental states are known by construction, enabling evaluation of attribution accuracy, local channel interventions, and trajectory-level counterfactual diagnostics. The resulting model, BDIBottleneck, improves over trajectory-only and unsupervised-context neural baselines. It is also competitive with a matched auxiliary-supervised context baseline while providing an editable mental-state interface that the unbottlenecked baseline lacks. Counterfactual experiments on desires confirm that substituting an agent's inferred preferences with a different activity type coherently shifts predicted destinations toward relevant locations. Counterfactual experiments on beliefs confirm that marking a location as unavailable in the agent's belief state reduces its predicted probability as a destination, though in a complex way. These results demonstrate, in a controlled navigation setting, that explicit BDI-structured supervision is a viable foundation for interpretable and counterfactual analysis of longitudinal trajectory data.
agent - arxiv:2611.07203 · cs.CLτ-bench-LILT: Evaluating Conversational Agents on Realistic Multilingual Tool-Use TasksKaden Uhlig, Kiarash Mokhtari, Anıl Arslan, Yunsu Kim +1
Frontier evaluations of tool-using conversational agents are conducted almost exclusively in English, leaving a critical gap for organizations deploying agents in other languages and regions. We present $τ$-bench-LILT, a multilingual extension of ${τ^3}$-bench covering German, Korean, Turkish, and Farsi in the airline, retail, telecom, and banking domains. We natively localize every layer of the benchmark: customer databases are populated with culturally appropriate records, user-simulator prompts are refined iteratively through contrastive annotation by native-language experts, and tasks are updated according to the localized database and to reflect realistic tasks in the target region. Evaluation across eight agent models shows the largest drops compared to English in the Korean and Farsi retail domains, by up to 61 points in pass^4. Code and data released at https://github.com/lilt/tau-bench-lilt.
agenttool-usebenchmark - arxiv:2611.07194 · cs.AISeVerus: Context-Aware, Gamified Benchmarking for Rust VerificationÁlvaro F. Silva, Martim Afonso, Nuno Saavedra, Emily First +2
Coding agents achieve high success rates on existing benchmarks for Verus, a verifier for Rust programs. These benchmarks typically provide the auxiliary lemmas needed to complete a proof. In practice, however, such lemmas must be discovered and proved. To capture this challenge, we present SEVERUS, a context-aware technique to generate Verus benchmarks that organizes verification tasks into quests, each divided into levelsin a gamified manner. Using the original proof dependencies in verified projects, SEVERUS constructs one quest per verification target and progressively removes proof context to create individual levels. These levels range from the easiest, requiring completing only a single missing lemma, to the hardest, requiring proving the target without any supporting lemmas. An independent checker rejects changes to executable code, weakened specifications, and unsound proof constructs. A noisy binary search estimates the highest level that an agent can solve reliably. Applied to six open-source Verus projects, SEVERUS generates 819 quests comprising 10,095 levels. Experiments with Codex using GPT-6 Sol and GPT-6 Luna, and with Claude Code using Sonnet 4.6, show that success decreases and cost increases as proof context is withdrawn. The strongest agent solves the final level of 52 of 60 sampled quests, yet can only reliably solve 35% of their levels, with failures concentrated in longer quests. Successful GPT-6 Sol proofs rarely reproduce the withheld lemmas, instead, using alternative proof decompositions. By systematically varying proof context, SEVERUS provides a foundation for constructing verification benchmarks that can continue to challenge increasingly capable coding agents.
agentbenchmark - arxiv:2611.07188 · cs.CVMammoSeek: A Mammography Agent with selective zooming for grounded reportsHaoyan Ding, Qian Wang, Lixia Wang, Debiao Li
Radiologists read a screening mammogram by first forming a global impression of the four views and then selectively magnifying suspicious regions before assigning a BI-RADS assessment. MLLMs, in contrast, see the whole image at a fixed, reduced resolution, which makes sub-millimeter calcifications invisible, and they perform close to chance on mammography. Therefore, we present MammoSeek, a mammography agent that reproduces this reading process. After forming an impression of each breast, the agent performs selective zooming: it decides by itself which detector candidates deserve a closer look and inspects only those on full-resolution views. Guided by a BI-RADS assessment skill, it then assigns a category and generates a localized, lexicon-structured report for each breast; its assessment is combined with image-model scores in a decision layer to decide recall. On an internal (EMBED) and an external (VinDr-Mammo) dataset, \ours reaches exam-level AUCs of 0.824 and 0.818, far above other MLLMs, mammography agent and CNN classifiers. Selective zooming also outperforms both no zooming and forced zooming in localization, structured-report F1 and recall decisions at low cost.
agent - arxiv:2611.07176 · cs.CVVINO: Revisiting 3D Representation Learning with Volume TransformersAdrian Kruse, Kadir Yilmaz, Daan de Geus, Francis Engelmann +1
Self-supervised learning has given rise to vision foundation models that learn general-purpose image representations transferable across downstream tasks, yet representation learning for 3D scenes has not achieved comparable success. Motivated by this, we identify key limitations of current 3D representation learning approaches and revisit their core design choices. We first replace the commonly used hierarchical PTv3 backbone with the Volume Transformer, a ViT-like architecture for 3D scenes. We then redesign two core objectives of modern approaches built around student-teacher networks: matching representations across views and predicting masked content from visible context. For the first objective, we separate the roles of teacher and student: the teacher provides clean, consistent targets while the student learns invariance to stronger perturbations. For the second objective, we prevent the student from accessing the masked geometry and instead require it to infer missing representations from visible context only. Finally, we redesign 2D-to-3D knowledge distillation, replacing direct regression of view-dependent image features with view-consistent supervision. Together, these changes form VINO, a framework that learns semantic representations of 3D scenes without labels. VINO substantially improves dense linear probing across indoor and outdoor benchmarks and transfers strongly to out-of-domain settings.
benchmark - arxiv:2611.07169 · cs.LGMeta-Frieren: Beyond Single Level Online LearningVahan Arsenyan, Arnak Dalalyan
We study a two-stage online optimization problem: at each round, a meta-learner first selects an agent-environment pair, then the chosen agent interacts with the environment, and receives a reward. We introduce Meta-Frieren, a general framework for maximizing cumulative reward in this setting. Our analysis yields a simple UCB-style criterion for meta-level selection and recovers prior guarantees as special cases while extending them to broader settings. We instantiate Meta-Frieren with GP-UCB under Bayesian and frequentist assumptions. We also derive, to our knowledge, the first high-probability regret bound for Feel-Good Thompson Sampling, specialize it to finite and parametric function classes, and integrate it into Meta-Frieren. Experiments support our theoretical results and demonstrate that Meta-Frieren can be successfully applied to hyperparameter optimization across multiple model classes.
