Assessing the impact of minor modifications on the interior structure of GRU: GRU1 and GRU2
In this study, two GRU variants named GRU1 and GRU2 are proposed by employing simple changes to the internal structure of the standard GRU, which is one of the popular RNN variants. Comparative experiments are conducted on four problems: language modeling, question answering, addition task, and sent...
Gespeichert in:
| Veröffentlicht in: | Concurrency and computation 2022-09, Vol.34 (20), p.n/a |
|---|---|
| Hauptverfasser: | , |
| Format: | Artikel |
| Sprache: | eng |
| Schlagworte: | |
| Online-Zugang: | Volltext |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| container_end_page | n/a |
|---|---|
| container_issue | 20 |
| container_start_page | |
| container_title | Concurrency and computation |
| container_volume | 34 |
| creator | Yigit, Gulsum Amasyali, Mehmet Fatih |
| description | In this study, two GRU variants named GRU1 and GRU2 are proposed by employing simple changes to the internal structure of the standard GRU, which is one of the popular RNN variants. Comparative experiments are conducted on four problems: language modeling, question answering, addition task, and sentiment analysis. Moreover, in the addition task, curriculum learning and anti‐curriculum learning strategies, which extend the training data having examples from easy to hard or from hard to easy, are comparatively evaluated. Accordingly, the GRU1 and GRU2 variants outperformed the standard GRU. In addition, the curriculum learning approach, in which the training data is expanded from easy to difficult, improves the performance considerably. |
| doi_str_mv | 10.1002/cpe.6775 |
| format | Article |
| fullrecord | <record><control><sourceid>proquest_cross</sourceid><recordid>TN_cdi_proquest_journals_2699544343</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>2699544343</sourcerecordid><originalsourceid>FETCH-LOGICAL-c2935-67168d132207c8efeec59ad7480f50d54bc3787e4c1604144da123382898ca613</originalsourceid><addsrcrecordid>eNp1kMFKxDAQhoMouK6Cj1Dw4qVrJkmT1NtS1lVYUMQ9eQgxTTXLtqlJi-zb21rx5mXmh_lmBj6ELgEvAGNyY1q74EJkR2gGGSUp5pQd_2XCT9FZjDuMATCFGXpdxmhjdM170n3YxNWtNl3iq6R2jQ9J7UtXOaM755uY-GaCms4GN0xjF3rT9cGOC-vn7e1YINFNOQZyjk4qvY_24rfP0fZu9VLcp5vH9UOx3KSG5DRLuQAuS6CEYGGkraw1Wa5LwSSuMlxm7M1QIYVlBjhmwFipgVAqicyl0RzoHF1Nd9vgP3sbO7XzfWiGl4rwPM8Yo4wO1PVEmeBjDLZSbXC1DgcFWI3q1KBOjeoGNJ3QL7e3h385VTytfvhvMLJtFA</addsrcrecordid><sourcetype>Aggregation Database</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype><pqid>2699544343</pqid></control><display><type>article</type><title>Assessing the impact of minor modifications on the interior structure of GRU: GRU1 and GRU2</title><source>Wiley-Blackwell Journals</source><creator>Yigit, Gulsum ; Amasyali, Mehmet Fatih</creator><creatorcontrib>Yigit, Gulsum ; Amasyali, Mehmet Fatih</creatorcontrib><description>In this study, two GRU variants named GRU1 and GRU2 are proposed by employing simple changes to the internal structure of the standard GRU, which is one of the popular RNN variants. Comparative experiments are conducted on four problems: language modeling, question answering, addition task, and sentiment analysis. Moreover, in the addition task, curriculum learning and anti‐curriculum learning strategies, which extend the training data having examples from easy to hard or from hard to easy, are comparatively evaluated. Accordingly, the GRU1 and GRU2 variants outperformed the standard GRU. In addition, the curriculum learning approach, in which the training data is expanded from easy to difficult, improves the performance considerably.