Location
MC 6460
Candidate
Ali Haghighatgooasiabar | Applied Mathematics, University of Waterloo
Title
Deep Learning for Early Warning Signals of Complex Dynamical Transitions in Epidemics
Abstract
The identification of tipping points, which are frequently caused by imperceptible alter-ations in internal structures or external disturbances, is crucial for forecasting dynamical systems such as vaccine-preventable infectious diseases. Based on empirical data, when either the vaccination rate or birth rate changes, the pediatric disease infection prevalence encounters a tipping point, causing a transition between different limit cycles such as annual to biennial or more complex cycles. Identifying this transition helps prevent the epidemic. In this work, our aim is to design deep learning models to construct an early warning system to provide advance warning of these tipping points.