{"id":"https://openalex.org/W4224932142","doi":"https://doi.org/10.1109/icassp43922.2022.9747344","title":"Being Greedy Does Not Hurt: Sampling Strategies for End-To-End Speech Recognition","display_name":"Being Greedy Does Not Hurt: Sampling Strategies for End-To-End Speech Recognition","publication_year":2022,"publication_date":"2022-04-27","ids":{"openalex":"https://openalex.org/W4224932142","doi":"https://doi.org/10.1109/icassp43922.2022.9747344"},"language":"en","primary_location":{"id":"doi:10.1109/icassp43922.2022.9747344","is_oa":false,"landing_page_url":"https://doi.org/10.1109/icassp43922.2022.9747344","pdf_url":null,"source":{"id":"https://openalex.org/S4363607702","display_name":"ICASSP 2022 - 2022 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)","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":"ICASSP 2022 - 2022 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)","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/A5006131416","display_name":"Jahn Heymann","orcid":null},"institutions":[{"id":"https://openalex.org/I4210089985","display_name":"Amazon (Germany)","ror":"https://ror.org/00b9ktm87","country_code":"DE","type":"company","lineage":["https://openalex.org/I1311688040","https://openalex.org/I4210089985"]}],"countries":["DE"],"is_corresponding":false,"raw_author_name":"Jahn Heymann","raw_affiliation_strings":["Amazon.com"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Amazon.com","institution_ids":["https://openalex.org/I4210089985"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5045428440","display_name":"Egor Lakomkin","orcid":null},"institutions":[{"id":"https://openalex.org/I4210089985","display_name":"Amazon (Germany)","ror":"https://ror.org/00b9ktm87","country_code":"DE","type":"company","lineage":["https://openalex.org/I1311688040","https://openalex.org/I4210089985"]}],"countries":["DE"],"is_corresponding":false,"raw_author_name":"Egor Lakomkin","raw_affiliation_strings":["Amazon.com"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Amazon.com","institution_ids":["https://openalex.org/I4210089985"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5018372535","display_name":"Leif Radel","orcid":null},"institutions":[{"id":"https://openalex.org/I4210089985","display_name":"Amazon (Germany)","ror":"https://ror.org/00b9ktm87","country_code":"DE","type":"company","lineage":["https://openalex.org/I1311688040","https://openalex.org/I4210089985"]}],"countries":["DE"],"is_corresponding":false,"raw_author_name":"Leif Radel","raw_affiliation_strings":["Amazon.com"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Amazon.com","institution_ids":["https://openalex.org/I4210089985"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":3,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":0.0,"has_fulltext":false,"cited_by_count":0,"citation_normalized_percentile":{"value":0.02327061,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"7787","last_page":"7791"},"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/T11309","display_name":"Music and Audio Processing","score":0.9987000226974487,"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/T10860","display_name":"Speech and Audio Processing","score":0.9973999857902527,"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":0.7456457018852234},{"id":"https://openalex.org/keywords/smoothing","display_name":"Smoothing","score":0.6272921562194824},{"id":"https://openalex.org/keywords/word-error-rate","display_name":"Word error rate","score":0.5803685188293457},{"id":"https://openalex.org/keywords/speech-recognition","display_name":"Speech recognition","score":0.573317289352417},{"id":"https://openalex.org/keywords/inference","display_name":"Inference","score":0.5535566806793213},{"id":"https://openalex.org/keywords/scalability","display_name":"Scalability","score":0.5511243343353271},{"id":"https://openalex.org/keywords/sampling","display_name":"Sampling (signal processing)","score":0.5355987548828125},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.4398406445980072},{"id":"https://openalex.org/keywords/machine-learning","display_name":"Machine learning","score":0.4156959354877472},{"id":"https://openalex.org/keywords/word","display_name":"Word (group theory)","score":0.41520610451698303},{"id":"https://openalex.org/keywords/mathematics","display_name":"Mathematics","score":0.08729887008666992}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7456457018852234},{"id":"https://openalex.org/C3770464","wikidata":"https://www.wikidata.org/wiki/Q775963","display_name":"Smoothing","level":2,"score":0.6272921562194824},{"id":"https://openalex.org/C40969351","wikidata":"https://www.wikidata.org/wiki/Q3516228","display_name":"Word error rate","level":2,"score":0.5803685188293457},{"id":"https://openalex.org/C28490314","wikidata":"https://www.wikidata.org/wiki/Q189436","display_name":"Speech recognition","level":1,"score":0.573317289352417},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.5535566806793213},{"id":"https://openalex.org/C48044578","wikidata":"https://www.wikidata.org/wiki/Q727490","display_name":"Scalability","level":2,"score":0.5511243343353271},{"id":"https://openalex.org/C140779682","wikidata":"https://www.wikidata.org/wiki/Q210868","display_name":"Sampling (signal processing)","level":3,"score":0.5355987548828125},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.4398406445980072},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.4156959354877472},{"id":"https://openalex.org/C90805587","wikidata":"https://www.wikidata.org/wiki/Q10944557","display_name":"Word (group