Complex Event Processing over RDF Triple Stores
Share this Session:
  Mark Gerken   Mark J Gerken
Chief Scientist, Strategic Initiatives Division
Intelligent Software Solutions, Inc.


Wednesday, June 6, 2012
04:30 PM - 05:15 PM
Level:  Business/Strategic

Location:  Plaza B

Researchers at Intelligent Software Solutions, Inc. have recently developed a description logic based Complex Event Processing (CEP) system capable of operating directly against RDF graphs and triple stores. This system was initially designed to fuse near real-time multi-source data against event-based models in support of predictive analysis but can also support causal/diagnostic analysis and association discovery. Missed detections are minimized via the application of Bayesian Nets to handle missing evidence as well using fuzzy Logic to handle partial evidence matching. False detections are minimized via dynamic context propagation which dynamically generates precise event descriptions based on the context provided by existing evidence. Because event data is inherently temporal and dynamic, several novel knowledge management approaches were introduced to facilitate temporal reasoning over fluents without requiring custom OWL/RDF extensions or a custom reasoner. This talk will also address planned enhancements and future capabilities.

• A domain independent Complex Event Processing (CEP) system that facilitates near real-time multi-INT fusion and predictive analysis against RDF graphs and triple stores
• Minimizes missed detections related to missing & partial data
• Bayesian Nets used to support missing evidence
• Fuzzy Logic used to handle partial evidence matching
• Minimizes false detections via dynamic context propagation in which precise event descriptions are generated dynamically based on the context provided by existing evidence
• Supports historical and real-time analysis and reasoning over changing data (truth maintenance)

Currently, Dr. Gerken functions as a Chief Scientist at ISS supporting advanced research and development of data access, machine learning, knowledge extraction and representation, and data fusion. For the past several years he has support research and development of knowledge exploitation systems based on both relational and ontological data. Prior to joining ISS, Dr. Gerken served as over 21 years as an Air Force officer with assignments ranging from research and development to test and evaluation. His last assignment was as a Deputy Department Head in the Department of Mathematical Sciences at the United States Air Force Academy where he taught a variety of courses including statistics, operations research, and engineering mathematics.

Close Window