Hi! This is a short experiment on Reinforcement learning using the OpenAI gymnasium library. We're going to train an agent to find the shortest path to the goal in the Frozen Lake game using Q-learning. This project was completed as part of the Machine Learning II course for the Big Data Master's Degree program at Comillas ICAI University.
Our team of contributors includes:
| Name | |
|---|---|
| Jorge Ayuso Martínez | [email protected] |
| Carlota Monedero Herranz | [email protected] |
| José Manuel Vega Gradit | [email protected] |
We've created a short animated gif showing the training process of our agent on evaluation:
And here's another one that showcases its performance on 100 evaluations:
If you want to check out our code and reproduce our results, head over to our GitHub repo: https://github.com/carlota-moh/frozen-lake