语言大模型 / LLM - arXivDaily 专题

2507.08108 2025-07-14 stat.ML cs.DS cs.LG math.PR math.ST stat.TH 57%

Mallows Model with Learned Distance Metrics: Sampling and Maximum Likelihood Estimation

Yeganeh Alimohammadi, Kiana Asgari

机构 * University of Southern California（南加州大学）； Stanford University（斯坦福大学）

专题命中后训练与偏好优化：language model(abstract)；分类 cs.LG

2507.06221 2025-07-09 cs.AI cs.GT 57%

Aligned Textual Scoring Rules

Yuxuan Lu, Yifan Wu, Jason Hartline, Michael J. Curry

机构 * Peking University（北京大学）； Northwestern University（西北大学）； University of Illinois Chicago（伊利诺伊大学香槟分校）

专题命中后训练与偏好优化：language model(abstract)；分类 cs.AI

2507.00828 2025-07-02 cs.CL 57%

ProxAnn: Use-Oriented Evaluations of Topic Models and Document Clustering

Alexander Hoyle, Lorena Calvo-Bartolomé, Jordan Boyd-Graber, Philip Resnik

机构 * ETH Zürich（苏黎世联邦理工学院）； University of Maryland（马里兰大学）； Universidad Carlos III de Madrid（马德里卡斯蒂利亚-拉曼恰大学）

专题命中后训练与偏好优化：LLM(abstract)；分类 cs.CL

Comments Accepted to ACL 2025 (Main)

2506.02355 2025-06-23 cs.LG 57%

Rewarding the Unlikely: Lifting GRPO Beyond Distribution Sharpening

Andre He, Daniel Fried, Sean Welleck

机构 * Carnegie Mellon University（卡内基梅隆大学）

专题命中后训练与偏好优化：language model(abstract)；分类 cs.LG

2506.14936 2025-06-19 cs.AI 57%

CALM: Contextual Analog Logic with Multimodality

Maxwell J. Jacobson, Corey J. Maley, Yexiang Xue

机构 * Purdue University（普渡大学）

专题命中后训练与偏好优化：LLM(abstract)；分类 cs.AI

2506.08998 2025-06-17 math.ST cs.LG stat.ML stat.TH 57%

On Monotonicity in AI Alignment

Gilles Bareilles, Julien Fageot, Lê-Nguyên Hoang, Peva Blanchard, Wassim Bouaziz, Sébastien Rouault, El-Mahdi El-Mhamdi

机构 * CTU in Prague, Tournesol（布拉格CTU，图尔诺索尔）； Tournesol（图尔诺索尔）； Calicarpa, Tournesol（卡利卡尔帕，图尔诺索尔）； Kleis Technology（克莱斯技术）； École Polytechnique（巴黎高等理工学院）； Calicarpa（卡利卡尔帕）

专题命中后训练与偏好优化：preference optimization(abstract)；分类 cs.LG

2306.13926 2025-06-09 cs.LG 57%

Quantifying the Optimization and Generalization Advantages of Graph Neural Networks Over Multilayer Perceptrons

Wei Huang, Yuan Cao, Haonan Wang, Xin Cao, Taiji Suzuki

机构 * RIKEN AIP（日本学术振兴会先进研究所）； The University of Hong Kong（香港大学）； National University of Singapore（新加坡国立大学）； The University of New South Wales（新南威尔士大学）； University of Tokyo（东京大学）

专题命中后训练与偏好优化：post-training(abstract)；分类 cs.LG

Journal ref AISTATS 2025

2506.03028 2025-06-04 cs.LG q-bio.BM 57%

Protein Inverse Folding From Structure Feedback

Junde Xu, Zijun Gao, Xinyi Zhou, Jie Hu, Xingyi Cheng, Le Song, Guangyong Chen, Pheng-Ann Heng, Jiezhong Qiu

机构 * CUHK（香港中文大学）； Hangzhou Institute of Medicine, CAS（杭州医学院，中国科学院）； Zhejiang Lab（浙江实验室）； MBZUAI

专题命中后训练与偏好优化：preference optimization(abstract)；分类 cs.LG

2505.22257 2025-06-02 cs.LG stat.ML 57%

Revisiting Group Relative Policy Optimization: Insights into On-Policy and Off-Policy Training

