{"id":"https://openalex.org/W3142589538","doi":"https://doi.org/10.1109/slt48900.2021.9383521","title":"On The Usefulness of Self-Attention for Automatic Speech Recognition with Transformers","display_name":"On The Usefulness of Self-Attention for Automatic Speech Recognition with Transformers","publication_year":2021,"publication_date":"2021-01-19","ids":{"openalex":"https://openalex.org/W3142589538","doi":"https://doi.org/10.1109/slt48900.2021.9383521","mag":"3142589538"},"language":"en","primary_location":{"id":"doi:10.1109/slt48900.2021.9383521","is_oa":false,"landing_page_url":"https://doi.org/10.1109/slt48900.2021.9383521","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2021 IEEE Spoken Language Technology Workshop (SLT)","raw_type":"proceedings-article"},"type":"conference-paper","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://www.research.ed.ac.uk/files/198509577/On_the_Usefulness_ZHANG_DOA20112020_AVF.pdf","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5003030422","display_name":"Shucong Zhang","orcid":null},"institutions":[{"id":"https://openalex.org/I98677209","display_name":"University of Edinburgh","ror":"https://ror.org/01nrxwf90","country_code":"GB","type":"education","lineage":["https://openalex.org/I98677209"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Shucong Zhang","raw_affiliation_strings":["Centre for Speech Technology Research, University of Edinburgh, Edinburgh, UK"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Centre for Speech Technology Research, University of Edinburgh, Edinburgh, UK","institution_ids":["https://openalex.org/I98677209"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5042184515","display_name":"Erfan Loweimi","orcid":"https://orcid.org/0000-0002-8761-021X"},"institutions":[{"id":"https://openalex.org/I98677209","display_name":"University of Edinburgh","ror":"https://ror.org/01nrxwf90","country_code":"GB","type":"education","lineage":["https://openalex.org/I98677209"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Erfan Loweimi","raw_affiliation_strings":["Centre for Speech Technology Research, University of Edinburgh, Edinburgh, UK"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Centre for Speech Technology Research, University of Edinburgh, Edinburgh, UK","institution_ids":["https://openalex.org/I98677209"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5102911387","display_name":"Peter Bell","orcid":"https://orcid.org/0000-0002-9597-9615"},"institutions":[{"id":"https://openalex.org/I98677209","display_name":"University of Edinburgh","ror":"https://ror.org/01nrxwf90","country_code":"GB","type":"education","lineage":["https://openalex.org/I98677209"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Peter Bell","raw_affiliation_strings":["Centre for Speech Technology Research, University of Edinburgh, Edinburgh, UK"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Centre for Speech Technology Research, University of Edinburgh, Edinburgh, UK","institution_ids":["https://openalex.org/I98677209"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5027442277","display_name":"Steve Renals","orcid":"https://orcid.org/0000-0002-8790-3389"},"institutions":[{"id":"https://openalex.org/I98677209","display_name":"University of Edinburgh","ror":"https://ror.org/01nrxwf90","country_code":"GB","type":"education","lineage":["https://openalex.org/I98677209"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Steve Renals","raw_affiliation_strings":["Centre for Speech Technology Research, University of Edinburgh, Edinburgh, UK"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Centre for Speech Technology Research, University of Edinburgh, Edinburgh, UK","institution_ids":["https://openalex.org/I98677209"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":["https://openalex.org/I98677209"],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":true,"cited_by_count":35,"citation_normalized_percentile":null,"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"89","last_page":"96"},"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/T10181","display_name":"Natural Language Processing Techniques","score":0.9987000226974487,"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":0.9979000091552734,"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/computer-science","display_name":"Computer science","score":0.6945286989212036},{"id":"https://openalex.org/keywords/speech-recognition","display_name":"Speech recognition","score":0.6721706986427307},{"id":"https://openalex.org/keywords/transformer","display_name":"Transformer","score":0.5898284912109375},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.40202248096466064},{"id":"https://openalex.org/keywords/natural-language-processing","display_name":"Natural