{"id":"https://openalex.org/W4372270753","doi":"https://doi.org/10.1109/icassp49357.2023.10096744","title":"Robust Knowledge Distillation from RNN-T Models with Noisy Training Labels Using Full-Sum Loss","display_name":"Robust Knowledge Distillation from RNN-T Models with Noisy Training Labels Using Full-Sum Loss","publication_year":2023,"publication_date":"2023-05-05","ids":{"openalex":"https://openalex.org/W4372270753","doi":"https://doi.org/10.1109/icassp49357.2023.10096744"},"language":"en","primary_location":{"id":"doi:10.1109/icassp49357.2023.10096744","is_oa":false,"landing_page_url":"https://doi.org/10.1109/icassp49357.2023.10096744","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"ICASSP 2023 - 2023 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)","raw_type":"proceedings-article"},"type":"conference-paper","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/A5060987315","display_name":"Mohammad Zeineldeen","orcid":null},"institutions":[{"id":"https://openalex.org/I887968799","display_name":"RWTH Aachen University","ror":"https://ror.org/04xfq0f34","country_code":"DE","type":"education","lineage":["https://openalex.org/I887968799"]}],"countries":["DE"],"is_corresponding":false,"raw_author_name":"Mohammad Zeineldeen","raw_affiliation_strings":["RWTH Aachen University,Human Language Technology and Pattern Recognition,Computer Science Department,Aachen,Germany,52074"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"RWTH Aachen University,Human Language Technology and Pattern Recognition,Computer Science Department,Aachen,Germany,52074","institution_ids":["https://openalex.org/I887968799"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5015927589","display_name":"Kartik Audhkhasi","orcid":"https://orcid.org/0000-0002-2340-1144"},"institutions":[{"id":"https://openalex.org/I1291425158","display_name":"Google (United States)","ror":"https://ror.org/00njsd438","country_code":"US","type":"company","lineage":["https://openalex.org/I1291425158","https://openalex.org/I4210128969"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Kartik Audhkhasi","raw_affiliation_strings":["Google LLC,New York","Google LLC, New York"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Google LLC,New York","institution_ids":["https://openalex.org/I1291425158"]},{"raw_affiliation_string":"Google LLC, New York","institution_ids":["https://openalex.org/I1291425158"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5064113992","display_name":"Murali Karthick Baskar","orcid":"https://orcid.org/0000-0002-9396-6319"},"institutions":[{"id":"https://openalex.org/I1291425158","display_name":"Google (United States)","ror":"https://ror.org/00njsd438","country_code":"US","type":"company","lineage":["https://openalex.org/I1291425158","https://openalex.org/I4210128969"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Murali Karthick Baskar","raw_affiliation_strings":["Google LLC,New York","Google LLC, New York"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Google LLC,New York","institution_ids":["https://openalex.org/I1291425158"]},{"raw_affiliation_string":"Google LLC, New York","institution_ids":["https://openalex.org/I1291425158"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5071715737","display_name":"Bhuvana Ramabhadran","orcid":null},"institutions":[{"id":"https://openalex.org/I1291425158","display_name":"Google (United States)","ror":"https://ror.org/00njsd438","country_code":"US","type":"company","lineage":["https://openalex.org/I1291425158","https://openalex.org/I4210128969"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Bhuvana Ramabhadran","raw_affiliation_strings":["Google LLC,New York","Google LLC, New York"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Google LLC,New York","institution_ids":["https://openalex.org/I1291425158"]},{"raw_affiliation_string":"Google LLC, New York","institution_ids":["https://openalex.org/I1291425158"]}]}],"institutions":[],"countries_distinct_count":2,"institutions_distinct_count":2,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":false,"cited_by_count":2,"citation_normalized_percentile":null,"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"1","last_page":"5"},"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":0.9972000122070312,"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":0.9958000183105469,"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/distillation","display_name":"Distillation","score":0.8361009359359741},{"id":"https://openalex.org/keywords/discriminative-model","display_name":"Discriminative model","score":0.7071077227592468},{"id":"https://openalex.org/keywords/recurrent-neural-network","display_name":"Recurrent neural network","score":0.6917333006858826},{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.6755945086479187},{"id":"https://openalex.org/keywords/sequence","display_name":"Sequence (biology)","score":0.5518175959587097},{"id":"https://openalex.org/keywords/artificial-neural-network","display_name":"Artificial neural network","score":0.52222740650177},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.5074674487113953},{"id":"https://openalex.org/keywords/machine-learning","display_name":"Machine learning","score":0.49903273582458496},{"id":"https://openalex.org/keywords/speech-recognition","display_name":"Speech recognition","score":0.33209991455078125},{"id":"https://openalex.org/keywords/chemistry","display_name":"Chemistry","score":0.08518141508102417},{"id":"https://openalex.org/keywords/chromatography","display_name":"Chromatography","score":0.08002403378486633}],"concepts":[{"id":"https://openalex.org/C204030448","wikidata":"https://www.wikidata.org/wiki/Q101017","display_name":"Distillation","level":2,"score":0.8361009359359741},{"id":"https://openalex.org/C97931131","wikidata":"https://www.wikidata.org/wiki/Q5282087","display_name":"Discriminative