Abstract:Markov chain Monte Carlo (MCMC) simulations have been widely used to generate samples from the complex posterior distribution in Bayesian inferences. However, these simulations often require multiple computations of the forward model in the likelihood function for each drawn sample. This computational burden renders MCMC sampling impractical when the forward model is computationally expensive, such as in the case of partial differential equation models. In this paper, we propose a novel sampling approach called the geometric optics approximation method (GOAM) for Bayesian inverse problems, which entirely circumvents the need for MCMC simulations. Our method is rooted in the problem of reflector shape design, which focuses on constructing a reflecting surface that redirects rays from a source, with a predetermined density, towards a target domain while achieving a desired density distribution. The key idea is to consider the unnormalized Bayesian posterior as the density on the target domain within the optical system and define a geometric optics approximation measure with respect to posterior by a reflecting surface. Consequently, once such a reflecting surface is obtained, we can utilize it to draw an arbitrary number of independent and uncorrelated samples from the posterior measure for Bayesian inverse problems. In theory, we have shown that the geometric optics approximation measure is well-posed. The efficiency and robustness of our proposed sampler, employing the geometric optics approximation method, are demonstrated through several numerical examples provided in this paper.

Consistency of the Bayes method for the inverse scattering problem

Oracle-type Posterior Contraction Rates in Bayesian Inverse Problems

Bayesian approach to inverse scattering with topological priors

A posterior contraction for Bayesian inverse problems in Banach spaces

Wavenumber-Explicit Well-Posedness of Bayesian Shape Inversion in Acoustic Scattering

Numerical Solution of an Inverse Medium Scattering Problem with a Stochastic Source

Numerical Solution of Inverse Scattering Problems with Multi-experimental Limited Aperture Data

INVERSE MEDIUM SCATTERING PROBLEMS IN NEAR-FIELD OPTICS

Inverse Scattering Problems with Multi-Frequencies

On the convergence of the Laplace approximation and noise-level-robustness of Laplace-based Monte Carlo methods for Bayesian inverse problems

A Bayesian approach for consistent reconstruction of inclusions

Increasing stability estimates for the inverse potential scattering problems

The Bayesian approach to inverse Robin problems

Inverse random source scattering for the Helmholtz equation in inhomogeneous media

Consistency of variational Bayesian inference for non-linear inverse problems of partial differential equations

Consistency of Bayesian inference with Gaussian process priors for a parabolic inverse problem

Advancing inverse scattering with surrogate modeling and Bayesian inference for functional inputs

Bayesian inference via geometric optics approximation

Inverse Random Source Scattering Problems in Several Dimensions

Numerical Solution of Inverse Scattering for Near-Field Optics.

Solving Schiffer's problem in inverse scattering theory almost surely