The Problem of Overlearning in High-Order ICA Approaches: Analysis and Solutions
Identifieur interne : 000806 ( Main/Corpus ); précédent : 000805; suivant : 000807The Problem of Overlearning in High-Order ICA Approaches: Analysis and Solutions
Auteurs : Jaakko S Rel ; Ricardo VigárioSource :
- Lecture Notes in Computer Science [ 0302-9743 ] ; 2001.
Abstract
Abstract: We consider a type of overlearning typical of independent component analysis algorithms. These can be seen to minimize the mutual information between source estimates. The overlearning causes spikelike signals if there are too few samples or there is a considerable amount of noise present. It is argued that if the data has flicker noise the problem is more severe and is better characterized by bumps instead of spikes. The problem is demonstrated using recorded magnetoencephalographic signals. Several methods are suggested that attempt to solve the overlearning problem or, at least, diminish reduce its effects.
Url:
DOI: 10.1007/3-540-45723-2_99
Links to Exploration step
ISTEX:EED15B332943C34B00EA62FAC243875440F299CBLe document en format XML
<record><TEI wicri:istexFullTextTei="biblStruct"><teiHeader><fileDesc><titleStmt><title xml:lang="en">The Problem of Overlearning in High-Order ICA Approaches: Analysis and Solutions</title><author><name sortKey="S Rel, Jaakko" sort="S Rel, Jaakko" uniqKey="S Rel J" first="Jaakko" last="S Rel">Jaakko S Rel</name><affiliation><mods:affiliation>Neural Network Research Centre, Helsinki University of Technology, FIN - 02015, HUT, Finland</mods:affiliation></affiliation><affiliation><mods:affiliation>E-mail: Jaakko.Sarela@hut.fi</mods:affiliation></affiliation></author><author><name sortKey="Vigario, Ricardo" sort="Vigario, Ricardo" uniqKey="Vigario R" first="Ricardo" last="Vigário">Ricardo Vigário</name><affiliation><mods:affiliation>GMD - FIRST, D- 12489, Kekuléstrasse 7, Berlin, Germany</mods:affiliation></affiliation><affiliation><mods:affiliation>E-mail: Ricardo.Vigario@hut.fi</mods:affiliation></affiliation></author></titleStmt><publicationStmt><idno type="wicri:source">ISTEX</idno><idno type="RBID">ISTEX:EED15B332943C34B00EA62FAC243875440F299CB</idno><date when="2001" year="2001">2001</date><idno type="doi">10.1007/3-540-45723-2_99</idno><idno type="url">https://api.istex.fr/document/EED15B332943C34B00EA62FAC243875440F299CB/fulltext/pdf</idno><idno type="wicri:Area/Main/Corpus">000806</idno><idno type="wicri:explorRef" wicri:stream="Main" wicri:step="Corpus" wicri:corpus="ISTEX">000806</idno></publicationStmt><sourceDesc><biblStruct><analytic><title level="a" type="main" xml:lang="en">The Problem of Overlearning in High-Order ICA Approaches: Analysis and Solutions</title><author><name sortKey="S Rel, Jaakko" sort="S Rel, Jaakko" uniqKey="S Rel J" first="Jaakko" last="S Rel">Jaakko S Rel</name><affiliation><mods:affiliation>Neural Network Research Centre, Helsinki University of Technology, FIN - 02015, HUT, Finland</mods:affiliation></affiliation><affiliation><mods:affiliation>E-mail: Jaakko.Sarela@hut.fi</mods:affiliation></affiliation></author><author><name sortKey="Vigario, Ricardo" sort="Vigario, Ricardo" uniqKey="Vigario R" first="Ricardo" last="Vigário">Ricardo Vigário</name><affiliation><mods:affiliation>GMD - FIRST, D- 12489, Kekuléstrasse 7, Berlin, Germany</mods:affiliation></affiliation><affiliation><mods:affiliation>E-mail: Ricardo.Vigario@hut.fi</mods:affiliation></affiliation></author></analytic><monogr></monogr><series><title level="s">Lecture Notes in Computer Science</title><imprint><date>2001</date></imprint><idno type="ISSN">0302-9743</idno><idno type="ISSN">0302-9743</idno></series><idno type="istex">EED15B332943C34B00EA62FAC243875440F299CB</idno><idno type="DOI">10.1007/3-540-45723-2_99</idno><idno type="ChapterID">99</idno><idno