{"id":"https://openalex.org/W2071310251","doi":"https://doi.org/10.1109/tasl.2013.2284378","title":"Optimization Techniques to Improve Training Speed of Deep Neural Networks for Large Speech Tasks","display_name":"Optimization Techniques to Improve Training Speed of Deep Neural Networks for Large Speech Tasks","publication_year":2013,"publication_date":"2013-10-03","ids":{"openalex":"https://openalex.org/W2071310251","doi":"https://doi.org/10.1109/tasl.2013.2284378","mag":"2071310251"},"language":"en","primary_location":{"id":"doi:10.1109/tasl.2013.2284378","is_oa":false,"landing_page_url":"https://doi.org/10.1109/tasl.2013.2284378","pdf_url":null,"source":{"id":"https://openalex.org/S199497470","display_name":"IEEE Transactions on Audio Speech and Language Processing","issn_l":"1558-7916","issn":["1558-7916","1558-7924"],"is_oa":false,"is_in_doaj":false,"is_core":true,"host_organization":"https://openalex.org/P4310319808","host_organization_name":"Institute of Electrical and Electronics Engineers","host_organization_lineage":["https://openalex.org/P4310319808"],"host_organization_lineage_names":["Institute of Electrical and Electronics Engineers"],"type":"journal"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"IEEE Transactions on Audio, Speech, and Language Processing","raw_type":"journal-article"},"type":"article","indexed_in":["crossref"],"open_access":{"is_oa":false,"oa_status":"closed","oa_url":null,"any_repository_has_fulltext":false},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5070513394","display_name":"Tara N. Sainath","orcid":"https://orcid.org/0000-0002-4126-6556"},"institutions":[{"id":"https://openalex.org/I1341412227","display_name":"IBM (United States)","ror":"https://ror.org/05hh8d621","country_code":"US","type":"company","lineage":["https://openalex.org/I1341412227"]},{"id":"https://openalex.org/I4210114115","display_name":"IBM Research - Thomas J. Watson Research Center","ror":"https://ror.org/0265w5591","country_code":"US","type":"facility","lineage":["https://openalex.org/I1341412227","https://openalex.org/I4210114115"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Tara N. Sainath","raw_affiliation_strings":["IBM Thomas J. Watson Research Center, NY, USA","IBM T. J. Watson Research Center, Yorktown Heights , NY, USA#TAB#"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"IBM Thomas J. Watson Research Center, NY, USA","institution_ids":["https://openalex.org/I4210114115"]},{"raw_affiliation_string":"IBM T. J. Watson Research Center, Yorktown Heights , NY, USA#TAB#","institution_ids":["https://openalex.org/I1341412227"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5003725957","display_name":"Brian Kingsbury","orcid":"https://orcid.org/0000-0002-1343-6837"},"institutions":[{"id":"https://openalex.org/I1341412227","display_name":"IBM (United States)","ror":"https://ror.org/05hh8d621","country_code":"US","type":"company","lineage":["https://openalex.org/I1341412227"]},{"id":"https://openalex.org/I4210114115","display_name":"IBM Research - Thomas J. Watson Research Center","ror":"https://ror.org/0265w5591","country_code":"US","type":"facility","lineage":["https://openalex.org/I1341412227","https://openalex.org/I4210114115"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Brian Kingsbury","raw_affiliation_strings":["IBM Thomas J. Watson Research Center, NY, USA","IBM T. J. Watson Research Center, Yorktown Heights , NY, USA#TAB#"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"IBM Thomas J. Watson Research Center, NY, USA","institution_ids":["https://openalex.org/I4210114115"]},{"raw_affiliation_string":"IBM T. J. Watson Research Center, Yorktown Heights , NY, USA#TAB#","institution_ids":["https://openalex.org/I1341412227"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5010878109","display_name":"Hagen Soltau","orcid":null},"institutions":[{"id":"https://openalex.org/I1341412227","display_name":"IBM (United States)","ror":"https://ror.org/05hh8d621","country_code":"US","type":"company","lineage":["https://openalex.org/I1341412227"]},{"id":"https://openalex.org/I4210114115","display_name":"IBM Research - Thomas J. Watson Research