跳到正文
研踪PaperTrail

论文库

检索计算机领域论文,按研究方向、作者或关键词筛选。

部分采集异常
正在读取账号与研究记录…

全站更新概况

+197新增论文
+151其中 AI 相关
+16已有论文修订
+91新增 AI 导读

arXiv 最近完成扫描:2026-11-22 06:41 · 全站最近入库:2026-11-21 12:19各信源进度 →

统计口径与数据新鲜度

统计 2026-11-21 06:58 — 2026-11-22 06:58(北京时间) · 点击数字查看对应简报;AI 相关包含在新增论文内。

最近完成检查:2026-11-22 06:41最近发现新增版本:2026-11-21 12:19库内论文最新更新时间:2026-11-20 02:59

每分钟刷新已保存的数据。空检查不代表采集异常;新收录可能包含历史论文。AI 导读按新增版本计数。

29,158 已收录论文24,181 AI 相关8,992 有修订112 HF 社区发现

浏览论文库

时间范围跟随排序:版本更新、本站收录或首次发表。最新收录可能包含历史论文。部分信源异常,查看状态 →

找到 21,490 篇论文 · 按版本更新时间排序上轮扫描起点 2026-11-22 06:39(北京时间)
版本更新 · v3arXiv:2510.06199

Learning Latent Volumetric Dynamics for Multimodal Goal-Conditioned Robot Manipulation of Hair

Chengyang Zhao, Uksang Yoo, Arkadeep Narayan Chaudhury, Giljoo Nam, Nathan Dennler, Andreea Bobu, Jonathan Francis, Jeffrey Ichnowski 等 9 位作者

原文摘要

Hair care is an essential daily activity for many, yet it remains inaccessible to individuals with limited mobility and challenging for robot systems due to the fine-grained physical structure and complex dynamics of hair. In this work, we present DYMO-Hair, a model-based robot hair care system for generalizable multimodal goal-conditioned hair styling. We introduce a volumetric dynamics learning method that models combing-induced hair deformations through action-conditioned editing in a pre-trained hierarchical 3D latent space. This latent space captures diverse hair states through volumetric occupancy and local strand orientations, while large-scale pre-training, supported by a physics-based hair-combing simulator, enables generalization to previously unseen hairstyles. Coupled with a Model Predictive Path Integral (MPPI) planner, the learned dynamics model enables closed-loop hair styling toward both visual references and natural language instructions as user input. In closed-loop experiments, DYMO-Hair achieves lower geometric error than baseline systems for visual goals and receives significantly higher ratings than VLM-based planners for language goals in a 58-participant user study. Real-world experiments further demonstrate zero-shot transfer to seven physical wigs varying in color, length, and texture, effectively styling hair toward both visual and language goals without real-world fine-tuning. Together, these results represent a step toward more generalizable, flexible, and accessible robot hair styling in the real world. More details are available on: https://dymohair.github.io/.

cs.RO

新论文 · v1arXiv:2611.20567

Calibrated Functional Prediction from Overlapping Nonlinear Observations

Peida Li

原文摘要

We study prediction of nonlinear audit functionals from scalar observations of a shared hidden table state. For Schur-rational operators, local Gaussian projections yield a common sparse kernel for overlapping tasks, repeated row identities, and cross-row dependence. Shared-state residual samples extend this kernel through a positive-semidefinite feature representation. Positive matrix integration and sample-value enclosures connect the original inputs to the actual saved hybrid prediction. A sorted calibration-role perturbation bound then certifies its complete finite-pool pinball risk. Independent complete-loss selection chooses the ridge, score scale, and residual weight, while task-rank calibration supplies an upper endpoint. Evaluation uses twenty new reference contexts per dataset at two task sizes, giving 400 conditions across ten benchmark populations. The mean dataset complete-loss ratios are 0.9977 to the degree-two kernel, 0.9961 to matched complete-MC shrinkage, and 0.9911 to the retrospective minimum over 174 public configurations. The last ratio has a paired reference-bootstrap interval of [0.9898,0.9927]. On 40 profiles at coefficient tolerance $10^{-7}$, Gamma quadrature has a geometric mean cold-time ratio of .0646 to strengthened Arb/Petras integration.

cs.LG · stat.ML

新论文 · v1arXiv:2611.20564

Strict Capabilities with a PNVI Escape Hatch

Yihe Li, Sai Dhawal Phaye, Roland H. C. Yap, Trevor E. Carlson, Gregory J. Duck

原文摘要

Capability systems enforce memory safety by tying pointers to their originating allocations, preventing forgery, use-after-free, and out-of-bounds accesses. Yet conventional capabilities are simultaneously too strict and not strict enough for C: they reject common low-level idioms involving integer casts or raw pointer bytes, while failing to enforce type correctness and sub-object bounds. Proposed Provenance Not Via Integers (PNVI) semantics expose this tension by permitting representation-level manipulation while retaining strict rules for ordinary pointer operations. We argue that these requirements call for two interoperable abstractions rather than a compromise within one. Strict capabilities enforce provenance, liveness, type correctness, and (sub-)object bounds; ordinary capability operations preserve these invariants. Exposed representations are non-dereferenceable, integer-like values that support arbitrary integer and byte-level manipulation under PNVI semantics while retaining unforgeability guarantees. They carry no directly usable memory-access authority: converting an exposed representation back to a strict capability reconstructs provenance under PNVI rules and validates type and bounds before access. Thus, low-level C idioms can use an explicit escape hatch without weakening ordinary capability derivation. We realize this design using Fully Randomized Pointers (FRP) and dynamic type metadata. Evaluation on low-level pointer idioms, Juliet tests, and existing C benchmarks shows that strict type and sub-object safety can coexist with low-level pointer manipulation.

