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

+0新增论文
+0其中 AI 相关
+2已有论文修订
+100新增 AI 导读

arXiv 最近完成扫描：2026-11-15 05:41 · 全站最近入库：2026-11-14 15:09各信源进度 →

统计口径与数据新鲜度

统计 2026-11-14 06:31 — 2026-11-15 06:31（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：2026-11-15 05:41最近发现新增版本：2026-11-14 15:09库内论文最新更新时间：2026-11-13 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。

23,348 已收录论文19,580 AI 相关6,729 有修订112 HF 社区发现

浏览论文库

研究日报 / 周报 →按信源浏览 →

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 33 篇论文 · 按版本更新时间排序上轮扫描起点 2026-11-15 05:40（北京时间）
新论文 · v1arXiv:2611.15160

Measuring, Revealing and Mitigating Harmful Content Propagation in Recommender Systems

Xiaokun Zhang, Youlin Wu, Junyu Lu, Jiamin Chen, Zhaochun Ren, Hongfei Lin, Fenglong Ma, Xue Liu 等 9 位作者

原文摘要

Recommender systems help users navigate vast amounts of information by filtering content according to their historical behaviors. As these systems increasingly shape everyday information access, they largely determine the content users are exposed to. However, current research typically prioritizes catering to user interest, while overlooking the existence and propagation of harmful content, such as violence, hate speech, and pornography. In this work, we take a systematic step toward investigating harmful content in recommender systems from three aspects: measuring, revealing, and mitigating. First, we construct datasets with harmfulness-related annotations and present new evaluation metrics to quantitatively measure the harmfulness of recommendation results. Second, our empirical analysis reveals that users tend to engage with harmful content and prevailing recommendation methods often contribute to the proliferation of such content. Third, to address this issue, we introduce a generic framework, MitiHarm, which integrates harmfulness-related signals into both item and user representation learning. Experiments show that MitiHarm can significantly mitigate the propagation of harmful content while preserving user satisfaction. Resources of this work are available at https://github.com/Zhang-xiaokun/MitiHarm.

cs.IR

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2511.23115

Robust Image Emotion Classification via Multi-Level Affective Captioning

Zibo Zhou, Zhengjun Zhai, Huimin Chen

原文摘要

Image emotion classification (IEC) has been studied for many years and has recently attracted increasing attention with the rapid development of deep learning. Although recent studies have benefited from knowledge embedded in pre-trained visual models, the performance is still limited by the "affective gap" between low-level visual features and high-level affective semantics. Psychological studies suggest that language can convey fine-grained and emotion-rich information. Inspired by this observation, we attempt to narrow the "affective gap" by introducing affective text. Specifically, we design a multi-level contrastive loss to detect emotional concepts from images and introduce an emotional-attribute-guided large language model approach to generate affective sentences. Then, a pre-trained language model is employed to combine emotional concepts with affective sentences for IEC. In addition, we introduce a contrastive loss with a semantic-similarity-based sampling strategy to address large intra-class differences and limited inter-class differences in affective datasets. Moreover, unlike many previous studies, we explicitly consider images containing embedded text. Extensive experiments on multiple benchmarks show that the proposed method helps narrow the "affective gap" and achieves competitive performance.

cs.CV

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.13612

Earth Embeddings in Environmental Remote Sensing

Yuchi Ma, Jiaqi Yang, Mengyu Liang, Ivan Higuera-Mendieta, Caleb Robinson, Anthony Ortiz, Min Chen, David B. Lobell 等 10 位作者

原文摘要

Earth observations from satellite platforms have profoundly transformed research and applications across numerous domains. However, uneven observation coverage and frequent data gaps continue to limit their ability to capture Earth's surface dynamics. Moreover, processing multi-modal Earth observations requires substantial computational resources and sophisticated data harmonization frameworks, which may not be readily accessible to many remote sensing practitioners. Earth embeddings have emerged as promising solutions to these challenges. As general-purpose representations converted from multi-source Earth observations, Earth embeddings can support a wide range of downstream tasks with no requirement of task-specific feature design and preprocessing. In this study, we first introduce the theoretical foundations of Earth embeddings. Then, we review several representative Earth embedding products. We next examine the applications of Earth embeddings across six major domains of environmental remote sensing. Finally, we systematically assess the performance of Earth embeddings in four downstream tasks, including hyperspectral image classification, individual tree crown segmentation, crop yield prediction, and aboveground biomass mapping. Our review and case studies suggest that, despite several advantages over conventional Earth observation data, current Earth embeddings do not always outperform remote sensing imagery or hand-crafted features. Earth embeddings should not be viewed as a direct replacement for conventional EO data, but rather as representations that provide stable information across space and time. Remote sensing practitioners may consider evaluating Earth embeddings' suitability for specific tasks and using them strategically, rather than assuming superior performance by default.

cs.AI

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.12420

Path-Level Structural and Semantic Alignment for Hierarchical Multi-Label Text Classification

Xindi Wang, Aifen Xiong, Baiyang Wang, Xuri Ge, Robert Mercer, Frank Rudzicz, Yifan Peng, Zhumin Chen 等 9 位作者

原文摘要

Hierarchical multi-label text classification (HMTC) requires models to capture both structural dependencies among labels and the fine-grained semantics of labels at different hierarchy levels. Existing methods often model hierarchy through prediction constraints or representations of individual labels and local relations. Structural and semantic supervision, however, are usually formulated separately. We instead use complete root-to-node label paths as a shared unit for both. Our framework introduces two auxiliary training objectives while retaining a standard multi-label classifier at inference. PathSimNCE aligns document representations with hierarchical paths using hierarchy-aware contrastive targets. These targets provide graded supervision for structurally related paths while preserving explicit supervision for the ground-truth paths. The complementary semantic objective uses LLM-generated, document-conditioned descriptions of the same positive paths. It thereby provides context-specific label semantics beyond static label representations. Experiments on three benchmarks show consistent improvements: structural alignment is particularly beneficial in deeper hierarchies, while semantic alignment provides additional gains for fine-grained labels.

