ReSearch: Learning to Reason with Search for LLMs via Reinforcement Learning

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Chen, Mingyang, Sun, Linzhuang, Li, Tianpeng, Sun, Haoze, Zhou, Yijie, Zhu, Chenzheng, Wang, Haofen, Pan, Jeff Z., Zhang, Wen, Chen, Huajun, Yang, Fan, Zhou, Zenan, Chen, Weipeng
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866909800970846208
author Chen, Mingyang
Sun, Linzhuang
Li, Tianpeng
Sun, Haoze
Zhou, Yijie
Zhu, Chenzheng
Wang, Haofen
Pan, Jeff Z.
Zhang, Wen
Chen, Huajun
Yang, Fan
Zhou, Zenan
Chen, Weipeng
author_facet Chen, Mingyang
Sun, Linzhuang
Li, Tianpeng
Sun, Haoze
Zhou, Yijie
Zhu, Chenzheng
Wang, Haofen
Pan, Jeff Z.
Zhang, Wen
Chen, Huajun
Yang, Fan
Zhou, Zenan
Chen, Weipeng
contents Large Language Models (LLMs) have shown remarkable capabilities in reasoning, exemplified by the success of OpenAI-o1 and DeepSeek-R1. However, integrating reasoning with external search processes remains challenging, especially for complex multi-hop questions requiring multiple retrieval steps. We propose ReSearch, a novel framework that trains LLMs to Reason with Search via reinforcement learning without using any supervised data on reasoning steps. Our approach treats search operations as integral components of the reasoning chain, where when and how to perform searches is guided by text-based thinking, and search results subsequently influence further reasoning. We train ReSearch on Qwen2.5-7B(-Instruct) and Qwen2.5-32B(-Instruct) models and conduct extensive experiments. Despite being trained on only one dataset, our models demonstrate strong generalizability across various benchmarks. Analysis reveals that ReSearch naturally elicits advanced reasoning capabilities such as reflection and self-correction during the reinforcement learning process.
format Preprint
id arxiv_https___arxiv_org_abs_2503_19470
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle ReSearch: Learning to Reason with Search for LLMs via Reinforcement Learning
Chen, Mingyang
Sun, Linzhuang
Li, Tianpeng
Sun, Haoze
Zhou, Yijie
Zhu, Chenzheng
Wang, Haofen
Pan, Jeff Z.
Zhang, Wen
Chen, Huajun
Yang, Fan
Zhou, Zenan
Chen, Weipeng
Artificial Intelligence
Computation and Language
Large Language Models (LLMs) have shown remarkable capabilities in reasoning, exemplified by the success of OpenAI-o1 and DeepSeek-R1. However, integrating reasoning with external search processes remains challenging, especially for complex multi-hop questions requiring multiple retrieval steps. We propose ReSearch, a novel framework that trains LLMs to Reason with Search via reinforcement learning without using any supervised data on reasoning steps. Our approach treats search operations as integral components of the reasoning chain, where when and how to perform searches is guided by text-based thinking, and search results subsequently influence further reasoning. We train ReSearch on Qwen2.5-7B(-Instruct) and Qwen2.5-32B(-Instruct) models and conduct extensive experiments. Despite being trained on only one dataset, our models demonstrate strong generalizability across various benchmarks. Analysis reveals that ReSearch naturally elicits advanced reasoning capabilities such as reflection and self-correction during the reinforcement learning process.
title ReSearch: Learning to Reason with Search for LLMs via Reinforcement Learning
topic Artificial Intelligence
Computation and Language
url https://arxiv.org/abs/2503.19470