Conference on Semantics in Healthcare and Life Sciences (C-SHALS)
Advanced Tech Talk
Presentation Vendor: IO Informatics
TITLE: The Semantic Automated Discovery and Integration (SADI) framework transparently exposes Web Service output as a Semantic Web resource – Application towards medical advances in cardiac diseases. (http://cardioshare.icapture.ubc.ca/)
SADI is a technology designed to expose Web Services and their output as queryable Semantic Web resources. We demonstrate SADI as a plug-in to the IO Informatics Knowledge Explorer user-interface as an example of generating disease-specific knowledgebases to advance understanding of heart disease. ABSTRACT: The "Deep Web' - all of the Web data accessible only through Web forms - is believed to contain hundreds to thousands of times more data than the Web itself. Here we demonstrate that Deep Web data, when exposed as Semantic Web Services, can be discovered, explored, and queried as if it were a typical Semantic Web resource. Our SADI framework interprets SPARQL queries, determines which Semantic Web Services are capable of generating the data needed to answer that query, and then executes those services in order to dynamically generate the database used for query resolution. Moreover, well-defined OWL-DL classes referred-to in the query are deconstructed to determine their defining property-restrictions, and these properties are also used for Semantic Web Service discovery.
This allows "naive" data to be analysed for potential classification into these OWL classes, and thus utilized for query answering. We have created an implementation of SADI as a plug-in to the Knowledge Explorer interface from IO Informatics. This provides an intuitive way for end-users to visualize and explore the data retrieved from SADI, as well as execute SADI Semantic Web Services through a simple query interface. The Use and application of this approach towards medical advances in the CardioSHARE (http://cardioshare.icapture.ubc.ca/) initiative are demostrated and discussed.
This presentation demonstrates the utility and power gained by utilizing the property restrictions in OWL-DL, rather than creating simple asserted class-hierarchies. Moreover, it demonstrates the ability of machines on the Semantic Web to re-interpret and utilize data in ways that were not anticipated by the data provider.