Yao Shunyu, a top student from Tsinghua University's Yao Class, recently announced his joining OpenAI. This brief news attracted attention and well-wishes from many industry professionals, including AI professors and investors such as Mark Chen, the head of frontier research at OpenAI.
Yao Shunyu's background includes:
- Graduate of Tsinghua Yao Class
- Co-chair of Yao Class Joint Association
- Co-founder of Tsinghua University Student Rap Society
- Ph.D. in Computer Science from Princeton University
His main research achievements include:
- Tree of Thoughts: A method to improve LLM reasoning abilities
- SWE-bench: A dataset for evaluating large model capabilities
- SWE-agent: An open-source AI programmer
Yao Shunyu's research primarily focuses on Language Agents, which is also the topic of his doctoral dissertation. Language Agents are a new category of intelligent agents that use language models for reasoning and action, enabling digital automation.
It involves three key technologies:
- ReAct: A method combining reasoning and action
- Tree of Thoughts: A tree search-based method to improve reasoning abilities
- CoALA: A conceptual framework for organizing and designing language agents
Besides Language Agents, Yao Shunyu is also interested in the theme of "computational thinking."
Overall, Yao Shunyu's research focuses on using language models to enhance AI systems' reasoning, planning, and decision-making capabilities, laying the foundation for more powerful AI agents. After joining OpenAI, he hopes to transform these research visions into practical applications.