{"id":"https://openalex.org/W3015315932","doi":"https://doi.org/10.1109/icassp40776.2020.9054663","title":"Exploring Pre-Training with Alignments for RNN Transducer Based End-to-End Speech Recognition","display_name":"Exploring Pre-Training with Alignments for RNN Transducer Based End-to-End Speech Recognition","publication_year":2020,"publication_date":"2020-04-09","ids":{"openalex":"https://openalex.org/W3015315932","doi":"https://doi.org/10.1109/icassp40776.2020.9054663","mag":"3015315932"},"language":"en","primary_location":{"id":"doi:10.1109/icassp40776.2020.9054663","is_oa":false,"landing_page_url":"https://doi.org/10.1109/icassp40776.2020.9054663","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"ICASSP 2020 - 2020 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)","raw_type":"proceedings-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/A5079582557","display_name":"Hu Hu","orcid":"https://orcid.org/0000-0003-2101-8317"},"institutions":[{"id":"https://openalex.org/I1290206253","display_name":"Microsoft (United States)","ror":"https://ror.org/00d0nc645","country_code":"US","type":"company","lineage":["https://openalex.org/I1290206253"]}],"countries":["US"],"is_corresponding":true,"raw_author_name":"Hu Hu","raw_affiliation_strings":["Microsoft Speech and Language Group, Redmond, WA, USA"],"affiliations":[{"raw_affiliation_string":"Microsoft Speech and Language Group, Redmond, WA, USA","institution_ids":["https://openalex.org/I1290206253"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5101987526","display_name":"Rui Zhao","orcid":"https://orcid.org/0000-0002-9699-9984"},"institutions":[{"id":"https://openalex.org/I1290206253","display_name":"Microsoft (United States)","ror":"https://ror.org/00d0nc645","country_code":"US","type":"company","lineage":["https://openalex.org/I1290206253"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Rui Zhao","raw_affiliation_strings":["Microsoft Speech and Language Group, Redmond, WA, USA"],"affiliations":[{"raw_affiliation_string":"Microsoft Speech and Language Group, Redmond, WA, USA","institution_ids":["https://openalex.org/I1290206253"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100365053","display_name":"Jinyu Li","orcid":"https://orcid.org/0000-0002-1089-9748"},"institutions":[{"id":"https://openalex.org/I1290206253","display_name":"Microsoft (United States)","ror":"https://ror.org/00d0nc645","country_code":"US","type":"company","lineage":["https://openalex.org/I1290206253"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Jinyu Li","raw_affiliation_strings":["Microsoft Speech and Language Group, Redmond, WA, USA"],"affiliations":[{"raw_affiliation_string":"Microsoft Speech and Language Group, Redmond, WA, USA","institution_ids":["https://openalex.org/I1290206253"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5101607148","display_name":"Liang Lu","orcid":"https://orcid.org/0000-0003-4005-679X"},"institutions":[{"id":"https://openalex.org/I1290206253","display_name":"Microsoft (United States)","ror":"https://ror.org/00d0nc645","country_code":"US","type":"company","lineage":["https://openalex.org/I1290206253"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Liang Lu","raw_affiliation_strings":["Microsoft Speech and Language Group, Redmond, WA, USA"],"affiliations":[{"raw_affiliation_string":"Microsoft Speech and Language Group, Redmond, WA, USA","institution_ids":["https://openalex.org/I1290206253"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5101928537","display_name":"Yifan Gong","orcid":"https://orcid.org/0000-0002-3912-097X"},"institutions":[{"id":"https://openalex.org/I1290206253","display_name":"Microsoft (United States)","ror":"https://ror.org/00d0nc645","country_code":"US","type":"company","lineage":["https://openalex.org/I1290206253"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Yifan Gong","raw_affiliation_strings":["Microsoft Speech and Language Group, Redmond, WA, USA"],"affiliations":[{"raw_affiliation_string":"Microsoft Speech and Language