Training a perceptron by a bit sequence: Storage capacity

M. Schröder; W. Kinzel; I. Kanter

doi:10.1088/0305-4470/29/24/020

Training a perceptron by a bit sequence: Storage capacity

M. Schröder, W. Kinzel, I. Kanter

Department of Physics

University of Würzburg

Research output: Contribution to journal › Article › peer-review

5 Scopus citations

Abstract

A perceptron is trained by a random bit sequence. In comparison with the corresponding classification problem, the storage capacity decreases to a_c = 1.70 ± 0.02 due to correlations between input and output bits. The numerical results are supported by a signal-to-noise analysis of Hebbian weights.

Original language	English
Pages (from-to)	7965-7972
Number of pages	8
Journal	Journal of Physics A: Mathematical and General
Volume	29
Issue number	24
DOIs	https://doi.org/10.1088/0305-4470/29/24/020
State	Published - 21 Dec 1996

Access to Document

10.1088/0305-4470/29/24/020

Cite this

@article{e9429563ef2b4f49aa2d7f1d920928e0,

title = "Training a perceptron by a bit sequence: Storage capacity",

abstract = "A perceptron is trained by a random bit sequence. In comparison with the corresponding classification problem, the storage capacity decreases to ac = 1.70 ± 0.02 due to correlations between input and output bits. The numerical results are supported by a signal-to-noise analysis of Hebbian weights.",

author = "M. Schr{\"o}der and W. Kinzel and I. Kanter",

year = "1996",

month = dec,

day = "21",

doi = "10.1088/0305-4470/29/24/020",

language = "אנגלית",

volume = "29",

pages = "7965--7972",

journal = "Journal of Physics A: Mathematical and General",

issn = "0305-4470",

publisher = "IOP Publishing Ltd.",

number = "24",

}

TY - JOUR

T1 - Training a perceptron by a bit sequence

T2 - Storage capacity

AU - Schröder, M.

AU - Kinzel, W.

AU - Kanter, I.

PY - 1996/12/21

Y1 - 1996/12/21

N2 - A perceptron is trained by a random bit sequence. In comparison with the corresponding classification problem, the storage capacity decreases to ac = 1.70 ± 0.02 due to correlations between input and output bits. The numerical results are supported by a signal-to-noise analysis of Hebbian weights.

AB - A perceptron is trained by a random bit sequence. In comparison with the corresponding classification problem, the storage capacity decreases to ac = 1.70 ± 0.02 due to correlations between input and output bits. The numerical results are supported by a signal-to-noise analysis of Hebbian weights.

UR - http://www.scopus.com/inward/record.url?scp=0030597486&partnerID=8YFLogxK

U2 - 10.1088/0305-4470/29/24/020

DO - 10.1088/0305-4470/29/24/020

M3 - ???researchoutput.researchoutputtypes.contributiontojournal.article???

AN - SCOPUS:0030597486

SN - 0305-4470

VL - 29

SP - 7965

EP - 7972

JO - Journal of Physics A: Mathematical and General

JF - Journal of Physics A: Mathematical and General

IS - 24

ER -

Training a perceptron by a bit sequence: Storage capacity

Abstract

Access to Document

Other files and links

Fingerprint

Cite this