自蒸馏框架让网页搜索智能体在可验证环境中自我进化，性能媲美开源模型。
DeepSearch-World: Self-Distillation for Deep Search Agents in a Verifiable Environment
- 用自蒸馏迭代训练，结合轨迹生成与过滤提升智能体能力。
- 在3个基准上表现优异：浏览任务31.2%，GAIA 61.5%，HotpotQA 93.4%。
- 适合研究长期交互、自我改进的深度搜索智能体的开发者与学者。
训练工具使用智能体从自身经验中持续改进仍具挑战性，因监督微调依赖固定教师轨迹，而稀疏奖励强化学习对长时程交互提供弱监督。本文提出DeepSearch-Evolve，一种基于DeepSearch-World（一个确定性且可验证的环境）的自蒸馏框架。该环境包含420K多跳问答任务，由实体级随机游走构建，支持进度验证、有依据反思和失败恢复等认知行为。DeepSearch-Evolve通过迭代生成轨迹、筛选、数据混合与微调，逐步训练更强智能体。无需来自更优模型的蒸馏，DeepSearch-World-9B已达到与开源模型相当的性能：在BrowseComp上31.2%，GAIA上61.5%，HotpotQA上93.4%。结果表明，可验证环境可实现长时程网页智能体的大规模自我演进。我们将发布环境、420K训练集、验证集、模型及代码，推动自进化深搜智能体研究。
原文摘要 · Abstract (English)
Training tool-use agents to improve from their own experience remains challenging, as supervised fine-tuning relies on fixed teacher-distilled trajectories, while sparse-reward reinforcement learning provides weak supervision for long-horizon interactions. We present DeepSearch-Evolve, a self-distillation framework for web agents built on DeepSearch-World, a deterministic and verifiable environment with reproducible search and page-reading tools. DeepSearch-World contains 420K multi-hop QA tasks constructed from entity-level random walks and supports key agentic cognitive behaviors useful for self-evolving, including progress verification, grounded reflection, and failure recovery. DeepSearch-Evolve iteratively performs trajectory generation, filtering, data mixing, and fine-tuning to train stronger agents. Without distillation from more capable models, DeepSearch-World-9B achieves competitive performance compared with open-source agents, reaching 31.2% on BrowseComp, 61.5% on GAIA, and 93.4% on HotpotQA, showing that verifiable environments enable scalable self-evolution for long-horizon web agents. We will release the environment, 420K training pool, validation set, model, and code to facilitate future research on self-improving deep search agents.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。