Training a perceptron by a bit sequence: Storage capacity
Condensed Matter
2009-10-28 v1
Abstract
A perceptron is trained by a random bit sequence. In comparison to the corresponding classification problem, the storage capacity decreases to alpha_c=1.70\pm 0.02 due to correlations between input and output bits. The numerical results are supported by a signal to noise analysis of Hebbian weights.
Cite
@article{arxiv.cond-mat/9607040,
title = {Training a perceptron by a bit sequence: Storage capacity},
author = {M. Schroeder and W. Kinzel and I. Kanter},
journal= {arXiv preprint arXiv:cond-mat/9607040},
year = {2009}
}
Comments
LaTeX, 13 pages incl. 4 figures and 1 table