On Coreference and the Semantic Web

Abstract

Much of the SemanticWeb relies upon open and unhindered interoperability between diverse systems; the successful convergence of multiple ontologies and referencing schemes is key. However, this is hampered by the difficult problem of coreference, which is the occurrence of multiple or inconsistent identifiers for a single resource. This paper investigates the origins of this phenomenon and how it is resolved in other fields. With this in mind, we have developed and tested an effective methodology for coreference resolution in the Semantic Web at large. This framework allows the user to a) record identified instances of coreference in a usable and retrievable manner b) integrate new and existing systems for reference management, and c) provide a thesaurus-like consistent reference service capable of providing on-tap resolutions to interested applications