Explanation in the DL-Lite Family of Description Logics

• Authors: Alexander Borgida;Diego Calvanese;Mariano Rodriguez-Muro
• Venue: OTM '08 Proceedings of the OTM 2008 Confederated International Conferences, CoopIS, DOA, GADA, IS, and ODBASE 2008. Part II on On the Move to Meaningful Internet Systems
• Year: 2008

In ontology-based data access (OBDA), access to (multiple)incomplete data sources is mediated by a conceptual layerconstituted by an ontology. In such a setting, to correctly computeanswers to queries, it is necessary to perform complex reasoningover the constraints expressed by the ontology. We consider thecase of ontologies expressed in DL - Lite, a familyof DLs that, in the context of OBDA, provide an optimal tradeoffbetween expressive power and computational complexity of reasoning;notably conjunctive query answering is LOGSPACE in the size of the data. However, queryanswering with reasoning comes at a price: the justification of thepresence of tuples in answers is no longer trivial, and requiresexplanation. In this paper, we characterize reasoning inDL - Lite, through deduction rules for buildingproofs, and we provide several novel contributions:(i) For standard ontology level reasoning, explanationis relatively simple, and our contribution comes mainly from anovel focus on brevity of proofs. (ii) Motivated by theuse of DL - Lite for OBDA, we analyze and provideexplanation for reasoning in finite models. (iii) Weprovide a facility for the explanation of an answer to aconjunctive query over a DL -¿ Lite ontology.This algorithm is able to exploit the relational query engine toextract from the data the information necessary for finding theexplanation more efficiently, and thus scales to large data sets.The presented approach has been implemented in a prototype forconstructing explanations.