Blackboard Pattern-Enabled Integrated Intelligent Toolset

To address the U.S. Army need for an automated toolset to identify and track high-value individuals (HVIs) and their organizations from unstructured HUMINT, COMINT, IMINT, and other intelligence sources, Physical Optics Corporation (POC) proposes to develop a new Blackboard-pattern-enabled Integrated Intelligent Toolset (BIIT) that detects and identifies multiple HVIs from unlinked, unorganized, diverse intelligence data. The BIIT, based on the synergetic integration of individual intelligent agents, provides a framework for fusing and analyzing all sources of data in associative memory, where the datasets are fused. In Phase I, POC has established the BIIT architecture, determined the criteria that characterize a “significant” individual, and demonstrated the BIIT performance on both Army-simulated HUMINT/COMINT data and Enron e-mail archives, which resulted in successful identification of HVIs from each test dataset based on significance criteria. In Phase II, POC will develop a near-real-time BIIT system with more accurate HVI/organization detection schemes with a user-friendly interface and processing to handle large data volumes, diverse persons of interest, and organization structures. The Phase II BIIT system will be tested and demonstrated on various HUMINT datasets with a user-adjustable testing environment, producing not only HVIs but also identification/tracking of diverse soft targets.