大模型推理能力 - arXivDaily 专题

2504.07416 2025-11-07 cs.CV cs.CL cs.LG 62%

RadZero: Similarity-Based Cross-Attention for Explainable Vision-Language Alignment in Chest X-ray with Zero-Shot Multi-Task Capability

Jonggwon Park, Byungmu Yoon, Soobum Kim, Kyoyun Choi

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

Comments NeurIPS 2025

2511.03206 2025-11-06 cs.CV cs.AI cs.LG 62%

QG-CoC: Question-Guided Chain-of-Captions for Large Multimodal Models

Kuei-Chun Kao, Hsu Tzu-Yin, Yunqi Hong, Ruochen Wang, Cho-Jui Hsieh

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

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

Comments 16 pages

Journal ref EMNLP 2025

2511.01323 2025-11-04 cs.CL cs.AI 62%

DEEPAMBIGQA: Ambiguous Multi-hop Questions for Benchmarking LLM Answer Completeness

Jiabao Ji, Min Li, Priyanshu Kumar, Shiyu Chang, Saloni Potdar

机构 * UCSB（UCSB大学）； Apple（苹果公司）

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

Comments 25 pages

2410.18032 2025-11-04 cs.AI cs.CL cs.MA 62%

GraphTeam: Facilitating Large Language Model-based Graph Analysis via Multi-Agent Collaboration

Xin Li, Qizhi Chu, Yubin Chen, Yang Liu, Yaoqi Liu, Zekai Yu, Weize Chen, Chen Qian, Chuan Shi, Cheng Yang

机构 * Beijing University of Posts and Telecommunications（北京邮电大学）； The Chinese University of Hong Kong（香港中文大学）； Tsinghua University（清华大学）

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

2511.00070 2025-11-04 cs.LG cs.AI 62%

Benchmarking Generative AI Against Bayesian Optimization for Constrained Multi-Objective Inverse Design

Muhammad Bilal Awan, Abdul Razzaq, Abdul Shahid

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

Comments 17 pages, 2 Figures

2510.24797 2025-10-31 cs.CL cs.AI 62%

Large Language Models Report Subjective Experience Under Self-Referential Processing

Cameron Berg, Diogo de Lucena, Judd Rosenblatt

机构 * AE Studio（AE工作室）

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

2503.09499 2025-10-31 cs.CV cs.AI cs.CL 62%

MindGYM: What Matters in Question Synthesis for Thinking-Centric Fine-Tuning?

Zhe Xu, Daoyuan Chen, Zhenqing Ling, Yaliang Li, Ying Shen

机构 * Sun Yat-sen University（中山大学）； Alibaba Group（阿里巴巴集团）； FSIETP

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

Comments Accepted by NeurIPS'25. 30 pages, 2 figures, 13 tables

2403.13002 2025-10-30 cs.HC cs.AI cs.CL 62%

AutoTRIZ: Automating Engineering Innovation with TRIZ and Large Language Models

Shuo Jiang, Weifeng Li, Yuping Qian, Yangjun Zhang, Jianxi Luo

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

Comments 28 pages, 12 figures

Journal ref Advanced Engineering Informatics 65 (2025): 103312

2510.24706 2025-10-29 cs.CL cs.AI cs.HC cs.SE 62%

ComboBench: Can LLMs Manipulate Physical Devices to Play Virtual Reality Games?

Shuqing Li, Jiayi Yan, Chenyu Niu, Jen-tse Huang, Yun Peng, Wenxuan Wang, Yepang Liu, Michael R. Lyu

机构 * Chinese University of Hong Kong（香港中文大学）； Southern University of Science and Technology（南方科技大学）

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

2510.24320 2025-10-29 cs.CL cs.AI 62%

Critique-RL: Training Language Models for Critiquing through Two-Stage Reinforcement Learning

Zhiheng Xi, Jixuan Huang, Xin Guo, Boyang Hong, Dingwen Yang, Xiaoran Fan, Shuo Li, Zehui Chen, Junjie Ye, Siyu Yuan, Zhengyin Du, Xuesong Yao, Yufei Xu, Jiecao Chen, Rui Zheng, Tao Gui, Qi Zhang, Xuanjing Huang

机构 * Fudan University（复旦大学）； ByteDance Seed（字节跳动种子）

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

Comments Preprint, 25 pages, 9 figures. Code: https://github.com/WooooDyy/Critique-RL

2510.23941 2025-10-29 cs.CL cs.AI 62%

Auto prompting without training labels: An LLM cascade for product quality assessment in e-commerce catalogs