cs.PL

版本更新 · v2arXiv:2611.16861

Visual Cipher Bench: Benchmarking LLM Agents on Long-Horizon Visual Reasoning

Deyao Hong, Jiayi Jiang, Yuchun Chen, Yixuan Shi, Wenyi Li, Zheng Jiang, Chengxing Xie, Mingyu Su 等 15 位作者

原文摘要

While language-model agents excel at long-horizon coding and terminal tasks, this success relies on two conditions: explicit text specifications and immediate execution feedback. However, neither condition holds in real-world problems where critical information exists only in images, leaving it unclear whether long-horizon capability transfers from text to visual reasoning. In such visual settings, because rules and clues resist text transcription and lack test-case feedback, agents must repeatedly inspect images, simulate spatial transformations, and autonomously cross-check clues. Despite this challenge, existing evaluations fail to capture such capability: visual reasoning benchmarks largely reduce tasks to single-turn, short-chain question answering, whereas agent benchmarks focus on text-based environments or interactive states that permit trial-and-error. To bridge this gap, we introduce Visual Cipher Bench, a collection of 40 original visual puzzles across four categories: spatial manipulation and simulation, logical constraints and planning, rule induction, and knowledge-based decoding. Each puzzle provides only one to three black-and-white line drawings and a minimal submission instruction, requiring agents to discover the underlying solving mechanism entirely from the images. Across 29 reasoning-effort configurations of eight frontier models, even the strongest model, GPT-6 Astra (max), achieves only 26.7% accuracy, while five of the eight models score below 1%. Yet on a ten-task subset, providing a single non-spoiler mechanism hint boosts accuracy from 12.2% to 51.1%. This sharp contrast shows that the primary bottleneck of current agents lies not in executing multi-step plans, but in autonomously discovering rules from visual evidence. Visual Cipher Bench thus provides a rigorous benchmark for measuring long-horizon visual reasoning in agents.

cs.CV

新论文 · v1arXiv:2611.20558

Vision-Language Grounded Semantic Navigation for Autonomous Mobile Robots

Ravindran Vasoor Saravanan, Chittin Varunav Ravi, Madhumita Singh, Thirulok Sundar Mohan Rasu

原文摘要

Language-guided navigation enables mobile robots to interpret natural language commands and navigate toward user-specified targets in unstructured environments. However, integrating open-vocabulary visual perception with classical robotics pipelines for mapping and navigation remains a challenging problem. In this work, we present a complete ROS2-based system that combines frontier-based exploration, open-vocabulary object detection using Grounding DINO, depth-based 3D localization, and semantic mapping to enable end-to-end vision-language grounded navigation in simulation. The system accepts natural language commands such as ``go to the yellow box,'' detects candidate objects from RGB images using text prompts, estimates their positions in the world frame, and generates navigation goals executed by the Nav2 stack. In a Gazebo-based indoor environment with a TurtleBot3 robot, the system autonomously explored the room, detected three of the five target objects, and successfully executed both natural language navigation commands tested, without task-specific training. The results indicate that open-vocabulary grounding models can be integrated into classical navigation pipelines, though performance is limited by detection uncertainty, depth-based localization error, and incomplete exploration coverage. These findings highlight both the feasibility and current limitations of deploying vision-language grounded navigation systems using lightweight, CPU-only architectures.

cs.RO · cs.CV

新论文 · v1arXiv:2611.20557

AsymEEF: Bimanual Manipulation with Asymmetric End Effectors

Xiaokang Sun, Ningyue Wang, Kunqi Xu, Guangqi Jiang, Jianglong Ye, Yong Wu, Xueyan Zou

原文摘要

Teleoperated demonstrations are costly to collect and rarely reusable across end-effectors with different kinematics and action spaces. This burden grows in bimanual manipulation, where each end-effector pair may require new data. We present AsymEEF, a framework for executable cross-end-effector demonstration generation and unified bimanual policy learning. Given a small source dataset, AsymEEF transfers demonstrations to heterogeneous combinations of two-, three-, and five-fingered end-effectors without additional teleoperation. Kinematic teacher anchors provide geometric supervision, while sparse real-world X-Calib anchors correct discrepancies in grasp execution. A shared encoder maps wrist poses and actuator states to a morphology-agnostic latent space, and a morphology-conditioned decoder reconstructs target commands. Generated trajectories undergo replay filtering for task-level executability. The representation also supports a unified bimanual visuomotor policy trained across multiple pairs. Across four simulated and six real-world tasks, AsymEEF improves replay success over adapted transfer baselines, achieves a mean replay-valid expansion gain of 16.6 times in simulation, and is evaluated through downstream policy training on real robots. Project website: https://asymeef.vercel.app/.

cs.RO · cs.AI · cs.CV · cs.LG

版本更新 · v2arXiv:2609.32083

When Should a Human Take Back Control? Optimal Delegation under Turbulent AI Risk

Haoze Yan, Julien Roze, Ved Upadhyay, Unal Tatar, Thibaut Mastrolia

原文摘要

When should a human take control from an autonomous AI agent, how much checking is worthwhile, and when should autonomy resume? These decisions become harder when incidents cluster and recent errors signal elevated risk. Exploratory analyses of AgentLogs coding-agent traces and ODCV-Bench task runs show higher held-out log likelihood for Hawkes incident-history models than matched no-history baselines, with positive gains whose conditional $90\%$ intervals exclude zero. AgentLogs evaluates continuous-time incident timing and type; ODCV evaluates failure occurrence at discrete decisions. Motivated by this predictive evidence, we formulate a continuous-time delegation problem that jointly chooses productive action, verification effort, and switching rates between human-directed, AI-assisted operation and autonomous AI operation. The model assumes that verification weakens incident propagation and balances project benefits against operating, supervision, risk, and handoff costs. Our Hawkes-PPO method summarizes observed incidents with exponential memory filters and learns these decisions from rewards. In synthetic experiments, it achieves a higher evaluated objective than either fixed regime and PPO without history filters, and approaches an approximate full-information benchmark. Further experiments show how incident excitation and supervision costs shape delegation.

