{"id":"https://openalex.org/W3007227084","doi":"https://doi.org/10.1109/asru46091.2019.9003906","title":"Improving RNN Transducer Modeling for End-to-End Speech Recognition","display_name":"Improving RNN Transducer Modeling for End-to-End Speech Recognition","publication_year":2019,"publication_date":"2019-12-01","ids":{"openalex":"https://openalex.org/W3007227084","doi":"https://doi.org/10.1109/asru46091.2019.9003906","mag":"3007227084"},"language":"en","primary_location":{"id":"doi:10.1109/asru46091.2019.9003906","is_oa":false,"landing_page_url":"https://doi.org/10.1109/asru46091.2019.9003906","pdf_url":null,"source":{"id":"https://openalex.org/S4306498489","display_name":"2019 IEEE Automatic Speech Recognition and Understanding Workshop (ASRU)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":true,"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"journal"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2019 IEEE Automatic Speech Recognition and Understanding Workshop (ASRU)","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/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":true,"raw_author_name":"Jinyu Li","raw_affiliation_strings":["Speech and Language Group, Microsoft"],"affiliations":[{"raw_affiliation_string":"Speech and Language Group, Microsoft","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":["Speech and Language Group, Microsoft"],"affiliations":[{"raw_affiliation_string":"Speech and Language Group, Microsoft","institution_ids":["https://openalex.org/I1290206253"]}]},{"author_position":"middle","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":false,"raw_author_name":"Hu Hu","raw_affiliation_strings":["Speech and Language Group, Microsoft"],"affiliations":[{"raw_affiliation_string":"Speech and Language Group, Microsoft","institution_ids":["https://openalex.org/I1290206253"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5077401426","display_name":"Yifan Gong","orcid":"https://orcid.org/0000-0001-8786-3391"},"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":["Speech and Language Group, Microsoft"],"affiliations":[{"raw_affiliation_string":"Speech and Language Group, Microsoft","institution_ids":["https://openalex.org/I1290206253"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":4,"corresponding_author_ids":["https://openalex.org/A5100365053"],"corresponding_institution_ids":["https://openalex.org/I1290206253"],"apc_list":null,"apc_paid":null,"fwci":18.1991,"has_fulltext":false,"cited_by_count":178,"citation_normalized_percentile":{"value":0.99380679,"is_in_top_1_percent":true,"is_in_top_10_percent":true},"cited_by_percentile_year":{"min":90,"max":100},"biblio":{"volume":null,"issue":null,"first_page":"114","last_page":"121"},"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/T11309","display_name":"Music and Audio Processing","score":0.9991000294685364,"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/T10860","display_name":"Speech and Audio Processing","score":0.9972000122070312,"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":0.9349644780158997},{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.7650325298309326},{"id":"https://openalex.org/keywords/encoder","display_name":"Encoder","score":0.5910576581954956},{"id":"https://openalex.org/keywords/speech-recognition","display_name":"Speech recognition","score":0.5837797522544861},{"id":"https://openalex.org/keywords/word-error-rate","display_name":"Word error rate","score":0.5368384122848511},{"id":"https://openalex.org/keywords/reduction","display_name":"Reduction (mathematics)","score":0.524610161781311},{"id":"https://openalex.org/keywords/language-model","display_name":"Language model","score":0.48464256525039673},{"id":"https://openalex.org/keywords/connectionism","display_name":"Connectionism","score":0.48165905475616455},{"id":"https://openalex.org/keywords/end-to-end-principle","display_name":"End-to-end