A connectionist architecture that adapts its representation to complex tasks
Identifieur interne :
000724 ( PascalFrancis/Corpus );
précédent :
000723;
suivant :
000725
A connectionist architecture that adapts its representation to complex tasks
Auteurs : Bruno ScherrerSource :
-
IEEE ... International Conference on Neural Networks [ 1098-7576 ] ; 2002.
RBID : Pascal:04-0132603
Descripteurs français
English descriptors
Abstract
This paper presents an original connectionist architecture that is capable of adapting its representation to one or various reinforcement problems. We briefly describe the generic reinforcement learning theory it is based on. We focus on distributed algorithms that enables efficient planning. In this specific framework, we define the notion of task-specialisation and propose a procedure for adapting a task model without increasing its complexity. It consists in a high-level learning of representation in problems with possibly delayed reinforcements. We show that such a single architecture can adapt to multiple tasks. Finally we stress its connectionist nature: most computations can be distributed and done in parallel. We illustrate and evaluate this adaptation paradigm on a navigation continuous-space environment.
Notice en format standard (ISO 2709)
Pour connaître la documentation sur le format Inist Standard.
| pA |
| A01 | 01 | 1 | | @0 1098-7576 |
|---|
| A08 | 01 | 1 | ENG | @1 A connectionist architecture that adapts its representation to complex tasks |
|---|
| A09 | 01 | 1 | ENG | @1 IJCNN'02 : international joint conference on neural networks : Honolulu HI, 12-17 May 2002 |
|---|
| A11 | 01 | 1 | | @1 SCHERRER (Bruno) |
|---|
| A14 | 01 | | | @1 CORTEX/MAIA Teams, LORIA, B.P. 239 @2 54506 Vandoeuvre-Les-Nancy @3 FRA @Z 1 aut. |
|---|
| A18 | 01 | 1 | | @1 IEEE. Neural Networks Society @3 USA @9 patr. |
|---|
| A18 | 02 | 1 | | @1 International Neural Network Society @3 USA @9 patr. |
|---|
| A20 | | | | @1 2929-2934 |
|---|
| A21 | | | | @1 2002 |
|---|
| A23 | 01 | | | @0 ENG |
|---|
| A26 | 01 | | | @0 0-7803-7278-6 |
|---|
| A43 | 01 | | | @1 INIST @2 Y 37961 @5 354000117750885180 |
|---|
| A44 | | | | @0 0000 @1 © 2004 INIST-CNRS. All rights reserved. |
|---|
| A45 | | | | @0 13 ref. |
|---|
| A47 | 01 | 1 | | @0 04-0132603 |
|---|
| A60 | | | | @1 P @2 C |
|---|
| A61 | | | | @0 A |
|---|
| A64 | 01 | 1 | | @0 IEEE ... International Conference on Neural Networks |
|---|
| A66 | 01 | | | @0 USA |
|---|
| C01 | 01 | | ENG | @0 This paper presents an original connectionist architecture that is capable of adapting its representation to one or various reinforcement problems. We briefly describe the generic reinforcement learning theory it is based on. We focus on distributed algorithms that enables efficient planning. In this specific framework, we define the notion of task-specialisation and propose a procedure for adapting a task model without increasing its complexity. It consists in a high-level learning of representation in problems with possibly delayed reinforcements. We show that such a single architecture can adapt to multiple tasks. Finally we stress its connectionist nature: most computations can be distributed and done in parallel. We illustrate and evaluate this adaptation paradigm on a navigation continuous-space environment. |