</description><identifier>ISSN: 1532-0626</identifier><identifier>EISSN: 1532-0634</identifier><identifier>DOI: 10.1002/cpe.6775</identifier><language>eng</language><publisher>Hoboken: Wiley Subscription Services, Inc</publisher><subject>Curricula ; curriculum learning ; Data mining ; gated recurrent units ; Learning ; recurrent neural networks ; Seq2seq ; short‐term dependency ; Training</subject><ispartof>Concurrency and computation, 2022-09, Vol.34 (20), p.n/a</ispartof><rights>2021 John Wiley & Sons Ltd.</rights><rights>2022 John Wiley & Sons, Ltd.</rights><lds50>peer_reviewed</lds50><woscitedreferencessubscribed>false</woscitedreferencessubscribed><citedby>FETCH-LOGICAL-c2935-67168d132207c8efeec59ad7480f50d54bc3787e4c1604144da123382898ca613</citedby><cites>FETCH-LOGICAL-c2935-67168d132207c8efeec59ad7480f50d54bc3787e4c1604144da123382898ca613</cites><orcidid>0000-0001-7010-169X</orcidid></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><linktopdf>$$Uhttps://onlinelibrary.wiley.com/doi/pdf/10.1002%2Fcpe.6775$$EPDF$$P50$$Gwiley$$H</linktopdf><linktohtml>$$Uhttps://onlinelibrary.wiley.com/doi/full/10.1002%2Fcpe.6775$$EHTML$$P50$$Gwiley$$H</linktohtml><link.rule.ids>314,776,780,1411,27901,27902,45550,45551</link.rule.ids></links><search><creatorcontrib>Yigit, Gulsum</creatorcontrib><creatorcontrib>Amasyali, Mehmet Fatih</creatorcontrib><title>Assessing the impact of minor modifications on the interior structure of GRU: GRU1 and GRU2</title><title>Concurrency and computation</title><description>In this study, two GRU variants named GRU1 and GRU2 are proposed by employing simple changes to the internal structure of the standard GRU, which is one of the popular RNN variants. Comparative experiments are conducted on four problems: language modeling, question answering, addition task, and sentiment analysis. Moreover, in the addition task, curriculum learning and anti‐curriculum learning strategies, which extend the training data having examples from easy to hard or from hard to easy, are comparatively evaluated. Accordingly, the GRU1 and GRU2 variants outperformed the standard GRU. In addition, the curriculum learning approach, in which the training data is expanded from easy to difficult, improves the performance considerably.</description><subject>Curricula</subject><subject>curriculum learning</subject><subject>Data mining</subject><subject>gated recurrent units</subject><subject>Learning</subject><subject>recurrent neural networks</subject><subject>Seq2seq</subject><subject>short‐term dependency</subject><subject>Training</subject><issn>1532-0626</issn><issn>1532-0634</issn><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2022</creationdate><recordtype>article</recordtype><recordid>eNp1kMFKxDAQhoMouK6Cj1Dw4qVrJkmT1NtS1lVYUMQ9eQgxTTXLtqlJi-zb21rx5mXmh_lmBj6ELgEvAGNyY1q74EJkR2gGGSUp5pQd_2XCT9FZjDuMATCFGXpdxmhjdM170n3YxNWtNl3iq6R2jQ9J7UtXOaM755uY-GaCms4GN0xjF3rT9cGOC-vn7e1YINFNOQZyjk4qvY_24rfP0fZu9VLcp5vH9UOx3KSG5DRLuQAuS6CEYGGkraw1Wa5LwSSuMlxm7M1QIYVlBjhmwFipgVAqicyl0RzoHF1Nd9vgP3sbO7XzfWiGl4rwPM8Yo4wO1PVEmeBjDLZSbXC1DgcFWI3q1KBOjeoGNJ3QL7e3h385VTytfvhvMLJtFA</recordid><startdate>20220910</startdate><enddate>20220910</enddate><creator>Yigit, Gulsum</creator><creator>Amasyali, Mehmet Fatih</creator><general>Wiley Subscription Services, Inc</general><scope>AAYXX</scope><scope>CITATION</scope><scope>7SC</scope><scope>8FD</scope><scope>JQ2</scope><scope>L7M</scope><scope>L~C</scope><scope>L~D</scope><orcidid>https://orcid.org/0000-0001-7010-169X</orcidid></search><sort><creationdate>20220910</creationdate><title>Assessing the impact of minor modifications on the interior structure of GRU: GRU1 and GRU2</title><author>Yigit, Gulsum ; Amasyali, Mehmet