跳到正文
研踪PaperTrail

YOUR RESEARCH, IN SIGHT

让下一篇重要论文，不再错过。

覆盖计算机全学科，以 AI 为快捷入口。关注方向、作者和关键词，把发现变成持续的研究积累。

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
+1已有论文修订
+100新增 AI 导读

arXiv 最近完成扫描：2026-11-07 23:54 · 全站最近入库：2026-11-07 17:43各信源进度 →

统计口径与数据新鲜度

统计 2026-11-07 00:27 — 2026-11-08 00:27（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：2026-11-07 23:54最近发现新增版本：2026-11-07 17:43库内论文最新更新时间：2026-11-06 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。

17,100 已收录论文14,661 AI 相关4,056 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 11 篇论文 · 按版本更新时间排序上轮扫描起点 2026-11-07 23:49（北京时间）
新论文 · v1arXiv:2611.06840

相同布局，不同数量：文生图模型中的数字词在超过十之后起什么作用

Same Layout, Different Count: What Number Words Do Past Ten in Text-to-Image Models

Liang Yue, Eric A. Vance

AI 速读 基于摘要

研究五个开源文生图模型在请求超过约十个物体后，数字词对生成数量的影响及其与初始噪声的关系。

cs.CV · cs.LG

arXiv ↗PDF ↗
查看原文摘要

Ask SD3.5-Medium for twelve of an object and then, from the same starting noise, for twenty-five. The picture keeps its layout, but the count changes about as much as it would under a new random seed, and in no consistent direction. We study what number words do once a request passes about ten objects, in five open text-to-image models. Holding the object and the starting noise fixed, we change one thing at a time, and people who never see the request count 3,088 of the images. Every model follows the number for small crowds, but only up to a point, its reach: SD3.5-M and FLUX.1-dev stop following it soon after ten, while FLUX.2 [klein] and Qwen-Image follow it loosely up to twenty-five. Beyond SD3.5-M's reach, the coarse part of the starting noise fixes the composition of the picture but not how many objects it holds: redrawing only the noise finer than an eighth of the image keeps the layout as well as a new number does, and already moves the count two-thirds as far. The number word still reaches the denoiser and acts in its first steps, but nothing we tried gives it a clear direction. Shown instead as a noised grid in the starting latent, the count is drawn: people find it in 59% of images, against 4% when it is only told. Past ten, an exact-match score barely separates a model that follows the number loosely from one that ignores it, and a relative-error score can even prefer a constant guess of sixteen. A same-noise test with a few requests per object tells them apart.

新论文 · v1arXiv:2611.06168

Reasoning-Aware Cross-Modal Association Unlearning in Multimodal Large Reasoning Models

Xinghao Li, Chenchen Tan, Youyang Qu, Shujie Cui, Xingliang Yuan, Cunjian Chen, Chi Liu, Longxiang Gao

原文摘要

Multimodal large reasoning models (MLRMs) integrate visual inputs with textual and parametric knowledge, allowing sensitive information to be accessed through cross-modal associations during reasoning. This raises privacy concerns when a visual input should no longer provide access to such information. Machine unlearning can remove this influence without retraining the model from scratch. However, multimodal unlearning methods mainly target samples, concepts, or knowledge content, without separating the knowledge from the cross-modal association through which it is accessed. In reasoning models, target information may disappear from the final answer while remaining in intermediate reasoning, or be replaced by incorrect substitute knowledge. We therefore study cross-modal association unlearning, where the forget target is a designated visual-to-knowledge association: the model should suppress target-bearing behavior under the designated visual context while preserving access to the same knowledge in permitted contexts. We propose Visual-Keyed Counterfactual Rebinding (VKCR), a parameter-efficient framework for cross-modal association editing. VKCR first localizes the edit to a compact association-conditioned parameter subspace. It then reduces support for the designated target at task outputs and counterfactually rebinds the affected reasoning behavior by discouraging target-bearing or unsupported continuations while favoring valid non-sensitive alternatives. We further introduce CLEAR-R and MLLMU-R to evaluate target leakage, hallucinated reassociation, reasoning preservation, and modality-specific utility. Experiments show that VKCR substantially reduces image-conditioned target leakage and hallucinated reassociation, including across heterogeneous task formulations, while preserving non-sensitive visual reasoning and permitted text-only knowledge.

cs.CV · cs.AI

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.06013

When Does a Cross-Device Run Justify PASS? Obligation-Level Evidence Qualification under Partial Observability

