Introduction and Motivation

Abstract

From a computational perspective, it is fascinating to see how fast humans or animals adapt to new, unforeseen situations. While an ordinary computer program can only handle foreseen situations it is applied to, humans are able to quickly estimate suitable behavior for new, unknown scenarios. A notable exception are Machine Learning (ML) algorithms that learn behavior based on some kind of optimization criterion. “Behavior of a computer program” may be a simple yes or no decision, or a rational choice of a robot’s next action.