OpAgent：网页导航的运算代理
OpAgent: Operator Agent for Web Navigation
AI总结：
本文提出OpAgent，通过在线强化学习优化网页导航策略，结合层级多任务微调和混合奖励机制，实现71.6%的高成功率。
摘要：
为了满足用户指令，自主网页代理必须应对现实网站的复杂性和波动性。传统方法主要依赖监督微调或离线强化学习，但存在分布偏移问题。本文提出一种鲁棒的在线强化学习网页代理，通过直接迭代交互优化策略。核心创新包括：1）层级多任务微调，构建具备强指令遵循能力的视觉语言模型；2）在线代理强化学习，结合WebJudge和决策树的混合奖励机制；3）运算代理框架，通过规划、定位、反思和总结模块实现鲁棒误差恢复，达到71.6%的SOTA成功率。
英文摘要：
To fulfill user instructions, autonomous web agents must contend with the inherent complexity and volatile nature of real-world websites. Conventional paradigms predominantly rely on Supervised Fine-Tuning (SFT) or Offline Reinforcement Learning (RL) using static datasets. However, these methods suffer from severe distributional shifts, as offline trajectories fail to capture the stochastic state transitions and real-time feedback of unconstrained wide web environments. In this paper, we propose a robust Online Reinforcement Learning WebAgent, designed to optimize its policy through direct, iterative interactions with unconstrained wide websites. Our approach comprises three core innovations: 1) Hierarchical Multi-Task Fine-tuning: We curate a comprehensive mixture of datasets categorized by functional primitives -- Planning, Acting, and Grounding -- establishing a Vision-Language Model (VLM) with strong instruction-following capabilities for Web GUI tasks. 2) Online Agentic RL in the Wild: We develop an online interaction environment and fine-tune the VLM using a specialized RL pipeline. We introduce a Hybrid Reward Mechanism that combines a ground-truth-agnostic WebJudge for holistic outcome assessment with a Rule-based Decision Tree (RDT) for progress reward. This system effectively mitigates the credit assignment challenge in long-horizon navigation. Notably, our RL-enhanced model achieves a 38.1\% success rate (pass@5) on WebArena, outperforming all existing monolithic baselines. 3) Operator Agent: We introduce a modular agentic framework, namely \textbf{OpAgent}, orchestrating a Planner, Grounder, Reflector, and Summarizer. This synergy enables robust error recovery and self-correction, elevating the agent's performance to a new State-of-the-Art (SOTA) success rate of \textbf{71.6\%}.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