Group, Redmond, WA, USA","institution_ids":["https://openalex.org/I1290206253"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":5,"corresponding_author_ids":["https://openalex.org/A5079582557"],"corresponding_institution_ids":["https://openalex.org/I1290206253"],"apc_list":null,"apc_paid":null,"fwci":4.3745,"has_fulltext":false,"cited_by_count":35,"citation_normalized_percentile":{"value":0.95262174,"is_in_top_1_percent":false,"is_in_top_10_percent":true},"cited_by_percentile_year":{"min":94,"max":99},"biblio":{"volume":null,"issue":null,"first_page":"7079","last_page":"7083"},"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/T10181","display_name":"Natural Language Processing Techniques","score":0.9987000226974487,"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/T11309","display_name":"Music and Audio Processing","score":0.9958999752998352,"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/recurrent-neural-network","display_name":"Recurrent neural network","score":0.9215217232704163},{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.8534023761749268},{"id":"https://openalex.org/keywords/initialization","display_name":"Initialization","score":0.813791811466217},{"id":"https://openalex.org/keywords/encoder","display_name":"Encoder","score":0.6323708295822144},{"id":"https://openalex.org/keywords/speech-recognition","display_name":"Speech recognition","score":0.6174535751342773},{"id":"https://openalex.org/keywords/word-error-rate","display_name":"Word error rate","score":0.5386955142021179},{"id":"https://openalex.org/keywords/end-to-end-principle","display_name":"End-to-end principle","score":0.5282757878303528},{"id":"https://openalex.org/keywords/language-model","display_name":"Language model","score":0.510513424873352},{"id":"https://openalex.org/keywords/leverage","display_name":"Leverage (statistics)","score":0.5033263564109802},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.49866509437561035},{"id":"https://openalex.org/keywords/connectionism","display_name":"Connectionism","score":0.44150885939598083},{"id":"https://openalex.org/keywords/artificial-neural-network","display_name":"Artificial neural network","score":0.3572843372821808}],"concepts":[{"id":"https://openalex.org/C147168706","wikidata":"https://www.wikidata.org/wiki/Q1457734","display_name":"Recurrent neural network","level":3,"score":0.9215217232704163},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8534023761749268},{"id":"https://openalex.org/C114466953","wikidata":"https://www.wikidata.org/wiki/Q6034165","display_name":"Initialization","level":2,"score":0.813791811466217},{"id":"https://openalex.org/C118505674","wikidata":"https://www.wikidata.org/wiki/Q42586063","display_name":"Encoder","level":2,"score":0.6323708295822144},{"id":"https://openalex.org/C28490314","wikidata":"https://www.wikidata.org/wiki/Q189436","display_name":"Speech recognition","level":1,"score":0.6174535751342773},{"id":"https://openalex.org/C40969351","wikidata":"https://www.wikidata.org/wiki/Q3516228","display_name":"Word error rate","level":2,"score":0.5386955142021179},{"id":"https://openalex.org/C74296488","wikidata":"https://www.wikidata.org/wiki/Q2527392","display_name":"End-to-end principle","level":2,"score":0.5282757878303528},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.510513424873352},{"id":"https://openalex.org/C153083717","wikidata":"https://www.wikidata.org/wiki/Q6535263","display_name":"Leverage (statistics)","level":2,"score":0.5033263564109802},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.49866509437561035},{"id":"https://openalex.org/C8521452","wikidata":"https://www.wikidata.org/wiki/Q203790","display_name":"Connectionism","level":3,"score":0.44150885939598083},{"id":"https://openalex.org/C50644808","wikidata":"https://www.wikidata.org/wiki/Q192776","display_name":"Artificial neural network","level":2,"score":0.3572843372821808},{"id":"https://openalex.org/C111919701","wikidata":"https://www.wikidata.org/wiki/Q9135","display_name":"Operating system","level":1,"score":0.0},{"id":"https://openalex.org/C199360897","wikidata":"https://www.wikidata.org/wiki/Q9143","display_name":"Programming language","level":1,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/icassp40776.2020.9054663","is_oa":false,"landing_page_url":"https://doi.org/10.1109/icassp40776.2020.9054663","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"ICASSP 2020 - 2020 