ACC-CS-CLUB-ML

Austin community college computer science club group machine learning project for poker and hidden information games

Welcome

Welcome to the ACC CS poker ML group projects. Do get started, please read contributing.md

Goals

The goal of NitDestroyer is to find a Nash Equilibrium for poker. We will begin with Kuhn Poker and gradually move up to Rhode Island Hold'em (RIH), Limit Texas Hold'em (LTH), and No Limit Texas Hold'em (NLTH). We are specifically targeting cash game play, but will also be looking at tournament play.

Rescources and References

Kuhn Poker Rules Rhode Island Hold'em Rules Limit Texas Hold'em Rules No Limit Texas Hold'em Rules Noam Brown's Libratus overview video Noam Brown's Libratus paper Meta's Pluribus paper Deep Counterfactual Regret Minimization

Roadmap

Github Project Board

• Create a repositorys
• Setup PR permisions
• Create version control
• Create contributing.md rules and PR template
• Choose a playing card emulator
• Build a poker module from scratch (player class, game class, etc.)
• Create a poker emulator from scratch
• Write unit tests
• Build a data logger
• Build Kuhn Poker
• Solve Kuhn Poker
• Brainstorm data collection and storage
• Create randomized player agents
• Begin ML training on randomized player agents (ultra simple algorithm)
• Improve ML algorithim (Monte Carlo Counterfactual Regret Minimization)
• Create Baselines