Plan-MCTS: Plan Exploration for Action Exploitation in Web Navigation
Fuente:
arXiv
Guardado en:
| Autores principales: | , , , , , , , , , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2026
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866915799607803904 |
|---|---|
| author | Zhang, Weiming Wang, Jihong Zhou, Jiamu Li, Qingyao Ma, Xinbei Zheng, Congmin Lou, Xingyu Liu, Weiwen Zhang, Zhuosheng Wang, Jun Yu, Yong Zhang, Weinan |
| author_facet | Zhang, Weiming Wang, Jihong Zhou, Jiamu Li, Qingyao Ma, Xinbei Zheng, Congmin Lou, Xingyu Liu, Weiwen Zhang, Zhuosheng Wang, Jun Yu, Yong Zhang, Weinan |
| contents | Large Language Models (LLMs) have empowered autonomous agents to handle complex web navigation tasks. While recent studies integrate tree search to enhance long-horizon reasoning, applying these algorithms in web navigation faces two critical challenges: sparse valid paths that lead to inefficient exploration, and a noisy context that dilutes accurate state perception. To address this, we introduce Plan-MCTS, a framework that reformulates web navigation by shifting exploration to a semantic Plan Space. By decoupling strategic planning from execution grounding, it transforms sparse action space into a Dense Plan Tree for efficient exploration, and distills noisy contexts into an Abstracted Semantic History for precise state awareness. To ensure efficiency and robustness, Plan-MCTS incorporates a Dual-Gating Reward to strictly validate both physical executability and strategic alignment and Structural Refinement for on-policy repair of failed subplans. Extensive experiments on WebArena demonstrate that Plan-MCTS achieves state-of-the-art performance, surpassing current approaches with higher task effectiveness and search efficiency. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2602_14083 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Plan-MCTS: Plan Exploration for Action Exploitation in Web Navigation Zhang, Weiming Wang, Jihong Zhou, Jiamu Li, Qingyao Ma, Xinbei Zheng, Congmin Lou, Xingyu Liu, Weiwen Zhang, Zhuosheng Wang, Jun Yu, Yong Zhang, Weinan Artificial Intelligence Large Language Models (LLMs) have empowered autonomous agents to handle complex web navigation tasks. While recent studies integrate tree search to enhance long-horizon reasoning, applying these algorithms in web navigation faces two critical challenges: sparse valid paths that lead to inefficient exploration, and a noisy context that dilutes accurate state perception. To address this, we introduce Plan-MCTS, a framework that reformulates web navigation by shifting exploration to a semantic Plan Space. By decoupling strategic planning from execution grounding, it transforms sparse action space into a Dense Plan Tree for efficient exploration, and distills noisy contexts into an Abstracted Semantic History for precise state awareness. To ensure efficiency and robustness, Plan-MCTS incorporates a Dual-Gating Reward to strictly validate both physical executability and strategic alignment and Structural Refinement for on-policy repair of failed subplans. Extensive experiments on WebArena demonstrate that Plan-MCTS achieves state-of-the-art performance, surpassing current approaches with higher task effectiveness and search efficiency. |
| title | Plan-MCTS: Plan Exploration for Action Exploitation in Web Navigation |
| topic | Artificial Intelligence |
| url | https://arxiv.org/abs/2602.14083 |