agentonline learning - arxiv:2611.07166 · cs.LGMulti-step SWE: Can Coding Agents Keep Building on Their Own Code Over Long Horizons?Xiangkai Zeng, Zheng Liu, Zihao Lin
Developers rarely give a coding agent a single request: they ask it to build a component and then to extend, upgrade, and optimize it, each time on code the agent itself wrote. Most benchmarks, however, score one task at a time. We introduce Multi-step SWE, a long-horizon benchmark of 69 software engineering tasks, each a chain of 5-15 dependent steps on a single codebase. Every step is a complete, multi-hour engineering request that starts from the code the agent produced in the preceding steps, so early design decisions and errors carry forward. The tasks span four families of engineering work and nine languages, include both from-scratch builds and existing open-source codebases, and ground every step in an existing source of truth that defines both the instruction and its verifier. Across nine frontier models, the strongest model reaches a solve rate of only 18.5%, and 38 of the 69 tasks are not solved by any model. We release Multi-step SWE publicly at https://multistepswe.com.
agentbenchmark - arxiv:2611.07165 · cs.LGStructural Conditional Independence Test via Neural Quantile ResolutionKhoa Nguyen, Bao Duong, Thin Nguyen
Conditional independence testing is difficult when dependence is weak, localized, or expressed through distributional changes rather than a global mean shift. Existing tests usually return a scalar statistic or p-value, which summarizes evidence against independence but does not describe where the conditional law changes. We propose QRCIT, a quantile-resolution conditional independence test that compares P(X|Z) and P(X|Y,Z) through a quantile deformation learned by a shared model. A lightweight residual quantile-flow model first learns the baseline conditional quantile law and then learns the additional deformation induced by Y, yielding a global predictive-gain statistic with conditional-randomization calibration and an inspectable deformation field over context and quantile level. We also introduce the Structural Dependency Benchmark (SDB), which evaluates dependence structures motivated by real applications, such as context-local, tail-only, sign-reversing, moving-ridge, and mean-neutral patterns. Across synthetic post-nonlinear (PNL) and SDB benchmarks, QRCIT improves detection of structured alternatives, with empirical Type-I behavior evaluated across settings. Higher quantile resolution improves merged-SDB power at moderate sample sizes, while diagnostics restricted to known affected regions improve detection of localized dependence. On Sachs, METABRIC breast cancer, and CCLE/PLX4720, QRCIT provides competitive real-data performance and interpretable deformation patterns. These results support quantile-resolved conditional-law deformation as a practical framework for testing and diagnosing structured conditional dependence.
benchmark - arxiv:2611.07155 · cs.LGDepth or Description? When a One-Step Latent Score Replaces Lookahead in Proactive Information AcquisitionBiswa Sengupta
A proactive agent must decide what to ask before it acts. Exact lookahead values combinations of questions but is expensive and leans on a fitted model; a one-step score is cheap but can miss them. Our main finding: the need for depth often lies in how the hidden cause is described. A one-step score on the label can be blind to an interaction among answers. A score on a suitable latent cause is not, and where that latent encodes the decision it replaces most of depth at a fraction of the search. Described differently, the same problem can defeat the score while exact values do not move, so keep the latent small and check it against exact values. Probabilities learned from one language model's answers do not carry to another unless planning allows for both; those read from tokens are worth only a few labels.
agent - arxiv:2611.07140 · cs.AIMore Agents, More Reliable? System-Level Reliability of Multi-Agent LLMs under Operational-Profile UncertaintyBurkay Tunçtürk, Koorosh Aslansefat, Yiannis Papadopoulos, Bao Nguyen
Teams increasingly build software from cooperating large language model (LLM) agents, often assuming more agents mean more reliability. Yet the usual evidence, average benchmark accuracy, ignores the input mix a deployed system meets and the rule merging the agents' answers. Building on HIP-LLM, a hierarchical imprecise-probability assessment of single LLMs, we propose a framework combining it with a Bayesian network of the system's failure logic for multi-agent systems. Agent failure-probability intervals propagate through this network, where the combining rule can itself fail, into a reliability envelope showing which evidence limits it. In a pipeline extracting regulated chemical-hazard statements, majority voting over three models protected against the weakest but did not clearly exceed the best, and a voter failing on 5\% of demands would largely erase its advantage over the middle model. More agents yield a more defensible reliability claim, not automatically a more reliable system.
agentmulti-agentagent systembenchmark - arxiv:2611.07132 · cs.LGA Proxy-Based Method for Measuring Racial Disparities in~Text-to-Image~GeneratorsNia Mucher, Ngan Pham, Jason Chin, Aaron Brown +1
We present an automated method for auditing text-to-image generators for representational bias against ground-truth demographic benchmarks. Our approach adapts the racial proxy variables used in U.S. antidiscrimination law to provide a fine-grained comparative analysis of variations in racial representation in generated images. We classify each generated image by race and then compare the model's distribution to publicly available demographic benchmark data to identify patterns of over- and under-representation. We also measure skin tone using CIELAB values and calculate cosine similarity amongst groups of images in order to capture biases that don't show up in demographic comparisons such as differences in image quality across races and skin tones. We report on experiments testing three models (Imagen 4 Fast, Stable Diffusion 3.5 Large Turbo, and Flux Schnell) using three categories of proxies (geographic, occupational, and surname) and 50 variants within each category. We find notable variation across the models as well as by proxy category. For example, Imagen 4 Fast overrepresents white classified images more than the other models tested and produces images with lighter, more homogeneous skin tones. Imagen 4 Fast also had the largest representational disparity for Black demographics among surname and geographical prompts. Stable Diffusion 3.5 Large Turbo produces results that overall diverge most from census data but yield the most diverse range of skin tones. Our method yields a generalizable and legally-grounded auditing tool that surfaces detailed information about model outputs without relying on implicit normative definitions of fairness.
benchmark - arxiv:2611.07130 · cs.LGTeaching Agents the "Principle", Not the "Reward"Shan Yang
Reinforcement learning has achieved strong performance in many tasks by letting agents learn from nothing more than a reward signal. However, in more complex tasks, agents still struggle to learn good, generalizable policies, even with carefully designed rewards and abundant samples. We argue that this difficulty stems from the reward decoding burden: designers encode their knowledge of the desired behavior into rewards and penalties, and agents must infer this knowledge from long sequences of noisy scalars. To address this burden, we propose prescriptive policy learning (PPL), which directly teaches agents higher-level knowledge about the tasks, termed principles of behavior, such as how to maintain balance, how to accumulate energy, and how to land smoothly. Through appropriately selected key quantities and references, PPL can directly indicate at every step whether the agent follows the principles and to what extent. Our experiments show that agents can learn good, generalizable policies without any reward. Taught only simple heuristic principles, PPL agents outperform reward-based methods with far fewer samples. Their policies also transfer directly to harder settings and keep performing well, demonstrating their strong generalization ability. In summary, PPL provides a new approach in which rough knowledge of the desired behavior is enough for agents to learn good, generalizable policies, even when the system dynamics cannot be accurately modeled.
