Kinesis

Kinesis is a digital sonic environment that explores the relationship between sound, space, performer,  and computation. A group of AI agents were trained using reinforcement learning to create generative spatial sound objects that react to human presence using their unpredictable behaviors. The result is an ever-changing digital audio-visual structure that transforms the space in an unrepeatable manner. The listeners/performers become active by walking and positioning themselves among the objects and 
indirectly navigating the composition. The agents constantly react to the performer's actions. 
These reactions affect the sonic features such as frequency pitch, duration, and amplitude. While the system interacts with the virtual plane of continuous differentiation of its components, every subjective experience becomes a part of the system by encountering it through sense-making.