{"id":"https://openalex.org/W4408355836","doi":"https://doi.org/10.1109/icassp49660.2025.10890266","title":"Knowledge Distillation Based Training of Unified Conformer CTC Models for Multi-form ASR","display_name":"Knowledge Distillation Based Training of Unified Conformer CTC Models for Multi-form ASR","publication_year":2025,"publication_date":"2025-03-12","ids":{"openalex":"https://openalex.org/W4408355836","doi":"https://doi.org/10.1109/icassp49660.2025.10890266"},"language":"en","primary_location":{"id":"doi:10.1109/icassp49660.2025.10890266","is_oa":false,"landing_page_url":"https://doi.org/10.1109/icassp49660.2025.10890266","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"ICASSP 2025 - 2025 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/A5102876935","display_name":"Takashi Fukuda","orcid":"https://orcid.org/0000-0002-4601-0955"},"institutions":[{"id":"https://openalex.org/I4210145865","display_name":"IBM Research - Tokyo","ror":"https://ror.org/04915qk43","country_code":"JP","type":"facility","lineage":["https://openalex.org/I1341412227","https://openalex.org/I4210114115","https://openalex.org/I4210145865"]}],"countries":["JP"],"is_corresponding":false,"raw_author_name":"Takashi Fukuda","raw_affiliation_strings":["IBM Research AI,Tokyo,Japan"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"IBM Research AI,Tokyo,Japan","institution_ids":["https://openalex.org/I4210145865"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5021294422","display_name":"Gakuto Kurata","orcid":null},"institutions":[{"id":"https://openalex.org/I4210145865","display_name":"IBM Research - Tokyo","ror":"https://ror.org/04915qk43","country_code":"JP","type":"facility","lineage":["https://openalex.org/I1341412227","https://openalex.org/I4210114115","https://openalex.org/I4210145865"]}],"countries":["JP"],"is_corresponding":false,"raw_author_name":"Gakuto Kurata","raw_affiliation_strings":["IBM Research AI,Tokyo,Japan"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"IBM Research AI,Tokyo,Japan","institution_ids":["https://openalex.org/I4210145865"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5079994647","display_name":"George Saon","orcid":"https://orcid.org/0009-0004-6837-5009"},"institutions":[{"id":"https://openalex.org/I1341412227","display_name":"IBM (United States)","ror":"https://ror.org/05hh8d621","country_code":"US","type":"company","lineage":["https://openalex.org/I1341412227"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"George Saon","raw_affiliation_strings":["IBM Research AI, IBM T. J.,Watson Research Center,NY,USA"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"IBM Research AI, IBM T. J.,Watson Research Center,NY,USA","institution_ids":["https://openalex.org/I1341412227"]}]}],"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":1,"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/T10876","display_name":"Fault Detection and Control Systems","score":0.9318000078201294,"subfield":{"id":"https://openalex.org/subfields/2207","display_name":"Control and Systems Engineering"},"field":{"id":"https://openalex.org/fields/22","display_name":"Engineering"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},"topics":[{"id":"https://openalex.org/T10876","display_name":"Fault Detection and Control Systems","score":0.9318000078201294,"subfield":{"id":"https://openalex.org/subfields/2207","display_name":"Control and Systems Engineering"},"field":{"id":"https://openalex.org/fields/22","display_name":"Engineering"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.6470173597335815},{"id":"https://openalex.org/keywords/distillation","display_name":"Distillation","score":0.5820899605751038},{"id":"https://openalex.org/keywords/training","display_name":"Training (meteorology)","score":0.5008111000061035},{"id":"https://openalex.org/keywords/conformational-isomerism","display_name":"Conformational isomerism","score":0.45582446455955505},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.36704957485198975},{"id":"https://openalex.org/keywords/natural-language-processing","display_name":"Natural language processing","score":0.32101374864578247},{"id":"https://openalex.org/keywords/chemistry","display_name":"Chemistry","score":0.12300476431846619},{"id":"https://openalex.org/keywords/organic-chemistry","display_name":"Organic chemistry","score":0.08802101016044617}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6470173597335815},{"id":"https://openalex.org/C204030448","wikidata":"https://www.wikidata.org/wiki/Q101017","display_name":"Distillation","level":2,"score":0.5820899605751038},{"id":"https://openalex.org/C2777211547","wikidata":"https://www.wikidata.org/wiki/Q17141490","display_name":"Training (meteorology)","level":2,"score":0.5008111000061035},{"id":"https://openalex.org/C18705241","wikidata":"https://www.wikidata.org/wiki/Q1128023","display_name":"Conformational