Application of Reinforcement Learning System to Interactive Digital Art

Abstract

The advance of science and technology has exerted huge impacts on arts and as result a new genre called digital art has emerged. In digital art, interaction is a very important element. Unique interactions, peculiar enjoyments or entertaining elements make digital art works look like a game. Various types of interaction are being implemented in digital art using touch, sound, movement, etc. Some digital art works are even called game art because game-like elements are expressed strongly. In general, interactive works adopting concepts such as ”goal” or ”contest” can be called game art. Digital art is quite similar to games by nature. An important fact is that those having the nature of game induce the spectators' involvement and immersion more than simple interactive works. The reason that artists include interactive components in their works is that they want the spectators' deeper appreciation and involvement. However, many of interactive works fail to induce interaction as much as the artists expect. If visible interaction is weak, spectators are not immersed much in the work. This study purposes to apply and implement to express digital art using the reinforcement learning technology, which is used in games that are highly interactive contents, for the interactivity of digital art works, and to evaluate and discuss the results. In this study, the reinforcement learning theory was applied to our art work ”The Flying,” a flight simulation game created as a digital art work using Kinect. This work is a game art for learning how to fly. It expresses the human desire to fly through the process of a bird's flying training, and in the process the spectator appreciates the work through reinforcement learning.