Gaussian process modelling of infectious diseases using the Greta software package and GPUs

TL;DR

This paper demonstrates how Greta software, leveraging GPU acceleration, enables efficient Gaussian process modeling of infectious disease spread, significantly reducing computation time and improving inference for spatio-temporal epidemiological data.

Contribution

It introduces a GPU-accelerated Gaussian process inference pipeline using Greta for modeling infectious disease dynamics, highlighting the impact of covariance kernel choices.

Findings

GPU acceleration reduces computation time by up to 70%

Covariance kernel choice affects inference accuracy and extrapolation

Applied to tuberculosis data in England with successful predictions

Abstract

Gaussian process are a widely-used statistical tool for conducting non-parametric inference in applied sciences, with many computational packages available to fit to data and predict future observations. We study the use of the Greta software for Bayesian inference to apply Gaussian process regression to spatio-temporal data of infectious disease outbreaks and predict future disease spread. Greta builds on Tensorflow, making it comparatively easy to take advantage of the significant gain in speed offered by GPUs. In these complex spatio-temporal models, we show a reduction of up to 70\% in computational time relative to fitting the same models on CPUs. We show how the choice of covariance kernel impacts the ability to infer spread and extrapolate to unobserved spatial and temporal units. The inference pipeline is applied to weekly incidence data on tuberculosis in the East and West Midlands regions of England over a period of two years.