DSR - Intro to Reinforcement Learning Class

Repo for the "Intro to Reinforcement Learning" Class @ DSR

Slides for the class

• The slides for the class are available here

Set-up of OpenAI gym, the python library for Reinforcement Learning on the command line

pip install gym

to check the installation, you can tape 'import gym' in a python terminal

if this does not work, you can clone directly the repository with the following commands:

• cd
• git clone https://github.com/openai/gym.git
• cd gym
• pip install -e '.[all]'

If you want, you can also create a conda environment and then install gym on it:

• conda create --name rl-env python=3.6 anaconda
• source activate rl-env
• pip install gym==0.7.0

Notebooks for the practical exercises

https://github.com/AMDonati/DSR-RL-class/tree/master/notebooks

Additional resources on RL

• https://github.com/AMDonati/RL-ressources-tutos-2018

Reading activity

Choose one article in the list below, read it in details, and prepare a summary of the article that you will present to the other students:

• Deep Reinforcement Learning doesn't work yet
• World Models
• Pong on Pixels
• AlphaGo original paper