ORACLE - Opportunistic Real-time Analysis, Content Labeling and Exploitation of Multimodal Sensory Data Navy SBIR FY2014.2 Sol No.: Navy SBIR FY2014.2 Topic No.: N142-122 Topic Title: ORACLE - Opportunistic Real-time Analysis, Content Labeling and Exploitation of Multimodal Sensory Data Proposal No.: N142-122-0506 Firm: DECISIVE ANALYTICS Corporation 1400 Crystal Drive Suite 1400 Arlington, Virginia 22202 Contact: Asaad Hakeem Phone: (703) 682-0748 Web Site: http://www.dac.us Abstract: Military and open-source data contain a wealth of information about entities of interest to US and coalition forces. This information is in the form of unstructured text, audio, imagery, video, SIGINT, and structured data. Intelligence analysts have access to these distributed, disparate data sources but are severely limited in exploiting the information. The success of military and security operations often depend on the level of situational awareness. A primary intelligence task is the development of patterns of life for the target of interest to determine target intent based on observation. To meet this challenge, DECISIVE ANALYTICS Corporation (DAC) and the Stevens Institute of Technology proposes a system for Opportunistic Real-time Analysis, Content Labeling and Exploitation (ORACLE) from multi-modal sensory data. The ORACLE system will enable real-time threat analysis and situational awareness from multi-modal data sources by automatically identifying the targets of interest, providing intent prediction based on patterns of life, and enabling the analyst to annotate targets of interest at both the (video) frame level and topic (event) level for future search. ORACLE will provide a distributed operational environment, where analysts from different departments, each with their own databases and perspectives, can analyze and collaborate on a common platform. Benefits: The goal of this effort is to develop the ORACLE system for DCGS-N ACAT IAM that will provide a distributed collaborative environment for automated analysis of unstructured text, audio, imagery, video, SIGINT, MASINT, GEOINT and structured data. ORACLE will enable real-time threat analysis by automatically identifying the targets of interest, providing intent prediction based on patterns of life, and enabling the analyst to annotate targets of interest, as well as conducting federated search. ORACLE also allows analysts from different departments, each with their own databases and perspectives, to analyze the data and automated analysis products, as well as collaborate on a common platform. The ORACLE system will provide the following benefits to the end-user: 1) A Common Representation for Multi-Modal Data Fusion and Target Identification; 2) Conduct Federated Search on the Aligned Distributed Knowledge Bases in a Cloud-based Collaborative Environment; 3) Apply Storytelling Algorithm to Represent Patterns of Life for Target Intent Prediction; and 4) Enhanced Predictive Analysis Algorithm that Models Target Behavior to Detect Anomalies. The Phase I effort will focus on a proof-of-concept system displaying these end-to-end capabilities on real-world data. Phase II & III will harden and operationalize the ORACLE system capabilities to provide an off-the-shelf distributed collaborative environment for automated analysis from multimodal data for automatically identifying the targets of interest and providing intent prediction based on patterns of life; that is secure, accurate and readily deployable at DCGS-N ACAT IAM s, as well as other military, government and commercial installations. Return