Learning Diffusion Priors from Observations by Expectation Maximization

François Rozet, Gérôme Andry, Francois Lanusse, Gilles Louppe

Keywords: diffusion, score-based, generative model, corrupted data, inverse problems, expectation-maximization, empirical bayes

Abstract: Diffusion models recently proved to be remarkable priors for Bayesian inverse problems. However, training these models typically requires access to large amounts of clean data, which could prove difficult in some settings. In this work, we present a novel method based on the expectation-maximization algorithm for training diffusion models from incomplete and noisy observations only. Unlike previous works, our method leads to proper diffusion models, which is crucial for downstream tasks. As part of our method, we propose and motivate an improved posterior sampling scheme for unconditional diffusion models. We present empirical evidence supporting the effectiveness of our method.

Primary Area: Diffusion based models

Submission Number: 10057