prof_pic.jpg

Aaron Lou

I lead the Strategic Explorations team at OpenAI, where we are pioneering the next frontier of large language models.

We are looking to recruit researchers and engineers interested in working on the bleeding edge of model development. If you think you may be a good fit, feel free to reach out to me at aaronlou@openai.com.

I was a CS PhD at Stanford advised by Stefano Ermon (currently on leave).

selected publications

  1. sedd.gif
    Discrete Diffusion Modeling by Estimating the Ratios of the Data Distribution
    Aaron Lou, Chenlin Meng, and Stefano Ermon
    In International Conference on Machine Learning, 2024
    Best Paper Award
    arXiv Blog Code
  2. evo.jpg
    Sequence Modeling and Design from Molecular to Genome Scale with Evo
    Eric Nguyen, Michael Poli, Matthew G. Durrant, Brian Kang, Dhruva Katrekar, David B. Li, Liam J. Bartie, Armin W. Thomas, Samuel H. King, Garyk Brixi, Jeremy Sullivan, Madelena Y. Ng, Ashley Lewis, Aaron Lou, Stefano Ermon, Stephen A. Baccus, Tina Hernandez-Boussard, Christopher Ré, Patrick D. Hsu, and Brian L. Hie
    In Science, 2024
    Cover Feature
    bioRXiv Code
  3. diffusiondpo.jpg
    Diffusion Model Alignment Using Direct Preference Optimization
    Bram Wallace, Meihua Dang, Rafael Rafailov, Linqi Zhou, Aaron Lou, Senthil Purushwalkam, Stefano Ermon, Caiming Xiong, Shafiq Joty, and Nikhil Naik
    In Conference on Computer Vision and Pattern Recognition, 2024
    arXiv Code
  4. refldiff.gif
    Reflected Diffusion Models
    Aaron Lou, and Stefano Ermon
    In International Conference on Machine Learning, 2023
    arXiv Blog Code
  5. nmode.gif
    Neural Manifold Ordinary Differential Equations
    Aaron Lou, Derek Lim, Isay Katsman, Leo Huang, Qingxuan Jiang, Ser-Nam Lim, and Christopher De Sa
    In Neural Information Processing Systems, 2020
    arXiv Code
  6. frechet.png
    Differentiating through the Fréchet Mean
    Aaron Lou, Isay Katsman, Qingxuan Jiang, Serge Belongie, Ser-Nam Lim, and Christopher De Sa
    In International Conference on Machine Learning, 2020
    arXiv Code