Regularization Methods for Reinforcement Learning

This repository contains implementations of selected regularization methods for reinforcement learning algorithms.

The base implementations of the reinforcement learning algorithms were adapted from the Stable-Baselines3 library.

This repository accompanies the master's thesis “Comparison of Regularization Methods in Reinforcement Learning Algorithms” submitted at Jagiellonian University.

• Official thesis record in the Jagiellonian University Repository
• Experimental results on Google Drive

Installation:

To install all dependencies, please run the following command:

install.sh

Run

To run the script, use the main.py file. `