What Secrets Do Your Manifolds Hold? Understanding the Local Geometry of Generative Models

ICLR 2025 Conference Submission13284 Authors

Keywords: Geometry, Diffusion models, VAE, Generative Models, Guidance, Memorization, Out-of-Distribution Detection

TL;DR: We show that the local geometry of generative models is indicative of generation aesthetics, artifacts, diversity, and memorization.

Abstract: Deep Generative Models are frequently used to learn continuous representations of complex data distributions using a finite number of samples. For any generative model, including pre-trained foundation models with GAN, Transformer or Diffusion architectures, generation performance can vary significantly based on which part of the learned data manifold is sampled. In this paper we study the post-training local geometry of the learned manifold and its relationship to generation outcomes for models ranging from toy settings to the latent decoder of the near state-of-the-art Stable Diffusion 1.4 Text-to-Image model. Building on the theory of continuous piecewise-linear (CPWL) generators, we characterize the local geometry in terms of three geometric descriptors - scaling ($\psi$), rank ($\nu$), and complexity ($\delta$). We provide quantitative and qualitative evidence showing that for a given latent, the local descriptors are indicative of generation aesthetics, artifacts, diversity, and memorization. Finally we demonstrate that training a reward model using the local geometry allows us to control the log-likelihood of a generated sample under the learned distribution, and improve the qualitative aspects of an image.

Primary Area: generative models

Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics.

Submission Guidelines: I certify that this submission complies with the submission instructions as described on https://iclr.cc/Conferences/2025/AuthorGuide.

Reciprocal Reviewing: I understand the reciprocal reviewing requirement as described on https://iclr.cc/Conferences/2025/CallForPapers. If none of the authors are registered as a reviewer, it may result in a desk rejection at the discretion of the program chairs. To request an exception, please complete this form at https://forms.gle/Huojr6VjkFxiQsUp6.

Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors’ identity.

No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.

Submission Number: 13284