cs.CL

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2611.04867

One Assistant, Many Memories: Benchmarking Real-World Group Assistance When Recall Is Not Enough

Zhanghao Hu, Linhai Zhang, Qian Zhao, Qi Zhu, Yuan Hua, Di Liang, Jiasheng Si, Xin Zhao 等 11 位作者

原文摘要

The value of memory lies not only in recall but in using it to provide appropriate assistance. Recent benchmarks test memory use in personalisation and agentic tasks, but seldom evaluate assistance without fixed answers or integration across multiple participants. Group chats make both challenges more acute: an assistant must integrate information from multiple members, track revisions and conflicts over time, and produce one response for the group. Yet existing group memory benchmarks rely on synthetic conversations and inherit constrained evaluation from general memory benchmarks, focusing on recall against fixed answers or decisions over a closed set. We introduce GroupAssistBench, the first memory benchmark from real-world group conversations with a deployed group assistant, covering 80,299 messages in 50 groups and costing over $53K. Over the same histories, GroupAssistBench evaluates memory recall (L1) with reference answers and memory use (L2) with service requests without fixed answers across single, distributed, and conflicting memories (M1--M3). Requests are verified to require group history, memories are anchored to their sources, and L2 uses tailored rubrics with penalties derived from real user feedback. Across evaluations of frontier models with full context, retrieval methods, and memory systems, the strongest memory system rivals full context on recall but falls far behind on use, especially for distributed and conflicting memories. Together, two additions that relate retrieved memories and bind remembered facts as response constraints bring consistent gains, showing that group memory systems must improve use, not recall alone.

cs.CL · cs.AI

查看论文
arXiv ↗PDF ↗
版本更新 · v3arXiv:2605.09586

DeformMaster: An Interactive Physics-Neural World Model for Deformable Objects from Videos

Can Li, Zhoujian Li, Ren Li, Jie Gu, Lei Lei, Jingmin Chen, Lei Sun

原文摘要

World models for deformable objects should recover not only geometry and appearance, but also underlying physical dynamics, interaction grounding, and material behavior. Learning such a model from real videos is challenging because deformable linear, planar, and volumetric objects evolve under high-dimensional deformation, noisy interactions, and complex material response. The model must infer a physical state from visual observations, roll it forward under new interactions, and render the resulting dynamics with high visual fidelity. We therefore present DeformMaster, a video-derived interactive physics-neural world model that turns real interaction videos into an online interactive model of deformable objects within a unified dynamics-and-appearance framework. DeformMaster preserves structured physical rollout while using a neural residual to compensate for unmodeled effects, grounds sparse hand motion as distributed compliant actuators for hand-continuum interaction, represents material response with spatially varying constitutive experts, and drives high-fidelity 4D appearance from the predicted physical evolution. Experiments on real-world deformable-object sequences demonstrate DeformMaster's ability to roll out future dynamics and render dynamic appearance, outperforming state-of-the-art baselines while supporting novel action rollout, material-parameter variation, and dynamic novel-view synthesis. Code and video demos are available at the project page: https://can-lee.github.io/deformmaster-web/

cs.CV · cs.RO

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.10484

Can General-Purpose Generative AI Supply Training Data for Medical Vision Models? A Multimodal Retrospective Study

Jiachuan Peng, Kyle Lam, Jierui Qu, Lin Li, Haining Zheng, Bang Zheng, Aniruddha Agarwal, Zaki Almallah 等 13 位作者

原文摘要

Background: Well-annotated, diverse medical imaging datasets remain a major constraint for vision-based medical AI. We evaluated Nano Banana, a general-purpose image generation and editing model, as a source of synthetic medical images for model training, subgroup augmentation, and model auditing. Methods: In this retrospective study across four imaging modalities, chest radiography, fundus photography, and CT models were trained only on synthetic images and externally validated on real-world data against real-data-trained models. Carotid ultrasound models were trained on a private Human Phenotype Project cohort (aged 40-75 years) with and without synthetic augmentation. Clinicians rated image plausibility. Evaluation used AUROC, AUPR, calibration, and decision curve analysis. Findings: Synthetic-only models achieved AUROCs of 0.877 (95% CI 0.843-0.910) for pneumonia (n=14863), 0.963 (0.948-0.975) for retinal detachment (n=1149), and 0.711 (0.600-0.823) for renal calculus (n=4122), comparable to real-data-trained models, but their calibration and net benefit were not consistently more favorable. With mixed synthetic-real training, a carotid ultrasound model produced an ordinal vascular age index that stratified 633 external subjects into subgroups differing in BMI and glucose (both p<0.01), linking vascular aging to cardiometabolic health. To stress-test shortcut learning, editing a moderate tessellated pattern into fundus photographs reduced the true-class probability for central serous chorioretinopathy from 0.996 to 0.822. Clinician-guided editing raised low-quality CT plausibility from 3.10/5 to 4.30/5. Interpretation: Synthetic data can complement real data in model training, subgroup representation, and model testing via controlled editing, but its clinical value should be assessed with calibration, decision curve analysis, and clinician review, not discrimination alone.

cs.CV · cs.AI

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.08985

NotaryMark: Multi-Bit Watermarking for Black-Box DatasetUsage Auditing in Text-to-Image Diffusion Models