Center","ror":"https://ror.org/0265w5591","country_code":"US","type":"facility","lineage":["https://openalex.org/I1341412227","https://openalex.org/I4210114115"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Hagen Soltau","raw_affiliation_strings":["IBM Thomas J. Watson Research Center, NY, USA","IBM T. J. Watson Research Center, Yorktown Heights , NY, USA#TAB#"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"IBM Thomas J. Watson Research Center, NY, USA","institution_ids":["https://openalex.org/I4210114115"]},{"raw_affiliation_string":"IBM T. J. Watson Research Center, Yorktown Heights , NY, USA#TAB#","institution_ids":["https://openalex.org/I1341412227"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5071715737","display_name":"Bhuvana Ramabhadran","orcid":null},"institutions":[{"id":"https://openalex.org/I1341412227","display_name":"IBM (United States)","ror":"https://ror.org/05hh8d621","country_code":"US","type":"company","lineage":["https://openalex.org/I1341412227"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Bhuvana Ramabhadran","raw_affiliation_strings":["Multilingual Analytics, IBM Research, NY, USA","Multilingual Analytics, IBM Res., Yorktown Heights, NY, USA"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Multilingual Analytics, IBM Research, NY, USA","institution_ids":[]},{"raw_affiliation_string":"Multilingual Analytics, IBM Res., Yorktown Heights, NY, USA","institution_ids":["https://openalex.org/I1341412227"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":2,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":14.6031,"has_fulltext":false,"cited_by_count":60,"citation_normalized_percentile":{"value":0.98880231,"is_in_top_1_percent":false,"is_in_top_10_percent":true},"cited_by_percentile_year":{"min":89,"max":99},"biblio":{"volume":"21","issue":"11","first_page":"2267","last_page":"2276"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10201","display_name":"Speech Recognition and Synthesis","score":0.9998999834060669,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},"topics":[{"id":"https://openalex.org/T10201","display_name":"Speech Recognition and Synthesis","score":0.9998999834060669,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T10860","display_name":"Speech and Audio Processing","score":0.9994999766349792,"subfield":{"id":"https://openalex.org/subfields/1711","display_name":"Signal Processing"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T11309","display_name":"Music and Audio Processing","score":0.9980999827384949,"subfield":{"id":"https://openalex.org/subfields/1711","display_name":"Signal Processing"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.8548800945281982},{"id":"https://openalex.org/keywords/speech-recognition","display_name":"Speech recognition","score":0.6136130690574646},{"id":"https://openalex.org/keywords/hidden-markov-model","display_name":"Hidden Markov model","score":0.5925135016441345},{"id":"https://openalex.org/keywords/deep-neural-networks","display_name":"Deep neural networks","score":0.5591785907745361},{"id":"https://openalex.org/keywords/artificial-neural-network","display_name":"Artificial neural network","score":0.5537044405937195},{"id":"https://openalex.org/keywords/task","display_name":"Task (project management)","score":0.5130301713943481},{"id":"https://openalex.org/keywords/vocabulary","display_name":"Vocabulary","score":0.5103158354759216},{"id":"https://openalex.org/keywords/gradient-descent","display_name":"Gradient descent","score":0.4815695285797119},{"id":"https://openalex.org/keywords/latency","display_name":"Latency (audio)","score":0.4648951590061188},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.4631914496421814},{"id":"https://openalex.org/keywords/regularization","display_name":"Regularization (linguistics)","score":0.42011579871177673},{"id":"https://openalex.org/keywords/computation","display_name":"Computation","score":0.4185031056404114},{"id":"https://openalex.org/keywords/stochastic-gradient-descent","display_name":"Stochastic gradient descent","score":0.4111085534095764},{"id":"https://openalex.org/keywords/algorithm","display_name":"Algorithm","score":0.10526415705680847}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8548800945281982},{"id":"https://openalex.org/C28490314","wikidata":"https://www.wikidata.org/wiki/Q189436","display_name":"Speech