Post-Bayes on the Beach

AI is accelerating scientific model development, while the traditional statistical practices of carefully checking model assumptions are becoming sidelined. At the same time, there is growing societal concern regarding the reliability of AI systems, prompting an urgent search for principled approaches to uncertainty quantification and robust decision-making. This confluence of rapid deployment and foundational gaps presents a clear opportunity for the statistical sciences to play a central role. It would be tempting to think that Bayesian inference offers a solution, combining a probabilistic model with prior knowledge to reason about uncertain parameters and to predict data that are yet to be observed. Sadly this is optimistic, as Bayesian inference can fail spectacularly when either the model is misspecified or when meaningful prior information cannot be exploited.

Motivated by these fundamental challenges, there has been a recent surge of interest in post-Bayesian statistical methodology. 

This course is intended for both postgraduate students and established practitioners who are familiar with Bayesian statistical methods and who would like to learn more about the emerging state-of-the-art in post-Bayesian methodology. 

Register to attend!

The name "Post-Bayes on the Beach" is a tribute to the wonderful "Bayes on the Beach" conference series organised by the statistics community in Australia, but the two events are not officially affiliated.