{"id":"https://openalex.org/W3094667432","doi":"https://doi.org/10.21437/interspeech.2020-3016","title":"Developing RNN-T Models Surpassing High-Performance Hybrid Models with Customization Capability","display_name":"Developing RNN-T Models Surpassing High-Performance Hybrid Models with Customization Capability","publication_year":2020,"publication_date":"2020-10-25","ids":{"openalex":"https://openalex.org/W3094667432","doi":"https://doi.org/10.21437/interspeech.2020-3016","mag":"3094667432"},"language":"en","primary_location":{"id":"doi:10.21437/interspeech.2020-3016","is_oa":false,"landing_page_url":"https://doi.org/10.21437/interspeech.2020-3016","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Interspeech 2020","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":false,"raw_author_name":"Jinyu Li","raw_affiliation_strings":["Microsoft Speech and Language Group"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Microsoft Speech and Language Group","institution_ids":["https://openalex.org/I1290206253"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5101414593","display_name":"Rui Zhao","orcid":"https://orcid.org/0000-0002-5147-2695"},"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"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Microsoft Speech and Language Group","institution_ids":["https://openalex.org/I1290206253"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5101749753","display_name":"Zhong Meng","orcid":"https://orcid.org/0000-0001-7814-5929"},"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":"Zhong Meng","raw_affiliation_strings":["Microsoft Speech and Language Group"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Microsoft Speech and Language Group","institution_ids":["https://openalex.org/I1290206253"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100360935","display_name":"Yanqing Liu","orcid":"https://orcid.org/0000-0003-0412-8805"},"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":"Yanqing Liu","raw_affiliation_strings":["Microsoft Speech and Language Group"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Microsoft Speech and Language Group","institution_ids":["https://openalex.org/I1290206253"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5055564488","display_name":"Wenning Wei","orcid":null},"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":"Wenning Wei","raw_affiliation_strings":["Microsoft Speech and Language Group"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Microsoft Speech and Language Group","institution_ids":["https://openalex.org/I1290206253"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5102376243","display_name":"Sarangarajan Parthasarathy","orcid":null},"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":"Sarangarajan Parthasarathy","raw_affiliation_strings":["Microsoft Speech and Language Group"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Microsoft Speech and Language Group","institution_ids":["https://openalex.org/I1290206253"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5042972227","display_name":"Vadim Mazalov","orcid":null},"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":"Vadim Mazalov","raw_affiliation_strings":["Microsoft Speech and Language Group"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Microsoft Speech and Language Group","institution_ids":["https://openalex.org/I1290206253"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5101767354","display_name":"Zhenghao Wang","orcid":"https://orcid.org/0000-0003-2676-7683"},"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":"Zhenghao Wang","raw_affiliation_strings":["Microsoft Speech and Language Group"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Microsoft Speech and Language Group","institution_ids":["https://openalex.org/I1290206253"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5061724331","display_name":"Lei He","orcid":"https://orcid.org/0000-0003-3028-6305"},"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":"Lei He","raw_affiliation_strings":["Microsoft Speech and Language Group"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Microsoft Speech and Language Group","institution_ids":["https://openalex.org/I1290206253"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100329353","display_name":"Sheng Zhao","orcid":"https://orcid.org/0000-0002-9624-5381"},"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":"Sheng Zhao","raw_affiliation_strings":["Microsoft Speech and Language Group"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Microsoft Speech and Language Group","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"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Microsoft Speech and Language Group","institution_ids":["https://openalex.org/I1290206253"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":11,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":11.3745,"has_fulltext":false,"cited_by_count":97,"citation_normalized_percentile":{"value":0.98787589,"is_in_top_1_percent":false,"is_in_top_10_percent":true},"cited_by_percentile_year":{"min":98,"max":100},"biblio":{"volume":null,"issue":null,"first_page":"3590","last_page":"3594"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10201","display_name":"Speech