MS&E338 Reinforcement Learning


rat in maze with gps 


This course offers an advanced introduction to reinforcement learning. Bandit learning and Markov decision processes will be covered during the first few weeks as background material. The focus will be on the design, analysis, and application of algorithms for complex reinforcement learning problems. The course will cover some recent developments in the field. Students are expected to have experience with optimization, stochastic processes, linear regression, coding for numerical computation, and writing mathematical proofs.


‘‘Other areas of machine learning are about minimization; reinforcement learning is about maximization.’’