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)","raw_type":"proceedings-article"}],"best_oa_location":null,"sustainable_development_goals":[{"display_name":"Quality Education","score":0.46000000834465027,"id":"https://metadata.un.org/sdg/4"}],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":37,"referenced_works":["https://openalex.org/W854541894","https://openalex.org/W1533416326","https://openalex.org/W1828163288","https://openalex.org/W2102113734","https://openalex.org/W2127141656","https://openalex.org/W2133564696","https://openalex.org/W2143612262","https://openalex.org/W2157331557","https://openalex.org/W2160815625","https://openalex.org/W2327501763","https://openalex.org/W2394932179","https://openalex.org/W2525778437","https://openalex.org/W2746192915","https://openalex.org/W2750499125","https://openalex.org/W2936123380","https://openalex.org/W2962742956","https://openalex.org/W2962760690","https://openalex.org/W2962784628","https://openalex.org/W2962826786","https://openalex.org/W2963211739","https://openalex.org/W2963414781","https://openalex.org/W2963827914","https://openalex.org/W2963970535","https://openalex.org/W2964272710","https://openalex.org/W2964308564","https://openalex.org/W2966163367","https://openalex.org/W2971840980","https://openalex.org/W3007227084","https://openalex.org/W3011207290","https://openalex.org/W4294619417","https://openalex.org/W6623517193","https://openalex.org/W6638575559","https://openalex.org/W6638749077","https://openalex.org/W6675365184","https://openalex.org/W6679434410","https://openalex.org/W6747158283","https://openalex.org/W6756286202"],"related_works":["https://openalex.org/W2916997151","https://openalex.org/W2495256954","https://openalex.org/W2259317772","https://openalex.org/W2949174760","https://openalex.org/W2594897229","https://openalex.org/W2151348424","https://openalex.org/W4221142855","https://openalex.org/W2050138804","https://openalex.org/W4398173524","https://openalex.org/W4290708361"],"abstract_inverted_index":{"Recently,":[0],"the":[1,38,51,70,83,123,144,156,161],"recurrent":[2],"neural":[3,44],"network":[4],"transducer":[5],"(RNN-T)":[6],"architecture":[7],"has":[8],"become":[9],"an":[10],"emerging":[11],"trend":[12],"in":[13],"end-to-end":[14],"automatic":[15],"speech":[16,29],"recognition":[17],"research":[18],"due":[19],"to":[20,49,55,68,81,93],"its":[21],"advantages":[22],"of":[23],"being":[24],"capable":[25],"for":[26],"online":[27],"streaming":[28],"recognition.":[30],"However,":[31],"RNN-T":[32,52,71,84,157],"training":[33,53],"is":[34,54],"made":[35],"difficult":[36],"by":[37],"huge":[39],"memory":[40],"requirements,":[41],"and":[42,97,130,143],"complicated":[43],"structure.":[45],"A":[46],"common":[47],"solution":[48,126],"ease":[50],"employ":[56],"connectionist":[57],"temporal":[58],"classification":[59],"(CTC)":[60],"model":[61,66,158],"along":[62],"with":[63,109,140],"RNN":[64],"language":[65],"(RNNLM)":[67],"initialize":[69],"parameters.":[72],"In":[73,121],"this":[74],"work,":[75],"we":[76],"conversely":[77],"leverage":[78],"external":[79],"alignments":[80],"seed":[82],"model.":[85],"Two":[86],"different":[87],"pre-training":[88,99,125],"solutions":[89,152],"are":[90],"explored,":[91],"referred":[92],"as":[94],"encoder":[95,124],"pre-training,":[96],"whole-network":[98],"respectively.":[100,150],"Evaluated":[101],"on":[102],"Microsoft":[103],"65,000":[104],"hours":[105],"anonymized":[106],"production":[107],"data":[108],"personally":[110],"identifiable":[111],"information":[112],"removed,":[113],"our":[114],"proposed":[115],"methods":[116],"can":[117],"obtain":[118],"significant":[119],"improvement.":[120],"particular,":[122],"achieved":[127],"a":[128,131],"10%":[129],"8%":[132],"relative":[133],"word":[134],"error":[135],"rate":[136],"reduction":[137],"when":[138],"compared":[139],"random":[141],"initialization":[142,148],"widely":[145],"used":[146],"CTC+RNNLM":[147],"strategy,":[149],"Our":[151],"also":[153],"significantly":[154],"reduce":[155],"latency":[159],"from":[160],"baseline.":[162]},"counts_by_year":[{"year":2024,"cited_by_count":2},{"year":2023,"cited_by_count":4},{"year":2022,"cited_by_count":3},{"year":2021,"cited_by_count":15},{"year":2020,"cited_by_count":11}],"updated_date":"2025-11-06T03:46:38.306776","created_date":"2025-10-10T00:00:00"}