Information Theory (for Data Science)
Ph.D. course, ISTA, 2023
I am excited to announce that I am co-teaching (along with Marco Mondelli) the Ph.D. class on Information Theory (for Data Science). The class covers a variety of classical and more modern results with an information-theoretic flavor. We will talk about: Information measures, compression, prediction, estimation, large deviations (Sanov style), multi-armed bandits, exponential families, exploration bias and generalisation error.