schliessen

Filtern

 

Bibliotheken

Learning to Remember, Forget and Ignore using Attention Control in Memory

Typical neural networks with external memory do not effectively separate capacity for episodic and working memory as is required for reasoning in humans. Applying knowledge gained from psychological studies, we designed a new model called Differentiable Working Memory (DWM) in order to specifically... Full description

Journal Title: arXiv.org Sep 28, 2018
Main Author: Jayram, T
Other Authors: Mcavoy, Ryan , Kornuta, Tomasz , Asseman, Alexis , Rocki, Kamil , Ozcan, Ahmet
Format: Electronic Article Electronic Article
Language: English
Subjects:
Quelle: © ProQuest LLC All rights reserved
Zum Text:
SendSend as email Add to Book BagAdd to Book Bag
Staff View
recordid: proquest2114922149
title: Learning to Remember, Forget and Ignore using Attention Control in Memory
format: Article
creator:
  • Jayram, T
  • Mcavoy, Ryan
  • Kornuta, Tomasz
  • Asseman, Alexis
  • Rocki, Kamil
  • Ozcan, Ahmet
subjects:
  • Memory
  • Psychology
  • Neural Networks
  • Machine Learning
  • Neural and Evolutionary Computation
  • Machine Learning
ispartof: arXiv.org, Sep 28, 2018
description: Typical neural networks with external memory do not effectively separate capacity for episodic and working memory as is required for reasoning in humans. Applying knowledge gained from psychological studies, we designed a new model called Differentiable Working Memory (DWM) in order to specifically emulate human working memory. As it shows the same functional characteristics as working memory, it robustly learns psychology inspired tasks and converges faster than comparable state-of-the-art models. Moreover, the DWM model successfully generalizes to sequences two orders of magnitude longer than the ones used in training. Our in-depth analysis shows that the behavior of DWM is interpretable and that it learns to have fine control over memory, allowing it to retain, ignore or forget information based on its relevance.
language: eng
source: © ProQuest LLC All rights reserved
identifier:
fulltext: fulltext_linktorsrc
url: Link


