Structured Sparsification of Gated Recurrent Neural Networks

Recently, a lot of techniques were developed to sparsify the weights of neural networks and to remove networks' structure units, e.g. neurons. We adjust the existing sparsification approaches to the gated recurrent architectures. Specifically, in addition to the sparsification of weights and ne...

Full description

Saved in:
Bibliographic Details
Published in:arXiv.org 2019-11
Main Authors: Lobacheva, Ekaterina, Chirkova, Nadezhda, Markovich, Alexander, Vetrov, Dmitry
Format: Article
Language:English
Subjects:
Neural networks
Neurons
Recurrent neural networks
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
cited_by
cites
container_end_page
container_issue
container_start_page
container_title arXiv.org
container_volume
creator Lobacheva, Ekaterina
Chirkova, Nadezhda
Markovich, Alexander
Vetrov, Dmitry
description Recently, a lot of techniques were developed to sparsify the weights of neural networks and to remove networks' structure units, e.g. neurons. We adjust the existing sparsification approaches to the gated recurrent architectures. Specifically, in addition to the sparsification of weights and neurons, we propose sparsifying the preactivations of gates. This makes some gates constant and simplifies LSTM structure. We test our approach on the text classification and language modeling tasks. We observe that the resulting structure of gate sparsity depends on the task and connect the learned structure to the specifics of the particular tasks. Our method also improves neuron-wise compression of the model in most of the tasks.
format article
fullrecord <record><control><sourceid>proquest</sourceid><recordid>TN_cdi_proquest_journals_2314447128</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>2314447128</sourcerecordid><originalsourceid>FETCH-proquest_journals_23144471283</originalsourceid><addsrcrecordid>eNqNissKwjAQAIMgWLT_EPBcaDep7cGb-Dh5sN5LSBNoLU3d7OLv24Mf4GlgZlYiAaWKrNYAG5HGOOR5DocKylIl4tgQsiVG18lmNhh731tDfZhk8PJqaPEPZxnRTSTvjtGMC-gT8BV3Yu3NGF3641bsL-fn6ZbNGN7sIrVDYJyW1IIqtNZVAbX67_oCdZE39Q</addsrcrecordid><sourcetype>Aggregation Database</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype><pqid>2314447128</pqid></control><display><type>article</type><title>Structured Sparsification of Gated Recurrent Neural Networks</title><source>Publicly Available Content Database</source><creator>Lobacheva, Ekaterina ; Chirkova, Nadezhda ; Markovich, Alexander ; Vetrov, Dmitry</creator><creatorcontrib>Lobacheva, Ekaterina ; Chirkova, Nadezhda ; Markovich, Alexander ; Vetrov, Dmitry</creatorcontrib><description>Recently, a lot of techniques were developed to sparsify the weights of neural networks and to remove networks' structure units, e.g. neurons. We adjust the existing sparsification approaches to the gated recurrent architectures. Specifically, in addition to the sparsification of weights and neurons, we propose sparsifying the preactivations of gates. This makes some gates constant and simplifies LSTM structure. We test our approach on the text classification and language modeling tasks. We observe that the resulting structure of gate sparsity depends on the task and connect the learned structure to the specifics of the particular tasks. Our method also improves neuron-wise compression of the model in most of the tasks.</description><identifier>EISSN: 2331-8422</identifier><language>eng</language><publisher>Ithaca: Cornell University Library, arXiv.org</publisher><subject>Neural networks ; Neurons ; Recurrent neural networks</subject><ispartof>arXiv.org, 2019-11</ispartof><rights>2019. This work is published under http://arxiv.org/licenses/nonexclusive-distrib/1.0/ (the “License”). Notwithstanding the ProQuest Terms and Conditions, you may use this content in accordance with the terms of the License.</rights><oa>free_for_read</oa><woscitedreferencessubscribed>false</woscitedreferencessubscribed></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><linktohtml>$$Uhttps://www.proquest.com/docview/2314447128?pq-origsite=primo$$EHTML$$P50$$Gproquest$$Hfree_for_read</linktohtml><link.rule.ids>776,780,25732,36991,44569</link.rule.ids></links><search><creatorcontrib>Lobacheva, Ekaterina</creatorcontrib><creatorcontrib>Chirkova, Nadezhda</creatorcontrib><creatorcontrib>Markovich, Alexander</creatorcontrib><creatorcontrib>Vetrov, Dmitry</creatorcontrib><title>Structured Sparsification of Gated Recurrent Neural Networks</title><title>arXiv.org</title><description>Recently, a lot of techniques were developed to sparsify the weights of neural networks and to remove networks' structure units, e.g. neurons. We adjust the existing sparsification approaches to the gated recurrent architectures. Specifically, in addition to the sparsification of weights and neurons, we propose sparsifying the preactivations of gates. This makes some gates constant and simplifies LSTM structure. We test our approach on the text classification and language modeling tasks. We observe that the resulting structure of gate sparsity depends on the task and connect the learned structure to the specifics of the particular tasks. Our method also improves neuron-wise compression of the model in most of the tasks.