Zimu Liu, Xiaoxi Guo, Xiaolu Tang, Liang Yu

原文摘要

A cross-device test can complete every command yet judge a business claim from evidence belonging to another device or attempt. We introduce an executable evidence contract that makes support qualification an explicit stage before predicate evaluation. It binds each obligation to admissible subjects, episodes, objects and observation scopes, and records unresolved support rather than treating an apparently successful observation as proof. We evaluate two implemented parts of the contract. A prospectively specified, 12-seed controlled study isolates association under reused identifiers across 2,304 claims. Given identical records and business policy, Atomic-Facts Flat produces 160 false PASS and 76.30% PASS recall; supplying computed associations reduces false PASS to five and increases PASS recall to 99.09%. Deterministic Closure produces zero false PASS and 100.00% PASS recall, improving on Atomic Flat in 11 seeds with one tie (two-sided sign p=0.00098). Most of the benefit comes from explicit association rather than graph notation. A separate prospective field characterization spans four workflows, 40 physical episodes and 80 claims. Against frozen reference labels, the physical-evidence verifier attains 95.00% accuracy and 100.00% PASS recall (27/27), with three false PASS judgments. The field results are descriptive: a policy-input mismatch prevents the planned matched-policy comparison, and post-hoc sensitivity does not establish superiority over a policy-informed textual baseline. Together, the studies characterize qualification on physical evidence and isolate the value of explicit association under a controlled matching contract.

cs.SE

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.05873

Can LLM Agents Learn to Test? Maintaining Scoped Experience as Testing Evidence for Mobile GUI Testing

Zimu Liu, Xiaoxi Guo, Xiaolu Tang, Liang Yu

原文摘要

Repeated mobile GUI tests can reuse routes, locators, and recovery strategies from earlier executions. However, a route that worked in one workflow or app version may be irrelevant or misleading in another. We study how an LLM-based tester can retain useful experience while withholding entries contradicted by current observations. We present a maintenance contract for AdaptiTest in which reusable experience has an explicit applicability scope, trace provenance, and a lifecycle for confirmation, revalidation, and withdrawal. The contract reserves persistence for a Reflector role and treats recalled experience as advice. The archived implementation does not establish end-to-end enforcement of every specified mechanism. The design changes retrieved context, not model weights. The evaluation uses read-only Android workflows and archived internal comparisons. A six-task, three-repeat ready subset reports 18/18 completed runs with memory versus 3/18 without it, but a grounded single-agent control also solves the earlier hard-task slice. A separate cross-session probe is excluded from efficacy conclusions because of confounded route schedules and unvalidated state labels. We examine 13 selected real-update hazard records across nine subjects, distinguishing stale-locator observations from policy-based withholding and offline retrieval diagnostics. The latter do not measure live-device failures or actions saved. These events are UI changes, not confirmed regression bugs. The evidence supports a bounded design and case study; it does not establish general efficacy, superiority over independent GUI-agent systems, or unrestricted exploration.

cs.SE

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.03868

EvoGroup: A Self-Evolving Harness for Group Agents in Shared Contexts

Xinyi Xu, Xiaoyang Li, Yucheng Feng, Wenqing Chen, Youliang Yuan, Wei Shi, Rongxiang Weng, Zibin Zheng

原文摘要

With the rise of Group Agent such as Claude Tag, agents increasingly serve as digital workers for long-term team collaboration. These agents operate over shared contexts that continuously accumulate through multi-user interactions. However, existing agent memory systems are mainly designed for individual users and struggle to capture group-level knowledge in shared contexts, including group relationships and collaboratively formed knowledge. Manual memory strategies further limit adaptation to diverse and evolving shared contexts. To this end, we propose \textbf{EvoGroup (EvoG)}, a self-evolving harness for group agents that starts from a minimal six-tool cold start and adapts through interaction with shared context. EvoG uses confidence and task outcomes to select informative trajectories, while also eliciting self-reflections on uncertainty in low-confidence cases. The analysis agent then examines selected trajectories through adaptive progressive disclosure, followed by bucketed analysis to synthesize high-level findings that guide the evolve agent in revising the Harness. Empirically, pass@1 on held-out split of EverMemBench improves from 71.07\% to 89.58\%, surpassing both self-evolving memory and human-designed harnesses, including Claude Code (85.83\%). To evaluate generalization, we freeze the evolved harness and evaluate it on the unseen GroupMemBench, where, with GPT-5.5, it demonstrates effective generalization and outperforms both memory baselines and human-designed harnesses.

cs.CL

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.02310

Detect, Repair and Audit: A Hierarchical Multi-Agent Defense Against Adversarial Attacks in Energy Forecasting