Xinyi Yu, Linkang Du, Zhou Su, Min Chen, Yuntao Wang, Zhikun Zhang

原文摘要

As generative artificial intelligence (AI) models increasingly require data, more publicly available yet licensed data is being harvested and used without authorization. Although several previous studies have proposed watermarking methods for dataset copyright auditing, their applicability to the text-to-image (T2I) setting remains notably limited. The core drawback is that watermarks embedded by existing techniques are often treated as noise and are gradually filtered out and attenuated during the latent noising-denoising process of diffusion model training. We propose NotaryMark, to our knowledge the first multi-bit watermarking method specifically designed for black-box auditing of T2I training dataset use. The core idea of NotaryMark is to align the watermark with the latent denoising of diffusion models, thereby overcoming the signal decay that plagues existing multi-bit methods during sampling. We evaluate NotaryMark across three datasets and three diffusion models spanning watermark capacities from 8 to 1024 bits. With a watermark capacity below 32 bits, it achieves a high average bit accuracy of 0.85-0.90. With a watermark capacity exceeding 32 bits, it maintains bit accuracy between 0.67 and 0.80, consistently outperforming existing multi-bit watermark methods whose bit accuracy is around 0.42-0.60. Furthermore, we design a voting-based aggregation scheme to boost the bit accuracy with a modest budget of 30 queries. Building on this, we incorporate error-correcting codes to enable reliable decoding and tracing of the embedded identifier information. Robustness evaluation against fifteen evasion attacks spanning seven categories shows that NotaryMark remains robust under most evaluated settings, while stronger attacks generally suppress watermark evidence only at the cost of substantial utility degradation, demonstrating its practical reliability against malicious tampering.

cs.CR

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.06910

FrontierPhysics：评估面向端到端物理研究的智能体

FrontierPhysics: Evaluating Agents for End-to-End Physics Research

Bingran You, Wenjun Ke, Ria Rosenauer, Bo Liu, Wenbo Chen, Min Chen, Shuqi Xu, Yue Ma 等 72 位作者

AI 速读 基于摘要

作者提出包含56个由物理学家基于自身研究设计的端到端物理研究基准，用确定性验证器与专家评分量表同时考察智能体研究的正确性与可发表性。

cs.AI

arXiv ↗PDF ↗
查看原文摘要

AI agents are playing an increasingly important role in physics research. However, existing physics benchmarks only evaluate fixed questions, isolated sub-tasks, or paper reproduction with given solutions, and rarely check research quality alongside correctness. We introduce FrontierPhysics, an end-to-end benchmark with 56 tasks made by 42 physicists based on their own research, covering nine domains across theoretical, computational, and experimental physics. Each task requires agents to review the literature, develop a research plan, conduct research and experiments, and produce structured outputs and scientific manuscripts. Combining deterministic verifiers with expert-authored rubrics, FrontierPhysics evaluates whether agents' research is correct and publishable. Across 22 models in 26 configurations, no configuration passes more than 46 of its 168 rollouts. Manual review of agent trajectories shows that frontier agents' main weaknesses are insufficiently rigorous reasoning and limited expert intuition, rather than gaps in physics knowledge or execution errors.

新论文 · v1arXiv:2611.06844

在哪里连接？通过跨所有 ViT 与 LLM 层的动态门控通路提升 MLLM

Where to Connect? Boosting MLLMs via Dynamic Gated Pathways across ALL ViT and LLM Layers

Yingying Yan, Jiaqi Tang, Wei Wei, Qianzhou Wang, Lei Zhang, Jianmin Chen, Yuyang Xia, Botong Geng 等 10 位作者

AI 速读 基于摘要

提出 DGP 动态门控通路，在每一层 ViT 与每一层 LLM 之间建立全连接，并让每个 LLM 层根据自身语义状态动态选择视觉粒度。

cs.CV

arXiv ↗PDF ↗
查看原文摘要

A central yet under-explored question in Multimodal Large Language Models (MLLMs) is where to connect---how visual information from a hierarchical vision encoder should be wired into the layer-wise semantics of a large language model. Mainstream MLLMs inject a fixed single-layer visual representation into the LLM and treat all LLM layers as a single consumer, ignoring that different layers demand different visual granularities. Recent multi-layer fusion methods either compress hierarchical features on the encoder side or rely on predefined sparse or hierarchical layer-to-layer connections, and therefore stop short of modeling adaptive cross-layer interactions between the ViT and LLM hierarchies. We address this gap with DGP (Dynamic Gated Pathways), a routing mechanism that establishes all-to-all connections between every ViT layer and every LLM layer, with each LLM layer dynamically gating its preference over visual layers conditioned on its own semantic state. Built on top of LLaVA-1.5-7B and 13B, DGP delivers consistent gains over the LLaVA-1.5-7B baseline (e.g., +4.2 on SQA$^{I}$, +3.3 on MM-Vet, +3.1 on VizWiz and LLaVA$^{W}$) and surpasses prior multi-layer visual fusion methods on most benchmarks. Layer-wise routing analyses further show that different LLM layers do select different visual granularities on demand, supporting both the effectiveness and interpretability of the proposed dynamic gated pathways. The code is available: https://github.com/YingyingYan1/DGP.

新论文 · v1arXiv:2611.06319

Verified Implementation of the Next-gen Borrow Checking Algorithm for Rust via Abstract Interpretation

Jinhua Wu, Yuting Wang, Zhemin Chen

原文摘要

Borrow checking is a central mechanism for Rust to ensure temporal memory safety. Its correctness is based on the intuition that the Aliasing XOR Mutability discipline guarantees that programs without borrow errors inherently could not access invalid references (i.e., pointers). However, existing work on formalizing borrow checking treats it either as a component of type systems or as symbolic execution describing resource transfer. None provides a direct explanation of the correctness of borrow checking algorithms such as Non-Lexical Lifetimes (NLL) or Polonius in the official Rust compiler that function more like static pointer analysis. We take a first step toward solving this problem by formalizing Polonius, the next-generation borrow- checking algorithm for the Rust compiler, on a core subset of Rust. We prove its correctness by showing that any program accepted by Polonius does not access invalid references at runtime. The key observation is that, when borrow checking is performed lazily, Polonius could be treated as a forward abstract interpretation on abstract domains recording pointer aliasing (borrowing) relations. By providing a concrete semantics that formally describes the Aliasing XOR Mutability discipline, we prove the soundness of this abstract interpretation, from which we derive memory safety of checked programs. This formulation of Polonius is implemented and formally verified in the abstract interpretation framework of CompCert. The correctness proofs are generated by an agentic AI under human guidance using paper proofs as proof outline. To the best of our knowledge, this is the first comprehensive formalization and verification of a borrow checking algorithm as manifested in the Rust compiler.