Soham Satyadharma, Fatemeh Sheikholeslami, Swati Kaul, Aziz Umit Batur, Suleiman A. Khan

机构 * Amazon Catalog AI（亚马逊目录人工智能）

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

2510.21727 2025-10-29 cs.IR cs.AI cs.LG 62%

Your Dense Retriever is Secretly an Expeditious Reasoner

Yichi Zhang, Jun Bai, Zhixin Cai, Shuhan Qin, Zhuofan Chen, Jinghua Guan, Wenge Rong

机构 * School of Computer Science and Engineering, Beihang University, China（北京航空航天大学计算机科学与工程学院）； Beijing Institute for General Artificial Intelligence, China（北京通用人工智能研究院）

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

Comments 16 pages, 11 figures

2510.23208 2025-10-28 cs.LG cs.AI 62%

Increasing LLM Coding Capabilities through Diverse Synthetic Coding Tasks

Amal Abed, Ivan Lukic, Jörg K. H. Franke, Frank Hutter

机构 * University of Freiburg（弗赖堡大学）； ELLIS Institute Tübingen（图宾根ELLIS研究所）； Open-Sci Collective（开放科学集体）； LAION ； Prior Labs（Prior实验室）

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

Comments Presented at the 39th Conference on Neural Information Processing Systems (NeurIPS 2025) Workshop: The 4th Deep Learning for Code Workshop (DL4C)

2510.22344 2025-10-28 cs.CL cs.AI cs.IR 62%

FAIR-RAG: Faithful Adaptive Iterative Refinement for Retrieval-Augmented Generation

Mohammad Aghajani Asl, Majid Asgari-Bidhendi, Behrooz Minaei-Bidgoli

机构 * Sharif University of Technology（谢里夫理工大学）； Iran University of Science and Technology（伊朗科学技术大学）； Noor Avaran Jelvehaye Maanaei Najm Co., Ltd.（诺尔·阿瓦兰·杰拉维耶·马纳埃尼纳姆公司）

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

Comments 30 pages, 5 figures, 5 tables. Keywords: Retrieval-Augmented Generation (RAG), Large Language Models (LLMs), Agentic AI, Multi-hop Question Answering, Faithfulness

2510.21817 2025-10-28 cs.RO cs.CL cs.LG 62%

VITA-E: Natural Embodied Interaction with Concurrent Seeing, Hearing, Speaking, and Acting

Xiaoyu Liu, Chaoyou Fu, Chi Yan, Chu Wu, Haihan Gao, Yi-Fan Zhang, Shaoqi Dong, Cheng Qian, Bin Luo, Xiuyong Yang, Guanwu Li, Yusheng Cai, Yunhang Shen, Deqiang Jiang, Haoyu Cao, Xing Sun, Caifeng Shan, Ran He

机构 * Nanjing University（南京大学）； Tencent Youtu Lab（腾讯云图实验室）； CASIA（中国科学院自动化研究所）； Fourier Intelligence Inc.（傅里叶智能公司）

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

Comments Homepage: https://lxysl.github.io/VITA-E/

2410.12854 2025-10-27 cs.CL cs.AI 62%

TPO: Aligning Large Language Models with Multi-branch & Multi-step Preference Trees

Weibin Liao, Xu Chu, Yasha Wang

机构 * School of Computer Science, Peking University（北京大学计算机科学系）； Center on Frontiers of Computing Studies, Peking University（北京大学前沿计算研究中心）； National Research and Engineering Center of Software Engineering, Peking University（北京大学软件工程研究中心）

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

Comments Accepted by ICLR 2025

2510.18053 2025-10-22 cs.LG cs.AI 62%

Adaptive Divergence Regularized Policy Optimization for Fine-tuning Generative Models

Jiajun Fan, Tong Wei, Chaoran Cheng, Yuxin Chen, Ge Liu

机构 * University of Illinois Urbana-Champaign（伊利诺伊大学厄巴纳-香槟分校）

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

Comments 30 pages

2505.19591 2025-10-22 cs.CL cs.AI cs.MA 62%

Multi-Agent Collaboration via Evolving Orchestration

Yufan Dang, Chen Qian, Xueheng Luo, Jingru Fan, Zihao Xie, Ruijie Shi, Weize Chen, Cheng Yang, Xiaoyin Che, Ye Tian, Xuantang Xiong, Lei Han, Zhiyuan Liu, Maosong Sun

机构 * Tsinghua University（清华大学）； Shanghai Jiao Tong University（上海交通大学）； Beijing University of Posts and Telecommunications（北京邮电大学）； Siemens（西门子）； Tencent Robotics X（腾讯机器人研究院）