Youssef Mroueh, Nicolas Dupuis, Brian Belgodere, Apoorva Nitsure, Mattia Rigotti, Kristjan Greenewald, Jiri Navratil, Jerret Ross, Jesus Rios

机构 * IBM Research（IBM研究院）； IBM Quantum（IBM量子实验室）； MIT-IBM Watson Lab（MIT-IBM沃森实验室）

专题命中后训练与偏好优化：post-training(abstract)；分类 cs.LG

2505.23363 2025-05-30 cs.CL 57%

Discriminative Policy Optimization for Token-Level Reward Models

Hongzhan Chen, Tao Yang, Shiping Gao, Ruijun Chen, Xiaojun Quan, Hongtao Tian, Ting Yao

机构 * Hongzhan Chen（洪湛 Chen）； Tao Yang（杨 Tao）； Shiping Gao（高世平）； Ruijun Chen（陈瑞俊）； Xiaojun Quan（钱小军）； Hongtao Tian（田洪涛）； Ting Yao（姚婷）

专题命中后训练与偏好优化：language model(abstract)；分类 cs.CL

Comments ICML 2025

2505.20966 2025-05-28 cs.CL cs.IR 57%

Personalized Query Auto-Completion for Long and Short-Term Interests with Adaptive Detoxification Generation

Zhibo Wang, Xiaoze Jiang, Zhiheng Qin, Enyun Yu, Han Li

专题命中后训练与偏好优化：preference optimization(abstract)；分类 cs.CL

Comments KDD 2025

2505.11755 2025-05-21 cs.RO cs.AI 57%

Reachability Barrier Networks: Learning Hamilton-Jacobi Solutions for Smooth and Flexible Control Barrier Functions

Matthew Kim, William Sharpless, Hyun Joe Jeong, Sander Tonkens, Somil Bansal, Sylvia Herbert

机构 * University of California, San Diego（加州大学圣地亚哥分校）； Stanford University（斯坦福大学）

专题命中后训练与偏好优化：post-training(abstract)；分类 cs.AI

Comments 15 pages, 7 figures

2505.10218 2025-05-16 cs.CL 57%

RAIDEN-R1: Improving Role-awareness of LLMs via GRPO with Verifiable Reward

Zongsheng Wang, Kaili Sun, Bowen Wu, Qun Yu, Ying Li, Baoxun Wang

机构 * Platform and Content Group, Tencent（腾讯平台与内容部）； School of Software & Microelectronics, Peking University（北京大学软件与微电子学院）

专题命中后训练与偏好优化：LLM(abstract)；分类 cs.CL

2505.03156 2025-05-07 cs.IT cs.AI math.IT 57%

Soft Best-of-n Sampling for Model Alignment

Claudio Mayrink Verdun, Alex Oesterling, Himabindu Lakkaraju, Flavio P. Calmon

机构 * Harvard University（哈佛大学）； Harvard Business School（哈佛商学院）

专题命中后训练与偏好优化：language model(abstract)；分类 cs.AI

Comments Accepted for presentation at the 2025 IEEE International Symposium on Information Theory (ISIT 2025)

2504.20520 2025-04-30 cs.RO cs.AI 57%

PRISM: Projection-based Reward Integration for Scene-Aware Real-to-Sim-to-Real Transfer with Few Demonstrations

Haowen Sun, Han Wang, Chengzhong Ma, Shaolong Zhang, Jiawei Ye, Xingyu Chen, Xuguang Lan

机构 * National Key Laboratory of Human-Machine Hybrid Augmented Intelligence（人机混合增强智能国家级实验室）； Institute of Artificial Intelligence and Robotics（人工智能与机器人研究院）； Xi’an Jiaotong University（西安交通大学）

专题命中后训练与偏好优化：language model(abstract)；分类 cs.AI

2504.20458 2025-04-30 cs.IR cs.CL 57%

Search-Based Interaction For Conversation Recommendation via Generative Reward Model Based Simulated User

Xiaolei Wang, Chunxuan Xia, Junyi Li, Fanzhe Meng, Lei Huang, Jinpeng Wang, Wayne Xin Zhao, Ji-Rong Wen

