Summaries - Office of Research & Innovation
Back Efficient Intelligence Processing from Large Networks
|Division||Research & Sponsored Programs|
|Department||Naval Research Program|
|Investigator(s)||Dimitrov, Nedialko B.|
|Sponsor||NPS Naval Research Program (Navy)|
The advent of sensing technologies -- from electro-optical devices to cyber interceptors -- has resulted in a plethora of sensors that collect and transmit an unprecedented glut of data. These data need to be processed and analyzed in a timely manner in order to produce useful information for operations. In a nutshell, and in view of the aforementioned challenge, we propose to answer the following question: Given a pile of intercepted communication (e.g. phone calls, emails, etc.), each between two (identified, partially identified, or unidentified) persons, and a limited amount of time, how should the intelligence processor screen the communications so that the maximum number of relevant items are passed onto analysts? We will develop a model and algorithm for optimally selecting intelligence items intercepted from a communication network, for screening in the presence of time and resource constraints. The intent is to create a method to concentrate an intelligence agency's focus in the face of a glut of information.
This research builds on an initial model developed in an NPS Master's thesis by Nevo (Nevo 2011) and incorporates advanced methods and techniques from graphical models (Koller & Friedman 2009) and Bayesian inference (Dagum & Luby 1997). The research focus will be to make that model more relevant, by incorporating new aspects of real-world intelligence collection that are not yet accurately captured in the model. In addition, a secondary goal will be to develop software to test the intelligence collection methodology on simulated intelligence queries. This research will aid the intelligence community in several ways: 1) it will create mathematical models for the deployment of limited resources to gather the most relevant information from a glut of collected data 2) it will test the mathematical models, to ensure they are relevant to intelligence analysts and produce quantifiably better results in intelligence collection.
|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|