cs.LG · cs.AI · math.OC

新论文 · v1arXiv:2611.20553

Learning a Task-Adaptive Low-Dimensional Semantic Space for Improved Visual Classification

Nikalal Kaldera, Wanqing Li, Chris Bunn, Philip O. Ogunbona

原文摘要

Semantic visual classification commonly aligns visual embeddings with semantic spaces induced by pretrained text encoders. However, such spaces are typically high-dimensional, weakly aligned with visual feature geometry, and insufficiently discriminative for downstream classification tasks. We propose TaSS, a framework for learning a task-adaptive low-dimensional semantic space that preserves semantic structure while explicitly optimizing class separability. TaSS is constructed through a parameterized geometry transformation of pretrained text embeddings that enlarges inter-class margins via distance reshaping and improves intra-class compactness through prototype-preserving semantic constraints. The resulting semantic space maintains meaningful semantic relationships while producing a discriminative geometry tailored for classification. Frozen visual encoder features are subsequently aligned to TaSS using a lightweight MLP, enabling improved discriminative learning without fine-tuning pretrained backbones. We further provide theoretical analysis showing that TaSS preserves semantic consistency while enforcing enhanced inter-class separation and controlled intra-class variance. Extensive experiments on skeleton-based human action recognition, video classification, and image classification benchmarks demonstrate consistent improvements over existing semantic alignment approaches, with gains exceeding 3 percentage points on average and up to 8.5 points on challenging datasets.

cs.CV

新论文 · v1arXiv:2611.20551

GranularGPT: Natural-Language-Driven End-to-End Discrete Element Method Simulations

Erlei Li, Sirui Tang, Ai Wang, Mingwei Hu, Yuchen Dai, Geoff Wang

原文摘要

Discrete element method (DEM) simulations are widely used to investigate granular and particulate systems, but their application often requires specialized expertise in particle physics, numerical modelling, and simulation post-processing. Large language models (LLMs) offer new opportunities to simplify these workflows by enabling users to interact with simulation tools through natural language. This study offers GranularGPT, an LLM-assisted agentic framework for end-to-end DEM simulation workflows, from natural-language problem descriptions to simulation execution and automated post-processing. The framework enables users to generate DEM simulation cases, execute and monitor simulations, and automatically extract and report physical and process-level metrics. Evaluation across 60 normal-condition cases covering 12 DEM scenarios and parameter variations achieve a simulation success rate of 91.7%. In the error-handling evaluation, all 10 repairable errors are detected and 9 are successfully resolved, while all 5 non-repairable cases are correctly identified and reported. The framework automatically extracts a range of DEM-relevant metrics, including particle size distributions, packing and density measures, velocity profiles, discharge rates, mixing indices, contact force characteristics, and thermal quantities. GranularGPT provides a foundation for improving the accessibility and efficiency of DEM simulation workflows while reducing the specialized expertise required for their operation.

physics.comp-ph · cs.AI

版本更新 · v2arXiv:2609.30856

Learning Chance-Constrained MDPs with Bellman Distributional Certificates

Chenbei Lu, Hongyu Yi

原文摘要

Safe reinforcement learning (RL) commonly enforces expected-cost constraints, but such expectation-form safety may fail to control the probability of rare high-cost trajectories. Chance-constrained MDPs (CCMDPs) impose a stronger probability-level requirement, but are widely viewed as harder because the chance constraint itself is nonconvex and depends on the full trajectory rather than a Bellman-linear expectation. In this paper, we reveal that this computational difficulty does not necessarily imply a higher statistical price. For tabular discounted CCMDPs with bounded successor support and access to a certified planning oracle, we establish a model-based upper bound, with a matching lower bound up to logarithmic terms. Technically, our key idea is the \emph{Bellman distributional certificate}, which constructs a Bellman recursion for constraint violation probabilities before policy selection. The certificate reduces trajectory-level safety analysis to local transition uncertainty, avoiding a union bound over the potentially exponential number of candidate policies. For stochastic policies, we give a model-free variance-reduced policy-gradient algorithm with a finite-sample expected KKT-residual guarantee and independent validation of every accepted policy. Numerical experiments on synthetic CCMDPs and an IEEE 14-bus energy storage control benchmark illustrate the safety and mechanism behavior of the proposed algorithms.

cs.LG

版本更新 · v2arXiv:2610.16346

CellChain: Benchmarking Cross-Scale Evidence Use for Perturbation Reasoning

Xue Xia, Yue Wang, Shuizhou Chen, Lang Yu, Hao Wu, Wenxuan Huang, Chengkai Yao, Zhangyang Gao 等 9 位作者

原文摘要

Recent advances in foundation models and large-scale cellular perturbation datasets have driven growing interest in using LLMs to predict and reason about cellular responses. Biological reasoning in this setting involves relating molecular mechanisms, cellular responses, and functional outcomes across scales. Evaluating foundation models therefore requires assessing not only whether their predictions are correct, but also how those predictions depend on the biological evidence provided. We introduce CellChain, an experimentally grounded benchmark comprising 590 biological cases and 2,360 matched evidence-condition instances across gene-dependency prediction, cytokine-axis identification, and drug-sensitivity ranking. CellChain links perturbation experiments and external biological resources through a provenance-aware evidence graph spanning molecular, pathway, transcriptional, and phenotypic information. Task-specific conditions vary evidence coverage, composition, or assignment while preserving questions, candidate answers, reference targets, and scoring rules. Across 18 foundation models, the highest accuracies under prespecified target-evidence conditions are 57.5%, 75.5%, and 70.8%, respectively, and broader evidence coverage does not consistently improve accuracy. On the drug-sensitivity benchmark cohort, matched reference histories outperform swapped histories by an average of 13.0 percentage points in cluster-weighted accuracy. Decision-transition analysis separates corrections, newly introduced errors, and wrong-answer switches, revealing substantial answer changes hidden by small net accuracy differences. CellChain provides controlled comparisons for attributing changes in model predictions to prespecified expansions of biological context, including changes in evidence coverage, composition, and biological assignment.

