ReST-MCTS*: LLM Self-Training via Process Reward Guided Tree Search

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Zhang, Dan, Zhoubian, Sining, Hu, Ziniu, Yue, Yisong, Dong, Yuxiao, Tang, Jie
Format: Preprint
Veröffentlicht: 2024
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866916484490461184
author Zhang, Dan
Zhoubian, Sining
Hu, Ziniu
Yue, Yisong
Dong, Yuxiao
Tang, Jie
author_facet Zhang, Dan
Zhoubian, Sining
Hu, Ziniu
Yue, Yisong
Dong, Yuxiao
Tang, Jie
contents Recent methodologies in LLM self-training mostly rely on LLM generating responses and filtering those with correct output answers as training data. This approach often yields a low-quality fine-tuning training set (e.g., incorrect plans or intermediate reasoning). In this paper, we develop a reinforced self-training approach, called ReST-MCTS*, based on integrating process reward guidance with tree search MCTS* for collecting higher-quality reasoning traces as well as per-step value to train policy and reward models. ReST-MCTS* circumvents the per-step manual annotation typically used to train process rewards by tree-search-based reinforcement learning: Given oracle final correct answers, ReST-MCTS* is able to infer the correct process rewards by estimating the probability this step can help lead to the correct answer. These inferred rewards serve dual purposes: they act as value targets for further refining the process reward model and also facilitate the selection of high-quality traces for policy model self-training. We first show that the tree-search policy in ReST-MCTS* achieves higher accuracy compared with prior LLM reasoning baselines such as Best-of-N and Tree-of-Thought, within the same search budget. We then show that by using traces searched by this tree-search policy as training data, we can continuously enhance the three language models for multiple iterations, and outperform other self-training algorithms such as ReST$^\text{EM}$ and Self-Rewarding LM. We release all code at https://github.com/THUDM/ReST-MCTS.
format Preprint
id arxiv_https___arxiv_org_abs_2406_03816
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle ReST-MCTS*: LLM Self-Training via Process Reward Guided Tree Search
Zhang, Dan
Zhoubian, Sining
Hu, Ziniu
Yue, Yisong
Dong, Yuxiao
Tang, Jie
Computation and Language
Recent methodologies in LLM self-training mostly rely on LLM generating responses and filtering those with correct output answers as training data. This approach often yields a low-quality fine-tuning training set (e.g., incorrect plans or intermediate reasoning). In this paper, we develop a reinforced self-training approach, called ReST-MCTS*, based on integrating process reward guidance with tree search MCTS* for collecting higher-quality reasoning traces as well as per-step value to train policy and reward models. ReST-MCTS* circumvents the per-step manual annotation typically used to train process rewards by tree-search-based reinforcement learning: Given oracle final correct answers, ReST-MCTS* is able to infer the correct process rewards by estimating the probability this step can help lead to the correct answer. These inferred rewards serve dual purposes: they act as value targets for further refining the process reward model and also facilitate the selection of high-quality traces for policy model self-training. We first show that the tree-search policy in ReST-MCTS* achieves higher accuracy compared with prior LLM reasoning baselines such as Best-of-N and Tree-of-Thought, within the same search budget. We then show that by using traces searched by this tree-search policy as training data, we can continuously enhance the three language models for multiple iterations, and outperform other self-training algorithms such as ReST$^\text{EM}$ and Self-Rewarding LM. We release all code at https://github.com/THUDM/ReST-MCTS.
title ReST-MCTS*: LLM Self-Training via Process Reward Guided Tree Search
topic Computation and Language
url https://arxiv.org/abs/2406.03816