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 08:44 · 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 10:26 — 2026-11-22 10:26(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 08:44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 7,943 篇论文 · 按版本更新时间排序上轮扫描起点 2026-11-22 08:42(北京时间)
新论文 · v1arXiv:2611.25889

EvoMem: Self-Evolving Skills for Time-Aware Long Video Memory Agents

Yutong Guan, Meixing Shi, Haoran Sun, Bing Sun, Yuxiang Cai, Yankai Jiang

原文摘要

Although memory-augmented architectures have expanded the ability of large vision-language models (LVLMs) to process long videos, existing retrieval processes still largely rely on global semantic matching and lack state-driven temporal control, often returning fragmented or redundant evidence and failing to maintain a coherent evidence chain. To address this challenge, we propose EvoMem, a time-aware long-video memory agent guided by self-evolving retrieval skills. EvoMem constructs a hierarchical time-indexed memory, where different types of evidence are aligned onto a unified video timeline; during retrieval, temporal cues from acquired evidence are further converted into explicit signals that constrain subsequent memory access. To address recurring procedural errors in long-video question answering, such as unclear retrieval targets, improper memory access paths, or premature stopping, EvoMem introduces retrieval skills to provide process-level guidance, enabling the agent to dynamically adjust retrieval content, memory operations, and target temporal ranges according to the current evidence state, while deciding when to stop. Furthermore, inspired by reinforcement learning, we formulate skill evolution as a reward-guided policy improvement process. Each candidate skill is treated as an external policy patch, and rewards are assigned according to answer-correctness state transitions on the validation set. In this way, effective skills are accepted and low-quality skills are rejected without updating model parameters. Experiments on four long-video question answering benchmarks show that EvoMem significantly outperforms existing baselines, achieving an average accuracy of 60.1% and an average improvement of 8.6 percentage points over the previous best methods, demonstrating its effectiveness for long-video reasoning.

cs.CV

版本更新 · v2arXiv:2609.35628

Arbitrary-Accuracy Neural Approximation with Optimal Neuron Count and Optimal Bit Complexity

Zilan Cheng, Li-Lian Wang, Zhongjian Wang

原文摘要

We study the minimum number of hidden neurons required for arbitrary-accuracy approximation of continuous functions on $[0,1]^d$ and the associated encoding complexity. For $d\geq 2$, we construct two fixed, explicitly defined activation functions and, for each, a closed-form network with two hidden layers of widths $d$ and $1$ that achieves arbitrary accuracy in the uniform norm. We prove that $d+1$ is the exact minimum total number of hidden neurons among standard feedforward networks with locally integrable activations and affine outputs. We further give a simpler construction using a single elementary activation that combines the floor and exponential functions. This construction requires three hidden layers of widths $d$, $1$, and $2$. If a skip connection is allowed, widths $d$, $1$, and $1$ suffice. For a bounded $α$-Hölder class with $α\in(0,1]$, the networks with the digit-extraction activation and with the elementary activation require $O(\varepsilon^{-d/α}\log(1/\varepsilon))$ bits, matching the metric-entropy lower bound up to a logarithmic factor. The other $(d,1)$ network, which stores differences of neighboring values along a spanning tree, needs only $Θ(\varepsilon^{-d/α})$ bits and is optimal in both neuron count and bit complexity.

cs.LG · cs.NE · stat.ML

版本更新 · v2arXiv:2609.37851

FlowMap-OPD: On-Policy Distillation for Few-Step Flow-Map Generators

Zhiqi Li, Cheng Wan, Bo Zhu

原文摘要

Few-step flow-map generators, such as MeanFlow and consistency models, sample in a few long-range jumps. On-policy distillation (OPD) trains a student on states drawn from its own sampler, but existing methods compare teacher and student through the local stochastic transitions of a multi-step sampler, which a flow-map student never deploys. We present FlowMap-OPD, the first systematic study of OPD for flow-map students. It separates state collection from supervision: the student collects states with its native few-step sampler, while the teacher supervises the flow map, velocities induced from it, or the instantaneous velocity linked to the map by a separately weighted consistency penalty. We find that instantaneous-velocity supervision with a small consistency weight, our chosen form, transfers the teacher's gains best, whereas supervising the deployed map, although the natural choice, transfers them worst when teacher and student start from different pretrained models. This gives a counter-intuitive rule: \emph{sample with the flow map, learn the velocity}. As a further benefit, this form applies to any teacher that defines a velocity, including multi-step ones, and to students initialized from multi-step models. On FLUX and SD3.5, a single few-step student consolidates three specialist teachers within 300 updates, matches or exceeds each on its own benchmark (e.g., OCR 0.883 versus 0.850 on FLUX), and outperforms both multi-reward reinforcement learning and LoRA merging.

cs.CV · cs.LG

新论文 · v1arXiv:2611.25873

WHU-Road3D: A dataset and benchmark towards fine-grained road surface mapping

Xiaoxin Mi, Zhen Cao, Chong Liu, Yuhao Li, Zhen Dong, Bisheng Yang, Yang Jia, Shengwu Xiong