cs.PL

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.05088

LatentTriad: Pre-Action Risk Representations for Malicious Agent Skill Detection

Zhe Yu, Yi Liu, Ying Zhang, Simin Chen

原文摘要

Large language model agents are increasingly extended with reusable skills, but untrusted skill content can induce operations outside the user's task. Detecting this contextual risk requires more than suspicious-pattern matching, while dynamic analysis incurs the cost of skill execution. We find that hidden states formed during prefill contain readable information about task-operation consistency, including across held-out expression forms in controlled evaluations. Matched Operation-state interventions also affect later Boundary states and shift fixed risk readouts toward the donor label. Based on these findings, we introduce LatentTriad, which combines contextual Boundary features, group-adjusted Operation directions, and departures from benign state relationships. All branches share one prefill: detection generates no output tokens, invokes no tools, and executes no skill code. Across 5,830 skill packages from five sources and seven backbones, LatentTriad achieves 99.0-99.2% pooled F1 and 93.0-94.6% source-macro F1. A separately measured compact implementation completes prefill and detection in a median of 259-284 ms. These results connect pre-action risk readability and functional influence to execution-free skill screening.

cs.AI

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2611.03260

The Cost of Mismatch: Noise Amplification in Zeroth-Order Reinforcement Learning

Lianmin Chen, Junbin Qiu, Chenxing Wei, Yao Shu, Kun He

原文摘要

Zeroth-order (ZO) optimization has emerged as a promising approach for memory-efficient large language model (LLM) post-training, yet its reliability in reinforcement learning with verifiable rewards (RLVR) remains largely unexplored. Motivated by the distinct data regimes of on-policy RL and fixed-data optimization, we develop a generic variance-theoretic framework to investigate what limits the reliability of ZO gradient estimation. Our analysis reveals that estimation reliability is structurally determined by whether the two finite-difference evaluations are performed on matched data. On-policy reinforcement learning (RL) under the mismatched-data regime produces a $1/μ^2$ oracle-noise amplification that sets an irreducible noise floor as the perturbation scale $μ$ decreases, whereas Boltzmann Targeted SFT (BOLT) under the matched-data regime eliminates this amplification via noise cancellation. We further derive convergence guarantees with explicit constants that cleanly separate the two regimes. These results motivate the ZO-BOLT algorithm, which builds on BOLT's matched-data advantage for reliable ZO optimization of an RLVR-aligned objective. Controlled perturbation-scale sweeps and matched-versus-mismatched end-to-end comparisons validate the predicted effect of data matching. Experiments across three models and two mathematical reasoning benchmarks further show that ZO-BOLT consistently improves over the pretrained baseline while retaining a near-inference memory footprint. The code is publicly available at https://github.com/lianmin-chen/the_cost_of_mismatch.

cs.LG

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.04356

CloakFEM: Agentic FEM Code Generation for Red-Teaming Computational Resource Exhaustion

Li Zheng, Simin Chen

原文摘要

Finite element method (FEM) solvers are foundational to scientific and engineering simulation, yet their susceptibility to \emph{efficiency attacks} remains largely unexplored. An adversary can craft syntactically valid and runnable FEM programs that complete successfully while forcing excessive computational work, enabling computational denial-of-service (DoS) attacks. Finding such inputs is challenging because FEM programs occupy a vast, structured space constrained by numerical, geometric, and constitutive requirements. We introduce CloakFEM, a domain-guided LLM-agent framework for automatically generating runnable adversarial FEM programs that amplify solver computation. CloakFEM combines LLM-guided search with an FEM-specific intermediate representation and a deterministic validity gate that enforces runnable constraints. Across three FEM stacks, 17 problem classes, and 1,068 seeds, CloakFEM discovers adversarial inputs that substantially amplify solver work while remaining valid and solvable. In the most extreme case, solver iterations increase from 43 to 247,367, a \(5{,}752.72\times\) amplification. CloakFEM outperforms five generic black-box baselines and retains an advantage over evolutionary search under a controlled bounded action space. Cross-solver analysis further reveals both problem-intrinsic and solver-specific sources of vulnerability. Our results uncover an underexplored efficiency-attack surface in FEM solvers and show that runnable FEM programs can serve as adversarial inputs.

cs.CR · cs.AI · cs.SE

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.04003

WASP-Mem: Let Agent Organize Procedural Memory. Human Spend Effort on Diagnosis

Peizhi Niu, Ge Shi, Vaibhav Gorde, Sarthak Ghosh, Christos Faloutsos, Min Cheng, Kevin Chen

原文摘要

LLM agents struggle to automate complex enterprise procedures because static, pre-built harness cannot capture proprietary policies, undocumented behaviors, or state-dependent recovery paths that surface only during execution. Procedural memory accrued from agent experience offers a training-free path to closing this gap -- yet unlike episodic records or semantic facts, it must consolidate repeated successes and diagnosed failures into reusable actions and recovery strategies. We introduce WASP-Mem (Wiki-based Agent-Synthesized Procedural Memory), a nimble but effective method that distills execution traces and feedback into persistent, interlinked wiki pages, enabling continual agent learning. On TheAgentCompany benchmark, WASP-Mem raises the checkpoint pass rate by 4.6-6.1% over the strongest baseline we compared across three models while requiring 57-61% less storage. We further present a systematic controlled study spanning consolidation method, feedback type, trace ingestion scope, and refinement rounds. Our results reveal that prescriptive knowledge organization such as human defined schemas and explicit deduplication rules hurts performance rather than closing the procedural-knowledge gap. In contrast, human diagnosis converts failed traces into corrective knowledge and drives sustained improvement: a single round of human diagnosis outperforms four rounds of self-judgment by 7.0-10.9%. These findings point to a simple division of labor: let agents organize procedural memory, and invest human effort where it matters most -- diagnosis.

