Dr. Bruce A. Cox, Ph.D.
Research Interests
My basic research interests involve the development, and foundational understanding of Neural Networks. This research is multipronged, examining both the characterization of Convolutional Neural Network loss response surfaces, as well as exploring how data augmentation techniques, such asGenerative Adversarial Networks,may be used to increase the diversity of training datasets and improve the generalizability of resulting models.Additionally, I perform basic research on very large and/or harddeterministic optimizationproblems. In particular, the design, development, and testing of new decompositiontechniques, and efficient heuristics with the aim of to expanding the scope and efficiency of deterministic optimization methods.
My applied research interests revolve around problems related to defense, to include computer vision, neural networks, and optimal network design (e.g.,supply chain networks, repair networks, etc.). I utilize numerous tools and techniques as informed by the underlying systems data and problem requirements, ranging from deterministic optimization, decision analysis, simulation, statistical stability and hypothesis testing, forecasting, and machine learning techniques, as appropriate.