原文摘要

Accurately extracting and mapping road surface elements is crucial for applications such as high-definition maps, road asset management, and intelligent transportation systems. This study first provides a comprehensive review of the benchmark datasets for extracting and mapping road elements, encompassing overhead-to-perspective data, large-to-local scales, and coarse-to-fine details. Subsequently, we thoroughly review the relevant approaches for extracting road surface elements. However, we argue that existing datasets still require enhancements in semantic richness, data modality diversity, and scenario complexity. To bridge these gaps, this paper presents WHU-Road3D, a large-scale multimodal dataset for fine-grained road surface mapping to address these shortcomings. WHU-Road3D provides approximately 100 km of diverse Chinese highways and urban streets. It features a rich combination of MLS point clouds, associated panoramic images, vehicle trajectories, detailed vectorized annotations, and pointwise instance annotations for road surface elements. The dataset carefully annotates 15 distinct categories of road elements, including common markings like road boundaries, zebra crossings, lane dividers, stop lines, as well as 10 types of direction symbols. To our knowledge, WHU-Road3D stands as the first multimodal dataset for mapping such a comprehensive range of road surface elements, designed to facilitate cutting-edge research in instance segmentation, full-element extraction and mapping on road surfaces. Beyond these immediate applications, WHU-Road3D also provides a valuable source for hierarchical road topology reasoning, map updating, and street scene generation. The dataset, along with the code for our benchmark methods, will be made publicly available at \href{https://github.com/WHU-USI3DV/WHU-Road3D.git}{https://github.com/WHU-USI3DV/WHU-Road3D.git}.

cs.CV

版本更新 · v4arXiv:2611.03043

VideoEditor-MoE: Scaling Complex Video Editing with Mixture of Experts

Xiao Cao, Fuchen Long, Yansong Qu, Wen Xiao, Xin Zhang, Xinlong Sun, Xi Chen, Zhiyong Huang

原文摘要

Complex video editing requires understanding complex instructions while jointly fulfilling multiple editing requests. Meeting these demands through dense model scaling substantially increases computational cost. We introduce VideoEditor-MoE, the first open-source mixture-of-experts framework for video editing. Beyond sparse MLP-MoE layers, MoE-Attention expands attention capacity through routed key--value experts while preserving global interactions among video tokens. Initialized from a 5B video editing model, VideoEditor-MoE scales to 14B and 56B parameters, with approximately 8B parameters activated per token in both variants. Comparisons with existing methods and two complementary baselines, Explicit-MoE and MLP-MoE, assess its benefits over explicit expert guidance and MLP-only expert architectures. Further analyses demonstrate that all experts remain active and reveal that learned routing responds more strongly to sample content and diffusion noise levels than to human-defined editing categories. Spatial analysis further indicates sensitivity to object-level semantic features. We also identify decoupled weight decay as a strategy for stabilizing expert scaling.

cs.CV · cs.AI

版本更新 · v2arXiv:2605.06404

FRInGe: Distribution-Space Integrated Gradients with Fisher--Rao Geometry

Gabriele Martino, Sebastian Tschiatschek

原文摘要

Although gradient-based attribution methods are model-faithful and scalable, a prominent variant, Integrated Gradients (IG), suffers from high sensitivity to design choices, particularly in baseline selection, path formulation, and discretization, as well as gradient saturation. To address these limitations, we propose Fisher--Rao Integrated Gradients (FRInGe), which defines both the reference and integration paths in the predictive distribution space. FRInGe replaces input baselines with a maximum-entropy predictive reference and follows a Fisher--Rao geodesic on the probability simplex. The corresponding input-space trajectory is realized through the pullback Fisher metric and stabilized by Kullback--Leibler (KL) and Euclidean trust regions; attributions are obtained by integrating input gradients along this trajectory. We further derive FRInGe-B, a variant that binarizes the predictive probability distribution, removing the computationally expensive Fisher products of FRInGe. Across six ImageNet architectures, FRInGe most clearly improves Magnitude Aligned Scoring (MAS), while remaining competitive on perturbation area-under-the-curve (AUC) metrics and infidelity.

cs.LG

新论文 · v1arXiv:2611.25861

Decoupled Knowledge Retrieval and Compact Memory Enable Reliable Lifelong Model Editing

Yue Che, Jingze Wang, Chen Guoyi, Yujie Wu, Xuming Ran, Mingkun Xu

原文摘要

Lifelong model editing requires continuously incorporating new facts into deployed large language models (LLMs) without expensive retraining or compromising existing capabilities. Memory-based editors offer an appealing solution by keeping the base parameters untouched, yet their practical use is hindered by linearly growing storage requirements and deteriorating retrieval performance as edits accumulate. To tackle these problems, we introduce BiSpace, a lifelong model editing framework that decouples memory retrieval from knowledge injection across two distinct representation spaces. BiSpace routes queries and identifies relevant edits in an external semantic space, while injecting the corresponding knowledge into the internal representation space of the LLM through sparse and storage-efficient updates. Extensive experiments show that BiSpace shrinks each entry's storage to only about 10.2% of the original and shows strong resilience to long edit sequences of up to 16K edits. On a more challenging benchmark, BiSpace further improves Overall Performance by more than 23% over existing methods. These results advance memory-based editing toward reliable long-term model maintenance at scale.

cs.AI

新论文 · v1arXiv:2611.25853

TableRoute: Capability-Aware Routing for Scalable Table Question Answering

Yuxiang Wang, Junhao Gan, Jianzhong Qi

原文摘要

Table question answering with large language models remains difficult to scale: tables range from compact to million-token sizes, while questions require precise grounding of columns, values, and computations. Existing methods often specialize in a fixed inference strategy, which is brittle across different table scales and model capabilities. Full-table reading is efficient for compact inputs but suffers from irrelevant context. Table reduction helps long inputs but depends on reliable evidence localization, while execution is required when tables exceed the model context window and remains sensitive to schema and value grounding errors. We propose TableRoute, a capability-aware routing framework with evidence-conditioned inference for scalable TableQA. TableRoute first generates validated question--table evidence and identifies the required operations, then routes each instance among these three strategies according to serialized table length and calibrated estimates of how reliably the base model handles these operations. The validated evidence supports the selected route by grounding columns, values, and conditions. Across WikiTQ, WTQ-L, BirdQA, and TableBench, TableRoute consistently improves accuracy through complementary routing and evidence conditioning across table scales, while reducing online token cost. Code will be made available upon publication.