cs.DB · q-bio.GN

版本更新 · v4arXiv:2606.08281

Toward Interaction Dynamics: A Predictive Framework for Safe Physical Human Robot Interaction

Yongyan Cao, Jinshan Tang

原文摘要

Physical human-robot interaction requires yielding transiently to contact yet recovering the commanded reference under sustained load. Finite-stiffness impedance control retains a static deflection there, while predictive alternatives typically optimize a nonlinear model online. Operational-space cancellation instead exposes a translational error double integrator with a fixed transition matrix and a configuration-scheduled input map, making interaction a predictive quantity. We build on it a compact offset-free interaction-error MPC for torque-controlled manipulators: a force-domain random-walk state estimates persistent interaction and model error, and a 30-variable convex QP maps the correction through the current task inertia while constraining the applied joint torque. Conditional results establish impedance equivalence of the unconstrained passive feedback, offset-free regulation at feasible frozen configurations, and quadratic stabilizability of the scheduled backbone. In a 1 kHz MuJoCo simulation of a 7-DOF Franka FR3, the estimator cuts steady-state error under a repeated 15 N step from 2.73 mm to 0.033 mm when added to the otherwise identical 100 Hz MPC, at unchanged peak joint power. A stiffness-and-damping-calibrated impedance baseline attains 2.63 mm at the same actuator cost, and 1.36 mm with ideal measured-force cancellation, so constant-load rejection is not unique to MPC; the sensorless controller still reaches 0.033 mm. The torque constraint holds the command inside a budget tightened to the lowest the task can meet, at no tracking cost. On a physical UR5e, five alternated pairs under a nominal 1.5 kg hanging load give 1.505 mm without the estimator; with it the controller is offset-free, its 0.023 mm error being the noise level of the encoder-based position measurement. A tuned anti-windup integral also removes the offset, with larger transients.

cs.RO · cs.HC · eess.SY · physics.med-ph

版本更新 · v4arXiv:2606.10380

CRADLE-Dialogue: Expert-Level Crisis Detection in Mental Health Conversations

Grace Byun, Abigail Lott, Rebecca Lipschutz, Sean T. Minton, Elizabeth A. Stinson, Jinho D. Choi

原文摘要

Real-world crisis intervention is inherently conversational, yet existing research largely focuses on static texts. When applied to multi-turn dialogues, current models exhibit significant performance degradation, struggling to track risk signals that emerge as context evolves. To address this gap, we introduce CRADLE-Dialogue, a clinician-annotated benchmark for turn-level crisis detection in conversational settings. The dataset features 600 dialogues with multi-label annotations across clinically grounded risks, including suicide ideation, self-harm, and child abuse, distinguishing past from ongoing risk. We further propose an Alert-Confirm evaluation protocol that distinguishes early warning signals (Alert) from turns where a specific crisis becomes explicitly identifiable (Confirm), reflecting the clinical need to intervene before risk becomes explicit. Experiments show that identifying when risk emerges is much harder than recognizing that it exists: models achieve only mid-40% to high-60% Micro F1. Additionally, we release a synthetic training corpus and a 32B-parameter model that substantially outperforms existing open-source models and achieves competitive or superior results against proprietary models across turn-level, dialogue-level, and confirm-only evaluation settings. We further validate our model on real patient interviews, demonstrating its ability to generalize beyond synthetic conversations.

cs.CL · cs.AI

新论文 · v1arXiv:2611.20546

A Gaussian Primitive Data Management Framework for Efficient 3DGS Trajectory Rendering

Haolun Ma, Ziyi Liu, Zizhuo Xu, Chuhan Hu, Lei Li

原文摘要

Trajectory rendering, which renders novel views of a reconstructed 3D scene along a camera path, is a core operation in digital twins and autonomous inspection. Structured 3D Gaussian Splatting (3DGS) renders a single view in real time by decoding sparse anchors into Gaussians on demand, but it renders a trajectory view by view: every view projects all anchors to select them, re-decodes anchors already requested by neighboring views, and runs all steps on the GPU while the CPU idles. We treat this as a spatial data management problem, where anchors are spatial objects, views are range queries, and decoded Gaussians are shared derived data, and propose GDM-GS, which removes redundancy at three levels. Within a view, an anchor index and an occluder index answer each view as a holed-frustum range query, so out-of-view and provably hidden anchors are never decoded. Across views, shared materialization decodes each requested anchor once per group of consecutive views while preserving each view's selection. Across processors, a hybrid pipeline balances selections between the CPU and the GPU before grouped decoding under bounded memory. Experiments on large-scale public scenes show that GDM-GS substantially reduces trajectory rendering latency while preserving image quality.

cs.DB · cs.CV · cs.GR

新论文 · v1arXiv:2611.20540

Changing a View Is Not Amending a Record: Timing and Visibility in a Model of Public Correction