type="ChapterID">Chap99</idno></biblStruct></sourceDesc><seriesStmt><idno type="ISSN">0302-9743</idno></seriesStmt></fileDesc><profileDesc><textClass></textClass><langUsage><language ident="en">en</language></langUsage></profileDesc></teiHeader><front><div type="abstract" xml:lang="en">Abstract: We consider a type of overlearning typical of independent component analysis algorithms. These can be seen to minimize the mutual information between source estimates. The overlearning causes spikelike signals if there are too few samples or there is a considerable amount of noise present. It is argued that if the data has flicker noise the problem is more severe and is better characterized by bumps instead of spikes. The problem is demonstrated using recorded magnetoencephalographic signals. Several methods are suggested that attempt to solve the overlearning problem or, at least, diminish reduce its effects.</div></front></TEI><istex><corpusName>springer</corpusName><author><json:item><name>Jaakko Särelä</name><affiliations><json:string>Neural Network Research Centre, Helsinki University of Technology, FIN - 02015, HUT, Finland</json:string><json:string>E-mail: Jaakko.Sarela@hut.fi</json:string></affiliations></json:item><json:item><name>Ricardo Vigário</name><affiliations><json:string>GMD - FIRST, D- 12489, Kekuléstrasse 7, Berlin, Germany</json:string><json:string>E-mail: Ricardo.Vigario@hut.fi</json:string></affiliations></json:item></author><language><json:string>eng</json:string></language><originalGenre><json:string>OriginalPaper</json:string></originalGenre><abstract>Abstract: We consider a type of overlearning typical of independent component analysis algorithms. These can be seen to minimize the mutual information between source estimates. The overlearning causes spikelike signals if there are too few samples or there is a considerable amount of noise present. It is argued that if the data has flicker noise the problem is more severe and is better characterized by bumps instead of spikes. The problem is demonstrated using recorded magnetoencephalographic signals. Several methods are suggested that attempt to solve the overlearning problem or, at least, diminish reduce its effects.</abstract><qualityIndicators><score>4.268</score><pdfVersion>1.3</pdfVersion><pdfPageSize>430.536 x 666.216 pts</pdfPageSize><refBibsNative>false</refBibsNative><keywordCount>0</keywordCount><abstractCharCount>627</abstractCharCount><pdfWordCount>3128</pdfWordCount><pdfCharCount>16458</pdfCharCount><pdfPageCount>8</pdfPageCount><abstractWordCount>95</abstractWordCount></qualityIndicators><title>The Problem of Overlearning in High-Order ICA Approaches: Analysis and Solutions</title><chapterId><json:string>99</json:string><json:string>Chap99</json:string></chapterId><genre><json:string>conference</json:string></genre><serie><editor><json:item><name>Gerhard Goos</name><affiliations><json:string>Karlsruhe University, Germany</json:string></affiliations></json:item><json:item><name>Juris Hartmanis</name><affiliations><json:string>Cornell University, NY, USA</json:string></affiliations></json:item><json:item><name>Jan van Leeuwen</name><affiliations><json:string>Utrecht University, The Netherlands</json:string></affiliations></json:item></editor><issn><json:string>0302-9743</json:string></issn><language><json:string>unknown</json:string></language><title>Lecture Notes in Computer Science</title><copyrightDate>2001</copyrightDate></serie><host><editor><json:item><name>José Mira</name><affiliations><json:string>Departamento de Inteligencia Artificial, Universidad Nacional de Educación a Distancia, Senda del Rey, s/n., 28040, Madrid, Spain</json:string><json:string>E-mail: jmira@dia.uned.es</json:string></affiliations></json:item><json:item><name>Alberto