AceSearcher: Bootstrapping Reasoning and Search for LLMs via Reinforced Self-Play

Fuente: arXiv
Enregistré dans:
Détails bibliographiques
Auteurs principaux: Xu, Ran, Zhuang, Yuchen, Dong, Zihan, Wang, Jonathan, Yu, Yue, Ho, Joyce C., Zhang, Linjun, Wang, Haoyu, Shi, Wenqi, Yang, Carl
Format: Preprint
Publié: 2025
Sujets:
Accès en ligne:
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866915520253526016
author Xu, Ran
Zhuang, Yuchen
Dong, Zihan
Wang, Jonathan
Yu, Yue
Ho, Joyce C.
Zhang, Linjun
Wang, Haoyu
Shi, Wenqi
Yang, Carl
author_facet Xu, Ran
Zhuang, Yuchen
Dong, Zihan
Wang, Jonathan
Yu, Yue
Ho, Joyce C.
Zhang, Linjun
Wang, Haoyu
Shi, Wenqi
Yang, Carl
contents Search-augmented LLMs often struggle with complex reasoning tasks due to ineffective multi-hop retrieval and limited reasoning ability. We propose AceSearcher, a cooperative self-play framework that trains a single large language model (LLM) to alternate between two roles: a decomposer that breaks down complex queries and a solver that integrates retrieved contexts for answer generation. AceSearcher couples supervised fine-tuning on a diverse mixture of search, reasoning, and decomposition tasks with reinforcement fine-tuning optimized for final answer accuracy, eliminating the need for intermediate annotations. Extensive experiments on three reasoning-intensive tasks across 10 datasets show that AceSearcher outperforms state-of-the-art baselines, achieving an average exact match improvement of 7.6%. Remarkably, on document-level finance reasoning tasks, AceSearcher-32B matches the performance of the DeepSeek-V3 model using less than 5% of its parameters. Even at smaller scales (1.5B and 8B), AceSearcher often surpasses existing search-augmented LLMs with up to 9x more parameters, highlighting its exceptional efficiency and effectiveness in tackling complex reasoning tasks. Our code will be published at https://github.com/ritaranx/AceSearcher and https://huggingface.co/AceSearcher.
format Preprint
id arxiv_https___arxiv_org_abs_2509_24193
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle AceSearcher: Bootstrapping Reasoning and Search for LLMs via Reinforced Self-Play
Xu, Ran
Zhuang, Yuchen
Dong, Zihan
Wang, Jonathan
Yu, Yue
Ho, Joyce C.
Zhang, Linjun
Wang, Haoyu
Shi, Wenqi
Yang, Carl
Computation and Language
Artificial Intelligence
Information Retrieval
Machine Learning
Search-augmented LLMs often struggle with complex reasoning tasks due to ineffective multi-hop retrieval and limited reasoning ability. We propose AceSearcher, a cooperative self-play framework that trains a single large language model (LLM) to alternate between two roles: a decomposer that breaks down complex queries and a solver that integrates retrieved contexts for answer generation. AceSearcher couples supervised fine-tuning on a diverse mixture of search, reasoning, and decomposition tasks with reinforcement fine-tuning optimized for final answer accuracy, eliminating the need for intermediate annotations. Extensive experiments on three reasoning-intensive tasks across 10 datasets show that AceSearcher outperforms state-of-the-art baselines, achieving an average exact match improvement of 7.6%. Remarkably, on document-level finance reasoning tasks, AceSearcher-32B matches the performance of the DeepSeek-V3 model using less than 5% of its parameters. Even at smaller scales (1.5B and 8B), AceSearcher often surpasses existing search-augmented LLMs with up to 9x more parameters, highlighting its exceptional efficiency and effectiveness in tackling complex reasoning tasks. Our code will be published at https://github.com/ritaranx/AceSearcher and https://huggingface.co/AceSearcher.
title AceSearcher: Bootstrapping Reasoning and Search for LLMs via Reinforced Self-Play
topic Computation and Language
Artificial Intelligence
Information Retrieval
Machine Learning
url https://arxiv.org/abs/2509.24193