language processing","score":0.32588624954223633},{"id":"https://openalex.org/keywords/engineering","display_name":"Engineering","score":0.15757739543914795},{"id":"https://openalex.org/keywords/electrical-engineering","display_name":"Electrical engineering","score":0.10267084836959839},{"id":"https://openalex.org/keywords/voltage","display_name":"Voltage","score":0.08878645300865173}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6945286989212036},{"id":"https://openalex.org/C28490314","wikidata":"https://www.wikidata.org/wiki/Q189436","display_name":"Speech recognition","level":1,"score":0.6721706986427307},{"id":"https://openalex.org/C66322947","wikidata":"https://www.wikidata.org/wiki/Q11658","display_name":"Transformer","level":3,"score":0.5898284912109375},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.40202248096466064},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.32588624954223633},{"id":"https://openalex.org/C127413603","wikidata":"https://www.wikidata.org/wiki/Q11023","display_name":"Engineering","level":0,"score":0.15757739543914795},{"id":"https://openalex.org/C119599485","wikidata":"https://www.wikidata.org/wiki/Q43035","display_name":"Electrical engineering","level":1,"score":0.10267084836959839},{"id":"https://openalex.org/C165801399","wikidata":"https://www.wikidata.org/wiki/Q25428","display_name":"Voltage","level":2,"score":0.08878645300865173}],"mesh":[],"locations_count":4,"locations":[{"id":"doi:10.1109/slt48900.2021.9383521","is_oa":false,"landing_page_url":"https://doi.org/10.1109/slt48900.2021.9383521","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2021 IEEE Spoken Language Technology Workshop (SLT)","raw_type":"proceedings-article"},{"id":"pmh:oai:pure.ed.ac.uk:publications/0083d6df-244a-45e7-bf25-9c112538c485","is_oa":true,"landing_page_url":"https://hdl.handle.net/20.500.11820/0083d6df-244a-45e7-bf25-9c112538c485","pdf_url":"https://www.research.ed.ac.uk/files/198509577/On_the_Usefulness_ZHANG_DOA20112020_AVF.pdf","source":{"id":"https://openalex.org/S4306400321","display_name":"Edinburgh Research Explorer (University of Edinburgh)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I98677209","host_organization_name":"University of Edinburgh","host_organization_lineage":["https://openalex.org/I98677209"],"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":"Zhang, S, Loweimi, E, Bell, P & Renals, S 2021, On the usefulness of self-attention for automatic speech recognition with transformers. in 2021 IEEE Spoken Language Technology Workshop (SLT). Institute of Electrical and Electronics Engineers, pp. 89 - 96, IEEE Spoken Language Technology Workshop, 19/01/21. https://doi.org/10.1109/SLT48900.2021.9383521","raw_type":"contributionToPeriodical"},{"id":"pmh:oai:pure.ed.ac.uk:openaire/0083d6df-244a-45e7-bf25-9c112538c485","is_oa":true,"landing_page_url":"https://www.research.ed.ac.uk/en/publications/0083d6df-244a-45e7-bf25-9c112538c485","pdf_url":null,"source":{"id":"https://openalex.org/S4306400321","display_name":"Edinburgh Research Explorer (University of Edinburgh)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I98677209","host_organization_name":"University of Edinburgh","host_organization_lineage":["https://openalex.org/I98677209"],"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":"Zhang, S, Loweimi, E, Bell, P & Renals, S 2021, On the usefulness of self-attention for automatic speech recognition with transformers. in 2021 IEEE Spoken Language Technology Workshop (SLT). Institute of Electrical and Electronics Engineers, pp. 89 - 96, IEEE Spoken Language Technology Workshop, 19/01/21. https://doi.org/10.1109/SLT48900.2021.9383521","raw_type":"contributionToPeriodical"},{"id":"pmh:oai:repository@napier.ac.uk:3585873","is_oa":false,"landing_page_url":"http://researchrepository.napier.ac.uk/Output/3585873","pdf_url":null,"source":{"id":"https://openalex.org/S4306402591","display_name":"Edinburgh Napier Research Repository (Edinburgh Napier University)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I251738","host_organization_name":"Edinburgh Napier University","host_organization_lineage":["https://openalex.org/I251738"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Presentation / Conference Contribution"}],"best_oa_location":{"id":"pmh:oai:pure.ed.ac.uk:publications/0083d6df-244a-45e7-bf25-9c112538c485","is_oa":true,"landing_page_url":"https://hdl.handle.net/20.500.11820/0083d6df-244a-45e7-bf25-9c112538c485","pdf_url":"https://www.research.ed.ac.uk/files/198509577/On_the_Usefulness_ZHANG_DOA20112020_AVF.pdf","source":{"id":"https://openalex.org/S4306400321","display_name":"Edinburgh Research Explorer (University of Edinburgh)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I98677209","host_organization_name":"University