Bayesian neural network priors for edge-preserving inversion

12/20/2021
by   Chen Li, et al.
13

We consider Bayesian inverse problems wherein the unknown state is assumed to be a function with discontinuous structure a priori. A class of prior distributions based on the output of neural networks with heavy-tailed weights is introduced, motivated by existing results concerning the infinite-width limit of such networks. We show theoretically that samples from such priors have desirable discontinuous-like properties even when the network width is finite, making them appropriate for edge-preserving inversion. Numerically we consider deconvolution problems defined on one- and two-dimensional spatial domains to illustrate the effectiveness of these priors; MAP estimation, dimension-robust MCMC sampling and ensemble-based approximations are utilized to probe the posterior distribution. The accuracy of point estimates is shown to exceed those obtained from non-heavy tailed priors, and uncertainty estimates are shown to provide more useful qualitative information.

READ FULL TEXT

page 2

page 19

page 20

page 21

research
07/19/2022

Horseshoe priors for edge-preserving linear Bayesian inversion

In many large-scale inverse problems, such as computed tomography and im...
research
04/23/2021

Exact priors of finite neural networks

Bayesian neural networks are theoretically well-understood only in the i...
research
12/20/2020

Dimension-robust Function Space MCMC With Neural Network Priors

This paper introduces a new prior on functions spaces which scales more ...
research
02/28/2021

Random tree Besov priors – Towards fractal imaging

We propose alternatives to Bayesian a priori distributions that are freq...
research
02/12/2021

Bayesian Neural Network Priors Revisited

Isotropic Gaussian priors are the de facto standard for modern Bayesian ...
research
05/14/2021

BNNpriors: A library for Bayesian neural network inference with different prior distributions

Bayesian neural networks have shown great promise in many applications w...
research
12/11/2022

Bayesian inversion with α-stable priors

We propose to use Lévy α-stable distributions for constructing priors fo...

Please sign up or login with your details

Forgot password? Click here to reset