Memorization and Regularization in Generative Diffusion Models

Join Professor Andrew Stuart from CALTECH (California Institute of Technology) for an in-depth seminar exploring memorization and regularization techniques in generative diffusion models. This lecture, part of the "Representing, calibrating & leveraging prediction uncertainty from statistics to machine learning" event series, delves into mathematical approaches to understanding how diffusion models learn and generalize. Taking place on May 15th, 2025, from 10:30 to 11:30 at the Isaac Newton Institute, this presentation offers valuable insights for researchers and practitioners working at the intersection of mathematics, statistics, and machine learning. Discover advanced concepts that help explain the behavior and optimization of these powerful generative models.