Information Gain-based Policy Optimization: A Simple and Effective Approach for Multi-Turn Search Agents

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Wang, Guoqing, Dai, Sunhao, Ye, Guangze, Gan, Zeyu, Yao, Wei, Deng, Yong, Wu, Xiaofeng, Ying, Zhenzhe
Formato: Preprint
Publicado: 2025
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866915885092962304
author Wang, Guoqing
Dai, Sunhao
Ye, Guangze
Gan, Zeyu
Yao, Wei
Deng, Yong
Wu, Xiaofeng
Ying, Zhenzhe
author_facet Wang, Guoqing
Dai, Sunhao
Ye, Guangze
Gan, Zeyu
Yao, Wei
Deng, Yong
Wu, Xiaofeng
Ying, Zhenzhe
contents Large language model (LLM)-based agents are increasingly trained with reinforcement learning (RL) to enhance their ability to interact with external environments through tool use, particularly in search-based settings that require multi-turn reasoning and knowledge acquisition. However, existing approaches typically rely on outcome-based rewards that are only provided exclusively upon generating the final answer. This reward sparsity becomes particularly problematic in multi-turn settings, where long trajectories exacerbate three critical issues: (i) advantage collapse, where all rollouts receive identical rewards and provide no useful learning signals; (ii) lack of fine-grained credit assignment, where the correctness of intermediate turns is obscured, especially in long-horizon tasks; and (iii) poor sample efficiency, where each rollout yields only a single outcome signal, leading to low data utilization. In this paper, we propose Information Gain-based Policy Optimization (IGPO), a simple yet effective RL framework that provides dense and intrinsic supervision for multi-turn agent training. IGPO models each interaction turn as an incremental process of acquiring information about the ground truth, and defines turn-level rewards as the marginal increase in the policy's probability of producing the correct answer. Unlike prior process-level reward approaches that depend on external reward models or costly Monte Carlo estimation, IGPO derives intrinsic rewards directly from the model's own belief updates. These intrinsic turn-level rewards are combined with outcome-level supervision to form dense reward signals. Extensive experiments on both in-domain and out-of-domain benchmarks demonstrate that IGPO consistently outperforms strong baselines in multi-turn scenarios, achieving higher accuracy and improved data efficiency. Our code is available at https://github.com/GuoqingWang1/IGPO.
format Preprint
id arxiv_https___arxiv_org_abs_2510_14967
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Information Gain-based Policy Optimization: A Simple and Effective Approach for Multi-Turn Search Agents
Wang, Guoqing
Dai, Sunhao
Ye, Guangze
Gan, Zeyu
Yao, Wei
Deng, Yong
Wu, Xiaofeng
Ying, Zhenzhe
Computation and Language
Artificial Intelligence
Machine Learning
Large language model (LLM)-based agents are increasingly trained with reinforcement learning (RL) to enhance their ability to interact with external environments through tool use, particularly in search-based settings that require multi-turn reasoning and knowledge acquisition. However, existing approaches typically rely on outcome-based rewards that are only provided exclusively upon generating the final answer. This reward sparsity becomes particularly problematic in multi-turn settings, where long trajectories exacerbate three critical issues: (i) advantage collapse, where all rollouts receive identical rewards and provide no useful learning signals; (ii) lack of fine-grained credit assignment, where the correctness of intermediate turns is obscured, especially in long-horizon tasks; and (iii) poor sample efficiency, where each rollout yields only a single outcome signal, leading to low data utilization. In this paper, we propose Information Gain-based Policy Optimization (IGPO), a simple yet effective RL framework that provides dense and intrinsic supervision for multi-turn agent training. IGPO models each interaction turn as an incremental process of acquiring information about the ground truth, and defines turn-level rewards as the marginal increase in the policy's probability of producing the correct answer. Unlike prior process-level reward approaches that depend on external reward models or costly Monte Carlo estimation, IGPO derives intrinsic rewards directly from the model's own belief updates. These intrinsic turn-level rewards are combined with outcome-level supervision to form dense reward signals. Extensive experiments on both in-domain and out-of-domain benchmarks demonstrate that IGPO consistently outperforms strong baselines in multi-turn scenarios, achieving higher accuracy and improved data efficiency. Our code is available at https://github.com/GuoqingWang1/IGPO.
title Information Gain-based Policy Optimization: A Simple and Effective Approach for Multi-Turn Search Agents
topic Computation and Language
Artificial Intelligence
Machine Learning
url https://arxiv.org/abs/2510.14967