{"id":"https://openalex.org/W3180338644","doi":"https://doi.org/10.1109/asru51503.2021.9688298","title":"Relaxed Attention: A Simple Method to Boost Performance of End-to-End Automatic Speech Recognition","display_name":"Relaxed Attention: A Simple Method to Boost Performance of End-to-End Automatic Speech Recognition","publication_year":2021,"publication_date":"2021-12-13","ids":{"openalex":"https://openalex.org/W3180338644","doi":"https://doi.org/10.1109/asru51503.2021.9688298","mag":"3180338644"},"language":"en","primary_location":{"id":"doi:10.1109/asru51503.2021.9688298","is_oa":false,"landing_page_url":"https://doi.org/10.1109/asru51503.2021.9688298","pdf_url":null,"source":{"id":"https://openalex.org/S4363606113","display_name":"2021 IEEE Automatic Speech Recognition and Understanding Workshop (ASRU)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"conference"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2021 IEEE Automatic Speech Recognition and Understanding Workshop (ASRU)","raw_type":"proceedings-article"},"type":"preprint","indexed_in":["arxiv","crossref","datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/2107.01275","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5039030019","display_name":"Timo Lohrenz","orcid":"https://orcid.org/0000-0001-7103-4302"},"institutions":[{"id":"https://openalex.org/I94509681","display_name":"Technische Universit\u00e4t Braunschweig","ror":"https://ror.org/010nsgg66","country_code":"DE","type":"education","lineage":["https://openalex.org/I94509681"]}],"countries":["DE"],"is_corresponding":true,"raw_author_name":"Timo Lohrenz","raw_affiliation_strings":["Technische Universität Braunschweig, Institute for Communications Technology,Braunschweig,Germany,38106","[Braunschweig University of Technology]"],"affiliations":[{"raw_affiliation_string":"Technische Universität Braunschweig, Institute for Communications Technology,Braunschweig,Germany,38106","institution_ids":["https://openalex.org/I94509681"]},{"raw_affiliation_string":"[Braunschweig University of Technology]","institution_ids":["https://openalex.org/I94509681"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5065268303","display_name":"Patrick Schwarz","orcid":"https://orcid.org/0000-0003-0383-1865"},"institutions":[{"id":"https://openalex.org/I94509681","display_name":"Technische Universit\u00e4t Braunschweig","ror":"https://ror.org/010nsgg66","country_code":"DE","type":"education","lineage":["https://openalex.org/I94509681"]}],"countries":["DE"],"is_corresponding":false,"raw_author_name":"Patrick Schwarz","raw_affiliation_strings":["Technische Universität Braunschweig, Institute for Communications Technology,Braunschweig,Germany,38106","[Braunschweig University of Technology]"],"affiliations":[{"raw_affiliation_string":"Technische Universität Braunschweig, Institute for Communications Technology,Braunschweig,Germany,38106","institution_ids":["https://openalex.org/I94509681"]},{"raw_affiliation_string":"[Braunschweig University of Technology]","institution_ids":["https://openalex.org/I94509681"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100736672","display_name":"Zhengyang Li","orcid":"https://orcid.org/0009-0007-6764-933X"},"institutions":[{"id":"https://openalex.org/I94509681","display_name":"Technische Universit\u00e4t Braunschweig","ror":"https://ror.org/010nsgg66","country_code":"DE","type":"education","lineage":["https://openalex.org/I94509681"]}],"countries":["DE"],"is_corresponding":false,"raw_author_name":"Zhengyang Li","raw_affiliation_strings":["Technische Universität Braunschweig, Institute for Communications Technology,Braunschweig,Germany,38106","[Braunschweig University of Technology]"],"affiliations":[{"raw_affiliation_string":"Technische Universität Braunschweig, Institute for Communications Technology,Braunschweig,Germany,38106","institution_ids":["https://openalex.org/I94509681"]},{"raw_affiliation_string":"[Braunschweig University of