Zewei Shi, Ruoxi Sun, Minhui Xue, Feng Liu, Hua Ma, Xingliang Yuan, Surya Nepal

原文摘要

Accurate electricity-demand forecasting is essential for the modern energy infrastructure. Recently, Times-Series Large Language Models (TS-LLMs) have emerged as a promising approach to forecasting more accurate demand from historical consumption data in zero- and few-shot settings, helping energy operators effectively plan supply and allocate energy resources. However, these models are vulnerable to compromised sensor inputs that can distort forecasts and lead to erroneous market orders or unnecessary load shedding. Existing defenses focus primarily on end-device forecasting, overlooking the distinct data-drift, latency, and audit requirements across energy infrastructure. To address this gap, we extract from major energy standards and regulatory frameworks a three-layer conceptual model that captures the security postures for each layer. Guided by this model, we develop a Multi-Agent Reinforcement Learning (MARL) detector coupled with signal repair at local control centers, designed to mitigate distributional data drift through complementary feature-based detection and clean-reference-calibrated repair while meeting latency budgets. We further deploy an evidence-grounded LLM auditing agent with human-in-the-loop validation for audit-critical Virtual Power Plants (VPPs). In our evaluation, the best MARL configuration achieves an F1 score above 0.91 across all dataset-attack pairs, while signal repair reduces adversarial-to-clean input Relative Squared Error (RSE) by up to 83%. The few-shot auditor detects 77.78% of attacks in the Sherlock dataset. On the evaluated hardware, detection and repair require 7.95 ms on average, with a worst-observed latency of 14.50 ms, remaining within the local-center latency budget informed by IEEE 1646.

cs.CR

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2610.38126

Answer-Only Disclosure: Private-Context LLM Assistance over Web APIs with a Leakage Ledger

Liang Yuan, Xiaohu Du, Ruikun Luo, Jinze Wang, Mashud Rana, Kaibin Wang, Jiangwen Zhou, Bo Li 等 9 位作者

原文摘要

Web assistants increasingly answer questions about a user's private documents by pairing a small on-device model with a cloud LLM behind a Web API. In collaborative decoding, the document stays on the client, and the cloud receives only a public view that the user approves, such as the question alone, but the two models exchange messages while the answer is generated. We show that the message transcript can disclose private content that neither the public view nor the answer contains. In a planted-secret audit, speculative decoding with private drafts did not include the secret code in its answers, yet sent the code to the cloud in rejected drafts. We present ADLER, a decoding protocol for streaming LLM APIs in which every message to the cloud depends only on the public view and the answer the user sees, regardless of how the client combines the two models or how the cloud behaves. A client-side leakage ledger then bounds how much the answer itself reveals relative to a public reference process built from the cloud's replies. In the audit, no ADLER configuration exposed the secret beyond the answer. On GSM8K math problems with numerical values masked in the cloud's input, ADLER with an unlimited leakage budget raises the exact-match accuracy of a 0.5B client from 0.41 to 0.83 and of a 1.5B client from 0.67 to 0.85 under lenient answer parsing, close to the 0.92 achieved by uploading the full problem. It also exceeds a zero-leakage program baseline (0.69) by 0.14 and 0.16, respectively. In this baseline, the cloud writes a program that the client runs offline using the private numbers. The ledger records the disclosure charge associated with this gain; budgets of 16 nats or less leave accuracy near the cloud-only baseline on GSM8K.

cs.CR · cs.AI

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2610.37286

LESSRec: Level-Specific History Modeling for Semantic ID-Based Recommendation

Shangxin Yang, Zongwei Wang, Jing Long, Junliang Yu, Min Gao

原文摘要

Semantic ID-based recommendation represents each item as a sequence of semantic tokens obtained through residual quantization. This paradigm generates the next item's tokens autoregressively from coarse to fine, entailing each prediction attending to historical tokens at all semantic levels. However, our inference-time analysis reveals that historical tokens at the target-token's semantic level tend to be more informative, whereas other levels can introduce interference. Motivated by these findings, we propose LESSRec that separately encodes the historical token sequence at each semantic level and restricts attention over historical tokens to the sequence at the target token's semantic level during prediction. To complement these level-specific histories, LESSRec explicitly provides an item-level semantic sequence that represents each historical item's overall semantics in chronological order. The decoder accesses the level-aligned token sequence and the item-level semantic sequence through two separate cross-attention branches and fuses their contexts. We further introduce LESSRec+, which uses separate encoders for different semantic-level histories with the same total parameter count to reduce encoder computation. Extensive experiments across three public datasets and three tokenizers demonstrate the recommendation effectiveness of LESSRec and LESSRec+, with efficiency analyses highlighting the computational benefits of LESSRec+.