|---|
| C02 | 01 | X | | @0 001D02C06 |
|---|
| C03 | 01 | X | FRE | @0 Planification @5 01 |
|---|
| C03 | 01 | X | ENG | @0 Planning @5 01 |
|---|
| C03 | 01 | X | SPA | @0 Planificación @5 01 |
|---|
| C03 | 02 | X | FRE | @0 Apprentissage renforcé @5 02 |
|---|
| C03 | 02 | X | ENG | @0 Reinforcement learning @5 02 |
|---|
| C03 | 02 | X | SPA | @0 Aprendizaje reforzado @5 02 |
|---|
| C03 | 03 | X | FRE | @0 Algorithme réparti @5 03 |
|---|
| C03 | 03 | X | ENG | @0 Distributed algorithm @5 03 |
|---|
| C03 | 03 | X | SPA | @0 Algoritmo repartido @5 03 |
|---|
| C03 | 04 | X | FRE | @0 Calcul réparti @5 04 |
|---|
| C03 | 04 | X | ENG | @0 Distributed computing @5 04 |
|---|
| C03 | 04 | X | SPA | @0 Cálculo repartido @5 04 |
|---|
| C03 | 05 | X | FRE | @0 Connexionnisme @5 05 |
|---|
| C03 | 05 | X | ENG | @0 Connectionism @5 05 |
|---|
| C03 | 05 | X | SPA | @0 Conexionismo @5 05 |
|---|
| C03 | 06 | X | FRE | @0 Décision Markov @5 06 |
|---|
| C03 | 06 | X | ENG | @0 Markov decision @5 06 |
|---|
| C03 | 06 | X | SPA | @0 Decisión Markov @5 06 |
|---|
| C03 | 07 | X | FRE | @0 Architecture réseau @5 07 |
|---|
| C03 | 07 | X | ENG | @0 Network architecture @5 07 |
|---|
| C03 | 07 | X | SPA | @0 Arquitectura red @5 07 |
|---|
| N21 | | | | @1 082 |
|---|
| N82 | | | | @1 PSI |
|---|
|
|---|
| pR |
| A30 | 01 | 1 | ENG | @1 2002 International joint conference on neural networks @3 Honolulu HI USA @4 2002-05-12 |
|---|
|
|---|
Format Inist (serveur)
| NO : | PASCAL 04-0132603 INIST |
|---|
| ET : | A connectionist architecture that adapts its representation to complex tasks |
|---|
| AU : | SCHERRER (Bruno) |
|---|
| AF : | CORTEX/MAIA Teams, LORIA, B.P. 239 /54506 Vandoeuvre-Les-Nancy/France (1 aut.) |
|---|
| DT : | Publication en série; Congrès; Niveau analytique |
|---|
| SO : | IEEE ... International Conference on Neural Networks; ISSN 1098-7576; Etats-Unis; Da. 2002; Pp. 2929-2934; Bibl. 13 ref. |
|---|
| LA : | Anglais |
|---|
| EA : | This paper presents an original connectionist architecture that is capable of adapting its representation to one or various reinforcement problems. We briefly describe the generic reinforcement learning theory it is based on. We focus on distributed algorithms that enables efficient planning. In this specific framework, we define the notion of task-specialisation and propose a procedure for adapting a task model without increasing its complexity. It consists in a high-level learning of representation in problems with possibly delayed reinforcements. We show that such a single architecture can adapt to multiple tasks. Finally we stress its connectionist nature: most computations can be distributed and done in parallel. We illustrate and evaluate this adaptation paradigm on a navigation continuous-space environment. |
|---|
| CC : | 001D02C06 |
|---|
| FD : | Planification; Apprentissage renforcé; Algorithme réparti; Calcul réparti; Connexionnisme; Décision Markov; Architecture réseau |
|---|
| ED : | Planning; Reinforcement learning; Distributed algorithm; Distributed computing; Connectionism; Markov decision; Network architecture |
