paper-with-me

홈 › Papers

RL-finetuning LLMs from on- and off-policy data with a single algorithm

2025-03-25 · Yunhao Tang, Taco Cohen, David W. Zhang, Michal Valko, Rémi Munos

We introduce a novel reinforcement learning algorithm (AGRO, for Any-Generation Reward Optimization) for fine-tuning large-language models. AGRO leverages the concept of generation consistency, which states that the optimal policy satisfies the notion of consistency across any possible generation of the model. We derive algorithms that find optimal solutions via the sample-based policy gradient and provide theoretical guarantees on their convergence. Our experiments demonstrate the effectiveness of AGRO in both on-policy and off-policy settings, showing improved performance on the mathematical reasoning dataset over baseline algorithms.

📄 PDF Abstract BibTeX arXiv:2503.19612

Code (0)

등록된 구현이 없습니다.

Tasks

Mathematical Reasoning

Similar Papers 제목 키워드 기반

Policy Finetuning: Bridging Sample-Efficient Offline and Online Reinforcement Learning

2021-06-09 · NeurIPS 2021 12 · Tengyang Xie, Nan Jiang, Huan Wang, Caiming Xiong 외

Recent theoretical work studies sample-efficient reinforcement learning (RL) extensively in two settings: learning interactively in the environment (online RL), or learning from an offline dataset (offline RL). However, …

Offline RLOpen-Ended Question Answeringreinforcement-learningReinforcement Learning+1

Contrastive Policy Gradient: Aligning LLMs on sequence-level scores in a supervised-friendly fashion

2024-06-27 · Yannis Flet-Berliac, Nathan Grinsztajn, Florian Strub, Bill Wu 외

Reinforcement Learning (RL) has been used to finetune Large Language Models (LLMs) using a reward model trained from preference data, to better align with human judgment. The recently introduced direct alignment methods,…

Code GenerationNatural Language InferenceReinforcement Learning (RL)

Policy Finetuning in Reinforcement Learning via Design of Experiments using Offline Data

2023-07-10 · NeurIPS 2023 11

In some applications of reinforcement learning, a dataset of pre-collected experience is already available but it is also possible to acquire some additional online data to help improve the quality of the policy. However…

Posterior Behavioral Cloning: Pretraining BC Policies for Efficient RL Finetuning

2025-12-18 · Andrew Wagenmaker, Perry Dong, Raymond Tsao, Chelsea Finn 외 arxiv

Standard practice across domains from robotics to language is to first pretrain a policy on a large-scale demonstration dataset, and then finetune this policy, typically with reinforcement learning (RL), in order to impr…

Reinforcement Learning

Finetuning from Offline Reinforcement Learning: Challenges, Trade-offs and Practical Solutions

2023-03-30 · Yicheng Luo, Jackie Kay, Edward Grefenstette, Marc Peter Deisenroth

Offline reinforcement learning (RL) allows for the training of competent agents from offline datasets without any interaction with the environment. Online finetuning of such offline models can further improve performance…

DiversityOffline RLreinforcement-learningReinforcement Learning (RL)