TGPO: Tree-Guided Preference Optimization for Robust Web Agent Reinforcement Learning

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Chen, Ziyuan, Zhao, Zhenghui, Han, Zhangye, Liu, Miancan, Ye, Xianhang, Li, Yiqing, Min, Hongbo, Ren, Jinkui, Zhang, Xiantao, Cao, Guitao
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866912594094194688
author Chen, Ziyuan
Zhao, Zhenghui
Han, Zhangye
Liu, Miancan
Ye, Xianhang
Li, Yiqing
Min, Hongbo
Ren, Jinkui
Zhang, Xiantao
Cao, Guitao
author_facet Chen, Ziyuan
Zhao, Zhenghui
Han, Zhangye
Liu, Miancan
Ye, Xianhang
Li, Yiqing
Min, Hongbo
Ren, Jinkui
Zhang, Xiantao
Cao, Guitao
contents With the rapid advancement of large language models and vision-language models, employing large models as Web Agents has become essential for automated web interaction. However, training Web Agents with reinforcement learning faces critical challenges including credit assignment misallocation, prohibitively high annotation costs, and reward sparsity. To address these issues, we propose Tree-Guided Preference Optimization (TGPO), an offline reinforcement learning framework that proposes a tree-structured trajectory representation merging semantically identical states across trajectories to eliminate label conflicts. Our framework incorporates a Process Reward Model that automatically generates fine-grained rewards through subgoal progress, redundancy detection, and action verification. Additionally, a dynamic weighting mechanism prioritizes high-impact decision points during training. Experiments on Online-Mind2Web and our self-constructed C-WebShop datasets demonstrate that TGPO significantly outperforms existing methods, achieving higher success rates with fewer redundant steps.
format Preprint
id arxiv_https___arxiv_org_abs_2509_14172
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle TGPO: Tree-Guided Preference Optimization for Robust Web Agent Reinforcement Learning
Chen, Ziyuan
Zhao, Zhenghui
Han, Zhangye
Liu, Miancan
Ye, Xianhang
Li, Yiqing
Min, Hongbo
Ren, Jinkui
Zhang, Xiantao
Cao, Guitao
Machine Learning
Artificial Intelligence
With the rapid advancement of large language models and vision-language models, employing large models as Web Agents has become essential for automated web interaction. However, training Web Agents with reinforcement learning faces critical challenges including credit assignment misallocation, prohibitively high annotation costs, and reward sparsity. To address these issues, we propose Tree-Guided Preference Optimization (TGPO), an offline reinforcement learning framework that proposes a tree-structured trajectory representation merging semantically identical states across trajectories to eliminate label conflicts. Our framework incorporates a Process Reward Model that automatically generates fine-grained rewards through subgoal progress, redundancy detection, and action verification. Additionally, a dynamic weighting mechanism prioritizes high-impact decision points during training. Experiments on Online-Mind2Web and our self-constructed C-WebShop datasets demonstrate that TGPO significantly outperforms existing methods, achieving higher success rates with fewer redundant steps.
title TGPO: Tree-Guided Preference Optimization for Robust Web Agent Reinforcement Learning
topic Machine Learning
Artificial Intelligence
url https://arxiv.org/abs/2509.14172