An Improved Natural Gradient ICA Algorithm Based on Lie Group Invariance

This paper proposes an improved natural gradient ICA (independent component analysis) algorithm based on the Riemannian structure of parameter space. The new algorithm introduces a scaling factor which makes the absolute value of the determinant of parameter matrix equal to one in the whole learning...

Full description

Saved in:
Bibliographic Details
Main Authors: Xianhua Zeng, Siwei Luo
Format: Conference Proceeding
Language:English
Subjects:
Acoustic signal processing
Analytical models
Array signal processing
Image restoration
Independent component analysis
Information technology
Neural networks
Probability density function
Signal processing algorithms
Signal restoration
Online Access:Request full text
Tags: Add Tag
No Tags, Be the first to tag this record!
cited_by
cites
container_end_page
container_issue
container_start_page
container_title
container_volume 3
creator Xianhua Zeng
Siwei Luo
description This paper proposes an improved natural gradient ICA (independent component analysis) algorithm based on the Riemannian structure of parameter space. The new algorithm introduces a scaling factor which makes the absolute value of the determinant of parameter matrix equal to one in the whole learning process. Therefore, the training process is faster and stable by restricting the drastic change of parameter matrix. In addition, the general criterion function is simplified by the scaling factor. In simulation experiment, we compare three algorithms including ordinary gradient ICA algorithm, natural gradient ICA algorithm and improved natural gradient ICA algorithm. Comparing the new improved natural gradient ICA with the natural gradient ICA, the mean relative error of restored signals is decreased 38.7%. The results show that the latter is better in restored-signal precision and convergence speed
doi_str_mv 10.1109/ICOSP.2006.345781
format conference_proceeding
fullrecord <record><control><sourceid>ieee_CHZPO</sourceid><recordid>TN_cdi_ieee_primary_4129222</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><ieee_id>4129222</ieee_id><sourcerecordid>4129222</sourcerecordid><originalsourceid>FETCH-LOGICAL-i175t-4a9fc3b0433b70bf5630a29d8073d1b95106ec81b19e27a92454279014cbd563</originalsourceid><addsrcrecordid>eNo1j1FLwzAUhSM6cM79APElf6D13ps0aR5ncVthOMG9j7RNNbK1I-0G_nsDKhw4HPg4nMPYA0KKCOapLLbvbykBqFTITOd4xeZG5xAljBY6v2Z3_0GJGzYlVDLJiHDCpqASQ9qQuWXzYfgCAIF5roSasvWi4-XxFPqLa_irHc_BHvgq2Ma7buRlseCLw0cf_Ph55M92iFDf8Y13kenPJ152Fxu87Wp3zyatPQxu_ucztlu-7Ip1stmuYs0m8aizMZHWtLWoQApRaajaTAmwZJr4RDRYmQxBuTrHCo0jbQ3JTMbpgLKumgjP2ONvrXfO7U_BH2343kskQ0TiB7VFTps</addsrcrecordid><sourcetype>Publisher</sourcetype><iscdi>true</iscdi><recordtype>conference_proceeding</recordtype></control><display><type>conference_proceeding</type><title>An Improved Natural Gradient ICA Algorithm Based on Lie Group Invariance</title><source>IEEE Xplore All Conference Series</source><creator>Xianhua Zeng ; Siwei Luo</creator><creatorcontrib>Xianhua Zeng ; Siwei Luo</creatorcontrib><description>This paper proposes an improved natural gradient ICA (independent component analysis) algorithm based on the Riemannian structure of parameter space. The new algorithm introduces a scaling factor which makes the absolute value of the determinant of parameter matrix equal to one in the whole learning process. Therefore, the training process is faster and stable by restricting the drastic change of parameter matrix. In addition, the general criterion function is simplified by the scaling factor. In simulation experiment, we compare three algorithms including ordinary gradient ICA algorithm, natural gradient ICA algorithm and improved natural gradient ICA algorithm. Comparing the new improved natural gradient ICA with the natural gradient ICA, the mean relative error of restored signals is decreased 38.7%. The results show that the latter is better in restored-signal precision and convergence