Speaker: Dr. Lijing Wang
Target Audience: Those interested in Artificial Intelligence (AI), Machine Learning (ML), Deep Learning (DL), and epidemiology.
This workshop covers the state-of-the-art for epidemic forecasting. It starts with a formal definition of an epidemic process and details the different aspects of disease spread dynamics. The problem of spatiotemporal epidemic forecasting is then formulated, and the central challenges are outlined. Subsequently, it covers major methodologies for epidemic forecasting including theory-based mechanistic methods and data-driven methods. Then a range of methods that have been developed will be described and the experience of our team will be discussed.
Forecasting the spatial and temporal evolution of epidemic dynamics has been an area of active research over the past couple of decades. The importance of the topic is evident: policymakers, citizens, and scientists would all like to get accurate and timely forecasts. In contrast to physical systems, the co-evolution of epidemics, individual and collective behavior, viral dynamics, and public policies make epidemic forecasting a problematic task. Data-driven methods are popular since they do not need explicit knowledge of the physical behavior of the system, and have been deployed successfully in multiple domains. For instance, deep learning-based predictive models have gained increasing prominence in epidemic forecasting. However, they are challenging to train due to sparse and noisy training data and the limited ability to explicitly incorporate mechanisms of disease spread. In recent times, theory-based mechanistic methods have become a mainstay of epidemic forecasting due to their ability to capture the underlying causal processes through mathematical and computational representations.