Introduction to Deep Reinforcement Learning

What it is about

You learn the fundamentals and advanced methods of Deep Reinforcement Learning (DRL). At the end you will be able to model problems as a Markov Decision Process, understand common DRL algorithms, and practically implement and evaluate a DRL method.

What you will be able to do

• Basic knowledge of search algorithms (graph and tree search)
• Problem formulation as Markov Decision Process (MDP)
• In-depth knowledge of Reinforcement Learning (e.g. Q-Learning, TD-Learning)
• Fundamental knowledge in Deep Learning (e.g. SGD, logistic regression, neural networks)
• Advanced knowledge in Deep Reinforcement Learning (e.g. DQN, PPO)
• Application of a DRL framework to a practical problem
• Evaluation of DRL methods regarding advantages and disadvantages
• Recognize typical pitfalls in practical applications and strategies to avoid them

What the module consists of

• VorlesungVermittlung der theoretischen Grundlagen von (Deep) Reinforcement Learning
• Übungen / Coding LabsAnwendung des Gelernten an praktischen Problemen durch Programmieraufgaben

Teaching method

• VorlesungenErklären der Theorie und Konzepte
• Übungen und Coding LabsPraktisches Einüben und Anwenden der Methoden