This thesis presents an extensive and thorough computational comparison featuring deep neural networks and kernel learning machines and successfully establishes that on both real-life datasets and artificially simulated ones, kernel learning machines tend to be just as good as deep neural networks and quite often far better predictively. It turns out from the findings of this thesis that while deep neural networks might have worked well on tasks for which millions of observations are available, kernel learning machines just happen to be predictively better on the wide variety of tasks with the kind of sample size that one should realistically expect to have in practice.
Library of Congress Subject Headings
Support vector machines; Neural networks (Computer science); Machine learning; Gaussian processes
Applied Statistics (MS)
Department, Program, or Center
School of Mathematical Sciences (COS)
Pei, Eddie, "On Some Similarities and Differences between Deep Neural Networks and Kernel Learning Machines" (2021). Thesis. Rochester Institute of Technology. Accessed from
RIT – Main Campus