agent - arxiv:2611.07127 · cs.LGScalable Bayesian Gaussian Process Classification for Cancer Histopathology: Calibrated Uncertainty Under Distribution ShiftYunru Zheng, Wun Ting Chan, Sampras Dsouza, Zhenyu Hong
Reliable uncertainty quantification is essential for clinical deployment of cancer diagnosis models, yet neural networks and variational Gaussian process methods can produce overconfident predictions, particularly under the distribution shifts that arise when models trained at one hospital are applied at another. We present a scalable Bayesian Gaussian Process Classification (GPC) framework that operates on frozen pathology foundation model embeddings (Phikon) and combines accelerated Randomly Pivoted Cholesky (RPCholesky) for low-rank kernel factorization, Hamiltonian Monte Carlo (HMC) with non-centered parameterization for posterior sampling, and a pivots-as-inducing-points scheme for prediction. This makes HMC posterior sampling in the low-rank latent representation tractable at N approximately 300,000 training points, enabling HC-based low-rank GPC on substantially larger datasets than prior MCMC-based demonstrations. We evaluate on two histopathology benchmarks for breast cancer metastasis detection: PCam and CAMELYON17-WILDS, where the latter tests calibration under hospital-level distribution shift. Compared to stochastic variational GP and neural network baselines, our method improves calibration and predictive performance - reducing expected calibration error by up to 88% and false negative rate by up to 90% - with the largest gains appearing under distribution shift, suggesting that HMC-based posterior sampling provides more trustworthy uncertainty estimates for cross-site pathology classification.
benchmark - arxiv:2611.07126 · cs.AIAdapting LLM Judges to Observed Human Annotation Policy through Soft Prompt TuningTamás Janusko, Sven Kleinkop, Ricardo Bochnia, Gunnar Hempel +3
Large Language Models (LLMs) are increasingly employed as automated judges for evaluating chatbot responses. While proprietary models can achieve high agreement with human annotators, they remain costly and difficult to deploy in practical settings. Lightweight open-source LLM judges provide an attractive alternative but typically exhibit lower agreement with human ratings. In this paper, we investigate whether soft prompt tuning can align open-source LLM judges more closely with human annotation policy. We train soft prompts on the FairBotBench benchmark and evaluate the adapted judges across the dimensions of bias, toxicity, and personalization. Our findings demonstrate that hybrid soft-prompt tuning, which combines textual prompts with learned continuous prompt embeddings, consistently improves agreement with human annotations over the corresponding untuned baselines. For Qwen 3.5 9B, the adapted judges surpass both hosted reference judges across all three evaluation dimensions.
benchmark - arxiv:2611.07124 · cs.LGMachine Learning for Live-Birth Prediction After IVF/ICSI From Structured Clinical Data: A Systematic Review, Meta-Analysis, and Critical AppraisalAbdul Raheem Shahzad, Muhammad Zaheer Khan, Umar Ajaib, Asad Ur Rehman +2
Machine-learning models predicting live birth after in vitro fertilisation (IVF) or intracytoplasmic sperm injection (ICSI) from routine clinical data report performance from near chance to near perfect. We systematically reviewed these models under PRISMA 2020, searching OpenAlex, Scopus and Web of Science to 22 October 2026, appraised them with PROBAST+AI and TRIPOD+AI, and pooled their discrimination with random-effects meta-analysis on the logit scale. We define what fixes the meaning of a model's output (decision point, endpoint, unit of analysis, data source) and map the 48 included studies by model family, decision point and validation design. Models evaluated only internally pooled to an area under the receiver operating characteristic curve (AUC) of 0.814, against 0.756 (95% prediction interval 0.546-0.889) for models evaluated on a later period or another centre, and 0.764 for studies not at high risk of bias; neither difference was significant, and heterogeneity was extreme. Models whose decision point was not stated pooled to 0.910, the only significant moderator in meta-regression. Against logistic regression fitted to the same data, the best machine-learning model gained a pooled 0.056 AUC (95% CI 0.017-0.090). Of 48 studies, 47 were at high or unclear risk of bias, only four assessed the calibration slope, and no model has evidence of clinical impact. The certainty of the pooled evidence is low to very low; the gains that survive validation came from staging predictions to decision points and local calibration, not deeper architectures. We provide a TRIPOD+AI-mapped checklist, a benchmark specification and a research agenda.
benchmark - arxiv:2611.07122 · cs.LGRegulating Autonomous Mobility-on-Demand Operators with Fixed and Adaptive Penalties: A Multi-Agent Reinforcement Learning ApproachKorbinian Kügler, Carolin Schmidt, Florian Dandl, Filipe Rodrigues
Profit-maximizing Autonomous Mobility-on-Demand (AMoD) operators may exploit consumer surplus through excessive prices and underserve less profitable regions by relocating vehicles away from them. We study how penalty-based regulation of an AMoD operator in a multimodal mobility system can limit surge pricing and improve regional service availability. We model regulation as an outer control loop that acts on a learning-based operator. Static regulation with fixed penalties is formulated as a Markov Decision Process for an operator that jointly controls pricing and rebalancing, and dynamic regulation as a Partially Observable Stochastic Game in which a regulator periodically adjusts region-specific penalties in closed loop. Operator and regulator policies are learned with graph-based multi-agent reinforcement learning. In experiments with real-world Manhattan data, fixed surge pricing penalties reduce the share of travelers facing excessive prices from 22% to 1%. For regional service availability, dynamic regulation improves the regulator's reward by up to 15.6% over fixed penalties of equal mean intensity. The gains in service availability are strongest when spare fleet capacity exists, whereas service deficits persist in peak periods with highly utilized fleets. The learned regulator policy responds by reducing penalties where limited fleet capacity restricts compliance. The results identify trade-offs between regulatory targets, operator profitability, and traveler utility.
multi-agent - arxiv:2611.07110 · cs.LGCredibility Clustering for Experience Rating: An Automated Approach to Risk Group Allocation in Group Life InsuranceThom van Rijn, Frank Weber, Michael Merz
In insurance, experience rating plays a crucial role in mitigating adverse selection by adjusting premiums according to historical claims experience. A key step in this process is the classification of insured collectives into homogeneous risk groups. Within group life insurance, this classification traditionally relies on manual decisions and arbitrary thresholds, which may lead to inefficiencies and inconsistent outcomes. This paper proposes a data-driven methodology for automating risk group allocation based on a combination of the multidimensional Bühlmann-Straub model and k-means clustering -- referred to as credibility clustering. By incorporating credibility weights, the method accounts for the varying statistical significance of experience data and ensuresthat risk classifications are both actuarially sound and practically implementable. Empirical results based on real-world group life insurance data from a major European insurer demonstrate that the proposed framework outperforms the two k-means benchmark variants considered in the empirical application in terms of viability, temporal stability and accounting for statistical reliability.
benchmark - arxiv:2611.07103 · cs.LGA Night at the LLM Zoo: Data Conditioned Flow Matching for Generating Pretrained LLM WeightsScott Biggs, Bedionita Soro, Saumya Gupta, Bruno Andreis +1
A challenge to Weight Space Learning as it expands from task specific and vision models to Large Language Models (LLMs) is the absence of a well curated and analyzed LLM population. To address this, we train such a model zoo for application to Weight Space Generation by systematically varying the data mixture composition between web text, books, mathematics, coding, and French across populations of 51M, 124M, and 355M parameter GPT-2 Models, each tier with count N = 100. We analyze these populations by cosine similarity, functional diversity, Leave-One-Mixture-Out modeling, and variance decomposition. We show that the zoo contains diverse models whose variation is structured by training mixture composition. We embed the training mixture composition to condition a Flow Matching model of complete pretrained LLM weights in Gram-PCA compressed weight space. By demonstrating zero-shot generalization to held out mixtures, measuring Continuous Ranked Probability Score skill, as well as finding minimal changes in downstream test-set perplexity and general capability benchmark scores, we show that our generator generalizes on the simplex of data mixtures to produce models that perform similarly to those in the zoo ex nihilo. Finally, we produce all source code and artifacts to support the continued expansion of Weight Space Learning and Generation into the LLM domain.
