CGCore: Common-Ground Contextual Operator-oriented REasoning: Machine Assisted Symbiotic Intelligence Frameworks for Representing, Understanding, Inte

CGCore will develop and demonstrate novel methods and algorithms which improve the ability of artificial intelligence and machine learning to communicate contextual understanding to human operators allowing them to better understand relevant information about entities, interactions, and their environment enabling more effective decision-making, command, and control. CGCore will enable an increased level of interoperability for human-machine and machine-machine interactions by preserving contextual meaning across differences in common ground, information formats, and domains enabling decision-makers to support naval missions with actionable intelligence from open source intelligence (OSINT) and multi-intelligence data sources despite differences in size, completeness, and structure. We will demonstrate CGCore on a large data set of multi-modal, maritime relevant, information on global transshipment operations, targeting adversarial actions related to AIS disabling events and other illegal activity. CGCore utilizes cutting edge developments from our prior work with the Air Force Research Labs to establish contextual knowledge dynamically and lexically with human operators, and our prior work with the Defense Advanced Research Projects Agency on the construction of representations of human contextual reasoning for symbiotic AI. CGCore will substantially improve human-machine and machine-machine interactions, shortening current ops-floor course of action decision timelines from a day, to less than an hour.