|---|
| SD : | Planificación; Aprendizaje reforzado; Algoritmo repartido; Cálculo repartido; Conexionismo; Decisión Markov; Arquitectura red |
|---|
| LO : | INIST-Y 37961.354000117750885180 |
|---|
| ID : | 04-0132603 |
|---|
Links to Exploration step
Pascal:04-0132603
Le document en format XML
<record><TEI><teiHeader><fileDesc><titleStmt><title xml:lang="en" level="a">A connectionist architecture that adapts its representation to complex tasks</title>
<author><name sortKey="Scherrer, Bruno" sort="Scherrer, Bruno" uniqKey="Scherrer B" first="Bruno" last="Scherrer">Bruno Scherrer</name>
<affiliation><inist:fA14 i1="01"><s1>CORTEX/MAIA Teams, LORIA, B.P. 239 </s1>
<s2>54506 Vandoeuvre-Les-Nancy</s2>
<s3>FRA</s3>
<sZ>1 aut.</sZ>
</inist:fA14><publicationStmt><idno type="wicri:source">INIST</idno>
<idno type="inist">04-0132603</idno>
<date when="2002">2002</date>
<idno type="stanalyst">PASCAL 04-0132603 INIST</idno>
<idno type="RBID">Pascal:04-0132603</idno>
<idno type="wicri:Area/PascalFrancis/Corpus">000724</idno>
</publicationStmt><sourceDesc><biblStruct><analytic><title xml:lang="en" level="a">A connectionist architecture that adapts its representation to complex tasks</title>
<author><name sortKey="Scherrer, Bruno" sort="Scherrer, Bruno" uniqKey="Scherrer B" first="Bruno" last="Scherrer">Bruno Scherrer</name>
<affiliation><inist:fA14 i1="01"><s1>CORTEX/MAIA Teams, LORIA, B.P. 239 </s1>
<s2>54506 Vandoeuvre-Les-Nancy</s2>
<s3>FRA</s3>
<sZ>1 aut.</sZ>
</inist:fA14><series><title level="j" type="main">IEEE ... International Conference on Neural Networks</title>
<idno type="ISSN">1098-7576</idno>
<imprint><date when="2002">2002</date>
</imprint><seriesStmt><title level="j" type="main">IEEE ... International Conference on Neural Networks</title>
<idno type="ISSN">1098-7576</idno>
</seriesStmt><profileDesc><textClass><keywords scheme="KwdEn" xml:lang="en"><term>Connectionism</term>
<term>Distributed algorithm</term>
<term>Distributed computing</term>
<term>Markov decision</term>
<term>Network architecture</term>
<term>Planning</term>
<term>Reinforcement learning</term>
</keywords><keywords scheme="Pascal" xml:lang="fr"><term>Planification</term>
<term>Apprentissage renforcé</term>
<term>Algorithme réparti</term>
<term>Calcul réparti</term>
<term>Connexionnisme</term>
<term>Décision Markov</term>
<term>Architecture réseau</term>
</keywords><front><div type="abstract" xml:lang="en">This paper presents an original connectionist architecture that is capable of adapting its representation to one or various reinforcement problems. We briefly describe the generic reinforcement learning theory it is based on. We focus on distributed algorithms that enables efficient planning. In this specific framework, we define the notion of task-specialisation and propose a procedure for adapting a task model without increasing its complexity. It consists in a high-level learning of representation in problems with possibly delayed reinforcements. We show that such a single architecture can adapt to multiple tasks. Finally we stress its connectionist nature: most computations can be distributed and done in parallel. We illustrate and evaluate this adaptation paradigm on a navigation continuous-space environment.</div>
