On the Bernstein-Von Mises Theorem for High Dimensional Nonlinear Bayesian Inverse Problems

Abstract

We prove a Bernstein-von Mises theorem for a general class of high dimensional nonlinear Bayesian inverse problems in the vanishing noise limit. We propose a sufficient condition on the growth rate of the number of unknown parameters under which the posterior distribution is asymptotically normal. This growth condition is expressed explicitly in terms of the model dimension, the degree of ill-posedness of the inverse problem and the noise parameter. The theoretical results are applied to a Bayesian estimation of the medium parameter in an elliptic problem.

15 pages

References in corpus (8)

• Bayesian inverse problems with Gaussian priors
• On the Bernstein-von Mises phenomenon for nonparametric Bayes procedures
• Hybrid inverse problems and internal functionals
• Bernstein von Mises Theorems for Gaussian Regression with increasing number of regressors
• Posterior Consistency for Bayesian Inverse Problems through Stability and Regression Results
• Bayesian Posterior Contraction Rates for Linear Severely Ill-posed Inverse Problems
• Bernstein - von Mises Theorem for growing parameter dimension
• Gaussian Approximations for Probability Measures on ${\mathbf{R}}^d$

Cited by in corpus (8)

• Frequentist Consistency of Variational Bayes
• Reduced and All-at-Once Approaches for Model Calibration and Discovery in Computational Solid Mechanics
• Non-asymptotic error estimates for the Laplace approximation in Bayesian inverse problems
• Gaussian Approximations for Probability Measures on ${\mathbf{R}}^d$
• Frequentist Consistency of Generalized Variational Inference
• Bayesian inference for nonlinear inverse problems
• The nonparametric LAN expansion for discretely observed diffusions
• Bayesian Generative Models for Knowledge Transfer in MRI Semantic Segmentation Problems