speed</description><identifier>ISSN: 2164-5221</identifier><identifier>ISBN: 0780397363</identifier><identifier>ISBN: 9780780397361</identifier><identifier>EISBN: 9780780397378</identifier><identifier>EISBN: 0780397371</identifier><identifier>DOI: 10.1109/ICOSP.2006.345781</identifier><identifier>LCCN: 06-927929</identifier><language>eng</language><subject>Acoustic signal processing ; Analytical models ; Array signal processing ; Image restoration ; Independent component analysis ; Information technology ; Neural networks ; Probability density function ; Signal processing algorithms ; Signal restoration</subject><ispartof>2006 8th international Conference on Signal Processing, 2006, Vol.3</ispartof><woscitedreferencessubscribed>false</woscitedreferencessubscribed></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><linktohtml>$$Uhttps://ieeexplore.ieee.org/document/4129222$$EHTML$$P50$$Gieee$$H</linktohtml><link.rule.ids>309,310,780,784,789,790,2057,4049,4050,27924,54554,54919,54931</link.rule.ids><linktorsrc>$$Uhttps://ieeexplore.ieee.org/document/4129222$$EView_record_in_IEEE$$FView_record_in_$$GIEEE</linktorsrc></links><search><creatorcontrib>Xianhua Zeng</creatorcontrib><creatorcontrib>Siwei Luo</creatorcontrib><title>An Improved Natural Gradient ICA Algorithm Based on Lie Group Invariance</title><title>2006 8th international Conference on Signal Processing</title><addtitle>ICOSP</addtitle><description>This paper proposes an improved natural gradient ICA (independent component analysis) algorithm based on the Riemannian structure of parameter space. The new algorithm introduces a scaling factor which makes the absolute value of the determinant of parameter matrix equal to one in the whole learning process. Therefore, the training process is faster and stable by restricting the drastic change of parameter matrix. In addition, the general criterion function is simplified by the scaling factor. In simulation experiment, we compare three algorithms including ordinary gradient ICA algorithm, natural gradient ICA algorithm and improved natural gradient ICA algorithm. Comparing the new improved natural gradient ICA with the natural gradient ICA, the mean relative error of restored signals is decreased 38.7%. The results show that the latter is better in restored-signal precision and convergence speed</description><subject>Acoustic signal processing</subject><subject>Analytical models</subject><subject>Array signal processing</subject><subject>Image restoration</subject><subject>Independent component analysis</subject><subject>Information technology</subject><subject>Neural networks</subject><subject>Probability density function</subject><subject>Signal processing algorithms</subject><subject>Signal restoration</subject><issn>2164-5221</issn><isbn>0780397363</isbn><isbn>9780780397361</isbn><isbn>9780780397378</isbn><isbn>0780397371</isbn><fulltext>true</fulltext><rsrctype>conference_proceeding</rsrctype><creationdate>2006</creationdate><recordtype>conference_proceeding</recordtype><sourceid>6IE</sourceid><recordid>eNo1j1FLwzAUhSM6cM79APElf6D13ps0aR5ncVthOMG9j7RNNbK1I-0G_nsDKhw4HPg4nMPYA0KKCOapLLbvbykBqFTITOd4xeZG5xAljBY6v2Z3_0GJGzYlVDLJiHDCpqASQ9qQuWXzYfgCAIF5roSasvWi4-XxFPqLa_irHc_BHvgq2Ma7buRlseCLw0cf_Ph55M92iFDf8Y13kenPJ152Fxu87Wp3zyatPQxu_ucztlu-7Ip1stmuYs0m8aizMZHWtLWoQApRaajaTAmwZJr4RDRYmQxBuTrHCo0jbQ3JTMbpgLKumgjP2ONvrXfO7U_BH2343kskQ0TiB7VFTps</recordid><startdate>2006</startdate><enddate>2006</enddate><creator>Xianhua Zeng</creator><creator>Siwei Luo</creator><scope>6IE</scope><scope>6IL</scope><scope>CBEJK</scope><scope>RIE</scope><scope>RIL</scope></search><sort><creationdate>2006</creationdate><title>An Improved Natural Gradient ICA Algorithm Based on Lie Group Invariance</title><author>Xianhua Zeng ; Siwei Luo</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-i175t-4a9fc3b0433b70bf5630a29d8073d1b95106ec81b19e27a92454279014cbd563</frbrgroupid><rsrctype>conference_proceedings</rsrctype><prefilter>conference_proceedings</prefilter><language>eng</language><creationdate>2006</creationdate><topic>Acoustic