MATH 6397 -- Probabilistic Methods in Reinforcement and Machine Learning
Course Meeting Times and Place: Tu Th 4 - 5:20 in SES 201
Office Hours: Th 3 - 4, and by appointment. Email me for an appointment.
Text:T here is no single textbook for this course, but I will follow the exposition in Kevin P. Murphy’s books entitled “Probabilistic Machine Learning”.
Course Description:
This course is an introduction to topics that are useful in designing learning algorithms with a focus on action, as well as understanding how makes these algorithms work, and how they could be implemented both in computers, and in brains. I will cover a number of different topics: 1) Introductory material with a focus on Bayesian inference, 2) Filtering, smoothing and prediction, 3) Markov decision processes, and reinforcement learning, 3) Machine learning and neuroscience, and, time permitting, 4) Information theory and information geometry with applications.