benchmark - arxiv:2611.07093 · cs.LGServer-Side Sparse Assignment for Federated Mixture-of-ExpertsXue Zheng, Aylin Yener
Large language models (LLMs) are increasingly recognized for their remarkable capabilities; federated learning (FL) offers a promising framework for collaboratively fine-tuning LLMs on edge devices while achieving data privacy. However, edge devices are inherently heterogeneous: clients may exhibit diverse data distributions and resource constraints, including limited memory and computation budgets. Recent federated Mixture-of-Experts (MoE) approaches address these challenges by using sparse expert activation or partial expert loading. However, existing partial-expert methods rely on heuristic full-pool observations or prioritize system-wide load balance, without guarantees for recovering client-specific relevant experts from partial observations. We propose $S^3$-MoE, a server-side sparse expert assignment framework that estimates client-specific expert relevance from partial observations and refines assignments through an exploit--explore mechanism. We also establish relevance-estimation error and exact Top-$K$ recovery guarantees. Experiments using a CNN-based MoE on CIFAR-100 show that $S^3$-MoE recovers task-dependent client--expert preferences and reduces conflicts caused by correlated expert selection across clients. Compared with Full-20 FedAvg, which gives each client access to all 20 experts for local Top-5 routing, $S^3$-MoE achieves 1.31 percentage points higher mean final accuracy while loading only $5$ experts per client.
memory - arxiv:2611.07089 · cs.AIPolynomial-Time Computation of Extensive-Form Proper Equilibria in Two-Player Zero-Sum GamesSam Ganzfried
We give a deterministic polynomial-time algorithm for computing an exact extensive-form proper equilibrium in a finite two-player zero-sum extensive-form game with perfect recall and rational data. Properness is understood in the agent normal form: an action at an information set is penalized relative to a better action while all other local behavioral choices, including the player's own future actions, remain fixed. Our algorithm assigns each information set a geometric permutahedron of allowable action probabilities. A scaled doubly stochastic representation embeds these restrictions in a polynomial-size sequence-form polytope; a saddle point then satisfies the local $\varepsilon$-properness inequalities. To obtain an exact equilibrium rather than a solution at an arbitrarily small numerical perturbation, we establish sign stability of a parametric linear system, construct a persistent feasible basis, and extract the limits of behavioral ratios. The resulting probabilities are rational and have polynomial encoding length, including at information sets whose realization probabilities vanish.
agent - arxiv:2611.07081 · cs.CVWide View Synthesis from Overlapping Perspective Cameras in Autonomous DrivingKaixuan Zhang, Zhipeng Xiong, Minxian Li, Mingwu Ren +1
Multi-perspective-camera systems are widely adopted since they provide broad visual coverage for autonomous driving, motivating the synthesis of a unified wide view from spatially separated observations. In this work, we formulate wide view synthesis (WVS), which aims to synthesize a geometrically consistent wide view at a specified virtual viewpoint from multiple perspective cameras. Unlike conventional image stitching, WVS requires reasoning about 3D geometry and visibility to handle depth-dependent parallax and occlusions arising from spatially separated camera centers. While vanilla 3D Gaussian Splatting (3DGS) enables high-fidelity view synthesis, its reliance on per-scene optimization limits its applicability to driving scenes. Furthermore, WVS introduces additional challenges, *e.g.*, peripheral stretching under perspective projection and photometric inconsistencies across cameras. To address these challenges, we propose **Wide-View Gaussian Splatting (WV-GS)**, a feed-forward Gaussian-splatting framework that synthesizes wide views without per-scene optimization. Specifically, we introduce Piecewise Perspective Rendering (PPR) to mitigate peripheral stretching through local perspective projections sharing a common camera center, and Spatial Affine Photometric Modeling (SAPM) to account for camera-dependent photometric variations during training. We further introduce **WideDrive**, a synthetic benchmark providing dense ground-truth wide views and adapt a LiDAR-projected sparse-reference protocol to real-world datasets. Extensive experiments on multiple datasets demonstrate the effectiveness of WV-GS for WVS. Dataset and code will be publicly released.
benchmark - arxiv:2611.07077 · physics.opticsReliability of Fisher-information bounds for dielectric metalenses: a site-resolved RCWA auditJ. Sumaya-Martinez, O. Olmos-Lopez
We assess the reliability of localization precision predicted for a 1716-nanopillar dielectric metalens using a site-resolved complex RCWA library and a vectorial Poisson Fisher model. An independent calculation reproduces the archived local axial Cramer-Rao bound of 20.7738 nm to within 0.0044 nm, yielding 20.7695 nm under the same ideal detector and 3000 detected signal photons. High-order periodic-cell calculations reveal a 0.1335 rad phase difference at a selected diameter; different substrate-index assumptions prevent attributing this difference solely to modal truncation. Replacing the recalculated coefficients at 39 actual sites changes the bound to 20.7522 nm. Restricted phase-mask benchmarks and pixel-integration checks further distinguish a stable statistical prediction from electromagnetic convergence. We provide a reproducible, model-conditional benchmark, not a physical-device performance claim.
benchmark - arxiv:2611.07066 · cs.AICooperative Multi-Agent Reinforcement Learning via Epigraph-Form Guided ExplorationSungil Son, Hoseong Jung, Dahyun Oh, H. Jin Kim
Discovering cooperative behavior in multi-agent reinforcement learning (MARL) is challenging due to the combinatorial complexity of joint state-action spaces, which hinders the emergence of coordinated behaviors from trial-and-error alone. Intrinsic rewards are often used to aid discovery, but naively combining them with team objectives can distort the learning signal, compromising task performance. In this paper, we propose EFXPLORER, a constrained exploration framework that maximizes exploration objectives subject to a constraint that preserves established task performance. We solve the resulting problem using an epigraph reformulation that introduces adaptive exploration budgets. This approach separates intrinsic rewards from task objectives and regulates exploration through task feasibility. To further encourage diverse and temporally extended exploration, we incorporate a successor distance-based intrinsic reward that captures long-horizon dependencies. Empirically, our method outperforms strong baselines and induces novel cooperative strategies across SMAX, VMAS, and MPE benchmark suites.
