Open Source Information Tactical Exploitation (ONSITE)

The discovery of textual information on digital media - emails, documents, messages, and the like - during tactical operations can provide evidence to the warfighter about possible future enemy operations, movements, or attacks. Unfortunately, the sheer volume of such information that may be discovered during tactical operations precludes extensive, on-site computational analysis. High throughput, but efficient, methods must be employed to analyze large volumes of text to reveal any usable information, which can then be stored and later analyzed using desktop computer workstations. To address this significant problem, I.D.E.A.L. Technology Corporation and Knexient, Inc. propose STREAM, a fast text analysis module designed to provide additional natural language processing capabilities for the System for TRIaging Key Evidence (STRIKE) platform. Specifically, STREAM finds named entities and a set of binary relations between entities in free text, which are extracted and made available in the STRIKE application interface to help warfighters quickly find mission-critical information in text sources discovered during exploitation operations. Additionally, STREAM will use the Hierarchical Document Classification system (HDC) from Knexient's Text Processing Engine (TPE) to initially classify text according to topic and primary event (the main event expressed in the discourse.