Towards Formal Concepts for Explanation Timing and Justifications

Abstract

Explainability of automated systems gains increasingly more importance, since these systems get more complex and at the same time we — individually and as a society — become more reliant on them. There is a vast amount of works on explainability in the field of computer science alone, but there is little work on formal concepts to model and design explanations. In order to contribute to paving the way for model based design of explanations, we formally approach two main questions "How to provide a justified explanation?" and "When to provide an explanation?".