multi-agentbenchmark - arxiv:2611.07063 · cs.CVSD-LoRA: Training-Time Structural Distillation into LoRA for Few-Shot Vision-Language AdaptationMohammed Rahman Sherif Khan Mohammad, Ardhendu Behera, Sandip Pradhan, Swagat Kumar +1
Few-shot adaptation of vision-language models must balance two competing goals: learning fine-grained visual distinctions from very limited supervision and preserving the efficiency that makes frozen CLIP-like models attractive for deployment. Existing lightweight adaptation methods, including prompt tuning, LoRA, and cache-based adapters, typically supervise global image embeddings and therefore underuse token-level structure. Conversely, structure-aware methods exploit local evidence but often introduce additional inference-time computation. We propose SD-LoRA, a structural distillation framework that uses token-level graph reasoning only during training and transfers its effects into a lightweight LoRA-adapted CLIP model. During training, a heterogeneous graph teacher models patch-patch, global-local, and visual-text relations over high-resolution ViT tokens. Rather than retaining this graph at test time, SD-LoRA transfers its structural bias into the shared CLIP-LoRA backbone through implicit distillation and Prototype Predictive Alignment, which aligns graph-conditioned representations with support-derived class prototypes. At inference, the graph teacher is discarded; prediction requires only a single CLIP-LoRA forward pass and a cache-based classifier. Across 11 few-shot benchmarks, SD-LoRA improves over a strong CLIP-LoRA baseline by 2.4/1.8/3.0/2.4/2.0 percentage points in the 1/2/4/8/16-shot settings, respectively, without additional graph computation at test time. Results on base-to-novel generalization, cross-dataset transfer, out-of-distribution robustness, and calibration further show that training-time structural supervision improves both accuracy and transfer reliability. Code is available at https://github.com/MR-Sherif/SD-LoRA.git.
benchmark - arxiv:2611.07055 · cs.CVExpanding the Quality-Coverage Frontier with Competitive Flow MatchingWenxuan Peng, Richard Zhang, Haitian Zheng, Jaskirat Singh +5
Flow matching (FM) generators often face a quality-coverage tradeoff. Inference-time techniques such as classifier-free guidance raise quality, but at the cost of coverage. We hypothesize this limitation stems from a basic property of the objective: an intermediate noisy image could correspond to many clean images, yet the model can only predict a single velocity, at best averaging over the multiple possibilities. To mitigate this, we introduce competition among predictions during training. We achieve this by augmenting the predictor with a randomly drawn low-dimensional conditioning code $z$. For each training example, we sample $K$ values of $z$ to produce $K$ candidate velocities that compete to explain the target, and apply the loss only to the best-matching candidate. This winner-take-all objective allows different modes to shape different predictions, rather than being averaged into a single compromise. Furthermore, this term naturally enables a new form of guidance, auxiliary conditioning scaling, which complements traditional guidance methods. Together, these two guidance mechanisms show that our method substantially improves the quality-coverage frontier. We validate on ImageNet-256, using two different testbeds, SiT and RAEv2. Given less than 3% of the original SiT pretraining compute for post-training, at the 67% Recall point, our method reduces $\mathrm{FD}_{r,6}$ from 19.22 to 5.91. On RAEv2, at the 70% Recall point, it reduces $\mathrm{FD}_{r,6}$ from 16.83 to 3.18.
post-training - arxiv:2611.07054 · cs.AIReclassification Is All You Missed: Surfacing Variant Reinterpretation Need in Online Health Communities at Population ScaleKamalakkannan Ravi, Mengshu Nie, Annapurna Poduri, Cong Liu
Genetic variants are routinely reclassified in ClinVar after a patient is tested, sometimes from uncertain to pathogenic, yet patient recontact is inconsistent, with no clear duty and responsibility diffused among labs, providers, and patients. Many affected people discuss their results in public online health communities that clinical recontact often does not reach. We built PRISM, an agentic large-language-model pipeline that screens these posts, checks whether the variant was later reclassified, and drafts a disclaimer-bearing outreach message, in a no-contact dry run. Across 3.9 million posts from 32 communities, agent-predicted reinterpretation-relevant discussion is rare (0.288%) but concentrated in specialist genetics communities (6.30% versus 0.10% in general medical-advice communities) and growing fivefold since 2014; PRISM surfaces 123 candidate classification differences, 5 of them from an uncertain submission to pathogenic or likely pathogenic. On a held-out set of 30 real-post drafts, a clinician judged 27 factually grounded and 15 fit to send: factual accuracy is necessary but not sufficient for clinically acceptable outreach. PRISM contributes a population-scale reinterpretation-need map with code and data to be released upon publication.
agentic - arxiv:2611.07051 · cs.AIClosed-Loop Agent Evaluation: Layered Measurement and Self-CalibrationRuifeng Ren, Gun Ahn
Once an agent is live, teams want to know whether it is doing its job, not just whether it throws errors. Yet most evaluation is one-off: a benchmark run, a look at traces, a single score. Its conclusions neither check themselves nor reach the agent. We propose closed-loop evaluation, an evaluation method for deployed agents. It models evaluation as performance estimation under biased raters and censored outcomes, and it requires every level's conclusion to trace back to atomic data and to be computed by versioned rules. The method covers four levels of analysis (span, trace, conversation and engagement), measures along a spine of Execution, Substance, Trajectory and Results, treats Conduct as a constraint throughout, and uses Provenance and Adjudication components to certify the evidence and the evaluators. Evaluator agents generated by an Agent Factory carry these components, each with a domain knowledge pack and an adapter harness. Two loops close the evaluation: an inner loop audits the evidence, rules and metrics themselves, and an outer loop uses per-component regression checks to drive agent improvement. We also give a pipeline that runs this process automatically. Using the business-analysis agent SharpInsight as a case and porting the method to public conversation data, we show how problems differ across levels and how evidence-and-rule judgment is more reliable than holistic scoring (ICC 0.86 vs. 0.81; kappa 0.81 vs. 0.63).
agentbenchmarkevaluator - arxiv:2611.07050 · cs.MAAuthorization Propagation in Agent2Agent Systems: An Experimental Security Evaluation of Multi-Hop DelegationGabriele Donno, Kashif Mehmood, Katina Kralevska
Delegated Agent2Agent (A2A) workflows create a security problem when the agent sending a request is different from the principal from which the authority originated. This paper evaluates a layered downstream-verification design in a synthetic four-party healthcare reimbursement workflow. The prototype separates current-caller context, transitive authorization evidence, per-hop Ed25519 request authentication, and independent replay state for requests and authorization artifacts. Of 15 adversarial conditions, 14 are rejected; an isolated delegation-path test exposes the absence of semantic authority attenuation. A separate authorized-intermediary experiment shows that a holder of the shared authorization-evidence HMAC key can modify the represented authority and generate valid evidence. In the fixed-depth local benchmark, the full configuration has a median security-processing latency of 5.11 ms and a 3.87 kB serialized payload. The results show the value of separating request authentication from inherited authority while identifying origin-verifiable delegation and semantic attenuation as requirements for stronger A2A authorization propagation.