cs.CL

新论文 · v1arXiv:2611.25849

EgoWitness: Benchmarking Instance-Level Spatial Memory and Reasoning in Egocentric Video Streams

Zhengru Fang, Zehao Liu, Mingyu Tsoi, Yaoyu Zhu, Junlong Li, Fengmao Lv, Sijie Cheng, Yuanliang Sun 等 10 位作者

原文摘要

Recalling where and when an object was last seen and who saw it requires recognizing it across days and people. Existing long-video benchmarks focus on events, object categories, or relative locations, leaving instance-level spatial memory largely untested in multimodal models. To address this gap, we present EgoWitness, a benchmark for instance-level spatial memory and reasoning in egocentric video. Its 8,453 human-verified questions span approximately 324 hours of week-long and short-horizon recordings and test seven abilities, from recognizing an object or room to recalling who saw it and reasoning about room visits. Over one million reviewed masks of 34 registered landmarks and 21 movable-object labels are linked through room timelines. Among twelve direct vision-language models (VLMs), open-weight models reach at most 46.97% accuracy, and none exceeds 35% on identifying who saw an object last. We therefore propose EgoClue, a spatial-memory agent that retrieves candidate intervals, verifies instance identity, and computes spatial and temporal relations with five tools. Two-stage training with supervised fine-tuning (SFT) and Group Relative Policy Optimization (GRPO) raises its 4B controller from 23.64% to 72.65% accuracy on 1,320 held-out questions. Evidence-trace synthesis matches the best direct VLM at 74.85% on week-long questions and achieves 80.56% in environments excluded from training.

cs.CV

版本更新 · v2arXiv:2610.32678

Probing the limits of materials design via a comprehensive crystal database of 14.6 million structures with 140 properties

Zhilong Song, Zijun Pang, Lixue Cheng

原文摘要

Forward screening and inverse design drive materials discovery, yet their limits remain unquantified because records of known crystal structures and properties are fragmented across databases. Consequently, forward screening leaves known crystals underexplored for new functions, whereas generative and autonomous discovery has repeatedly reported known phases or close variants as new materials. Here we match structures across 55 databases to assemble MateriaViva, a unified database of 14.6 million distinct crystals and over one billion property values. To provide the missing properties of every structure, a task we term property completion, we develop an equivariant Crystal Graph Transformer Network with state-of-the-art accuracy on nine benchmark tasks and apply 140 property-prediction models across the database. These predictions reveal the underestimated potential of forward screening, identifying tens of thousands of high-performance candidates among previously synthesized or computationally studied materials that lack recorded target-property values. Conversely, comparison with MateriaViva shows that the novelty of inverse design has been overestimated, as 55-100% of inverse-designed structures are not new. Under equal density functional theory verification budgets for bulk-modulus design, inverse design yields a higher fraction of qualifying candidates, whereas forward screening reaches the higher bulk modulus. MateriaViva thus provides a comprehensive reference for probing the limits of both paradigms, revealing unrecognized performance in known crystals and directing inverse design beyond reported structures.

cond-mat.mtrl-sci · cs.LG

新论文 · v1arXiv:2611.25841

Whose Risks Count? What 1,901 AI Incidents Reveal About Corporate Disclosures

Simeon Allmendinger, Edyta Paulina Bogucka, Niklas Kühl, Daniele Quercia

原文摘要

AI systems can harm people, yet the AI actors that provide or deploy them are rarely obliged to disclose these risks. US securities regulation is an exception. Companies registered with the Securities and Exchange Commission (SEC) must disclose their material risks, including those of their AI systems, in filings that carry liability for misstatement. Materiality, however, is defined by what a reasonable investor would consider important, so harms to third parties enter disclosure only insofar as they become financial risks for the company. We test whether AI harms do so by matching 9,001 AI incidents to 18,229 SEC filings of 328 AI actors by risk type. Of the 1,901 incidents with filings before and after them, half belong to a risk type that the AI actor never disclosed, and this share varies with the kind of harm. Filings disclose the risk types of 67\% of system failures but only 2\% of social system harms, and AI actors themselves report fewer than 2\% of incidents. Securities disclosure thus records the risks that AI actors face rather than the harms that people subject to AI bear. Public authorities that hold AI actors accountable for such harms could make them financially material. We release DISC, the Disclosure--Incident Securities Corpus, to support this work.

cs.AI

新论文 · v1arXiv:2611.25838

ContextRenderer: Dynamic-Context Video Generation for Interactive Game Worlds

Runze Wang, Fei Zha, Bing He, Ke Xu, Zhongyuan Wu, Yuxuan Xie, Wenzheng Wu, Zhongyuan Liu 等 10 位作者

