{"id":"https://openalex.org/W4392903564","doi":"https://doi.org/10.1109/icassp48485.2024.10447520","title":"Multilingual Distilwhisper: Efficient Distillation of Multi-Task Speech Models Via Language-Specific Experts","display_name":"Multilingual Distilwhisper: Efficient Distillation of Multi-Task Speech Models Via Language-Specific Experts","publication_year":2024,"publication_date":"2024-03-18","ids":{"openalex":"https://openalex.org/W4392903564","doi":"https://doi.org/10.1109/icassp48485.2024.10447520"},"language":"en","primary_location":{"id":"doi:10.1109/icassp48485.2024.10447520","is_oa":false,"landing_page_url":"https://doi.org/10.1109/icassp48485.2024.10447520","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"ICASSP 2024 - 2024 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)","raw_type":"proceedings-article"},"type":"article","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://telecom-paris.hal.science/hal-04463381/document","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5068900854","display_name":"Thomas Palmeira Ferraz","orcid":"https://orcid.org/0000-0002-5385-9164"},"institutions":[{"id":"https://openalex.org/I60922564","display_name":"Naver (South Korea)","ror":"https://ror.org/04nzrnx83","country_code":"KR","type":"company","lineage":["https://openalex.org/I60922564"]}],"countries":["KR"],"is_corresponding":false,"raw_author_name":"Thomas Palmeira Ferraz","raw_affiliation_strings":["Naver Labs,Europe,France","Naver Labs, Europe, France"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Naver Labs,Europe,France","institution_ids":["https://openalex.org/I60922564"]},{"raw_affiliation_string":"Naver Labs, Europe, France","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5043744108","display_name":"Marcely Zanon Boito","orcid":"https://orcid.org/0000-0003-0134-6719"},"institutions":[{"id":"https://openalex.org/I60922564","display_name":"Naver (South Korea)","ror":"https://ror.org/04nzrnx83","country_code":"KR","type":"company","lineage":["https://openalex.org/I60922564"]}],"countries":["KR"],"is_corresponding":false,"raw_author_name":"Marcely Zanon Boito","raw_affiliation_strings":["Naver Labs,Europe,France","Naver Labs, Europe, France"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Naver Labs,Europe,France","institution_ids":["https://openalex.org/I60922564"]},{"raw_affiliation_string":"Naver Labs, Europe, France","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5105462436","display_name":"Caroline Brun","orcid":"https://orcid.org/0000-0003-2824-5221"},"institutions":[{"id":"https://openalex.org/I60922564","display_name":"Naver (South Korea)","ror":"https://ror.org/04nzrnx83","country_code":"KR","type":"company","lineage":["https://openalex.org/I60922564"]}],"countries":["KR"],"is_corresponding":false,"raw_author_name":"Caroline Brun","raw_affiliation_strings":["Naver Labs,Europe,France","Naver Labs, Europe, France"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Naver Labs,Europe,France","institution_ids":["https://openalex.org/I60922564"]},{"raw_affiliation_string":"Naver Labs, Europe, France","institution_ids":[]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5020441209","display_name":"Vassilina Nikoulina","orcid":null},"institutions":[{"id":"https://openalex.org/I60922564","display_name":"Naver (South Korea)","ror":"https://ror.org/04nzrnx83","country_code":"KR","type":"company","lineage":["https://openalex.org/I60922564"]}],"countries":["KR"],"is_corresponding":false,"raw_author_name":"Vassilina Nikoulina","raw_affiliation_strings":["Naver Labs,Europe,France","Naver Labs, Europe, France"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Naver Labs,Europe,France","institution_ids":["https://openalex.org/I60922564"]},{"raw_affiliation_string":"Naver Labs, Europe, France","institution_ids":[]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":4,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":2.7056,"has_fulltext":true,"cited_by_count":9,"citation_normalized_percentile":{"value":0.90939463,"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":"10716","last_page":"10720"},"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/T10028","display_name":"Topic Modeling","score":0.9979000091552734,"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":0.996399998664856,"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":0.8819432854652405},{"id":"https://openalex.org/keywords/modular-design","display_name":"Modular design","score":0.6469428539276123},{"id":"https://openalex.org/keywords/inference","display_name":"Inference","score":0.5958768129348755},{"id":"https://openalex.org/keywords/robustness","display_name":"Robustness (evolution)","score":0.5847313404083252},{"id":"https://openalex.org/keywords/language-model","display_name":"Language model","score":0.5190832614898682},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.5055496692657471},{"id":"https://openalex.org/keywords/task","display_name":"Task (project management)","score":0.4861114025115967},{"id":"https://openalex.org/keywords/boosting","display_name":"Boosting (machine