{"id":"https://openalex.org/W3196784225","doi":"https://doi.org/10.21437/interspeech.2021-1887","title":"Dissecting User-Perceived Latency of On-Device E2E Speech Recognition","display_name":"Dissecting User-Perceived Latency of On-Device E2E Speech Recognition","publication_year":2021,"publication_date":"2021-08-27","ids":{"openalex":"https://openalex.org/W3196784225","doi":"https://doi.org/10.21437/interspeech.2021-1887","mag":"3196784225"},"language":"en","primary_location":{"id":"doi:10.21437/interspeech.2021-1887","is_oa":false,"landing_page_url":"https://doi.org/10.21437/interspeech.2021-1887","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":"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/A5047358828","display_name":"Yuan Shangguan","orcid":null},"institutions":[{"id":"https://openalex.org/I4210114444","display_name":"Meta (United States)","ror":"https://ror.org/01zbnvs85","country_code":"US","type":"company","lineage":["https://openalex.org/I4210114444"]}],"countries":["US"],"is_corresponding":true,"raw_author_name":"Yuan Shangguan","raw_affiliation_strings":["Facebook AI, USA"],"affiliations":[{"raw_affiliation_string":"Facebook AI, USA","institution_ids":["https://openalex.org/I4210114444"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5032640894","display_name":"Rohit Prabhavalkar","orcid":"https://orcid.org/0000-0001-5331-6058"},"institutions":[{"id":"https://openalex.org/I4210114444","display_name":"Meta (United States)","ror":"https://ror.org/01zbnvs85","country_code":"US","type":"company","lineage":["https://openalex.org/I4210114444"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Rohit Prabhavalkar","raw_affiliation_strings":["Facebook AI, USA"],"affiliations":[{"raw_affiliation_string":"Facebook AI, USA","institution_ids":["https://openalex.org/I4210114444"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100341891","display_name":"Hang Su","orcid":"https://orcid.org/0000-0002-6877-6783"},"institutions":[{"id":"https://openalex.org/I4210114444","display_name":"Meta (United States)","ror":"https://ror.org/01zbnvs85","country_code":"US","type":"company","lineage":["https://openalex.org/I4210114444"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Hang Su","raw_affiliation_strings":["Facebook AI, USA"],"affiliations":[{"raw_affiliation_string":"Facebook AI, USA","institution_ids":["https://openalex.org/I4210114444"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5074237839","display_name":"Jay Mahadeokar","orcid":null},"institutions":[{"id":"https://openalex.org/I4210114444","display_name":"Meta (United States)","ror":"https://ror.org/01zbnvs85","country_code":"US","type":"company","lineage":["https://openalex.org/I4210114444"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Jay Mahadeokar","raw_affiliation_strings":["Facebook AI, USA"],"affiliations":[{"raw_affiliation_string":"Facebook AI, USA","institution_ids":["https://openalex.org/I4210114444"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5103247973","display_name":"Yangyang Shi","orcid":"https://orcid.org/0000-0001-5297-4155"},"institutions":[{"id":"https://openalex.org/I4210114444","display_name":"Meta (United States)","ror":"https://ror.org/01zbnvs85","country_code":"US","type":"company","lineage":["https://openalex.org/I4210114444"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Yangyang Shi","raw_affiliation_strings":["Facebook AI, USA"],"affiliations":[{"raw_affiliation_string":"Facebook AI, USA","institution_ids":["https://openalex.org/I4210114444"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5043926301","display_name":"Jiatong Zhou","orcid":null},"institutions":[{"id":"https://openalex.org/I4210114444","display_name":"Meta (United States)","ror":"https://ror.org/01zbnvs85","country_code":"US","type":"company","lineage":["https://openalex.org/I4210114444"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Jiatong Zhou","raw_affiliation_strings":["Facebook AI, USA"],"affiliations":[{"raw_affiliation_string":"Facebook AI, USA","institution_ids":["https://openalex.org/I4210114444"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5101579713","display_name":"Chunyang Wu","orcid":"https://orcid.org/0000-0002-0269-3555"},"institutions":[{"id":"https://openalex.org/I4210114444","display_name":"Meta (United States)","ror":"https://ror.org/01zbnvs85","country_code":"US","type":"company","lineage":["https://openalex.org/I4210114444"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Chunyang Wu","raw_affiliation_strings":["Facebook AI, USA"],"affiliations":[{"raw_affiliation_string":"Facebook AI, USA","institution_ids":["https://openalex.org/I4210114444"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5103098852","display_name":"Duc Le","orcid":"https://orcid.org/0000-0001-9490-2563"},"institutions":[{"id":"https://openalex.org/I4210114444","display_name":"Meta (United States)","ror":"https://ror.org/01zbnvs85","country_code":"US","type":"company","lineage":["https://openalex.org/I4210114444"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Duc Le","raw_affiliation_strings":["Facebook AI, USA"],"affiliations":[{"raw_affiliation_string":"Facebook AI, USA","institution_ids":["https://openalex.org/I4210114444"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5066166549","display_name":"Ozlem Kalinli","orcid":null},"institutions":[{"id":"https://openalex.org/I4210114444","display_name":"Meta (United States)","ror":"https://ror.org/01zbnvs85","country_code":"US","type":"company","lineage":["https://openalex.org/I4210114444"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Ozlem Kalinli","raw_affiliation_strings":["Facebook AI, USA"],"affiliations":[{"raw_affiliation_string":"Facebook AI, USA","institution_ids":["https://openalex.org/I4210114444"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5047073253","display_name":"Christian Fuegen","orcid":null},"institutions":[{"id":"https://openalex.org/I4210114444","display_name":"Meta (United States)","ror":"https://ror.org/01zbnvs85","country_code":"US","type":"company","lineage":["https://openalex.org/I4210114444"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Christian Fuegen","raw_affiliation_strings":["Facebook AI, USA"],"affiliations":[{"raw_affiliation_string":"Facebook AI, USA","institution_ids":["https://openalex.org/I4210114444"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5041313589","display_name":"Michael L. Seltzer","orcid":"https://orcid.org/0000-0003-3474-2451"},"institutions":[{"id":"https://openalex.org/I4210114444","display_name":"Meta (United States)","ror":"https://ror.org/01zbnvs85","country_code":"US","type":"company","lineage":["https://openalex.org/I4210114444"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Michael L. Seltzer","raw_affiliation_strings":["Facebook AI, USA"],"affiliations":[{"raw_affiliation_string":"Facebook AI, USA","institution_ids":["https://openalex.org/I4210114444"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":11,"corresponding_author_ids":["https://openalex.org/A5047358828"],"corresponding_institution_ids":["https://openalex.org/I4210114444"],"apc_list":null,"apc_paid":null,"fwci":3.0978,"has_fulltext":false,"cited_by_count":26,"citation_normalized_percentile":{"value":0.92866291,"is_in_top_1_percent":false,"is_in_top_10_percent":true},"cited_by_percentile_year":{"min":94,"max":99},"biblio":{"volume":null,"issue":null,"first_page":"4553","last_page":"4557"},"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/T10860","display_name":"Speech and Audio Processing","score":0.9997000098228455,"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/T11309","display_name":"Music 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":0.8566689491271973},{"id":"https://openalex.org/keywords/security-token","display_name":"Security token","score":0.6743841171264648},{"id":"https://openalex.org/keywords/latency","display_name":"Latency (audio)","score":0.6661185026168823},{"id":"https://openalex.org/keywords/speech-recognition","display_name":"Speech recognition","score":0.6166288256645203},{"id":"https://openalex.org/keywords/word-error-rate","display_name":"Word error rate","score":0.512829601764679},{"id":"https://openalex.org/keywords/decoding-methods","display_name":"Decoding methods","score":0.42763757705688477},{"id":"https://openalex.org/keywords/computation","display_name":"Computation","score":0.420566588640213},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.32539600133895874},{"id":"https://openalex.org/keywords/computer-network","display_name":"Computer network","score":0.0998128354549408}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8566689491271973},{"id":"https://openalex.org/C48145219","wikidata":"https://www.wikidata.org/wiki/Q1335365","display_name":"Security token","level":2,"score":0.6743841171264648},{"id":"https://openalex.org/C82876162","wikidata":"https://www.wikidata.org/wiki/Q17096504","display_name":"Latency (audio)","level":2,"score":0.6661185026168823},{"id":"https://openalex.org/C28490314","wikidata":"https://www.wikidata.org/wiki/Q189436","display_name":"Speech recognition","level":1,"score":0.6166288256645203},{"id":"https://openalex.org/C40969351","wikidata":"https://www.wikidata.org/wiki/Q3516228","display_name":"Word error rate","level":2,"score":0.512829601764679},{"id":"https://openalex.org/C57273362","wikidata":"https://www.wikidata.org/wiki/Q576722","display_name":"Decoding methods","level":2,"score":0.42763757705688477},{"id":"https://openalex.org/C45374587","wikidata":"https://www.wikidata.org/wiki/Q12525525","display_name":"Computation","level":2,"score":0.420566588640213},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.32539600133895874},{"id":"https://openalex.org/C31258907","wikidata":"https://www.wikidata.org/wiki/Q1301371","display_name":"Computer