cs.AI · cs.CL · cs.LG

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.03982

SLATE: A Large-Scale Benchmark for Long-Horizon API Selection in E-commerce Tool Use

Rongzhe Wei, Ge Shi, Min Cheng, Sarthak Ghosh, Vaibhav Gorde, Leman Akoglu, Pan Li, Na Zhang

原文摘要

Enterprises increasingly want LLM agents to automate standard operating procedures over their software stacks -- mapping each step of a procedure to the right call among thousands of internal APIs, without hand-authored step-to-API mappings. The core difficulty is API selection at scale: choosing and composing the correct tools over a long horizon, judged only by whether the overall task succeeds. Current tool-use benchmarks do not test this regime -- they expose only a handful of tools, evaluate single-step invocation, or grade multi-step behavior with subjective LLM-as-a-judge scores that penalize valid alternative trajectories. We introduce SLATE (Synthetic Large-scale API Toolkit for E-commerce), a large-scale benchmark that tests large-toolset, long-horizon API selection in a realistic enterprise domain. SLATE couples (i) queries decomposed into hierarchical, tool-grounded procedures over a ~1,000-tool e-commerce library, (ii) a deterministic, context-aware simulator that returns recorded outputs for matching calls and default responses otherwise, and (iii) an end-to-end task-success metric -- a task succeeds when it reaches every reference outcome -- that credits diverse-but-valid trajectories while remaining fully automatic and reproducible. Experiments with six open decision models and the language models they are built on, from 3B to 12B parameters, show that even the best model completes only 5.0% of the tasks, establishing SLATE as a challenging testbed for reliable enterprise tool use. We release SLATE, its simulator, and the evaluation harness at https://geshijoker.github.io/slate/

cs.CL · cs.AI

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.03223

Do Agents Still Follow the Rules? Benchmarking Rule Compliance under Agent Context Compaction

Yiwei Zha, Qingyu Meng, Herbert Bos, Min Chen

原文摘要

LLM agents compact long histories to fit their context windows, but the compacted context must still carry the operating rules that govern later actions. Prior work shows that violations rise after compaction, but measures compliance outside a complete agent harness and does not identify where each violation comes from. We introduce ARCBench(Agent Rule Compliance Benchmark), a configurable benchmark built on the open-source pi agent harness. Its tool-calling agent loop lets us measure rule compliance from actions taken during execution, bringing the evaluation closer to deployed agent workflows. Using ARCBench, we examine where violations after compaction come from, and how different compaction configurations and adversarial settings change their sources. We find that compaction rarely deletes the rule: most violations after compaction occur while the rule is still in the compacted history, with the agent treating a later request as permission, ignoring the rule, and proceeding anyway. Where violations come from depends on the configuration: how the rule is delivered decides whether the agent would break it even without compaction, while the compaction strategy and budget decide whether the rule survives. The strongest compactor-targeted attack more than doubles violations by corrupting the rule in the summary and changing how the agents behave, whereas the strongest defense restores the rule in most summary but still violates. Finally, retaining a rule does not guarantee compliance: compacted runs that retain the rule violate more than twice as often as runs with the full history.

cs.CR · cs.AI · cs.CL

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.02765

Recast: Refinement-Centered Approximate Nearest Neighbor Search Optimized for Storage

Zehai Yang, Fengyi Tao, Wenzhuo Dong, Shimin Chen

原文摘要

Large-scale ANNS systems often exploit storage in addition to DRAM for supporting vector search. The central question is how to divide the work between memory and storage for achieving both high query throughput-recall and good index maintenance performance. Existing storage-optimized ANNS paradigms fall short of providing a complete answer: disk-resident graph indexes have difficulty in supporting efficient vector deletions, while posting-list systems often incur more disk I/Os for queries. In this paper, we introduce a refinement-centered paradigm for storage-optimized ANNS systems. We propose Recast as a concrete design of this paradigm. Recast stores raw vectors on disk and keeps an IVF index with our RecastQ quantization encoding in memory. For an ANNS query, Recast selects nprobe inverted lists and performs SIMD-based FastScan on their RecastQ codes to identify refinement candidates in memory. It then retrieves candidate vectors from disk, computes exact distances, and refines the top-K result. To reduce disk I/Os, RecastQ computes uncertainty intervals containing true distances with high probability. Our QUIC policy selects a minimal refinement set based on these intervals, minimizing candidate-retrieval I/Os. Both the on-disk raw vectors and the in-memory IVF structure support insertions and deletions. Experiments on real-world data sets show that Recast processes batched insertions and deletions 3.10-37.27x faster than SPFresh and OdinANN, while its QPS at 0.99 recall changes by at most 6.9% after updating 50% of the corpus. For static queries at 0.99 recall, Recast achieves 1.56-2.01x the QPS of the strongest baseline with the matched memory budget on each data set, while requiring 1.22-6.75x less external storage than the evaluated systems. At matched in-memory space, RecastQ also provides a better Recall-refinement frontier than RaBitQ.

cs.DB

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2509.25894

Red Teaming Program Repair Agents: When Correct Patches can Hide Vulnerabilities

Simin Chen, Yixin He, Suman Jana, Baishakhi Ray

原文摘要

LLM-based agents are increasingly deployed for software maintenance tasks such as automated program repair (APR). APR agents automatically fetch GitHub issues and use backend LLMs to generate patches that fix the reported bugs. However, existing work primarily focuses on the functional correctness of APR-generated patches, whether they pass hidden or regression tests, while largely ignoring potential security risks. Given the openness of platforms like GitHub, where any user can raise issues and participate in discussions, an important question arises: Can an adversarial user submit a valid issue on GitHub that misleads an LLM-based agent into generating a functionally correct but vulnerable patch? To answer this question, we propose SWExploit, which generates adversarial issue statements designed to make APR agents produce patches that are functionally correct yet vulnerable. SWExploit operates in three main steps: (1) program analysis to identify potential injection points for vulnerable payloads; (2) adversarial issue generation to provide misleading reproduction and error information while preserving the original issue semantics; and (3) iterative refinement of the adversarial issue statements based on the outputs of the APR agents. Empirical evaluation on three agent pipelines and five backend LLMs shows that SWExploit can produce patches that are both functionally correct and vulnerable (the attack success rate on the correct patch could reach 0.91, whereas the baseline ASRs are all below 0.20). Based on our evaluation, we are the first to challenge the traditional assumption that a patch passing all tests is inherently reliable and secure, highlighting critical limitations in the current evaluation paradigm for APR agents.