Ziming Liu

原文摘要

An accusation may remain on the public record after its author has changed their mind. Observers can respond either to what a person says now or to an earlier statement that has never been explicitly amended. We examine the consequences of these two signals with a generative agent-based model that separates private evidence updates, current expression, and revisions linked to a specific declaration. In paired worlds, evidence, prehistory, and opportunities to revisit that declaration remain fixed. When expression and amendment are synchronized and fully observed, the feedback channels produce identical trajectories. With sparse review, record-based feedback leaves 19.44 percentage points more eligible accusations unamended after 24 rounds (paired-world 95 percent bootstrap interval 13.89 to 25.31). The comparison averages 50 eligible worlds, containing 556 agents, from 128 newly generated worlds. Event traces show how a persistent public cue changes subsequent amendment propensity; an exact small-state calculation independently establishes the timing boundary. Reducing record salience can erase or reverse the increment, and a generic held-state process reproduces the dynamics. The findings show when current expression ceases to track correction of a particular public statement.

cs.SI

新论文 · v1arXiv:2611.20538

Effective and Efficient Rumor Assertion Over Large-Scale Social Networks (Technical Report)

Qi Wen, Vipin Chaudhary, Xiang Lian, Yutong Ye, Mingsong Chen

原文摘要

Nowadays, social networks, like Facebook, X, Reddit, YouTube, and so on, have become essential online platforms for users to connect with friends daily, post/share news, and express their opinions/comments. Meanwhile, however, the rapid and widespread dissemination of rumors and misinformation across social networks has become a serious concern. Such misleading information can distort public opinion and potentially cause significant social and economic consequences in real-world scenarios, including product promotion, elections and voting, and even cyberbullying. Therefore, accurately and efficiently verifying whether news spread on social networks is rumor or truth has become an increasingly important yet challenging problem. While previous works usually identified rumors by their contents, in this paper, we propose a novel \textit{user behavior model} to capture the users' judgment/discriminative abilities, which is learned via our well-designed evidence-aware probability reconstructor. We develop a general framework for tackling \textit{rumor assertion over large-scale social networks} (RA-SN), which can efficiently and effectively determine whether a given news item is a rumor with high confidence, based on users' behaviors, in a scalable manner. Extensive experiments demonstrate the efficiency and effectiveness of our proposed RA-SN approach on both real-world and synthetic graphs.

cs.SI · cs.CY

新论文 · v1arXiv:2611.20537

SmoothShift: Disrupting Visual Adversarial Attacks on LVLMs at Test Time

Xiaoqing Chen, Jiayi Ouyang, Dongrui Wu

原文摘要

Visual adversarial perturbations can substantially impair large vision-language models (LVLMs) despite introducing only subtle input changes. To understand how these small changes cause substantial errors, we investigate their optimization and propagation through the model, revealing two key findings. First, stronger attacks rely on finer iterative optimization, while subtle input perturbations induce representation discrepancies that are amplified layer by layer through the vision encoder and persist in the language model. Second, disrupting the alignment between within-patch input patterns and the model's computation reduces clean-adversarial representation discrepancies and substantially weakens attacks. Together, these findings suggest that adversarial effectiveness depends on both fine-grained perturbation patterns and their coupling with the model's computation. Motivated by these findings, we propose SmoothShift, a lightweight, training-free defense that combines local smoothing and spatial transformations. Smoothing weakens fine-grained perturbation patterns, while spatial transformations disrupt their alignment with the encoder's patch grid. SmoothShift requires no model updates, auxiliary networks, or additional model inference. Experiments cover three LVLM families on four datasets and robustly fine-tuned Contrastive Language-Image Pre-training (CLIP) models on eight classification datasets. Evaluations across eight attack methods and multiple perturbation budgets demonstrate improved robustness while largely preserving clean performance. SmoothShift also provides robustness gains under defense-aware attacks using mixed-gradient BPDA with EOT.

cs.CV

新论文 · v1arXiv:2611.20535

VAHBench: A Harness Benchmark for Agentic Long-Video Generation

Hongbo Zhang, Liuyang Song, Hang Zhou, Xinchen Liu, Wei He, Kai Han, Yunhe Wang

原文摘要

AI agents are commonly used for everyday information retrieval and coding. We ask whether they can also interpret human requests and turn them into complete videos. We introduce VAHBench, a benchmark that tests this ability through five short-drama briefs specifying story, characters, dialogue and duration. Agents plan and carry out production using shared generation tools. Evaluation combines video completeness and basic quality with checks of production reliability and reporting accuracy. Across 240 runs covering 48 configurations of general-purpose and video-specific agents, higher quality scores do not necessarily mean more complete delivery. Missing episodes, repeated footage and static filler reveal gaps between producing a video file and fulfilling the brief. Added composition tools help some agents but do not consistently improve performance. VAHBench measures how well agents translate user intent into finished video content and helps identify where that process breaks down.

cs.AI

新论文 · v1arXiv:2611.20533

RemeshingAgent: An Iterative Agent for Repair-Based Decimation of Generated Building Meshes

Haipeng Wang

原文摘要

Aggressively simplifying a generated high-poly building mesh to 0.5-1% of its faces confronts a hard limit. The surface looks complete, yet the mesh is topologically shattered: 10^3-10^4 connected components and non-manifold thin walls. Below a certain budget, every practical decimator "melts": it deletes parts, welds thin walls, or breaks the error tolerance. We attribute this to instance-specific pathology: which components are significant and where the cliff sits vary from mesh to mesh, so no single fixed algorithm suffices, and decimation becomes an iterative loop of cleaning, simplifying, and verifying. We present RemeshingAgent, an agent that turns this loop into a validated process. The melting follows from a triangle-budget decomposition into curvature, crease, component, and genus terms: below the component floor, no output satisfies the repair contract, and cleaning first is the theory-mandated response. At a 0.5% budget the agent stays topology-clean where QEM and vertex clustering melt, matches a commercial remesher on 20 unseen city buildings, and on the 24 training buildings a nine-participant single-workstation blind study prefers our outputs 168:48 (p = 0.0195, subject-level sign test). The trained skill deploys as retrieval replay: 11.7 seconds per mesh on CPU, with no LLM calls.

