yfletberliac / rlss-2019

Materials for the Practical Sessions of the Reinforcement Learning Summer School 2019: Bandits, RL & Deep RL (PyTorch).

RLSS 2019: Pratical Sessions

Setup and Installation

Two choices are available to you:

• running the notebooks on Google Colab 📙 if you want to take advantage of the GPU acceleration it offers;
• running the notebooks elsewhere (locally or on a server).

Google Colab

It has it's own VM so you only have to install the necessary packages from inside the notebooks.

Elsewhere

You can use the rlss2019-docker image. Here you'll find the instructions for installing and running the rlss2019-docker image on Linux, MacOS or Windows.

Materials

Bandits

• Stochastic Bandits 📙
• Recommender Systems 📙

Reinforcement Learning

• Dynamic Programming + QLearning + SARSA 📙

Deep Reinforcement Learning

• REINFORCE + A2C 📙
• DQN 📙
• Model-Based 📙

Final Project

• Intro 📙
• TextWorld 📙
• MiniWoB

Misc./Known issues

You are running Windows and want to install a Virtual Machine running Ubuntu 18.04? Here is a tutorial. You may also want to directly install the Ubuntu terminal on Windows 10.

Contributors

• Raphaël Avalos
• Geoffrey Cideron
• Omar Darwiche Domingues
• Yannis Flet-Berliac
• Emilie Kaufmann
• Max Lapan
• Edouard Leurent
• Odalric-Ambrym Maillard
• Jérémie Mary
• Mathieu Seurin