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

+2新增论文
+2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：2026-11-12 21:36 · 全站最近入库：2026-11-12 06:37各信源进度 →

统计口径与数据新鲜度

统计 2026-11-11 22:28 — 2026-11-12 22:28（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：2026-11-12 21:36最近发现新增版本：2026-11-12 06:37库内论文最新更新时间：2026-11-11 08:00

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。

21,675 已收录论文18,323 AI 相关6,056 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 2026-11-12 21:30（北京时间）
版本更新 · v9arXiv:2502.18770

Reward Shaping to Mitigate Reward Hacking in RLHF

Jiayi Fu, Xuandong Zhao, Chengyuan Yao, Heng Wang, Qi Han, Yanghua Xiao, Yuxia Wang

原文摘要

Reinforcement learning from human feedback (RLHF) is widely used to align large language models with human preferences, but it remains vulnerable to \emph{reward hacking}, whereby a policy exploits flaws in the reward function at the expense of the intended behavior. Reward shaping can stabilize training and partially mitigate reward hacking, yet its design principles remain underexplored. Our systematic study of common shaping methods suggests that shaped rewards should be bounded and should rise rapidly before gradually saturating. We introduce \emph{Preference as Reward} (PAR), which applies a sigmoid to centered rewards to produce a signal that reflects the reward model's latent preferences. We show that PAR has a variance-reduction property that stabilizes training and substantially widens the window for effective early stopping. Using Proximal Policy Optimization (PPO) and Gemma2-2B, PAR consistently outperforms the other shaping methods evaluated and demonstrates high data efficiency and robustness. Comparisons with unshaped rewards across four base models and four reinforcement learning algorithms show that PAR is effective across models and particularly beneficial for actor--critic methods when value estimates become unstable. The code is available at https://github.com/PorUna-byte/PAR.

cs.LG · cs.AI · cs.CL

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.11801

Proactive Dimension-Adaptive Denoising for Graph Collaborative Filtering via Representation Coherence

Wenjun Li, Dong Zhao, Jiaying Shi, Juan Zhang, Zhaoming Pan, Zhimin Zhang

原文摘要

Implicit feedback in recommender systems contains noisy interactions (e.g., accidental clicks, impulse purchases) that propagate through user-item graphs and contaminate learned representations. Existing explicit denoising methods uniformly rely on training loss to estimate interaction reliability. However, loss is a fundamentally passive signal: it detects noise only after it has caused prediction errors, and it is heavily confounded by training stage, negative sampling, and hard positives. We propose Noise-aware Dimension Learning (NDL), a proactive dimension-adaptive denoising framework. NDL introduces Representation Coherence Detection (RCD), which evaluates the coherence between an interaction representation and the user's historical preference centroid in embedding space, providing an active noise signal independent of loss. RCD and the loss signal are fused via activity-aware weighting (active users rely more on coherence while cold-start users automatically fall back to loss), and the fused estimate drives an interaction-level dimension-adaptive mask that limits the effective representation capacity of unreliable interactions. A lightweight contrastive regularizer further improves embedding uniformity, stabilizing the geometry on which RCD discriminates. On three public benchmarks NDL consistently outperforms state-of-the-art denoising methods, lifting Recall@5 by +17.95% on Amazon-book, +12.33% on Yelp, and +5.63% on MovieLens-1M over the strongest baseline DAR; the core variant without contrastive learning already surpasses DAR on sparser datasets, confirming representation coherence as the primary driver of these gains. NDL is confined to the training loop and is serving-compatible with LightGCN-based production pipelines, adding no online latency or parameters at inference time. Code: https://github.com/GordonDongZHAO/NDL

cs.IR

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2607.21394

CRAFT: Exploring Wearable Creative AI on Smart Glasses for Fiction Writing in Real-World Contexts

Runze Cai, Yuxuan Huang, Lin-Ping Yuan, Kexin Xiang, David Hsu, Collier Nogues, Jussi Holopainen, Shengdong Zhao

原文摘要

Creative writing increasingly integrates AI assistance, yet current tools miss in-situ moments when writers draw inspiration from real-world experiences. We envision Context-aware Reality-Fiction Transformation (CRAFT), an approach for AI glasses that translates daily experiences into fiction narratives. We explored its desirability, feasibility, and potential viability through three studies. Interviews with nine writers yielded desires and three design goals: 1) augmenting in-situ perception to bridge reality-fiction gaps, 2) promoting authenticity grounded in real-world experiences while maintaining fictionalization, and 3) preserving creative agency, enjoyment, and life-art boundaries. Co-design workshops with 16 writers and researchers operationalized these goals into concrete interaction mechanisms using a technology probe. We then conducted supported field trials with eight writers across 24 sessions using a refined probe, revealing writer-perceived benefits (e.g., enriched fictional ideas from serendipitous encounters), emergent practices (e.g., micro-creation), and design considerations for future sustained use. We contribute design explorations for the CRAFT approach, offering design implications and empirical insights on ubiquitous human-AI creative collaboration in everyday life.