Technology]","institution_ids":["https://openalex.org/I94509681"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5002593702","display_name":"Tim Fingscheidt","orcid":"https://orcid.org/0000-0002-8895-5041"},"institutions":[{"id":"https://openalex.org/I94509681","display_name":"Technische Universit\u00e4t Braunschweig","ror":"https://ror.org/010nsgg66","country_code":"DE","type":"education","lineage":["https://openalex.org/I94509681"]}],"countries":["DE"],"is_corresponding":false,"raw_author_name":"Tim Fingscheidt","raw_affiliation_strings":["Technische Universität Braunschweig, Institute for Communications Technology,Braunschweig,Germany,38106"],"affiliations":[{"raw_affiliation_string":"Technische Universität Braunschweig, Institute for Communications Technology,Braunschweig,Germany,38106","institution_ids":["https://openalex.org/I94509681"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":4,"corresponding_author_ids":["https://openalex.org/A5039030019"],"corresponding_institution_ids":["https://openalex.org/I94509681"],"apc_list":null,"apc_paid":null,"fwci":0.0,"has_fulltext":true,"cited_by_count":0,"citation_normalized_percentile":{"value":0.05858978,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":"15","issue":null,"first_page":"177","last_page":"184"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10201","display_name":"Speech Recognition and Synthesis","score":1.0,"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":1.0,"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":0.9984999895095825,"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":0.9983999729156494,"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":0.7493869662284851},{"id":"https://openalex.org/keywords/end-to-end-principle","display_name":"End-to-end principle","score":0.6994369029998779},{"id":"https://openalex.org/keywords/decoding-methods","display_name":"Decoding methods","score":0.6624847650527954},{"id":"https://openalex.org/keywords/speech-recognition","display_name":"Speech recognition","score":0.6514438390731812},{"id":"https://openalex.org/keywords/language-model","display_name":"Language model","score":0.633897066116333},{"id":"https://openalex.org/keywords/transformer","display_name":"Transformer","score":0.6137785911560059},{"id":"https://openalex.org/keywords/encoder","display_name":"Encoder","score":0.6069101095199585},{"id":"https://openalex.org/keywords/word-error-rate","display_name":"Word error rate","score":0.522125780582428},{"id":"https://openalex.org/keywords/benchmark","display_name":"Benchmark (surveying)","score":0.4922458231449127},{"id":"https://openalex.org/keywords/hyperparameter","display_name":"Hyperparameter","score":0.481444776058197},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.3430543541908264},{"id":"https://openalex.org/keywords/algorithm","display_name":"Algorithm","score":0.17197933793067932},{"id":"https://openalex.org/keywords/engineering","display_name":"Engineering","score":0.07124286890029907}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7493869662284851},{"id":"https://openalex.org/C74296488","wikidata":"https://www.wikidata.org/wiki/Q2527392","display_name":"End-to-end principle","level":2,"score":0.6994369029998779},{"id":"https://openalex.org/C57273362","wikidata":"https://www.wikidata.org/wiki/Q576722","display_name":"Decoding methods","level":2,"score":0.6624847650527954},{"id":"https://openalex.org/C28490314","wikidata":"https://www.wikidata.org/wiki/Q189436","display_name":"Speech recognition","level":1,"score":0.6514438390731812},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.633897066116333},{"id":"https://openalex.org/C66322947","wikidata":"https://www.wikidata.org/wiki/Q11658","display_name":"Transformer","level":3,"score":0.6137785911560059},{"id":"https://openalex.org/C118505674","wikidata":"https://www.wikidata.org/wiki/Q42586063","display_name":"Encoder","level":2,"score":0.6069101095199585},{"id":"https://openalex.org/C40969351","wikidata":"https://www.wikidata.org/wiki/Q3516228","display_name":"Word error