机构 * Gaoling School of Artificial Intelligence, Renmin University of China（中国人民大学人工智能学院）

专题命中后训练与偏好优化：instruction tuning(abstract)；分类 cs.CL

Comments Accepted by SIGIR 2025

2504.15674 2025-04-23 cs.CR cs.LG 57%

TrojanDam: Detection-Free Backdoor Defense in Federated Learning through Proactive Model Robustification utilizing OOD Data

Yanbo Dai, Songze Li, Zihan Gan, Xueluan Gong

机构 * Southeast University（东南大学）； HKUST(GZ)（香港科技大学（广州））； Nanyang Technological University（南洋理工大学）

专题命中后训练与偏好优化：post-training(abstract)；分类 cs.LG

2504.13630 2025-04-21 cs.CL 57%

Remedy: Learning Machine Translation Evaluation from Human Preferences with Reward Modeling

Shaomu Tan, Christof Monz

机构 * Language Technology Lab University of Amsterdam（语言技术实验室阿姆斯特丹大学）

专题命中后训练与偏好优化：prompting(abstract)；分类 cs.CL

2504.11337 2025-04-16 cs.CL 57%

REWARD CONSISTENCY: Improving Multi-Objective Alignment from a Data-Centric Perspective

Zhihao Xu, Yongqi Tong, Xin Zhang, Jun Zhou, Xiting Wang

专题命中后训练与偏好优化：language model(abstract)；分类 cs.CL

2504.07954 2025-04-11 cs.CV cs.CL 57%

Perception-R1: Pioneering Perception Policy with Reinforcement Learning

En Yu, Kangheng Lin, Liang Zhao, Jisheng Yin, Yana Wei, Yuang Peng, Haoran Wei, Jianjian Sun, Chunrui Han, Zheng Ge, Xiangyu Zhang, Daxin Jiang, Jingyu Wang, Wenbing Tao

专题命中后训练与偏好优化：post-training(abstract)；分类 cs.CL

Comments Github page: https://github.com/linkangheng/PR1

2504.03153 2025-04-07 cs.LG 57%

MORAL: A Multimodal Reinforcement Learning Framework for Decision Making in Autonomous Laboratories

Natalie Tirabassi, Sathish A. P. Kumar, Sumit Jha, Arvind Ramanathan

专题命中后训练与偏好优化：language model(abstract)；分类 cs.LG

Comments 9 pages, 14 figures and 3 tables

2503.02948 2025-03-06 cs.CL cs.IR 57%

ExpertGenQA: Open-ended QA generation in Specialized Domains

Haz Sameen Shahgir, Chansong Lim, Jia Chen, Evangelos E. Papalexakis, Yue Dong

专题命中后训练与偏好优化：LLM(abstract)；分类 cs.CL

2411.11761 2025-02-21 cs.LG cs.HC 57%

Mapping out the Space of Human Feedback for Reinforcement Learning: A Conceptual Framework

Yannick Metz, David Lindner, Raphaël Baur, Mennatallah El-Assady

专题命中后训练与偏好优化：RLHF(abstract)；分类 cs.LG

2502.06812 2025-02-19 cs.LG cs.GR 57%

Harness Local Rewards for Global Benefits: Effective Text-to-Video Generation Alignment with Patch-level Reward Models

Shuting Wang, Haihong Tang, Zhicheng Dou, Chenyan Xiong

专题命中后训练与偏好优化：post-training(abstract)；分类 cs.LG

2411.01696 2025-02-11 cs.LG stat.ML 57%

Conformal Risk Minimization with Variance Reduction

Sima Noorani, Orlando Romero, Nicolo Dal Fabbro, Hamed Hassani, George J. Pappas

专题命中后训练与偏好优化：post-training(abstract)；分类 cs.LG

2502.04751 2025-02-10 cs.IR cs.CL 57%

Holistically Guided Monte Carlo Tree Search for Intricate Information Seeking

Ruiyang Ren, Yuhao Wang, Junyi Li, Jinhao Jiang, Wayne Xin Zhao, Wenjie Wang, Tat-Seng Chua

