Обучение с подкреплением (Reinforcement Learning)

Reinforcement Learning is a widely used machine learning technique for autonomous systems, based on learning from reward and punishment. Many of the recent big AI successes are at least in part based on RL technology (e.g. Google Deep Mind). This course will introduce the fundamentals of RL and many of the state-of-the-art techniques. Students will learn both the theory of RL and get an opportunity to implement RL methods and apply them to popular benchmark domains. Topics covered include dynamic programming, monte carlo methods temporal difference learning, model-based RL, value function approximation, hierarchical RL, inverse RL, transfer learning.