Cycles Of SiTuation Update with ModEls (COSTUME)

An extensible system for level 2 and 3 data fusion utilizing new classes of Behavioral, Situation and Threat Models will be developed. The framework for this system is based on a model of the human analytic process in which cycles of situation update and revision are performed as evidence arrives incrementally. We argue that people learn a combination of localized information, that we call a mental space, which provides the models people need to update their understanding of the current situation as information arrives incrementally. People gain computational efficiency and flexibility by using a hierarchy of these mental spaces and rules for transitioning between them. We have already developed portions of a system for performing cycles of situation update on incrementally arriving intelligence reports using an iGENTM cognitive model. In this new effort, we will create an autonomous and flexible data fusion system that performs Cycles Of SiTuation Update with ModEls (COSTUME) by adding mental spaces and mechanisms for utilizing them to the existing model developed as part of our effort on Rome Labs' EAGLE program. We will also use an interactive gaming environment to create a set of simple threat scenarios against which COSTUME can be tested.