In this project we attempt to create AI agents with unexpected, novel and interesting behaviors. We focus on evolving colonies of learning artificial agents to achieve this.
Search algorithms that produce an entire solution before the agent takes its first action lead to increasing action delays as the search graph size increases. Real-time search addresses the problem in a fundamentally different way. Instead of computing a complete, possibly abstract, solution before the first action is to be taken, real-time search algorithms compute (or plan) only the few first actions for the agent to take.
This project covers a wide variety of research, all revolving around deep learning and interacting with sound, such as video game accents, bird song, and musical instrumentation.