NEURAL NETWORK DISTRIBUTIONAL INITIAL CONDITION ROBUSTNESS IN POWER SYSTEMS
Smith, Philip B.
Gannon, Anthony J.
MetadataShow full item record
How can we measure and classify neural network robustness across differently distributed data to avoid misuse of machine learning tools? This thesis adopts several metrics to measure the initial condition robustness of feedforward neural networks, allowing the creators of such networks to measure and refine their robustness and performance. This could allow highly robust neural networks to be used reliably on untrained data distributions and prevent the use of less robust networks as a black box in a poor environment. We test this measurement of robustness on a series of differently sized neural networks trained to detect and classify microgrid power system faults, giving examples of both robust and nonrobust networks, along with suggestions on how to maximize robustness. The analysis reveals that collecting data from segments along trajectories enhances the robustness of neural networks. In such data sets, the distribution of data points is dominated by the dynamics of the system, not the initial state distribution.
RightsThis publication is a work of the U.S. Government as defined in Title 17, United States Code, Section 101. Copyright protection is not available for this work in the United States.
Showing items related by title, author, creator and subject.
Kozak, Elana P. (Monterey, CA; Naval Postgraduate School, 2022-06);The applications of machine learning are broad and useful for a variety of industry and military objectives, but are the current methods robust? Robustness requires more than accuracy in ideal conditions; it means the ...
Brunger, Clifford A. (Monterey, California. Naval Postgraduate School, 1994-03);Aircraft design and control techniques rely on the proper modeling of the aircraft's equations of motion. Many of the variables used in these equations are aerodynamic coefficients which are ...
Alderson, D.; Roughan, M.; Shalunov, S.; Tanaka, R.; Willinger, W.; Doyle, J.; Li, L.; Low, S. (2005);The search for unifying properties of complex networks is popular, challenging, and important. For modeling approaches that focus on robustness and fragility as unifying concepts, the Internet is an especially attractive ...