signal processing</topic><topic>Analytical models</topic><topic>Array signal processing</topic><topic>Image restoration</topic><topic>Independent component analysis</topic><topic>Information technology</topic><topic>Neural networks</topic><topic>Probability density function</topic><topic>Signal processing algorithms</topic><topic>Signal restoration</topic><toplevel>online_resources</toplevel><creatorcontrib>Xianhua Zeng</creatorcontrib><creatorcontrib>Siwei Luo</creatorcontrib><collection>IEEE Electronic Library (IEL) Conference Proceedings</collection><collection>IEEE Proceedings Order Plan All Online (POP All Online) 1998-present by volume</collection><collection>IEEE Xplore All Conference Proceedings</collection><collection>IEEE Electronic Library (IEL)</collection><collection>IEEE Proceedings Order Plans (POP All) 1998-Present</collection></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext_linktorsrc</fulltext></delivery><addata><au>Xianhua Zeng</au><au>Siwei Luo</au><format>book</format><genre>proceeding</genre><ristype>CONF</ristype><atitle>An Improved Natural Gradient ICA Algorithm Based on Lie Group Invariance</atitle><btitle>2006 8th international Conference on Signal Processing</btitle><stitle>ICOSP</stitle><date>2006</date><risdate>2006</risdate><volume>3</volume><issn>2164-5221</issn><isbn>0780397363</isbn><isbn>9780780397361</isbn><eisbn>9780780397378</eisbn><eisbn>0780397371</eisbn><abstract>This paper proposes an improved natural gradient ICA (independent component analysis) algorithm based on the Riemannian structure of parameter space. The new algorithm introduces a scaling factor which makes the absolute value of the determinant of parameter matrix equal to one in the whole learning process. Therefore, the training process is faster and stable by restricting the drastic change of parameter matrix. In addition, the general criterion function is simplified by the scaling factor. In simulation experiment, we compare three algorithms including ordinary gradient ICA algorithm, natural gradient ICA algorithm and improved natural gradient ICA algorithm. Comparing the new improved natural gradient ICA with the natural gradient ICA, the mean relative error of restored signals is decreased 38.7%. The results show that the latter is better in restored-signal precision and convergence speed</abstract><doi>10.1109/ICOSP.2006.345781</doi></addata></record>
fulltext fulltext_linktorsrc
identifier ISSN: 2164-5221
ispartof 2006 8th international Conference on Signal Processing, 2006, Vol.3
issn 2164-5221
language eng
recordid cdi_ieee_primary_4129222
source IEEE Xplore All Conference Series
subjects Acoustic signal processing
Analytical models
Array signal processing
Image restoration
Independent component analysis
Information technology
Neural networks
Probability density function
Signal processing algorithms
Signal restoration
title An Improved Natural Gradient ICA Algorithm Based on Lie Group Invariance
url http://sfxeu10.hosted.exlibrisgroup.com/loughborough?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-01-08T15%3A11%3A53IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-ieee_CHZPO&rft_val_fmt=info:ofi/fmt:kev:mtx:book&rft.genre=proceeding&rft.atitle=An%20Improved%20Natural%20Gradient%20ICA%20Algorithm%20Based%20on%20Lie%20Group%20Invariance&rft.btitle=2006%208th%20international%20Conference%20on%20Signal%20Processing&rft.au=Xianhua%20Zeng&rft.date=2006&rft.volume=3&rft.issn=2164-5221&rft.isbn=0780397363&rft.isbn_list=9780780397361&rft_id=info:doi/10.1109/ICOSP.2006.345781&rft.eisbn=9780780397378&rft.eisbn_list=0780397371&rft_dat=%3Cieee_CHZPO%3E4129222%3C/ieee_CHZPO%3E%3Cgrp_id%3Ecdi_FETCH-LOGICAL-i175t-4a9fc3b0433b70bf5630a29d8073d1b95106ec81b19e27a92454279014cbd563%3C/grp_id%3E%3Coa%3E%3C/oa%3E%3Curl%3E%3C/url%3E&rft_id=info:oai/&rft_id=info:pmid/&rft_ieee_id=4129222&rfr_iscdi=true