Recognition and Synthesis","score":0.9988999962806702,"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":0.9988999962806702,"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":0.9768999814987183,"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/T10028","display_name":"Topic Modeling","score":0.9639999866485596,"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"}}],"keywords":[{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.7032973766326904},{"id":"https://openalex.org/keywords/personalization","display_name":"Personalization","score":0.6437182426452637},{"id":"https://openalex.org/keywords/computer-architecture","display_name":"Computer architecture","score":0.3820405900478363},{"id":"https://openalex.org/keywords/world-wide-web","display_name":"World Wide Web","score":0.12335425615310669}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7032973766326904},{"id":"https://openalex.org/C183003079","wikidata":"https://www.wikidata.org/wiki/Q1000371","display_name":"Personalization","level":2,"score":0.6437182426452637},{"id":"https://openalex.org/C118524514","wikidata":"https://www.wikidata.org/wiki/Q173212","display_name":"Computer architecture","level":1,"score":0.3820405900478363},{"id":"https://openalex.org/C136764020","wikidata":"https://www.wikidata.org/wiki/Q466","display_name":"World Wide Web","level":1,"score":0.12335425615310669}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.21437/interspeech.2020-3016","is_oa":false,"landing_page_url":"https://doi.org/10.21437/interspeech.2020-3016","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Interspeech 2020","raw_type":"proceedings-article"}],"best_oa_location":null,"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":34,"referenced_works":["https://openalex.org/W854541894","https://openalex.org/W1736701665","https://openalex.org/W1828163288","https://openalex.org/W1915251500","https://openalex.org/W2003123121","https://openalex.org/W2121879602","https://openalex.org/W2133564696","https://openalex.org/W2160815625","https://openalex.org/W2327501763","https://openalex.org/W2396384435","https://openalex.org/W2402268235","https://openalex.org/W2507699225","https://openalex.org/W2750499125","https://openalex.org/W2896457183","https://openalex.org/W2903853691","https://openalex.org/W2916997151","https://openalex.org/W2939248538","https://openalex.org/W2956480774","https://openalex.org/W2962760690","https://openalex.org/W2963414781","https://openalex.org/W2963523217","https://openalex.org/W2963827914","https://openalex.org/W2965373594","https://openalex.org/W2973122799","https://openalex.org/W2973138437","https://openalex.org/W2987019345","https://openalex.org/W3007227084","https://openalex.org/W3015315932","https://openalex.org/W3015686596","https://openalex.org/W3015752032","https://openalex.org/W3016155933","https://openalex.org/W3095311338","https://openalex.org/W4294619240","https://openalex.org/W4385245566"],"related_works":["https://openalex.org/W2899084033","https://openalex.org/W2748952813","https://openalex.org/W2109940557","https://openalex.org/W2466832359","https://openalex.org/W3040708694","https://openalex.org/W3172493050","https://openalex.org/W2390279801","https://openalex.org/W1582019636","https://openalex.org/W1499005795","https://openalex.org/W2593798266"],"abstract_inverted_index":{"Because":[0],"of":[1,35,61],"its":[2],"streaming":[3],"nature,":[4],"recurrent":[5],"neural":[6],"network":[7],"transducer":[8],"(RNN-T)":[9],"is":[10,96,130],"a":[11,70,92],"very":[12,71],"promising":[13],"end-to-end":[14],"(E2E)":[15],"model":[16,23,68,75],"that":[17,117],"may":[18],"replace":[19],"the":[20,65,112,131],"popular":[21],"hybrid":[22,74],"for":[24,98],"automatic":[25],"speech":[26],"recognition.In":[27],"this":[28],"paper,":[29],"we":[30,115],"describe":[31],"our":[32],"recent":[33],"development":[34],"RNN-T":[36,67,89],"models":[37,90,101],"with":[38,52,56,76],"reduced":[39],"GPU":[40],"memory":[41],"consumption":[42],"during":[43],"training,":[44],"better":[45,78],"initialization":[46],"strategy,":[47],"and":[48,81,121],"advanced":[49],"encoder":[50],"modeling":[51],"future":[53],"lookahead.When":[54],"trained":[55,73],"Microsoft's":[57],"65":[58],"thousand":[59],"hours":[60],"anonymized":[62],"training":[63],"data,":[64],"developed":[66],"surpasses":[69],"well":[72],"both":[77],"recognition":[79],"accuracy":[80],"lower":[82],"latency.We":[83],"further":[84],"study":[85],"how":[86],"to":[87,91,102],"customize":[88],"new":[93,113],"domain,":[94,114],"which":[95],"important":[97],"deploying":[99],"E2E":[100],"practical":[103],"scenarios.By":[104],"comparing":[105],"several":[106],"methods":[107],"leveraging":[108],"text-only":[109],"data":[110],"in":[111],"found":[116],"updating":[118],"RNN-T's":[119],"prediction":[120],"joint":[122],"networks":[123],"using":[124],"text-to-speech":[125],"generated":[126],"from":[127],"domain-specific":[128],"text":[129],"most":[132],"effective.":[133]},"counts_by_year":[{"year":2025,"cited_by_count":6},{"year":2024,"cited_by_count":7},{"year":2023,"cited_by_count":26},{"year":2022,"cited_by_count":20},{"year":2021,"cited_by_count":31},{"year":2020,"cited_by_count":7}],"updated_date":"2026-06-11T09:08:48.828518","created_date":"2025-10-10T00:00:00"}