cs.GR

新论文 · v1arXiv:2611.20528

ReconPotBench: Evaluating LLM Honeypots in Multi-Round Linux Reconnaissance

Qianli Wang, Junchao Wu, Binkai Ou, Lidia S. Chao, Derek F. Wong

原文摘要

Large language models (LLMs) are increasingly used to build interactive honeypots, yet existing evaluations mainly assess isolated command outputs and cannot determine whether a system remains believable when later commands depend on earlier results. We present ReconPotBench, a multi-round benchmark for Linux reconnaissance honeypots. It contains 1{,}032 three-round reconnaissance interactions across eight reconnaissance categories, with reference outputs collected from a real Linux environment. Under a unified testing protocol, we compare Cowrie, open-source LLMs, and closed-source LLMs along three dimensions: round-level authenticity, cross-round contextual consistency, and hacker-oriented reconnaissance value. Results show that multi-round evaluation exposes failures hidden by single-round testing, as authenticity declines with interaction depth and both open-source and closed-source LLMs struggle to maintain coherent state across rounds. LLM-based honeypots nevertheless outperform Cowrie in authenticity and reconnaissance value, while Cowrie remains more consistent because of its rigid templated behavior. Category-level analysis further shows that tasks requiring systems to preserve network relationships across rounds are harder than tasks centered on explicit entities such as ports. ReconPotBench enables reproducible evaluation of LLM honeypots across authenticity, entity consistency, and hacker-perspective utility in dependent multi-round interactions. The dataset is available at: https://github.com/qianli-wang/ReconPotBench

cs.CR · cs.CL

新论文 · v1arXiv:2611.20524

GaugeRoPE: RoPE-Compatible Executable Latent Attention

Yuanhang Gao, Xiangrui Yang, Lingyuan Meng, Yuanfeng Chen, Dongsheng Li

原文摘要

Latent KV cache compression reduces memory footprint by mapping historical keys and values into compact representations. However, generic compressed key states require reconstruction to native dimension before applying RoPE and computing attention scores, adding sequence-scaling overhead to the decode path. We characterize when a latent KV representation can serve directly as the execution state of pretrained RoPE. For RoPE-based transformers, exact static linear execution of the rotary key path requires the latent-to-native map to preserve the pretrained RoPE action through an intertwining relation, restricting exact mixing to coordinates that share the same RoPE frequency. Guided by this characterization, we build GaugeRoPE, a training-free executable latent KV representation. It combines a RoPE-compatible carrier that preserves the rotary action with an approximate position-independent residual key--value (RKV) correction for complementary key and value information, both evaluated without per-token dense historical K/V reconstruction. On Mistral-7B, GaugeRoPE achieves $5.85\times$ KV compression while maintaining quality close to the full-cache baseline. Additional MHA results validate the construction beyond GQA. GaugeRoPE achieves up to $2.48\times$ the decode throughput of the vanilla uncompressed baseline through direct latent execution.

cs.LG · cs.CL

版本更新 · v4arXiv:2602.23530

Unified Ranking Across Retrieval Channels: A Large-Scale E-Commerce Approach

Aditya Gaydhani, Guangyue Xu, Dhanush Kamath, Ankit Singh, Alex Fengjie Li

原文摘要

Large-scale e-commerce search and recommendation systems rely on multiple specialized retrieval channels to surface diverse products, each optimized for a distinct objective. A key challenge is merging candidates from these heterogeneous channels into a single ranked list under strict latency constraints while optimizing business KPIs such as user conversion. Static rank-fusion methods such as Reciprocal Rank Fusion (RRF) and Weighted Interleaving use the same fusion behavior across queries, limiting their ability to adapt to context or model interactions among channel signals. We reformulate multi-channel fusion as a channel-aware learning-to-rank problem and propose a unified ranking model trained on a scalar, funnel-weighted graded-relevance label that combines clicks, add-to-carts, and purchases. The model incorporates channel-specific scores together with recent aggregate query-item behavior to capture near-term population demand. The evaluated channel set combines conventional relevance retrieval with query-conditioned, recommendation-oriented channels emphasizing newness, trending demand, regional relevance, and seasonality. A randomized online experiment shows a 2.85% increase in user conversion over the deployed weighted-interleaving control. The model achieves a p95 reranking-stage latency below 50 ms and has been deployed on Target.com.

cs.IR

新论文 · v1arXiv:2611.20522

AFRNet: Adaptive Frequency Redistribution Network for Image Deraining

Dezhi Zhang, Shangrong Yang, Husheng Guo, Zhijie Shen, Wenjian Wang

原文摘要

Image deraining aims to recover rain-free images from rain-degraded observations. Recent frequency-aware deraining methods have demonstrated promising performance by removing rain artifacts in the spectral domain. However, since rain artifacts and background textures often appear in several correlated subbands, prior frequency-domain deraining models still enhance frequency subbands separately. Such band-wise processing limits information exchange across subbands. Without context from other subbands, rain and clean content that appear together in one subband cannot be separated. This leads to either residual rain streaks or over-smoothing of fine details. To address this limitation, we propose AFRNet, an Adaptive Frequency Redistribution Network for image deraining. First, we introduce an Adaptive Frequency Recombination Module (AFRM) that performs cross-band modulation among wavelet subbands, enabling information interaction and adaptive recombination of degraded frequency components. We then develop a Subband Compensation Block (SCB) that compensates for decoder subband degradation by leveraging encoder subband features through inter-subband attention in the wavelet domain, alleviating detail loss during reconstruction. Extensive experiments demonstrate that AFRNet achieves competitive performance and consistent results across diverse deraining benchmarks.

