arXiv subjects

Paul J. Birrell

Publications and source records attributed to Paul J. Birrell.

6 recordsLinked to original sources

arXiv2026

GENIE: Generative Neural Inference for Epidemics

The SARS-CoV-2 pandemic highlighted the ongoing risk infectious diseases pose to society and the value of reliable information on the likely future burden. When forecasting an epidemic at fine spatial resolution, traditionally used mechanistic compartmental model struggle to capture highly complex granular transmission dynamics, resulting in inaccurate and overconfident forecasts. However, detailed Agent-Based Models (ABMs), are challenging to calibrate and are too computationally expensive to use in real-time. Amortized simulation-based inference promises to overcome this difficulty by exploiting the power of machine learning (ML) to perform approximate forecasting at near-real-time using arbitrarily complex models of epidemics. In this work we introduce Generative Neural Inference for Epidemics (GENIE), a spatio-temporal ML-based framework for high-resolution forecasting of the burden of respiratory pathogens. GENIE is designed to reflect two key characteristics of outbreaks: (i) shared biological mechanisms across locations and (ii) location-specific characteristics affecting transmission dynamics. This results in the model architecture having two modules: (i) a Local Infection Encoder - which learns to represent disease dynamics shared across all locations and (ii) a Local Profile Encoder - which learns location-specific representations. Using simulations from a high-resolution spatio-temporal ABM, GENIE is trained to generate samples from an approximate posterior predictive distribution of future epidemic trajectories. Benchmarked against established statistical and ML models, GENIE demonstrates superior performance across a range of measures including the timing and magnitude of peak hospitalisations.

stat.ME↗

arXiv2024

The Lifebelt Particle Filter for robust estimation from low-valued count data

Particle filtering methods can be applied to estimation problems in discrete spaces on bounded domains, to sample from and marginalise over unknown hidden states. As in continuous settings, problems such as particle degradation can arise: proposed particles can be incompatible with the data, lying in low probability regions or outside the boundary constraints, and the discrete system could result in all particles having weights of zero. In this paper we introduce the Lifebelt Particle Filter (LBPF), a novel method for robust likelihood estimation in low-valued count problems. The LBPF combines a standard particle filter with one (or more) lifebelt particles which, by construction, lie within the boundaries of the discrete random variables, and therefore are compatible with the data. A mixture of resampled and non-resampled particles allows for the preservation of the lifebelt particle, which, together with the remaining particle swarm, provides samples from the filtering distribution, and can be used to generate unbiased estimates of the likelihood. The main benefit of the LBPF is that only one or few, wisely chosen, particles are sufficient to prevent particle collapse. Differently from other methods, there is no need to increase the number of particles, and therefore the computational effort, in regions of the parameter space that generate less likely hidden states. The LBPF can be used within a pseudo-marginal scheme to draw inferences on static parameters, $ \boldsymbolθ $, governing the system. We address here the estimation of a parameter governing probabilities of deaths and recoveries of hospitalised patients during an epidemic.

stat.CO↗

arXiv2024

Sample-efficient neural likelihood-free Bayesian inference of implicit HMMs

Likelihood-free inference methods based on neural conditional density estimation were shown to drastically reduce the simulation burden in comparison to classical methods such as ABC. When applied in the context of any latent variable model, such as a Hidden Markov model (HMM), these methods are designed to only estimate the parameters, rather than the joint distribution of the parameters and the hidden states. Naive application of these methods to a HMM, ignoring the inference of this joint posterior distribution, will thus produce an inaccurate estimate of the posterior predictive distribution, in turn hampering the assessment of goodness-of-fit. To rectify this problem, we propose a novel, sample-efficient likelihood-free method for estimating the high-dimensional hidden states of an implicit HMM. Our approach relies on learning directly the intractable posterior distribution of the hidden states, using an autoregressive-flow, by exploiting the Markov property. Upon evaluating our approach on some implicit HMMs, we found that the quality of the estimates retrieved using our method is comparable to what can be achieved using a much more computationally expensive SMC algorithm.

