Artificial metaplasticity and the challenge to train ANNS with reduced pattern availability

Authors:
Diego Andina
Affiliations:
Group for Automation in Signal and Communications, Technical University of Madrid, Spain
Venue:
CIMMACS'07 Proceedings of the 6th WSEAS international conference on Computational intelligence, man-machine systems and cybernetics
Year:
2007

Citing 3
Cited 0

Computational Intelligence: for Engineering and Manufacturing

Computational Intelligence: for Engineering and Manufacturing
Advances in Neyman-Pearson Neural Detectors Design

IWANN '03 Proceedings of the 7th International Work-Conference on Artificial and Natural Neural Networks: Part II: Artificial Neural Nets Problem Solving Methods
Quick simulation: a review of importance sampling techniques in communications systems

IEEE Journal on Selected Areas in Communications

Quantified Score

Hi-index	0.00

Visualization

Abstract

Artificial implementation of Biological Metaplasticity property promise to improve Artificial Neural Networks (ANN) design. This upgrade of existing models claims a much more efficient information extraction from the patterns available to train the ANN. The hypothesis has been tested as an application example in the Multilayer Perceptron (MLP) case, probably the most widely ANN applied through the ANN history. The results show a much more efficient training that is of crucial relevance when few training patterns are the only information font for the ANN design.