{"id":"https://openalex.org/W4406858434","doi":"https://doi.org/10.1109/apsipaasc63619.2025.10849163","title":"Block Refinement Learning for Improving Early Exit in Autoregressive ASR","display_name":"Block Refinement Learning for Improving Early Exit in Autoregressive ASR","publication_year":2024,"publication_date":"2024-12-03","ids":{"openalex":"https://openalex.org/W4406858434","doi":"https://doi.org/10.1109/apsipaasc63619.2025.10849163"},"language":"en","primary_location":{"id":"doi:10.1109/apsipaasc63619.2025.10849163","is_oa":false,"landing_page_url":"https://doi.org/10.1109/apsipaasc63619.2025.10849163","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2024 Asia Pacific Signal and Information Processing Association Annual Summit and Conference (APSIPA ASC)","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/A5016817412","display_name":"Naotaka Kawata","orcid":null},"institutions":[{"id":"https://openalex.org/I2251713219","display_name":"NTT (Japan)","ror":"https://ror.org/00berct97","country_code":"JP","type":"company","lineage":["https://openalex.org/I2251713219"]}],"countries":["JP"],"is_corresponding":true,"raw_author_name":"Naotaka Kawata","raw_affiliation_strings":["NTT Corporation,NTT Human Informatics Laboratories,Japan"],"affiliations":[{"raw_affiliation_string":"NTT Corporation,NTT Human Informatics Laboratories,Japan","institution_ids":["https://openalex.org/I2251713219"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5013009514","display_name":"Shota Orihashi","orcid":null},"institutions":[{"id":"https://openalex.org/I2251713219","display_name":"NTT (Japan)","ror":"https://ror.org/00berct97","country_code":"JP","type":"company","lineage":["https://openalex.org/I2251713219"]}],"countries":["JP"],"is_corresponding":false,"raw_author_name":"Shota Orihashi","raw_affiliation_strings":["NTT Corporation,NTT Human Informatics Laboratories,Japan"],"affiliations":[{"raw_affiliation_string":"NTT Corporation,NTT Human Informatics Laboratories,Japan","institution_ids":["https://openalex.org/I2251713219"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5063110877","display_name":"Satoshi Suzuki","orcid":"https://orcid.org/0000-0002-1423-3767"},"institutions":[{"id":"https://openalex.org/I2251713219","display_name":"NTT (Japan)","ror":"https://ror.org/00berct97","country_code":"JP","type":"company","lineage":["https://openalex.org/I2251713219"]}],"countries":["JP"],"is_corresponding":false,"raw_author_name":"Satoshi Suzuki","raw_affiliation_strings":["NTT Corporation,NTT Human Informatics Laboratories,Japan"],"affiliations":[{"raw_affiliation_string":"NTT Corporation,NTT Human Informatics Laboratories,Japan","institution_ids":["https://openalex.org/I2251713219"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5009756622","display_name":"Tomohiro Tanaka","orcid":"https://orcid.org/0000-0001-7442-4912"},"institutions":[{"id":"https://openalex.org/I2251713219","display_name":"NTT (Japan)","ror":"https://ror.org/00berct97","country_code":"JP","type":"company","lineage":["https://openalex.org/I2251713219"]}],"countries":["JP"],"is_corresponding":false,"raw_author_name":"Tomohiro Tanaka","raw_affiliation_strings":["NTT Corporation,NTT Human Informatics Laboratories,Japan"],"affiliations":[{"raw_affiliation_string":"NTT Corporation,NTT Human Informatics Laboratories,Japan","institution_ids":["https://openalex.org/I2251713219"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5027118468","display_name":"Mana Ihori","orcid":null},"institutions":[{"id":"https://openalex.org/I2251713219","display_name":"NTT (Japan)","ror":"https://ror.org/00berct97","country_code":"JP","type":"company","lineage":["https://openalex.org/I2251713219"]}],"countries":["JP"],"is_corresponding":false,"raw_author_name":"Mana Ihori","raw_affiliation_strings":["NTT Corporation,NTT Human Informatics Laboratories,Japan"],"affiliations":[{"raw_affiliation_string":"NTT Corporation,NTT Human Informatics Laboratories,Japan","institution_ids":["https://openalex.org/I2251713219"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5116046293","display_name":"Naoki Maikishima","orcid":null},"institutions":[{"id":"https://openalex.org/I2251713219","display_name":"NTT (Japan)","ror":"https://ror.org/00berct97","country_code":"JP","type":"company","lineage":["https://openalex.org/I2251713219"]}],"countries":["JP"],"is_corresponding":false,"raw_author_name":"Naoki Maikishima","raw_affiliation_strings":["NTT Corporation,NTT Human Informatics Laboratories,Japan"],"affiliations":[{"raw_affiliation_string":"NTT Corporation,NTT Human Informatics