Application of Artificial Intelligence in Playing Cards

Artificial intelligence has invaded into fields of chess playing with the weapon of supervised, deep and reinforcement learning. However, the problem whether artificial intelligence could qualify for the incomplete information system such as playing cards is still open. Efforts have made towards numerous games all of which have professional experience, so it remains the problem of developing a system from scratch without human’s help. In this study, a model combining the Monte-Carlo method and decorated self-Reinforcement Learning was proposed to tackle such problem. To be more specific, a card game called "Zhengshangyou" was chose to verify the proposed method. It requires two players competing each other with a pack of card including the Jokers. In addition, the rules of this game and the designed Artificial Intelligence (AI) algorithm are presented and explained in this study to provide a clear mechanism. The method has promising through theoretical and experimental analysis.