Variational Calibration of Computer Models

TL;DR

This paper introduces a scalable calibration framework for computer models using approximate Deep Gaussian processes and variational inference, addressing computational and statistical challenges of traditional Bayesian methods.

Contribution

It proposes a novel, practical approach that replaces Gaussian process emulation and MCMC with deep Gaussian processes and variational inference for calibration.

Findings

Achieves competitive calibration performance

Reduces computational complexity

Demonstrates scalability to complex models

Abstract

Bayesian calibration of black-box computer models offers an established framework to obtain a posterior distribution over model parameters. Traditional Bayesian calibration involves the emulation of the computer model and an additive model discrepancy term using Gaussian processes; inference is then carried out using MCMC. These choices pose computational and statistical challenges and limitations, which we overcome by proposing the use of approximate Deep Gaussian processes and variational inference techniques. The result is a practical and scalable framework for calibration, which obtains competitive performance compared to the state-of-the-art.