Institutions | About Us | Help | Gaeilge
rian logo


Mark
Go Back
Deep Evolution of Feature Representations for Handwritten Digit Recognition
Agapitos, Alexandros; O'Neill, Michael; Nicolau, Miguel; Fagan, David; Kattan, Ahmed; Curran, Kathleen M.
2015 IEEE Congress on Evolutionary Computation (CEC), Sendai, Japan, 25-28 May 2015 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.
Keyword(s): Object recognition; Feature extraction; Genetic algorithms; Handwritten character recognition
Publication Date:
2017
Type: Other
Peer-Reviewed: Unknown
Language(s): English
Institution: University College Dublin
Publisher(s): IEEE
First Indexed: 2017-01-17 05:17:59 Last Updated: 2018-10-11 15:04:45