On gradient adaptation with unit-norm constraints

Scott C. Douglas; Shun Ichi Amari; S. Y. Kung

doi:10.1109/78.845952

On gradient adaptation with unit-norm constraints

Scott C. Douglas, Shun Ichi Amari, S. Y. Kung

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

49 Scopus citations

Abstract

In this correspondence, we describe gradient-based adaptive algorithms within parameter spaces that are specified by ∥w∥ = 1, where ∥·∥ is any vector norm. We provide several algorithm forms and relate them to true gradient procedures via their geometric structures. We also give algorithms that mitigate an inherent numerical instability for L₂-norm-constrained optimization tasks. Simulations showing the performance of the techniques for independent component analysis are provided.

Original language	English (US)
Pages (from-to)	1843-1847
Number of pages	5
Journal	IEEE Transactions on Signal Processing
Volume	48
Issue number	6
DOIs	https://doi.org/10.1109/78.845952
State	Published - 2000
Externally published	Yes

All Science Journal Classification (ASJC) codes

Signal Processing
Electrical and Electronic Engineering

Access to Document

10.1109/78.845952

Cite this

@article{8e4f4bce7e374a57a7a49ca610eebd40,

title = "On gradient adaptation with unit-norm constraints",

abstract = "In this correspondence, we describe gradient-based adaptive algorithms within parameter spaces that are specified by ∥w∥ = 1, where ∥·∥ is any vector norm. We provide several algorithm forms and relate them to true gradient procedures via their geometric structures. We also give algorithms that mitigate an inherent numerical instability for L2-norm-constrained optimization tasks. Simulations showing the performance of the techniques for independent component analysis are provided.",

author = "Douglas, {Scott C.} and Amari, {Shun Ichi} and Kung, {S. Y.}",

year = "2000",

doi = "10.1109/78.845952",

language = "English (US)",

volume = "48",

pages = "1843--1847",

journal = "IRE Transactions on Audio",

issn = "1053-587X",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

number = "6",

}

TY - JOUR

T1 - On gradient adaptation with unit-norm constraints

AU - Douglas, Scott C.

AU - Amari, Shun Ichi

AU - Kung, S. Y.

PY - 2000

Y1 - 2000

N2 - In this correspondence, we describe gradient-based adaptive algorithms within parameter spaces that are specified by ∥w∥ = 1, where ∥·∥ is any vector norm. We provide several algorithm forms and relate them to true gradient procedures via their geometric structures. We also give algorithms that mitigate an inherent numerical instability for L2-norm-constrained optimization tasks. Simulations showing the performance of the techniques for independent component analysis are provided.

AB - In this correspondence, we describe gradient-based adaptive algorithms within parameter spaces that are specified by ∥w∥ = 1, where ∥·∥ is any vector norm. We provide several algorithm forms and relate them to true gradient procedures via their geometric structures. We also give algorithms that mitigate an inherent numerical instability for L2-norm-constrained optimization tasks. Simulations showing the performance of the techniques for independent component analysis are provided.

UR - http://www.scopus.com/inward/record.url?scp=0033738185&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0033738185&partnerID=8YFLogxK

U2 - 10.1109/78.845952

DO - 10.1109/78.845952

M3 - Article

AN - SCOPUS:0033738185

SN - 1053-587X

VL - 48

SP - 1843

EP - 1847

JO - IRE Transactions on Audio

JF - IRE Transactions on Audio

IS - 6

ER -

On gradient adaptation with unit-norm constraints

Abstract

All Science Journal Classification (ASJC) codes

Access to Document

Other files and links

Fingerprint

Cite this