principle","score":0.42263007164001465},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.4084226191043854},{"id":"https://openalex.org/keywords/artificial-neural-network","display_name":"Artificial neural network","score":0.15686172246932983},{"id":"https://openalex.org/keywords/mathematics","display_name":"Mathematics","score":0.09006699919700623}],"concepts":[{"id":"https://openalex.org/C147168706","wikidata":"https://www.wikidata.org/wiki/Q1457734","display_name":"Recurrent neural network","level":3,"score":0.9349644780158997},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7650325298309326},{"id":"https://openalex.org/C118505674","wikidata":"https://www.wikidata.org/wiki/Q42586063","display_name":"Encoder","level":2,"score":0.5910576581954956},{"id":"https://openalex.org/C28490314","wikidata":"https://www.wikidata.org/wiki/Q189436","display_name":"Speech recognition","level":1,"score":0.5837797522544861},{"id":"https://openalex.org/C40969351","wikidata":"https://www.wikidata.org/wiki/Q3516228","display_name":"Word error rate","level":2,"score":0.5368384122848511},{"id":"https://openalex.org/C111335779","wikidata":"https://www.wikidata.org/wiki/Q3454686","display_name":"Reduction (mathematics)","level":2,"score":0.524610161781311},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.48464256525039673},{"id":"https://openalex.org/C8521452","wikidata":"https://www.wikidata.org/wiki/Q203790","display_name":"Connectionism","level":3,"score":0.48165905475616455},{"id":"https://openalex.org/C74296488","wikidata":"https://www.wikidata.org/wiki/Q2527392","display_name":"End-to-end principle","level":2,"score":0.42263007164001465},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.4084226191043854},{"id":"https://openalex.org/C50644808","wikidata":"https://www.wikidata.org/wiki/Q192776","display_name":"Artificial neural network","level":2,"score":0.15686172246932983},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.09006699919700623},{"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/C2524010","wikidata":"https://www.wikidata.org/wiki/Q8087","display_name":"Geometry","level":1,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/asru46091.2019.9003906","is_oa":false,"landing_page_url":"https://doi.org/10.1109/asru46091.2019.9003906","pdf_url":null,"source":{"id":"https://openalex.org/S4306498489","display_name":"2019 IEEE Automatic Speech Recognition and Understanding Workshop (ASRU)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":true,"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"journal"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2019 IEEE Automatic Speech Recognition and Understanding Workshop (ASRU)","raw_type":"proceedings-article"}],"best_oa_location":null,"sustainable_development_goals":[{"display_name":"Quality Education","score":0.41999998688697815,"id":"https://metadata.un.org/sdg/4"}],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":55,"referenced_works":["https://openalex.org/W854541894","https://openalex.org/W1533416326","https://openalex.org/W1600744878","https://openalex.org/W1665214252","https://openalex.org/W1828163288","https://openalex.org/W1924770834","https://openalex.org/W2102113734","https://openalex.org/W2127141656","https://openalex.org/W2133564696","https://openalex.org/W2157331557","https://openalex.org/W2193413348","https://openalex.org/W2291522532","https://openalex.org/W2327501763","https://openalex.org/W2396144419","https://openalex.org/W2396384435","https://openalex.org/W2475988411","https://openalex.org/W2513938599","https://openalex.org/W2545177271","https://openalex.org/W2586630418","https://openalex.org/W2734724284","https://openalex.org/W2739427748","https://openalex.org/W2746192915","https://openalex.org/W2750499125","https://openalex.org/W2911544293","https://openalex.org/W2912691643","https://openalex.org/W2936123380","https://openalex.