Convergence of a hebbian-type learning algorithm

Qingfu Zhang; Yiu-Wing Leung

doi:10.1109/82.746680

Convergence of a hebbian-type learning algorithm

Qingfu Zhang^*, Yiu-Wing Leung

^*Corresponding author for this work

Department of Computer Science

Research output: Contribution to journal › Journal article › peer-review

7 Citations (Scopus)

22 Downloads (Pure)

Abstract

A Hebbian-type learning algorithm was proposed in [1] for extracting the minor components of the input signals. In this paper, we demonstrate that some solutions of the averaging differential equation of this algorithm can become unbounded in a finite time. We derive five sufficient conditions to ensure that the solutions of its averaging differential equation are bounded and can be extended to the time interval [0, ∞). Any one of these conditions can guarantee that this algorithm can be used to lind the minor components of the input signals.

Original language	English
Pages (from-to)	1599-1601
Number of pages	3
Journal	IEEE Transactions on Circuits and Systems II: Analog and Digital Signal Processing
Volume	45
Issue number	12
DOIs	https://doi.org/10.1109/82.746680
Publication status	Published - Dec 1998

Access to Document

10.1109/82.746680

Full TextAccepted author manuscript, 3.6 MB

Cite this

@article{0687bf61ef834a2da0e5d14e997131fd,

title = "Convergence of a hebbian-type learning algorithm",

abstract = "A Hebbian-type learning algorithm was proposed in [1] for extracting the minor components of the input signals. In this paper, we demonstrate that some solutions of the averaging differential equation of this algorithm can become unbounded in a finite time. We derive five sufficient conditions to ensure that the solutions of its averaging differential equation are bounded and can be extended to the time interval [0, ∞). Any one of these conditions can guarantee that this algorithm can be used to lind the minor components of the input signals.",

author = "Qingfu Zhang and Yiu-Wing Leung",

year = "1998",

month = dec,

doi = "10.1109/82.746680",

language = "English",

volume = "45",

pages = "1599--1601",

journal = "IEEE Transactions on Circuits and Systems II: Analog and Digital Signal Processing",

issn = "1057-7130",

publisher = "IEEE",

number = "12",

}

TY - JOUR

T1 - Convergence of a hebbian-type learning algorithm

AU - Zhang, Qingfu

AU - Leung, Yiu-Wing

PY - 1998/12

Y1 - 1998/12

N2 - A Hebbian-type learning algorithm was proposed in [1] for extracting the minor components of the input signals. In this paper, we demonstrate that some solutions of the averaging differential equation of this algorithm can become unbounded in a finite time. We derive five sufficient conditions to ensure that the solutions of its averaging differential equation are bounded and can be extended to the time interval [0, ∞). Any one of these conditions can guarantee that this algorithm can be used to lind the minor components of the input signals.

AB - A Hebbian-type learning algorithm was proposed in [1] for extracting the minor components of the input signals. In this paper, we demonstrate that some solutions of the averaging differential equation of this algorithm can become unbounded in a finite time. We derive five sufficient conditions to ensure that the solutions of its averaging differential equation are bounded and can be extended to the time interval [0, ∞). Any one of these conditions can guarantee that this algorithm can be used to lind the minor components of the input signals.

UR - http://www.scopus.com/inward/record.url?scp=0032280813&partnerID=8YFLogxK

U2 - 10.1109/82.746680

DO - 10.1109/82.746680

M3 - Journal article

AN - SCOPUS:0032280813

SN - 1057-7130

VL - 45

SP - 1599

EP - 1601

JO - IEEE Transactions on Circuits and Systems II: Analog and Digital Signal Processing

JF - IEEE Transactions on Circuits and Systems II: Analog and Digital Signal Processing

IS - 12

ER -

Convergence of a hebbian-type learning algorithm

Abstract

Access to Document

Other files and links

Fingerprint

Cite this