The course has a final project to apply the knowledge gathered throughout the course to a specific problem.
In this project, you will apply the algorithms you've learned in the course to complex Reinforcement Learning environments. The project consists of two parts. In Part 1 you will select two environments from the proposed list, where each environment has a target episode reward and some have several complexity levels. The goal of Part 1 is to train and fine-tune two algorithms from the exercises to reach the target episode reward for the two selected environments. You will then compare the performance of that two algorithms on selected environments.
In Part 2 you will research one paper from the proposed list of papers and implement this paper, answer the questions about this paper. Then you will train the algorithm from this paper and compare its performance with the ones from Part 1.
Alternatively, students can also propose their own project topic. This option is mainly aimed at PhD students that want to apply Reinforcement Learning to their own field, but Master's students are also encouraged.
The project proposal needs to be submitted and will be evaluated by the staff of the course, and can be started once the project is approved. The deadline for the alternative course project proposal is 08.11.2022 at 23:55.
The course project grade accounts for 20% of the final grade of the course.