160 likes | 170 Views
Demonstrating Semantic Mediation for Scientific Applications. Developing a Mediated System. Assume that the mediator software exists Mediation Engineer creates/imports an ontology We have imported an ontology called UMLS from the National Library of Medicine
E N D
Developing a Mediated System • Assume that the mediator software exists • Mediation Engineer creates/imports an ontology • We have imported an ontology called UMLS from the National Library of Medicine • We show a scientists’ tool that augments this ontology by adding “spatial knowledge” to UMLS
Brain Atlas Data (Spatial Vector Models) Ilya Zaslavsky and Joshua Tran
Geometry information was saved successfully! Unified Medical Language System (UMLS): Ontologies developed by the National Library of Medicine to serve as “Knowledge Sources” in the biomedical sciences www.nlm.nih.gov/research/umls
Purkinje neuron Registering My Data
Source Registration • An individual source must “register” itself with the mediator to join the federation • This will be done in two steps • Use a wrapper tool to construct the conceptual schema that should be exported to the mediator • We show a wrapper tool that creates a conceptual model on top of an RDBMS • Physically connect to the mediator and go through the “register me” process
Querying the Source • Once a source is registered • The mediator can query the source based on its schema • This query will go to the wrapper of the source • The wrapper will • Translate the query into the source’s language • Send the query to the source and get results • Translate results back in the mediator’s language and send it to the mediator • We show this for our relational wrapper
The Final Outcome • We show an older prototype of a mediated neuroscience system called KIND