Evolutionary Computation for Poker AI

Our goal in this project was to teach a computer how to play Texas hold 'em poker using principles of artificial intelligence. The plan was to give the computer the rules of poker and, by using artificial neural networks and genetic algorithms, have the computer teach itself how to play poker at a respectable level. So far as possible, we avoided using huge amounts of domain knowledge (human knowledge on how to play poker well). Through many generations, storing and reproducing the strong genes and discarding the weak genes, our aim was to evolve better and smarter poker players. Varying and fine-tuning our techniques allowed us to find what worked best.