Big Data and Deep Learning for Defense Acquisition Visibility Environment (DAVE)—Developing NPS Student Thesis Research

Abstract

The U.S. Department of Defense (DoD) acquisition process is extremely complex. There are three key processes that must work in concert to deliver capabilities: determining warfighters’ requirements and needs, planning the DoD budget, and procuring final products. Each process produces large amounts of information (big data). There is a critical need for automation, validation, and discovery to help acquisition professionals, decision-makers, and researchers understand the important content within large data sets and optimize DoD resources. Lexical link analysis (LLA) and collaborative learning agents (CLAs) have been applied to reveal and depict—to decision-makers—the correlations, associations, and program gaps across acquisition programs examined over many years. This enables strategic understanding of data gaps and potential trends, and it can inform managers which areas might be exposed to higher program risk and how resource and big data management might affect the desired return on investment (ROI) among projects. In last year’s research, a Naval Postgraduate School (NPS) thesis started using LLA and data from the Defense Acquisition Visibility Environment (DAVE). The goal of the thesis was to discover the correlation of the vendors’ capabilities and the requirements of a logistics application. LLA also used visualization capabilities, which was planned to be used in the student thesis. In the same time, a conference paper about discovering high-value information using LLA and the associated new visualizations was published. There is an interesting connection between the LLA and CLA computing theory and quantum game theory. LLA/CLA/SSA was introduced in the context of quantum game and quantum intelligence, which is an interesting connection that can help systems of systems, such as DoD acquisition systems, reach stable states of Nash equilibria and at the same time be Pareto optimal. This theory is capable of making competitive systems cooperate, such as DoD acquisition systems. For example, the theory can be applied to the current acquisition research to select systems of systems by balancing the authoritative attributes (i.e., the system attributes that help to reach Nash equilibrium) and expertise attributes (i.e., the system attributes that help to reach Pareto optimality).