The CHHS Webinar Series presents Infectious Disease Forecasting with Digital Data Streams: Comparing AI Transformers with Classical Statistical Methods, featuring Dr. Shihao Yang, Harold L. Smalley Career Professor and Assistant Professor in the School of Industrial & Systems Engineering at Georgia Tech.

Accurate and timely forecasting of infectious diseases is critical for effective public health decision-making. Over the past decade, digital data streams—such as internet search queries, electronic health records, and pharmacy sales data—have emerged as powerful tools to enhance traditional surveillance systems.

In this talk, Dr. Yang will share a comprehensive overview of his research on leveraging these data sources for infectious disease prediction. He will explore classical statistical approaches, including autoregressive time series models enhanced with penalized regression and their extensions to spatial-temporal and multi-disease forecasting, as well as applications to COVID-19 adaptation.

On the AI front, Dr. Yang will discuss recent advancements in attention-based transformer architectures for time series forecasting, including methods for multivariate dependency modeling, contextual learning, and efficient linear attention. Drawing from his experience with the CDC FluSight forecasting initiative, he will compare the performance, strengths, and limitations of both statistical and AI-driven approaches.

This session aims to provide practical insights into when and how each methodology can be applied, offering a balanced perspective on combining statistical rigor with deep learning flexibility to improve disease forecasting outcomes.