Implementation of the article Online Multi-Task Learning for Policy Gradient Methods by Ammar & al.(2014)

Getting Started

Git clone the repository using

git clone https://github.com/charlesdognin/Online_Multi_Task_Learning

Use

Online_Multi_Task_Learning

And

pip install -r requirements.txt 

Running the models and visualising the learned optimal policy

To choose the number of pg-ella iterations, the number of episodes and the horizon (max 45):

nano parameters.py

Save your file and finally run:

python3 main.py