{"id":"https://openalex.org/W3137489363","doi":"https://doi.org/10.21437/interspeech.2021-1743","title":"Transformer-Based ASR Incorporating Time-Reduction Layer and Fine-Tuning with Self-Knowledge Distillation","display_name":"Transformer-Based ASR Incorporating Time-Reduction Layer and Fine-Tuning with Self-Knowledge Distillation","publication_year":2021,"publication_date":"2021-08-27","ids":{"openalex":"https://openalex.org/W3137489363","doi":"https://doi.org/10.21437/interspeech.2021-1743","mag":"3137489363"},"language":"en","primary_location":{"id":"doi:10.21437/interspeech.2021-1743","is_oa":false,"landing_page_url":"https://doi.org/10.21437/interspeech.2021-1743","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Interspeech 2021","raw_type":"proceedings-article"},"type":"conference-paper","indexed_in":["arxiv","crossref","datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/2103.09903","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5006199508","display_name":"Akmal Haidar","orcid":null},"institutions":[{"id":"https://openalex.org/I2250955327","display_name":"Huawei Technologies (China)","ror":"https://ror.org/00cmhce21","country_code":"CN","type":"company","lineage":["https://openalex.org/I2250955327"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Md. Akmal Haidar","raw_affiliation_strings":["Huawei#TAB#"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Huawei#TAB#","institution_ids":["https://openalex.org/I2250955327"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5107981608","display_name":"Chao Xing","orcid":"https://orcid.org/0000-0003-1120-8803"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Chao Xing","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5028862918","display_name":"Mehdi Rezagholizadeh","orcid":"https://orcid.org/0000-0003-4014-6007"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Mehdi Rezagholizadeh","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":true,"cited_by_count":0,"citation_normalized_percentile":null,"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"2102","last_page":"2106"},"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":0.9984999895095825,"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":0.9977999925613403,"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/transformer","display_name":"Transformer","score":0.7942585945129395},{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.7750405073165894},{"id":"https://openalex.org/keywords/encoder","display_name":"Encoder","score":0.7325440049171448},{"id":"https://openalex.org/keywords/inference","display_name":"Inference","score":0.6497613787651062},{"id":"https://openalex.org/keywords/speech-recognition","display_name":"Speech recognition","score":0.5040292143821716},{"id":"https://openalex.org/keywords/language-model","display_name":"Language model","score":0.4745025336742401},{"id":"https://openalex.org/keywords/word-error-rate","display_name":"Word error rate","score":0.46429309248924255},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.4586929678916931},{"id":"https://openalex.org/keywords/engineering","display_name":"Engineering","score":0.08130228519439697},{"id":"https://openalex.org/keywords/voltage","display_name":"Voltage","score":0.07148706912994385}],"concepts":[{"id":"https://openalex.org/C66322947","wikidata":"https://www.wikidata.org/wiki/Q11658","display_name":"Transformer","level":3,"score":0.7942585945129395},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7750405073165894},{"id":"https://openalex.org/C118505674","wikidata":"https://www.wikidata.org/wiki/Q42586063","display_name":"Encoder","level":2,"score":0.7325440049171448},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.6497613787651062},{"id":"https://openalex.org/C28490314","wikidata":"https://www.wikidata.org/wiki/Q189436","display_name":"Speech recognition","level":1,"score":0.5040292143821716},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.4745025336742401},{"id":"https://openalex.org/C40969351","wikidata":"https://www.wikidata.org/wiki/Q3516228","display_name":"Word error rate","level":2,"score":0.46429309248924255},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.4586929678916931},{"id":"https://openalex.org/C127413603","wikidata":"https://www.wikidata.org/wiki/Q11023","display_name":"Engineering","level":0,"score":0.08130228519439697},{"id":"https://openalex.org/C165801399","wikidata":"https://www.wikidata.org/wiki/Q25428","display_name":"Voltage","level":2,"score":0.07148706912994385},{"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/C111919701","wikidata":"https://www.wikidata.org/wiki/Q9135","display_name":"Operating system","level":1,"score":0.0}],"mesh":[],"locations_count":4,"locations":[{"id":"doi:10.21437/interspeech.2021-1743","is_oa":false,"landing_page_url":"https://doi.org/10.21437/interspeech.2021-1743","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Interspeech 