{"id":"https://openalex.org/W3096297644","doi":"https://doi.org/10.21437/interspeech.2020-1179","title":"Distilling the Knowledge of BERT for Sequence-to-Sequence ASR","display_name":"Distilling the Knowledge of BERT for Sequence-to-Sequence ASR","publication_year":2020,"publication_date":"2020-10-25","ids":{"openalex":"https://openalex.org/W3096297644","doi":"https://doi.org/10.21437/interspeech.2020-1179","mag":"3096297644"},"language":"en","primary_location":{"id":"doi:10.21437/interspeech.2020-1179","is_oa":false,"landing_page_url":"https://doi.org/10.21437/interspeech.2020-1179","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":"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/A5034216052","display_name":"Hayato Futami","orcid":null},"institutions":[{"id":"https://openalex.org/I22299242","display_name":"Kyoto University","ror":"https://ror.org/02kpeqv85","country_code":"JP","type":"education","lineage":["https://openalex.org/I22299242"]}],"countries":["JP"],"is_corresponding":false,"raw_author_name":"Hayato Futami","raw_affiliation_strings":["Graduate School of Informatics, Kyoto University, Sakyo-ku, Kyoto, Japan"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Graduate School of Informatics, Kyoto University, Sakyo-ku, Kyoto, Japan","institution_ids":["https://openalex.org/I22299242"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5040282669","display_name":"Hirofumi Inaguma","orcid":"https://orcid.org/0000-0003-0610-1251"},"institutions":[{"id":"https://openalex.org/I22299242","display_name":"Kyoto University","ror":"https://ror.org/02kpeqv85","country_code":"JP","type":"education","lineage":["https://openalex.org/I22299242"]}],"countries":["JP"],"is_corresponding":false,"raw_author_name":"Hirofumi Inaguma","raw_affiliation_strings":["Graduate School of Informatics, Kyoto University, Sakyo-ku, Kyoto, Japan"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Graduate School of Informatics, Kyoto University, Sakyo-ku, Kyoto, Japan","institution_ids":["https://openalex.org/I22299242"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5103835523","display_name":"Sei Ueno","orcid":"https://orcid.org/0000-0002-2255-3154"},"institutions":[{"id":"https://openalex.org/I22299242","display_name":"Kyoto University","ror":"https://ror.org/02kpeqv85","country_code":"JP","type":"education","lineage":["https://openalex.org/I22299242"]}],"countries":["JP"],"is_corresponding":false,"raw_author_name":"Sei Ueno","raw_affiliation_strings":["Graduate School of Informatics, Kyoto University, Sakyo-ku, Kyoto, Japan"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Graduate School of Informatics, Kyoto University, Sakyo-ku, Kyoto, Japan","institution_ids":["https://openalex.org/I22299242"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5102851028","display_name":"Masato Mimura","orcid":"https://orcid.org/0000-0002-2403-0680"},"institutions":[{"id":"https://openalex.org/I22299242","display_name":"Kyoto University","ror":"https://ror.org/02kpeqv85","country_code":"JP","type":"education","lineage":["https://openalex.org/I22299242"]}],"countries":["JP"],"is_corresponding":false,"raw_author_name":"Masato Mimura","raw_affiliation_strings":["Graduate School of Informatics, Kyoto University, Sakyo-ku, Kyoto, Japan"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Graduate School of Informatics, Kyoto University, Sakyo-ku, Kyoto, Japan","institution_ids":["https://openalex.org/I22299242"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5110763152","display_name":"Shinsuke Sakai","orcid":null},"institutions":[{"id":"https://openalex.org/I22299242","display_name":"Kyoto University","ror":"https://ror.org/02kpeqv85","country_code":"JP","type":"education","lineage":["https://openalex.org/I22299242"]}],"countries":["JP"],"is_corresponding":false,"raw_author_name":"Shinsuke Sakai","raw_affiliation_strings":["Graduate School of Informatics, Kyoto University, Sakyo-ku, Kyoto, Japan"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Graduate School of Informatics, Kyoto University, Sakyo-ku, Kyoto, Japan","institution_ids":["https://openalex.org/I22299242"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5038044080","display_name":"Tatsuya Kawahara","orcid":"https://orcid.org/0000-0002-2686-2296"},"institutions":[{"id":"https://openalex.org/I22299242","display_name":"Kyoto University","ror":"https://ror.org/02kpeqv85","country_code":"JP","type":"education","lineage":["https://openalex.org/I22299242"]}],"countries":["JP"],"is_corresponding":false,"raw_author_name":"Tatsuya Kawahara","raw_affiliation_strings":["Graduate School of Informatics, Kyoto University, Sakyo-ku, Kyoto, Japan"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Graduate School of Informatics, Kyoto University, Sakyo-ku, Kyoto, Japan","institution_ids":["https://openalex.org/I22299242"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":["https://openalex.org/I22299242"],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":false,"cited_by_count":40,"citation_normalized_percentile":null,"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"3635","last_page":"3639"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10181","display_name":"Natural Language Processing Techniques","score":0.9990000128746033,"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/T10181","display_name":"Natural Language Processing