paper-with-me

Papers

Learning to Clarify by Reinforcement Learning Through Reward-Weighted Fine-Tuning

2025-06-08 · Subhojyoti Mukherjee, Viet Dac Lai, Raghavendra Addanki, Ryan Rossi, Seunghyun Yoon, Trung Bui, Anup Rao, Jayakumar Subramanian, Branislav Kveton

Question answering (QA) agents automatically answer questions posed in natural language. In this work, we learn to ask clarifying questions in QA agents. The key idea in our method is to simulate conversations that contain clarifying questions and learn from them using reinforcement learning (RL). To make RL practical, we propose and analyze offline RL objectives that can be viewed as reward-weighted supervised fine-tuning (SFT) and easily optimized in large language models. Our work stands in a stark contrast to recently proposed methods, based on SFT and direct preference optimization, which have additional hyper-parameters and do not directly optimize rewards. We compare to these methods empirically and report gains in both optimized rewards and language quality.

📄 PDF Abstract BibTeX arXiv:2506.06964

Code (0)

등록된 구현이 없습니다.

Tasks

Offline RLQuestion AnsweringReinforcement Learning (RL)

Methods 이 논문이 사용한 방법론

SFT Shrink and Fine-Tune, or SFT, is a type of distillation that avoids explicit distillation by copying parameters to a student student model and then fine-tuning.…

Similar Papers 제목 키워드 기반

Supervised Fine Tuning on Curated Data is Reinforcement Learning (and can be improved)

2025-07-17 · Chongli Qin, Jost Tobias Springenberg

Behavior Cloning (BC) on curated (or filtered) data is the predominant paradigm for supervised fine-tuning (SFT) of large language models; as well as for imitation learning of control policies. Here, we draw on a connect…

continuous-controlContinuous ControlImitation LearningReinforcement Learning (RL)

A Survey of Process Reward Models: From Outcome Signals to Process Supervisions for Large Language Models

2025-10-09 · Congmin Zheng, Jiachen Zhu, Zhuoying Ou, Yuxiang Chen 외 arxiv

Although Large Language Models (LLMs) exhibit advanced reasoning ability, conventional alignment remains largely dominated by outcome reward models (ORMs) that judge only final answers. Process Reward Models(PRMs) addres…

Reinforcement LearningMultimodal Reasoning

Beyond Pairwise Preferences: Listwise Reward-Aware Alignment for Diffusion Models

2026-05-26 · Austin Wang, Jiaqi Han, Stefano Ermon, Yisong Yue arxiv

Preference optimization has emerged as an efficient alternative to online reinforcement learning from human feedback (RLHF) for aligning text-to-image diffusion models. However, existing methods largely reduce supervisio…

Text-to-Image GenerationReinforcement LearningImage Editing

CLARIFY: Contrastive Preference Reinforcement Learning for Untangling Ambiguous Queries

2025-05-31 · Ni Mu, Hao Hu, Xiao Hu, Yiqin Yang 외

Preference-based reinforcement learning (PbRL) bypasses explicit reward engineering by inferring reward functions from human preference comparisons, enabling better alignment with human intentions. However, humans often …

Contrastive Learningreinforcement-learningReinforcement Learning

Towards Understanding Self-play for LLM Reasoning

2025-10-31 · Justin Yang Chae, Md Tanvirul Alam, Nidhi Rastogi arxiv

Recent advances in large language model (LLM) reasoning, led by reinforcement learning with verifiable rewards (RLVR), have inspired self-play post-training, where models improve by generating and solving their own probl…

Reinforcement Learning