stat.ML↗

arXiv2022

An approximate diffusion process for environmental stochasticity in infectious disease transmission modelling

Modelling the transmission dynamics of an infectious disease is a complex task. Not only it is difficult to accurately model the inherent non-stationarity and heterogeneity of transmission, but it is nearly impossible to describe, mechanistically, changes in extrinsic environmental factors including public behaviour and seasonal fluctuations. An elegant approach to capturing environmental stochasticity is to model the force of infection as a stochastic process. However, inference in this context requires solving a computationally expensive ``missing data" problem, using data-augmentation techniques. We propose to model the time-varying transmission-potential as an approximate diffusion process using a path-wise series expansion of Brownian motion. This approximation replaces the ``missing data" imputation step with the inference of the expansion coefficients: a simpler and computationally cheaper task. We illustrate the merit of this approach through two examples: modelling influenza using a canonical SIR model, and the modelling of COVID-19 pandemic using a multi-type SEIR model.

stat.CO↗

arXiv2017

Exploiting routinely collected severe case data to monitor and predict influenza outbreaks

Influenza remains a significant burden on health systems. Effective responses rely on the timely understanding of the magnitude and the evolution of an outbreak. For monitoring purposes, data on severe cases of influenza in England are reported weekly to Public Health England. These data are both readily available and have the potential to provide valuable information to estimate and predict the key transmission features of seasonal and pandemic influenza. We propose an epidemic model that links the underlying unobserved influenza transmission process to data on severe influenza cases. Within a Bayesian framework, we infer retrospectively the parameters of the epidemic model for each seasonal outbreak from 2012 to 2015, including: the effective reproduction number; the initial susceptibility; the probability of admission to intensive care given infection; and the effect of school closure on transmission. The model is also implemented in real time to assess whether early forecasting of the number of admission to intensive care is possible. Our model of admissions data allows reconstruction of the underlying transmission dynamics revealing: increased transmission during the season 2013/14 and a noticeable effect of Christmas school holiday on disease spread during season 2012/13 and 2014/15. When information on the initial immunity of the population is available, forecasts of the number of admissions to intensive care can be substantially improved. Readily available severe case data can be effectively used to estimate epidemiological characteristics and to predict the evolution of an epidemic, crucially allowing real-time monitoring of the transmission and severity of the outbreak.

stat.AP↗

arXiv2015

Synthesising evidence to estimate pandemic (2009) A/H1N1 influenza severity in 2009-2011

Knowledge of the severity of an influenza outbreak is crucial for informing and monitoring appropriate public health responses, both during and after an epidemic. However, case-fatality, case-intensive care admission and case-hospitalisation risks are difficult to measure directly. Bayesian evidence synthesis methods have previously been employed to combine fragmented, under-ascertained and biased surveillance data coherently and consistently, to estimate case-severity risks in the first two waves of the 2009 A/H1N1 influenza pandemic experienced in England. We present in detail the complex probabilistic model underlying this evidence synthesis, and extend the analysis to also estimate severity in the third wave of the pandemic strain during the 2010/2011 influenza season. We adapt the model to account for changes in the surveillance data available over the three waves. We consider two approaches: (a) a two-stage approach using posterior distributions from the model for the first two waves to inform priors for the third wave model; and (b) a one-stage approach modelling all three waves simultaneously. Both approaches result in the same key conclusions: (1) that the age-distribution of the case-severity risks is "u"-shaped, with children and older adults having the highest severity; (2) that the age-distribution of the infection attack rate changes over waves, school-age children being most affected in the first two waves and the attack rate in adults over 25 increasing from the second to third waves; and (3) that when averaged over all age groups, case-severity appears to increase over the three waves. The extent to which the final conclusion is driven by the change in age-distribution of those infected over time is subject to discussion.

stat.AP↗