{"id":"https://openalex.org/W3160337317","doi":"https://doi.org/10.1109/icassp39728.2021.9413493","title":"Gaussian Kernelized Self-Attention for Long Sequence Data and its Application to CTC-Based Speech Recognition","display_name":"Gaussian Kernelized Self-Attention for Long Sequence Data and its Application to CTC-Based Speech Recognition","publication_year":2021,"publication_date":"2021-05-13","ids":{"openalex":"https://openalex.org/W3160337317","doi":"https://doi.org/10.1109/icassp39728.2021.9413493","mag":"3160337317"},"language":"en","primary_location":{"id":"doi:10.1109/icassp39728.2021.9413493","is_oa":false,"landing_page_url":"https://doi.org/10.1109/icassp39728.2021.9413493","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"ICASSP 2021 - 2021 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/A5109644502","display_name":"Yosuke Kashiwagi","orcid":null},"institutions":[{"id":"https://openalex.org/I1304132090","display_name":"Sony (Taiwan)","ror":"https://ror.org/0214y7014","country_code":"TW","type":"company","lineage":["https://openalex.org/I1304132090","https://openalex.org/I4210143797"]},{"id":"https://openalex.org/I2800278093","display_name":"Sony Corporation (United States)","ror":"https://ror.org/05k91zb11","country_code":"US","type":"company","lineage":["https://openalex.org/I2800278093"]},{"id":"https://openalex.org/I4210121096","display_name":"Dexerials (Japan)","ror":"https://ror.org/03p330m15","country_code":"JP","type":"company","lineage":["https://openalex.org/I4210121096"]}],"countries":["JP","TW","US"],"is_corresponding":false,"raw_author_name":"Yosuke Kashiwagi","raw_affiliation_strings":["Sony Corporation,Japan","Sony Corporation, Japan"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Sony Corporation,Japan","institution_ids":["https://openalex.org/I1304132090","https://openalex.org/I2800278093","https://openalex.org/I4210121096"]},{"raw_affiliation_string":"Sony Corporation, Japan","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5011937407","display_name":"Emiru Tsunoo","orcid":null},"institutions":[{"id":"https://openalex.org/I1304132090","display_name":"Sony (Taiwan)","ror":"https://ror.org/0214y7014","country_code":"TW","type":"company","lineage":["https://openalex.org/I1304132090","https://openalex.org/I4210143797"]},{"id":"https://openalex.org/I2800278093","display_name":"Sony Corporation (United States)","ror":"https://ror.org/05k91zb11","country_code":"US","type":"company","lineage":["https://openalex.org/I2800278093"]},{"id":"https://openalex.org/I4210121096","display_name":"Dexerials (Japan)","ror":"https://ror.org/03p330m15","country_code":"JP","type":"company","lineage":["https://openalex.org/I4210121096"]}],"countries":["JP","TW","US"],"is_corresponding":false,"raw_author_name":"Emiru Tsunoo","raw_affiliation_strings":["Sony Corporation,Japan","Sony Corporation, Japan"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Sony Corporation,Japan","institution_ids":["https://openalex.org/I1304132090","https://openalex.org/I2800278093","https://openalex.org/I4210121096"]},{"raw_affiliation_string":"Sony Corporation, Japan","institution_ids":[]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5001291873","display_name":"Shinji Watanabe","orcid":"https://orcid.org/0000-0002-5970-8631"},"institutions":[{"id":"https://openalex.org/I145311948","display_name":"Johns Hopkins University","ror":"https://ror.org/00za53h95","country_code":"US","type":"education","lineage":["https://openalex.org/I145311948"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Shinji Watanabe","raw_affiliation_strings":["Johns Hopkins University,USA","Johns Hopkins University, USA"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Johns Hopkins University,USA","institution_ids":["https://openalex.org/I145311948"]},{"raw_affiliation_string":"Johns Hopkins University, USA","institution_ids":["https://openalex.org/I145311948"]}]}],"institutions":[],"countries_distinct_count":3,"institutions_distinct_count":4,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":0.4396,"has_fulltext":false,"cited_by_count":3,"citation_normalized_percentile":{"value":0.59732294,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":{"min":90,"max":96},"biblio":{"volume":null,"issue":null,"first_page":"6214","last_page":"6218"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10201","display_name":"Speech