Learn++.MT: A New Approach to Incremental Learning
Identifieur interne : 001513 ( Main/Exploration ); précédent : 001512; suivant : 001514Learn++.MT: A New Approach to Incremental Learning
Auteurs : Michael Muhlbaier [États-Unis] ; Apostolos Topalis [États-Unis] ; Robi Polikar [États-Unis]Source :
- Lecture Notes in Computer Science [ 0302-9743 ] ; 2004.
Abstract
Abstract: An ensemble of classifiers based algorithm, Learn++, was recently introduced that is capable of incrementally learning new information from datasets that consecutively become available, even if the new data introduce additional classes that were not formerly seen. The algorithm does not require access to previously used datasets, yet it is capable of largely retaining the previously acquired knowledge. However, Learn++ suffers from the inherent ”out-voting” problem when asked to learn new classes, which causes it to generate an unnecessarily large number of classifiers. This paper proposes a modified version of this algorithm, called Learn++.MT that not only reduces the number of classifiers generated, but also provides performance improvements. The out-voting problem, the new algorithm and its promising results on two benchmark datasets as well as on one real world application are presented.
Url:
DOI: 10.1007/978-3-540-25966-4_5
Affiliations:
Links toward previous steps (curation, corpus...)
- to stream Istex, to step Corpus: 001493
- to stream Istex, to step Curation: 001407
- to stream Istex, to step Checkpoint: 000D48
- to stream Main, to step Merge: 001564
- to stream Main, to step Curation: 001513
Le document en format XML
<record><TEI wicri:istexFullTextTei="biblStruct"><teiHeader><fileDesc><titleStmt><title xml:lang="en">Learn++.MT: A New Approach to Incremental Learning</title><author><name sortKey="Muhlbaier, Michael" sort="Muhlbaier, Michael" uniqKey="Muhlbaier M" first="Michael" last="Muhlbaier">Michael Muhlbaier</name></author><author><name sortKey="Topalis, Apostolos" sort="Topalis, Apostolos" uniqKey="Topalis A" first="Apostolos" last="Topalis">Apostolos Topalis</name></author><author><name sortKey="Polikar, Robi" sort="Polikar, Robi" uniqKey="Polikar R" first="Robi" last="Polikar">Robi Polikar</name></author></titleStmt><publicationStmt><idno type="wicri:source">ISTEX</idno><idno type="RBID">ISTEX:A225C5D713C63CEA13AA61314046B4C09D2E7DEC</idno><date when="2004" year="2004">2004</date><idno type="doi">10.1007/978-3-540-25966-4_5</idno><idno type="url">https://api.istex.fr/document/A225C5D713C63CEA13AA61314046B4C09D2E7DEC/fulltext/pdf</idno><idno type="wicri:Area/Istex/Corpus">001493</idno><idno type="wicri:Area/Istex/Curation">001407</idno><idno type="wicri:Area/Istex/Checkpoint">000D48</idno><idno type="wicri:doubleKey">0302-9743:2004:Muhlbaier M:learn:mt:a</idno><idno type="wicri:Area/Main/Merge">001564</idno><idno type="wicri:Area/Main/Curation">001513</idno><idno type="wicri:Area/Main/Exploration">001513</idno></publicationStmt><sourceDesc><biblStruct><analytic><title level="a" type="main" xml:lang="en">Learn++.MT: A New Approach to Incremental Learning</title><author><name sortKey="Muhlbaier, Michael" sort="Muhlbaier, Michael" uniqKey="Muhlbaier M" first="Michael" last="Muhlbaier">Michael Muhlbaier</name><affiliation wicri:level="2"><country xml:lang="fr">États-Unis</country><wicri:regionArea>Electrical and Computer Engineering Department, Rowan University, 201 Mullica Hill Rd., 08028, Glassboro, NJ</wicri:regionArea><placeName><region type="state">New Jersey</region></placeName></affiliation><affiliation wicri:level="1"><country wicri:rule="url">États-Unis</country></affiliation></author><author><name sortKey="Topalis, Apostolos" sort="Topalis, Apostolos" uniqKey="Topalis