原文摘要

Interactive game world models separate executable world evolution from visual generation, using video models to render engine-provided scene representations into detailed observations. During sustained gameplay, however, existing generative renderers struggle to follow changing conditions while preserving visual continuity across diverse proxies and target styles. We therefore introduce ContextRenderer, an autoregressive video diffusion framework designed for both dynamic conditioning and cross-proxy generalization. It couples Dynamic-Context Learning (DCL) for chunk-level interaction supervision with Recurrent History Compression (RHC) to follow condition updates while preserving visual continuity within a bounded token budget. Across proxy representations, Cross-Representation Flow Alignment (CRFA) connects proxy reconstruction with appearance and interaction variation through a shared high-fidelity reconstruction. WorldPrism, our interaction-centric dataset constructed from agent-built worlds, AAA gameplay, and open-source videos, supplies the corresponding renderings and chunk-level annotations for these learning tasks. Experiments on GWM-Bench and Solar-Bench show improved reconstruction and visual quality, while qualitative results demonstrate minute-long interactive rendering under changing prompts and subject references. Project page: https://rainzor.github.io/contextrenderer/

cs.CV · cs.GR

版本更新 · v2arXiv:2609.29589

Online Bin Packing with Per-Bin Maximum Delay

Tianhang Lu, Runtian Ren, Shengcai Liu

原文摘要

In online bin packing with delay, items of sizes in $(0,1]$ arrive over time and may wait before being packed into unit-capacity bins. A service selects a nonempty feasible batch of pending items and seals it in a new bin, at a service cost of one. We study the objective that adds one maximum waiting time for each sealed bin to the total service cost. Thus a bin's delay is determined by its oldest item, and younger items that fit can join it without increasing that delay. We give a randomized online algorithm with strict competitive ratio $2$ against an oblivious adversary, running in polynomial time in the ideal random-real model. It maintains one Next-Fit buffer and uses a single random silence threshold. Its analysis combines an offline-bin fragmentation bound with a two-unit saving per silence cluster. The ratio is tight for this policy; $b$ fair bits suffice for ratio $2+2^{-b}$. A fixed threshold gives deterministic ratio $1+\sqrt3$. The lower bounds $2$ and $e/(e-1)$ are attained when capacity is nonbinding. Offline, the problem minimizes the number of bins plus their total arrival span. We prove strong NP-hardness, an absolute approximation barrier of $3/2$, a deterministic polynomial $2$-approximation, and exact algorithms for several special cases. We also give stronger online ratios with exact endpoint-packing oracles and an AFPTAS for the fixed-length weighted endpoint subproblem. Finally, for Poisson arrivals of half-size items, we determine the offline and optimal causal long-run cost rates. Two natural asymptotic ratio measures coincide and have maximum $4/3$.

cs.DS

版本更新 · v2arXiv:2609.16546

GPUThor: Amplifying Rowhammer Attacks via Non-Uniform Patterns to Exploit ECC-Protected GPUs

Chris S. Lin, Joyce Qu, Aditya Rajeev, Gururaj Saileshwar

原文摘要

GDDR memory in GPUs is vulnerable to Rowhammer attacks, where rapid memory accesses induce bit flips in adjacent cells, enabling data tampering and privilege escalation. However, prior GPU Rowhammer attacks trigger only tens to hundreds of bit flips, orders of magnitude fewer than CPU attacks, severely limiting their practical impact. This gap stems from the reliance of existing GPU Rowhammer attacks on uniform hammering patterns that activate aggressor and decoy rows equally, which results in low hammering intensity for aggressor rows. We present GPUThor, a high-intensity Rowhammer attack on NVIDIA GPUs leveraging non-uniform hammering. GPUThor reverse engineers GPU memory-access coalescing behavior to enable non-uniform hammering patterns on GPUs, that activate aggressor rows more intensely than decoy rows. Additionally, by identifying refresh instances when in-DRAM mitigations are applied, it constructs longer attack patterns that escape mitigation across refresh intervals, further increasing hammering intensity. Together, these techniques yield 500X to 23,500X more bit flips than prior GPU Rowhammer attacks, across several NVIDIA GPUs (A4000, A4500, A5000, A6000), reaching bit flip rates close to state-of-the-art CPU Rowhammer attacks. GPUThor also enables the first Rowhammer exploits on ECC-protected GPUs, inducing uncorrectable double and triple bit flips, making denial-of-service and privilege-escalation attacks practical even on GPUs with ECC enabled.

cs.CR

版本更新 · v2arXiv:2611.21799

Controllable Chains of Thought Can Reveal Less, Even Without Evasion Training

Zheng-Xin Yong, Benjamin Arnav, Miles Wang, Tomek Korbak, Olivia Watkins

原文摘要

Chain-of-thought (CoT) monitoring is a promising tool for detecting hidden goals, off-task behavior, and other failures in AI agents. One risk is that models could deliberately shape their reasoning to make it opaque to monitors. We stress-test this risk through persona-control training: supervised fine-tuning on mathematical reasoning traces rewritten in the voices of diverse personas. This induces a broader ability to control CoT, even though training contains no monitor-evasion examples. In covert side-task evaluations, traces in which a gpt-oss-20b model organism completes the hidden side task contain less monitorable evidence than the base model's. Averaged across evaluations and CoT lengths, weighted by side-task-success counts, the CoT monitorability metric g-mean$^2$ falls from 0.94 for the base model to 0.68 for the model organism without a hiding instruction, and to 0.22 when explicitly instructed to hide. In intervention evaluations, explicit hiding often drives monitorability near zero. The effect extends to a Kimi-K2.6 model organism in agentic environments, though hidden objectives requiring extended reasoning or longer trajectories remain more monitorable than simple single-step side tasks. We further discover that targeted transparency prompts can substantially recover monitorability, sometimes surpassing the base model. These model organisms provide a testbed for studying monitorability failures and potential mitigations. Our code can be found on https://github.com/yongzx/cot-control-model-organism.

cs.AI · cs.LG

新论文 · v1arXiv:2611.25819

SCCV-6D: Sparse Canonical Coordinate Voting for Holistic Multi-View RGB-D 6D Pose Estimation in Robotic Bin Picking