agentagent systembenchmark - arxiv:2611.07048 · cs.AIAdapting the Process of Adaptation: Darwinian, Lamarckian, and Agent-Directed Evolution in AIAndrew W. Lo, Chaoyi Zhao
Evolution has shaped intelligence for billions of years. Artificial intelligence (AI) now raises a striking possibility: intelligence may begin to direct its own evolution. As AI agents become capable of modifying and evaluating their own behavior, how quickly could they improve, and how can that improvement remain safe? We develop a theoretical framework comparing Darwinian, Lamarckian, and agent-directed evolution following environmental change. Darwinian evolution relies on natural selection, Lamarckian evolution adds inherited learning, and agent-directed evolution allows agents to generate and evaluate behavioral variants and choose how extensively to search. When learning is sufficiently effective, these mechanisms exhibit progressively faster adaptation. Generating more variants accelerates adaptation, while optimal search effort depends on learning ability, problem difficulty, and experimentation costs. However, an agent's ability to direct its improvement does not ensure that improvements are evaluated against the intended objective. When the evaluation target is misspecified, additional search accelerates convergence toward the wrong behavior, even as measured performance improves. In recursive self-improvement, such errors may persist across generations, posing challenges for AI safety. Independent evaluation and oversight must keep pace with autonomous AI development. The central challenge is not simply to accelerate the evolution of intelligence, but to preserve humanity's ability to guide its direction, limit its speed, and intervene when necessary.
ai agentself-improvement - arxiv:2611.07042 · cs.CLFollow the Policy, Fail the Benchmark: Auditing and Revising Policy-Based Decision BenchmarksYanru Xiao, Yifei Cai
Policy-based decision benchmarks evaluate whether systems apply written rules to financial and operational cases. Yet a rule-following response can be marked wrong when the benchmark misstates the required decision. We audit nine public policy-based decision benchmarks through source tracing, program execution, and controlled checks. We find unsupported targets, reference errors that survive replay or propagate as solved steps, and evaluators that reject acceptable answers or credit incomplete decisions. To measure the scoring impact, we rescore unchanged model responses. On the 96 LegalBench SARA numeric cases, correcting the scorer raises acceptance by more than 80 percentage points for each of three models, without changing references or responses. In RuleArena Tax, restoring a required zero floor changes 27 of 300 references and narrows an observed model gap. A controlled federal-exemption task with explicit ownership and eligibility assumptions also shows that errors in supplied intermediate answers can affect final decisions. We present BAR (Benchmark Audit and Revision), a versioned collection of audit evidence and revisions for reevaluation.
benchmarkevaluator - arxiv:2611.07025 · cs.LGLowRAM: Low-Rank Keys and Amortised Values for Composable KV Cache Compression at Test-TimeTeresa Delgado, Hugo Hazard, Martin Benfeghoul, Jun Wang +1
Autoregressive decoding relies on the KV cache, whose memory footprint grows linearly with context length and caps batch size and throughput. Existing post-training methods exploit redundancy across tokens, layers, and feature dimensions, or lower numerical precision. We identify a complementary axis: within each attention head, values are largely predictable from their corresponding keys. We exploit this with value cache amortisation, which fits a lightweight per-head predictor at test time and stores residuals only for the tokens it reconstructs worst, giving $4\times$ near-lossless value compression. The predictor can be initialised by meta-learning for substantially reduced prefill overhead, or fitted from scratch in settings that allow a one-off cost, such as multi-query inference. We pair value amortisation with cross-layer low-rank key compression to produce LowRAM, an asymmetric compression pipeline that stays close to lossless across three model families. By reconstructing only a selected subset of KV pairs at each decoding step, LowRAM can keep decoding overhead minimal and achieve up to $5\times$ higher peak decode throughput at $4\times$ compression. LowRAM further composes with quantisation and eviction, reaching $26.9\times$ compression at $94.5\%$ of baseline performance on long contexts. These results show that treating keys and values as structurally distinct enables efficient, aggressive, and composable KV cache compression.
memorylong contextpost-training - arxiv:2611.07009 · cs.CVTraCE: Trajectory-Aware Concepts for Explainable Video UnderstandingWooil Lee, Junpyo Hong, Jongseo Lee, Tuomas Oikarinen +2
Video understanding models often rely on motion, yet their explanations can obscure which motion sources support a prediction. We identify motion-source under-specification as a limitation of existing video XAI: region-centric explanations can be source-ambiguous, while motion vocabularies restricted to single-person pose sequences are source-incomplete. We propose TraCE (Trajectory-Aware Concepts for Explainable Video Understanding), a concept-based framework that discovers source-flexible motion concepts by clustering optical-flow trajectories from training videos. These concepts represent dynamic evidence from people, objects, and interactions, while object and scene concepts account for static context. The resulting typed trajectory/object/scene interface supports class prediction for action recognition and question-conditioned concept matching for video question answering (VQA). We evaluate TraCE across five random seeds on six benchmarks: NTU RGB+D Mutual, Chiral SSv2, UCF-101, KTH, TOMATO, and STAR. Relative to non-interpretable baselines, TraCE maintains competitive task performance, with an average accuracy drop of 1.0 points for action recognition and a nominal gain of 0.4 points for VQA. Motion perturbation and concept-removal tests provide complementary evidence of directional semantic alignment and decision faithfulness, while sample-level interventions illustrate targeted model debugging. Our project page is available at https://leewooil.github.io/TraCE_project_page/.
benchmark - arxiv:2611.06996 · cs.AITrust Responsive Runtime Authorization for Agentic AI with Empirical Characterization of Policy Behavior and Parameter SensitivityAhsan Habib Shihab, Imre Felde, Ramis Akhmedov, Stefan Gubo +3
The policy that converts runtime estimates of an autonomous agent's condition into operational authority is rarely specified or measured. We empirically characterize the Trust Responsive Authorization Policy (TRAP), which maps an Authorization Readiness Score (ARS), a weighted combination of runtime trust, operational stability, and contextual risk, onto discrete authorization levels through an asymmetric rule: escalation requires consecutive observations above a band edge plus a margin, while a sharp drop in readiness restricts authority at once. On one matched set of 100 Monte Carlo traces, we compare TRAP with static, trust-only, risk-only, and symmetric policies. Under abrupt degradation, TRAP restricts at onset and re-escalates sooner after recovery than the symmetric policy, with the highest and, descriptively, least variable stability among the adaptive policies, although more of its post-recovery escalations occur below the selected trust threshold. Paired bootstrap intervals exclude zero for the remaining differences. In one-factor sweeps, recovery time rises with the escalation margin and run length, shifts with the score weights, and is unchanged by the restriction threshold. In an exploratory post-freeze experiment, when the same degradation is spread over 5-20 intervals, TRAP restricts only on the 4 traces that retain a single-interval drop above that threshold; averaged over all 100 traces, it holds authority above what readiness supports for 15.9-18.9 of the 20 degradation intervals, against 0.8 under abrupt degradation and about 10 for the symmetric policy. Its restriction advantage therefore depends on degradation being abrupt. All reported values are traceable to the publicly available, versioned research release.
autonomous agentagentic - arxiv:2611.06995 · cs.AIAutonomous AI Agents Systematically Diverge From Human Preferences in Proof FormalizationBálint Gyevnár, Simon DeDeo
Artificial intelligence (AI) systems have ushered in a new age of proof abundance, but helping humans understand these proofs requires new attention to the differences in how AI agents and humans pursue a proof-making task. We present "simulated collaboration", a methodology that allows us to examine two key questions: (1) the relative attention AI agents pay to surface syntax or, conversely, deeper underlying semantics, and (2) how and where AI agent preferences for what to do next diverge from human preferences. On a set of nine major human collaborations in Lean with 4.7M definitions across more than 6K commits, we find that, over the course of proof construction, AI agents (1) are far more strongly drawn than humans to explicitly marked gaps; (2) show a relative aversion to densely connected proof centers and to complex or surprising proof goals; and (3) are relatively attracted to surprising proof bodies, which they tend to regularize toward baseline. Our results highlight important blind spots and potential interventions for future human-AI collaboration in mathematics.
