Special Topics in Deep Reinforcement Learning

Team

Timo P. Gros, Prof. Dr. Verena Wolf (lecturers)
Joschka Groß, Felix Jahn, Nicola J. Müller (assistants)

Content

Reinforcement Learning (RL) is a popular sub-discipline of Machine Learning for problems that require strategies to solve complex tasks such as board games, scheduling problems, or other discrete optimization problems. Deep Reinforcement Learning (DRL) is the combination of Reinforcement Learning with Neural Networks.

In this seminar, we first covered the most important algorithms of Deep Reinforcement Learning. Afterwards, we looked at special topics that our group is conducting research on: exploration in DRL, safe reinforcement learning, and causal reinforcement learning.

Topics

Deep Reinforcement Learning

1. Rainbow (paper)
2. A2C & A3C (paper)
3. PPO (paper)
4. Model-based RL & MCTS (book, especially chapters 8.1, 8.10 and 8.11; original MCTS paper; paper with more details)
5. Alpha(Go) Zero (paper, paper 2)

Safe Reinforcement Learning

6. Conservative Safety Critics for Exploration (paper)
7. Constrained Policy Optimization (paper)
8. Safe Reinforcement Learning via Shielding (paper)

Causal Reinforcement Learning

9. Invariant Causal Prediction for Block MDP (paper)
10. Causal Reinforcement Learning Using Observational and Interventional Data (paper)
11. Can Humans Be Out of the Loop? (paper)

Exploration Techniques

12. Random Network Distillation (paper)
13. Go-Explore (paper)

Organization

All information and materials are available in the CMS.