isomerism","level":3,"score":0.45582446455955505},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.36704957485198975},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.32101374864578247},{"id":"https://openalex.org/C185592680","wikidata":"https://www.wikidata.org/wiki/Q2329","display_name":"Chemistry","level":0,"score":0.12300476431846619},{"id":"https://openalex.org/C178790620","wikidata":"https://www.wikidata.org/wiki/Q11351","display_name":"Organic chemistry","level":1,"score":0.08802101016044617},{"id":"https://openalex.org/C153294291","wikidata":"https://www.wikidata.org/wiki/Q25261","display_name":"Meteorology","level":1,"score":0.0},{"id":"https://openalex.org/C32909587","wikidata":"https://www.wikidata.org/wiki/Q11369","display_name":"Molecule","level":2,"score":0.0},{"id":"https://openalex.org/C121332964","wikidata":"https://www.wikidata.org/wiki/Q413","display_name":"Physics","level":0,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/icassp49660.2025.10890266","is_oa":false,"landing_page_url":"https://doi.org/10.1109/icassp49660.2025.10890266","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"ICASSP 2025 - 2025 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)","raw_type":"proceedings-article"}],"best_oa_location":null,"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":28,"referenced_works":["https://openalex.org/W2117486215","https://openalex.org/W2127141656","https://openalex.org/W2329068866","https://openalex.org/W2741256201","https://openalex.org/W2747909401","https://openalex.org/W2802623045","https://openalex.org/W2936774411","https://openalex.org/W2937657912","https://openalex.org/W2963587345","https://openalex.org/W2963882470","https://openalex.org/W2973229104","https://openalex.org/W3008898571","https://openalex.org/W3015194534","https://openalex.org/W3096518646","https://openalex.org/W3097777922","https://openalex.org/W3160766462","https://openalex.org/W3162249256","https://openalex.org/W3162649911","https://openalex.org/W3163300396","https://openalex.org/W3197148831","https://openalex.org/W3197898596","https://openalex.org/W3203453034","https://openalex.org/W4375869224","https://openalex.org/W4391021406","https://openalex.org/W6600213771","https://openalex.org/W6754005058","https://openalex.org/W6756974721","https://openalex.org/W6840509714"],"related_works":["https://openalex.org/W2023036309","https://openalex.org/W1972863456","https://openalex.org/W2089371831","https://openalex.org/W2052399476","https://openalex.org/W2810163160","https://openalex.org/W3083970380","https://openalex.org/W2037964119","https://openalex.org/W4385077663","https://openalex.org/W4220905048","https://openalex.org/W3204019825"],"abstract_inverted_index":{"There":[0],"is":[1,64],"an":[2,47,61,111],"on-going":[3],"body":[4],"of":[5,43,68,130,163],"research":[6],"on":[7,25,83,134],"training":[8,79],"separate":[9],"dedicated":[10],"models":[11,20,82,144],"for":[12,78,122],"either":[13],"short-form":[14,31,137],"or":[15],"long-form":[16,29],"utterances.":[17],"Multi-form":[18],"acoustic":[19],"that":[21,58,120,173],"are":[22],"simply":[23],"trained":[24,148],"combined":[26],"data":[27,86],"from":[28,35,60,88],"and":[30,49,91,115,136],"utterances":[32,151],"often":[33],"suffer":[34],"various":[36,153],"negative":[37],"impacts":[38],"due":[39],"to":[40,93,165],"the":[41,69,117,128],"diversity":[42],"a":[44,50,55,99,160,176],"speaking":[45],"style,":[46],"accent,":[48],"recording":[51],"condition.":[52],"In":[53,71],"addition,":[54],"linguistic":[56],"mismatch":[57],"comes":[59],"utterance":[62],"length":[63],"also":[65],"another":[66],"factor":[67],"degradation.":[70],"this":[72],"paper":[73],"we":[74],"investigate":[75],"novel":[76],"techniques":[77],"unified":[80,124],"Conformer-based":[81],"multi-form":[84],"speech":[85],"obtained":[87],"diverse":[89],"domains":[90],"sources":[92],"serve":[94],"multiple":[95],"downstream":[96],"applications":[97],"with":[98,110,152],"single":[100,123],"model.":[101],"Our":[102],"approach":[103],"incorporates":[104],"chunk-wise":[105],"short-term":[106],"discriminative":[107],"knowledge":[108],"distillation":[109],"encoder":[112],"embedding":[113],"masking":[114],"mitigates":[116],"aforementioned":[118],"problems":[119],"appear":[121],"models.":[125],"We":[126],"show":[127],"benefit":[129],"our":[131,143],"proposed":[132,157],"technique":[133,158],"long":[135],"ASR":[138],"test":[139],"sets":[140],"by":[141,149],"comparing":[142],"against":[145],"several":[146],"variants":[147],"mixing":[150],"audio":[154],"lengths.":[155],"The":[156],"provides":[159],"significant":[161],"improvement":[162],"up":[164],"8.5%":[166],"relative":[167],"WER":[168],"reduction":[169],"over":[170],"baseline":[171],"systems":[172],"operate":[174],"at":[175],"similar":[177],"decoding":[178],"cost.":[179]},"counts_by_year":[{"year":2026,"cited_by_count":1}],"updated_date":"2026-07-14T23:27:15.235271","created_date":"2025-10-10T00:00:00"}