Optimal approximation error approach in Bayesian inverse problems

DOI10.48550/ARXIV.2610.38434MaRDI QIDQ7409131arXivFDO

Authors Fabian Schneider, Tapio Helin

Publication date 28 October 2026

Published in arXiv

Tldr We propose an optimal approximation error (OAE) framework, which determines the model error via minimising the expected information loss between true and approximate posteriors; this approach yields better performance than existing methods in numerical experiments on CT and Darcy-flow inverse problems.

Full work available at URL https://doi.org/10.48550/ARXIV.2610.38434

arXiv classification

math.NAstat.COstat.ME

Cites work

• A block coordinate descent method for regularized multiconvex optimization with applications to nonnegative tensor factorization and completion
• A convergent least-squares regularized blind deconvolution approach
• A Convex Approximation Framework for Neural Likelihood-Based Bayesian Inverse Problems
• Adaptive construction of surrogates for the Bayesian solution of inverse problems
• Approximation errors and model reduction with an application in optical diffusion tomography
• Asymptotic Statistics
• Bayesian calibration of computer models. (With discussion)
• Can one use total variation prior for edge-preserving Bayesian inversion?
• Choosing observation operators to mitigate model error in Bayesian inverse problems
• Choosing observation operators to mitigate model error in Bayesian inverse problems
• Convergence of Proximal Iteratively Reweighted Nuclear Norm Algorithm for Image Processing
• Convex optimization
• Discretization-invariant Bayesian inversion and Besov space priors
• EMBEDDED MODEL ERROR REPRESENTATION FOR BAYESIAN MODEL CALIBRATION
• Estimation of the Robin coefficient field in a Poisson problem with uncertain conductivity field
• Gaussian approximations for probability measures on \(\mathbb R^d\)
• Inverse problems: a Bayesian perspective
• Iterative updating of model error for Bayesian inversion
• Kullback-Leibler approximation for probability measures on infinite dimensional spaces
• Learning about physical parameters: the importance of model discrepancy
• Linear inverse problems for generalised random variables
• MARGINALIZATION OF UNINTERESTING DISTRIBUTED PARAMETERS IN INVERSE PROBLEMS APPLICATION TO DIFFUSE OPTICAL TOMOGRAPHY
• Model Reduction for Large-Scale Systems with High-Dimensional Parametric Input Space
• On gradients of functions definable in o-minimal structures
• On learned operator correction in inverse problems
• Optimal experimental design: formulations and computations
• Posterior consistency for Gaussian process approximations of Bayesian posterior distributions
• Statistical inverse problems: discretization, model reduction and inverse crimes
• Stochastic Equations in Infinite Dimensions
• Stochastic spectral methods for efficient Bayesian solution of inverse problems

This page was built for publication: Optimal approximation error approach in Bayesian inverse problems

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q7409131)