郭大亚 Daya GuoDaya Guo
DeepSeek
DeepSeek-R1 第一作者DeepSeek-R1 First Author
nlp
香港中文大学（深圳）博士。DeepSeek 核心研究员。DeepSeek-R1（2025）第一作者，该推理模型证明了强化学习可以在无需监督微调的情况下教会 LLM 推理，并以极低成本达到与 OpenAI o1 竞争的性能。同时是 DeepSeek-Coder 和 DeepSeek-V2/V3 的关键贡献者。
Chinese University of Hong Kong (Shenzhen) PhD. Core researcher at DeepSeek. First author of DeepSeek-R1 (2025), the reasoning model that showed reinforcement learning could teach LLMs to reason without supervised fine-tuning, reaching performance competitive with OpenAI o1 at much lower cost. Also key contributor to DeepSeek-Coder and DeepSeek-V2/V3.
在关系图谱中查看 郭大亚 →View Daya Guo in the graph →
时间线Timeline
- 2025 DeepSeek-R1: Incentivizing Reasoning in LLMs via RL
- 2024 DeepSeek-V2: A Strong, Economical MoE LLM
- 2024 DeepSeek-Coder-V2