In this project, we will develop an infrastructure, Expozé-DATALATOR, by extending our existing Expozé tool suite for discovery of mapping relationships. In addition, in order to realize our goals for building effective transformation routines for the Expozé-discovered mappings, we will utilize the fuselet development support environment of Apollo a Rome Labs-developed JBI (Joint Battlespace Infosphere) reference implementation to test emerging integration technologies for dynamic air operations planning and scheduling. We propose to demonstrate the feasibility of crafting mapping ontologies as an abstract layer over JBI information object types through Expozé-DATALATORs framework. The proposed mappings will be authored in a high-level language such as OWL/CL and will unfold as XSLT-based transformations underneath the hood. We will also provide a graphical front-end for authoring the complex relationships such that the information transformation components of the JBI can be made much more generic and reusable. These components will be useful for composition into information transformation modules that operate over various relationships from several datasets, thus obviating the need to hard code the mappings into each and every transformation program. Intelligence operations can thus utilize the relationships-based information generated by Expozé-DATALATOR to compare the expected effects of ongoing operations with the actual effects observed during operations so that commanders can reorient and adapt their future actions quickly. We will utilize Apollos benchmarks to test out the efficacy of our joint technologies.
Keywords: Clustering, Mapping Relationships, Discovery, Knowledge Capture, Semantic-Mediation, Apollo, Fuselets
