Toolkit for Agent-based Knowledge Extraction (TAKE)

We propose to research, design, and develop an innovative Toolkit for Agent-based Knowledge Extraction (TAKE). We propose a hybrid approach that combines emerging knowledge discovery techniques with traditional ontology engineering methods to extractknowledge from emerging distributed and disparate data sources and rapidly deploy this knowledge for the analysis, reengineering, and redesign of complex multi-agent systems. The Phase I effort will (i) establish TAKE requirements, (ii) design TAKEarchitecture, (iii) build and demonstrate a prototype TAKE software. The Phase II project will harden the software and demonstrate its benefits on focused military and commercial applications. Key innovations include: (i) hybrid approach to combine softcomputing techniques with ontology engineering methods for automated and rapid knowledge extraction and analysis, (ii) innovative self-adaptation mechanisms that ensures automated revisions to knowledge base and the knowledge analysis strategies inresponse to rapidly changing application environments, (iii) scalable and re-configurable distributed agent architecture that enables rapid and cost effective knowledge based systems reengineering and redesign. The principal project benefits are: (i)significant reductions in time and effort needed for extracting knowledge from distributed data and knowledge sources; (ii) significant reductions in time and effort needed to reengineer and redesign complex software systems; and (iii) the ability toexplore a significantly larger number of system design alternatives at the same cost and time. Applications include systems reengineering, knowledge management, knowledge discovery, and document management