Deep Evolution of Feature Representations for Handwritten Digit Recognition

Created by W.Langdon from gp-bibliography.bib Revision:1.4504

  author =       "Alexandros Agapitos and Michael O'Neill and 
                 Miguel Nicolau and David Fagan and Ahmed Kattan and 
                 Kathleen Curran",
  title =        "Deep Evolution of Feature Representations for
                 Handwritten Digit Recognition",
  booktitle =    "Proceedings of 2015 IEEE Congress on Evolutionary
                 Computation (CEC 2015)",
  editor =       "Yadahiko Murata",
  pages =        "2452--2459",
  year =         "2015",
  address =      "Sendai, Japan",
  publisher =    "IEEE Press",
  keywords =     "genetic algorithms, genetic programming",
  DOI =          "doi:10.1109/CEC.2015.7257189",
  abstract =     "A training protocol for learning deep neural networks,
                 called greedy layer-wise training, is applied to the
                 evolution of a hierarchical, feed-forward Genetic
                 Programming based system for feature construction and
                 object recognition. Results on a popular handwritten
                 digit recognition benchmark clearly demonstrate that
                 two layers of feature transformations improves
                 generalisation compared to a single layer. In addition,
                 we show that the proposed system outperforms several
                 standard Genetic Programming systems, which are based
                 on hand-designed features, and use different program
                 representations and fitness functions.",
  notes =        "CEC2015",

Genetic Programming entries for Alexandros Agapitos Michael O'Neill Miguel Nicolau David Fagan Ahmed Kattan Kathleen Curran