recognition","level":1,"score":0.6136130690574646},{"id":"https://openalex.org/C23224414","wikidata":"https://www.wikidata.org/wiki/Q176769","display_name":"Hidden Markov model","level":2,"score":0.5925135016441345},{"id":"https://openalex.org/C2984842247","wikidata":"https://www.wikidata.org/wiki/Q197536","display_name":"Deep neural networks","level":3,"score":0.5591785907745361},{"id":"https://openalex.org/C50644808","wikidata":"https://www.wikidata.org/wiki/Q192776","display_name":"Artificial neural network","level":2,"score":0.5537044405937195},{"id":"https://openalex.org/C2780451532","wikidata":"https://www.wikidata.org/wiki/Q759676","display_name":"Task (project management)","level":2,"score":0.5130301713943481},{"id":"https://openalex.org/C2777601683","wikidata":"https://www.wikidata.org/wiki/Q6499736","display_name":"Vocabulary","level":2,"score":0.5103158354759216},{"id":"https://openalex.org/C153258448","wikidata":"https://www.wikidata.org/wiki/Q1199743","display_name":"Gradient descent","level":3,"score":0.4815695285797119},{"id":"https://openalex.org/C82876162","wikidata":"https://www.wikidata.org/wiki/Q17096504","display_name":"Latency (audio)","level":2,"score":0.4648951590061188},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.4631914496421814},{"id":"https://openalex.org/C2776135515","wikidata":"https://www.wikidata.org/wiki/Q17143721","display_name":"Regularization (linguistics)","level":2,"score":0.42011579871177673},{"id":"https://openalex.org/C45374587","wikidata":"https://www.wikidata.org/wiki/Q12525525","display_name":"Computation","level":2,"score":0.4185031056404114},{"id":"https://openalex.org/C206688291","wikidata":"https://www.wikidata.org/wiki/Q7617819","display_name":"Stochastic gradient descent","level":3,"score":0.4111085534095764},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.10526415705680847},{"id":"https://openalex.org/C187736073","wikidata":"https://www.wikidata.org/wiki/Q2920921","display_name":"Management","level":1,"score":0.0},{"id":"https://openalex.org/C162324750","wikidata":"https://www.wikidata.org/wiki/Q8134","display_name":"Economics","level":0,"score":0.0},{"id":"https://openalex.org/C41895202","wikidata":"https://www.wikidata.org/wiki/Q8162","display_name":"Linguistics","level":1,"score":0.0},{"id":"https://openalex.org/C138885662","wikidata":"https://www.wikidata.org/wiki/Q5891","display_name":"Philosophy","level":0,"score":0.0},{"id":"https://openalex.org/C76155785","wikidata":"https://www.wikidata.org/wiki/Q418","display_name":"Telecommunications","level":1,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/tasl.2013.2284378","is_oa":false,"landing_page_url":"https://doi.org/10.1109/tasl.2013.2284378","pdf_url":null,"source":{"id":"https://openalex.org/S199497470","display_name":"IEEE Transactions on Audio Speech and Language Processing","issn_l":"1558-7916","issn":["1558-7916","1558-7924"],"is_oa":false,"is_in_doaj":false,"is_core":true,"host_organization":"https://openalex.org/P4310319808","host_organization_name":"Institute of Electrical and Electronics Engineers","host_organization_lineage":["https://openalex.org/P4310319808"],"host_organization_lineage_names":["Institute of Electrical and Electronics Engineers"],"type":"journal"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"IEEE Transactions on Audio, Speech, and Language Processing","raw_type":"journal-article"}],"best_oa_location":null,"sustainable_development_goals":[{"id":"https://metadata.un.org/sdg/4","display_name":"Quality