Introduction to (Deep) Reinforcement Learning

Here are the exercises for the first/second session of our RL course.

Requirements

Create a virtual environment, with python or conda, e.g.

conda create -n rl python=3.10
conda activate rl

Then, update your pip!

pip install --upgrade pip

Clone the repository and get into RL-course-DSR. Now you can install the required packages which are listed in requirements.txt, for example by

pip install -r requirements.txt

Presentation

The presentation of the course can be found here.

Some references

The old testament bible, God him/her/theirselves.

Deep Reinforcement Learning, Sergey Levine.

Introduction to Reinforcement Learning, David Silver.

Reinforcement Learning Virtual School, 2021.

Finally, the real bible!

Some places to go

Environments

• OpenAI Gym environments a collection of RL Hello worlds environments.
• A short list of interesting environments
• A more exhaustive list of environments
• InstaDeep's Jumanji

RL framework

• Stable Baselines is a set of improved implementations of Reinforcement Learning (RL) algorithms based on OpenAI Baselines. You can find good papers here!

• RLlib is an open-source library for reinforcement learning.

• InstaDeep's Mava