BayesDiff: Estimating Pixel-wise Uncertainty in Diffusion via Bayesian Inference

arXiv:2310.11142v2 Announce Type: replace-cross
Abstract: Diffusion models have impressive image generation capability, but low-quality generations still exist, and their identification remains challenging due to the lack of a proper sample-wise metric. To address this, we propose BayesDiff, a pixel-wise uncertainty estimator for generations from diffusion models based on Bayesian inference. In particular, we derive a novel uncertainty iteration principle to characterize the uncertainty dynamics in diffusion, and leverage the last-layer Laplace approximation for efficient Bayesian inference. The estimated pixel-wise uncertainty …

abstract arxiv bayesian bayesian inference capability cs.cv cs.lg diffusion diffusion models identification image image generation inference low pixel quality sample type uncertainty via wise