Prieto</name><affiliations><json:string>Departamento de Arquitectura y Tecnología de Computadores, Universidad de Granada, Campus Fuentenueva, 18071, Granada, Spain</json:string><json:string>E-mail: aprieto@atc.ugr.es</json:string></affiliations></json:item></editor><subject><json:item><value>Computer Science</value></json:item><json:item><value>Computer Science</value></json:item><json:item><value>Computation by Abstract Devices</value></json:item><json:item><value>Algorithm Analysis and Problem Complexity</value></json:item><json:item><value>Artificial Intelligence (incl. Robotics)</value></json:item><json:item><value>Neurosciences</value></json:item><json:item><value>Neurology</value></json:item><json:item><value>Computer Appl. in Life Sciences</value></json:item></subject><isbn><json:string>978-3-540-42237-2</json:string></isbn><language><json:string>unknown</json:string></language><title>Bio-Inspired Applications of Connectionism</title><bookId><json:string>3-540-45723-2</json:string></bookId><volume>2085</volume><pages><last>825</last><first>818</first></pages><issn><json:string>0302-9743</json:string></issn><genre><json:string>book-series</json:string></genre><eisbn><json:string>978-3-540-45723-7</json:string></eisbn><copyrightDate>2001</copyrightDate><doi><json:string>10.1007/3-540-45723-2</json:string></doi></host><publicationDate>2001</publicationDate><copyrightDate>2001</copyrightDate><doi><json:string>10.1007/3-540-45723-2_99</json:string></doi><id>EED15B332943C34B00EA62FAC243875440F299CB</id><score>0.15858607</score><fulltext><json:item><original>true</original><mimetype>application/pdf</mimetype><extension>pdf</extension><uri>https://api.istex.fr/document/EED15B332943C34B00EA62FAC243875440F299CB/fulltext/pdf</uri></json:item><json:item><original>false</original><mimetype>application/zip</mimetype><extension>zip</extension><uri>https://api.istex.fr/document/EED15B332943C34B00EA62FAC243875440F299CB/fulltext/zip</uri></json:item><istex:fulltextTEI uri="https://api.istex.fr/document/EED15B332943C34B00EA62FAC243875440F299CB/fulltext/tei"><teiHeader><fileDesc><titleStmt><title level="a" type="main" xml:lang="en">The Problem of Overlearning in High-Order ICA Approaches: Analysis and Solutions</title><respStmt xml:id="ISTEX-API" resp="Références bibliographiques récupérées via GROBID" name="ISTEX-API (INIST-CNRS)"></respStmt><respStmt><resp>Références bibliographiques récupérées via GROBID</resp><name resp="ISTEX-API">ISTEX-API (INIST-CNRS)</name></respStmt></titleStmt><publicationStmt><authority>ISTEX</authority><publisher>Springer Berlin Heidelberg</publisher><pubPlace>Berlin, Heidelberg</pubPlace><availability><p>SPRINGER</p></availability><date>2001</date></publicationStmt><sourceDesc><biblStruct type="inbook"><analytic><title level="a" type="main" xml:lang="en">The Problem of Overlearning in High-Order ICA Approaches: Analysis and Solutions</title><author><persName><forename type="first">Jaakko</forename><surname>Särelä</surname></persName><email>Jaakko.Sarela@hut.fi</email><affiliation>Neural Network Research Centre, Helsinki University of Technology, FIN - 02015, HUT, Finland</affiliation></author><author><persName><forename type="first">Ricardo</forename><surname>Vigário</surname></persName><email>Ricardo.Vigario@hut.fi</email><affiliation>GMD - FIRST, D- 12489, Kekuléstrasse 7, Berlin, Germany</affiliation></author></analytic><monogr><title level="m">Bio-Inspired Applications of Connectionism</title><title level="m" type="sub">6th International Work-Conference on Artificial and Natural Neural Networks, IWANN 2001 Granada, Spain, June 13–15, 2001 Proceedings, Part II</title><idno type="pISBN">978-3-540-42237-2</idno><idno type="eISBN">978-3-540-45723-7</idno><idno type="pISSN">0302-9743</idno><idno type="DOI">10.1007/3-540-45723-2</idno><idno