Fatih</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-c2935-67168d132207c8efeec59ad7480f50d54bc3787e4c1604144da123382898ca613</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2022</creationdate><topic>Curricula</topic><topic>curriculum learning</topic><topic>Data mining</topic><topic>gated recurrent units</topic><topic>Learning</topic><topic>recurrent neural networks</topic><topic>Seq2seq</topic><topic>short‐term dependency</topic><topic>Training</topic><toplevel>peer_reviewed</toplevel><toplevel>online_resources</toplevel><creatorcontrib>Yigit, Gulsum</creatorcontrib><creatorcontrib>Amasyali, Mehmet Fatih</creatorcontrib><collection>CrossRef</collection><collection>Computer and Information Systems Abstracts</collection><collection>Technology Research Database</collection><collection>ProQuest Computer Science Collection</collection><collection>Advanced Technologies Database with Aerospace</collection><collection>Computer and Information Systems Abstracts Academic</collection><collection>Computer and Information Systems Abstracts Professional</collection><jtitle>Concurrency and computation</jtitle></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext</fulltext></delivery><addata><au>Yigit, Gulsum</au><au>Amasyali, Mehmet Fatih</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>Assessing the impact of minor modifications on the interior structure of GRU: GRU1 and GRU2</atitle><jtitle>Concurrency and computation</jtitle><date>2022-09-10</date><risdate>2022</risdate><volume>34</volume><issue>20</issue><epage>n/a</epage><issn>1532-0626</issn><eissn>1532-0634</eissn><abstract>In this study, two GRU variants named GRU1 and GRU2 are proposed by employing simple changes to the internal structure of the standard GRU, which is one of the popular RNN variants. Comparative experiments are conducted on four problems: language modeling, question answering, addition task, and sentiment analysis. Moreover, in the addition task, curriculum learning and anti‐curriculum learning strategies, which extend the training data having examples from easy to hard or from hard to easy, are comparatively evaluated. Accordingly, the GRU1 and GRU2 variants outperformed the standard GRU. In addition, the curriculum learning approach, in which the training data is expanded from easy to difficult, improves the performance considerably.</abstract><cop>Hoboken</cop><pub>Wiley Subscription Services, Inc</pub><doi>10.1002/cpe.6775</doi><tpages>12</tpages><orcidid>https://orcid.org/0000-0001-7010-169X</orcidid></addata></record> |
| fulltext | fulltext |
| identifier | ISSN: 1532-0626 |
| ispartof | Concurrency and computation, 2022-09, Vol.34 (20), p.n/a |
| issn | 1532-0626 1532-0634 |
| language | eng |
| recordid | cdi_proquest_journals_2699544343 |
| source | Wiley-Blackwell Journals |
| subjects | Curricula curriculum learning Data mining gated recurrent units Learning recurrent neural networks Seq2seq short‐term dependency Training |
| title | Assessing the impact of minor modifications on the interior structure of GRU: GRU1 and GRU2 |
| url | https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-02-04T14%3A15%3A54IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-proquest_cross&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=Assessing%20the%20impact%20of%20minor%20modifications%20on%20the%20interior%20structure%20of%20GRU:%20GRU1%20and%20GRU2&rft.jtitle=Concurrency%20and%20computation&rft.au=Yigit,%20Gulsum&rft.date=2022-09-10&rft.volume=34&rft.issue=20&rft.epage=n/a&rft.issn=1532-0626&rft.eissn=1532-0634&rft_id=info:doi/10.1002/cpe.6775&rft_dat=%3Cproquest_cross%3E2699544343%3C/proquest_cross%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_pqid=2699544343&rft_id=info:pmid/&rfr_iscdi=true |