cs.SE

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2610.39782

Assessing AI Companion Chatbot Safety for Vulnerable Users: A Crisis-Stage Multi-Platform Evaluation

Huanyue Liao, Seok Min Lim, Yan Ling Lim, Shi Min Cheng, Kellin Pelrine, Chen Hui Ong

原文摘要

AI companion chatbots are increasingly used by vulnerable individuals, including adolescents in suicidal crisis, for emotional support. Existing safety evaluations focus on base models, leaving gaps in product-level assessment, crisis-stage sensitivity, and comparisons between companion apps and general-purpose LLMs. We introduce MIRROR-Eval (Multi-turn Interaction Risk Review: Observing suicidal-crisis Responses), a dataset of 960 annotated 20-turn conversations across eight deployed platform configurations, six clinically grounded personas, and two crisis stages: ideation and explicit planning. An LLM-as-a-Judge pipeline applies an expert-reviewed rubric covering user risk severity, chatbot risk awareness, and intervention behavior. We observe a three-tier harm structure: companion apps (Character.AI and Replika) and Grok 4.3 form a high-harm tier; Gemini 3.1 Pro is mid-range; GPT-5.5 and Claude Sonnet 5 show substantially lower harm. At planning, companion apps show increased glorification and normalisation, advancement toward harmful action, and harmful methods and instructions. Crisis recognition and empathetic language do not ensure safe responses. These findings highlight the need for product-level safety evaluation across crisis stages. Dataset and code: https://huggingface.co/datasets/imda-biztech/MIRROR-Eval. Warning: this paper includes examples that may be offensive or harmful.

cs.HC

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2610.36831

Right Click, Wrong Box: Grading GUI Grounding by What a Click Does

Xiaokang Yuan, Binhui Yao, Luwei Yang, Ruimin Chen, Zeng Wang

原文摘要

Leading GUI grounders score 93 to 96% on ScreenSpot-v2, and reported gaps between new and prior models are about a point. At this level the useful question is what the remaining errors are. Benchmarks count a click as correct only inside one annotated box, but interfaces often expose one action through several elements: a duplicated link, the label of a checkbox, the title next to a logo. The box is not wrong. The rule is incomplete. We call a click correct when executing it produces the same outcome as clicking the annotated target. A probe that executes and fingerprints about a thousand clicks per page turns this definition into outcome maps and an execution judge for live web applications, with no new annotation. We execute the predictions of three strong 7--8B grounders on a code forge and an online shop. About one in five and one in three of the clicks that box scoring rejects reach the intended outcome, and the shares remain one in seven and one in three when the instructions are rewritten as natural tasks. A blind audit of 11 models on ScreenSpot-v2 (three 7--8B grounders and eight 3B variants) finds the same effect on web screenshots, with two independent reviewers on a 60-unit subset. The share is highest for the strongest grounders. Training seeds and test-set sampling add noise that more runs and paired intervals control. Box scoring instead under-counts correct clicks, which repetition does not remove, and the audit-based correction of 0.2 to 1.8 points on ScreenSpot-v2 is as large as the gaps the field reports. The correction is needed in evaluation. In our 3B GRPO study, crediting equivalent clicks in the reward gives no stable gain over box rewards across three seeds and four test sets. We release the probe, the outcome maps, the audit and every prediction.

cs.CV

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2610.36623

God Does Not Play Dice: Enforcing Batch Invariance in Agentic Kernel Generation

Youwei Bao, Chengpeng Wang, Simin Chen

原文摘要

Large language model (LLM) serving systems dynamically batch user requests for GPU execution, so the same request may be processed alongside different requests across runs. If the GPU kernels lack batch invariance (BI), this changing batch context can alter the numerical output for an unchanged input, undermining reproducibility. Existing kernel generation agents focus on correctness and efficiency but \emph{overlook BI}. We find 19 confirmed BI violations among 83 batch-applicable released artifacts. To address this limitation, we propose \tool, which combines contract-guided static dependency analysis with dynamic validation of bitwise consistency across batch contexts. \tool is designed as a plug-in component for general kernel agents, allowing their original optimization capabilities to be reused while adding BI enforcement. We integrate it as successive acceptance gates in the optimization loop, rejecting structurally risky or dynamically non-invariant candidates before performance benchmarking and returning feedback to the agent. Experiments across three agent frameworks under baseline and BI-aware generation on real world LLM serving tasks show that \tool improves the BI pass rate from 67.6\% to 91.9\% while retaining 91.8\% of the baseline geometric mean speedup. These results demonstrate that \tool can be incorporated into kernel optimization by agents as an additional reliability objective alongside correctness and speed.

cs.AI · cs.LG · cs.SE

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2610.36499

ProCerDCL: Query-Efficient Black-Box Probabilistic Verification of Deep Learning Compiler Optimization

