Explaining Non-Entailment by Model Transformation for the Description Logic EL

Research output: Contribution to book/conference proceedings/anthology/reportConference contributionContributedpeer-review

Abstract

Reasoning results computed by description logic systems can be hard to comprehend. When an ontology does not entail an expected subsumption relationship, generating an explanation of this non-entailment becomes necessary. In this paper, we use countermodels to explain non-entailments. More precisely, we devise relevant parts of canonical models of EL ontologies that serve as explanations and discuss the computational complexity of extracting these parts by means of model transformations. Furthermore, we provide an implementation of these transformations and evaluate it using real ontologies.

Details

Original languageEnglish
Title of host publicationIJCKG '22: Proceedings of the 11th International Joint Conference on Knowledge Graphs
EditorsAlessandro Artale, Diego Calvanese, Haofen Wang, Xiaowang Zhang
Pages1-9
Number of pages9
ISBN (electronic)9781450399876
Publication statusPublished - 13 Feb 2023
Peer-reviewedYes

External IDs

Scopus 85148543350
dblp conf/jist/AlrabbaaH22
Mendeley 8c330099-4ff9-3904-a1a2-76a5bafba995

Keywords

Keywords

  • Model Transformation, Explainable AI, Description Logics