Why this and not that? A Logic-based Framework for Contrastive Explanations

2025/07/11 by Geibinger, Tobias, Jaakkola, Reijo, Kuusisto, Antti +2 · 1 citation
#03B05 #68T27 #Artificial Intelligence (cs.AI) #F.4.1 #FOS: Computer and information sciences #I.2.3 #Logic in Computer Science (cs.LO) #Machine Learning (cs.LG)

paper · doi:10.48550/arxiv.2507.08454

Abstract

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.

Citations

• Local vs. Global Interpretability: A Computational Complexity Perspective
• Logic for Explainable AI
• Logic-Based Explainability in Machine Learning
• Explainability via Short Formulas: the Case of Propositional Logic with Implementation
• Model Interpretability through the Lens of Computational Complexity
• On The Reasons Behind Decisions
• Abduction-Based Explanations for Machine Learning Models
• Explanation in Artificial Intelligence: Insights from the Social\n Sciences
• Explanation in artificial intelligence: Insights from the social sciences

Cited by

• Additive Models Explained: A Computational Complexity Approach

Related

• Can You Tell the Difference? Contrastive Explanations for ABox Entailments
• Efficient computation of contrastive explanations
• Contrastive Explanations for Argumentation-Based Conclusions
• Prompting Contrastive Explanations for Commonsense Reasoning Tasks
• Reason to explain: Interactive contrastive explanations (REASONX)