CILVR Seminar: Generalization in diffusion models arises from geometry adaptive harmonic representations

Date: Wednesday, October 9, 2024, 2PM
Location: 60FA , Room
Speaker: Zahra Kadkhodaie

Here, we show that two DNNs trained on non-overlapping subsets of a dataset learn nearly the same score function, and thus the same density, when the number of training images is large enough. 

Watch the recording here.