</description><subject>Neural networks</subject><subject>Neurons</subject><subject>Recurrent neural networks</subject><issn>2331-8422</issn><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2019</creationdate><recordtype>article</recordtype><sourceid>PIMPY</sourceid><recordid>eNqNissKwjAQAIMgWLT_EPBcaDep7cGb-Dh5sN5LSBNoLU3d7OLv24Mf4GlgZlYiAaWKrNYAG5HGOOR5DocKylIl4tgQsiVG18lmNhh731tDfZhk8PJqaPEPZxnRTSTvjtGMC-gT8BV3Yu3NGF3641bsL-fn6ZbNGN7sIrVDYJyW1IIqtNZVAbX67_oCdZE39Q</recordid><startdate>20191113</startdate><enddate>20191113</enddate><creator>Lobacheva, Ekaterina</creator><creator>Chirkova, Nadezhda</creator><creator>Markovich, Alexander</creator><creator>Vetrov, Dmitry</creator><general>Cornell University Library, arXiv.org</general><scope>8FE</scope><scope>8FG</scope><scope>ABJCF</scope><scope>ABUWG</scope><scope>AFKRA</scope><scope>AZQEC</scope><scope>BENPR</scope><scope>BGLVJ</scope><scope>CCPQU</scope><scope>DWQXO</scope><scope>HCIFZ</scope><scope>L6V</scope><scope>M7S</scope><scope>PIMPY</scope><scope>PQEST</scope><scope>PQQKQ</scope><scope>PQUKI</scope><scope>PRINS</scope><scope>PTHSS</scope></search><sort><creationdate>20191113</creationdate><title>Structured Sparsification of Gated Recurrent Neural Networks</title><author>Lobacheva, Ekaterina ; Chirkova, Nadezhda ; Markovich, Alexander ; Vetrov, Dmitry</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-proquest_journals_23144471283</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2019</creationdate><topic>Neural networks</topic><topic>Neurons</topic><topic>Recurrent neural networks</topic><toplevel>online_resources</toplevel><creatorcontrib>Lobacheva, Ekaterina</creatorcontrib><creatorcontrib>Chirkova, Nadezhda</creatorcontrib><creatorcontrib>Markovich, Alexander</creatorcontrib><creatorcontrib>Vetrov, Dmitry</creatorcontrib><collection>ProQuest SciTech Collection</collection><collection>ProQuest Technology Collection</collection><collection>Materials Science &amp; Engineering Collection</collection><collection>ProQuest Central (Alumni)</collection><collection>ProQuest Central UK/Ireland</collection><collection>ProQuest Central Essentials</collection><collection>ProQuest Central</collection><collection>Technology Collection</collection><collection>ProQuest One Community College</collection><collection>ProQuest Central Korea</collection><collection>SciTech Premium Collection (Proquest) (PQ_SDU_P3)</collection><collection>ProQuest Engineering Collection</collection><collection>Engineering Database</collection><collection>Publicly Available Content Database</collection><collection>ProQuest One Academic Eastern Edition (DO NOT USE)</collection><collection>ProQuest One Academic</collection><collection>ProQuest One Academic UKI Edition</collection><collection>ProQuest Central China</collection><collection>Engineering collection</collection></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext</fulltext></delivery><addata><au>Lobacheva, Ekaterina</au><au>Chirkova, Nadezhda</au><au>Markovich, Alexander</au><au>Vetrov, Dmitry</au><format>book</format><genre>document</genre><ristype>GEN</ristype><atitle>Structured Sparsification of Gated Recurrent Neural Networks</atitle><jtitle>arXiv.org</jtitle><date>2019-11-13</date><risdate>2019</risdate><eissn>2331-8422</eissn><abstract>Recently, a lot of techniques were developed to sparsify the weights of neural networks and to remove networks' structure units, e.g. neurons. We adjust the existing sparsification approaches to the gated recurrent architectures. Specifically, in addition to the sparsification of weights and neurons, we propose sparsifying the preactivations of gates. This makes some gates constant and simplifies LSTM structure. We test our approach on the text classification and language modeling tasks. We observe that the resulting structure of gate sparsity depends on the task and connect the learned structure to the specifics of the particular tasks. Our method also improves neuron-wise compression of the model in most of the tasks.</abstract><cop>Ithaca</cop><pub>Cornell University Library, arXiv.org</pub><oa>free_for_read</oa></addata></record>
fulltext fulltext
identifier EISSN: 2331-8422
ispartof arXiv.org, 2019-11
issn 2331-8422
language eng
recordid cdi_proquest_journals_2314447128
source Publicly Available Content Database
subjects Neural networks
Neurons
Recurrent neural networks
title Structured Sparsification of Gated Recurrent Neural Networks
url http://sfxeu10.hosted.exlibrisgroup.com/loughborough?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-01-22T07%3A37%3A53IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-proquest&rft_val_fmt=info:ofi/fmt:kev:mtx:book&rft.genre=document&rft.atitle=Structured%20Sparsification%20of%20Gated%20Recurrent%20Neural%20Networks&rft.jtitle=arXiv.org&rft.au=Lobacheva,%20Ekaterina&rft.date=2019-11-13&rft.eissn=2331-8422&rft_id=info:doi/&rft_dat=%3Cproquest%3E2314447128%3C/proquest%3E%3Cgrp_id%3Ecdi_FETCH-proquest_journals_23144471283%3C/grp_id%3E%3Coa%3E%3C/oa%3E%3Curl%3E%3C/url%3E&rft_id=info:oai/&rft_pqid=2314447128&rft_id=info:pmid/&rfr_iscdi=true