agentai agent - arxiv:2611.06993 · cs.CVBatch-Conditioned Semantic Anchors for Robust Transductive Adaptation of Vision--Language ModelsMohammed Rahman Sherif Khan Mohammad, Ardhendu Behera, Sandip Pradhan, Swagat Kumar +1
Transductive adaptation improves vision-language models at test time by refining predictions on an unlabeled batch, but current methods face a key trade-off. Aggressive approaches exploit batch structure but can drift when batches are sparse, have prior shift, or cover little of the label space. Conservative methods better preserve CLIP text semantics, but fixed semantic references can under-adapt even when the batch provides strong corrective evidence. We propose JPTA (Joint Prior and Transport Adaptation), a prior-aware framework that casts transductive VLM adaptation as batch-conditioned semantic reference estimation. Instead of treating CLIP text prototypes as fixed priors or freely replaceable, JPTA views them as semantic references that shift toward image-side batch structure only under calibrated evidence. It estimates batch priors and soft image-side prototypes to form transported semantic references whose impact is scaled by batch reliability. Across standard benchmarks, low effective-class regimes, all-class evaluation, and online streams, JPTA consistently outperforms TransCLIP and StatA, with the largest gains on sparse and prior-shifted batches. Anchor-drift, active-support, and absent-class-mass diagnostics show that JPTA fixes effective-class mismatch without unconstrained transductive drift, indicating that robust transductive VLM adaptation should recalibrate semantic references using reliable batch evidence rather than choose between unrestricted transduction and fixed prototypes. Code is available at https://github.com/MR-Sherif/JPTA.git.
benchmark - arxiv:2611.06989 · cs.AIThe Two-Ledger Model: Transaction-Grain Cost Accounting for Multi-Agent AI Systems Across Clouds, Models, and Systems of RecordVivek Acharya
Enterprises run business transactions through multi-agent AI systems spanning clouds, large language models (LLMs), orchestration harnesses, and systems of record. Finance receives AI cost only at settlement grain, aggregated by SKU and charge period. The business question (what did this order, claim, or case cost in AI?) cannot be answered from billing data, because providers never see enterprise transaction boundaries. The FinOps Open Cost and Usage Specification (FOCUS) gives billing data one schema across providers but remains settlement-grain even as it adds AI token and requester attribution. We formalize this grain mismatch, derive seven requirements, and propose the Two-Ledger Model: a settlement ledger of FOCUS-conformant billing data joined to an attribution ledger of invocation-grain usage records built on OpenTelemetry GenAI telemetry. A layered correlation spine (trace context, A2A/MCP metadata, business-key correlation, probabilistic matching) assigns each record to its business transaction with recorded method and confidence; a versioned rate card prices usage in near real time; period reconciliation ties both ledgers to the invoice, and the unexplained residual identifies shadow AI. Costs are reported per order, claim, or outcome with agentic cost metrics. A simulation with known ground truth shows the spine sustaining cost coverage above 99% as trace propagation falls to zero, the reconciliation residual recovering planted unmetered spend within 0.24 percentage points while flagging provider billing errors, and mid-flight gating cutting p99 transaction cost by up to 61% at a 2-4% false-pause rate. We propose the attribution ledger as an invocation-grain FOCUS profile for standardization.
multi-agentagentic - arxiv:2611.06988 · cs.LGBig Models, Small Gains: Frozen Forecasters as Classifiers on MONSTERRishit Jain, Rishit Kar, Chetashri Bhadane, Nilesh Patil
Forecasting foundation models are increasingly used as frozen feature extractors for time series classification, but it remains unclear whether their pretrained representations offer an advantage over strong classifiers that do not require pretraining. This raises an important question: do these representations provide useful information beyond what the model architecture alone can capture, particularly when labelled data are scarce? We present a controlled benchmarking of two forecasting models, Reverso and Chronos-Bolt, on six large datasets from the MONSTER benchmark, using 1%, 10%, and 100% of the available training labels. We compare their frozen representations against four established time series classification methods and random-weight versions of the same forecasting models. Our results show that pretraining improves classification accuracy over random-weight counterparts on five of the six datasets, with gains of up to 23 percentage points. However, the frozen representations do not reliably outperform the strongest non-pretrained baseline in any setting: they perform worse in 15 of 17 settings and tie in the remaining two. Further experiments show that restoring amplitude information recovers part of the performance gap, while fine-tuning and forecast-error features do not close it. These findings highlight the importance of strong task specific baselines and random weight controls when evaluating forecasting foundation models for classification. Our code is available at https://github.com/Rachet2006/frozen-forecasters-monster.
benchmark - arxiv:2611.06984 · cs.AIResource-Efficient Reinforcement Learning Post-Training through Elastic Expert ParallelismJia Wei, Ziyi Qiu, Yi Xu, Xuehai Qian
Reinforcement learning post-training improves the capabilities of large language models, but generating training responses is a major computational bottleneck. In mixture-of-experts models, workers that finish their assigned responses may remain occupied because their experts are still needed by workers processing longer responses. Consequently, approximately one-third of occupied worker time in the baseline occurs after local response generation finishes. Here we present FlexFloor, a system that enables earlier resource release during synchronous response generation while preserving the expert computation required by ongoing responses. FlexFloor jointly adapts response scheduling and the active group of workers. It uses historical response lengths and memory constraints to guide batching and prompt placement, then progressively releases workers while maintaining expert availability and correct communication. Compared with the verl baseline, FlexFloor generates approximately 65.40% more response tokens per unit of occupied computing time and increases response generation throughput by 52.74%, with little observed variation in rollout reward. By making reinforcement learning post-training more efficient, FlexFloor supports faster development and iterative improvement of large language models and AI agents.
memoryai agentpost-training - arxiv:2611.06977 · cs.AIBlockLock: Reproducible LLM Inference with Fixed Batch GeometryXiuyuan Qi, Pedro Reviriego, Javier Conde, Shanshan Liu
Reproducible LLM inference requires that a request produces the same output regardless of the other requests in the batch. Existing solutions pay for this in different ways: batch-invariant kernels constrain every operator, and isolating prefill with fixed-shape decode ({our CoRun-style baseline}) gives up shared prefill compute and pads decode. We present BlockLock, which keeps shared batched prefill and pins only the numerical path of the protected request: it fixes the geometry that the request sees, with {two modes: BlockLock-Rows uses fixed-width rows and retains batched decode of all requests; BlockLock-Pack uses terminal-anchored packing with an isolated dummy segment, then extracts the protected request's KV state for target-only decode. Under a stated per-deployment contract, both modes preserve the target's output bitwise when its companions change.} Across five {dense} models (Qwen3-8B/14B, Llama-3.1-8B-Instruct, and Gemma-3-4B/12B-IT) and three single-GPU deployments (NVIDIA A6000, A100, and H100), we evaluate our approach on both tinyBenchmarks and BFCL{; we also include initial tensor-parallel experiments on two NVIDIA A6000 GPUs. Both modes preserve target sequences and Top20-logit bits in the reported tests.} At equal delivered work, {BlockLock-Rows} reduces latency by 53.5--55.0\% and improves throughput by 115.2--122.2\% relative to vLLM Batch-Invariant; compared with batch-invariant kernels, it needs no custom kernel implementations and constrains only the protected requests.
