Fiscal Year	2011
Division	Research & Sponsored Programs
Department	Cebrowski Institute
Investigator(s)	Martell, Craig H.
Sponsor	National Reconnaissance Office (DoD)
Summary	State-of-the-art techniques in topic detection and authorship detection are becoming powerful enough to help with problems of interest to the NRO/IAO. In particular, the natural-language processing community has made large strides in being able to reliably detect the topic and author of Twitter messages, SMS, chat posts, and blogs. We are also able to correlate the content of these data with other openly available signals, such as YouTube videos. Despite these recent advances, the state of the art still requires large-scale computational infrastructure. We propose to build topic and authorship detection software that can be run solely on a constrained device of interest, for example, a smart phone. Further, we propose to study methods for on-demand transfer, balance, and coordination of the computational analysis as mandated by current device conditions (e.g. an overly taxed cellular phone).
Keywords	Data Exfiltration Data Tagging Insider Threat
Publications	Publications, theses (not shown) and data repositories will be added to the portal record when information is available in FAIRS and brought back to the portal
Data	Publications, theses (not shown) and data repositories will be added to the portal record when information is available in FAIRS and brought back to the portal