Semantic Middleware

Investigating fundamental issues in entity/relationship extraction, disambiguation (matching & mapping) and annotation.

Semantic annotation of resources supported by an existing world model (the ontology schema that provides an agreed upon and unambiguous model for capturing data and metadata) and knowledge base (ontology instances) follows three primary steps: entity identification, entity disambiguation and annotation. These three steps vary depending on the kind of resource one is trying to annotate. Building a semantic middleware infrastructure with individual functionalities exposed as Web services is one of the goals of this project. By using and evaluating different existing and new approaches, we hope to provide users with an end-to-end system that given resources and one or more ontolgies, facilitates annotation of entities and relationships in the resource.

Pictorial representation of the process of semantic annotation of documents.

• Book chapter on Semantic Annotations in Web services. Abstract
• Entity disambiguation using propagation of reconciliation decisions. "Semantic Analytics on Social Networks: Experiences in Addressing the Problem of Conflict of Interest Detection", 15th International World Wide Web Conference. (pdf)
• Data transformations / mappings between Web services in a Web process. Appendix on WSDL-S, a W3C member submission for Semantic Web services.
• Extracting relationship: Use of ontologies to extract relationships from a collection of documents, thereby providing a more intuitive organization of these documents based on named relationships between entity instances in these documents. Details coming soon. Contact: Cartic Ramakrishnan
• Reading material

Details on individual components coming soon.

Contact: Meena Nagarajan