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 in generative diffusion models. This 78-minute talk, scheduled for May 15th, 2025 from 10:30 to 11:30, is part of the "Representing, calibrating & leveraging prediction uncertainty from statistics to machine learning" (RCL) event series. Discover key insights into how diffusion models store and process information, and learn about regularization techniques that impact their performance. The seminar will take place at INI Seminar Room 2, offering attendees the opportunity to engage with cutting-edge research at the intersection of machine learning, statistics, and uncertainty quantification.