Sutton and Barto book implementation

[ivanbelenky]

https://github.com/ivanbelenky/RL

[vlad]

The authors were professor and grad student at UMass Amherst, and are the current winners of the Turing Award.

https://www.cics.umass.edu/

https://www.nsf.gov/news/ai-pioneers-andrew-barto-richard-su...

[mark_l_watson]

I want to add a second comment:

Professors White & White (a husband and wife team) have a very good set of courses on RL on Coursera:

https://www.coursera.org/specializations/reinforcement-learn...

[ivanbelenky → mark_l_watson]

Lovely!

[ultrasounder]

Super helpful while I come upto speed with this field in general. Currently taking the XCS234(RL @ Stanford online) and this book is referenced for everything.

[mark_l_watson]

Very nice, thanks for doing this.

I have experimented a lot with the "official" Common Lisp and Python examples for the Sutton/Barto RL book, and I will enjoy your implementations also!

For reference, original examples in Lisp and Python: http://incompleteideas.net/book/code/code2nd.html

A bunch of implementations with all kinds of use cases (e.g., using OpenAI RL Gym, etc.):

Here are some resources with code examples and implementations related to the Sutton and Barto "Reinforcement Learning: An Introduction" book:

Code for Sutton & Barto Book: Reinforcement Learning: An Introduction: The official website for the book provides links to various software and re-implementations in different languages, including Python, Julia, and Lisp. This is a great starting point to find code directly associated with the book's examples and exercises.

Link: http://incompleteideas.net/book/code/code2nd.html jovsa/rl-examples-sutton-and-barto-book on GitHub: This repository offers Python implementations of examples from the book, organized by chapter. It includes code for figures and examples from various chapters, covering topics like Gridworld, Blackjack, and the Mountain Car task.

Link: https://github.com/jovsa/rl-examples-sutton-and-barto-book kamenbliznashki/sutton_barto on GitHub: This repository provides Python implementations of RL algorithms for the examples and figures in the Sutton and Barto book. It covers a wide range of topics from multi-armed bandits to policy gradient methods.

Link: https://github.com/kamenbliznashki/sutton_barto boldyshev/sutton on GitHub: This repository contains Python implementations of example experiments (figures) and programming exercises from the second edition of the book. Chapters are added as the author studies the book, making it a potentially growing resource.

Link: https://github.com/boldyshev/sutton AntonioSerrano/Implementation-of-RL-algorithms-from-Sutton-and-Barto-2018 on GitHub: This repository offers implementations in Python using OpenAI Gym and Tensorflow, covering exercises and solutions to complement the book and David Silver's RL course. It includes various algorithms like Dynamic Programming, Monte Carlo, Temporal Difference, and Policy Gradient methods.

Link: https://github.com/AntonioSerrano/Implementation-of-RL-algor...

[ivanbelenky → mark_l_watson]

my code is not as good as anything above most probably. Ive done this exploring while studying. No linter no typechecker, grug engineer mentality. But thanks nevertheless for the comment :)

[mark_l_watson → ivanbelenky]

well, it looks good to me.

[AndrewKemendo]

Let me know if anyone fills out the true online Sarsa section with a working example in a robot

[sage76]

Damn this is a lot of work. Bookmarked.

[ivanbelenky → sage76]

It has not been stress tested, or optimized, tread lightly and thanks a lot for appreciating the work.