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

Comments accepted at NeurIPS 2025

2510.16555 2025-10-21 cs.AI cs.LG 62%

Urban-R1: Reinforced MLLMs Mitigate Geospatial Biases for Urban General Intelligence

Qiongyan Wang, Xingchen Zou, Yutian Jiang, Haomin Wen, Jiaheng Wei, Qingsong Wen, Yuxuan Liang

机构 * The Hong Kong University of Science and Technology (Guangzhou)（香港科技大学（广州））； Carnegie Mellon University（卡内基梅隆大学）； Squirrel Ai Learning

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

2510.16363 2025-10-21 cs.CL cs.AI 62%

End-to-End Argument Mining through Autoregressive Argumentative Structure Prediction

Nilmadhab Das, Vishal Vaibhav, Yash Sunil Choudhary, V. Vijaya Saradhi, Ashish Anand

机构 * Applied Machine Learning (AMaL) Lab（应用机器学习（AMaL）实验室）； Department of Computer Science and Engineering, Indian Institute of Technology (IIT), Guwahati, India（计算机科学与工程系，印度理工学院（IIT），古瓦哈蒂，印度）； North Eastern Regional Institute of Science and Technology (NERIST), Arunachal Pradesh, India（北东部科学与技术研究所（NERIST），阿鲁纳恰尔邦，印度）

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

Comments Accepted version. To appear in IJCNN 2025

2510.15992 2025-10-21 cs.LG cs.AI 62%

Stratos: An End-to-End Distillation Pipeline for Customized LLMs under Distributed Cloud Environments

Ziming Dai, Tuo Zhang, Fei Gao, Xingyi Cai, Xiaofei Wang, Cheng Zhang, Wenyu Wang, Chengjie Zang

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

2505.18573 2025-10-21 cs.LG cs.CL 62%

Enhancing Efficiency and Exploration in Reinforcement Learning for LLMs

Mengqi Liao, Xiangyu Xi, Ruinian Chen, Jia Leng, Yangen Hu, Ke Zeng, Shuai Liu, Huaiyu Wan

机构 * School of Computer Science and Technology, Beijing Jiaotong University（计算机科学与技术学院，北京交通大学）； Meituan（美团）； Beijing Key Laboratory of Traffic Data Mining and Embodied Intelligence（北京交通数据挖掘与具身智能重点实验室）

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

Comments Accept by EMNLP 2025 main

2510.15374 2025-10-20 cs.AI cs.LG 62%

Towards Flash Thinking via Decoupled Advantage Policy Optimization

Zezhong Tan, Hang Gao, Xinhong Ma, Feng Zhang, Ziqiang Dong

机构 * Alibaba Inc.（阿里巴巴公司）； Peking University（北京大学）

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

2510.15280 2025-10-20 cs.LG cs.AI cs.DL 62%

Foundation Models for Scientific Discovery: From Paradigm Enhancement to Paradigm Transition

Fan Liu, Jindong Han, Tengfei Lyu, Weijia Zhang, Zhe-Rui Yang, Lu Dai, Cancheng Liu, Hao Liu

机构 * The Hong Kong University of Science and Technology (Guangzhou)（香港科技大学（广州））； The Hong Kong University of Science and Technology（香港科技大学）

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

Comments NeurIPS 2025

2510.13253 2025-10-20 cs.CV cs.AI cs.LG 62%

End-to-End Multi-Modal Diffusion Mamba

Chunhao Lu, Qiang Lu, Meichen Dong, Jake Luo

机构 * China University of Petroleum-Beijing（中国石油大学（北京））； Leyard Optoelectronic（莱亚德光电）； University of Wisconsin-Milwaukee（威斯康星大学密尔沃基分校）

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

Comments Accepted by ICCV 2025

2509.20186 2025-10-20 cs.CL cs.LG 62%

Thinking Augmented Pre-training

Liang Wang, Nan Yang, Shaohan Huang, Li Dong, Furu Wei

机构 * Microsoft Research（微软研究院）

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

Comments 19 pages; v4 fixes an issue for HumanEval scores

2505.14714 2025-10-20 cs.CV cs.AI cs.CL 62%

KGAlign: Joint Semantic-Structural Knowledge Encoding for Multimodal Fake News Detection

Tuan-Vinh La, Minh-Hieu Nguyen, Minh-Son Dao

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

Comments Withdrawn by the authors due to lack of explicit agreement from all co-authors to post this version publicly on arXiv

