Visualization of Fighting Game Player Skills Using Imitation Learning Agents

If it is possible to develop a machine-learning algorithm capable of accurately predicting human behavior in a given situation, it is possible to use the machine-learned behavior as an agent that behaves in a human-like manner. Such imitation learning agents can be employed for a variety of educational purposes. Moreover, the use of XAI technology enables the visualization of the key elements that are important to the imitation learning agent. Such a visualization may prove useful for training players of fighting games.