Similar to content on the web, scientific data is highly heterogeneous and can benefit from rich semantic descriptions. We are particularly interested in developing an infrastructure for expressing explicit semantic descriptions of ecological data (and life-sciences data in general), and exploiting these descriptions to provide support for automated data integration and transformation within scientific workflows. Using semantic descriptions, our goal is to provide scientists with: (1) tools to easily search for and retrieve datasets relevant to their study (i.e., data procurement), (2) the ability to select a subset of returned datasets as input to a scientific workflow, and (3) automated integration and restructuring of the selected datasets for seamless workflow execution. As part of this effort, we are developing the Semantic Mediation System (SMS) within the SEEK project, which aims at combining knowledge representation and semantic-web technologies (e.g., OWL and RDF) with traditional data-integration techniques. We observe that along with these traditional approaches, mediation of ecological data also requires external, special-purpose services for accessing information not easily or conveniently expressed using conceptual modeling languages, such as description logics. The following are two specific examples of ecologically relevant, external services that can be exploited for scientific-data integration and transformation.
展开▼