Reinforcement Learning (RL) is a subfield of Machine Learning where an agent learns to make decisions by taking actions in an environment towards a specific learning goal. RL relies on exploration methods, using feedback from its actions to improve its future behavior. This Lab Course addresses problems that are commonly faced when working with, researching and/or applying RL techniques.
The course especially focuses on topics such as
- Safety: How can we make sure that an agent refrains from taking bad decisions?
- Sparse rewards: How can agents learn to fulfill their targets even if their environments seldomly provide feedback to learn from?
- Sim2Real: It is more convenient, safer, and faster to first train agents in virtual environments instead of the real world. But how can we bridge the differences between simulated and real environments?
- Interpretable and Explainable RL: How can we make the decision-making process of RL agents transparent and understandable to humans?
- Exploration vs. Exploitation: In order to learn better decisions, agents have to explore their options. But how can it be done efficiently and/or safely?
- Causality: How can RL agents understand and use causal relations in their environment for enhanced learning and generalization?
- Kursleiter/in: Nils Jansen
Semester: WiSe 2025/26