Nico Leuze, Samed Dogan, Maximilian Hoh, Henry Schaub, Alfred Schöttl

原文摘要

Industrial bin picking requires pose estimates that are accurate under clutter, partial occlusion, and the reflective, texture-less nature of manufactured parts. Most recent approaches to 6D pose estimation follow multi-stage pipelines. These achieve high accuracy, but incur memory and runtime footprints that do not cope with the operational constraints of many real-world industrial deployments. Holistic pipelines estimate poses in a single pass, with cost controlled through indiscriminate subsampling or lowered resolution. Continuing this line, we present SCCV-6D, a fully sparse multi-view RGB-D estimator. We reconstruct the 3D space at millimeter precision, prune background via a staged heatmap mechanism, and predict class, translation offset and rotation at highest resolution for an arbitrary number of objects simultaneously. Rotation is formulated as canonical coordinate voting. Points vote for their canonical object coordinates with an associated confidence, enabling closed-form pose recovery with a symmetry-aware weighted Procrustes solve. Along an in-cluster point recovery induced by offset votes, this removes time-consuming post-processing stages. We reuse the task-supervised attention priors from the heatmap formulation for a novel attention-gated distillation objective that reweights teacher-student feature alignment. On the IPD test split we report competitive 0.826 AP at 1.13 s per scene, significantly faster than the leading entries. On MV-YCB our approach reaches 96.1 ADD-S, and remains competitive even on the densely packed bins of ROBI.

cs.CV

版本更新 · v2arXiv:2609.40119

Function Contracts and Beyond: Machine-Checked Assurance Cases for Conformance of C Libraries to Their Requirements

Haokun Li, Zhongyi Wang, Mingshuai Chen, Ling Zhang, Guanyan Li, Depeng Liu, Xiao Yi, Shengchao Qin 等 9 位作者

原文摘要

Large language models (LLMs) have shown promise in automating interactive theorem proving, yet verification of real-world C codebases requires more than discharging individual proof goals. The task involves jointly constructing expressive function specifications and their proofs, and ensuring that library interfaces compose along intended call sequences even without a designated client. We present Accord, an LLM-assisted framework for building machine-checked assurance cases: structured, auditable artifacts supporting the claim that a C codebase meets its intended requirements. To model intended cross-interface use in open libraries, Accord constructs an interface protocol that exposes permitted call sequences and resource assumptions for review, with a conditional safety guarantee under verified contracts and caller obligations. Accord coordinates two complementary phases: (i) requirement-guided analysis and bottom-up construction of candidate specifications and protocols; and (ii) modular proof construction with feedback that revises the specifications and proofs. Implemented using VST in Rocq, Accord establishes memory safety for 259 function definitions across seven implementation cases from six benchmark families, including industrial cryptographic modules, and additional functional correctness for red--black trees and ML-DSA. Human review took an estimated 0.17--8.00 hours per case. The guarantees depend on disclosed contracts and assumptions; human review supplies the conformance judgments connecting the formal artifacts to the intended requirements.

cs.PL · cs.SE

新论文 · v1arXiv:2611.25801

Fast and Accurate Body Estimation in Pixels, World Space and Scenes

Jinkun Cao

原文摘要

We present FAB (Fast and Accurate Body estimation), a human body estimator that jointly improves inference efficiency and body-recovery fidelity. Rather than predicting only parametric mesh-model parameters, FAB directly predicts human skeleton joints and dense surface points. In images, it predicts dense 2D and 3D body points alongside a parametric Momentum Human Rig (MHR) head, then fits SMPL in one learned pass followed by two Gauss-Newton steps. Its ViT-L student, FAB-L, distilled from a fine-tuned SAM 3D Body, runs in real time for one person on an RTX 4070 Ti SUPER, from detection to a fitted body, at 4.6x the speed of SAM 3D Body's eager-PyTorch body-only path, while improving accuracy on most camera-space benchmarks. The same network predicts heel, toe, hand, and knee contacts, which connect the body to the scene. FAB also estimates world-space motion, fuses static-scene depth around the visible person, and aligns body and scene using predicted contacts. For human-to-humanoid real-to-sim transfer, an in-model G1 head and FAB-Retarget, a motion-level retargeter that reads dense points or the fitted mesh, map motion to a Unitree G1. FAB-Retarget produces references that follow the person more closely than specialized retargeters including SOMA, GMR, UMR, and OmniRetarget, and substantially more closely than Fast SAM 3D Body retargeted with our sequence IK. Project page: https://www.jinkuncao.com/FAB.

cs.CV · cs.RO

新论文 · v1arXiv:2611.25798

Design and Validation of a Telescopic Chassis-Assisted Stair-Climbing Mechanism for Powered Wheelchairs

Yasuhiro Fukuoka, Haruya Shirado, Yuto Nakamura, Haruto Okabe, Ahmad Najmuddin Ibrahim, Takahiro Fukui

原文摘要

We propose a simple stair traversal mechanism that augments a four-wheel wheelchair with an auxiliary telescopic chassis actuator. The proposed mechanism provides the following key contributions during stair ascent: (A) enhanced safety by ensuring that either the front or rear wheels remain stationary at all times; (B) primary reliance of the required force and mechanical work on the telescopic chassis rather than the wheels; (C) insensitivity of the required force and mechanical work to stair inclination; and (D) reduced risk of wheel slip on steeper stairs. In addition, the telescopic chassis is expected to improve level-ground mobility by enabling tight turning with the shortest wheelbase and enhancing stability with the longest wheelbase.

