Analyzing and Guiding Zero-Shot Posterior Sampling in Diffusion Models
This work offers a rigorous foundation and practical guidance for tuning diffusion-based inverse problem solvers, addressing a key bottleneck in their application.
The paper provides a theoretical analysis of zero-shot posterior sampling in diffusion models for inverse problems, deriving closed-form expressions under a Gaussianity assumption. It introduces a principled framework for parameter design that replaces heuristic tuning, achieving a consistent balance between perceptual quality and signal fidelity.
Recovering a signal from its degraded measurements is a long standing challenge in science and engineering. Recently, zero-shot diffusion based methods have been proposed for such inverse problems, offering a posterior sampling based solution that leverages prior knowledge. Such algorithms incorporate the observations through inference, often leaning on manual tuning and heuristics. In this work we propose a rigorous analysis of these approximate posterior samplers, relying on a Gaussianity assumption of the prior. Under this regime, we show that both the ideal posterior sampler and diffusion-based reconstruction algorithms can be expressed in closed-form, enabling their thorough analysis and comparisons in the spectral domain. Building on these representations, we introduce a principled framework for parameter design, replacing heuristic selection strategies used to date. The proposed approach is method-agnostic and yields tailored parameter choices that jointly account for the characteristics of the prior, the degraded signal, and the diffusion dynamics. We show that our spectral recommendations differ structurally from standard heuristics and vary with the diffusion step size, resulting in a consistent balance between perceptual quality and signal fidelity.