network","level":1,"score":0.0998128354549408},{"id":"https://openalex.org/C76155785","wikidata":"https://www.wikidata.org/wiki/Q418","display_name":"Telecommunications","level":1,"score":0.0},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.21437/interspeech.2021-1887","is_oa":false,"landing_page_url":"https://doi.org/10.21437/interspeech.2021-1887","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"}],"best_oa_location":null,"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":34,"referenced_works":["https://openalex.org/W1734538896","https://openalex.org/W1828163288","https://openalex.org/W1995562189","https://openalex.org/W2120375264","https://openalex.org/W2143612262","https://openalex.org/W2293634267","https://openalex.org/W2327501763","https://openalex.org/W2407080277","https://openalex.org/W2497335382","https://openalex.org/W2572102653","https://openalex.org/W2625979394","https://openalex.org/W2935756939","https://openalex.org/W2962760690","https://openalex.org/W2963250244","https://openalex.org/W2963414781","https://openalex.org/W2975044525","https://openalex.org/W2987019345","https://openalex.org/W3007227084","https://openalex.org/W3015194534","https://openalex.org/W3015654466","https://openalex.org/W3015927303","https://openalex.org/W3016010032","https://openalex.org/W3016155933","https://openalex.org/W3095996511","https://openalex.org/W3096104971","https://openalex.org/W3096702180","https://openalex.org/W3096888553","https://openalex.org/W3097777922","https://openalex.org/W3140235797","https://openalex.org/W3149509723","https://openalex.org/W3151287998","https://openalex.org/W3162639446","https://openalex.org/W3162665866","https://openalex.org/W3163907627"],"related_works":["https://openalex.org/W4388335561","https://openalex.org/W2970530566","https://openalex.org/W4288261899","https://openalex.org/W4307309205","https://openalex.org/W2967478618","https://openalex.org/W4385009901","https://openalex.org/W4385572700","https://openalex.org/W4389471172","https://openalex.org/W4389115113","https://openalex.org/W50892825"],"abstract_inverted_index":{"As":[0],"speech-enabled":[1],"devices":[2],"such":[3,33,56],"as":[4,34,45,69,71],"smartphones":[5],"and":[6,39,54,88,155,170],"smart":[7],"speakers":[8],"become":[9],"increasingly":[10],"ubiquitous,":[11],"there":[12],"is":[13],"growing":[14],"interest":[15],"in":[16,135],"building":[17],"automatic":[18],"speech":[19,30,61],"recognition":[20,31],"(ASR)":[21],"systems":[22,57],"that":[23,95,111,149],"can":[24],"run":[25],"directly":[26],"on-device;":[27],"end-to-end":[28],"(E2E)":[29],"models":[32,141],"recurrent":[35],"neural":[36],"network":[37],"transducers":[38],"their":[40],"variants":[41],"have":[42,158],"recently":[43,183],"emerged":[44],"prime":[46],"candidates":[47],"for":[48],"this":[49],"task.Apart":[50],"from":[51],"being":[52],"accurate":[53],"compact,":[55],"need":[58],"to":[59,116],"decode":[60],"with":[62,125,178],"low":[63],"user-perceived":[64],"latency":[65,130,138,169],"(UPL),":[66],"producing":[67],"words":[68],"soon":[70],"they":[72],"are":[73,120,142],"spoken.This":[74],"work":[75],"examines":[76],"the":[77,113,165,182],"impact":[78,161],"of":[79,97,106],"various":[80],"techniques":[81],"-model":[82],"architectures,":[83],"training":[84],"criteria,":[85],"decoding":[86],"hyperparameters,":[87],"endpointer":[89],"parameters":[90],"-on":[91],"UPL.Our":[92],"analyses":[93],"suggest":[94],"measures":[96,105],"model":[98],"size":[99],"(parameters,":[100],"input":[101,118],"chunk":[102],"sizes),":[103],"or":[104],"computation":[107],"(e.g.,":[108],"FLOPS,":[109],"RTF)":[110],"reflect":[112],"model's":[114],"ability":[115],"process":[117],"frames":[119],"not":[121],"always":[122],"strongly":[123],"correlated":[124],"observed":[126,139],"UPL.Thus,":[127],"conventional":[128],"algorithmic":[129],"measurements":[131],"might":[132],"be":[133],"inadequate":[134],"accurately":[136],"capturing":[137],"when":[140,174],"deployed":[143],"on":[144,162],"embedded":[145],"devices.Instead,":[146],"we":[147],"find":[148],"factors":[150],"affecting":[151],"token":[152],"emission":[153],"latency,":[154],"endpointing":[156],"behavior":[157],"a":[159],"larger":[160],"UPL.We":[163],"achieve":[164],"best":[166],"trade-off":[167],"between":[168],"word":[171],"error":[172],"rate":[173],"performing":[175],"ASR":[176],"jointly":[177],"endpointing,":[179],"while":[180],"utilizing":[181],"proposed":[184],"alignment":[185],"regularization":[186],"mechanism.":[187]},"counts_by_year":[{"year":2025,"cited_by_count":4},{"year":2024,"cited_by_count":6},{"year":2023,"cited_by_count":11},{"year":2022,"cited_by_count":2},{"year":2021,"cited_by_count":3}],"updated_date":"2026-04-03T22:45:19.894376","created_date":"2025-10-10T00:00:00"}