Inference, Information Theory, Learning and Statistical Mechanics

Harvard FAS

This course focuses on the modern applications of Statistical Mechanics. We will learn the basics of information theory, coding and compression. We will next learn about Bayesian Inference, priors and maximizing entropy, which will naturally lead us to regularization and compressed sensing. We will then cover learning: support vector machines, vc dimension, supervised, reinforcement and unsupervised learning. These topics, which build on each other, will be taught using examples in…