Adaptive Toolkit for Pattern Discovery (ATPD)

The ATPD technology is combining new machine learning techniques with advanced rule based methods for automated data utilization pattern discovery.  The technology allows users to monitor information use, discover use patterns, and develop ontology models of these relations and patterns.

This capability enables users to monitor information use across systems, discover usage patterns, analyze patterns for potential new (and useful) concepts and relations, and offer intelligent assistance to enterprise knowledge modelers about the integration of newly learned concepts into evolving reference knowledge models.

Phase II Development

In Phase I of the ATPD initiative, KBSI developed a proof-of-concept ATPD system and demonstrated the system by integrating it with the Air Force Knowledge Management Framework (KMF).  The Phase II initiative extended the ATPD technology and built a focused and field-able ATPD application at the 45th Space Wing, Cape Canaveral Air Force Station (CCAFS), with the goal of helping the Air Force to federate their diverse, large-scale systems.  This ATPD application provides automated support for extracting ontologies from multi-source text data using statistical and text mining techniques, extracting different types of document metadata by applying (data-usage-based) pattern learning techniques and ontology analysis and visualization support for human-in-the-loop ontology editing and enhancement.

KBSI is currently enhancing the ATPD Phase II results to demonstrate the effectiveness of automated knowledge (metadata and ontology) extraction and ontology analysis for key Ballistic Missile Defense Systems (BMDS) semantic technology applications.  The resulting semantic technology will enable automated Enterprise Application Integration (EAI) and systems of systems engineering (SoSE) analysis and design.  In particular, KBSI will demonstrate the ATPD results and its extensions for EAI and SoSE within the context of Missile Defense Agency (MDA) experimental semantic interoperability applications for BMDS.