Extending Matching in Description Logics

Talk by Pavlos Marantidis

Location: APB 2026
Start: 27. September 2018 at 1:00 pm
End: 27. September 2018 at 2:30 pm
Research group: Computational Logic
Research group: Automata Theory
Research group: Knowledge-Based Systems
Event series: KBS Seminar
iCal

Matching concept descriptions against concept patterns was introduced as a non-standard inference task in Description Logics motivated by applications in knowledge bases. It was first investigated for the DL FL0, for which a polynomial-time algorithm was obtained. In this talk, we will present recent results that extend matching in FL0 in two aspects. Initially, we investigate approximate matching, that increases the recall of the method by checking whether the input description and pattern can “almost” match, rather than become equivalent. The meaning of almost is formalised using distance measures between concepts. Afterwards, we examine matching in the presence of TBoxes, i.e., background knowledge of concept inclusions. It is well-known that subsumption reasoning in FL0 rises from P to ExpTime when TBoxes are considered, and we prove that the same is true for matching. Finally, we present some preliminary ideas on approximate matching in the presence of TBoxes.