Predictive Bayesian Inference

National AI Research Lab (NAIRL) hosted an invited talk by Dr. Edwin Fong, Assistant Professor in the Department of Statistics and Actuarial Science at the University of Hong Kong, on June 3, 2025, at the Seoul AI Hub.

Under the theme “Predictive Bayesian Inference,” Dr. Fong introduced a predictive view of Bayesian inference, in which prediction, rather than the traditional likelihood-prior specification, serves as the starting point of statistical inference. The talk covered how this perspective connects the Bayesian posterior with predictive distributions on unobserved data, and how it opens up new approaches to longstanding challenges such as scalability and model misspecification.

Dr. Fong completed his Ph.D. at the University of Oxford and the Alan Turing Institute, and previously worked as a data scientist at Novo Nordisk, developing statistical methods for healthcare applications. His research focuses on Bayesian methods driven by prediction, spanning Bayesian nonparametrics, scalable inference, and model selection, with connections to machine learning, conformal prediction, and causal inference in clinical trials. NAIRL will continue to connect Korea’s research community with leading AI experts from around the world through its invited talk series.