of Edinburgh","host_organization_lineage":["https://openalex.org/I98677209"],"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":"Zhang, S, Loweimi, E, Bell, P & Renals, S 2021, On the usefulness of self-attention for automatic speech recognition with transformers. in 2021 IEEE Spoken Language Technology Workshop (SLT). Institute of Electrical and Electronics Engineers, pp. 89 - 96, IEEE Spoken Language Technology Workshop, 19/01/21. https://doi.org/10.1109/SLT48900.2021.9383521","raw_type":"contributionToPeriodical"},"sustainable_development_goals":[{"score":0.7200000286102295,"display_name":"Peace, Justice and strong institutions","id":"https://metadata.un.org/sdg/16"}],"awards":[],"funders":[],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W3142589538.pdf","grobid_xml":"https://content.openalex.org/works/W3142589538.grobid-xml"},"referenced_works_count":50,"referenced_works":["https://openalex.org/W1522301498","https://openalex.org/W1524333225","https://openalex.org/W1600744878","https://openalex.org/W1924770834","https://openalex.org/W2024490156","https://openalex.org/W2064675550","https://openalex.org/W2085628288","https://openalex.org/W2095705004","https://openalex.org/W2107878631","https://openalex.org/W2127141656","https://openalex.org/W2133564696","https://openalex.org/W2166637769","https://openalex.org/W2302255633","https://openalex.org/W2526425061","https://openalex.org/W2767601419","https://openalex.org/W2802023636","https://openalex.org/W2892009249","https://openalex.org/W2899771611","https://openalex.org/W2936706905","https://openalex.org/W2962778134","https://openalex.org/W2962780374","https://openalex.org/W2962784628","https://openalex.org/W2963312585","https://openalex.org/W2963403868","https://openalex.org/W2964121744","https://openalex.org/W2964308564","https://openalex.org/W2970565456","https://openalex.org/W2972389417","https://openalex.org/W2972451902","https://openalex.org/W2972818416","https://openalex.org/W2981857663","https://openalex.org/W3007328579","https://openalex.org/W3008191852","https://openalex.org/W3027687786","https://openalex.org/W3097973766","https://openalex.org/W4288347855","https://openalex.org/W4385245566","https://openalex.org/W6631190155","https://openalex.org/W6631362777","https://openalex.org/W6640212811","https://openalex.org/W6674330103","https://openalex.org/W6679434410","https://openalex.org/W6698183232","https://openalex.org/W6739901393","https://openalex.org/W6745543685","https://openalex.org/W6756040250","https://openalex.org/W6762945437","https://openalex.org/W6769557084","https://openalex.org/W6780226713","https://openalex.org/W6887576000"],"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":{"Self-attention":[0],"models":[1,84],"such":[2],"as":[3],"Transformers,":[4,100],"which":[5],"can":[6],"capture":[7],"temporal":[8],"relationships":[9],"without":[10],"being":[11],"limited":[12],"by":[13],"the":[14,27,30,35,67,72,116,123,129,138],"distance":[15],"between":[16],"events,":[17],"have":[18],"given":[19],"competitive":[20],"speech":[21,60],"recognition":[22],"results.":[23],"However,":[24],"we":[25,82],"note":[26],"range":[28],"of":[29,66,115,118],"learned":[31,124],"context":[32],"increases":[33,127],"from":[34,128],"lower":[36,86,130],"to":[37,56,98,131],"upper":[38,73,132,145],"self-attention":[39,74,134],"layers,":[40],"whilst":[41],"acoustic":[42],"events":[43],"often":[44],"happen":[45],"within":[46],"short":[47],"time":[48],"spans":[49],"in":[50,77,143],"a":[51,57,63,112],"left-to-right":[52],"order.":[53],"This":[54],"leads":[55],"question:":[58],"for":[59,71],"recognition,":[61],"is":[62,141],"global":[64,139],"view":[65,140],"entire":[68],"sequence":[69],"useful":[70],"encoder":[75,133,146],"layers":[76,89],"Transformers?":[78],"To":[79],"investigate":[80],"this,":[81],"train":[83],"with":[85],"self-attention/upper":[87],"feed-forward":[88],"encoders":[90],"on":[91],"Wall":[92],"Street":[93],"Journal":[94],"and":[95,121],"Switchboard.":[96],"Compared":[97],"baseline":[99],"no":[101],"performance":[102],"drop":[103],"but":[104],"minor":[105],"gains":[106],"are":[107],"observed.":[108],"We":[109,136],"further":[110],"developed":[111],"novel":[113],"metric":[114],"diagonality":[117,125],"attention":[119],"matrices":[120],"found":[122],"indeed":[126],"layers.":[135,147],"conclude":[137],"unnecessary":[142],"training":[144]},"counts_by_year":[{"year":2026,"cited_by_count":2},{"year":2025,"cited_by_count":7},{"year":2024,"cited_by_count":9},{"year":2023,"cited_by_count":8},{"year":2022,"cited_by_count":7},{"year":2021,"cited_by_count":1},{"year":2020,"cited_by_count":1}],"updated_date":"2026-07-14T23:27:15.235271","created_date":"2025-10-10T00:00:00"}