Reinforcement Learning

The participants have a thorough understanding of the key concepts and are able to discriminate between the different types of Reinforcement Learning (RL). They can understand the noteworthy algorithms and methods in the field and are able to implement them programmatically. The participants are able to identify tasks that can be solved with Reinforcement Learning.

1. Classical Reinforcement Learning: Cross-Entropy Method, Dynamic Programming, Monte Carlo Methods, Temporal-Difference Learning
2. Value Based methods: Sarsa, Q-Learning
3. Policy Based methods: REINFORCE, Proximal Policy Optimization
4. Deep Reinforcement Learning: Q-Networks
5. Hybrid Methods: A2C, A3C