The SDML book club started discussing Reinforcement Learning: An Introduction (second edition) by Richard Sutton and Andrew Barto in May 2021. A PDF copy of the book and other related material can be found on Richard Sutton's book website.
Below are the notes/slides and meetup recordings:
Session 1: Introduction to Reinforcement Learning
Here is the full Meetup event description.
Most of the material was drawn from chapters 1 and 3 of the Sutton & Barto book.
Content was covered over two weeks.
Slides are available in PowerPoint (large file)
or PDF (smaller file),
and Meetup video from the first week and part two video are both up.
Session 2: Dynamic Programming and Monte Carlo Methods for RL
Here is the detailed Meetup event description.
The main material for this session came from chapters 4 and 5 of the Sutton & Barto book.
This session also was covered over two weeks.
The first week discussed dynamic programming. Here are the slides,
and here is the dynamic programming video.
The second week discussed Monte Carlo methods in RL. This is the link to the slides in Ted's repo,
and this is the Monte Carlo methods video.
Session 3: Temporal-Difference Learning
Here is the Meetup event description.
The majority of the material for this session comes from chapter 6 of the Sutton & Barto book.
Here are the slides and video will be uploaded soon.
Session 4: Programming Reinforcement Learning Algorithms
Full details in the Meetup event description.
Here are the slides.
Session 5: Planning and Learning in Reinforcement Learning
Here are the full details of the Meetup event
The material for this session primarily is from chapter 8 of the Sutton & Barto book.
Link to the slides.
Session 6: Policy Gradient Methods for Reinforcement Learning
The full description is in the Meetup event details
The core concepts for this discussion came from chapter 13 of the Sutton & Barto book.
GitHub has the slides for the session. \
We clearly have a lot of catching up to get all the videos uploaded to YouTube!
An implementation of the examples in the book in Python has been shared in this repo: https://github.com/ShangtongZhang/reinforcement-learning-an-introduction. I haven't had a chance to review it to validate all of the code runs without errors.