type="BookID">3-540-45723-2</idno><idno type="BookTitleID">69374</idno><idno type="BookSequenceNumber">2085</idno><idno type="BookVolumeNumber">2085</idno><idno type="BookChapterCount">101</idno><editor><persName><forename type="first">José</forename><surname>Mira</surname></persName><email>jmira@dia.uned.es</email><affiliation>Departamento de Inteligencia Artificial, Universidad Nacional de Educación a Distancia, Senda del Rey, s/n., 28040, Madrid, Spain</affiliation></editor><editor><persName><forename type="first">Alberto</forename><surname>Prieto</surname></persName><email>aprieto@atc.ugr.es</email><affiliation>Departamento de Arquitectura y Tecnología de Computadores, Universidad de Granada, Campus Fuentenueva, 18071, Granada, Spain</affiliation></editor><imprint><publisher>Springer Berlin Heidelberg</publisher><pubPlace>Berlin, Heidelberg</pubPlace><date type="published" when="2001"></date><biblScope unit="volume">2085</biblScope><biblScope unit="page" from="818">818</biblScope><biblScope unit="page" to="825">825</biblScope></imprint></monogr><series><title level="s">Lecture Notes in Computer Science</title><editor><persName><forename type="first">Gerhard</forename><surname>Goos</surname></persName><affiliation>Karlsruhe University, Germany</affiliation></editor><editor><persName><forename type="first">Juris</forename><surname>Hartmanis</surname></persName><affiliation>Cornell University, NY, USA</affiliation></editor><editor><persName><forename type="first">Jan</forename><surname>van Leeuwen</surname></persName><affiliation>Utrecht University, The Netherlands</affiliation></editor><biblScope><date>2001</date></biblScope><idno type="pISSN">0302-9743</idno><idno type="seriesId">558</idno></series><idno type="istex">EED15B332943C34B00EA62FAC243875440F299CB</idno><idno type="DOI">10.1007/3-540-45723-2_99</idno><idno type="ChapterID">99</idno><idno type="ChapterID">Chap99</idno></biblStruct></sourceDesc></fileDesc><profileDesc><creation><date>2001</date></creation><langUsage><language ident="en">en</language></langUsage><abstract xml:lang="en"><p>Abstract: We consider a type of overlearning typical of independent component analysis algorithms. These can be seen to minimize the mutual information between source estimates. The overlearning causes spikelike signals if there are too few samples or there is a considerable amount of noise present. It is argued that if the data has flicker noise the problem is more severe and is better characterized by bumps instead of spikes. The problem is demonstrated using recorded magnetoencephalographic signals. Several methods are suggested that attempt to solve the overlearning problem or, at least, diminish reduce its effects.</p></abstract><textClass><keywords scheme="Book Subject Collection"><list><label>SUCO11645</label><item><term>Computer Science</term></item></list></keywords></textClass><textClass><keywords scheme="Book Subject Group"><list><label>I</label><label>I16013</label><label>I16021</label><label>I21017</label><label>B18006</label><label>H36001</label><label>L17004</label><item><term>Computer Science</term></item><item><term>Computation by Abstract Devices</term></item><item><term>Algorithm Analysis and Problem Complexity</term></item><item><term>Artificial Intelligence (incl. Robotics)</term></item><item><term>Neurosciences</term></item><item><term>Neurology</term></item><item><term>Computer Appl. in Life Sciences</term></item></list></keywords></textClass></profileDesc><revisionDesc><change when="2001">Published</change><change xml:id="refBibs-istex" who="#ISTEX-API" when="2016-3-20">References added</change><change xml:id="refBibs-istex" who="#ISTEX-API" when="2016-07-26">References