Recognition and Synthesis","score":0.9998000264167786,"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":0.9998000264167786,"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/T11309","display_name":"Music and Audio Processing","score":0.9987999796867371,"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"}},{"id":"https://openalex.org/T10181","display_name":"Natural Language Processing Techniques","score":0.9950000047683716,"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":0.7397373914718628},{"id":"https://openalex.org/keywords/kernel","display_name":"Kernel (algebra)","score":0.616923451423645},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.5573644638061523},{"id":"https://openalex.org/keywords/pattern-recognition","display_name":"Pattern recognition (psychology)","score":0.5297858715057373},{"id":"https://openalex.org/keywords/speech-recognition","display_name":"Speech recognition","score":0.5276452302932739},{"id":"https://openalex.org/keywords/inference","display_name":"Inference","score":0.5075355768203735},{"id":"https://openalex.org/keywords/gaussian-function","display_name":"Gaussian function","score":0.49279072880744934},{"id":"https://openalex.org/keywords/gaussian","display_name":"Gaussian","score":0.46972692012786865},{"id":"https://openalex.org/keywords/search-engine-indexing","display_name":"Search engine indexing","score":0.41145333647727966},{"id":"https://openalex.org/keywords/invariant","display_name":"Invariant (physics)","score":0.4113362431526184},{"id":"https://openalex.org/keywords/mathematics","display_name":"Mathematics","score":0.1485157012939453}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7397373914718628},{"id":"https://openalex.org/C74193536","wikidata":"https://www.wikidata.org/wiki/Q574844","display_name":"Kernel (algebra)","level":2,"score":0.616923451423645},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5573644638061523},{"id":"https://openalex.org/C153180895","wikidata":"https://www.wikidata.org/wiki/Q7148389","display_name":"Pattern recognition (psychology)","level":2,"score":0.5297858715057373},{"id":"https://openalex.org/C28490314","wikidata":"https://www.wikidata.org/wiki/Q189436","display_name":"Speech recognition","level":1,"score":0.5276452302932739},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.5075355768203735},{"id":"https://openalex.org/C7218915","wikidata":"https://www.wikidata.org/wiki/Q1054475","display_name":"Gaussian function","level":3,"score":0.49279072880744934},{"id":"https://openalex.org/C163716315","wikidata":"https://www.wikidata.org/wiki/Q901177","display_name":"Gaussian","level":2,"score":0.46972692012786865},{"id":"https://openalex.org/C75165309","wikidata":"https://www.wikidata.org/wiki/Q2258979","display_name":"Search engine indexing","level":2,"score":0.41145333647727966},{"id":"https://openalex.org/C190470478","wikidata":"https://www.wikidata.org/wiki/Q2370229","display_name":"Invariant (physics)","level":2,"score":0.4113362431526184},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.1485157012939453},{"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/C114614502","wikidata":"https://www.wikidata.org/wiki/Q76592","display_name":"Combinatorics","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/C37914503","wikidata":"https://www.wikidata.org/wiki/Q156495","display_name":"Mathematical physics","level":1,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/icassp39728.2021.9413493","is_oa":false,"landing_page_url":"https://doi.org/10.1109/icassp39728.2021.9413493","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"ICASSP 2021 - 2021 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)","raw_type":"proceedings-article"}],"best_oa_location":null,"sustainable_development_goals":[{"score":0.46000000834465027,"display_name":"Quality