cs.HC

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.07682

DEQ-Det: Dual Evidence Queries for Few-Shot Analog IC Device Detection

Zidong Zhao, Rong Huang, Chengxing Fan, Jie Hu, Malik Haris, Ghazal A. Fahmy, Morteza Mousazadeh, Guoqiang Zhang

原文摘要

Detecting analog integrated circuit (IC) devices in layout images supports circuit analysis and inspection. With only a few annotated examples, variation in device geometry and visual similarity to dummy or routing structures make reliable detection difficult. Foreground similarity alone can highlight plausible targets without distinguishing them from competing layout patterns. We propose Dual Evidence Queries for Detection (DEQ-Det), which constructs complementary foreground and background references from the same support images without additional annotations. Foreground references capture target morphology, while selected background references characterize surrounding layout patterns. Both sources guide query selection and region verification, combining target appearance with layout context. Across five held-out device categories and three random seeds, DEQ-Det achieves macro AP50 of 39.41 and 53.03 at 10 and 30 shots, exceeding the strongest listed baselines by 3.97 and 9.71 points, respectively. A matched 30-shot ablation further shows that RoI background evidence reduces pooled false positives at comparable recall, supporting its role in interference suppression.

cs.CV

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.

版本更新 · v2arXiv:2605.08441

DUET: Optimize Token-Budget Allocation for Reinforcement Learning with Verifiable Rewards

Haoyu Hu, Xuandong Zhao, Xuhai "Orson'' Xu, Nori Jacoby

原文摘要

Reinforcement learning with verifiable rewards (RLVR) generates hundreds of thousands of tokens per training step, with rollout generation dominating the computational cost. The overall token budget can be controlled along two main dimensions: (i) deciding which prompts to allocate rollouts to, and (ii) deciding how long each rollout should be. Prior work has generally controlled only one of these dimensions at a time. We show that jointly tuning both decisions under a shared compute budget improves both reasoning quality and wall-clock training time. We instantiate this view as \textbf{DU}al-controlled tok\textbf{E}n alloca\textbf{T}ion (DUET), a computationally efficient layer over GRPO that uses a lightweight pre-rollout surrogate of prompt informativeness to set how many rollouts each prompt receives, and a marker-gated abort rule with importance reweighting to set when to stop them. On Qwen3-1.7B trained on MATH, DUET outperforms full-budget GRPO and the other three budget-aware baseline methods. DUET's advantage further generalizes to other benchmarks across math and coding, and is on par with the best baseline on the scientific Q\&A domain, while also achieving a $1.62\times$ wall-clock speedup. More notably, using only 50\% of the token budget, DUET still outperforms all baseline methods at their full budget, achieving an even higher $2.51\times$ speedup over full-budget GRPO. We verify the high performance of DUET on other backbone LLMs, including Qwen3-4B to 14B and Llama-3.2-3B-Instruct. Notably, the gap between DUET and the strongest baseline \emph{widens} as the budget tightens, contrary to the usual pattern in which efficient methods trade off quality as compute decreases. More broadly, these results suggest that DUET budget-aware control strategies are valuable not only for accelerating training, but also for improving the quality of the learning signal.

cs.LG · cs.AI

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2610.32046HF 社区推荐

Can Test-Time Memory Evolution Outperform Test-Time Training for Scientific Discovery?

Yepeng Liu, Shengmai Chen, Xuandong Zhao, Chengzhi Liu, Dawn Song, Xin Eric Wang, Yuheng Bu

原文摘要

Recent \emph{test-time training} (TTT) frameworks have enabled open-weight large language models (LLMs) to outperform inference-only frameworks powered by frontier models on scientific discovery tasks. However, TTT introduces additional computational cost for online weight updates during the discovery process, making computational efficiency an important consideration when scaling to larger models. This motivates us to explore an efficient \emph{test-time adaptation} method for scientific discovery. We introduce \emph{Test-Time Memory Evolution} (TTME), a plug-and-play test-time adaptation method, as an alternative to TTT. TTME continually updates the external memory by distilling the experience to guide subsequent discovery without updating model weights. To have a clean comparison, we evaluate TTME and TTT under the same discovery framework, replacing only the test-time adaptation method. Our evaluation shows that TTME outperforms TTT on all five mathematical tasks, while reducing computational cost by $82.8$-$84.6\%$ in FLOPs and $47.7$-$58.4\%$ in runtime. These results establish TTME as an effective and computationally efficient alternative to TTT in the evaluated discovery settings.

cs.AI · cs.LG

查看论文
arXiv ↗PDF ↗
首次观测 · 记录 #1ACL Anthology · ACL 2026:acl:2026.acl-long.115ACL Anthology · ACL 2026

Toward Secure Tuning: Mitigating Security Risks from Instruction Fine-Tuning

Yanrui Du, Fenglei Fan City University of Hong Kong, Sendong Zhao, Jiawei Cao, Ming Ma, Danyang Zhao Harbin Institute of Technology, Shuren Qi City University of Hong Kong, Ting Liu Harbin Institute of Technology 等 9 位作者