Zexin Li, Weibo Zhang, Linyi Li, Arvind Easwaran, Simin Chen

原文摘要

We present \tool, a query-efficient framework for black-box probabilistic verification. Given a fixed, sampleable input distribution and a binary property oracle, \tool assesses the property's violation probability $p$ against a threshold $θ$. Under independent sampling and stateless oracle behavior, acceptance asserts $p\leθ$, rejection asserts $p>θ+η$ for a user-specified separation margin, and an inconclusive verdict indicates insufficient evidence for either conclusion. The probability of each type of erroneous verdict is bounded by its specified error budget. To support this three-valued contract efficiently, \tool uses a single sequential test, eliminating the interval family required by the baseline. The test accounts for repeated sampling checks through either corrected exact binomial confidence intervals or an anytime-valid confidence sequence. We apply \tool to probabilistic local model semantic consistency, where the property oracle detects decision-level inconsistencies between reference and compiled executions under a specified perturbation distribution. Our evaluation covers 24 tasks across TorchInductor and TensorRT and examines sampling efficiency, conclusiveness, and diagnostic behavior. For sampling efficiency, we use paired one-model clean-label radius searches: the baseline requires 3.6-4.2 times as many unique samples under matched numerical parameters while providing a weaker decision contract. For model semantic consistency, checks at $θ=10^{-3}$ yield 817 acceptances, 19 inconclusive verdicts, and 4 rejections across 840 input--backend instances from 21 tasks. Case studies of two documented compiler faults further demonstrate what these local verdicts can diagnose.

cs.SE · cs.AI · cs.PL

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2610.24816

First-Order LoRA Training with Phone NPUs

Xu Zheng, Wei Cheng, Zhuomin Chen, Dongkuan Xu, Mo Sha, Jingchao Ni, Dongsheng Luo

原文摘要

On-device fine-tuning lets a large language model (LLM) learn from its user's private data without uploading it. Low-rank adaptation (LoRA) reduces the trainable parameters, but each first-order training step still multiplies by the frozen base-model weights during the forward pass, activation recomputation, and input-gradient computation. A phone's neural processing unit (NPU) already holds these weights for inference; using it for training also requires computing input gradients, which its inference interface does not support. We present a system that runs all three frozen-weight products on the NPU while keeping the training graph, adapters, and optimizer on the CPU. Its input-gradient kernels reuse the resident weight tiles and the forward dequantizer, avoiding a transposed weight copy and ensuring forward and backward use the same effective weights. To preserve usable gradients, the system selects precision by gradient sensitivity and applies a per-row range guard for half-precision (FP16) outputs. Its runtime fuses the remaining CPU operators and overlaps NPU execution with adapter computation, while resumable execution and thermal pacing let training proceed as a background job. We evaluate it on two Qualcomm Snapdragon platforms using four LLMs from the Qwen and Llama families. Our workloads cover nightly personalization with labels supplied on the phone and the public LaMP-2 benchmark; comparisons include an optimized CPU trainer, a phone-GPU trainer, forward-only training, and GPU-trained references.

cs.LG · cs.AI

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2610.36113

Scaling and Renormalization in Neural Networks: An Effective Renormalization-Group Description of Grokking

Min Chen, Junqi Wang, Zeguan Wu, Wenhao Huang, Junyu Liu

原文摘要

Grokking is a delayed transition from memorization to generalization in deep neural networks (NNs). We examine whether scaling and renormalization group (RG) theories can describe this dynamics, establishing a quantitative physical framework for neural generalization transitions. For modular addition on a two-layer NN, an ensemble of trained models defines an order parameter (the squared weight norm) and fluctuation proxies for the heat capacity and the susceptibility, with training epoch and weight decay serving as thermal and field control coordinates. To bridge neural architectures to real-space RG, we construct an exact coupling matrix via a second-order expansion of the loss with respect to per-neuron norms and use its Fiedler vector to define an effective geometry. This spectral ordering yields a monotonic spatial decay of coupling strength that is absent under random permutations. We then execute real-space RG by block-spin aggregating adjacent neurons and parameterizing the coupling matrix into off-diagonal band strengths. Linearizing this local RG map yields relevant scaling exponents and five derived critical exponents, with direct empirical estimates matching RG predictions within 1-10%. Overall, our framework provides a consistent effective description of grokking, suggesting the scaling and renormalization group formalism as a tool to investigate NN dynamics.

cs.LG · cs.AI · hep-th · quant-ph

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2601.06640

Agentic AI Empowered Intent-Based Networking for 6G

Genze Jiang, Kezhi Wang, Xiaomin Chen, Yizhou Huang

原文摘要

The transition towards sixth-generation (6G) wireless networks necessitates autonomous orchestration mechanisms capable of translating high-level operational intents into executable network configurations. Existing approaches to Intent-Based Networking (IBN) rely upon either rule-based systems that struggle with linguistic variation or end-to-end neural models that lack interpretability and fail to enforce operational constraints. This paper presents a hierarchical multi-agent framework where Large Language Model (LLM) based agents autonomously decompose natural language intents, consult domain-specific specialists, and synthesise technically feasible network slice configurations through iterative reasoning-acting (ReAct) cycles. The proposed architecture employs an orchestrator agent coordinating two specialist agents, namely Radio Access Network (RAN) and Core Network agents, via ReAct-style reasoning, grounded in structured network state representations. Experimental evaluation across diverse benchmark scenarios demonstrates that the proposed system outperforms monolithic agents, rule-based systems, and direct LLM prompting, with architectural principles applicable to Open RAN (O-RAN) deployments. The results also demonstrate that whilst contemporary LLMs possess substantial general telecommunications knowledge, structured prompt engineering encoding domain-specific decision thresholds and operational policies is essential for reliable network automation performance.

cs.AI · cs.NI

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2610.35051

VISTA: Active Exploration for Self-Improvement in Embodied Agents

Zhang Zhang, Shengjun Zhang, Simin Huang, Min Chen, Chong Xia, Xin Fei, Yueqi Duan

