Adaptive Toolkit for the Discovery of Threats (ATDT)

We propose to design, develop, and deploy an adaptive toolkit for automated threat discovery. The toolkit will provide innovative mechanisms to rapidly discover emerging threat situations. The proposed approach integrates intelligent knowledge discoverytechniques with knowledge-based methods to facilitate robust and cost-effective threat detection. The main product of this project will be the Adaptive Toolkit for the Discovery of Threats (ATDT), an agent-based decision support system that facilitatesthe automated generation of information from disparate and distributed data to support early detection of emerging threats. The Phase I effort will (i) establish ATDT requirements, (ii) design the ATDT architecture, and (iii) demonstrate prototype ATDTsoftware. The Phase II project will harden the software and demonstrate its benefits on focused military and commercial applications. Key innovations include (i) hybrid approach that combines knowledge discovery methods with classical rule basedtechniques, (ii) extendible and reconfigurable agent-based architecture, and (iii) intelligent learning mechanisms that adapt to changing data patterns. Key benefits include (a) ability to rapidly and accurately predict emerging threats, (b) significantreductions in the time and effort for converting distributed data to actionable decision information, and (c) robust threat prediction in dynamic and uncertain environments. The principal benefit is the capability to rapidly, reliably, and accurately -detect and predict emerging threats. Applications include terrorist threat prediction, emergency response management, target/decoy recognition and discrimination, industrial espionage detection, financial fraud detection, computer network intrusiondetection, and business intelligence.