Reinforcement Learning

Undergraduate

Multi-armed bandits, Markov decision processes, model-based and model-free methods (such as dynamic programming, Monte Carlo methods, and temporal-difference methods) for learning value and policy functions. Approximation solutions including deep reinforcement learning. Course Format: Lectures, lab, 5 hours per week. Restrictions: Open to COSC single or combined, GAME, NEUR, and Data Science programs. Prerequisite(s): COSC 3P71 (minimum 60 percent), MATH 1P02 or MATH 1P06 Course Notes: This course may be offered in multiple modes of delivery. The method of delivery will be listed on the academic timetable, in the applicable term.