</front><inist><standard h6="B"><pA><fA01 i1="01" i2="1"><s0>1098-7576</s0>
</fA01><fA08 i1="01" i2="1" l="ENG"><s1>A connectionist architecture that adapts its representation to complex tasks</s1>
</fA08><fA09 i1="01" i2="1" l="ENG"><s1>IJCNN'02 : international joint conference on neural networks : Honolulu HI, 12-17 May 2002</s1>
</fA09><fA11 i1="01" i2="1"><s1>SCHERRER (Bruno)</s1>
</fA11><fA14 i1="01"><s1>CORTEX/MAIA Teams, LORIA, B.P. 239 </s1>
<s2>54506 Vandoeuvre-Les-Nancy</s2>
<s3>FRA</s3>
<sZ>1 aut.</sZ>
</fA14><fA18 i1="01" i2="1"><s1>IEEE. Neural Networks Society</s1>
<s3>USA</s3>
<s9>patr.</s9>
</fA18><fA18 i1="02" i2="1"><s1>International Neural Network Society</s1>
<s3>USA</s3>
<s9>patr.</s9>
</fA18><fA20><s1>2929-2934</s1>
</fA20><fA21><s1>2002</s1>
</fA21><fA23 i1="01"><s0>ENG</s0>
</fA23><fA26 i1="01"><s0>0-7803-7278-6</s0>
</fA26><fA43 i1="01"><s1>INIST</s1>
<s2>Y 37961</s2>
<s5>354000117750885180</s5>
</fA43><fA44><s0>0000</s0>
<s1>© 2004 INIST-CNRS. All rights reserved.</s1>
</fA44><fA45><s0>13 ref.</s0>
</fA45><fA47 i1="01" i2="1"><s0>04-0132603</s0>
</fA47><fA60><s1>P</s1>
<s2>C</s2>
</fA60><fA64 i1="01" i2="1"><s0>IEEE ... International Conference on Neural Networks</s0>
</fA64><fA66 i1="01"><s0>USA</s0>
</fA66><fC01 i1="01" l="ENG"><s0>This paper presents an original connectionist architecture that is capable of adapting its representation to one or various reinforcement problems. We briefly describe the generic reinforcement learning theory it is based on. We focus on distributed algorithms that enables efficient planning. In this specific framework, we define the notion of task-specialisation and propose a procedure for adapting a task model without increasing its complexity. It consists in a high-level learning of representation in problems with possibly delayed reinforcements. We show that such a single architecture can adapt to multiple tasks. Finally we stress its connectionist nature: most computations can be distributed and done in parallel. We illustrate and evaluate this adaptation paradigm on a navigation continuous-space environment.</s0>
</fC01><fC02 i1="01" i2="X"><s0>001D02C06</s0>
</fC02><fC03 i1="01" i2="X" l="FRE"><s0>Planification</s0>
<s5>01</s5>
</fC03><fC03 i1="01" i2="X" l="ENG"><s0>Planning</s0>
<s5>01</s5>
</fC03><fC03 i1="01" i2="X" l="SPA"><s0>Planificación</s0>
<s5>01</s5>
</fC03><fC03 i1="02" i2="X" l="FRE"><s0>Apprentissage renforcé</s0>
<s5>02</s5>
</fC03><fC03 i1="02" i2="X" l="ENG"><s0>Reinforcement learning</s0>
<s5>02</s5>
</fC03><fC03 i1="02" i2="X" l="SPA"><s0>Aprendizaje reforzado</s0>
<s5>02</s5>
</fC03><fC03 i1="03" i2="X" l="FRE"><s0>Algorithme réparti</s0>
<s5>03</s5>
</fC03><fC03 i1="03" i2="X" l="ENG"><s0>Distributed algorithm</s0>
<s5>03</s5>
</fC03><fC03 i1="03" i2="X" l="SPA"><s0>Algoritmo repartido</s0>
<s5>03</s5>
</fC03><fC03 i1="04" i2="X" l="FRE"><s0>Calcul réparti</s0>
<s5>04</s5>
</fC03><fC03 i1="04" i2="X" l="ENG"><s0>Distributed computing</s0>
<s5>04</s5>
</fC03><fC03 i1="04" i2="X" l="SPA"><s0>Cálculo repartido</s0>
<s5>04</s5>
</fC03><fC03 i1="05" i2="X" l="FRE"><s0>Connexionnisme</s0>
<s5>05</s5>
</fC03><fC03 i1="05" i2="X" l="ENG"><s0>Connectionism</s0>