cs.IR

查看论文
arXiv ↗PDF ↗
首次观测 · 记录 #1ACL Anthology · ACL 2026:acl:2026.acl-long.139ACL Anthology · ACL 2026

CriticLean: Critic-Guided Reinforcement Learning for Mathematical Formalization

Zhongyuan Peng, Yifan Yao, Kaijing Ma, Shuyue Guo, Yizhe Li, Yichi Zhang, Chenchen Zhang, Yifan Zhang Princeton University 等 19 位作者

原文摘要

Translating natural language mathematical statements into formal, executable code is a fundamental challenge in automated theorem proving. While prior work has focused on generation and compilation success, little attention has been paid to the critic phase—the evaluation of whether generated formalizations truly capture the semantic intent of the original problem. In this paper, we introduce CriticLean, a novel critic-guided reinforcement learning framework that elevates the role of the critic from a passive validator to an active learning component. Specifically, first, we propose the CriticLeanGPT, trained via supervised fine-tuning and reinforcement learning, to rigorously assess the semantic fidelity of Lean 4 formalizations. Then, we introduce CriticLeanBench, a benchmark designed to measure models’ ability to distinguish semantically correct from incorrect formalizations, and demonstrate that our trained CriticLeanGPT models can significantly outperform strong open- and closed-source baselines. Building on the CriticLean framework, we construct FineLeanCorpus, a dataset comprising over 509K problems that exhibits rich domain diversity, broad difficulty coverage, and high correctness based on human evaluation.Overall, our findings highlight that optimizing the critic phase is essential for producing reliable formalizations and we hope our CriticLean will provide valuable insights for future advances in formal mathematical reasoning.

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.529ACL Anthology · ACL 2026

SHAPE: Unifying Safety, Helpfulness and Pedagogy for Educational LLMs

Sihang Zhao, Kangrui Yu, Youliang Yuan The Chinese University of Hong Kong-Shenzhen, Pinjia He The Chinese University of Hong Kong, Shenzhen, Hongyi Wen New York University, Shanghai

原文摘要

Large Language Models (LLMs) have been widely explored in educational scenarios. We identify a critical vulnerability in current educational LLMs, pedagogical jailbreaks, where students use answer-inducing prompts to elicit solutions rather than scaffolded instructions. To enable systematic study, we unify and formalize safe, helpful, and pedagogical behaviors with a knowledge-mastery graph and introduce SHAPE, a benchmark of 9,087 student-question pairs for evaluating tutoring behavior under adversarial pressure. We propose a graph-augmented tutoring pipeline that infers prerequisite concepts from queries, identifies mastery gaps, and routes generation between instructing and problem-solving via explicit gating. Experiments across multiple LLMs show that our method yields significantly improved safety under two pedagogical jailbreak settings, while maintaining near-ceiling helpfulness under the same evaluation protocol. Our code and data are available at https://github.com/MAPS-research/SHaPE

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.844ACL Anthology · ACL 2026

Curing Miracle Steps in LLM Mathematical Reasoning with Rubric Rewards

Youliang Yuan The Chinese University of Hong Kong-Shenzhen, Qiuyang Mang, Jingbang Chen The Chinese University of Hong Kong, Hong Wan, Xiaoyuan Liu Xiaohongshu and The Chinese University of Hong Kong, Shenzhen, Junjielong Xu Tencent and The Chinese University of Hong Kong, Shenzhen, Jen-tse Huang, Wenxuan Wang Renmin University of China 等 10 位作者

原文摘要

In this paper, we observe that current models are susceptible to reward hacking, leading to a substantial overestimation of a model’s reasoning ability. This is evidenced by a high incidence of “false positives”—solutions that reach the correct answer through an unsound process.Through a systematic analysis with human verification, we establish a taxonomy of these failure modes, identifying patterns like Miracle Steps—abrupt jumps to a correct output without a valid preceding derivation. Probing experiments suggest that these Miracle Steps are linked to answer-recall shortcuts, including memorization from pretraining, where the model accesses the correct answer independently of its reasoning chain.To mitigate this systemic issue, we introduce the Rubric Reward Model (RRM), a process-oriented reward function that evaluates the entire reasoning trajectory against problem-specific rubrics.The RRM explicitly penalizes logical flaws and encourages rigorous deduction.When integrated into an RL pipeline, RRM-based training consistently outperforms outcome-only supervision across four math benchmarks.Notably, it boosts Verified Pass@1024 on AIME2024 from 26.7% to 62.6% and reduces the incidence of Miracle Steps by 71%.Our work demonstrates that rewarding the solution process is crucial for building accurate and reliable models.

cs.CL

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