2021","raw_type":"proceedings-article"},{"id":"pmh:oai:arXiv.org:2103.09903","is_oa":true,"landing_page_url":"http://arxiv.org/abs/2103.09903","pdf_url":"https://arxiv.org/pdf/2103.09903","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":"cc-by-nc-nd","license_id":"https://openalex.org/licenses/cc-by-nc-nd","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},{"id":"mag:3137489363","is_oa":true,"landing_page_url":"https://arxiv.org/pdf/2103.09903.pdf","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.2103.09903","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2103.09903","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":"Preprint"}],"best_oa_location":{"id":"pmh:oai:arXiv.org:2103.09903","is_oa":true,"landing_page_url":"http://arxiv.org/abs/2103.09903","pdf_url":"https://arxiv.org/pdf/2103.09903","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":"cc-by-nc-nd","license_id":"https://openalex.org/licenses/cc-by-nc-nd","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"sustainable_development_goals":[{"display_name":"Quality Education","score":0.46000000834465027,"id":"https://metadata.un.org/sdg/4"}],"awards":[],"funders":[],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W3137489363.pdf","grobid_xml":"https://content.openalex.org/works/W3137489363.grobid-xml"},"referenced_works_count":37,"referenced_works":["https://openalex.org/W1494198834","https://openalex.org/W1821462560","https://openalex.org/W1828163288","https://openalex.org/W2102113734","https://openalex.org/W2127141656","https://openalex.org/W2193413348","https://openalex.org/W2327501763","https://openalex.org/W2508418541","https://openalex.org/W2795935804","https://openalex.org/W2803023299","https://openalex.org/W2885185669","https://openalex.org/W2888930363","https://openalex.org/W2892008152","https://openalex.org/W2892009249","https://openalex.org/W2936774411","https://openalex.org/W2938974662","https://openalex.org/W2941814890","https://openalex.org/W2962760690","https://openalex.org/W2962835968","https://openalex.org/W2962988733","https://openalex.org/W2963362078","https://openalex.org/W2963382687","https://openalex.org/W2963403868","https://openalex.org/W2964121744","https://openalex.org/W2972389417","https://openalex.org/W2972818416","https://openalex.org/W2983128379","https://openalex.org/W2991213871","https://openalex.org/W2992632249","https://openalex.org/W3015974384","https://openalex.org/W3016010032","https://openalex.org/W3021515889","https://openalex.org/W3025165719","https://openalex.org/W3036049548","https://openalex.org/W3037029661","https://openalex.org/W3103005696","https://openalex.org/W3156828761"],"related_works":["https://openalex.org/W3136786651","https://openalex.org/W2913566369","https://openalex.org/W2804884650","https://openalex.org/W3197976821","https://openalex.org/W3105804405","https://openalex.org/W2883586237","https://openalex.org/W3037029661","https://openalex.org/W2992448548","https://openalex.org/W3095773170","https://openalex.org/W1504449689","https://openalex.org/W3148654612","https://openalex.org/W2572023240","https://openalex.org/W3186546663","https://openalex.org/W2992632249","https://openalex.org/W2922078472","https://openalex.org/W3194687854","https://openalex.org/W2980690446","https://openalex.org/W3209376089","https://openalex.org/W3204687321","https://openalex.org/W2899829828"],"abstract_inverted_index":{"End-to-end":[0],"automatic":[1],"speech":[2,18,25,47],"recognition":[3],"(ASR),":[4],"unlike":[5],"conventional":[6],"ASR,":[7,55],"does":[8],"not":[9,60],"have":[10],"modules":[11],"to":[12,30,50,76,118,122],"learn":[13,31],"the":[14,21,28,32,37,43,56,72,77,101,128,135,139,166],"semantic":[15,33,66],"representation":[16,26,34,67],"from":[17],"encoder.":[19],"Moreover,":[20,149],"higher":[22],"frame-rate":[23,45,58],"of":[24,46,84,138,168],"prevents":[27],"model":[29,99,190],"properly.":[35],"Therefore,":[36],"models":[38,158,205],"that":[39,125,177],"are":[40],"constructed":[41],"by":[42],"lower":[44,57],"encoder":[48,114],"lead":[49],"better":[51,65],"performance.":[52],"For":[53],"Transformer-based":[54,97,184,203],"is":[59],"only":[61],"important":[62],"for":[63,70,142,155,202],"learning":[64],"but":[68],"also":[69],"reducing":[71,134],"computational":[73,136],"complexity":[74,85],"due":[75],"self-attention":[78,140],"mechanism":[79],"which":[80,106,163],"has":[81],"O(n^2)":[82],"order":[83],"in":[86,105,116,133],"both":[87],"training":[88,143],"and":[89,144],"inference.":[90],"In":[91],"this":[92],"paper,":[93],"we":[94,107,150,193],"propose":[95],"a":[96,152],"ASR":[98,157,170,185,204],"with":[100,146,188,206],"time":[102,109],"reduction":[103,110],"layer,":[104],"incorporate":[108],"layer":[111],"inside":[112],"transformer":[113],"layers":[115],"addition":[117],"traditional":[119],"sub-sampling":[120],"methods":[121,180],"input":[123],"features":[124],"further":[126,164],"reduce":[127],"frame-rate.":[129],"This":[130],"can":[131],"help":[132],"cost":[137],"process":[141],"inference":[145],"performance":[147,167],"improvement.":[148],"introduce":[151],"fine-tuning":[153],"approach":[154],"pre-trained":[156],"using":[159],"self-knowledge":[160],"distillation":[161],"(S-KD)":[162],"improves":[165],"our":[169,178],"model.":[171],"Experiments":[172],"on":[173],"LibriSpeech":[174],"datasets":[175],"show":[176],"proposed":[179],"outperform":[181],"all":[182],"other":[183],"systems.":[186],"Furthermore,":[187],"language":[189],"(LM)":[191],"fusion,":[192],"achieve":[194],"new":[195],"state-of-the-art":[196],"word":[197],"error":[198],"rate":[199],"(WER)":[200],"results":[201],"just":[207],"30":[208],"million":[209],"parameters":[210],"trained":[211],"without":[212],"any":[213],"external":[214],"data.":[215]},"counts_by_year":[],"updated_date":"2026-07-14T23:27:15.235271","created_date":"2025-10-10T00:00:00"}