Education","score":0.7300000190734863}],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":45,"referenced_works":["https://openalex.org/W44815768","https://openalex.org/W196761320","https://openalex.org/W217970951","https://openalex.org/W1533861849","https://openalex.org/W1964155876","https://openalex.org/W1970092758","https://openalex.org/W1991083751","https://openalex.org/W2006903949","https://openalex.org/W2076794394","https://openalex.org/W2097428361","https://openalex.org/W2098318492","https://openalex.org/W2101593431","https://openalex.org/W2102017903","https://openalex.org/W2105767123","https://openalex.org/W2106554350","https://openalex.org/W2109722477","https://openalex.org/W2112984492","https://openalex.org/W2114016253","https://openalex.org/W2114766824","https://openalex.org/W2120432001","https://openalex.org/W2130984546","https://openalex.org/W2136922672","https://openalex.org/W2140833774","https://openalex.org/W2155117693","https://openalex.org/W2160306971","https://openalex.org/W2160815625","https://openalex.org/W2166706236","https://openalex.org/W2168231600","https://openalex.org/W2172097686","https://openalex.org/W2296748324","https://openalex.org/W2394932179","https://openalex.org/W2395981916","https://openalex.org/W2403195671","https://openalex.org/W2951282416","https://openalex.org/W2979307260","https://openalex.org/W4285719527","https://openalex.org/W6608133726","https://openalex.org/W6608710415","https://openalex.org/W6631943919","https://openalex.org/W6674634876","https://openalex.org/W6676367512","https://openalex.org/W6677103964","https://openalex.org/W6680887930","https://openalex.org/W6684249991","https://openalex.org/W6684859321"],"related_works":["https://openalex.org/W4206903459","https://openalex.org/W2754816816","https://openalex.org/W4366280654","https://openalex.org/W3160167280","https://openalex.org/W4362706668","https://openalex.org/W4231621013","https://openalex.org/W3171021120","https://openalex.org/W3008318776","https://openalex.org/W2041416246","https://openalex.org/W3020853991"],"abstract_inverted_index":{"While":[0],"Deep":[1],"Neural":[2],"Networks":[3],"(DNNs)":[4],"have":[5],"achieved":[6],"tremendous":[7],"success":[8],"for":[9,69],"large":[10,46],"vocabulary":[11],"continuous":[12],"speech":[13,55],"recognition":[14],"(LVCSR)":[15],"tasks,":[16],"training":[17,49,62,66,86,125],"these":[18],"networks":[19],"is":[20,32,190],"slow.":[21],"Even":[22],"to":[23,29,83,156],"date,":[24],"the":[25,45,92,104,109,117,150,183,188,193],"most":[26],"common":[27],"approach":[28],"train":[30,157],"DNNs":[31,158],"via":[33],"stochastic":[34],"gradient":[35,93],"descent,":[36],"serially":[37],"on":[38,135,159],"one":[39],"machine.":[40],"Serial":[41],"training,":[42,99],"coupled":[43],"with":[44,143],"number":[47,105,184],"of":[48,79,91,106,133,185,187],"parameters":[50,107,186],"(i.e.,":[51,59],"10-50":[52],"million)":[53],"and":[54,97,165],"data":[56],"set":[57],"sizes":[58],"20-100":[60],"million":[61],"points)":[63],"makes":[64],"DNN":[65,85,124,189],"very":[67],"slow":[68],"LVCSR":[70],"tasks.":[71],"In":[72],"this":[73],"work,":[74],"we":[75,121,153],"explore":[76],"a":[77,112,131,136,160,166,177],"variety":[78],"different":[80],"optimization":[81,119],"techniques":[82],"improve":[84],"speed.":[87],"This":[88],"includes":[89],"parallelization":[90],"computation":[94],"during":[95],"cross-entropy":[96],"sequence":[98],"as":[100,102],"well":[101],"reducing":[103],"in":[108,146],"network":[110],"using":[111,149],"low-rank":[113],"matrix":[114],"factorization.":[115],"Applying":[116],"proposed":[118,151],"techniques,":[120,152],"show":[122],"that":[123],"can":[126],"be":[127],"sped":[128],"up":[129],"by":[130],"factor":[132],"3":[134],"50-hour":[137],"English":[138,168],"Broadcast":[139],"News":[140],"(BN)":[141],"task":[142,164],"no":[144],"loss":[145],"accuracy.":[147],"Furthermore,":[148],"are":[154],"able":[155],"300-hr":[161],"Switchboard":[162],"(SWB)":[163],"400-hr":[167],"BN":[169],"task,":[170],"showing":[171],"improvements":[172],"between":[173],"9-30%":[174],"relative":[175],"over":[176],"state-of-the":[178],"art":[179],"GMM/HMM":[180,194],"system":[181],"while":[182],"smaller":[191],"than":[192],"system.":[195]},"counts_by_year":[{"year":2025,"cited_by_count":2},{"year":2024,"cited_by_count":1},{"year":2022,"cited_by_count":2},{"year":2021,"cited_by_count":5},{"year":2020,"cited_by_count":5},{"year":2019,"cited_by_count":5},{"year":2018,"cited_by_count":6},{"year":2017,"cited_by_count":3},{"year":2016,"cited_by_count":6},{"year":2015,"cited_by_count":10},{"year":2014,"cited_by_count":10},{"year":2013,"cited_by_count":4},{"year":2012,"cited_by_count":1}],"updated_date":"2025-11-06T03:46:38.306776","created_date":"2025-10-10T00:00:00"}