Wednesday, June 6, 2012
04:30 PM - 05:15 PM
|Level: ||Case Study|
|Location: ||Franciscan B|
Tactical communications provide commanders with significant intelligence during military operations. In general, however, the situational updates contained in these "voice communications" are not automatically translated into structured knowledge that can be used to enhance situational awareness. We present results of research performed for the US Navy in (1) analyzing the semantics of (unclassified) tactical communications transcripts, (2) designing an ontology, vocabulary and grammar for automated knowledge extraction from transcripts, and (3) developing a prototype knowledge extraction capability for automated situational updates. Specifically, we present:
• An example (unclassified) tactical communications transcript;
• Semantic analysis of the transcript metadata, contents, and process flow;
• Development of the domain vocabulary, ontology and text extraction grammars;
• Prototype results, including knowledge extracted from the transcript and represented as RDF in Allegrograph;
• Current status and next steps.
Dr. Bimson is the Vice President of Research at Modus Operandi, Inc., where he oversees the company’s semantic technology research activities. His technical specialties include linguistics, semantic analysis, ontology modeling, enterprise information integration (EII) and fusion, natural language processing, and decision support systems. Dr. Bimson has pioneered work in Ontology Life Cycle Management and has authored and delivered workshops on ontology design and development and information service architectures to numerous DoD, state, and commercial organizations. He is a founder of the USAF 45th Space Wing’s Knowledge Management Initiative (KMI), the goal of which is to enrich operational information stores semantically to improve the information integration and understanding needed to enhance decision quality. Dr. Bimson is on the faculty of Intervista Institute, where he teaches workshops that relate semantics to business architectures. He holds a Ph.D. in Linguistics from UCLA and a M.S. in Computer Science from California State University, Sacramento.
Dr. Teresa Nieten is a Scientist at Modus Operandi, Inc, where she serves as Principal Investigator on several Phase I and II SBIRs. Her research has focused on the extraction of entities and events of interest from natural language text, semantic analysis, and fusion of multi-intelligence data for situational awareness and decision support. Dr. Nieten has 20 years of experience in commercial, DoD and aerospace industries. Prior to her work at Modus Operandi, Dr. Nieten was a lead software engineer at Interoptek, Inc., and Intelligent Innovations, Inc., where she led the development of a framework for autonomous, collaborative behaviors for unmanned systems. She has also held positions with SAIC and Harris Corp. Dr. Nieten graduated from the University of Florida in 2011 with a Ph.D. in Computer Engineering, with a concentration on artificial intelligence.