Predictive-based Bayesian learning

There is a renewed interest for predictive-based Bayesian learning, as a foundational concept as well in its methodological and computational implications. Our motivation stems from providing Bayesian understanding, with the consequent quantification of uncertainty, of predictive algorithms; I would give an example of that for DP mixture models with streaming data. The focus in the talk will however be on explaining foundational concepts and illustrating overlooked implications in predictive-based sampling. If time permits, I will also present recent (ongoing) results on predictive-based asymptotic approximations of the posterior distribution that are quite illuminating - we believe - on the link between prediction and frequentist coverage of credible intervals.

Zoom link: LINK