added</change></revisionDesc></teiHeader></istex:fulltextTEI><json:item><original>false</original><mimetype>text/plain</mimetype><extension>txt</extension><uri>https://api.istex.fr/document/EED15B332943C34B00EA62FAC243875440F299CB/fulltext/txt</uri></json:item></fulltext><metadata><istex:metadataXml wicri:clean="Springer, Publisher found" wicri:toSee="no header"><istex:xmlDeclaration>version="1.0" encoding="UTF-8"</istex:xmlDeclaration><istex:docType PUBLIC="-//Springer-Verlag//DTD A++ V2.4//EN" URI="http://devel.springer.de/A++/V2.4/DTD/A++V2.4.dtd" name="istex:docType"></istex:docType><istex:document><Publisher><PublisherInfo><PublisherName>Springer Berlin Heidelberg</PublisherName><PublisherLocation>Berlin, Heidelberg</PublisherLocation></PublisherInfo><Series><SeriesInfo SeriesType="Series" TocLevels="0"><SeriesID>558</SeriesID><SeriesPrintISSN>0302-9743</SeriesPrintISSN><SeriesTitle Language="En">Lecture Notes in Computer Science</SeriesTitle></SeriesInfo><SeriesHeader><EditorGroup><Editor AffiliationIDS="Aff1"><EditorName DisplayOrder="Western"><GivenName>Gerhard</GivenName><FamilyName>Goos</FamilyName></EditorName></Editor><Editor AffiliationIDS="Aff2"><EditorName DisplayOrder="Western"><GivenName>Juris</GivenName><FamilyName>Hartmanis</FamilyName></EditorName></Editor><Editor AffiliationIDS="Aff3"><EditorName DisplayOrder="Western"><GivenName>Jan</GivenName><Particle>van</Particle><FamilyName>Leeuwen</FamilyName></EditorName></Editor><Affiliation ID="Aff1"><OrgName>Karlsruhe University</OrgName><OrgAddress><Country>Germany</Country></OrgAddress></Affiliation><Affiliation ID="Aff2"><OrgName>Cornell University</OrgName><OrgAddress><State>NY</State><Country>USA</Country></OrgAddress></Affiliation><Affiliation ID="Aff3"><OrgName>Utrecht University</OrgName><OrgAddress><Country>The Netherlands</Country></OrgAddress></Affiliation></EditorGroup></SeriesHeader><Book Language="En"><BookInfo BookProductType="Proceedings" ContainsESM="No" Language="En" MediaType="eBook" NumberingStyle="Unnumbered" TocLevels="0"><BookID>3-540-45723-2</BookID><BookTitle>Bio-Inspired Applications of Connectionism</BookTitle><BookSubTitle>6th International Work-Conference on Artificial and Natural Neural Networks, IWANN 2001 Granada, Spain, June 13–15, 2001 Proceedings, Part II</BookSubTitle><BookVolumeNumber>2085</BookVolumeNumber><BookSequenceNumber>2085</BookSequenceNumber><BookDOI>10.1007/3-540-45723-2</BookDOI><BookTitleID>69374</BookTitleID><BookPrintISBN>978-3-540-42237-2</BookPrintISBN><BookElectronicISBN>978-3-540-45723-7</BookElectronicISBN><BookChapterCount>101</BookChapterCount><BookHistory><OnlineDate><Year>2001</Year><Month>6</Month><Day>12</Day></OnlineDate></BookHistory><BookCopyright><CopyrightHolderName>Springer-Verlag Berlin Heidelberg</CopyrightHolderName><CopyrightYear>2001</CopyrightYear></BookCopyright><BookSubjectGroup><BookSubject Code="I" Type="Primary">Computer Science</BookSubject><BookSubject Code="I16013" Priority="1" Type="Secondary">Computation by Abstract Devices</BookSubject><BookSubject Code="I16021" Priority="2" Type="Secondary">Algorithm Analysis and Problem Complexity</BookSubject><BookSubject Code="I21017" Priority="3" Type="Secondary">Artificial Intelligence (incl. Robotics)</BookSubject><BookSubject Code="B18006" Priority="4" Type="Secondary">Neurosciences</BookSubject><BookSubject Code="H36001" Priority="5" Type="Secondary">Neurology</BookSubject><BookSubject Code="L17004" Priority="6" Type="Secondary">Computer Appl. in Life Sciences</BookSubject><SubjectCollection Code="SUCO11645">Computer Science</SubjectCollection></BookSubjectGroup><BookContext><SeriesID>558</SeriesID></BookContext></BookInfo><BookHeader><EditorGroup><Editor AffiliationIDS="Aff4"><EditorName DisplayOrder="Western"><GivenName>José</GivenName><FamilyName>Mira</FamilyName></EditorName><Contact><Email>jmira@dia.uned.es</Email></Contact></Editor><Editor AffiliationIDS="Aff5"><EditorName DisplayOrder="Western"><GivenName>Alberto</GivenName><FamilyName>Prieto</FamilyName></EditorName><Contact><Email>aprieto@atc.ugr.es</Email></Contact></Editor><Affiliation ID="Aff4"><OrgDivision>Departamento de Inteligencia Artificial</OrgDivision><OrgName>Universidad Nacional de Educación a Distancia</OrgName><OrgAddress><Street>Senda del Rey, s/n.