2510.13913 2025-10-17 cs.CL cs.AI 62%

Synthesizing Agentic Data for Web Agents with Progressive Difficulty Enhancement Mechanisms

Shrey Pandit, Xuan-Phi Nguyen, Yifei Ming, Austin Xu, Jiayu Wang, Caiming Xiong, Shafiq Joty

机构 * Salesforce AI Research（Salesforce AI研究院）； University of Wisconsin-Madison（威斯康星大学麦迪逊分校）

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

Comments Preprint. ICLR 26 submission

2507.18671 2025-10-17 cs.LG cs.AI 62%

Innovator: Scientific Continued Pretraining with Fine-grained MoE Upcycling

Ning Liao, Xiaoxing Wang, Zehao Lin, Weiyang Guo, Feng Hong, Shixiang Song, Geng Yu, Zihua Zhao, Sitao Xie, Longxuan Wei, Xiangqi Jin, Xiaohan Qin, Jiale Ma, Kai Chen, Jiangchao Yao, Zhouhan Lin, Junchi Yan, Zhiyu Li, Feiyu Xiong, Yanfeng Wang, Linfeng Zhang

机构 * Shanghai Jiao Tong University（上海交通大学）； MemTensor (Shanghai) Technology Co., Ltd.（MemTensor（上海）技术有限公司）； Institute for Advanced Algorithms Research, Shanghai.（上海先进算法研究所）

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

Comments Technical Report

2506.21638 2025-10-17 cs.IR cs.AI cs.LG 62%

R1-Ranker: Teaching LLM Rankers to Reason

Tao Feng, Zhigang Hua, Zijie Lei, Yan Xie, Shuang Yang, Bo Long, Jiaxuan You

机构 * University of Illinois Urbana-Champaign（伊利诺伊大学厄巴纳-香槟分校）； Meta Monetization AI（Meta 资源 monetization AI）

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

1. 复杂问题求解 5001 篇

RadZero: Similarity-Based Cross-Attention for Explainable Vision-Language Alignment in Chest X-ray with Zero-Shot Multi-Task Capability

QG-CoC: Question-Guided Chain-of-Captions for Large Multimodal Models

DEEPAMBIGQA: Ambiguous Multi-hop Questions for Benchmarking LLM Answer Completeness

GraphTeam: Facilitating Large Language Model-based Graph Analysis via Multi-Agent Collaboration

Benchmarking Generative AI Against Bayesian Optimization for Constrained Multi-Objective Inverse Design

Large Language Models Report Subjective Experience Under Self-Referential Processing

MindGYM: What Matters in Question Synthesis for Thinking-Centric Fine-Tuning?

AutoTRIZ: Automating Engineering Innovation with TRIZ and Large Language Models

ComboBench: Can LLMs Manipulate Physical Devices to Play Virtual Reality Games?

Critique-RL: Training Language Models for Critiquing through Two-Stage Reinforcement Learning

Auto prompting without training labels: An LLM cascade for product quality assessment in e-commerce catalogs

Your Dense Retriever is Secretly an Expeditious Reasoner

Increasing LLM Coding Capabilities through Diverse Synthetic Coding Tasks

FAIR-RAG: Faithful Adaptive Iterative Refinement for Retrieval-Augmented Generation

VITA-E: Natural Embodied Interaction with Concurrent Seeing, Hearing, Speaking, and Acting

TPO: Aligning Large Language Models with Multi-branch & Multi-step Preference Trees

Adaptive Divergence Regularized Policy Optimization for Fine-tuning Generative Models

Multi-Agent Collaboration via Evolving Orchestration

Urban-R1: Reinforced MLLMs Mitigate Geospatial Biases for Urban General Intelligence

End-to-End Argument Mining through Autoregressive Argumentative Structure Prediction

Stratos: An End-to-End Distillation Pipeline for Customized LLMs under Distributed Cloud Environments

Enhancing Efficiency and Exploration in Reinforcement Learning for LLMs

Towards Flash Thinking via Decoupled Advantage Policy Optimization

Foundation Models for Scientific Discovery: From Paradigm Enhancement to Paradigm Transition

End-to-End Multi-Modal Diffusion Mamba

Thinking Augmented Pre-training

KGAlign: Joint Semantic-Structural Knowledge Encoding for Multimodal Fake News Detection

Synthesizing Agentic Data for Web Agents with Progressive Difficulty Enhancement Mechanisms

Innovator: Scientific Continued Pretraining with Fine-grained MoE Upcycling

R1-Ranker: Teaching LLM Rankers to Reason