A" first="Apostolos" last="Topalis">Apostolos Topalis</name><affiliation wicri:level="2"><country xml:lang="fr">États-Unis</country><wicri:regionArea>Electrical and Computer Engineering Department, Rowan University, 201 Mullica Hill Rd., 08028, Glassboro, NJ</wicri:regionArea><placeName><region type="state">New Jersey</region></placeName></affiliation><affiliation wicri:level="1"><country wicri:rule="url">États-Unis</country></affiliation></author><author><name sortKey="Polikar, Robi" sort="Polikar, Robi" uniqKey="Polikar R" first="Robi" last="Polikar">Robi Polikar</name><affiliation wicri:level="2"><country xml:lang="fr">États-Unis</country><wicri:regionArea>Electrical and Computer Engineering Department, Rowan University, 201 Mullica Hill Rd., 08028, Glassboro, NJ</wicri:regionArea><placeName><region type="state">New Jersey</region></placeName></affiliation><affiliation wicri:level="1"><country wicri:rule="url">États-Unis</country></affiliation></author></analytic><monogr></monogr><series><title level="s">Lecture Notes in Computer Science</title><imprint><date>2004</date></imprint><idno type="ISSN">0302-9743</idno><idno type="eISSN">1611-3349</idno><idno type="ISSN">0302-9743</idno></series><idno type="istex">A225C5D713C63CEA13AA61314046B4C09D2E7DEC</idno><idno type="DOI">10.1007/978-3-540-25966-4_5</idno><idno type="ChapterID">5</idno><idno type="ChapterID">Chap5</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: An ensemble of classifiers based algorithm, Learn++, was recently introduced that is capable of incrementally learning new information from datasets that consecutively become available, even if the new data introduce additional classes that were not formerly seen. The algorithm does not require access to previously used datasets, yet it is capable of largely retaining the previously acquired knowledge. However, Learn++ suffers from the inherent ”out-voting” problem when asked to learn new classes, which causes it to generate an unnecessarily large number of classifiers. This paper proposes a modified version of this algorithm, called Learn++.MT that not only reduces the number of classifiers generated, but also provides performance improvements. The out-voting problem, the new algorithm and its promising results on two benchmark datasets as well as on one real world application are presented.</div></front></TEI><affiliations><list><country><li>États-Unis</li></country><region><li>New Jersey</li></region></list><tree><country name="États-Unis"><region name="New Jersey"><name sortKey="Muhlbaier, Michael" sort="Muhlbaier, Michael" uniqKey="Muhlbaier M" first="Michael" last="Muhlbaier">Michael Muhlbaier</name></region><name sortKey="Muhlbaier, Michael" sort="Muhlbaier, Michael" uniqKey="Muhlbaier M" first="Michael" last="Muhlbaier">Michael Muhlbaier</name><name sortKey="Polikar, Robi" sort="Polikar, Robi" uniqKey="Polikar R" first="Robi" last="Polikar">Robi Polikar</name><name sortKey="Polikar, Robi" sort="Polikar, Robi" uniqKey="Polikar R" first="Robi" last="Polikar">Robi Polikar</name><name sortKey="Topalis, Apostolos" sort="Topalis, Apostolos" uniqKey="Topalis A" first="Apostolos" last="Topalis">Apostolos Topalis</name><name sortKey="Topalis, Apostolos" sort="Topalis, Apostolos" uniqKey="Topalis A" first="Apostolos" last="Topalis">Apostolos Topalis</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 001513 | SxmlIndent | more
Ou
HfdSelect -h $EXPLOR_AREA/Data/Main/Exploration/biblio.hfd -nk 001513 | 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:A225C5D713C63CEA13AA61314046B4C09D2E7DEC
|texte= Learn++.MT: A New Approach to Incremental Learning
}}
| This area was generated with Dilib version V0.6.32. |  |