Entropy and Margin Maximization for Structured Output Learning
Identifieur interne : 000623 ( Main/Exploration ); précédent : 000622; suivant : 000624Entropy and Margin Maximization for Structured Output Learning
Auteurs : Patrick Pletscher [Suisse] ; Soon Ong [Suisse] ; M. Buhmann [Suisse]Source :
- Lecture Notes in Computer Science [ 0302-9743 ] ; 2010.
Abstract
Abstract: We consider the problem of training discriminative structured output predictors, such as conditional random fields (CRFs) and structured support vector machines (SSVMs). A generalized loss function is introduced, which jointly maximizes the entropy and the margin of the solution. The CRF and SSVM emerge as special cases of our framework. The probabilistic interpretation of large margin methods reveals insights about margin and slack rescaling. Furthermore, we derive the corresponding extensions for latent variable models, in which training operates on partially observed outputs. Experimental results for multiclass, linear-chain models and multiple instance learning demonstrate that the generalized loss can improve accuracy of the resulting classifiers.
Url:
DOI: 10.1007/978-3-642-15939-8_6
Affiliations:
Links toward previous steps (curation, corpus...)
- to stream Istex, to step Corpus: 001E22
- to stream Istex, to step Curation: 001D00
- to stream Istex, to step Checkpoint: 000203
- to stream Main, to step Merge: 000628
- to stream Main, to step Curation: 000623
Le document en format XML
<record><TEI wicri:istexFullTextTei="biblStruct:series"><teiHeader><fileDesc><titleStmt><title xml:lang="en">Entropy and Margin Maximization for Structured Output Learning</title><author><name sortKey="Pletscher, Patrick" sort="Pletscher, Patrick" uniqKey="Pletscher P" first="Patrick" last="Pletscher">Patrick Pletscher</name></author><author><name sortKey="Ong, Soon" sort="Ong, Soon" uniqKey="Ong S" first="Soon" last="Ong">Soon Ong</name></author><author><name sortKey="Buhmann, M" sort="Buhmann, M" uniqKey="Buhmann M" first="M." last="Buhmann">M. Buhmann</name></author></titleStmt><publicationStmt><idno type="wicri:source">ISTEX</idno><idno type="RBID">ISTEX:20560ABF2EBDAD9BBEC72D0EC1347CA518344653</idno><date when="2010" year="2010">2010</date><idno type="doi">10.1007/978-3-642-15939-8_6</idno><idno type="url">https://api.istex.fr/document/20560ABF2EBDAD9BBEC72D0EC1347CA518344653/fulltext/pdf</idno><idno type="wicri:Area/Istex/Corpus">001E22</idno><idno type="wicri:Area/Istex/Curation">001D00</idno><idno type="wicri:Area/Istex/Checkpoint">000203</idno><idno type="wicri:doubleKey">0302-9743:2010:Pletscher P:entropy:and:margin</idno><idno type="wicri:Area/Main/Merge">000628</idno><idno type="wicri:Area/Main/Curation">000623</idno><idno type="wicri:Area/Main/Exploration">000623</idno></publicationStmt><sourceDesc><biblStruct><analytic><title level="a" type="main" xml:lang="en">Entropy and Margin Maximization for Structured Output Learning</title><author><name sortKey="Pletscher, Patrick" sort="Pletscher, Patrick" uniqKey="Pletscher P" first="Patrick" last="Pletscher">Patrick Pletscher</name><affiliation wicri:level="4"><country xml:lang="fr">Suisse</country><wicri:regionArea>Department of Computer Science, ETH Zürich</wicri:regionArea><placeName><settlement type="city">Zurich</settlement><region nuts="3" type="region">Canton de Zurich</region></placeName><orgName type="university">École polytechnique fédérale de Zurich</orgName></affiliation></author><author><name sortKey="Ong, Soon" sort="Ong, Soon" uniqKey="Ong S" first="Soon" last="Ong">Soon Ong</name><affiliation wicri:level="4"><country xml:lang="fr">Suisse</country><wicri:regionArea>Department of Computer Science, ETH Zürich</wicri:regionArea><placeName><settlement type="city">Zurich</settlement><region nuts="3" type="region">Canton de Zurich</region></placeName><orgName type="university">École polytechnique fédérale de Zurich</orgName></affiliation></author><author><name sortKey="Buhmann, M" sort="Buhmann, M" uniqKey="Buhmann M" first="M." last="Buhmann">M. Buhmann</name><affiliation wicri:level="4"><country xml:lang="fr">Suisse</country><wicri:regionArea>Department of Computer Science, ETH Zürich</wicri:regionArea><placeName><settlement type="city">Zurich</settlement><region nuts="3" type="region">Canton de Zurich</region></placeName><orgName type="university">École polytechnique fédérale de Zurich</orgName></affiliation></author></analytic><monogr></monogr><series><title level="s">Lecture Notes in Computer Science</title><imprint><date>2010</date></imprint><idno type="ISSN">0302-9743</idno><idno type="eISSN">1611-3349</idno><idno type="ISSN">0302-9743</idno></series><idno type="istex">20560ABF2EBDAD9BBEC72D0EC1347CA518344653</idno><idno type="DOI">10.1007/978-3-642-15939-8_6</idno><idno type="ChapterID">6</idno><idno type="ChapterID">Chap6</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 the problem of training discriminative structured output predictors, such as conditional random fields (CRFs) and structured support vector machines (SSVMs). A generalized loss function is introduced, which jointly maximizes the entropy and the margin of the solution. The CRF and SSVM emerge as special cases of our framework. The probabilistic interpretation of large margin methods reveals insights about margin and slack rescaling. Furthermore, we derive the corresponding extensions for latent variable models, in which training operates on partially observed outputs. Experimental results for multiclass, linear-chain models and multiple instance learning demonstrate that the generalized loss can improve accuracy of the resulting classifiers.</div></front></TEI><affiliations><list><country><li>Suisse</li></country><region><li>Canton de Zurich</li></region><settlement><li>Zurich</li></settlement><orgName><li>École polytechnique fédérale de Zurich</li></orgName></list><tree><country name="Suisse"><region name="Canton de Zurich"><name sortKey="Pletscher, Patrick" sort="Pletscher, Patrick" uniqKey="Pletscher P" first="Patrick" last="Pletscher">Patrick Pletscher</name></region><name sortKey="Buhmann, M" sort="Buhmann, M" uniqKey="Buhmann M" first="M." last="Buhmann">M. Buhmann</name><name sortKey="Ong, Soon" sort="Ong, Soon" uniqKey="Ong S" first="Soon" last="Ong">Soon Ong</name></country></tree></affiliations></record>Pour manipuler ce document sous Unix (Dilib)
EXPLOR_STEP=$WICRI_ROOT/Ticri/CIDE/explor/OcrV1/Data/Main/Exploration
HfdSelect -h $EXPLOR_STEP/biblio.hfd -nk 000623 | SxmlIndent | more
Ou
HfdSelect -h $EXPLOR_AREA/Data/Main/Exploration/biblio.hfd -nk 000623 | SxmlIndent | more
Pour mettre un lien sur cette page dans le réseau Wicri
{{Explor lien
|wiki= Ticri/CIDE
|area= OcrV1
|flux= Main
|étape= Exploration
|type= RBID
|clé= ISTEX:20560ABF2EBDAD9BBEC72D0EC1347CA518344653
|texte= Entropy and Margin Maximization for Structured Output Learning
}}
| This area was generated with Dilib version V0.6.32. |  |