cs.CV

新论文 · v1arXiv:2611.20521

LINK: Lightweight and Inductive Node-to-Node Propagation Rule Learning for Traffic Kriging

You Shang, Yu-Kwan Ching, Weifeng Su, Guanyao Li, Ruiyuan Li, Gary Shueng-Han Chan

原文摘要

Spatio-temporal traffic kriging aims to reconstruct traffic time series at unobserved locations with no historical traffic observations by leveraging spatial and temporal dependencies across the road network. Existing Graph Neural Network (GNN)-based approaches primarily rely on neighborhood aggregation to capture spatial dependencies, thereby modeling traffic interactions implicitly and making it difficult to accurately recover traffic magnitudes at locations without historical observations. To address this limitation, we propose LINK, a Lightweight and Inductive Node-to-node propagation rule learning paradigm for traffic Kriging. LINK introduces three key designs: 1) a propagation-centric traffic kriging paradigm that explicitly learns a shared node-to-node transition rule; 2) a context-aware Spectral Propagation Network that performs traffic-state propagation in both the spectral and time domains; and 3) TraffiKrig, a realistic traffic kriging benchmark that preserves heterogeneous road-segment characteristics and road-specific metadata while enabling systematic comparison of representative kriging methods under a unified metadata-enhanced setting. Extensive experiments demonstrate that LINK achieves strong reconstruction accuracy and computational efficiency. Compared with the strongest baseline on each subset, LINK achieves an average MAE reduction of 54.7% over original baselines. LINK also achieves substantial efficiency improvements, reducing training time by 66.5% and inference time by 86.6% compared with the fastest competing methods. Zero-shot experiments further demonstrate the transferability of the learned propagation rule across regions.

cs.LG · cs.AI

版本更新 · v2arXiv:2609.22308

GameReplica: A Benchmark for Black-Box Visual Game Replication by Vision-Language Agents

Boyu Qiao, Zixin Tang, Kaige Li, Xiaoshuai Hao, Dongyu She, Xiaochun Cao

原文摘要

Benchmarks for coding agents typically assess implementation from explicit specifications, leaving open whether agents can discover and reproduce systems whose rules are undocumented and accessible only through interaction. We study this capability in web games, which offer controlled environments with observable actions and outcomes, and introduce GameReplica, a benchmark for end-to-end black-box game replication. It comprises 125 instances spanning 25 web games and five core mechanism families, with each game instantiated at five difficulty levels. Under container isolation, an agent accesses the target game only through an action--screenshot interface, may iterate between probing the game, implementing a replica, and checking it, and delivers a self-contained runnable replica together with a rule checklist. A reference-grounded evaluation then scores the deliverables separately on rule consistency, implementation consistency, and visual fidelity. Experiments across six models and five agent runtimes show that black-box replication remains challenging: Claude Opus 4.8 achieves the highest Overall score, 71.6%, while the other models obtain 4.0%--43.0%. Every model scores highest on visual fidelity and lowest on rule consistency, suggesting that reproducing appearance is easier than recovering mechanics. Difficulty widens the gap: from L1 to L5, the four lower-scoring models show relative Overall score declines of 33%--63%, whereas the strongest model declines by less than 5%. In the evaluated configurations, outcomes vary more across models than across agent runtimes, and most attributed failures of weaker models arise during exploration, before replica code is written. Resources and interactive demonstrations are available at https://qqqqqqby.github.io/GameReplica/.

cs.CV · cs.AI

版本更新 · v2arXiv:2603.27928

Generalizable Social Bot Detection via Multi-Granularity Summarization and Domain-Invariant Learning

Boyu Qiao, Zixin Tang, Yunman Chen, Xiaoshuai Hao, Dongyu She, Wei Zhou, Yunfeng Diao, Songlin Hu

原文摘要

Social bots are becoming increasingly sophisticated in mimicking human users, posing persistent threats to online information ecosystems. Despite substantial progress in social bot detection, existing methods often generalize poorly to unseen datasets. Two challenges underlie this limitation: heterogeneous user information across datasets makes consistent representation difficult, while distribution shifts encourage reliance on domain-specific patterns. To address these challenges, we propose Multi-Granularity Summarization and Domain-Invariant Learning (MGDIL), a unified framework for generalizable social bot detection. MGDIL first converts profile attributes into structured textual descriptions and summarizes historical posts into user-level behavioral representations, providing a shared semantic interface across heterogeneous datasets. It then combines task-oriented instruction tuning, domain-adversarial learning, and cross-domain contrastive learning to reduce dependence on domain-specific patterns while preserving discriminative information between bots and human users. We construct a unified instruction-tuning corpus from 15 social bot datasets, train MGDIL on 13 earlier datasets, and evaluate it on two later-collected unseen datasets under a temporally ordered cross-dataset protocol. Comparisons with representative conventional, neural, LLM-based, and transfer-based baselines, together with ablation and representation analyses, demonstrate the effectiveness of MGDIL in improving generalization across datasets. The code is available at: https://github.com/QQQQQQBY/MGDIL.

cs.SI · cs.HC

新论文 · v1arXiv:2611.20514

Beyond the Patch: Reconstructing an Incomplete-Fix Patch Bypass in LLM Agent Frameworks with Control-Gated, Benign Probes

