通义深度研究:开源AI研究新范式 episode artwork

EPISODE · Sep 21, 2025 · 24 MIN

通义深度研究:开源AI研究新范式

from Daily LLM Papers

通义 DeepResearch:开源 AI 智能体的新纪元阿里巴巴通义实验室在开发自主信息搜索和推理代理模型方面的进展,集中于三个相关的项目:Tongyi DeepResearch、WebDancer 和 WebSailor。这些研究的核心目标是构建能够执行复杂、多步骤任务的大型语言模型(LLMs),方法包括生成高质量、高不确定性的训练数据,例如通过构建知识图谱和模糊信息来合成问题(SailorFog-QA, CRAWLQA),以及使用如 ReAct 框架和强化学习(RL)等后训练方法。特别是,WebSailor 引入了 DUPO 算法来提高 RL 训练的效率和有效性,而 WebWatcher 则将代理的能力扩展到了多模态(Vision-Language, VL)推理任务,并在 BrowseComp-VL 和 HLE 等挑战性基准测试中展现出超越现有开源和专有系统的性能。前往小宇宙评论区与主播互动

Episode metadata supplied by the publisher feed · Published Sep 21, 2025

Embed this episode

Ready to play

通义深度研究:开源AI研究新范式

0:00 24:25

No transcript for this episode yet

We transcribe on demand. Request one and we'll notify you when it's ready — usually under 10 minutes.

No similar episodes found.

No similar podcasts found.

Frequently Asked Questions

How long is this episode of Daily LLM Papers?

This episode is 24 minutes long.

When was this Daily LLM Papers episode published?

This episode was published on September 21, 2025.

Can I download this Daily LLM Papers episode?

Yes. Use the download control on the episode player to save the publisher-provided media file.
URL copied to clipboard!