learning)","score":0.4735117256641388},{"id":"https://openalex.org/keywords/overhead","display_name":"Overhead (engineering)","score":0.45565107464790344},{"id":"https://openalex.org/keywords/speech-recognition","display_name":"Speech recognition","score":0.44805479049682617},{"id":"https://openalex.org/keywords/transformer","display_name":"Transformer","score":0.439087450504303},{"id":"https://openalex.org/keywords/natural-language-processing","display_name":"Natural language processing","score":0.43172362446784973},{"id":"https://openalex.org/keywords/distillation","display_name":"Distillation","score":0.41676419973373413},{"id":"https://openalex.org/keywords/machine-learning","display_name":"Machine learning","score":0.3493419885635376},{"id":"https://openalex.org/keywords/programming-language","display_name":"Programming language","score":0.1546417772769928}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8819432854652405},{"id":"https://openalex.org/C101468663","wikidata":"https://www.wikidata.org/wiki/Q1620158","display_name":"Modular design","level":2,"score":0.6469428539276123},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.5958768129348755},{"id":"https://openalex.org/C63479239","wikidata":"https://www.wikidata.org/wiki/Q7353546","display_name":"Robustness (evolution)","level":3,"score":0.5847313404083252},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.5190832614898682},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5055496692657471},{"id":"https://openalex.org/C2780451532","wikidata":"https://www.wikidata.org/wiki/Q759676","display_name":"Task (project management)","level":2,"score":0.4861114025115967},{"id":"https://openalex.org/C46686674","wikidata":"https://www.wikidata.org/wiki/Q466303","display_name":"Boosting (machine learning)","level":2,"score":0.4735117256641388},{"id":"https://openalex.org/C2779960059","wikidata":"https://www.wikidata.org/wiki/Q7113681","display_name":"Overhead (engineering)","level":2,"score":0.45565107464790344},{"id":"https://openalex.org/C28490314","wikidata":"https://www.wikidata.org/wiki/Q189436","display_name":"Speech recognition","level":1,"score":0.44805479049682617},{"id":"https://openalex.org/C66322947","wikidata":"https://www.wikidata.org/wiki/Q11658","display_name":"Transformer","level":3,"score":0.439087450504303},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.43172362446784973},{"id":"https://openalex.org/C204030448","wikidata":"https://www.wikidata.org/wiki/Q101017","display_name":"Distillation","level":2,"score":0.41676419973373413},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.3493419885635376},{"id":"https://openalex.org/C199360897","wikidata":"https://www.wikidata.org/wiki/Q9143","display_name":"Programming language","level":1,"score":0.1546417772769928},{"id":"https://openalex.org/C55493867","wikidata":"https://www.wikidata.org/wiki/Q7094","display_name":"Biochemistry","level":1,"score":0.0},{"id":"https://openalex.org/C62520636","wikidata":"https://www.wikidata.org/wiki/Q944","display_name":"Quantum mechanics","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/C187736073","wikidata":"https://www.wikidata.org/wiki/Q2920921","display_name":"Management","level":1,"score":0.0},{"id":"https://openalex.org/C104317684","wikidata":"https://www.wikidata.org/wiki/Q7187","display_name":"Gene","level":2,"score":0.0},{"id":"https://openalex.org/C178790620","wikidata":"https://www.wikidata.org/wiki/Q11351","display_name":"Organic chemistry","level":1,"score":0.0},{"id":"https://openalex.org/C121332964","wikidata":"https://www.wikidata.org/wiki/Q413","display_name":"Physics","level":0,"score":0.0},{"id":"https://openalex.org/C185592680","wikidata":"https://www.wikidata.org/wiki/Q2329","display_name":"Chemistry","level":0,"score":0.0},{"id":"https://openalex.org/C165801399","wikidata":"https://www.wikidata.org/wiki/Q25428","display_name":"Voltage","level":2,"score":0.0}],"mesh":[],"locations_count":2,"locations":[{"id":"doi:10.1109/icassp48485.2024.10447520","is_oa":false,"landing_page_url":"https://doi.org/10.1109/icassp48485.2024.10447520","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"ICASSP 2024 - 2024 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)","raw_type":"proceedings-article"},{"id":"pmh:oai:HAL:hal-04463381v1","is_oa":true,"landing_page_url":"https://telecom-paris.hal.science/hal-04463381","pdf_url":"https://telecom-paris.hal.science/hal-04463381/document","source":{"id":"https://openalex.org/S4306402512","display_name":"HAL (Le Centre pour la Communication Scientifique Directe)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I1294671590","host_organization_name":"Centre National de la Recherche Scientifique","host_organization_lineage":["https://openalex.org/I1294671590"],"host_organization_lineage_names":[],"type":"repository"},"license":"other-oa","license_id":"https://openalex.org/licenses/other-oa","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":"ICASSP 2024 - 2024 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), IEEE, Apr 2024, Seoul (Korea), South Korea","raw_type":"Conference