Machine Learning in Adversarial Game Using Flight Chess
Yu Liu, Dan Li, Yingsong Hu · 2011
Game playing is a perfect domain of the study of machine learning for its simplicity that allows the researchers to focus on the learning problems themselves and ignore marginal factors. Many learning techniques derived from games have been applied successfully in other learning problems. In this paper, we introduce a Minimax Recurrence Learning algorithm to reinforce the intelligence of a game agent and a supervised learning technique to train the agent. It proves that our intelligent flight chess agent defeat human players in the flight chess game with high probability. Theory deduction proves that combination of the reinforcement learning and supervised learning techniques used in our agent can learn the essential knowledge in an adversarial game. The infrastructure and the algorithm of our agent can be extended in other learning problems also.