NOTE: Most recent work is the cartpole file in top directory. Previous versions are there, the SARSA implementation is functional as well, and can successfully beat simple maps.

below requires update, but may still be relevant. If you're a newbie interested in Reinforcement Learning (like me) this project may be worth looking at! I'd love to answer any questions/discuss the topic: danekbenjamin@gmail.com

Updates to my process of learning reinforcement learning: The motivation behind undergoing through this rigerous self teaching process is preparing for a project hinged on reinforcement learning. This project requires that I know the underlying math (even the gradient-zation parts) like the back of my hand, and be able to implement it. My math is good, my programming is visible here.

I'm using all kinds of sources to learn this. CS231N Lecture 14 is the foundation, but only covers the theory behind RL https://www.youtube.com/watch?v=lvoHnicueoE

I've found this source to be extremely helpful when it comes to the overview and practical knowledge/heuristics: https://simoninithomas.github.io/Deep_reinforcement_learning_Course/

I'll add other sources as I go along.

Fundamental Dependancies: python 3.5.1 (OpenAI) gym 0.10.9 keras: 2.2.4 (maybe won't need) tensorflow: 1.12.0 (maybe won't need)

...and all other dependancies that come with it (the unmodified list is below) Keras 2.2.4 Keras-Applications 1.0.6 Keras-Preprocessing 1.0.5 Markdown 3.0.1 PyYAML 3.13 Werkzeug 0.14.1 absl-py 0.6.1 astor 0.7.1 certifi 2018.11.29 chardet 3.0.4 future 0.17.1 gast 0.2.0 grpcio 1.17.1 gym 0.10.9 h5py 2.8.0 idna 2.8 numpy 1.15.4 pip 10.0.1 protobuf 3.6.1 pyglet 1.3.2 requests 2.21.0 scipy 1.1.0 setuptools 39.1.0 six 1.12.0 tensorboard 1.12.1 tensorflow 1.12.0 termcolor 1.1.0 urllib3 1.24.1 wheel 0.32.3

I can upload my venv at some point as well, but since it's a large file I choose to only do that once somebody asks: danekbenjamin@gmail.com

[yes I know its a bit untraditional/unusual] version name nomenclature:

[flag]xx.y[name-from-gym]; example a00.0CartPole-v0 [flag] - flag (alpha, represents learning and work-in-progress); other flags tbd xx - version number .y - minor augment to version number [name-from-gym] - the name of import/problem from openai gym

TODO

• update keras and tf dependancies
• update full dependancies list
• check for somebody asking about venv
• update any good sources (OneNote?)