A1298
Title: Scalable calibration of individual-based epidemic models through categorical approximations
Authors: Lorenzo Rimella - Università degli studi di Bergamo (Italy) [presenting]
Abstract: Traditional compartmental models capture population-level dynamics but fail to characterize individual-level risk. The computational cost of exact likelihood evaluation for partially observed individual-based models, however, grows exponentially with the population size, necessitating approximate inference. Existing sampling-based methods usually require multiple simulations of the individuals in the population and rely on bespoke proposal distributions or summary statistics. A deterministic approach to approximating the likelihood using categorical distributions is proposed. The approximate likelihood is amenable to automatic differentiation so that parameters can be estimated by maximization or posterior sampling using standard software libraries such as TensorFlow with little user effort. Consistency of the maximum approximate likelihood estimator is proven. The approach is empirically tested on several classes of individual-based models for epidemiology: different sets of disease states, individual-specific transition rates, spatial interactions, under-reporting and misreporting. Ground truth recovery and comparable log-likelihood values are demonstrated at substantially reduced cost compared to competitor methods. Finally, scalability and effectiveness of the approach are shown with a real-world application on the 2001 UK Foot-and-Mouth outbreak, where the simplicity of the CAL allows inclusion of 162,775 farms.
