大模型推理能力 - arXivDaily 专题

2505.20013 2025-09-19 cs.CL 88%

WebCoT: Enhancing Web Agent Reasoning by Reconstructing Chain-of-Thought in Reflection, Branching, and Rollback

Minda Hu, Tianqing Fang, Jianshu Zhang, Junyu Ma, Zhisong Zhang, Jingyan Zhou, Hongming Zhang, Haitao Mi, Dong Yu, Irwin King

机构 * Chinese University of Hong Kong（香港中文大学）； Tencent AI Lab（腾讯AI实验室）

专题命中复杂问题求解：reasoning(title,abstract)；chain-of-thought(title,abstract)；分类 cs.CL

Comments 18 pages

2509.12275 2025-09-19 cs.SD cs.AI eess.AS 83%

Omni-CLST: Error-aware Curriculum Learning with guided Selective chain-of-Thought for audio question answering

Jinghua Zhao, Hang Su, Lichun Fan, Zhenbo Luo, Hui Wang, Haoqin Sun, Yong Qin

机构 * TMCC, College of Computer Science, Nankai University, Tianjin, China（TMCC，计算机科学学院，南开大学，天津，中国）； MiLM Plus, Xiaomi Inc.（MiLM Plus，小米公司）

专题命中复杂问题求解：chain-of-thought(title,abstract)；reasoning(abstract)；分类 cs.AI

Comments 5 pages, 1 figure, 2 tables submitted to icassp, under prereview

2506.05128 2025-09-19 cs.CL cs.AI cs.LG 82%

DiCoRe: Enhancing Zero-shot Event Detection via Divergent-Convergent LLM Reasoning

Tanmay Parekh, Kartik Mehta, Ninareh Mehrabi, Kai-Wei Chang, Nanyun Peng

机构 * Computer Science Department, University of California, Los Angeles（加州大学洛杉矶分校计算机科学系）

专题命中复杂问题求解：reasoning(title,abstract)；分类 cs.CL、cs.AI、cs.LG

Comments Accepted at EMNLP 2025 Main

2509.14284 2025-09-19 cs.CR cs.AI cs.CL 73%

The Sum Leaks More Than Its Parts: Compositional Privacy Risks and Mitigations in Multi-Agent Collaboration

Vaidehi Patil, Elias Stengel-Eskin, Mohit Bansal

机构 * UNC Chapel Hill（北卡罗来纳大学教堂山分校）； The University of Texas at Austin（德克萨斯大学奥斯汀分校）

专题命中复杂问题求解：reasoning(abstract)；chain-of-thought(abstract)；分类 cs.CL、cs.AI

Comments Code: https://github.com/Vaidehi99/MultiAgentPrivacy

2502.17651 2025-09-19 cs.CV cs.AI cs.CL 73%

METAL: A Multi-Agent Framework for Chart Generation with Test-Time Scaling

Bingxuan Li, Yiwei Wang, Jiuxiang Gu, Kai-Wei Chang, Nanyun Peng

机构 * University of California, Los Angeles（加州大学洛杉矶分校）； University of California, Merced（加州大学默塞德分校）； Adobe Research（Adobe研究）

专题命中复杂问题求解：reasoning(abstract)；self-correction(abstract)；分类 cs.CL、cs.AI

Comments ACL2025 Main

2508.15213 2025-09-19 cs.CL 57%

Select to Know: An Internal-External Knowledge Self-Selection Framework for Domain-Specific Question Answering

Bolei He, Xinran He, Run Shao, Shanfu Shu, Xianwei Xue, Mingquan Cheng, Haifeng Li, Zhenhua Ling

机构 * University of Science and Technology of China（中国科学技术大学）； Baidu Inc.（百度公司）； Central South University（中南大学）； Chongqing University（重庆大学）

专题命中复杂问题求解：reasoning(abstract)；分类 cs.CL

Comments EMNLP2025 Findings

2509.14889 2025-09-19 cs.RO 50%

CollabVLA: Self-Reflective Vision-Language-Action Model Dreaming Together with Human

Nan Sun, Yongchang Li, Chenxu Wang, Huiying Li, Huaping Liu

机构 * Department of Computer Science and Technology, Tsinghua University（计算机科学与技术系，清华大学）

专题命中复杂问题求解：reasoning(abstract)

Comments 8 pages, 5 figures, 3 tables

1. 复杂问题求解 7 篇

WebCoT: Enhancing Web Agent Reasoning by Reconstructing Chain-of-Thought in Reflection, Branching, and Rollback

Omni-CLST: Error-aware Curriculum Learning with guided Selective chain-of-Thought for audio question answering

DiCoRe: Enhancing Zero-shot Event Detection via Divergent-Convergent LLM Reasoning

The Sum Leaks More Than Its Parts: Compositional Privacy Risks and Mitigations in Multi-Agent Collaboration

METAL: A Multi-Agent Framework for Chart Generation with Test-Time Scaling

Select to Know: An Internal-External Knowledge Self-Selection Framework for Domain-Specific Question Answering

CollabVLA: Self-Reflective Vision-Language-Action Model Dreaming Together with Human