Combinatorial Methods for Explainable AI

Abstract

This short paper introduces an approach to producing explanations or justifications of decisions made by artificial intelligence and machine learning (AI/ML) systems, using methods derived from fault location in combinatorial testing. We use a conceptually simple scheme to make it easy to justify classification decisions: identifying combinations of features that are present in members of the identified class and absent or rare in non-members. The method has been implemented in a prototype tool, and examples of its application are given.

Topics & Concepts