Research Summaries

Back Deep Learning for Detecting Anomalous Activity

Fiscal Year 2019
Division Graduate School of Engineering & Applied Science
Department Mechanical & Aerospace Engineering
Investigator(s) Ross, Isaac M.
Karpenko, Mark
Sponsor Department of Defense Space (DoD)
Summary We propose to design a parallelizable deep learning algorithm for detecting anomalous activity that meets or exceeds human levels of accuracy. The technical objective will be met by a new machine learning approach that separates the optimization process from the backpropagation step in deep neural nets (DNNs). In sharp contrast to the current state of practice that involves sweeping back the gradients to determine the differential change in the loss function, our proposed algorithm minimizes a function that is related to the loss function in machine learning.
Keywords Deep Learning deep neural networks supervised learning algorithms
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