<s5>05</s5>
</fC03><fC03 i1="05" i2="X" l="SPA"><s0>Conexionismo</s0>
<s5>05</s5>
</fC03><fC03 i1="06" i2="X" l="FRE"><s0>Décision Markov</s0>
<s5>06</s5>
</fC03><fC03 i1="06" i2="X" l="ENG"><s0>Markov decision</s0>
<s5>06</s5>
</fC03><fC03 i1="06" i2="X" l="SPA"><s0>Decisión Markov</s0>
<s5>06</s5>
</fC03><fC03 i1="07" i2="X" l="FRE"><s0>Architecture réseau</s0>
<s5>07</s5>
</fC03><fC03 i1="07" i2="X" l="ENG"><s0>Network architecture</s0>
<s5>07</s5>
</fC03><fC03 i1="07" i2="X" l="SPA"><s0>Arquitectura red</s0>
<s5>07</s5>
</fC03><fN21><s1>082</s1>
</fN21><fN82><s1>PSI</s1>
</fN82><pR><fA30 i1="01" i2="1" l="ENG"><s1>2002 International joint conference on neural networks</s1>
<s3>Honolulu HI USA</s3>
<s4>2002-05-12</s4>
</fA30><server><NO>PASCAL 04-0132603 INIST</NO>
<ET>A connectionist architecture that adapts its representation to complex tasks</ET>
<AU>SCHERRER (Bruno)</AU>
<AF>CORTEX/MAIA Teams, LORIA, B.P. 239 /54506 Vandoeuvre-Les-Nancy/France (1 aut.)</AF>
<DT>Publication en série; Congrès; Niveau analytique</DT>
<SO>IEEE ... International Conference on Neural Networks; ISSN 1098-7576; Etats-Unis; Da. 2002; Pp. 2929-2934; Bibl. 13 ref.</SO>
<LA>Anglais</LA>
<EA>This paper presents an original connectionist architecture that is capable of adapting its representation to one or various reinforcement problems. We briefly describe the generic reinforcement learning theory it is based on. We focus on distributed algorithms that enables efficient planning. In this specific framework, we define the notion of task-specialisation and propose a procedure for adapting a task model without increasing its complexity. It consists in a high-level learning of representation in problems with possibly delayed reinforcements. We show that such a single architecture can adapt to multiple tasks. Finally we stress its connectionist nature: most computations can be distributed and done in parallel. We illustrate and evaluate this adaptation paradigm on a navigation continuous-space environment.</EA>
<CC>001D02C06</CC>
<FD>Planification; Apprentissage renforcé; Algorithme réparti; Calcul réparti; Connexionnisme; Décision Markov; Architecture réseau</FD>
<ED>Planning; Reinforcement learning; Distributed algorithm; Distributed computing; Connectionism; Markov decision; Network architecture</ED>
<SD>Planificación; Aprendizaje reforzado; Algoritmo repartido; Cálculo repartido; Conexionismo; Decisión Markov; Arquitectura red</SD>
<LO>INIST-Y 37961.354000117750885180</LO>
<ID>04-0132603</ID>
</server>
Pour manipuler ce document sous Unix (Dilib)
EXPLOR_STEP=$WICRI_ROOT/Wicri/Lorraine/explor/InforLorV4/Data/PascalFrancis/Corpus
HfdSelect -h $EXPLOR_STEP/biblio.hfd -nk 000724 | SxmlIndent | more
Ou
HfdSelect -h $EXPLOR_AREA/Data/PascalFrancis/Corpus/biblio.hfd -nk 000724 | SxmlIndent | more
Pour mettre un lien sur cette page dans le réseau Wicri
{{Explor lien
|wiki= Wicri/Lorraine
|area= InforLorV4
|flux= PascalFrancis
|étape= Corpus
|type= RBID
|clé= Pascal:04-0132603
|texte= A connectionist architecture that adapts its representation to complex tasks
}}
| This area was generated with Dilib version V0.6.33.
Data generation: Mon Jun 10 21:56:28 2019. Site generation: Fri Feb 25 15:29:27 2022 |  |
