Semantic Web

SPONSORSHIP

DARPA, NSF/ITR

BRIEF DESCRIPTION

The Semantic Web  (SW) is a collaborative effort led by W3C with participation from a large number of researchers and industrial partners, aimed at providing a common framework that allows data to be shared and reused across application, enterprise, and community boundaries by defining and liniking data in a way that it can be used and understood by machines as well as humans. It is based on the Resource Description Framework (RDF), which integrates a variety of applications using XML for syntax and URIs for naming. Its Web Ontology Language (OWL) , based on description logics, has been recommended by W3C and become the  industry standard for ontology definition.
    The primary interest of my research in this area is uncertainty in SW: how to represent uncertain relations between concepts/classes of interest, how to handle partial subsumptions in ontological reasoning, and most imporantly, how to mapp/translate concepts between independently defined ontologies. To address these issues, we take Bayesian approach. We are currently working on developing a framework called Bayes OWL, which extends OWL and translates an OWL ontology to a Bayesian network (BN) on which the ontological reasoning tasks can be cast as probabilistic inferences with existing BN algorithms. We are also working on formulating ontology mapping/translation between two ontologies as evidential reasoning between the BNs translated from the two ontologies.
    Other research activities include incorporating SW with agent technology, SW applications to E-Commerce, and the development of Swoogle, a search and metadata engine for the semantic web.

RELATED PAPERS

FOR MORE INFORMATION

Contact Yun Peng, ypeng@umbc.edu .