{"id":"https://openalex.org/W4372346832","doi":"https://doi.org/10.1109/icassp49357.2023.10095377","title":"Peak-First CTC: Reducing the Peak Latency of CTC Models by Applying Peak-First Regularization","display_name":"Peak-First CTC: Reducing the Peak Latency of CTC Models by Applying Peak-First Regularization","publication_year":2023,"publication_date":"2023-05-05","ids":{"openalex":"https://openalex.org/W4372346832","doi":"https://doi.org/10.1109/icassp49357.2023.10095377"},"language":"en","primary_location":{"id":"doi:10.1109/icassp49357.2023.10095377","is_oa":false,"landing_page_url":"https://doi.org/10.1109/icassp49357.2023.10095377","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"ICASSP 2023 - 2023 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/A5101817056","display_name":"Zhengkun Tian","orcid":"https://orcid.org/0000-0002-0469-3049"},"institutions":[{"id":"https://openalex.org/I4210087373","display_name":"Meizu (China)","ror":"https://ror.org/0067g4302","country_code":"CN","type":"company","lineage":["https://openalex.org/I4210087373"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Zhengkun Tian","raw_affiliation_strings":["Meituan,Beijing,China","Meituan, Beijing, China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Meituan,Beijing,China","institution_ids":["https://openalex.org/I4210087373"]},{"raw_affiliation_string":"Meituan, Beijing, China","institution_ids":["https://openalex.org/I4210087373"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5024508647","display_name":"Hongyu Xiang","orcid":"https://orcid.org/0000-0001-5953-1293"},"institutions":[{"id":"https://openalex.org/I4210087373","display_name":"Meizu (China)","ror":"https://ror.org/0067g4302","country_code":"CN","type":"company","lineage":["https://openalex.org/I4210087373"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Hongyu Xiang","raw_affiliation_strings":["Meituan,Beijing,China","Meituan, Beijing, China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Meituan,Beijing,China","institution_ids":["https://openalex.org/I4210087373"]},{"raw_affiliation_string":"Meituan, Beijing, China","institution_ids":["https://openalex.org/I4210087373"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100400746","display_name":"Min Li","orcid":"https://orcid.org/0000-0002-5773-8976"},"institutions":[{"id":"https://openalex.org/I4210087373","display_name":"Meizu (China)","ror":"https://ror.org/0067g4302","country_code":"CN","type":"company","lineage":["https://openalex.org/I4210087373"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Min Li","raw_affiliation_strings":["Meituan,Beijing,China","Meituan, Beijing, China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Meituan,Beijing,China","institution_ids":["https://openalex.org/I4210087373"]},{"raw_affiliation_string":"Meituan, Beijing, China","institution_ids":["https://openalex.org/I4210087373"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5101827252","display_name":"Feifei Lin","orcid":"https://orcid.org/0000-0002-2388-5011"},"institutions":[{"id":"https://openalex.org/I4210087373","display_name":"Meizu (China)","ror":"https://ror.org/0067g4302","country_code":"CN","type":"company","lineage":["https://openalex.org/I4210087373"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Feifei Lin","raw_affiliation_strings":["Meituan,Beijing,China","Meituan, Beijing, China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Meituan,Beijing,China","institution_ids":["https://openalex.org/I4210087373"]},{"raw_affiliation_string":"Meituan, Beijing, China","institution_ids":["https://openalex.org/I4210087373"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5086348845","display_name":"Ke Ding","orcid":"https://orcid.org/0000-0003-0653-8816"},"institutions":[{"id":"https://openalex.org/I4210087373","display_name":"Meizu (China)","ror":"https://ror.org/0067g4302","country_code":"CN","type":"company","lineage":["https://openalex.org/I4210087373"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Ke