专题命中后训练与偏好优化：LLM(abstract)；分类 cs.CL

2408.08307 2025-02-07 cs.LG cs.CV 57%

What Secrets Do Your Manifolds Hold? Understanding the Local Geometry of Generative Models

Ahmed Imtiaz Humayun, Ibtihel Amara, Cristina Vasconcelos, Deepak Ramachandran, Candice Schumann, Junfeng He, Katherine Heller, Golnoosh Farnadi, Negar Rostamzadeh, Mohammad Havaei

专题命中后训练与偏好优化：foundation model(abstract)；分类 cs.LG

Comments Accepted for publication at ICLR 2025

2412.19055 2025-01-31 cs.CV cs.LG 57%

SpectralKD: A Unified Framework for Interpreting and Distilling Vision Transformers via Spectral Analysis

Huiyuan Tian, Bonan Xu, Shijian Li, Gang Pan

专题命中后训练与偏好优化：post-training(abstract)；分类 cs.LG

2312.14436 2025-01-22 cs.RO cs.LG 57%

REBEL: Reward Regularization-Based Approach for Robotic Reinforcement Learning from Human Feedback

Souradip Chakraborty, Anukriti Singh, Amisha Bhaskar, Pratap Tokekar, Dinesh Manocha, Amrit Singh Bedi

专题命中后训练与偏好优化：RLHF(abstract)；分类 cs.LG

2410.02711 2025-01-14 cs.LG cond-mat.stat-mech hep-lat 57%

NETS: A Non-Equilibrium Transport Sampler

Michael S. Albergo, Eric Vanden-Eijnden

专题命中后训练与偏好优化：post-training(abstract)；分类 cs.LG

1. 后训练与偏好优化 4541 篇

Mallows Model with Learned Distance Metrics: Sampling and Maximum Likelihood Estimation

Aligned Textual Scoring Rules

ProxAnn: Use-Oriented Evaluations of Topic Models and Document Clustering

Rewarding the Unlikely: Lifting GRPO Beyond Distribution Sharpening

CALM: Contextual Analog Logic with Multimodality

On Monotonicity in AI Alignment

Quantifying the Optimization and Generalization Advantages of Graph Neural Networks Over Multilayer Perceptrons

Protein Inverse Folding From Structure Feedback

Revisiting Group Relative Policy Optimization: Insights into On-Policy and Off-Policy Training

Discriminative Policy Optimization for Token-Level Reward Models

Personalized Query Auto-Completion for Long and Short-Term Interests with Adaptive Detoxification Generation

Reachability Barrier Networks: Learning Hamilton-Jacobi Solutions for Smooth and Flexible Control Barrier Functions

RAIDEN-R1: Improving Role-awareness of LLMs via GRPO with Verifiable Reward

Soft Best-of-n Sampling for Model Alignment

PRISM: Projection-based Reward Integration for Scene-Aware Real-to-Sim-to-Real Transfer with Few Demonstrations

Search-Based Interaction For Conversation Recommendation via Generative Reward Model Based Simulated User

TrojanDam: Detection-Free Backdoor Defense in Federated Learning through Proactive Model Robustification utilizing OOD Data

Remedy: Learning Machine Translation Evaluation from Human Preferences with Reward Modeling

REWARD CONSISTENCY: Improving Multi-Objective Alignment from a Data-Centric Perspective

Perception-R1: Pioneering Perception Policy with Reinforcement Learning

MORAL: A Multimodal Reinforcement Learning Framework for Decision Making in Autonomous Laboratories

ExpertGenQA: Open-ended QA generation in Specialized Domains

Mapping out the Space of Human Feedback for Reinforcement Learning: A Conceptual Framework

Harness Local Rewards for Global Benefits: Effective Text-to-Video Generation Alignment with Patch-level Reward Models

Conformal Risk Minimization with Variance Reduction

Holistically Guided Monte Carlo Tree Search for Intricate Information Seeking

What Secrets Do Your Manifolds Hold? Understanding the Local Geometry of Generative Models

SpectralKD: A Unified Framework for Interpreting and Distilling Vision Transformers via Spectral Analysis

REBEL: Reward Regularization-Based Approach for Robotic Reinforcement Learning from Human Feedback

NETS: A Non-Equilibrium Transport Sampler