PAC Learning of Concept Inclusions for Ontology-Mediated Query Answering

Sergei ObiedkovSergei Obiedkov, Baris SertkayaBaris Sertkaya

ISSN: 0888-613X
Download
DOI Link

Sergei Obiedkov, Baris Sertkaya
PAC Learning of Concept Inclusions for Ontology-Mediated Query Answering
International Journal of Approximate Reasoning, 186:109523, November 2025

Details
Bibtex

KurzfassungAbstract
We present a probably approximately correct algorithm for learning the terminological part of a description-logic knowledge base via subsumption queries. The axioms we learn are concept inclusions between conjunctions of concepts from a specified set of concept descriptions. By varying the distribution of queries posed to the oracle, we adapt the algorithm to improve the recall when using the resulting TBox for ontology-mediated query answering. Experimental evaluation on OWL 2 EL ontologies suggests that our approach helps significantly improve recall while maintaining a high precision of query answering.
Weitere Informationen unter:Further Information: Link
Projekt:Project: Cfaed, SECAI, ScaDS.AI
Forschungsgruppe:Research Group: Wissensbasierte SystemeKnowledge-Based Systems

@article{OBIEDKOV2025109523,
title = {PAC learning of concept inclusions for ontology-mediated query answering},
journal = {International Journal of Approximate Reasoning},
volume = {186},
pages = {109523},
year = {2025},
issn = {0888-613X},
doi = {https://doi.org/10.1016/j.ijar.2025.109523},
url = {https://www.sciencedirect.com/science/article/pii/S0888613X25001641},
author = {Sergei Obiedkov and Barış Sertkaya},
keywords = {Ontologies, Description logics, Active learning, Knowledge acquisition, PAC learning},
abstract = {We present a probably approximately correct algorithm for learning the terminological part of a description-logic knowledge base via subsumption queries. The axioms we learn are concept inclusions between conjunctions of concepts from a specified set of concept descriptions. By varying the distribution of queries posed to the oracle, we adapt the algorithm to improve the recall when using the resulting TBox for ontology-mediated query answering. Experimental evaluation on OWL 2 EL ontologies suggests that our approach helps significantly improve recall while maintaining a high precision of query answering.}
}