Reinforcement learning is a subfield of AI concerned with learning from repeated interactions with an environment, and is the basis for game-playing algorithms (Chess, Go, Backgammon, StarCraft) and methods across robotics and operations research. Covers Dynamic Programming, Markov Decision Processes, Value/Policy Iteration, Temporal Difference and Monte Carlo methods, and function approximation.

Level

advanced

Department

ENG

Credits

4

Prerequisites

MA225+EK103+EK381 or DS120/121/122