Ding","raw_affiliation_strings":["Meituan,Beijing,China","Meituan, Beijing, China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Meituan,Beijing,China","institution_ids":["https://openalex.org/I4210087373"]},{"raw_affiliation_string":"Meituan, Beijing, China","institution_ids":["https://openalex.org/I4210087373"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5072537055","display_name":"Guanglu Wan","orcid":"https://orcid.org/0009-0003-1061-3724"},"institutions":[{"id":"https://openalex.org/I4210087373","display_name":"Meizu (China)","ror":"https://ror.org/0067g4302","country_code":"CN","type":"company","lineage":["https://openalex.org/I4210087373"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Guanglu Wan","raw_affiliation_strings":["Meituan,Beijing,China","Meituan, Beijing, China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Meituan,Beijing,China","institution_ids":["https://openalex.org/I4210087373"]},{"raw_affiliation_string":"Meituan, Beijing, China","institution_ids":["https://openalex.org/I4210087373"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":["https://openalex.org/I4210087373"],"apc_list":null,"apc_paid":null,"fwci":0.4839,"has_fulltext":false,"cited_by_count":3,"citation_normalized_percentile":{"value":0.6931612,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":{"min":90,"max":96},"biblio":{"volume":null,"issue":null,"first_page":"1","last_page":"5"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10201","display_name":"Speech Recognition and Synthesis","score":0.9993000030517578,"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":0.9993000030517578,"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":0.9972000122070312,"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/T10028","display_name":"Topic Modeling","score":0.9909999966621399,"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":0.7373227477073669},{"id":"https://openalex.org/keywords/regularization","display_name":"Regularization (linguistics)","score":0.6659702062606812},{"id":"https://openalex.org/keywords/latency","display_name":"Latency (audio)","score":0.6637349128723145},{"id":"https://openalex.org/keywords/security-token","display_name":"Security token","score":0.5907340049743652},{"id":"https://openalex.org/keywords/inference","display_name":"Inference","score":0.5586888790130615},{"id":"https://openalex.org/keywords/algorithm","display_name":"Algorithm","score":0.46065470576286316},{"id":"https://openalex.org/keywords/probability-distribution","display_name":"Probability distribution","score":0.4219302535057068},{"id":"https://openalex.org/keywords/pattern-recognition","display_name":"Pattern recognition (psychology)","score":0.35950973629951477},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.3594258427619934},{"id":"https://openalex.org/keywords/mathematics","display_name":"Mathematics","score":0.18338972330093384},{"id":"https://openalex.org/keywords/statistics","display_name":"Statistics","score":0.12459692358970642}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7373227477073669},{"id":"https://openalex.org/C2776135515","wikidata":"https://www.wikidata.org/wiki/Q17143721","display_name":"Regularization (linguistics)","level":2,"score":0.6659702062606812},{"id":"https://openalex.org/C82876162","wikidata":"https://www.wikidata.org/wiki/Q17096504","display_name":"Latency (audio)","level":2,"score":0.6637349128723145},{"id":"https://openalex.org/C48145219","wikidata":"https://www.wikidata.org/wiki/Q1335365","display_name":"Security token","level":2,"score":0.5907340049743652},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.5586888790130615},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.46065470576286316},{"id":"https://openalex.org/C149441793","wikidata":"https://www.wikidata.org/wiki/Q200726","display_name":"Probability distribution","level":2,"score":0.4219302535057068},{"id":"https://openalex.org/C153180895","wikidata":"https://www.wikidata.org/wiki/Q7148389","display_name":"Pattern recognition (psychology)","level":2,"score":0.35950973629951477},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.3594258427619934},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.18338972330093384},{"id":"https://openalex.org/C105795698","wikidata":"https://www.wikidata.org/wiki/Q12483","display_name":"Statistics","level":1,"score":0.12459692358970642},{"id":"https://openalex.org/C76155785","wikidata":"https://www.wikidata.org/wiki/Q418","display_name":"Telecommunications","level":1,"score":0.0},{"id":"https://openalex.org/C38652104","wikidata":"https://www.wikidata.org/wiki/Q3510521","display_name":"Computer