ONSITE: Open Source Information Tactical Exploitation

Modus Operandi proposes the development of ONSITE, a set of scalable and accurate algorithms for extracting semantic relations from open source information. This algorithm and process supports warfighter tactical operations by providing near real-time situational awareness. Warfighters performing missions involving the hostage rescue, insurgent extraction and targeting may only have minutes to assess a tactical situation and then to act. Today, a large amount of intelligence is unavailable because it resides in the form of natural language texts (news stories, blogs, press releases and other unstructured data). Current approaches, coupled with appropriate extraction algorithms and applied to textual data, can greatly increase the warfighter’s understanding of his operating environment if those approaches can be made faster and more accurate. We propose to adapt state of the art algorithms in rote extraction to work within a robust, distributed computing framework. We also plan to extend a pattern discovery methodology to identify more specific relation contexts. These innovative capabilities will significantly improve the warfighter’s access to timely and accurate information necessary for tactical operations.