Fabio Zanasi, UCL, UK
A Categorical Perspective on Bayesian Reasoning
I will give an overview of recent work providing an algebraic and logical foundation to Bayesian probability theory. Within this perspective, Bayesian networks are treated as categorical objects (string diagrams) and reasoned about using methods inspired from programming language semantics. As a proof of concept, we will see how to compute causal inference and counterfactual reasoning in this setting. This is joint work with Bart Jacobs and Aleks Kissinger.