model","level":2,"score":0.7071077227592468},{"id":"https://openalex.org/C147168706","wikidata":"https://www.wikidata.org/wiki/Q1457734","display_name":"Recurrent neural network","level":3,"score":0.6917333006858826},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6755945086479187},{"id":"https://openalex.org/C2778112365","wikidata":"https://www.wikidata.org/wiki/Q3511065","display_name":"Sequence (biology)","level":2,"score":0.5518175959587097},{"id":"https://openalex.org/C50644808","wikidata":"https://www.wikidata.org/wiki/Q192776","display_name":"Artificial neural network","level":2,"score":0.52222740650177},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5074674487113953},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.49903273582458496},{"id":"https://openalex.org/C28490314","wikidata":"https://www.wikidata.org/wiki/Q189436","display_name":"Speech recognition","level":1,"score":0.33209991455078125},{"id":"https://openalex.org/C185592680","wikidata":"https://www.wikidata.org/wiki/Q2329","display_name":"Chemistry","level":0,"score":0.08518141508102417},{"id":"https://openalex.org/C43617362","wikidata":"https://www.wikidata.org/wiki/Q170050","display_name":"Chromatography","level":1,"score":0.08002403378486633},{"id":"https://openalex.org/C55493867","wikidata":"https://www.wikidata.org/wiki/Q7094","display_name":"Biochemistry","level":1,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/icassp49357.2023.10096744","is_oa":false,"landing_page_url":"https://doi.org/10.1109/icassp49357.2023.10096744","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"ICASSP 2023 - 2023 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)","raw_type":"proceedings-article"}],"best_oa_location":null,"sustainable_development_goals":[{"display_name":"Reduced inequalities","score":0.5699999928474426,"id":"https://metadata.un.org/sdg/10"}],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":29,"referenced_works":["https://openalex.org/W1494198834","https://openalex.org/W1821462560","https://openalex.org/W1828163288","https://openalex.org/W2125336414","https://openalex.org/W2127141656","https://openalex.org/W2617258110","https://openalex.org/W2892008152","https://openalex.org/W2914699162","https://openalex.org/W2936774411","https://openalex.org/W2963444698","https://openalex.org/W2995181338","https://openalex.org/W3097777922","https://openalex.org/W3139918052","https://openalex.org/W3162649911","https://openalex.org/W3166226836","https://openalex.org/W3203453034","https://openalex.org/W4224330094","https://openalex.org/W4226033575","https://openalex.org/W4287828713","https://openalex.org/W4295364049","https://openalex.org/W6638523607","https://openalex.org/W6638749077","https://openalex.org/W6678809451","https://openalex.org/W6728994542","https://openalex.org/W6757720242","https://openalex.org/W6774433549","https://openalex.org/W6791904447","https://openalex.org/W6795595584","https://openalex.org/W6810807250"],"related_works":["https://openalex.org/W4389116644","https://openalex.org/W2153315159","https://openalex.org/W3103844505","https://openalex.org/W259157601","https://openalex.org/W4205463238","https://openalex.org/W2761785940","https://openalex.org/W1482209366","https://openalex.org/W2110523656","https://openalex.org/W2521627374","https://openalex.org/W2981954115"],"abstract_inverted_index":{"This":[0],"work":[1],"studies":[2],"knowledge":[3,88,138],"distillation":[4,35,59,116,132],"(KD)":[5],"and":[6,156,158],"addresses":[7],"its":[8],"constraints":[9],"for":[10,118,122],"recurrent":[11],"neural":[12],"network":[13],"transducer":[14],"(RNN-T)":[15],"models.":[16],"In":[17,69],"hard":[18],"distillation,":[19,113],"a":[20,31,109,128],"teacher":[21,48,141],"model":[22],"transcribes":[23],"large":[24],"amounts":[25],"of":[26,46,54,80,103,130,139],"unlabelled":[27],"speech":[28],"to":[29,51,85,100,143],"train":[30],"student":[32],"model.":[33,49],"Soft":[34],"is":[36,67],"another":[37],"popular":[38],"KD":[39],"method":[40],"that":[41,108,133],"distills":[42,134],"the":[43,47,52,78,101,135,140],"output":[44],"logits":[45],"Due":[50],"nature":[53],"RNN-T":[55,119],"alignments,":[56],"applying":[57],"soft":[58],"between":[60],"RNNT":[61],"architectures":[62],"having":[63,73],"different":[64],"posterior":[65],"distributions":[66],"challenging.":[68],"addition,":[70],"bad":[71,123],"teachers":[72],"high":[74],"word-error-rate":[75],"(WER)":[76],"reduce":[77],"efficacy":[79],"KD.":[81],"We":[82,106,125,148],"investigate":[83],"how":[84],"effectively":[86],"distill":[87],"from":[89],"variable":[90],"quality":[91],"ASR":[92],"teachers,":[93],"which":[94],"has":[95],"not":[96],"been":[97],"studied":[98],"before":[99],"best":[102],"our":[104],"knowledge.":[105],"show":[107],"sequence-level":[110],"KD,":[111],"full-sum":[112,131],"outperforms":[114],"other":[115],"methods":[117],"models,":[120],"especially":[121],"teachers.":[124],"also":[126],"propose":[127],"variant":[129],"sequence":[136],"discriminative":[137],"leading":[142],"further":[144],"improvement":[145],"in":[146],"WER.":[147],"conduct":[149],"experiments":[150],"on":[151,159],"public":[152],"datasets":[153],"namely":[154],"SpeechStew":[155],"LibriSpeech,":[157],"in-house":[160],"production":[161],"data.":[162]},"counts_by_year":[{"year":2026,"cited_by_count":1},{"year":2024,"cited_by_count":1}],"updated_date":"2026-07-14T23:27:15.235271","created_date":"2025-10-10T00:00:00"}