Laboratories,Japan","institution_ids":["https://openalex.org/I2251713219"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5102511251","display_name":"Taiga Yamane","orcid":null},"institutions":[{"id":"https://openalex.org/I2251713219","display_name":"NTT (Japan)","ror":"https://ror.org/00berct97","country_code":"JP","type":"company","lineage":["https://openalex.org/I2251713219"]}],"countries":["JP"],"is_corresponding":false,"raw_author_name":"Taiga Yamane","raw_affiliation_strings":["NTT Corporation,NTT Human Informatics Laboratories,Japan"],"affiliations":[{"raw_affiliation_string":"NTT Corporation,NTT Human Informatics Laboratories,Japan","institution_ids":["https://openalex.org/I2251713219"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5060644399","display_name":"Ryo Masumura","orcid":"https://orcid.org/0000-0002-2415-4149"},"institutions":[{"id":"https://openalex.org/I2251713219","display_name":"NTT (Japan)","ror":"https://ror.org/00berct97","country_code":"JP","type":"company","lineage":["https://openalex.org/I2251713219"]}],"countries":["JP"],"is_corresponding":false,"raw_author_name":"Ryo Masumura","raw_affiliation_strings":["NTT Corporation,NTT Human Informatics Laboratories,Japan"],"affiliations":[{"raw_affiliation_string":"NTT Corporation,NTT Human Informatics Laboratories,Japan","institution_ids":["https://openalex.org/I2251713219"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":8,"corresponding_author_ids":["https://openalex.org/A5016817412"],"corresponding_institution_ids":["https://openalex.org/I2251713219"],"apc_list":null,"apc_paid":null,"fwci":0.0,"has_fulltext":false,"cited_by_count":0,"citation_normalized_percentile":{"value":0.32736768,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"1","last_page":"6"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10876","display_name":"Fault Detection and Control Systems","score":0.8876000046730042,"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":0.8876000046730042,"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"}},{"id":"https://openalex.org/T11447","display_name":"Blind Source Separation Techniques","score":0.8263000249862671,"subfield":{"id":"https://openalex.org/subfields/1711","display_name":"Signal Processing"},"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/autoregressive-model","display_name":"Autoregressive model","score":0.8466599583625793},{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.7414036989212036},{"id":"https://openalex.org/keywords/block","display_name":"Block (permutation group theory)","score":0.6231780052185059},{"id":"https://openalex.org/keywords/speech-recognition","display_name":"Speech recognition","score":0.47335100173950195},{"id":"https://openalex.org/keywords/star-model","display_name":"STAR model","score":0.45763447880744934},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.4203832447528839},{"id":"https://openalex.org/keywords/algorithm","display_name":"Algorithm","score":0.3726944923400879},{"id":"https://openalex.org/keywords/machine-learning","display_name":"Machine learning","score":0.2980891466140747},{"id":"https://openalex.org/keywords/autoregressive-integrated-moving-average","display_name":"Autoregressive integrated moving average","score":0.222543865442276},{"id":"https://openalex.org/keywords/time-series","display_name":"Time series","score":0.18680593371391296},{"id":"https://openalex.org/keywords/econometrics","display_name":"Econometrics","score":0.18638932704925537},{"id":"https://openalex.org/keywords/mathematics","display_name":"Mathematics","score":0.13036945462226868}],"concepts":[{"id":"https://openalex.org/C159877910","wikidata":"https://www.wikidata.org/wiki/Q2202883","display_name":"Autoregressive model","level":2,"score":0.8466599583625793},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7414036989212036},{"id":"https://openalex.org/C2777210771","wikidata":"https://www.wikidata.org/wiki/Q4927124","display_name":"Block (permutation group theory)","level":2,"score":0.6231780052185059},{"id":"https://openalex.org/C28490314","wikidata":"https://www.wikidata.org/wiki/Q189436","display_name":"Speech recognition","level":1,"score":0.47335100173950195},{"id":"https://openalex.org/C194657046","wikidata":"https://www.wikidata.org/wiki/Q7394685","display_name":"STAR model","level":4,"score":0.45763447880744934},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.4203832447528839},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.3726944923400879},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.2980891466140747},{"id":"https://openalex.org/C24338571","wikidata":"https://www.wikidata.org/wiki/Q2566298","display_name":"Autoregressive integrated moving