Generative modelling meets Bayesian inference: a new paradigm for inverse problems
Abstract
This special issue addresses Bayesian inverse problems using data-driven priors derived from deep generative models (DGMs) and the convergence of generative modelling techniques and Bayesian inference methods. Conventional Bayesian priors often fail to accurately capture the properties and the underlying geometry of complex, real-world data distributions. In contrast, deep generative models (DGMs), which include generative adversarial networks (GANs), variational auto-encoders (VAEs), normalizing flows and diffusion models (DMs), have demonstrated tremendous success in capturing detailed data representations learned directly from empirical observations. As a result, these models produce priors endowed with superior accuracy, increased perceptual realism and enhanced capacities for uncertainty quantification within inverse problem contexts. This paradigm emerged in the late 2010s, when pioneering efforts were made to explicitly formulate Bayesian inverse problems using conditional Wasserstein generative adversarial networks (GANs). These advances have greatly improved methods for quantifying uncertainties, especially in large-scale imaging applications. Building on these fundamental insights, posterior sampling techniques utilizing DMs have demonstrated remarkable efficiency and robustness, highlighting their potential to effectively tackle complex and diverse inverse problems. The articles collected herein provide essential theoretical breakthroughs and significant algorithmic innovations, collectively demonstrating how deep generative priors mitigate traditional limitations and profoundly enrich both the practical applicability and theoretical foundations of Bayesian inversion.This article is part of the theme issue 'Generative modelling meets Bayesian inference: a new paradigm for inverse problems'.
Related
- Uncertainty Quantification of State Variables Trajectories in the Context of Inverse Problems: An Approach from Bayesian Inference and FDA
- Variational objectives for amortized Bayesian inference in inverse problems: The role of posterior conditioning
- FB-GDM: Fully-Bayesian Guided Diffusion Models for High-Dimensional Linear Inverse Problems via Unsupervised Variational Inference
- Structure-Driven Inversion: A New Paradigm for Solving Inverse Problems
- Causal Inference with Generative Artificial Intelligence: Application to Texts as Treatments
BibTeX
@article{olivierodurmus2026generative,
title = {{Generative modelling meets Bayesian inference: a new paradigm for inverse problems}},
author = {Alain Oliviero-Durmus and Yazid Janati and Eric Moulines and Marcelo Pereyra and Sebastian Reich},
journal = {Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences},
year = {2026},
month = sep,
doi = {10.1098/rsta.2024.0334},
url = {https://hal.science/hal-05755652v1},
note = {CC BY},
}
RIS
TY - JOUR
TI - Generative modelling meets Bayesian inference: a new paradigm for inverse problems
AU - Alain Oliviero-Durmus
AU - Yazid Janati
AU - Eric Moulines
AU - Marcelo Pereyra
AU - Sebastian Reich
PY - 2026
DA - 2026/09/19
JO - Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences
DO - 10.1098/rsta.2024.0334
UR - https://hal.science/hal-05755652v1
AB - This special issue addresses Bayesian inverse problems using data-driven priors derived from deep generative models (DGMs) and the convergence of generative modelling techniques and Bayesian inference methods. Conventional Bayesian priors often fail to accurately capture the properties and the underlying geometry of complex, real-world data distributions. In contrast, deep generative models (DGMs), which include generative adversarial networks (GANs), variational auto-encoders (VAEs), normalizing flows and diffusion models (DMs), have demonstrated tremendous success in capturing detailed data representations learned directly from empirical observations. As a result, these models produce priors endowed with superior accuracy, increased perceptual realism and enhanced capacities for uncertainty quantification within inverse problem contexts. This paradigm emerged in the late 2010s, when pioneering efforts were made to explicitly formulate Bayesian inverse problems using conditional Wasserstein generative adversarial networks (GANs). These advances have greatly improved methods for quantifying uncertainties, especially in large-scale imaging applications. Building on these fundamental insights, posterior sampling techniques utilizing DMs have demonstrated remarkable efficiency and robustness, highlighting their potential to effectively tackle complex and diverse inverse problems. The articles collected herein provide essential theoretical breakthroughs and significant algorithmic innovations, collectively demonstrating how deep generative priors mitigate traditional limitations and profoundly enrich both the practical applicability and theoretical foundations of Bayesian inversion.This article is part of the theme issue 'Generative modelling meets Bayesian inference: a new paradigm for inverse problems'.
SN - 1364-503X
SN - 1471-2962
ER -