Demonstration of Bayesian inference and Bayesian experimental design in a model film/substrate inference problem

Abstract

In this thesis, we implement Bayesian inference and Bayesian experiment design in a model materials science problem. We demonstrate that by observing the behavior of a film deposited on a substrate, certain features of the substrate may be inferred, with quantified uncertainty. We show that Bayesian experimental design can be used to design efficient experiments. The substrate in this model problem is a Gaussian random field, and the film is a phase separating mixture modeled by the Cahn-Hilliard equation. A key feature of the inference and the experiment design is a stochastic reduced-order model.