Chaitanya Garware

原文摘要

Replaying a vulnerability's original proof-of-concept after patching tests whether its reported trigger remains reproducible, but does not establish that the underlying security invariant holds on other code paths. We present a benign, control-gated reconstruction method and a pilot study across four LLM agent frameworks and four invariant classes. For PraisonAI, we compare original-trigger replay (B0) with a distinct sibling-path probe (M). We also perform five before/after reconstructions across Langroid, Agno, and CrewAI. Each run reports VIOLATION, NO_VIOLATION_OBSERVED, or INCONCLUSIVE, gated by a positive authorized control. Probes use inert markers or builtins-reachability checks. At PraisonAI 4.6.31, the original tool-loading trigger is blocked while the tool_override path still violates the invariant. This reproduces the publicly documented CVE-2026-44334 bypass of CVE-2026-40287; neither tested path shows a violation at 4.6.32. For CrewAI, we observe a violation in tested release 1.11.0 and confinement in 1.11.1, without claiming the earliest fixed release. Fourteen harness-validation scenarios produced their expected verdicts. Only one same-version B0-versus-M divergence is demonstrated. The PraisonAI probes invoke loader functions directly and establish boundary behavior under the supplied setup, not a full end-to-end exploit. A non-violation verdict applies only to the tested probe, version, and configuration and is not a proof of safety. This pilot reconstructs a known incomplete fix; it neither introduces a new testing technique nor estimates vulnerability prevalence.

cs.CR

新论文 · v1arXiv:2611.20511

Optimizing Query Decomposition for Reliable Validation

Shuo Zhang, Sudeepa Roy, Jun Yang

原文摘要

Large language models (LLMs) have made querying databases more accessible, but cannot guarantee that the SQL they generate is correct. Without a trusted reference query, a validator (human or LLM) must decide whether a generated query answers the intended question. Enterprise questions often require queries with multiple joins, nested subqueries, and multi-level aggregation, and validating such a query in one judgment is error-prone. Decomposition simplifies this task by rewriting a query into smaller queries (blocks) that recompose to the original, each checked against its own sub-question. Decompositions differ, however: smaller blocks make each judgment easier but require more judgments, and the best balance depends on the validator's ability. We formulate validation-aware query decomposition as an optimization problem. Just as a query optimizer chooses among equivalent plans by estimated cost, we choose among decompositions by how reliably a given validator is expected to check them. We model how task difficulty and validator ability jointly affect reliability and define an objective that balances reliability against the number of judgments. Over a space of decompositions derived from logical-plan rewrites, our search finds an optimal one. We implement it in DecoVal, a pipeline that presents each block with its sub-question and input interface, combines local verdicts into a query-level decision, and localizes errors to blocks. A user study shows that decomposition helps human validators catch and locate errors: with DecoVal, participants caught 91% of the errors, up from 36% without decomposition, and located the error in nearly every rejection. Experiments on BEAVER and Spider 2.0 with LLM validators show that DecoVal also raises error-detection recall over whole-query validation while keeping precision high, and that the best granularity depends on the validator.

cs.DB

版本更新 · v2arXiv:2602.07859

Dynamic Load Model for Data Centers with Pattern-Consistent Calibration

Siyu Lu, Chenhan Xiao, Yang Weng

原文摘要

The rapid growth of large-scale data centers has made accurate dynamic data center load modeling increasingly important for power system analysis. Such loads are characterized by fast workload-driven variability and protection-driven disconnection and reconnection behavior that are not captured by conventional load models. Existing approaches to data center load modeling include physics-based methods, which provide interpretable structure for grid simulation, and data-driven approaches, which capture empirical workload variability from data. However, physics-based models are typically uncalibrated to facility-level operation, while trajectory alignment in data-driven methods often leads to overfitting and unrealistic dynamic behavior. To resolve these limitations, we design the framework to leverage both physics-based structure and data-driven adaptability. We further show that trajectory-level alignment is limited for inherently stochastic data center loads. Therefore, we design the calibration to align temporal and statistical patterns using temporal contrastive learning (TCL). This calibration is performed locally at the facility, and only calibrated parameters are shared with utilities, reducing the need to share full-resolution operational time-series traces with utilities. The proposed load model is calibrated by real-world operational load data from the MIT Supercloud, ASU Sol, Blue Waters, and ASHRAE datasets. Then it is integrated into the ANDES platform and evaluated on the IEEE 39-bus, NPCC 140-bus, and WECC 179-bus systems. We simulate the collective dynamic responses of multiple heterogeneous data centers under system-level disturbances. We find that interactions among data centers can fundamentally alter post-disturbance recovery behavior, producing compound disconnection-reconnection dynamics and delayed stabilization that are not captured by uncalibrated load models.

cs.LG · eess.SY

新论文 · v1arXiv:2611.20510

WTR-Drive: World-Task Decomposition with Structured Risk Supervision for End-to-End Driving

Haoyu Xing, Shaobing Ding, Guanglei Zhao, Guanhua Zhao, Nawu Peng, Mingyang Hou

原文摘要

Proposal-conditioned scene predictions can lack a common object-level reference across candidate trajectories. Aggregate planning scores also leave the objects and times associated with violations implicit. We present WTR-Drive (World-Task-Risk Driving), a Drive-JEPA-based planner that separates shared scene estimation from proposal-conditioned reasoning. A proposal-independent World Memory estimates object states, map structure, and future motion by re-reading the original visual features at each refinement round. Proposal feedback updates a residual Task branch without writing back to World Memory. This branch guides refinement through candidate-specific retrieval and gated visual feedback. A final Task readout supplies candidate-conditioned features for scoring. We derive structured risk targets from existing candidate-scoring rollouts without additional simulator calls. For at-fault collision and time-to-collision violations, the targets encode event existence, association with shared object slots, and the first violation time bin. On the NAVSIM navtest split, WTR-Drive achieves 94.66 PDMS on v1 and 88.24 EPDMS on v2 (expert-frame sequential evaluation). These scores are 0.96 and 0.44 points above the published Drive-JEPA values, respectively. Candidate-level analysis shows that higher best-candidate scores can coexist with lower selected-trajectory scores, revealing a gap between candidate quality and selection effectiveness.

cs.RO