---
id: 20260711-T0-08
title: "DeepSearch-World：让搜索 Agent 在可验证环境中通过自蒸馏自我进化"
title_en: "DeepSearch-World: Self-Distillation for Deep Search Agents in a Verifiable Environment"
url: https://ai.daily.yangsir.net/daily/20260711-T0-08
issue_date: 2026-07-11
publish_date: 2026-07-10T04:00:00.000Z
category: research
source_name: "arXiv cs.CL (NLP)"
source_url: https://arxiv.org/abs/2607.07820
---

# DeepSearch-World：让搜索 Agent 在可验证环境中通过自蒸馏自我进化

一项新研究提出了 DeepSearch-World 框架，解决了工具使用型 Agent 训练的难题。监督微调依赖固定的教师轨迹，而强化学习的奖励稀疏。该方法让 Agent 在可验证的搜索环境中不断尝试，然后从自己的成功轨迹中进行自我蒸馏（self-distillation），从而持续提升性能，无需依赖外部教师模型。

## English Version

**DeepSearch-World: Self-Distillation for Deep Search Agents in a Verifiable Environment**

A new study introduces DeepSearch-World, a framework addressing the challenge of training tool-use agents. Supervised fine-tuning relies on fixed teacher trajectories, while RL suffers from sparse rewards. This method allows agents to explore in a verifiable search environment, then self-distill from their own successful trajectories to continuously improve performance without relying on an external teacher model.

---

**来源**：[arXiv cs.CL (NLP)](https://arxiv.org/abs/2607.07820)

**详情页**：https://ai.daily.yangsir.net/daily/20260711-T0-08

---

*智语观潮 · Daily — https://ai.daily.yangsir.net/llms.txt*