RL: Tutorials

The course offers weekly, 1 hour tutorial sessions starting in week 2 of the second semester. There are 9 tutorials running in weeks 2 till 10. Each tutorial has a corresponding tutorial sheet usually consisting of two questions. While the labs provide hands-on experience with developing Reinforcement Learning solutions, the tutorials are aimed at stimulating a deeper understanding of the key concepts. The tutorials should help you understand when and why to build what, and prepare you for the exam.

Tutorials this year will run in-person in AT 5.04 on Wednesdays (you should only attend one tutorial session per week). The complete list of sessions for a course can be found on the course's DRPS page timetable, which for this course is available here. Students will be randomly allocated to their tutorial group but can request a change by submitting a Tutorial Change form, which will soon be available here (once the form has been submitted, or if it is just for one week only, feel free to join an alternative session without registration).

The tutorials will involve discussing the weekly exercises with other students and with a tutor to answer questions and discuss solutions. You should come to each tutorial having reviewed the current week's tutorial sheet, and ideally also having prepared a few questions that you like to discuss.

We will be releasing tutorial materials on the Tuesday of the previous week of the tutorial session (around noon), and the answers by Tuesday noon of the following week. For example, the Week 2 tutorial will have its questions released on Tuesday by 12:00 (UK time) of Week 1. Solutions and hints for answers will then be posted by Tuesday 12:00 of Week 3. You will find all materials right below on this page, as they are released.

The tutorials in weeks 9 and 10 will give you an opportunity to revise the main content of the course based on exam questions from previous years. If you have any suggestions for additional topics to be discussed in the revision tutorials or any other questions, corrections, and comments, please contact michael dot herrmann at ed.

License
All rights reserved The University of Edinburgh