</Street><Postcode>28040</Postcode><City>Madrid</City><Country>Spain</Country></OrgAddress></Affiliation><Affiliation ID="Aff5"><OrgDivision>Departamento de Arquitectura y Tecnología de Computadores</OrgDivision><OrgName>Universidad de Granada</OrgName><OrgAddress><Street>Campus Fuentenueva</Street><Postcode>18071</Postcode><City>Granada</City><Country>Spain</Country></OrgAddress></Affiliation></EditorGroup></BookHeader><Part ID="Part6"><PartInfo TocLevels="0"><PartID>6</PartID><PartSequenceNumber>6</PartSequenceNumber><PartTitle>General Applications</PartTitle><PartChapterCount>38</PartChapterCount><PartContext><SeriesID>558</SeriesID><BookID>3-540-45723-2</BookID><BookTitle>Bio-Inspired Applications of Connectionism</BookTitle></PartContext></PartInfo><Chapter ID="Chap99" Language="En"><ChapterInfo ChapterType="OriginalPaper" ContainsESM="No" Language="En" NumberingStyle="Unnumbered" TocLevels="0"><ChapterID>99</ChapterID><ChapterDOI>10.1007/3-540-45723-2_99</ChapterDOI><ChapterSequenceNumber>99</ChapterSequenceNumber><ChapterTitle Language="En">The Problem of Overlearning in High-Order ICA Approaches: Analysis and Solutions</ChapterTitle><ChapterFirstPage>818</ChapterFirstPage><ChapterLastPage>825</ChapterLastPage><ChapterCopyright><CopyrightHolderName>Springer-Verlag Berlin Heidelberg</CopyrightHolderName><CopyrightYear>2001</CopyrightYear></ChapterCopyright><ChapterHistory><RegistrationDate><Year>2001</Year><Month>6</Month><Day>11</Day></RegistrationDate><OnlineDate><Year>2001</Year><Month>6</Month><Day>12</Day></OnlineDate></ChapterHistory><ChapterGrants Type="Regular"><MetadataGrant Grant="OpenAccess"></MetadataGrant><AbstractGrant Grant="OpenAccess"></AbstractGrant><BodyPDFGrant Grant="Restricted"></BodyPDFGrant><BodyHTMLGrant Grant="Restricted"></BodyHTMLGrant><BibliographyGrant Grant="Restricted"></BibliographyGrant><ESMGrant Grant="Restricted"></ESMGrant></ChapterGrants><ChapterContext><SeriesID>558</SeriesID><PartID>6</PartID><BookID>3-540-45723-2</BookID><BookTitle>Bio-Inspired Applications of Connectionism</BookTitle></ChapterContext></ChapterInfo><ChapterHeader><AuthorGroup><Author AffiliationIDS="Aff6"><AuthorName DisplayOrder="Western"><GivenName>Jaakko</GivenName><FamilyName>Särelä</FamilyName></AuthorName><Contact><Email>Jaakko.Sarela@hut.fi</Email></Contact></Author><Author AffiliationIDS="Aff7"><AuthorName DisplayOrder="Western"><GivenName>Ricardo</GivenName><FamilyName>Vigário</FamilyName></AuthorName><Contact><Email>Ricardo.Vigario@hut.fi</Email></Contact></Author><Affiliation ID="Aff6"><OrgDivision>Neural Network Research Centre</OrgDivision><OrgName>Helsinki University of Technology</OrgName><OrgAddress><Postcode>FIN - 02015</Postcode><City>HUT</City><Country>Finland</Country></OrgAddress></Affiliation><Affiliation ID="Aff7"><OrgName>GMD - FIRST</OrgName><OrgAddress><City>Kekuléstrasse 7</City><Postcode>D- 12489</Postcode><State>Berlin</State><Country>Germany</Country></OrgAddress></Affiliation></AuthorGroup><Abstract ID="Abs1" Language="En"><Heading>Abstract</Heading><Para>We consider a type of overlearning typical of independent component analysis algorithms. These can be seen to minimize the mutual information between source estimates. The overlearning causes spikelike signals if there are too few samples or there is a considerable amount of noise present. It is argued that if the data has flicker noise the problem is more severe and is better characterized by bumps instead of spikes. The problem is demonstrated using recorded magnetoencephalographic signals. Several methods are suggested that attempt to solve the overlearning problem or, at least, diminish reduce its effects.