Saved in:
Bibliographic Details
Main Authors: Yang, Yiming, Cheng, Xiaoyuan, He, Yi, Li, Kaiyu, Yuan, Wenxuan, Sun, Zhuo
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2602.02045
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866917476264050688
author Yang, Yiming
Cheng, Xiaoyuan
He, Yi
Li, Kaiyu
Yuan, Wenxuan
Sun, Zhuo
author_facet Yang, Yiming
Cheng, Xiaoyuan
He, Yi
Li, Kaiyu
Yuan, Wenxuan
Sun, Zhuo
contents Diffusion models have emerged as powerful learned priors for Bayesian inverse problems (BIPs). Diffusion-based solvers rely on a presumed likelihood for the observations in BIPs to guide the generation process. Likelihood misspecification is common in practical BIPs and is known to degrade recovery performance, particularly under outlier contamination. We investigate this problem by first characterizing the induced posterior deviation and proving the stability of diffusion-based solvers for linear BIPs. Our stability analysis further reveals potential robustness deficiencies of existing diffusion-based solvers under outlier-contaminated measurements. To address this issue, we propose a simple yet effective solution: robust diffusion posterior sampling, which is provably outlier-robust for linear BIPs and compatible with existing gradient-based posterior samplers. Empirical results from scientific inverse problems and natural image tasks demonstrate the effectiveness and robustness of our method, with consistent performance gains in challenging scenarios involving outlier contamination for both linear and nonlinear tasks.
format Preprint
id arxiv_https___arxiv_org_abs_2602_02045
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Outlier-robust Diffusion Posterior Sampling for Bayesian Inverse Problems
Yang, Yiming
Cheng, Xiaoyuan
He, Yi
Li, Kaiyu
Yuan, Wenxuan
Sun, Zhuo
Machine Learning
Diffusion models have emerged as powerful learned priors for Bayesian inverse problems (BIPs). Diffusion-based solvers rely on a presumed likelihood for the observations in BIPs to guide the generation process. Likelihood misspecification is common in practical BIPs and is known to degrade recovery performance, particularly under outlier contamination. We investigate this problem by first characterizing the induced posterior deviation and proving the stability of diffusion-based solvers for linear BIPs. Our stability analysis further reveals potential robustness deficiencies of existing diffusion-based solvers under outlier-contaminated measurements. To address this issue, we propose a simple yet effective solution: robust diffusion posterior sampling, which is provably outlier-robust for linear BIPs and compatible with existing gradient-based posterior samplers. Empirical results from scientific inverse problems and natural image tasks demonstrate the effectiveness and robustness of our method, with consistent performance gains in challenging scenarios involving outlier contamination for both linear and nonlinear tasks.
title Outlier-robust Diffusion Posterior Sampling for Bayesian Inverse Problems
topic Machine Learning
url https://arxiv.org/abs/2602.02045