Exploiting the Small-Worlds of the Semantic Web to Connect Heterogeneous, Local Ontologies

Abstract

In this paper, a parallel is drawn between the semantic Web search problem of software agents finding the right ontology definition to how people are able to find strangers using a surprisingly short chain of acquaintances – a result from the “six degrees of separation” experiment. The experiment relied on shared understanding of the phrase, “someone you know on a first name basis” to define an acquaintance relationship. Web searching relies on standardized use of the hyperlink relationship. Hyperlinks are constituted from universally accepted meta-data: Anchor and bookmark HTML markups. Say that heterogeneous local ontologies are all marked-up using standard meta-data. Then, the meta-data and some universally accepted semantics constitute a shared ontology, which can be used to bridge local ontologies, much as highly connected people who belonged to many cliques (small-worlds) were used disproportionately often in the search for strangers. This paper outlines the framework for approaching the semantic Web search problem using meta-data based shared ontologies inspired from small-worlds theory of sociology. This approach is exciting because it (1) enables data sharing over the semantic Web without post hoc modifications to local ontologies, and (2) uses meta-data, which in many situations are already commonly available and implemented in XML.