原文摘要

Instruction Fine-Tuning (IFT) has emerged as a critical technique for customizing Large Language Models (LLMs) to meet diverse downstream applications. However, recent studies have revealed that IFT can compromise the built-in security mechanisms of LLMs, thereby posing significant security risks. Although defense methods targeting various training stages have been proposed, they either face challenges in practical deployment or exhibit instability and limited performance gains. In our study, we propose a novel SWAT method that introduces a key idea: shifting more of the learning burden onto security-robust parameters. To this end, our study investigates how module-level parameters affect LLMs’ internal security feature space, aiming to uncover robustness patterns in parameters. Guided by this analysis, we identify a robust module set (Mods_Rob) that exhibits minimal effects on LLMs’ security feature space. Leveraging this insight, SWAT proceeds in two phases: (1) a warm-up phase that preferentially trains Mods_Rob to learn low-level features with minimal security risk, followed by (2) standard tuning to achieve optimal task performance. Across diverse knowledge-intensive datasets, scenarios, and LLMs, SWAT substantially reduces security risks without sacrificing task performance gains.

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.167ACL Anthology · ACL 2026

Collaborative Chain-of-Agents for Parametric-Retrieved Knowledge Synergy

Yi Jiang, Sendong Zhao, Jianbo Li, Haochun Wang Harbin Institute of Technology, Lizhe Zhang, Yan Liu, Bing Qin Harbin Institute of Technology

原文摘要

Retrieval-Augmented Generation (RAG) enhances Large Language Models (LLMs), especially for knowledge-intensive tasks. Despite its advantages, current RAG methods often struggle to fully exploit knowledge during generation. In particular, the synergy between the model’s internal parametric knowledge and external retrieved knowledge remains limited. Retrieved contents may sometimes mislead generation, while certain generated content can guide the model toward more accurate outputs. In this work, we propose Collaborative Chain-of-Agents, a framework designed to enhance explicitly synergy over both parametric and retrieved knowledge. Specifically, we first introduce CoCoA-zero, a multi-agent RAG framework that first performs conditional knowledge induction and then reasons answers. Building on this, we develop CoCoA, a long-chain training strategy that synthesizes extended multi-agent reasoning trajectories from CoCoA-zero to fine-tune the LLM. This strategy enhances the model’s capability to explicitly integrate and jointly leverage parametric and retrieved knowledge. Experimental results demonstrate the superiority of CoCoA in open-domain QA and multi-hop QA.

cs.CL · cs.MA

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.1682ACL Anthology · ACL 2026

Turning Failures into Value: Negative Experience Replay for RLVR via Confidence Gating and Boundary Failure Sampling

Jialiang Guo, Fucheng Xiong, Xu He, Haodong Zhao Meituan, Xingyang li Meituan, Ke Zeng, Xunliang Cai Meituan

原文摘要

Reinforcement Learning with Verifiable Rewards (RLVR) has become the standard paradigm for enhancing reasoning capabilities in Large Language Models, yet on-policy algorithms like GRPO suffer from sample inefficiency. Current experience replay methods for RLVR typically replay correct trajectories to consolidate learned reasoning patterns and accelerate convergence, but overlook the vast failure space. This work investigates how to effectively replay failure trajectories. We find that the high heterogeneity of failures renders random replay ineffective, and that high-value negatives should be both gradient-efficient and structurally proximal to correct solutions. To this end, we propose NexGRPO, which employs mid-confidence gating to filter invalid noise and saturated errors, and utilizes boundary failure sampling to retrieve boundary errors semantically similar to correct solutions for targeted refinement. Extensive experiments on mathematical and general reasoning benchmarks demonstrate that NexGRPO outperforms strong baaselines and achieves improved out-of-distribution generalization.

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.395ACL Anthology · ACL 2026

When Correct Beliefs Collapse: Epistemic Resilience of LLMs under Clinical Pressure

Boyu Xiao, Xiuqi Tian, Xuwen Song Harbin Institute of Technology, Haochun Wang Harbin Institute of Technology, Guanchun Song, Sendong Zhao, Bing Qin Harbin Institute of Technology

原文摘要

Despite strong medical benchmark accuracy, LLMs can exhibit severe multi-turn sycophancy in clinical dialogue, abandoning initial correct diagnosis under escalating pressure. We propose Med-Stress, a targeted stress test framework that evaluates belief stability under escalating pressure. Across nine frontier large language models (LLMs), we find a clear dissociation between medical knowledge and robustness: high initial diagnostic capability does not imply high belief stability, yielding large knowledge-robustness gaps for several LLMs. To mitigate this failure mode, we propose a lightweight inference-time defense, RBED (Role-Based Epistemic Defense), and R-FT (Resilience-oriented Fine-Tuning), a training-time approach that internalizes evidence-based resistance to pressure. Experiments show that R-FT nearly eliminates belief change and substantially improves robustness.

cs.CL

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