Shengran Hu

hsr.jpg

Hello! I am Shengran Hu (胡圣然). I am a PhD student at University of British Columbia, advised by Dr. Jeff Clune. I am currently interning at Sakana AI as a research scientist.

I want to understand the emergence of complexity, such as natural evolution and human scientific discoveries. I study this by building open-ended AI systems that can accumulate complexity in language.

Email / Google Scholar / Github / Twitter

news

Apr 10, 2025 Introducing The AI Scientist-v2, which produced the 1st fully AI-generated paper to pass peer review at a workshop level!
Feb 1, 2025 We introduce Automated Capability Discovery (ACD). ACD automatically discovers surprising new capabilities and failure modes in foundation models via self-exploration.
Jan 23, 2025 Two of our works Automated Design of Agentic Systems and Intelligent Go-Explore were accepted to ICLR 2025. Thank you to my incredible collaborators Cong and Jeff!
Dec 16, 2024 Our work Automated Design of Agentic Systems got Outstanding Paper at NeurIPS 2024 OWA Workshop and Oral at LanGame Workshop!
Aug 19, 2024 Introducing Automated Design of Agentic Systems (ADAS)! [tweet thread].

selected publications

  1. ADAS.png
    Automated Design of Agentic Systems
    Shengran HuCong Lu, and Jeff Clune
    International Conference on Learning Representations (ICLR), 2025
    🏆 Outstanding Paper (NeurIPS 2024 Open-World Agent Workshop)
  2. ACD.png
    Automated Capability Discovery via Foundation Model Self-Exploration
    Cong Lu*Shengran Hu*, and Jeff Clune
    arXiv preprint arXiv:2502.07577 , 2025
  3. AISCI-v2.png
    The AI Scientist-v2: Workshop-Level Automated Scientific Discovery via Agentic Tree Search
    Yutaro YamadaRobert Tjarko LangeCong LuShengran Hu, and 4 more authors
    arXiv preprint arXiv:2504.08066 , 2025
  4. ige_framework.png
    Intelligent Go-Explore: Standing on the Shoulders of Giant Foundation Models
    Cong LuShengran Hu, and Jeff Clune
    International Conference on Learning Representations (ICLR), 2025
  5. Thought_Cloning.gif
    Thought Cloning: Learning to Think while Acting by Imitating Human Thinking
    Shengran Hu, and Jeff Clune
    Advances in Neural Information Processing Systems (NeurIPS), 2023
    Spotlight (top 3.1% in 12,343)
  6. emo2021.jpg
    Multi-objective Neural Architecture Search with Almost No Training
    Shengran Hu, Ran Cheng, Cheng He, and Zhichao Lu
    In Evolutionary Multi-Criterion Optimization , 2021