BIMC: The Bayesian Inverse Monte Carlo method for goal-oriented uncertainty quantification. Part I

TL;DR

This paper introduces BIMC, an importance sampling method based on Bayesian inverse problems, to efficiently estimate rare event probabilities in complex systems with uncertain parameters, significantly reducing computational costs.

Contribution

The paper presents BIMC, a novel importance sampling scheme that uses Bayesian inverse problem solutions to improve rare event probability estimation in expensive systems.

Findings

BIMC achieves several orders of magnitude computational savings.

It provides conditions for optimality and failure modes of the method.

Demonstrated effectiveness on multiple example problems.

Abstract

We consider the problem of estimating rare event probabilities, focusing on systems whose evolution is governed by differential equations with uncertain input parameters. If the system dynamics is expensive to compute, standard sampling algorithms such as the Monte Carlo method may require infeasible running times to accurately evaluate these probabilities. We propose an importance sampling scheme (which we call BIMC) that relies on solving an auxiliary, fictitious Bayesian inverse problem. The solution of the inverse problem yields a posterior PDF, a local Gaussian approximation to which serves as the importance sampling density. We apply BIMC to several problems and demonstrate that it can lead to computational savings of several orders of magnitude over the Monte Carlo method. We delineate conditions under which BIMC is optimal, as well as conditions when it can fail to yield an effective IS density.