average","level":3,"score":0.222543865442276},{"id":"https://openalex.org/C151406439","wikidata":"https://www.wikidata.org/wiki/Q186588","display_name":"Time series","level":2,"score":0.18680593371391296},{"id":"https://openalex.org/C149782125","wikidata":"https://www.wikidata.org/wiki/Q160039","display_name":"Econometrics","level":1,"score":0.18638932704925537},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.13036945462226868},{"id":"https://openalex.org/C2524010","wikidata":"https://www.wikidata.org/wiki/Q8087","display_name":"Geometry","level":1,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/apsipaasc63619.2025.10849163","is_oa":false,"landing_page_url":"https://doi.org/10.1109/apsipaasc63619.2025.10849163","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2024 Asia Pacific Signal and Information Processing Association Annual Summit and Conference (APSIPA ASC)","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":30,"referenced_works":["https://openalex.org/W2126105956","https://openalex.org/W2183341477","https://openalex.org/W2936774411","https://openalex.org/W2962677625","https://openalex.org/W2998183051","https://openalex.org/W3019321633","https://openalex.org/W3035038672","https://openalex.org/W3098576111","https://openalex.org/W3154971029","https://openalex.org/W3196295870","https://openalex.org/W3196733780","https://openalex.org/W4200216396","https://openalex.org/W4224918069","https://openalex.org/W4225285622","https://openalex.org/W4372347449","https://openalex.org/W4385571751","https://openalex.org/W4389523983","https://openalex.org/W4401597536","https://openalex.org/W6601563604","https://openalex.org/W6638523607","https://openalex.org/W6638632666","https://openalex.org/W6739901393","https://openalex.org/W6746023985","https://openalex.org/W6751979845","https://openalex.org/W6767164110","https://openalex.org/W6779313456","https://openalex.org/W6789394017","https://openalex.org/W6796551075","https://openalex.org/W6839719377","https://openalex.org/W6840061620"],"related_works":["https://openalex.org/W2439807930","https://openalex.org/W2009692134","https://openalex.org/W1972271943","https://openalex.org/W2019155478","https://openalex.org/W2024529895","https://openalex.org/W2168175994","https://openalex.org/W1902630399","https://openalex.org/W2120434453","https://openalex.org/W3120578569","https://openalex.org/W1487412319"],"abstract_inverted_index":{"While":[0],"the":[1,13,33,41,71,91,94,105,114,134,162,168,172,199,202,208],"autoregressive":[2],"transformer":[3],"models":[4],"of":[5,19,43,93,107,118,136,156,171,177,201,210],"automatic":[6],"speech":[7],"recognition":[8],"(ASR)":[9],"are":[10],"highly":[11],"accurate,":[12],"inference":[14,34,51,63,84,128],"time":[15],"is":[16,25,57,74,87,124,139,152],"long":[17],"because":[18,111,133],"their":[20],"sequential":[21],"decoding.":[22],"Early":[23],"exit":[24,100],"a":[26,54,121,153],"technique":[27],"that":[28,68],"aims":[29],"to":[30,77,82,89,126],"speed":[31],"up":[32],"process":[35],"by":[36,174],"terminating":[37],"it":[38,86,123],"early":[39,99],"on":[40,103,212],"basis":[42],"output":[44],"from":[45],"intermediate":[46,61,96,109,131,164,181,186,192,204],"decoder":[47],"layers.":[48,97,187,205],"When":[49],"an":[50],"result":[52],"with":[53],"high-confidence":[55],"token":[56],"obtained":[58],"in":[59,129],"low":[60,95,130,163,180,191],"layers,":[62],"can":[64,194],"be":[65,195],"terminated":[66],"at":[67],"point":[69],"and":[70],"computational":[72],"complexity":[73],"reduced":[75],"compared":[76],"not":[78,178],"terminating.":[79],"In":[80,188],"order":[81],"terminate":[83,127],"early,":[85],"necessary":[88],"improve":[90],"accuracy":[92,106,116,170,200],"However,":[98],"often":[101],"focuses":[102],"improving":[104],"high":[108,185,203],"layers":[110,132,138,165,182,193],"they":[112],"determine":[113],"upper":[115],"limit":[117],"ASR.":[119],"As":[120],"result,":[122],"difficult":[125],"confidence":[135],"those":[137],"low.":[140],"To":[141],"solve":[142],"this":[143,189],"problem,":[144],"we":[145],"propose":[146],"block":[147],"refinement":[148],"learning":[149],"(BRL),":[150],"which":[151],"re-training":[154],"method":[155],"existing":[157],"early-exiting":[158],"models.":[159],"BRL":[160,211],"trains":[161],"while":[166,197],"maintaining":[167,198],"overall":[169],"model":[173],"considering":[175],"gradients":[176],"only":[179],"but":[183],"also":[184],"way,":[190],"trained":[196],"We":[206],"demonstrated":[207],"effectiveness":[209],"Japanese":[213],"discourse":[214],"ASR":[215],"tasks.":[216]},"counts_by_year":[],"updated_date":"2025-12-27T23:08:20.325037","created_date":"2025-10-10T00:00:00"}