Learning multilayer perceptrons efficiently
Disordered Systems and Neural Networks
2007-05-23 v1
Abstract
A learning algorithm for multilayer perceptrons is presented which is based on finding the principal components of a correlation matrix computed from the example inputs and their target outputs. For large networks our procedure needs far fewer examples to achieve good generalization than traditional on-line algorithms.
Keywords
Cite
@article{arxiv.cond-mat/0101132,
title = {Learning multilayer perceptrons efficiently},
author = {C. Bunzmann and M. Biehl and R. Urbanczik},
journal= {arXiv preprint arXiv:cond-mat/0101132},
year = {2007}
}
Comments
5 pages, 3 figures, to appear: Phys. Rev. Letts