cs.RO

版本更新 · v2arXiv:2608.00608

From field-scale to large-scale spectral libraries: Tabular foundation models in soil spectroscopy

Viacheslav Barkov, Jonas Schmidinger, Robin Gebbers, Martin Atzmueller

原文摘要

Visible and near-infrared (vis-NIR) and mid-infrared (MIR) spectroscopy enable rapid, cost-effective prediction of soil properties. Yet, translating high-dimensional, highly collinear spectra into accurate soil property predictions remains challenging, particularly when employing machine learning. We systematically investigated regression models and dimensionality reduction approaches for spectroscopic modeling across 85 soil property prediction tasks from open benchmark datasets in pedometrics, spanning field-scale digital soil mapping and a global soil spectral library. We compared an in-context learning tabular foundation model (TabPFN), a convolutional neural network (CNN), rule-based regression (Cubist), Random Forest, and partial least squares regression (PLSR) using full spectra as well as features derived from principal component analysis (PCA) and partial least squares (PLS) latent variables. TabPFN consistently delivered the best overall performance across spectral regions, properties, and scales, including large spectral library tasks with tens of thousands of soil samples. Notably, TabPFN applied directly to full spectra already surpassed all classical baselines, showing that explicit dimensionality reduction is not strictly required for strong performance. Further improvements were achieved through PLS, which proved to be an effective dimensionality reduction strategy for all models. Combining PLS latent variables with TabPFN yielded the best predictions overall, lowering the root mean squared error (RMSE) relative to the strongest classical baseline on 44 of the 46 large-scale spectral library tasks, by a median of 14%. Our findings provide evidence-based guidance for spectroscopic calibration model selection across operational scales, demonstrating that the long-standing advantages of PLSR and modern tabular foundation models complement each other in chemometrics.

cs.LG

新论文 · v1arXiv:2611.25793

When Safety Rationales Hallucinate: Diagnosing and Grounding Multimodal Safety Guards

Yingshuo Wang, Weixiang Zhao, Guancheng Li, Jiahe Guo, Yu Zhang, Yanyan Zhao

原文摘要

Multimodal safety guards increasingly generate rationales alongside safety predictions, yet whether these rationales are reliable remains largely unexplored. We systematically study *rationale hallucination* from two perspectives: visual rationale hallucination, where rationales contain unsupported visual facts, and safety attribution hallucination, where grounded content is assigned an incorrect safety interpretation. Across rationale-generation methods, multiple vision-language backbones, and safety datasets, we find that rationale hallucination is widespread and does not consistently diminish with higher classification accuracy or increased model scale. Further analysis of the Think-with-Image paradigm shows that hallucination is concentrated in fine-grained attributes and relations, while visual tools are often invoked infrequently or without sufficient spatial focus. Based on these findings, we propose *Tool-Contrastive Grounding Reinforcement Learning* (TCG-RL), which rewards tool use only when it improves rationale grounding over a tool-free reference. TCG-RL substantially reduces rationale hallucination while maintaining competitive safety prediction performance. By revealing the gap between prediction correctness and rationale reliability, our study motivates a shift in multimodal safety guarding from outcome-only evaluation toward safety decisions that are both accurate and verifiably grounded.

cs.AI

新论文 · v1arXiv:2611.25778

ECHO: In-Context Behavior foundation model for Humanoid Whole-Body Control

Xingyi Wang, Bin Li, Jiyuan Shi, Chenyun Zhang, Weiji Xie, Xiangyu Miao, Junsong Wu, Wei Song 等 10 位作者

原文摘要

Humanoid behavior foundation models (BFMs) acquire diverse skills through large-scale motion imitation and massively parallel reinforcement learning, yet broad behavioral capability does not ensure precise execution across physical conditions. Without distinguishing the underlying dynamics, policies may favor conservative compromises over environment-specific precision. We introduce ECHO, an in-context BFM that turns interaction experience into environment-specific control, separating general behavioral capability from dynamics-specific execution. Mixed and Annealed Sampling first establishes a behavior prior in a single pretraining stage by balancing targeted practice with broad motion coverage. Building on this prior, Context Representation and Residual Adaptation learns structured dynamics representations from interaction histories to guide residual corrections to the frozen prior, enabling adaptation without test-time parameter updates. Extensive experiments show that ECHO outperforms all baseline methods across diverse motions and physical conditions. Moreover, deployment on the Unitree G1 demonstrates whole-body teleoperation spanning extreme acrobatics and terrain interaction, alongside payload adaptation in manipulation and whole-body control. Page is available at https://mwondering.github.io/echo-bfm/

cs.RO

新论文 · v1arXiv:2611.25773

Duopra: From Full-Duplex Conversation to Full-Duplex Agency

Chiyu Chen, Pengzhou Cheng, Yuezhao Ma, Xiaocheng Li, Can Liu, Haonan Hu, Yao Yao, Yifei Yang 等 15 位作者

原文摘要

Full-duplex interaction models exhibit strong conversational agency, enabling natural, low-latency turn-taking, backchannels, and interruptions. However, completing tasks further requires task agency: timely tool invocation and argument revision following user corrections. Frontend-backend agents separate initiation from execution: the frontend decides whether and what to request, while the backend determines how. Execution can be offloaded, but unsent requests cannot be executed, and incomplete ones can misdirect the backend. On Full-Duplex-Bench v3 (FDB v3), most failures of Realtime-Venus-Audio + Harness in our streaming pipeline occur before a domain tool executes. We introduce Duopra, a full-duplex interaction model that learns task execution alongside conversational control. Duopra treats the backend as a tool, accepting a self-contained natural-language request. Direct calls and backend delegation share an action interface and requirements for faithful initiation and correct execution. Its three-stage training combines protocol SFT, full-duplex RL rewarding timely and complete execution, and projector-only repair of talker conditioning after thinker updates. With the same backend, Pass@1 rises from 4.7% to 44.7%; Duopra achieves 47.7% calling tools directly (26.0% after SFT) and 49.3% with a DeepSeek backend. It improves three FDB v1.0 takeover rates over the base model. These results show that task agency can coexist with key conversational agency.