@attributes
ID1085563267
RANK0.07
NO1
SEARCH_ENGINEprimo_central_multiple_fe
SEARCH_ENGINE_TYPEPrimo Central Search Engine
LOCALfalse
PrimoNMBib
record
control
sourcerecordid2114922149
sourceidproquest
recordidTN_proquest2114922149
sourcesystemPC
pqid2114922149
display
typearticle
titleLearning to Remember, Forget and Ignore using Attention Control in Memory
creatorJayram, T ; Mcavoy, Ryan ; Kornuta, Tomasz ; Asseman, Alexis ; Rocki, Kamil ; Ozcan, Ahmet
contributorOzcan, Ahmet (pacrepositoryorg)
ispartofarXiv.org, Sep 28, 2018
subjectMemory ; Psychology ; Neural Networks ; Machine Learning ; Neural and Evolutionary Computation ; Machine Learning
descriptionTypical neural networks with external memory do not effectively separate capacity for episodic and working memory as is required for reasoning in humans. Applying knowledge gained from psychological studies, we designed a new model called Differentiable Working Memory (DWM) in order to specifically emulate human working memory. As it shows the same functional characteristics as working memory, it robustly learns psychology inspired tasks and converges faster than comparable state-of-the-art models. Moreover, the DWM model successfully generalizes to sequences two orders of magnitude longer than the ones used in training. Our in-depth analysis shows that the behavior of DWM is interpretable and that it learns to have fine control over memory, allowing it to retain, ignore or forget information based on its relevance.
languageeng
source
0© ProQuest LLC All rights reserved
1Engineering Database
2Publicly Available Content Database
3ProQuest Engineering Collection
4ProQuest Technology Collection
5ProQuest SciTech Collection
6Materials Science & Engineering Database
7ProQuest Central (new)
8ProQuest Central Korea
9SciTech Premium Collection
10Technology Collection
11ProQuest Central Essentials
12ProQuest One Academic
13Engineering Collection (ProQuest)
oafree_for_read
links
openurl$$Topenurl_article
openurlfulltext$$Topenurlfull_article
linktorsrc$$Uhttp://search.proquest.com/docview/2114922149/?pq-origsite=primo$$EView_record_in_ProQuest_(subscribers_only)
search
creatorcontrib
0Jayram, T
1Mcavoy, Ryan
2Kornuta, Tomasz
3Asseman, Alexis
4Rocki, Kamil
5Ozcan, Ahmet
titleLearning to Remember, Forget and Ignore using Attention Control in Memory
descriptionTypical neural networks with external memory do not effectively separate capacity for episodic and working memory as is required for reasoning in humans. Applying knowledge gained from psychological studies, we designed a new model called Differentiable Working Memory (DWM) in order to specifically emulate human working memory. As it shows the same functional characteristics as working memory, it robustly learns psychology inspired tasks and converges faster than comparable state-of-the-art models. Moreover, the DWM model successfully generalizes to sequences two orders of magnitude longer than the ones used in training. Our in-depth analysis shows that the behavior of DWM is interpretable and that it learns to have fine control over memory, allowing it to retain, ignore or forget information based on its relevance.
subject
0Memory
1Psychology
2Neural Networks
3Machine Learning
4Neural and Evolutionary Computation
general
0English
1Cornell University Library, arXiv.org
2Engineering Database
3Publicly Available Content Database
4ProQuest Engineering Collection
5ProQuest Technology Collection
6ProQuest SciTech Collection
7Materials Science & Engineering Database
8ProQuest Central (new)
9ProQuest Central Korea
10SciTech Premium Collection
11Technology Collection
12ProQuest Central Essentials
13ProQuest One Academic
14Engineering Collection (ProQuest)
sourceidproquest
recordidproquest2114922149
rsrctypearticle
creationdate2018
addtitlearXiv.org
searchscope
01007853
11008875
21008886
31009127
41009240
510000041
610000053
710000120
810000250
910000255
1010000256
1110000258
1210000260
1310000265
1410000268
1510000281
1610000348
1710000356
1810000360
19proquest
scope
01007853
11008875
21008886
31009127
41009240
510000041
610000053
710000120
810000250
910000255
1010000256
1110000258
1210000260
1310000265
1410000268
1510000281
1610000348
1710000356
1810000360
19proquest
lsr43
01007853true
11008875true
21008886true
31009127true
41009240true
510000041true
610000053true
710000120true
810000250true
910000255true
1010000256true
1110000258true
1210000260true
1310000265true
1410000268true
1510000281true
1610000348true
1710000356true
1810000360true
startdate20180928
enddate20180928
sort
titleLearning to Remember, Forget and Ignore using Attention Control in Memory
authorJayram, T ; Mcavoy, Ryan ; Kornuta, Tomasz ; Asseman, Alexis ; Rocki, Kamil ; Ozcan, Ahmet
creationdate20180928
lso0120180928
facets
frbrgroupid7124747721321072052
frbrtype6
newrecords20181004
languageeng
creationdate2018
topic
0Memory
1Psychology
2Neural Networks
3Machine Learning
4Neural and Evolutionary Computation
collection
0Engineering Database
1Publicly Available Content Database
2ProQuest Engineering Collection
3ProQuest Technology Collection
4ProQuest SciTech Collection
5Materials Science & Engineering Database
6ProQuest Central (new)
7ProQuest Central Korea
8SciTech Premium Collection
9Technology Collection
10ProQuest Central Essentials
11ProQuest One Academic
12Engineering Collection (ProQuest)
prefilterarticles
rsrctypearticles
creatorcontrib
0Jayram, T
1Mcavoy, Ryan
2Kornuta, Tomasz
3Asseman, Alexis
4Rocki, Kamil
5Ozcan, Ahmet
jtitlearXiv.org
delivery
delcategoryRemote Search Resource
fulltextfulltext_linktorsrc
addata
aulast
0Jayram
1McAvoy
2Kornuta
3Asseman
4Rocki
5Ozcan
aufirst
0T
1Ryan
2Tomasz
3Alexis
4Kamil
5Ahmet
au
0Jayram, T
1Mcavoy, Ryan
2Kornuta, Tomasz
3Asseman, Alexis
4Rocki, Kamil
5Ozcan, Ahmet
atitleLearning to Remember, Forget and Ignore using Attention Control in Memory
jtitlearXiv.org
risdate20180928
formatjournal
genrearticle
ristypeJOUR
abstractTypical neural networks with external memory do not effectively separate capacity for episodic and working memory as is required for reasoning in humans. Applying knowledge gained from psychological studies, we designed a new model called Differentiable Working Memory (DWM) in order to specifically emulate human working memory. As it shows the same functional characteristics as working memory, it robustly learns psychology inspired tasks and converges faster than comparable state-of-the-art models. Moreover, the DWM model successfully generalizes to sequences two orders of magnitude longer than the ones used in training. Our in-depth analysis shows that the behavior of DWM is interpretable and that it learns to have fine control over memory, allowing it to retain, ignore or forget information based on its relevance.
copIthaca
pubCornell University Library, arXiv.org
urlhttp://search.proquest.com/docview/2114922149/
oafree_for_read
date2018-09-28