Junhua Chen

junhua.jpg

I enjoy solving problems, especially if they involve math and computers. Currently, I work in machine learning, with a particular emphasis on generative modelling and its applications. Problems that interest me include:

  • What are the most effective approaches for modelling different (multi)-modalities of data — diffusion, autoregressive, or more novel alternatives?

  • How can generative models be fine-tuned, adapted, and steered to address practical objectives?

  • How can we algorithmically improve the efficiency and scalability of generative models?

I completed my studies in mathematics and computer science at Trinity College, Cambridge University. I also work in quantitative finance.

news

Sep 09, 2025 New Website in development!

latest posts

Dec 21, 2020 On Tree graphs

selected publications

  1. Sequential Controlled Langevin Diffusions
    Junhua Chen*, Lorenz Richter*, Julius Berner*, Denis Blessing*, Gerhard Neumann, and Anima Anandkumar
    In ICLR 2025, 2025
  2. Straight but not so fast: Challenges with Rectified Flows in Protein Design.
    Junhua Chen, Simon V Mathis, Charles Harris, Kieran Didi, and Pietro Lio
    In ICML 2025 Generative AI and Biology (GenBio) Workshop, 2025