papers"}],"best_oa_location":{"id":"pmh:oai:HAL:hal-04463381v1","is_oa":true,"landing_page_url":"https://telecom-paris.hal.science/hal-04463381","pdf_url":"https://telecom-paris.hal.science/hal-04463381/document","source":{"id":"https://openalex.org/S4306402512","display_name":"HAL (Le Centre pour la Communication Scientifique Directe)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I1294671590","host_organization_name":"Centre National de la Recherche Scientifique","host_organization_lineage":["https://openalex.org/I1294671590"],"host_organization_lineage_names":[],"type":"repository"},"license":"other-oa","license_id":"https://openalex.org/licenses/other-oa","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":"ICASSP 2024 - 2024 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), IEEE, Apr 2024, Seoul (Korea), South Korea","raw_type":"Conference papers"},"sustainable_development_goals":[],"awards":[{"id":"https://openalex.org/G1982687888","display_name":"Bias, fairness and ethics of compressed NLP models","funder_award_id":"ANR-21-CE23-0026","funder_id":"https://openalex.org/F4320320883","funder_display_name":"Agence Nationale de la Recherche"}],"funders":[{"id":"https://openalex.org/F4320320883","display_name":"Agence Nationale de la Recherche","ror":"https://ror.org/00rbzpz17"}],"has_content":{"grobid_xml":false,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4392903564.pdf"},"referenced_works_count":40,"referenced_works":["https://openalex.org/W2958953787","https://openalex.org/W2978017171","https://openalex.org/W2979826702","https://openalex.org/W3035390927","https://openalex.org/W3036601975","https://openalex.org/W3172698324","https://openalex.org/W3185293939","https://openalex.org/W3198429080","https://openalex.org/W3201225328","https://openalex.org/W3209059054","https://openalex.org/W3209984917","https://openalex.org/W3213029956","https://openalex.org/W4225274946","https://openalex.org/W4285158119","https://openalex.org/W4287890953","https://openalex.org/W4307322847","https://openalex.org/W4319862635","https://openalex.org/W4321348716","https://openalex.org/W4323066294","https://openalex.org/W4323066695","https://openalex.org/W4378105483","https://openalex.org/W4385570255","https://openalex.org/W4385570643","https://openalex.org/W4385573385","https://openalex.org/W4385807419","https://openalex.org/W4385823097","https://openalex.org/W4390874621","https://openalex.org/W6638523607","https://openalex.org/W6765469073","https://openalex.org/W6768851824","https://openalex.org/W6771467084","https://openalex.org/W6780218876","https://openalex.org/W6790547227","https://openalex.org/W6796581206","https://openalex.org/W6846004400","https://openalex.org/W6847363464","https://openalex.org/W6850218400","https://openalex.org/W6850321100","https://openalex.org/W6850360519","https://openalex.org/W6852909395"],"related_works":["https://openalex.org/W4288365749","https://openalex.org/W2936497627","https://openalex.org/W3013624417","https://openalex.org/W4287826556","https://openalex.org/W3098382480","https://openalex.org/W4287598411","https://openalex.org/W3100913109","https://openalex.org/W3198458223","https://openalex.org/W3126642501","https://openalex.org/W2964413124"],"abstract_inverted_index":{"Whisper":[0],"is":[1,119],"a":[2,20,32,38,144],"multitask":[3,69,110],"and":[4,70,88,111,137],"multilingual":[5,71,112],"speech":[6,15],"model":[7,28,43],"covering":[8],"99":[9],"languages.":[10],"It":[11],"yields":[12],"commendable":[13],"automatic":[14],"recognition":[16],"(ASR)":[17],"results":[18],"in":[19,41,59,130],"subset":[21],"of":[22,35,68,83],"its":[23],"covered":[24],"languages,":[25,37],"but":[26],"the":[27,56,66,105,109,131],"still":[29],"underperforms":[30],"on":[31],"non-negligible":[33],"number":[34],"under-represented":[36],"problem":[39],"exacerbated":[40],"smaller":[42],"versions.":[44],"In":[45],"this":[46],"work,":[47],"we":[48],"propose":[49],"DistilWhisper,":[50],"an":[51],"approach":[52,74,95,118],"able":[53],"to":[54,98],"bridge":[55],"performance":[57,102,129],"gap":[58],"ASR":[60,81,101],"for":[61,134],"these":[62],"languages":[63,133],"while":[64,103,141],"retaining":[65],"advantages":[67],"capabilities.":[72],"Our":[73],"involves":[75],"two":[76],"key":[77],"strategies:":[78],"lightweight":[79],"modular":[80],"fine-tuning":[82,124],"whisper-small":[84],"using":[85],"language-specific":[86],"experts,":[87],"knowledge":[89],"distillation":[90],"from":[91,108],"whisper-large-v2.":[92],"This":[93],"dual":[94],"allows":[96],"us":[97],"effectively":[99],"boost":[100],"keeping":[104],"robustness":[106],"inherited":[107],"pre-training.":[113],"Results":[114],"demonstrate":[115],"that":[116],"our":[117],"more":[120],"effective":[121],"than":[122],"standard":[123],"or":[125],"LoRA":[126],"adapters,":[127],"boosting":[128],"targeted":[132],"both":[135],"in-":[136],"out-of-domain":[138],"test":[139],"sets,":[140],"introducing":[142],"only":[143],"negligible":[145],"parameter":[146],"overhead":[147],"at":[148],"inference.":[149]},"counts_by_year":[{"year":2026,"cited_by_count":1},{"year":2025,"cited_by_count":6},{"year":2024,"cited_by_count":2}],"updated_date":"2026-06-20T22:02:38.213706","created_date":"2025-10-10T00:00:00"}