cs.AI

新论文 · v1arXiv:2611.25761

AdaptReg: Skip-level FiLM Adapter for Task-Specific Adaptation of a Frozen Registration Foundation Model

Shuyue Shi, Rong Yuan, Lin Tian

原文摘要

Longitudinal PSMA-PET/CT registration is a key step in assessing treatment response in prostate cancer, and it requires more than generic alignment accuracy: the deformation must preserve lesion volume, so that measured changes in metabolic tumor volume (MTV) and total lesion glycolysis (TLG) reflect treatment response rather than registration artifacts. We propose AdaptReg, which adapts a frozen registration foundation model to such task-specific requirements through a lightweight trainable adapter branch with identity-preserving initialization. Keeping uniGradICON fully frozen, we apply FiLM-style per-voxel affine modulation ($γ$, $β$) to its decoder skip connections; the modulation is generated from anatomical semantic features of a frozen TotalSegmentator encoder and supervised by a lesion volume-preservation loss. On the official Codabench validation set (20 pairs), AdaptReg achieves DSC 0.7585, HD95 6.33 mm, MTV/TLG errors of 1.68%/2.11%, and 0.0022% non-diffeomorphic voxels, improving on the uniGradICON + instance optimization baseline by $+0.05$ DSC and reducing the MTV/TLG errors by 70.9%/67.7%. On the hidden official test set (160 pairs), AdaptReg achieves CT DSC 0.816, PET DSC 0.598, and MTV/TLG errors of 4.02%/3.46%, ranking fourth among 16 teams in the final PSMAReg evaluation. Code: https://github.com/shishuyue/AdaptReg

cs.CV · cs.LG

新论文 · v1arXiv:2611.25760

Characterizing and Detecting Unfaithful Work Reports by Coding Agents

Wenhao Yang, Runzhi He, Minghui Zhou, Pengfei Gao, Dayi Lin, Qingwei Lin, Dongmei Zhang

原文摘要

Coding agents increasingly carry out long-horizon development tasks, autonomously planning and executing multi-step actions that span files, tools, and sessions. Because developers rarely follow each action, and agents resuming the work depend on records of what was done, the agent's work report often becomes the primary account of what was changed and established. When the report contains false or unsupported claims, it gives recipients false assurance: incomplete work may be accepted without verification, incorrect assumptions about the current state can propagate to later steps and agents, and evaluators that credit stated success may reward the report rather than the work. Yet little is known about what unfaithful reporting looks like in real development work, or how developers and agents handle it once it surfaces. Starting from real developer--agent conversations in SWE-chat, we conduct a systematic mixed-methods analysis of self-admitted unfaithful reports. We identify eleven subtypes in four categories: misreported observations, misattributed evidence, overstated task success, and faulty technical explanations, with overstated task success the most common. Handling these problems involves on average 4.3 further user turns and 12.71 minutes of recorded response and execution time, and users express dissatisfaction more often. Among 435 work reports that we separately sample and annotate, \NaturalPositiveShare{}\% contain at least one unfaithful claim. We also develop FaithDetector, a detector for unfaithful reporting informed by the observed patterns. On naturally sampled reports, FaithDetector attains \NaturalDetectorMacro{}\% macro-F1, about 13 percentage points above the baseline. Together, our findings and FaithDetector provide empirical evidence and an automated method for checking reporting faithfulness, informing agent evaluation and training feedback.

cs.SE · cs.AI

版本更新 · v2arXiv:2606.05263

Learning Which Evidence-Seeking Decisions Matter: Calibrated Counterfactual Replay for Language Agents

Renwei Meng, Chengyi Sun, Hongyi Zhang, Xiaoyi Hou

原文摘要

Language agents answer questions by searching, reading, and checking claims. A reward on the final answer says little about which of those decisions earned it, since an agent can reach the right string after redundant searches or with half the evidence missing. We treat the question as an experiment. At a saved decision point, PCCC-RL, for prefix-conditioned counterfactual credit, restores the tool state, runs the action the agent chose alongside a feasible alternative, and lets the same frozen policy finish both branches. The difference in final score then samples what the decision was worth. Replaying everything is too expensive, so the method predicts each action's credit up front, checks a random subset, and uses the recorded selection probabilities to correct the errors of that prediction, spending the budget where a further check can still move the training signal. Alternatives are drawn from the prefix alone, which keeps the comparison a usable gradient baseline. On HotpotQA and MuSiQue, joint exact match, requiring a correct answer together with complete supporting evidence, reaches 56.71% and 36.82%, gains of 4.19 and 3.86 points over a baseline with the same replay access. The gain concentrates in complete evidence chains and accumulates over many updates. Single-update effects stay uncertain, and citation coverage and scope distinctions such as "none" versus "not all" remain open.