theory)","level":2,"score":0.41520610451698303},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.08729887008666992},{"id":"https://openalex.org/C2524010","wikidata":"https://www.wikidata.org/wiki/Q8087","display_name":"Geometry","level":1,"score":0.0},{"id":"https://openalex.org/C106131492","wikidata":"https://www.wikidata.org/wiki/Q3072260","display_name":"Filter (signal processing)","level":2,"score":0.0},{"id":"https://openalex.org/C31972630","wikidata":"https://www.wikidata.org/wiki/Q844240","display_name":"Computer vision","level":1,"score":0.0},{"id":"https://openalex.org/C77088390","wikidata":"https://www.wikidata.org/wiki/Q8513","display_name":"Database","level":1,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/icassp43922.2022.9747344","is_oa":false,"landing_page_url":"https://doi.org/10.1109/icassp43922.2022.9747344","pdf_url":null,"source":{"id":"https://openalex.org/S4363607702","display_name":"ICASSP 2022 - 2022 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)","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":"ICASSP 2022 - 2022 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)","raw_type":"proceedings-article"}],"best_oa_location":null,"sustainable_development_goals":[{"display_name":"Peace, Justice and strong institutions","score":0.4000000059604645,"id":"https://metadata.un.org/sdg/16"}],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":41,"referenced_works":["https://openalex.org/W648786980","https://openalex.org/W854541894","https://openalex.org/W1494198834","https://openalex.org/W1522301498","https://openalex.org/W1828163288","https://openalex.org/W1850531616","https://openalex.org/W2064675550","https://openalex.org/W2108677974","https://openalex.org/W2127141656","https://openalex.org/W2327501763","https://openalex.org/W2577366047","https://openalex.org/W2626177496","https://openalex.org/W2626967530","https://openalex.org/W2900209846","https://openalex.org/W2912526802","https://openalex.org/W2936774411","https://openalex.org/W2962765220","https://openalex.org/W2962824709","https://openalex.org/W2963144852","https://openalex.org/W2963250244","https://openalex.org/W2963620441","https://openalex.org/W2963727906","https://openalex.org/W2963747784","https://openalex.org/W3173510162","https://openalex.org/W4289440825","https://openalex.org/W6621543089","https://openalex.org/W6623517193","https://openalex.org/W6631190155","https://openalex.org/W6638749077","https://openalex.org/W6639056794","https://openalex.org/W6676315081","https://openalex.org/W6738974909","https://openalex.org/W6739638286","https://openalex.org/W6739651123","https://openalex.org/W6745628346","https://openalex.org/W6746892368","https://openalex.org/W6746924902","https://openalex.org/W6747046791","https://openalex.org/W6754576867","https://openalex.org/W6758654326","https://openalex.org/W6779059162"],"related_works":["https://openalex.org/W1978572805","https://openalex.org/W2383807498","https://openalex.org/W1997992934","https://openalex.org/W1987225439","https://openalex.org/W4238188170","https://openalex.org/W2125114371","https://openalex.org/W2019977573","https://openalex.org/W2149980199","https://openalex.org/W3125766170","https://openalex.org/W2944691285"],"abstract_inverted_index":{"Maximum":[0],"Likelihood":[1],"Estimation":[2],"(MLE)":[3],"is":[4,40,73,95,194],"currently":[5],"the":[6,48,71,96,169,180],"most":[7,157],"common":[8],"approach":[9,155],"to":[10,59,92,196],"train":[11],"large":[12,85,136],"scale":[13,137],"speech":[14,86],"recognition":[15],"systems.":[16],"While":[17],"it":[18],"has":[19],"significant":[20],"practical":[21],"advantages,":[22],"MLE":[23,78,145,181],"exhibits":[24],"several":[25,108,127],"drawbacks":[26],"known":[27,99],"in":[28,100],"literature:":[29],"training":[30,53],"and":[31,36,80,106,114,119,134,140,146,159,187],"inference":[32],"conditions":[33],"are":[34],"mismatched":[35],"a":[37,76,161],"proxy":[38],"objective":[39],"optimized":[41],"instead":[42],"of":[43,62,116],"word":[44,164],"error":[45,165],"rate.":[46],"Recently,":[47],"Optimal":[49],"Completion":[50],"Distillation":[51],"(OCD)":[52],"method":[54,72,177],"was":[55],"proposed":[56,149],"which":[57,91],"attempts":[58],"address":[60],"some":[61],"those":[63],"issues.":[64],"In":[65,102],"this":[66],"paper,":[67],"we":[68,104,173],"analyze":[69,107],"if":[70],"competitive":[74],"over":[75,168,179],"strong":[77],"baseline":[79],"investigate":[81],"its":[82],"scalability":[83],"towards":[84],"data":[87,133,139],"beyond":[88],"read":[89],"speech,":[90],"our":[93],"knowledge":[94],"first":[97],"attempt":[98],"literature.":[101],"addition,":[103],"propose":[105],"sampling":[109,151],"strategies":[110],"trading":[111],"off":[112],"exploration":[113],"exploitation":[115],"unseen":[117],"prefixes":[118],"their":[120],"effect":[121],"on":[122,129],"ASR":[123],"accuracy.":[124],"We":[125],"conduct":[126],"experiments":[128],"both":[130],"public":[131],"LibriSpeech":[132],"in-house":[135],"far-field":[138],"compare":[141],"models":[142],"trained":[143],"with":[144,152],"OCD.":[147],"Our":[148],"greedy":[150],"soft":[153],"targets":[154],"proves":[156],"effective":[158],"yields":[160],"9%":[162],"rel.":[163],"rate":[166],"improvement":[167],"i.i.d":[170],"sampling.":[171],"Finally,":[172],"note":[174],"that":[175],"OCD":[176],"improves":[178],"without":[182],"label":[183,192],"smoothing":[184,193],"by":[185,189],"12%,":[186],"underperform":[188],"6%":[190],"once":[191],"introduced":[195],"MLE.":[197]},"counts_by_year":[],"updated_date":"2026-06-11T09:08:48.828518","created_date":"2025-10-10T00:00:00"}