Human-Agent Communication in Poker Game

This paper considers both the emotion recognition model under various poker game situations and the partner agent of a seven-card stud poker game, and realize a human-agent communication in a poker game. A human player plays a poker game in cooperation with the partner agent having the emotion recognition model. In a game, the partner agent presents some advices and various facial expressions according to current game situations and human player's emotion recognized by the emotion recognition model when a human player calls the agent for some advice on a game. A neural network model is applied to the emotion recognition model and the soft computing technology including fuzzy theory is applied to the partner agent. This paper also performs subject experiments to confirm whether the communication between a human player and the partner agent is smooth or not and whether the response of the partner agent for human partner's emotion is good or not, where each subject plays games with the partner agent having the emotion recognition model and with the partner agent not having the emotion recognition model.