THE FUTURE IS HERE

I Trained an AI to Play Balatro Using Reinforcement Learning

In this video, I learned Reinforcement Learning by training an AI play Balatro. I used lovely-injector to help modify the game and used Stable-Baslines3 to help me train the agent.

If you’re curious about how reinforcement learning works, then this is a good video to learn because I don’t just show the results but instead I describe how reinforcement learning works from beginning to end. This is not an expert deep dive into reinforcement learning, but this can be a decent/fun introduction video

Hope you enjoy!

Resources:
What is Proximal Policy Optimization – https://www.youtube.com/watch?v=5P7I-xPq8u8
lovely-injector – https://github.com/ethangreen-dev/lovely-injector
Stable-Baslines3 (SB3) docs – https://stable-baselines3.readthedocs.io/en/master/

Credits:
https://www.youtube.com/watch?v=xAXvfVTgqr0
https://www.youtube.com/watch?v=kQ2bqz3HPJE
https://www.youtube.com/watch?v=dp8RgqcENXs

Chapters:
0:00 Introduction
0:11 Disclaimers
0:35 What is Reinforcement Learning
1:32 Going over system
13:59 Training Results
16:53 The AI’s best game