rate","level":2,"score":0.522125780582428},{"id":"https://openalex.org/C185798385","wikidata":"https://www.wikidata.org/wiki/Q1161707","display_name":"Benchmark (surveying)","level":2,"score":0.4922458231449127},{"id":"https://openalex.org/C8642999","wikidata":"https://www.wikidata.org/wiki/Q4171168","display_name":"Hyperparameter","level":2,"score":0.481444776058197},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.3430543541908264},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.17197933793067932},{"id":"https://openalex.org/C127413603","wikidata":"https://www.wikidata.org/wiki/Q11023","display_name":"Engineering","level":0,"score":0.07124286890029907},{"id":"https://openalex.org/C111919701","wikidata":"https://www.wikidata.org/wiki/Q9135","display_name":"Operating system","level":1,"score":0.0},{"id":"https://openalex.org/C119599485","wikidata":"https://www.wikidata.org/wiki/Q43035","display_name":"Electrical engineering","level":1,"score":0.0},{"id":"https://openalex.org/C205649164","wikidata":"https://www.wikidata.org/wiki/Q1071","display_name":"Geography","level":0,"score":0.0},{"id":"https://openalex.org/C165801399","wikidata":"https://www.wikidata.org/wiki/Q25428","display_name":"Voltage","level":2,"score":0.0},{"id":"https://openalex.org/C13280743","wikidata":"https://www.wikidata.org/wiki/Q131089","display_name":"Geodesy","level":1,"score":0.0}],"mesh":[],"locations_count":4,"locations":[{"id":"doi:10.1109/asru51503.2021.9688298","is_oa":false,"landing_page_url":"https://doi.org/10.1109/asru51503.2021.9688298","pdf_url":null,"source":{"id":"https://openalex.org/S4363606113","display_name":"2021 IEEE Automatic Speech Recognition and Understanding Workshop (ASRU)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"conference"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2021 IEEE Automatic Speech Recognition and Understanding Workshop (ASRU)","raw_type":"proceedings-article"},{"id":"pmh:oai:arXiv.org:2107.01275","is_oa":true,"landing_page_url":"http://arxiv.org/abs/2107.01275","pdf_url":"https://arxiv.org/pdf/2107.01275","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":"public-domain","license_id":"https://openalex.org/licenses/public-domain","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},{"id":"mag:3180338644","is_oa":true,"landing_page_url":"http://export.arxiv.org/pdf/2107.01275","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":"arXiv (Cornell University)","raw_type":null},{"id":"doi:10.48550/arxiv.2107.01275","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2107.01275","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":null,"is_accepted":false,"is_published":null,"raw_source_name":null,"raw_type":"article"}],"best_oa_location":{"id":"pmh:oai:arXiv.org:2107.01275","is_oa":true,"landing_page_url":"http://arxiv.org/abs/2107.01275","pdf_url":"https://arxiv.org/pdf/2107.01275","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":"public-domain","license_id":"https://openalex.org/licenses/public-domain","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"sustainable_development_goals":[{"display_name":"Peace, Justice and strong institutions","score":0.6100000143051147,"id":"https://metadata.un.org/sdg/16"}],"awards":[{"id":"https://openalex.org/G3353889084","display_name":null,"funder_award_id":"414091002","funder_id":"https://openalex.org/F4320320879","funder_display_name":"Deutsche Forschungsgemeinschaft"}],"funders":[{"id":"https://openalex.org/F4320320879","display_name":"Deutsche