Sequence Labelling SVMs Trained in One Pass
Identifieur interne : 000B27 ( Main/Exploration ); précédent : 000B26; suivant : 000B28Sequence Labelling SVMs Trained in One Pass
Auteurs : Antoine Bordes [France, États-Unis] ; Nicolas Usunier [France] ; Léon Bottou [États-Unis]Source :
- Lecture Notes in Computer Science [ 0302-9743 ] ; 2008.
Abstract
Abstract: This paper proposes an online solver of the dual formulation of support vector machines for structured output spaces. We apply it to sequence labelling using the exact and greedy inference schemes. In both cases, the per-sequence training time is the same as a perceptron based on the same inference procedure, up to a small multiplicative constant. Comparing the two inference schemes, the greedy version is much faster. It is also amenable to higher order Markov assumptions and performs similarly on test. In comparison to existing algorithms, both versions match the accuracies of batch solvers that use exact inference after a single pass over the training examples.
Url:
DOI: 10.1007/978-3-540-87479-9_28
Affiliations:
Links toward previous steps (curation, corpus...)
- to stream Istex, to step Corpus: 000B95
- to stream Istex, to step Curation: 000B80
- to stream Istex, to step Checkpoint: 000593
- to stream Main, to step Merge: 000B39
- to stream Main, to step Curation: 000B27
Le document en format XML
<record><TEI wicri:istexFullTextTei="biblStruct:series"><teiHeader><fileDesc><titleStmt><title xml:lang="en">Sequence Labelling SVMs Trained in One Pass</title><author><name sortKey="Bordes, Antoine" sort="Bordes, Antoine" uniqKey="Bordes A" first="Antoine" last="Bordes">Antoine Bordes</name></author><author><name sortKey="Usunier, Nicolas" sort="Usunier, Nicolas" uniqKey="Usunier N" first="Nicolas" last="Usunier">Nicolas Usunier</name></author><author><name sortKey="Bottou, Leon" sort="Bottou, Leon" uniqKey="Bottou L" first="Léon" last="Bottou">Léon Bottou</name></author></titleStmt><publicationStmt><idno type="wicri:source">ISTEX</idno><idno type="RBID">ISTEX:A99CCAE9C28CA662661AA10D92F1EC81149AF5AF</idno><date when="2008" year="2008">2008</date><idno type="doi">10.1007/978-3-540-87479-9_28</idno><idno type="url">https://api.istex.fr/document/A99CCAE9C28CA662661AA10D92F1EC81149AF5AF/fulltext/pdf</idno><idno type="wicri:Area/Istex/Corpus">000B95</idno><idno type="wicri:Area/Istex/Curation">000B80</idno><idno type="wicri:Area/Istex/Checkpoint">000593</idno><idno type="wicri:doubleKey">0302-9743:2008:Bordes A:sequence:labelling:svms</idno><idno type="wicri:Area/Main/Merge">000B39</idno><idno type="wicri:Area/Main/Curation">000B27</idno><idno type="wicri:Area/Main/Exploration">000B27</idno></publicationStmt><sourceDesc><biblStruct><analytic><title level="a" type="main" xml:lang="en">Sequence Labelling SVMs Trained in One Pass</title><author><name sortKey="Bordes, Antoine" sort="Bordes, Antoine" uniqKey="Bordes A" first="Antoine" last="Bordes">Antoine Bordes</name><affiliation wicri:level="3"><country xml:lang="fr">France</country><wicri:regionArea>LIP6, Université Paris 6, 104 Avenue du Pdt Kennedy, 75016, Paris</wicri:regionArea><placeName><region type="region" nuts="2">Île-de-France</region><settlement type="city">Paris</settlement></placeName></affiliation><affiliation wicri:level="1"><country xml:lang="fr">États-Unis</country><wicri:regionArea>NEC Laboratories America, Inc., 4 Independence Way, NJ08540, Princeton</wicri:regionArea><wicri:noRegion>Princeton</wicri:noRegion></affiliation><affiliation wicri:level="1"><country wicri:rule="url">France</country></affiliation></author><author><name