Variational Inference over Non-differentiable Cardiac Simulators using Bayesian Optimization

TL;DR

This paper introduces a likelihood-free inference method using Bayesian optimization to estimate parameters of a non-differentiable cardiac simulator, improving its fit to real ECG data.

Contribution

It presents a novel approach combining variational inference and Bayesian optimization for parameter inference in complex, non-differentiable simulators.

Findings

Enhanced fit of the cardiac simulator to real ECG data

Demonstrated effectiveness of likelihood-free inference in complex models

Applicable to other non-differentiable simulation-based inference tasks

Abstract

Performing inference over simulators is generally intractable as their runtime means we cannot compute a marginal likelihood. We develop a likelihood-free inference method to infer parameters for a cardiac simulator, which replicates electrical flow through the heart to the body surface. We improve the fit of a state-of-the-art simulator to an electrocardiogram (ECG) recorded from a real patient.