Information spreading in diffusion models from effective field theory

Navonil Neogi, Nabil Iqbal

Keywords: diffusion, diffusion models, generative, field theory, information

Abstract: We study score-matching diffusion models with a convolutional architecture. We argue that the inductive bias of locality means that the machinery of effective field theory from physics can be usefully applied to describe the denoising dynamics. As a simple application we study a toy example and show that the mutual information between two points grows in a manner predicted by a simple effective field theory.

Anonymization: This submission has been anonymized for double-blind review via the removal of identifying information such as names, affiliations, and identifying URLs.

Style Files: I have used the style files.

Submission Number: 76