Accurate and timely forecasting of infectious diseases is critical for 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 supplements to traditional surveillance systems.

This seminar presents a synthesis of research focused on leveraging these data sources for infectious disease prediction, spanning from classical statistical approaches to modern AI architectures.

Key Areas of Discussion:

• Statistical Methodologies: The presentation covers a family of models that combine autoregressive time series structure with penalized regression on Google search data. This includes extensions to spatial-temporal modeling, multi-disease settings, and COVID-19 adaptation.
• AI Architectures: Recent work on attention-based transformer architectures for time series forecasting will be discussed, highlighting methods for multivariate dependency modeling, in-context learning, and efficient linear attention.
• Practical Application: Drawing from ongoing participation in the CDC FluSight forecasting initiative, the session compares the strengths and limitations of both paradigms and shares practical lessons learned from real-time deployment.

The talk aims to offer perspective on when and how statistical rigor and deep learning flexibility each contribute to reliable disease forecasting.

About the Speaker

Dr. Shihao Yang is the Harold E. Smalley Early Career Professor and Assistant Professor in the School of Industrial and Systems Engineering at Georgia Tech. He completed his PhD in statistics and post-doc in Biomedical Informatics at Harvard University. Dr. Yang’s research focuses on data science, with special interest in time series, dynamical systems, and applications in infectious disease transmission forecasting.

To attend, please register online via Zoom.