Diffusion models as plug-and-play priors

TL;DR

This paper introduces a method to use diffusion models as flexible priors in high-dimensional inference tasks, enabling diverse applications like conditional generation and segmentation by iteratively differentiating through the denoising process.

Contribution

It demonstrates how diffusion models can serve as plug-and-play priors for various inference problems, expanding their utility beyond generation.

Findings

Enables high-dimensional inference with diffusion models as priors.

Allows adaptation to new tasks and domains via differentiable constraints.

Introduces a novel search mechanism using multiple noised versions for optimization.

Abstract

We consider the problem of inferring high-dimensional data $\mathbf{x}$ in a model that consists of a prior $p(\mathbf{x})$ and an auxiliary differentiable constraint $c(\mathbf{x},\mathbf{y})$ on $x$ given some additional information $\mathbf{y}$. In this paper, the prior is an independently trained denoising diffusion generative model. The auxiliary constraint is expected to have a differentiable form, but can come from diverse sources. The possibility of such inference turns diffusion models into plug-and-play modules, thereby allowing a range of potential applications in adapting models to new domains and tasks, such as conditional generation or image segmentation. The structure of diffusion models allows us to perform approximate inference by iterating differentiation through the fixed denoising network enriched with different amounts of noise at each step. Considering many noised versions of $\mathbf{x}$ in evaluation of its fitness is a novel search mechanism that may lead to new algorithms for solving combinatorial optimization problems.