Neural Network based Constraint Satisfaction in Ontology Mapping

Ontology mapping seeks to find semantic correspondences between similar elements of different ontologies. Ontology mapping is critical to achieve semantic interoperability in the WWW. Due to the fact that ubiquitous constraints (e.g., hierarchical restrictions in RDFS) caused by the characteristics of ontologies and their representations exist in ontologies, constraints satisfaction has become an intriguing research problem in ontology mapping area. Though different techniques have been examined to find mappings, little work is made to solve constraint satisfaction problem for ontology mapping. Currently most approaches simply validate ontology constraints using isolate heuristic rules instead of comprehensively considering them in a global view. This paper proposes a neural network based approach to search for a global optimal solution that can satisfy ontology constraints as many as possible. Experimental results on OAEI benchmark tests #248-#266 show the approach is promising. It dramatically improves the performance of preliminary mapping results.