security","level":1,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/icassp49357.2023.10095377","is_oa":false,"landing_page_url":"https://doi.org/10.1109/icassp49357.2023.10095377","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"ICASSP 2023 - 2023 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)","raw_type":"proceedings-article"}],"best_oa_location":null,"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":25,"referenced_works":["https://openalex.org/W2127141656","https://openalex.org/W2143612262","https://openalex.org/W2144499799","https://openalex.org/W2291513470","https://openalex.org/W2327501763","https://openalex.org/W2526425061","https://openalex.org/W2899663614","https://openalex.org/W2911291251","https://openalex.org/W2936774411","https://openalex.org/W2962760690","https://openalex.org/W2962780374","https://openalex.org/W2973215447","https://openalex.org/W3015537910","https://openalex.org/W3097874139","https://openalex.org/W3160766462","https://openalex.org/W3163907627","https://openalex.org/W3197956343","https://openalex.org/W3198260227","https://openalex.org/W3198643121","https://openalex.org/W4296069356","https://openalex.org/W4306705025","https://openalex.org/W4385245566","https://openalex.org/W6687566353","https://openalex.org/W6755977528","https://openalex.org/W6845608264"],"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/W4389115113","https://openalex.org/W3128807919","https://openalex.org/W3176411177"],"abstract_inverted_index":{"The":[0,42,95,180],"CTC":[1,33,46,131,148,177],"model":[2,54,132],"has":[3],"been":[4],"widely":[5],"applied":[6],"to":[7,55,61,100,124,133,196],"many":[8,24],"application":[9],"scenarios":[10],"because":[11],"of":[12,45,80,129,141,147,168,203],"its":[13],"simple":[14,110],"structure,":[15],"excellent":[16],"performance,":[17],"and":[18,35,75,111,150,175],"fast":[19],"inference":[20],"speed.":[21],"There":[22],"are":[23,98,155],"peaks":[25,57],"in":[26,71],"the":[27,32,53,66,72,76,86,104,126,130,137,144,153,166,169,184,189],"probability":[28,127],"distribution":[29,128],"predicted":[30],"by":[31,51,91,193],"models,":[34],"each":[36],"peak":[37,105,191],"represents":[38],"a":[39,109,119,158],"non-blank":[40],"token.":[41],"recognition":[43,204],"latency":[44,63,192],"models":[47,178],"can":[48,187],"be":[49],"reduced":[50],"encouraging":[52],"predict":[56],"earlier.":[58],"Existing":[59],"methods":[60,82,97],"reduce":[62,103,188],"require":[64],"modifying":[65,143],"transition":[67],"relationship":[68],"between":[69],"tokens":[70],"forward-backward":[73],"algorithm,":[74],"gradient":[77],"calculation.":[78],"Some":[79],"these":[81],"even":[83],"depend":[84],"on":[85,157,172],"forced":[87],"alignment":[88],"results":[89,181],"provided":[90],"other":[92],"pretrained":[93],"models.":[94],"above":[96],"complex":[99],"implement.":[101],"To":[102],"latency,":[106],"we":[107],"propose":[108],"novel":[112],"method":[113,186],"named":[114],"peak-first":[115],"regularization,":[116],"which":[117],"utilizes":[118],"frame-wise":[120],"knowledge":[121],"distillation":[122],"function":[123],"force":[125],"shift":[134],"left":[135],"along":[136],"time":[138],"axis":[139],"instead":[140],"directly":[142],"calculation":[145],"process":[146],"loss":[149],"gradients.":[151],"All":[152],"experiments":[154],"conducted":[156],"Chinese":[159],"Mandarin":[160],"dataset":[161],"AISHELL-1.":[162],"We":[163],"have":[164],"verified":[165],"effectiveness":[167],"proposed":[170,185],"regularization":[171],"both":[173],"streaming":[174],"non-streaming":[176],"respectively.":[179],"show":[182],"that":[183],"average":[190],"about":[194],"100":[195],"200":[197],"milliseconds":[198],"with":[199],"almost":[200],"no":[201],"degradation":[202],"accuracy.":[205]},"counts_by_year":[{"year":2025,"cited_by_count":2},{"year":2024,"cited_by_count":1}],"updated_date":"2026-06-26T08:34:08.712188","created_date":"2025-10-10T00:00:00"}