TY - GEN
T1 - Diverse evolutionary neural networks based on information theory
AU - Kim, Kyung Joong
AU - Cho, Sung Bae
PY - 2008
Y1 - 2008
N2 - There is no consensus on measuring distances between two different neural network architectures. Two folds of methods are used for that purpose: Structural and behavioral distance measures. In this paper, we focus on the later one that compares differences based on output responses given the same input. Usually neural network output can be interpreted as a probabilistic function given the input signals if it is normalized to 1. Information theoretic distance measures are widely used to measure distances between two probabilistic distributions. In the framework of evolving diverse neural networks, we adopted information-theoretic distance measures to improve its performance. Experimental results on UCI benchmark dataset show the promising possibility of the approach.
AB - There is no consensus on measuring distances between two different neural network architectures. Two folds of methods are used for that purpose: Structural and behavioral distance measures. In this paper, we focus on the later one that compares differences based on output responses given the same input. Usually neural network output can be interpreted as a probabilistic function given the input signals if it is normalized to 1. Information theoretic distance measures are widely used to measure distances between two probabilistic distributions. In the framework of evolving diverse neural networks, we adopted information-theoretic distance measures to improve its performance. Experimental results on UCI benchmark dataset show the promising possibility of the approach.
UR - http://www.scopus.com/inward/record.url?scp=54049090462&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=54049090462&partnerID=8YFLogxK
U2 - 10.1007/978-3-540-69162-4_105
DO - 10.1007/978-3-540-69162-4_105
M3 - Conference contribution
AN - SCOPUS:54049090462
SN - 3540691596
SN - 9783540691594
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 1007
EP - 1016
BT - Neural Information Processing - 14th International Conference, ICONIP 2007, Revised Selected Papers
T2 - 14th International Conference on Neural Information Processing, ICONIP 2007
Y2 - 13 November 2007 through 16 November 2007
ER -