</Para></Abstract></ChapterHeader><NoBody></NoBody></Chapter></Part></Book></Series></Publisher></istex:document></istex:metadataXml><mods version="3.6"><titleInfo lang="en"><title>The Problem of Overlearning in High-Order ICA Approaches: Analysis and Solutions</title></titleInfo><titleInfo type="alternative" contentType="CDATA" lang="en"><title>The Problem of Overlearning in High-Order ICA Approaches: Analysis and Solutions</title></titleInfo><name type="personal"><namePart type="given">Jaakko</namePart><namePart type="family">Särelä</namePart><affiliation>Neural Network Research Centre, Helsinki University of Technology, FIN - 02015, HUT, Finland</affiliation><affiliation>E-mail: Jaakko.Sarela@hut.fi</affiliation><role><roleTerm type="text">author</roleTerm></role></name><name type="personal"><namePart type="given">Ricardo</namePart><namePart type="family">Vigário</namePart><affiliation>GMD - FIRST, D- 12489, Kekuléstrasse 7, Berlin, Germany</affiliation><affiliation>E-mail: Ricardo.Vigario@hut.fi</affiliation><role><roleTerm type="text">author</roleTerm></role></name><typeOfResource>text</typeOfResource><genre type="conference" displayLabel="OriginalPaper"></genre><originInfo><publisher>Springer Berlin Heidelberg</publisher><place><placeTerm type="text">Berlin, Heidelberg</placeTerm></place><dateIssued encoding="w3cdtf">2001-06-12</dateIssued><copyrightDate encoding="w3cdtf">2001</copyrightDate></originInfo><language><languageTerm type="code" authority="rfc3066">en</languageTerm><languageTerm type="code" authority="iso639-2b">eng</languageTerm></language><physicalDescription><internetMediaType>text/html</internetMediaType></physicalDescription><abstract lang="en">Abstract: We consider a type of overlearning typical of independent component analysis algorithms. These can be seen to minimize the mutual information between source estimates. The overlearning causes spikelike signals if there are too few samples or there is a considerable amount of noise present. It is argued that if the data has flicker noise the problem is more severe and is better characterized by bumps instead of spikes. The problem is demonstrated using recorded magnetoencephalographic signals. Several methods are suggested that attempt to solve the overlearning problem or, at least, diminish reduce its effects.</abstract><relatedItem type="host"><titleInfo><title>Bio-Inspired Applications of Connectionism</title><subTitle>6th International Work-Conference on Artificial and Natural Neural Networks, IWANN 2001 Granada, Spain, June 13–15, 2001 Proceedings, Part II</subTitle></titleInfo><name type="personal"><namePart type="given">José</namePart><namePart type="family">Mira</namePart><affiliation>Departamento de Inteligencia Artificial, Universidad Nacional de Educación a Distancia, Senda del Rey, s/n., 28040, Madrid, Spain</affiliation><affiliation>E-mail: jmira@dia.uned.es</affiliation><role><roleTerm type="text">editor</roleTerm></role></name><name type="personal"><namePart type="given">Alberto</namePart><namePart type="family">Prieto</namePart><affiliation>Departamento de Arquitectura y Tecnología de Computadores, Universidad de Granada, Campus Fuentenueva, 18071, Granada, Spain</affiliation><affiliation>E-mail: aprieto@atc.ugr.es</affiliation><role><roleTerm type="text">editor</roleTerm></role></name><genre type="book-series" displayLabel="Proceedings"></genre><originInfo><copyrightDate encoding="w3cdtf">2001</copyrightDate><issuance>monographic</issuance></originInfo><subject><genre>Book-Subject-Collection</genre><topic authority="SpringerSubjectCodes" authorityURI="SUCO11645">Computer