Mauro Maggioni: Some mathematical perspectives on AI, and on its application to scientific discovery

Mauro Maggioni
Bloomberg Distinguished Professor in Mathematics and Applied Mathematics and Statistics,
Johns Hopkins University

I will discuss at very high level, and lots of pictures, what some of the mathematical ideas behind some of the recent advancements in AI are.  As an example of an application, I will discuss recent work in deep learning in the context personalized cardiology, with the goal of predicting medically-relevant electrophysiological features of digital twins of the hearts of patients.  I will conclude with discussing what are present challenges and opportunities in making AI useful for scientific discovery.