Assertional Removed Sets Merging of DL-Lite Knowledge Bases

Salem Benferhat (1) , Zied Bouraoui (1) , Odile Papini (2, 3) , Eric Würbel (2, 3)

1 CRIL - Centre de Recherche en Informatique de Lens
2 LIS - Laboratoire d'Informatique et des Systèmes (LIS) (Marseille, Toulon)
3 LIRICA - Logique, Interaction, Raisonnement et Inférence, Complexité, Algèbre

DL-Lite is a tractable family of Description Logics that underlies the OWL-QL profile of the ontology web language, which is specifically tailored for query answering. In this paper, we consider the setting where the queried data are provided by several and potentially conflicting sources. We propose a merging approach, called "Assertional Removed Sets Fusion" (ARSF) for merging DL-Lite assertional bases. This approach stems from the inconsistency minimization principle and consists in determining the minimal subsets of assertions, called assertional removed sets, that need to be dropped from the original assertional bases in order to resolve conflicts between them. We give several merging strategies based on different definitions of minimality criteria, and we characterize the behaviour of these strategies with respect to rational properties. The last part of the paper shows how to use the notion of hitting sets for computing the assertional removed sets, and the merging outcome.

Domaines: Intelligence artificielle [cs.AI]; Logique en informatique [cs.LO]; Complexité [cs.CC]

Salem Benferhat, Zied Bouraoui, Odile Papini, Eric Würbel. Assertional Removed Sets Merging of DL-Lite Knowledge Bases. The 13th international conference on Scalable Uncertainty Management, Dec 2019, Compiègne, France. ⟨hal-02392974⟩