Computing instantiated explanations in OWL DL

Abstract

Finding explanations for queries to Description Logics (DL) theories is a non-standard reasoning service originally defined for debugging purposes but recently found useful for answering queries to probabilistic theories. In the latter case, besides the axioms that are used to entail the query, it is necessary to record also the individuals to which the axioms are applied. We refer, in this case, to instantiated explanations. The system BUNDLE computes the probability of queries to probabilistic ALC knowledge bases by first finding instantiated explanations for the query and then applying a dynamic programming algorithm. In order to apply BUNDLE to more expressive DLs, such as SHOIN(D) that is at the basis of OWL DL, instantiated explanations must be found. In this paper, we discuss how we extended BUNDLE in order to compute instantiated explanations for SHOIN(D). © Springer International Publishing Switzerland 2013.