Why this and not that? A Logic-based Framework for Contrastive Explanations
This work addresses the need for structured contrastive explanations in logic-based systems, which is incremental as it builds on existing methods by formalizing and analyzing specific problems.
The paper tackles the problem of generating contrastive explanations by defining canonical problems that answer 'Why P but not Q?' and computing causes for both P and Q to compare differences, showing that their framework captures a cardinality-minimal version of existing explanations and analyzing computational complexities with implementation examples.
We define several canonical problems related to contrastive explanations, each answering a question of the form ''Why P but not Q?''. The problems compute causes for both P and Q, explicitly comparing their differences. We investigate the basic properties of our definitions in the setting of propositional logic. We show, inter alia, that our framework captures a cardinality-minimal version of existing contrastive explanations in the literature. Furthermore, we provide an extensive analysis of the computational complexities of the problems. We also implement the problems for CNF-formulas using answer set programming and present several examples demonstrating how they work in practice.