Context Based Data Abstraction

Conflicts in the world have shifted toward asymmetric warfare. As a result, the methodology for attaining situational awareness has been drastically altered as the number of variables directly impacting situational awareness (i.e. behavioral patterns and social networks), their rates of change, and the rate of data creation have dramatically increased. Providing commanders with only the most vital information has become a key challenge. Therefore, highly adaptive, automated, and flexible tools and systems are required that will provide solutions (i.e. actionable intelligence and situational awareness) on an ongoing basis. To address these issues, we propose the Automated Data Abstraction and Modeling (ADAM) system. ADAM is an automated approach that constructs abstract models for task completion backed by Bayesian Networks – predictive models capable of capturing and analyzing the impact of individual variables on the entire model. Each model will be focused on the completion of a specific task with the minimum amount of data required to ensure accurate results. ADAM will be domain agnostic and capable of producing a wide variety of models for the completion of highly specific tasks across any domain. These models will provide a highly adaptive, real-time analytic engine capable of reducing commanders’ and their staffs’ cognitive load.