benchmark - arxiv:2611.06976 · cs.CLOn the Opportunities and Challenges of Prefill-to-Autoregressive Cross-Model TransferZhenyi Shen, Junru Lu, Wei Liu, Jiangnan Ye +7
Prefill-to-Autoregressive Cross-model Transfer (PACT), a form of latent communication between language models, combines a larger model's prompt representations with a smaller model's efficient generation, offering a promising way to balance capability and inference cost. However, existing systems provide limited support for independently evaluating large-small model collaboration, and improved task performance alone does not establish whether the receiver benefits from the sender's information. We introduce PLatour, an open-source PACT framework that connects frozen models through lightweight, zero-initialized linear bridges trained directly with task rewards and supports hybrid-attention architectures. PLatour significantly outperforms open-source baselines on generation tasks and surpasses Mostik's reported results on shared mathematics and coding benchmarks using a substantially smaller sender. However, controlled experiments show that the gains on some tasks can be reproduced by uninformative sender signals, self-communication, or receiver-only training. We find that trainable bridges can act as parameter-efficient adaptation mechanisms without exploiting useful sender information. We therefore propose PACT-Check to quantify how much of a trained PACT system's gain genuinely comes from communication. Finally, we discuss the opportunities and challenges of PACT systems. Code is available at https://github.com/zhenyi4/platour.
benchmark - arxiv:2611.06962 · cs.AIFormalizing inverse problems in the age of AI: possibilities, pitfalls, and an invitationChristian Clason, Tapio Helin, Medet Nursultanov, Lauri Oksanen
Proof assistants such as Lean check mathematical proofs down to the axioms, and AI agents have recently made writing such proofs cheap. We describe AutoIP, a Lean library in which classical linear regularization theory, together with the functional analysis it rests on, has been autoformalized by AI agents. A machine-checked proof certifies only the formal statement it proves, so the remaining human task is to confirm that each formal statement says what was intended. We organize the library around this task: a reviewer compares the formal statements with their natural-language counterparts and never needs to read a proof. Our experience is that the risks lie in the statements rather than in the proofs, and that simple automated tests on the statements can expose problems even in the established literature. We illustrate this with a textbook theorem whose hypothesis is more restrictive than asserted. We invite the inverse problems community to extend and build on the library.
ai agent - arxiv:2611.06956 · cs.LGOctopus: One Brain, Many Branches for Latent-KV Multi-Agent Reinforcement LearningTianming Sha
Multi-agent LLM systems explore in parallel by spawning sub-agents with isolated context windows, whereas single-agent systems append every search and verification step into one growing trajectory. However, existing multi-agent frameworks duplicate prefix KV caches across branches, exchange findings by re-prefilling text summaries, and flatten or drop cross-branch dependencies during reinforcement learning (RL). We introduce Octopus, a single-model multi-agent RL framework in which all threads share one backbone and coordinate over a GPU-resident directed acyclic graph (DAG) of KV states through five primitives (fork, read_thread, wait_threads, rewind, commit). Octopus combines three co-designed mechanisms: (1) Zero-Copy Radix Forking, which spawns branches as $\mathcal{O}(1)$ cursor views over shared physical KV pages; (2) Attention-Only KV Mounting, which mounts a sibling thread's terminal KV blocks directly into full-attention layers while bypassing linear-recurrent states (Gated DeltaNet / Mamba-2), cutting off-task hijacking by $21\times$ ($57.9\% \to 2.7\%$) while retaining $92.2\%$ of full-text mutual information at zero extra KV allocation; and (3) a Packed DAG-GRPO Trainer, which packs all episode branches into one block-sparse attention pass and assigns counterfactual branch credit via leave-one-branch-out marginal rewards. Trained end-to-end with GRPO on Qwen3.5-4B over a sparse BM25 corpus, Octopus reaches $49.6\%$ accuracy on BrowseComp-Plus ($+17.5$ points over single-agent GRPO at 143K context and $+12.5$ points over text-summary multi-agent RL) and $51.2\%$ on OpenResearcher, outperforming dense-retrieval agents up to $250\times$ larger while increasing training throughput $3.3\times$ and cutting peak KV memory $2.8\times$.
memorymulti-agentagent frameworkagent system
02 US SEMI · SEC 8-K FILINGS
11 itemsscanned: NVDA / AVGO / MRVL / COHR / LITE / AMD / TSM / SMCI / ANET / CRDO / POWL / VECO
- $COHR · 8-K · filed 2026-11-04Coherent CorpItems: 2.02,7.01,9.018-K
- $VECO · 8-K · filed 2026-11-04Veeco Instruments IncItems: 2.02,9.018-K
- $POWL · 8-K · filed 2026-11-04Powell Industries IncItems: 8.01,9.018-K
- $ANET · 8-K · filed 2026-11-03Arista Networks IncItems: 2.02,9.018-K
- $SMCI · 8-K · filed 2026-11-03Super Micro Computer IncItems: 2.02,8.01,9.018-K
- $AMD · 8-K · filed 2026-11-03Advanced Micro Devices IncItems: 2.02,7.01,9.018-K
- $LITE · 8-K · filed 2026-11-02Lumentum Holdings IncItems: 2.02,9.018-K
- $LITE · 8-K · filed 2026-11-02Lumentum Holdings IncItems: 1.01,5.02,7.01,9.018-K
- $MRVL · 8-K · filed 2026-11-02Marvell Technology IncItems: 1.01,9.018-K
- $MRVL · 8-K · filed 2026-11-02Marvell Technology IncItems: 2.02,7.01,9.018-K
- $VECO · 8-K · filed 2026-11-02Veeco Instruments IncItems: 7.01,8.01,9.01FORM 8-K
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 RoboticsJonathan Hurst answers your questions about building Digit 5InsightOctober 23, 2026
- Agility RoboticsThe Realistic Pathway to HomeInsightMay 26, 2026
- Agility RoboticsAgility and AIInsightMarch 16, 2026
- Agility RoboticsAgility Gets a New BrandInsightMarch 5, 2026
Physical Intelligence (8)
- Physical IntelligenceFrontier 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 AgiBotSix AGIBOT G2 Robots Complete 1,730 Tasks in 24-Hour Shanghai Convenience Store Trial2026-11-06
- 智元 AgiBotAGIBOT Unveils GO-3 Preview, Takes Top Spot on Three Embodied AI Benchmarks2026-10-29
- 智元 AgiBotHow Much Work Do Robots Really Deliver? The Thinking Behind Worken2026-10-29
- 智元 AgiBotAGIBOT Introduces Worken to Measure Productive Robot Work at AGIBOT Innovation Day 20262026-10-28
- 智元 AgiBotAGIBOT Launches Genie Studio 2.0 to Bring Robots into Real Workplaces2026-10-26