Situational Awareness as a Man-Machine Map Reduce Job

This effort is focused on the development of a system called Situational Awareness via Mixed-initiative Universal Recognition, Analysis, and Inference (SAMURAI). SAMURAI will provide a single cloud-enabled end-to-end workflow covering the full range of data analysis from data ingest to situational assessment and decision support. As part of this workflow SAMURAI will provide the ability for automated processing to provide rapid extraction of knowledge while allowing manual corrections and additions at multiple levels. These manual changes will then be utilized to improve the underlying statistical models utilized within the automated processes. The SAMURAI system will also provide a behavior/intent prediction capability that has been demonstrated to operate on millions of entities and billions of relationships. Moreover, the prediction algorithms will have the ability to be updated according to analyst feedback and use analyst-provided information not present in the raw data. The overall efficiency and accuracy of a large scale, mixed initiative analytical system will be heavily affected by when and how users provide feedback to the algorithms, the system incorporates this feedback, and users interact with each other. SAMURAI identifies when such interactions have the greatest benefit, and advises the user accordingly.