Towards Scalable Information Integration with Instance Coreferences

Identifying data instances, that come from differ-ent sources but denote the same entity is neces-sary for effective integration of Semantic Web data. This “instance coreference ” identification problem has gained attention in recent years. Although this is a critical aspect of the overall information inte-gration problem in the Semantic Web, we put for-ward that information integration algorithms also need to be extended in order to work effectively and efficiently in the presence of these coreferenced entities (whether they are discovered by a tool or explicitly stated with an owl:sameAs assertion). We describe such an extension to our Goal Node Search algorithm for Semantic Web information in-tegration.