sortKey="Usunier, Nicolas" sort="Usunier, Nicolas" uniqKey="Usunier N" first="Nicolas" last="Usunier">Nicolas Usunier</name><affiliation wicri:level="3"><country xml:lang="fr">France</country><wicri:regionArea>LIP6, Université Paris 6, 104 Avenue du Pdt Kennedy, 75016, Paris</wicri:regionArea><placeName><region type="region" nuts="2">Île-de-France</region><settlement type="city">Paris</settlement></placeName></affiliation><affiliation wicri:level="1"><country wicri:rule="url">France</country></affiliation></author><author><name sortKey="Bottou, Leon" sort="Bottou, Leon" uniqKey="Bottou L" first="Léon" last="Bottou">Léon Bottou</name><affiliation wicri:level="1"><country xml:lang="fr">États-Unis</country><wicri:regionArea>NEC Laboratories America, Inc., 4 Independence Way, NJ08540, Princeton</wicri:regionArea><wicri:noRegion>Princeton</wicri:noRegion></affiliation><affiliation><wicri:noCountry code="no comma">E-mail: leonb@nec-labs.com</wicri:noCountry></affiliation></author></analytic><monogr></monogr><series><title level="s">Lecture Notes in Computer Science</title><imprint><date>2008</date></imprint><idno type="ISSN">0302-9743</idno><idno type="eISSN">1611-3349</idno><idno type="ISSN">0302-9743</idno></series><idno type="istex">A99CCAE9C28CA662661AA10D92F1EC81149AF5AF</idno><idno type="DOI">10.1007/978-3-540-87479-9_28</idno><idno type="ChapterID">28</idno><idno type="ChapterID">Chap28</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: This paper proposes an online solver of the dual formulation of support vector machines for structured output spaces. We apply it to sequence labelling using the exact and greedy inference schemes. In both cases, the per-sequence training time is the same as a perceptron based on the same inference procedure, up to a small multiplicative constant. Comparing the two inference schemes, the greedy version is much faster. It is also amenable to higher order Markov assumptions and performs similarly on test. In comparison to existing algorithms, both versions match the accuracies of batch solvers that use exact inference after a single pass over the training examples.</div></front></TEI><affiliations><list><country><li>France</li><li>États-Unis</li></country><region><li>Île-de-France</li></region><settlement><li>Paris</li></settlement></list><tree><country name="France"><region name="Île-de-France"><name sortKey="Bordes, Antoine" sort="Bordes, Antoine" uniqKey="Bordes A" first="Antoine" last="Bordes">Antoine Bordes</name></region><name sortKey="Bordes, Antoine" sort="Bordes, Antoine" uniqKey="Bordes A" first="Antoine" last="Bordes">Antoine Bordes</name><name sortKey="Usunier, Nicolas" sort="Usunier, Nicolas" uniqKey="Usunier N" first="Nicolas" last="Usunier">Nicolas Usunier</name><name sortKey="Usunier, Nicolas" sort="Usunier, Nicolas" uniqKey="Usunier N" first="Nicolas" last="Usunier">Nicolas Usunier</name></country><country name="États-Unis"><noRegion><name sortKey="Bordes, Antoine" sort="Bordes, Antoine" uniqKey="Bordes A" first="Antoine" last="Bordes">Antoine Bordes</name></noRegion><name sortKey="Bottou, Leon" sort="Bottou, Leon" uniqKey="Bottou L" first="Léon" last="Bottou">Léon Bottou</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 000B27 | SxmlIndent | more
Ou
HfdSelect -h $EXPLOR_AREA/Data/Main/Exploration/biblio.hfd -nk 000B27 | 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:A99CCAE9C28CA662661AA10D92F1EC81149AF5AF
|texte= Sequence Labelling SVMs Trained in One Pass
}}
| This area was generated with Dilib version V0.6.32. |  |