|Toolkit for Agent-based Knowledge Extraction (TAKE™)|
KBSI used a hybrid approach combining ontology engineering methods with cutting-edge knowledge discovery techniques to extract, analyze, and map ontologies from knowledge and data sources. The TAKE™ technology helps users to rapidly and cost effectively analyze, redesign, integrate and/or reengineer their enterprise systems.
This initiative developed an innovative Toolkit for Agent-based Knowledge Extraction (TAKE™). The TAKE™ technology applies a hybrid approach that combines ontology engineering methods with cutting-edge knowledge discovery techniques to extract, analyze, and map ontologies from distributed and disparate knowledge and data sources. These maps can then be used for the analysis and redesign of complex, multi-agent systems.
The TAKE™ technology provides innovative solutions that address the problems associated with rapidly designing, developing, and deploying large scale knowledge based applications. KBSI's overall approach in the initiative conceptualized and validated a framework and a methodology for sustainable knowledge management that encompasses the entire knowledge life cycle: capture/discovery, representation, analysis, validation, storage, retrieval, (in-context) delivery, and evolution over extended periods of time.
In the early phase of the initiative, KBSI established a novel TAKE™ ontology reuse and application method, designed innovative ontology analysis and ontology extraction algorithms, and designed the TAKE™ architecture. These innovations include the novel application of information fusion techniques for aggregated ontology similarity assessment and mapping the application of advanced knowledge discovery techniques for automated ontology extraction and mapping from multi-source text data. A demonstration of a prototype TAKE™ software tool was also demonstrated on a focused, high payoff Navy transition application scenario.
In Phase III, KBSI provided the requirements definition, design, research, advanced development, modeling, test, and training/mentoring for integrating the TAKE™ technology in Navy shipyards. This work included demonstrating the application of TAKE™ ontology-based technology capabilities to shipbuilding production planning challenges and the migration of the shipyard part catalog in accordance with Navy standards.
KBSI first applied TAKE™ semantic mapping to the production of an initial shipbuilding production process and plan description lexicon and meta-ontology, to the development of modeling and semantic mapping methods, and to address the shipyard CPC legacy part migration. These requirements were validated at Tier 1 and Tier 2 shipyards and ship design software vendors who support interfaces to the Tier 2 CPC system. KBSI next researched method and algorithm extensions to support the defined requirements, and then designed and developed the TAKE™ architecture, algorithms, methods, and user interface enhancements in collaboration with the Tier 1 and Tier 2 community-of-interest. The enhancements developed in Phase III demonstrated how TAKE™ technology can be leveraged to help U.S. shipbuilders better achieve the goals of decreasing costs and increasing productivity.
Phase II.5 Development
The TAKE™ Phase II.5 initiative is transitioning TAKE™ initiative results to KBSI's ONR-funded Intelligence Rapid Product Generator (iRPG) technology with the goal of addressing technology gaps in iRPG.
The iRPG framework provides warfighter-enabled integration of disparate services for generating actionable intelligence products.This capability significantly improves real time decision making through integrated algorithms and knowledge sharing, provides a “common understanding” framework for effective information integration, and supports services-based interoperability of component products.The current configuration of the iRPG prototype provides intuitive user interfaces to enable the creation of war-fighter evolved applications that broaden accessibility to analytical algorithms and promote development of applications to meet particular intelligence product requirements.
Specifically, the TAKE™ II.5 initiative is addressing several iRPG technical voids: dynamic ontology knowledge management, dynamic process knowledge management, and dynamic semantic knowledge discovery. At the heart of this capability is KBSI’s MODELMOSAIC® Knowledge Management Framework, produced during the TAKE™ initiative. The MODELMOSAIC® framework facilitates dynamic ontology and workflow knowledge discovery, creation, analysis and maintenance. The framework provides workflow capture, modeling, and analysis; ontology capture, modeling, and analysis; the automated generation of enterprise analysis models; and the automated generation of translators that enable information integration.
The TAKE™ II.5 initiative is enhancing the MODELMOSAIC® tool’s ontology knowledge management capabilities, which includes the capability to suggest ontologies for extension, advanced ontology development, verification and validation support, and the ability to extract concepts and relationships from text sources.
In addition, TAKE™ II.5 is developing theMODELMOSAIC® tool’s process knowledge management capabilities, including the integration of IDEF3 modeling with IDEF5 ontology modeling and the ability to describe workflows as semantic services, to extract process concepts and ordering from text sources, and to support workflow validation and archiving. With respect to dynamic semantic knowledge discovery, enhancements to theMODELMOSAIC® tool include the capability to suggest semantic service descriptions and to provide semantic service description validation and repository refinement and maintenance. The tool will also provide advanced semantic service mappings and mappings repository support.
The initiative is currently implementing an initial capability for modeling and compiling Oozie workflows from high-level semantic workflow descriptions. This includes registering Oozie workflows and creating a “workflow of workflows” that contains registered Oozie workflows. KBSI is also testing methods for making the creation of workflows easier for analysts. TheMODELMOSAIC® tool’s IDEF0 (function/activity modeling), IDEF3 (process modeling) and IDEF5 (ontology modeling) methods provides a means for achieving this and integrating these methods. KBSI has also developed the initial implementation of a capability to semi-automatically fill in the details (service mappings, data flow mappings) of a workflow template—another method simplifying workflow creation.
The innovative TAKE™ knowledge extraction, analysis, and reuse methods and semantic mapping and search methods will provide immediate short-term benefits to the Navy, particularly in collaboration with the ONR-funded iRPG project activities.