Education","id":"https://metadata.un.org/sdg/4"}],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":38,"referenced_works":["https://openalex.org/W170242364","https://openalex.org/W2089372326","https://openalex.org/W2125378448","https://openalex.org/W2127141656","https://openalex.org/W2152624700","https://openalex.org/W2159549127","https://openalex.org/W2603621985","https://openalex.org/W2799473636","https://openalex.org/W2802023636","https://openalex.org/W2892009249","https://openalex.org/W2911291251","https://openalex.org/W2936774411","https://openalex.org/W2941814890","https://openalex.org/W2963066927","https://openalex.org/W2963403868","https://openalex.org/W2963925437","https://openalex.org/W2963979492","https://openalex.org/W2964089206","https://openalex.org/W2972451902","https://openalex.org/W2972818416","https://openalex.org/W2981857663","https://openalex.org/W2994673210","https://openalex.org/W3007328579","https://openalex.org/W3008191852","https://openalex.org/W3015834770","https://openalex.org/W3026709049","https://openalex.org/W3034729383","https://openalex.org/W3097030750","https://openalex.org/W3101648800","https://openalex.org/W4295838474","https://openalex.org/W4385245566","https://openalex.org/W6606829109","https://openalex.org/W6678556256","https://openalex.org/W6739901393","https://openalex.org/W6761563299","https://openalex.org/W6771626834","https://openalex.org/W6777191834","https://openalex.org/W6779469704"],"related_works":["https://openalex.org/W2375370983","https://openalex.org/W2369557298","https://openalex.org/W2545232906","https://openalex.org/W2011212036","https://openalex.org/W3044685708","https://openalex.org/W2065131986","https://openalex.org/W1545294102","https://openalex.org/W2127643120","https://openalex.org/W4229068138","https://openalex.org/W4321262185"],"abstract_inverted_index":{"ISelf-attention":[0],"(SA)":[1],"based":[2,151],"models":[3],"have":[4],"recently":[5],"achieved":[6],"significant":[7,174],"performance":[8],"improvements":[9],"in":[10,176,184,186],"hybrid":[11],"and":[12,53,104,163],"end-to-end":[13],"automatic":[14],"speech":[15],"recognition":[16],"(ASR)":[17],"systems":[18],"owing":[19],"to":[20,38,46,108,146,182],"their":[21],"flexible":[22],"context":[23],"modeling":[24],"capability.":[25],"However,":[26],"it":[27],"is":[28,43,79,88,106],"also":[29],"known":[30],"that":[31,96,168],"the":[32,47,51,57,83,119,123,129,157,169],"accuracy":[33,177],"degrades":[34],"when":[35],"applying":[36],"SA":[37,143,171],"long":[39,187],"sequence":[40,188],"data.":[41],"This":[42],"mainly":[44],"due":[45],"length":[48],"mismatch":[49],"between":[50],"inference":[52],"training":[54,58],"data":[55,59,189],"because":[56],"are":[60],"usually":[61],"divided":[62],"into":[63],"short":[64],"segments":[65],"for":[66,102],"efficient":[67],"training.":[68],"To":[69],"mitigate":[70],"this":[71,115],"mismatch,":[72],"we":[73,93],"propose":[74],"a":[75,80,89,109,135,173],"new":[76],"architecture,":[77],"which":[78,86],"variant":[81],"of":[82,159],"Gaussian":[84,121,141],"kernel,":[85,122],"itself":[87],"shift-invariant":[90,127],"kernel.":[91],"First,":[92],"mathematically":[94],"demonstrate":[95],"self-attention":[97],"with":[98,118,128,156],"shared":[99],"weight":[100],"parameters":[101],"queries":[103],"keys":[105],"equivalent":[107],"normalized":[110],"kernel":[111,116],"function.":[112],"By":[113],"replacing":[114],"function":[117],"proposed":[120,140,170],"architecture":[124],"becomes":[125],"completely":[126],"relative":[130],"position":[131],"information":[132],"embedded":[133],"using":[134],"frame":[136],"indexing":[137],"technique.":[138],"The":[139],"kernelized":[142],"was":[144],"applied":[145],"connectionist":[147],"temporal":[148],"classification":[149],"(CTC)":[150],"ASR.":[152],"An":[153],"experimental":[154],"evaluation":[155],"Corpus":[158],"Spontaneous":[160],"Japanese":[161],"(CSJ)":[162],"TEDLIUM":[164],"3":[165],"benchmarks":[166],"shows":[167],"achieves":[172],"improvement":[175],"(e.g.,":[178],"from":[179],"24.0%":[180],"WER":[181],"6.0%":[183],"CSJ)":[185],"without":[190],"any":[191],"windowing":[192],"techniques.":[193]},"counts_by_year":[{"year":2024,"cited_by_count":1},{"year":2022,"cited_by_count":2}],"updated_date":"2026-07-29T14:22:42.915294","created_date":"2025-10-10T00:00:00"}