The National Academies Logo
Research Associateship Programs
Fellowships Office
Policy and Global Affairs

Participating Agencies - NAWCWD

  Sign InPrintable View

Opportunity at Naval Air Warfare Center Weapons Division (NAWCWD)

Learning Dynamic Representations of Large Data Sets for Pattern of Life Analysis


Naval Air Warfare Center Weapons Division Research Dept., Physics Division

RO# Location
34.01.02.B8010 China Lake, CA 93555


Name E-mail Phone
Flenner, Arjuna 760.939.2359


Advancements in computer memory, imaging sensors, and unmanned vehicles has introduced a rapid increase in information collected and stored by the Navy. This data supports decision-making and operations across diverse mission areas that are physically large, exhibit dynamics that conatin many objects, events, and activities, but currently many data analysis tasks are completed manually. Thus, much of the possible actionable information content of the collection is unexploited. Over the last ten years, there have been many advancements in applied probability and statistics that organize static data collections such as still images and large collections of text documents. This project will develop algorithms that can learn and segment the repetitive structure of dynamic patterns, which involves characterizing the routine behaviors of objects of interest, along natural time scales to organize dynamic activities and events into patterns of life.

Organizing large data collections into patterns of life is a challenging problem. The large quantity of data necessitates fast and efficient algorithms, but the data variability demands careful modeling. Furthermore, patterns of life are not completely determined by dynamical tracks and events, but social and communication networks are important in predicting patterns of life. For this reason, methods that can combine track information, image information, and interaction information are particularly appropriate. This project will build upon representation learning algorithms by including dynamical and graphical constraints into the algorithm. The Associate will develop statistical or variational machine learning algorithms with dynamical and or graphical contraints. They will need a strong statistical modeling or optimization background and MATLAB or C++ programming is required.


Machine learning; Image processing; Signal processing; Compressed sensing; Sparse representations; Dictionary learning; Applied mathematics; Pattern recognition; Computer vision;


Citizenship:  Open to U.S. citizens
Level:  Open to Postdoctoral applicants
Copyright © 2014. National Academy of Sciences. All rights reserved. 500 Fifth St. N.W., Washington, D.C. 20001.
Terms of Use and Privacy Statement.