Techniques","score":0.9990000128746033,"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":0.9987999796867371,"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/T10201","display_name":"Speech Recognition and Synthesis","score":0.9934999942779541,"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/sequence","display_name":"Sequence (biology)","score":0.7837306261062622},{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.7411082983016968},{"id":"https://openalex.org/keywords/natural-language-processing","display_name":"Natural language processing","score":0.37152111530303955},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.33834993839263916},{"id":"https://openalex.org/keywords/programming-language","display_name":"Programming language","score":0.3316161036491394},{"id":"https://openalex.org/keywords/genetics","display_name":"Genetics","score":0.05414685606956482}],"concepts":[{"id":"https://openalex.org/C2778112365","wikidata":"https://www.wikidata.org/wiki/Q3511065","display_name":"Sequence (biology)","level":2,"score":0.7837306261062622},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7411082983016968},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.37152111530303955},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.33834993839263916},{"id":"https://openalex.org/C199360897","wikidata":"https://www.wikidata.org/wiki/Q9143","display_name":"Programming language","level":1,"score":0.3316161036491394},{"id":"https://openalex.org/C54355233","wikidata":"https://www.wikidata.org/wiki/Q7162","display_name":"Genetics","level":1,"score":0.05414685606956482},{"id":"https://openalex.org/C86803240","wikidata":"https://www.wikidata.org/wiki/Q420","display_name":"Biology","level":0,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.21437/interspeech.2020-1179","is_oa":false,"landing_page_url":"https://doi.org/10.21437/interspeech.2020-1179","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":33,"referenced_works":["https://openalex.org/W648786980","https://openalex.org/W854541894","https://openalex.org/W1522301498","https://openalex.org/W1821462560","https://openalex.org/W1828163288","https://openalex.org/W1915251500","https://openalex.org/W2062988874","https://openalex.org/W2183341477","https://openalex.org/W2327501763","https://openalex.org/W2577366047","https://openalex.org/W2752047430","https://openalex.org/W2886025712","https://openalex.org/W2888779557","https://openalex.org/W2889504751","https://openalex.org/W2896457183","https://openalex.org/W2935811960","https://openalex.org/W2936774411","https://openalex.org/W2937649809","https://openalex.org/W2952650870","https://openalex.org/W2962739339","https://openalex.org/W2962784628","https://openalex.org/W2963362078","https://openalex.org/W2963827914","https://openalex.org/W2965373594","https://openalex.org/W2970597249","https://openalex.org/W2972991710","https://openalex.org/W2990391581","https://openalex.org/W2993326934","https://openalex.org/W3034729383","https://openalex.org/W3034775979","https://openalex.org/W3035317912","https://openalex.org/W4287824654","https://openalex.org/W4385245566"],"related_works":["https://openalex.org/W4391375266","https://openalex.org/W2899084033","https://openalex.org/W2748952813","https://openalex.org/W2390279801","https://openalex.org/W4391913857","https://openalex.org/W2358668433","https://openalex.org/W4396701345","https://openalex.org/W2376932109","https://openalex.org/W2001405890","https://openalex.org/W3204019825"],"abstract_inverted_index":{"Attention-based":[0],"sequence-to-sequence":[1],"(seq2seq)":[2],"models":[3,15],"have":[4,24],"achieved":[5],"promising":[6],"results":[7],"in":[8,17],"automatic":[9],"speech":[10],"recognition":[11],"(ASR).However,":[12],"as":[13,40,72,130],"these":[14],"decode":[16],"a":[18,103],"left-to-right":[19],"way,":[20],"they":[21],"do":[22],"not":[23,139],"access":[25],"to":[26,45,58,74],"context":[27,36,67,118],"on":[28,90],"the":[29,60,69,83,87,91,114,120],"right.We":[30],"leverage":[31,66],"both":[32],"left":[33,110],"and":[34,133],"right":[35],"by":[37],"applying":[38],"BERT":[39,54,99],"an":[41],"external":[42],"language":[43],"model":[44],"seq2seq":[46,63,88],"ASR":[47,84],"through":[48],"knowledge":[49],"distillation.In":[50],"our":[51,79],"proposed":[52],"method,":[53],"generates":[55],"soft":[56],"labels":[57],"guide":[59],"training":[61],"of":[62,93,116],"ASR.Furthermore,":[64],"we":[65],"beyond":[68,119],"current":[70,121],"utterance":[71],"input":[73],"BERT.Experimental":[75],"evaluations":[76],"show":[77,113],"that":[78,101,106],"method":[80,123],"significantly":[81],"improves":[82],"performance":[85],"from":[86,98,102],"baseline":[89],"Corpus":[92],"Spontaneous":[94],"Japanese":[95],"(CSJ).Knowledge":[96],"distillation":[97],"outperforms":[100,124],"transformer":[104],"LM":[105,126],"only":[107],"looks":[108],"at":[109],"context.We":[111],"also":[112],"effectiveness":[115],"leveraging":[117],"utterance.Our":[122],"other":[125],"application":[127],"approaches":[128],"such":[129],"n-best":[131],"rescoring":[132],"shallow":[134],"fusion,":[135],"while":[136],"it":[137],"does":[138],"require":[140],"extra":[141],"inference":[142],"cost.":[143]},"counts_by_year":[{"year":2026,"cited_by_count":2},{"year":2025,"cited_by_count":1},{"year":2024,"cited_by_count":7},{"year":2023,"cited_by_count":13},{"year":2022,"cited_by_count":8},{"year":2021,"cited_by_count":9}],"updated_date":"2026-07-14T23:27:15.235271","created_date":"2025-10-10T00:00:00"}