Memorization and Regularization in Generative Diffusion Models

Add to your list(s) Download to your calendar using vCal

• Andrew Stuart (CALTECH (California Institute of Technology))
• Thursday 15 May 2025, 10:30-11:30
• Seminar Room 2, Newton Institute.

If you have a question about this talk, please contact nobody.

RCL - Representing, calibrating & leveraging prediction uncertainty from statistics to machine learning

Ricardo Baptista, Agnimitra Dasgupta, Nikola B. Kovachki, Assad Oberai, Andrew M. Stuart Diffusion models have emerged as a powerful framework for generative modeling. At the heart of the methodology is score matching: learning gradients of families of log-densities for noisy versions of the data distribution at different scales. When the loss function adopted in score matching is evaluated using empirical data, rather than the population loss, the minimizer corresponds to the score of a time-dependent Gaussian mixture. However, use of this analytically tractable minimizer leads to data memorization: in both unconditioned and conditioned settings, the generative model returns the training samples. This talk explains the dynamical mechanism underlying memorization. The analysis highlights the need for regularization to avoid reproducing the analytically tractable minimizer; and, in so doing, lays the foundations for a principled understanding of how to regularize. Numerical experiments investigate the properties of: (i) Tikhonov regularization; (ii) regularization designed to promote asymptotic consistency; and (iii) regularizations induced by under-parameterization of a neural network or by early stopping when training a neural network. These experiments are evaluated in the context of memorization, and directions for future development of regularization are highlighted.

This talk is part of the Isaac Newton Institute Seminar Series series.

This talk is included in these lists:

• All CMS events
• Featured lists
• INI info aggregator
• Isaac Newton Institute Seminar Series
• School of Physical Sciences
• Seminar Room 2, Newton Institute
• bld31

Note that ex-directory lists are not shown.