Ontologies for Knowledge Graphs: Breaking the Rules

From korrekt.org

Markus Krötzsch, Veronika Thost

Ontologies for Knowledge Graphs: Breaking the Rules

Abstract. Large-scale knowledge graphs (KGs) abound in industry and academia. They provide a unified format for integrating information sources, aided by standards such as, e.g., the W3C RDB to RDF Mapping Language. Meaningful semantic integration, however, is much harder than syntactic alignment. Ontologies could be an interoperable and declarative solution to this task. At a closer look, however, we find that popular ontology languages, such as OWL and Datalog, cannot express even the most basic relationships on the normalised data format of KGs. Existential rules are more powerful, but may make reasoning undecidable, and normalising them to suit KGs can destroy syntactic restrictions that ensure decidability and low complexity. We study this issue for several classes of existential rules and derive more general syntactic criteria to recognise well-behaved rule-based ontologies over knowledge graphs.

Published at ISWC 2016 (Conference paper)

Download PDF (last update: July 2, 2016)

Citation details


Semantic Web, Rule languages