Beskjeder
If you have any questions during the exam, please visit my personal meeting room:
https://uio.zoom.us/my/christos.dimitrakakis
The April 14th group session is moved to April 13th for personal reasons (hopefully I can update the official schedule as well)
Chapter 9: 9.1, 9.2, 9.4, 9.5
Chapter 10.
Optionally read 9.3 if you are interested in continuous state spaces
The project description is available now. You should submit a draft proposal, which we can discuss the following week, before you all start the project. I expect there will be 4-5 groups, so we can devote at least 15 minutes per group in each session, but it's probably best if all the groups are there, as I expect there to be similar questions.
If you want to submit a previous assignment late, send me an email. All assignment work should be handed in by next week, though.
Chapter 8, especially Approximate Value Iteration, Approximate Policy Iteration and Policy Gradient.
External reading. If you have access to some of these books by Bertsekas
"Neurodynamic programming", Chapter 6
The book "Reinforcement Learning and Optimal Control"
"Dynamic Programming and Optimal Control", Chapter 6
Sec 6.3.1
Section 6.4
From 6.5:
- 6.5.1
- 6.5.2
- 6.5.3
- 6.5.4 (optionally the linear programming section)
You should submit assignments through devilry:
https://devilry.ifi.uio.no/
If you do not have access, then submit by email by the deadline with the subject:
IN-STK5100 Assignment #number
Chapter 1
From Chapter 2
- 2.1 (excluding 2.1.5)
- 2.3 (excldung 2.3.4)
From Chapter 3
- 3.1
- 3.2
- 3.3
- Optionally, 3.4 3.5
From Chapter 4
- 4.1
- 4.2
- 4.3.1 (the other models will be discussed later)
- 4.4
- 4.5
- 4.6.4
Please post any questions to padlet
Since external people could not access the public uio repo, I made a new one, available here:
https://github.com/olethrosdc/rldmuu
Please use that one
The course book is available here
In the first week, we go through chapters 2 and 3.
Sometimes I'll post online lecture quizzes (rather than zoom polls) To do those, go to
https://b.socrative.com/student/
and use room INSTK5100
The course will involve some in-class coding, please clone this repo:
https://github.uio.no/chridim/in-stk-5100
Apart from registering for the course, you should register for the online lectures here:
Lecture/lab sessions are twice-weekly until March. Then it is once-weekly project discussions in March, April and May.