org/W2939248538","https://openalex.org/W2962742956","https://openalex.org/W2962760690","https://openalex.org/W2962784628","https://openalex.org/W2962824709","https://openalex.org/W2962826786","https://openalex.org/W2962962542","https://openalex.org/W2963211739","https://openalex.org/W2963326356","https://openalex.org/W2963414781","https://openalex.org/W2963827914","https://openalex.org/W2963917928","https://openalex.org/W2963970535","https://openalex.org/W2964084166","https://openalex.org/W2964272710","https://openalex.org/W2964308564","https://openalex.org/W2966163367","https://openalex.org/W4294619417","https://openalex.org/W6623517193","https://openalex.org/W6637242042","https://openalex.org/W6638749077","https://openalex.org/W6675365184","https://openalex.org/W6679434410","https://openalex.org/W6687566353","https://openalex.org/W6741807409","https://openalex.org/W6747398299","https://openalex.org/W6758661737","https://openalex.org/W6758956981","https://openalex.org/W6761230321"],"related_works":["https://openalex.org/W2160451571","https://openalex.org/W2916997151","https://openalex.org/W2949174760","https://openalex.org/W2495256954","https://openalex.org/W1566315437","https://openalex.org/W4221142855","https://openalex.org/W2594897229","https://openalex.org/W2151348424","https://openalex.org/W2050138804","https://openalex.org/W767271433"],"abstract_inverted_index":{"In":[0,61],"the":[1,14,32,43,66,75,82,110,128,149,161,180],"last":[2],"few":[3],"years,":[4],"an":[5],"emerging":[6],"trend":[7],"in":[8,69,187],"automatic":[9],"speech":[10],"recognition":[11],"research":[12],"is":[13,50,157],"study":[15],"of":[16,78,184],"end-to-end":[17],"(E2E)":[18],"systems.":[19],"Connectionist":[20],"Temporal":[21],"Classification":[22],"(CTC),":[23],"Attention":[24],"Encoder-Decoder":[25],"(AED),":[26],"and":[27,54,123,175],"RNN":[28],"Transducer":[29],"(RNN-T)":[30],"are":[31],"most":[33],"popular":[34],"three":[35,39],"methods.":[36],"Among":[37],"these":[38],"methods,":[40],"RNN-T":[41,67,79,107,130,151,155],"has":[42],"advantages":[44],"to":[45,52,80],"do":[46],"online":[47],"streaming":[48],"which":[49],"challenging":[51],"AED":[53],"it":[55],"doesn't":[56],"have":[57,89],"CTC's":[58],"frame-independence":[59],"assumption.":[60],"this":[62],"paper,":[63],"we":[64,73,87,98,105],"improve":[65],"training":[68,76,91,95],"two":[70],"aspects.":[71],"First,":[72],"optimize":[74],"algorithm":[77],"reduce":[81],"memory":[83],"consumption":[84],"so":[85,103],"that":[86,104],"can":[88],"larger":[90],"minibatch":[92],"for":[93],"faster":[94],"speed.":[96],"Second,":[97],"propose":[99],"better":[100,159],"model":[101,131,135,156,164,183],"structures":[102],"obtain":[106],"models":[108],"with":[109,118,132,165],"very":[111],"good":[112],"accuracy":[113],"but":[114],"small":[115],"footprint.":[116],"Trained":[117],"30":[119],"thousand":[120],"hours":[121],"anonymized":[122],"transcribed":[124],"Microsoft":[125],"production":[126],"data,":[127],"best":[129,154],"even":[133],"smaller":[134],"size":[136,167],"(216":[137],"Megabytes)":[138],"achieves":[139],"up-to":[140,170],"11.8%":[141],"relative":[142,172],"word":[143],"error":[144],"rate":[145],"(WER)":[146],"reduction":[147],"from":[148],"baseline":[150],"model.":[152],"This":[153],"significantly":[158],"than":[160],"device":[162],"hybrid":[163,182],"similar":[166,177],"by":[168],"achieving":[169],"15.0%":[171],"WER":[173],"reduction,":[174],"obtains":[176],"WERs":[178],"as":[179],"server":[181],"5120":[185],"Megabytes":[186],"size.":[188]},"counts_by_year":[{"year":2026,"cited_by_count":2},{"year":2025,"cited_by_count":9},{"year":2024,"cited_by_count":10},{"year":2023,"cited_by_count":32},{"year":2022,"cited_by_count":34},{"year":2021,"cited_by_count":61},{"year":2020,"cited_by_count":29},{"year":2019,"cited_by_count":1}],"updated_date":"2026-04-02T15:55:50.835912","created_date":"2025-10-10T00:00:00"}