Raytheon to develop tools to understand natural language

CAMBRIDGE, Mass., 5 Dec. 2012. Raytheon BBN Technologies has been awarded $16 million under the Deep Exploration and Filter of Text (DEFT) program, sponsored by the Defense Advanced Research Projects Agency (DARPA) and the U.S. Air Force Research Laboratory. BBN Technologies is a wholly owned subsidiary of Raytheon Co. (NYSE: RTN).

The DEFT program aims to enhance currently available analytical tools by developing a deep natural language exploration and filtering capability to find operationally significant information and deliver relevant alerts to users.

Raytheon BBN scientists will tackle two significant goals of the DEFT program: to develop new techniques to infer meaning, discover relationships, and detect anomalies in text; and to research and develop tools to integrate new, individual facts into existing large information stores. The understanding enabled by these capabilities will allow analysts and other users to exploit not just the surface information in massive data archives, but also to discover the interconnections between disparate data elements and the tactically and strategically significant nuances.

Government analysts collect and process massive amounts of data from a variety of sources to create and assess plans and to execute missions. Advances in speech and language technologies have provided them with robust tools for their tasks, but discovering information that is implied but not explicitly stated remains a challenge for automated tools.

The DEFT program aims to strip away ambiguity and also to make it easier to integrate new facts into existing knowledge bases in a meaningful way.