Science</topic></subject><subject><genre>Book-Subject-Group</genre><topic authority="SpringerSubjectCodes" authorityURI="I">Computer Science</topic><topic authority="SpringerSubjectCodes" authorityURI="I16013">Computation by Abstract Devices</topic><topic authority="SpringerSubjectCodes" authorityURI="I16021">Algorithm Analysis and Problem Complexity</topic><topic authority="SpringerSubjectCodes" authorityURI="I21017">Artificial Intelligence (incl. Robotics)</topic><topic authority="SpringerSubjectCodes" authorityURI="B18006">Neurosciences</topic><topic authority="SpringerSubjectCodes" authorityURI="H36001">Neurology</topic><topic authority="SpringerSubjectCodes" authorityURI="L17004">Computer Appl. in Life Sciences</topic></subject><identifier type="DOI">10.1007/3-540-45723-2</identifier><identifier type="ISBN">978-3-540-42237-2</identifier><identifier type="eISBN">978-3-540-45723-7</identifier><identifier type="ISSN">0302-9743</identifier><identifier type="BookTitleID">69374</identifier><identifier type="BookID">3-540-45723-2</identifier><identifier type="BookChapterCount">101</identifier><identifier type="BookVolumeNumber">2085</identifier><identifier type="BookSequenceNumber">2085</identifier><identifier type="PartChapterCount">38</identifier><part><date>2001</date><detail type="part"><title>General Applications</title></detail><detail type="volume"><number>2085</number><caption>vol.</caption></detail><extent unit="pages"><start>818</start><end>825</end></extent></part><recordInfo><recordOrigin>Springer-Verlag Berlin Heidelberg, 2001</recordOrigin></recordInfo></relatedItem><relatedItem type="series"><titleInfo><title>Lecture Notes in Computer Science</title></titleInfo><name type="personal"><namePart type="given">Gerhard</namePart><namePart type="family">Goos</namePart><affiliation>Karlsruhe University, Germany</affiliation><role><roleTerm type="text">editor</roleTerm></role></name><name type="personal"><namePart type="given">Juris</namePart><namePart type="family">Hartmanis</namePart><affiliation>Cornell University, NY, USA</affiliation><role><roleTerm type="text">editor</roleTerm></role></name><name type="personal"><namePart type="given">Jan</namePart><namePart type="family">van Leeuwen</namePart><affiliation>Utrecht University, The Netherlands</affiliation><role><roleTerm type="text">editor</roleTerm></role></name><originInfo><copyrightDate encoding="w3cdtf">2001</copyrightDate><issuance>serial</issuance></originInfo><identifier type="ISSN">0302-9743</identifier><identifier type="SeriesID">558</identifier><recordInfo><recordOrigin>Springer-Verlag Berlin Heidelberg, 2001</recordOrigin></recordInfo></relatedItem><identifier type="istex">EED15B332943C34B00EA62FAC243875440F299CB</identifier><identifier type="DOI">10.1007/3-540-45723-2_99</identifier><identifier type="ChapterID">99</identifier><identifier type="ChapterID">Chap99</identifier><accessCondition type="use and reproduction" contentType="copyright">Springer-Verlag Berlin Heidelberg, 2001</accessCondition><recordInfo><recordContentSource>SPRINGER</recordContentSource><recordOrigin>Springer-Verlag Berlin Heidelberg, 2001</recordOrigin></recordInfo></mods></metadata></istex></record>Pour manipuler ce document sous Unix (Dilib)
EXPLOR_STEP=$WICRI_ROOT/Wicri/France/explor/AussoisV1/Data/Main/Corpus
HfdSelect -h $EXPLOR_STEP/biblio.hfd -nk 000806 | SxmlIndent | more
Ou
HfdSelect -h $EXPLOR_AREA/Data/Main/Corpus/biblio.hfd -nk 000806 | SxmlIndent | more
Pour mettre un lien sur cette page dans le réseau Wicri
{{Explor lien
|wiki= Wicri/France
|area= AussoisV1
|flux= Main
|étape= Corpus
|type= RBID
|clé= ISTEX:EED15B332943C34B00EA62FAC243875440F299CB
|texte= The Problem of Overlearning in High-Order ICA Approaches: Analysis and Solutions
}}
| This area was generated with Dilib version V0.6.29. |  |