原文摘要

Agentic embodied intelligence plays a foundational role in open-ended, dynamic physical environments. Rather than treating policy learning as an end-to-end mapping from instructions to actions, recent studies reveal that embodied systems depend not only on pre-trained foundation models, but also on harness. Prior approaches largely treat rollouts as static logs for post-hoc reflection, producing harness updates without autonomously testing and refining them through subsequent embodied interaction. To address this challenge, we propose VISTA, a meta-agent framework that transforms post-hoc rollout reflection into active self-exploration for embodied harness improvement. Specifically, VISTA maintains a harness graph that records the lineage of proposed harnesses, which enables the meta-agent to retrieve prior hypotheses, build on effective mechanisms, and avoid repeatedly exploring failed directions. Each iteration begins from a retained harness version and its execution evidence. Rather than directly committing an update from these historical trajectories, the meta-agent actively explores in the environment. It autonomously generates new rollouts, probes execution failures and recovery behaviors, and iteratively modifies the harness based on the resulting evidence. The refined harness is then evaluated afresh and added to the harness graph. We evaluate VISTA on the LIBERO-Pro, RoboCasa, and RoboTwin benchmarks. Experimental results demonstrate that VISTA progressively improves task success and is more rollout-efficient than static reflection, achieving higher success rates with fewer simulator rollouts.

cs.RO

查看论文
arXiv ↗PDF ↗
首次观测 · 记录 #1ACL Anthology · ACL 2026:acl:2026.acl-industry.33ACL Anthology · ACL 2026

Learning from Textual Radiology Reports: A Benchmark Dataset for Coronary CT Angiography

Sudharshan Balaji, Zhiyu Liu, Zhengyuan Jiang, Shuo Lei, Yimin Chen, Yang Xiao, Shone O. Almeida, Mathew Joseph Karivelil 等 10 位作者

原文摘要

While coronary imaging is widely used for anatomical assessment, CCTA reports play a distinct last-mile role in clinical care. Ratherthan serving as an intermediate signal, CCTA provides an assessment of coronary disease severity (known as the CAD-RADS score) toguide patient management. However, real-world clinical text exhibits substantial heterogeneity in terminology and structure, leadingto inconsistent interpretation by automated systems, even for clinically similar cases. Recent work leverages a direct application ofLLMs for automated CAD-RADS scoring, but is limited by small, non-public, and homogeneous clinical data. We introduce CCTA-RADS, the largest publicly available dataset of 940 real-world CCTA reports from a major cardiovascular center, each annotated with CAD-RADS scores. Our analysis reveals that direct approaches, including state-of-the-art LLMs (GPT-4o, GPT-o3) and fine-tuned BERT models underperform on diverse real-world clinical data. To address these limitations, we propose a two-stage pipeline that decouples structuring from classification: an LLM-based parser normalizes heterogeneous reports into structured format, followed by fine-tuned BERT classification. This approach substantially improves the F1-score by 6%-13% compared with direct methods. We deploy our system as an interactive web interface that allows clinicians to upload CCTA reports for automated CAD-RADS assessment with SHAP and LIME explainability visualizations.

cs.CL

Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 6: Industry Track) · 2026 · 正式发表

首次观测 · 记录 #1ACL Anthology · ACL 2026:acl:2026.acl-long.1384ACL Anthology · ACL 2026

Spectral Characterization and Mitigation of Sequential Knowledge Editing Collapse

Chi Zhang, Mengqi Zhang Shandong University, Xiaotian Ye, Runxi Cheng, Zisheng Zhou, Ying Zhou Shandong University, Pengjie Ren Shandong University, Zhumin Chen Shandong University

原文摘要

Sequential knowledge editing in large language models often causes catastrophic collapse of the model’s general abilities, especially for parameter-modifying methods. Existing approaches mitigate this issue through heuristic constraints on parameter updates, the mechanisms underlying such degradation remain insufficiently understood. In this work, we present a systematic spectral analysis of sequential knowledge editing and show that a model’s general abilities are closely associated with dominant singular directions of pretrained weight matrices. These directions are highly sensitive to perturbations and are progressively disrupted by repeated edits, closely tracking the collapse in both editing efficacy and general performance. Building on this insight, we propose REVIVE, a plug-and-play framework that prevents model collapse by explicitly preserving this dominant subspace. REVIVE analyzes parameter updates in the spectral basis of the original weights and filters out components that would interfere with the dominant subspace. Extensive experiments across multiple models and benchmarks show that REVIVE consistently improves editing efficacy while substantially preserving general abilities under long-horizon sequential editing, including extreme settings with up to 20,000 edits.

cs.CL

Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · 2026 · 正式发表

首次观测 · 记录 #1ACL Anthology · ACL 2026:acl:2026.acl-long.1754ACL Anthology · ACL 2026

Revealing Procedural Reasoning Structures in Chain-of-Thought Training via Span-Level Gradient Organization

Jia Liu South China University of Technology, Pazhou Laboratory and Huazhong University of Science and Technology, Jiaxin Luo, Weiwen Xu Alibaba Group, Jonathan M. Garibaldi The University of Nottingham Ningbo China and University of Nottingham, Xiao-Kun Wu Renmin University of China, Yixue Hao, Min Chen South China University of Technology and Pazhou Laboratory

原文摘要

Chain-of-Thought (CoT) prompting enables large language models to produce multi-step reasoning, yet how such reasoning-related structure is expressed during training remains poorly understood. We present Gradient-based Structural Developer (GSD), an unsupervised framework with a principled gradient aggregation view that tracks span-level gradient during fine-tuning on reasoning benchmarks to understand how models develop structured, step-by-step reasoning capabilities. Our analysis shows that while gradients at the level of individual tokens are often noisy, aggregating gradients over contiguous reasoning-related spans reveals stable and recurring directional alignment across samples. We refer to these directionally aligned patterns as aligned sequential stresses, reflecting consistent gradient organization associated with similar reasoning procedures. Beyond capturing semantically similar reasoning instances, such gradient alignment also reveals structurally similar but semantically diverse cases that share common procedural organization. These findings position GSD as a diagnostic framework for analyzing how procedural reasoning structures emerge during training, with downstream selection results serving as auxiliary evidence correlating gradient alignment with adaptation efficiency.

cs.CL

Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · 2026 · 正式发表