Skip to content

Reinforcement Learning - Following Sutton and Barton Book

Notifications You must be signed in to change notification settings

bluephoton/rl-insanity

Repository files navigation

This repo

This repo is created to capture the code I write during my adventure into the world of reinforcement learning. In this adventure I'll be following the tracks left by Sutton and Barto in their fantastic book: Reinforcement Learning, an introduction

This repo also will be the code behind a Medium series I'm working on, where I summarize these concepts so I, or anyone else interested in the subject, can visit later, refresh his memory, or check something he/she found confusing.

Code Structure

So far I organized code into two main parts

Insanity

Bardon the naming, I got bit excited! Is a framework that models the components of RL allowing me to perform experiments in a modular and extensible way. So far it is work in progress but the results are promising.

Idea of Insanity is to isolate specifics of experiment, environment...etc form the core logic of RL and to provide a way to supply different learning algorithms, different ways to initialize or select actions...etc

I don't know much yet, so the more I learn the more updates I'll be adding.

Experiment Code

Specifics about environment, agent, data collected, plotting, analysis...etc

Caveat

My python experience is very limited. And as I'm going through this work - like it or not - I found python to be an aging language that needs complete rehaul and its community and echo system are the only reason it is still alive! So I may switch to something else if I can find it.

About

Reinforcement Learning - Following Sutton and Barton Book

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages