Top-Down Abstraction Learning Using Prediction as a Supervisory Signal

We present a top-down approach for learning abstractions whereby a robot begins with a coarse representation of the world and incrementally finds new distinctions as they enable the robot to better predict its environment. The approach has been implemented on a simulated robot that learns new distinctions in the form of variable discretizations through autonomous exploration. This paper discusses how to generalize this approach to learning broader abstractions.