This repository contains a Texas Hold'em bot trained using a genetic algorithm. The bot was built by George Han and myself as one part of our final project for Math 153 (Evolutionary Dynamics) at Harvard University. The bot significantly outperforms our heuristic-based baseline bot and a random bot, but is not competitive with human poker players.
Additional resources for poker-playing bots, evolutionary algorithms, and reinforcement learning may be found below: Links:
- Evolutionary Dynamics of Reinforcement Learning Algorithms
- DeepStack
- DeepStack implementation in Torch (not PyTorch)
- PyTorch reinforcement learning tutorial
- PyTorch implementation of top reinforcement learning algorithms
We thank our TF Alex Heyde for his assistance and advice.