cs.LG · cs.AI

新论文 · v1arXiv:2611.25742

Goal-Guided Reinforcement Learning via Fine-Tuning Vision-Language Models

Yajie Qi, Wei Wei, Lin Li, Lijun Zhang, Jianye Hao

原文摘要

Exploration remains a major challenge for reinforcement learning (RL) agents in Craftax-style environments due to sparse rewards and long horizons. Large language models (LLMs) can guide exploration by generating high-level goals, but methods using models without explicit task adaptation often rely on carefully designed prompts and may produce inconsistent or contextually inappropriate goals. We introduce CCGP-2M, a multimodal goal-preference dataset constructed from interaction trajectories in Craftax-Classic, and propose a vision-language model (VLM)-guided RL framework that learns environment-specific goal guidance from this dataset. The framework consists of two stages: offline VLM fine-tuning and online goal-guided RL. In the offline stage, we align the VLM with goal preferences using Kullback-Leibler Token-level Direct Preference Optimization (KLTDPO), which combines token-level KL regularization with distributionally robust reweighting. Experiments on Craftax-Classic show that our method collects diamonds earlier than state-of-the-art baselines, achieves the highest score and reward, and attains the highest success rates on most achievements.

cs.LG

新论文 · v1arXiv:2611.25733

Can LLMs Resolve Entity Conflicts in Domain Specific NLU Pipelines?

Akansha Bhardwaj, Mingbo Cui, Victor Guimaraes, Dominik Kohler, Diana Nicoleta Popa, Meghana M. Reddy, Paolo Rosso

原文摘要

Resolving entity conflicts is a critical step in Pipeline-based NLU architectures, especially when multiple extractors provide overlapping or conflicting outputs. Traditional rule-based entity conflict resolution methods often lack flexibility and are unscalable. In this work, we aim to study if the broad knowledge of LLMs can be leveraged to resolve entity conflicts in domain specific NLU pipelines. We present an experimental comparison between a traditional rule-based conflict resolution method, and two LLM-based approaches: a multi-shot prompting method using the Llama3.1-8B-Instruct model and a LoRA fine-tuned Llama3.1-8B-Instruct model. The evaluation was conducted under two setups: (1) a broad assessment of LLM's capability to match existing NLU pipeline output, and (2) a more controlled scenario assessing the LLM accuracy on human-labeled data. Results show that while multi-shot prompting offers ease of adaptation and rapid behavioral changes, the fine-tuned model delivers superior performance on complex, ambiguous cases. Ultimately, the best approach appears to be a combination of both strategies, leveraging multi-shot prompting's flexibility alongside the broad accuracy gains of LoRA fine-tuning for a more comprehensive performance.

cs.IR · cs.LG

新论文 · v1arXiv:2611.25721

MemMend: A Selective Revision Framework for Memory Question Answering

Sheng Shi, Geng Li

原文摘要

Revising every answer from a long-term memory can consume substantial inference budget and damage facts that were already correct. We introduce MemMend, a modular selective revision framework that routes potentially incomplete answers to an editor, represents corrections as bounded exact-text replacements, and accepts only edits with resolvable source coordinates and valid edit operations. The editor preserves the current answer when no legal repair is proposed; accepted edits require no additional semantic-verification call. On 1,986 LoCoMo development questions, MemMend achieves 67.17 conversation-macro F1 and 88.35 mean Judge score, improving over a verified revision chain by 0.82 and 1.24 points while using 62.37% fewer revision calls. Routing uses 54.15% fewer revision calls than ungated local review, with similar F1 but a 0.67-point lower Judge score. Applied directly to native PropMem, Mem0, and SimpleMem answers, the editing interface raises F1 by 3.79, 6.44, and 0.42 points. It outperforms same-gate rewriting and saves 49.85-83.72% of editing calls relative to ungated editing, which retains higher F1 in this adaptation. A source-by-editing factorial supports local editing for F1 but leaves an independent quality benefit from source obligations unresolved. Transfer to 451 LongMemEval-V2 Small questions also favors revision efficiency, while an external PerLTQA dialogue subset reveals a completeness gap. These results support selective editing as a reusable quality-cost operating point.

cs.AI

新论文 · v1arXiv:2611.25717

VR-SparseQuant: Accurate FP4 Quantization with 2:4 Sparsity on Vera Rubin

Samir Khaki, Wei-Chen Wang, Song Han, Ryan Hanrui Wang

原文摘要

In recent years, GPUs have incorporated increasingly specialized support for low-precision computation, from Hopper's FP8 to Blackwell's FP4 Tensor Cores. Precision formats have evolved alongside this hardware, from MXFP4 and NVFP4 on Blackwell to E2M1 FP4 with direct UE5M3 block scaling on Vera Rubin, which we term as $\boldsymbol{VRFP4}$. These advances help serve larger language models at lower cost, but four-bit activation quantization can substantially degrade model quality. We introduce VR-SparseQuant, which uses Vera Rubin's elementwise 2:4 sparsity to recover accuracy while retaining native FP4 execution. By allowing two sparse streams to overlap, we use the same FP4 value budget as dense quantization to refine activations with larger errors. Offline scale-ratio calibration avoids runtime search, while a shared-weight GEMM accumulates both streams efficiently. On Nemotron-3-Nano-30B-A3B, VR-SparseQuant reduces WikiText-2 perplexity from $\boldsymbol{6.85}$ to $\boldsymbol{5.57}$ versus dense RTN W4A4. Including activation preparation, our native Rubin kernels for VRFP4 achieve up to $\boldsymbol{3.1\times}$ GEMM and $\boldsymbol{2.7\times}$ MoE speedup over BF16. To the best of our knowledge, this is the first study of adaptive activation quantization using Vera Rubin's elementwise 2:4 structured sparse FP4 instructions.

cs.CL