Forschungsgemeinschaft","ror":"https://ror.org/018mejw64"}],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W3180338644.pdf","grobid_xml":"https://content.openalex.org/works/W3180338644.grobid-xml"},"referenced_works_count":53,"referenced_works":["https://openalex.org/W854541894","https://openalex.org/W1494198834","https://openalex.org/W1524333225","https://openalex.org/W1915251500","https://openalex.org/W2024490110","https://openalex.org/W2024490156","https://openalex.org/W2102113734","https://openalex.org/W2143612262","https://openalex.org/W2183341477","https://openalex.org/W2303024254","https://openalex.org/W2327501763","https://openalex.org/W2526425061","https://openalex.org/W2577366047","https://openalex.org/W2626778328","https://openalex.org/W2774812352","https://openalex.org/W2787248535","https://openalex.org/W2885185669","https://openalex.org/W2886180730","https://openalex.org/W2892009249","https://openalex.org/W2933138175","https://openalex.org/W2936774411","https://openalex.org/W2948210185","https://openalex.org/W2950300355","https://openalex.org/W2962824709","https://openalex.org/W2963250244","https://openalex.org/W2963362078","https://openalex.org/W2964199361","https://openalex.org/W2964308564","https://openalex.org/W2970971581","https://openalex.org/W2972389417","https://openalex.org/W2972818416","https://openalex.org/W2974321942","https://openalex.org/W3007528493","https://openalex.org/W3008191852","https://openalex.org/W3008912312","https://openalex.org/W3015974384","https://openalex.org/W3096318498","https://openalex.org/W3097777922","https://openalex.org/W3142529478","https://openalex.org/W3163169798","https://openalex.org/W3163221127","https://openalex.org/W3163793923","https://openalex.org/W3198871389","https://openalex.org/W6623517193","https://openalex.org/W6631362777","https://openalex.org/W6638523607","https://openalex.org/W6674330103","https://openalex.org/W6675365184","https://openalex.org/W6679434410","https://openalex.org/W6732696085","https://openalex.org/W6739901393","https://openalex.org/W6754299077","https://openalex.org/W6766978945"],"related_works":["https://openalex.org/W3041914762","https://openalex.org/W2526425061","https://openalex.org/W3104384728","https://openalex.org/W3155343771","https://openalex.org/W3142529478","https://openalex.org/W2982640351","https://openalex.org/W3148986242","https://openalex.org/W3133764785","https://openalex.org/W2775304348","https://openalex.org/W3008370744","https://openalex.org/W3095773170","https://openalex.org/W3034198685","https://openalex.org/W2936706905","https://openalex.org/W2968981126","https://openalex.org/W2776984718","https://openalex.org/W3180992661","https://openalex.org/W2900676547","https://openalex.org/W2767601419","https://openalex.org/W3132002396","https://openalex.org/W2766058732"],"abstract_inverted_index":{"Recently,":[0],"attention-based":[1],"encoder-decoder":[2,45],"(AED)":[3],"models":[4,94],"have":[5],"shown":[6],"high":[7],"performance":[8],"for":[9,109],"end-to-end":[10,111],"automatic":[11],"speech":[12,112],"recognition":[13,113],"(ASR)":[14],"across":[15,66],"several":[16],"tasks.":[17],"Addressing":[18],"overconfidence":[19],"in":[20,23],"such":[21],"models,":[22],"this":[24],"paper":[25],"we":[26,104],"introduce":[27],"the":[28,44,61,91,124],"concept":[29],"of":[30,39,57,63,119,123],"relaxed":[31,64,88],"attention,":[32],"which":[33],"is":[34,51],"a":[35,40,106,115,133],"simple":[36],"gradual":[37],"injection":[38],"uniform":[41],"distribution":[42],"to":[43],"attention":[46,65,89],"weights":[47],"during":[48,96],"training":[49],"that":[50,84],"easily":[52],"implemented":[53],"with":[54,87,98,114],"two":[55,72],"lines":[56],"code.":[58],"We":[59,82],"investigate":[60],"effect":[62],"different":[67],"AED":[68],"model":[69],"architectures":[70],"and":[71,80],"prominent":[73],"ASR":[74],"tasks,":[75],"Wall":[76],"Street":[77],"Journal":[78],"(WSJ)":[79],"Librispeech.":[81],"found":[83],"transformers":[85],"trained":[86],"outperform":[90],"standard":[92],"baseline":[93],"consistently":[95],"decoding":[97],"external":[99],"language":[100],"models.":[101],"On":[102],"WSJ,":[103],"set":[105],"new":[107],"benchmark":[108],"transformer-based":[110],"word":[116],"error":[117],"rate":[118],"3.65%,":[120],"outperforming":[121],"state":[122],"art":[125],"(4.20%)":[126],"by":[127],"13.1%":[128],"relative,":[129],"while":[130],"introducing":[131],"only":[132],"single":[134],"hyperparameter.":[135]},"counts_by_year":[],"updated_date":"2026-04-09T08:11:56.329763","created_date":"2025-10-10T00:00:00"}