Absorption for ABoxes with Local Universal Restrictions.
Jiewen Wu, Taras Kinash, David Toman, Grant Weddell · 2013
Abstract. We elaborate on earlier work in which we developed a novel method for evaluating instance queries over DL knowledge bases that derives from binary absorption. An important feature of this earlier method and its refinement in this paper is that they avoid the need to check explicitly for consistency, a property that is desirable, for example, in SPARQL query evaluation over RDF data sets that can dynamically include sophisticated ontologies. In particular, we resolve a number of outstanding issues with the earlier method that limited its capabilities for knowledge bases that involve an extensive use of typing constraints expressed as axioms of the form A ⊑ ∀R.B, or that require and use both role hierarchies and transitive roles. We also show how our more general method supports a safe use of nominals in instance queries, and how the method can therefore be used to evaluate basic graph patterns in the SPARQL query language. Finally, we present the results of a preliminary experimental evaluation that validates the efficacy of our more refined method for instance checking. 1