{"id":"https://openalex.org/W3197956343","doi":"https://doi.org/10.21437/interspeech.2021-322","title":"Reducing Streaming ASR Model Delay with Self Alignment","display_name":"Reducing Streaming ASR Model Delay with Self Alignment","publication_year":2021,"publication_date":"2021-08-27","ids":{"openalex":"https://openalex.org/W3197956343","doi":"https://doi.org/10.21437/interspeech.2021-322","mag":"3197956343"},"language":"en","primary_location":{"id":"doi:10.21437/interspeech.2021-322","is_oa":false,"landing_page_url":"https://doi.org/10.21437/interspeech.2021-322","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/A5100344655","display_name":"Jaeyoung Kim","orcid":"https://orcid.org/0000-0002-4830-0454"},"institutions":[],"countries":[],"is_corresponding":true,"raw_author_name":"Jaeyoung Kim","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5101836793","display_name":"Lu Han","orcid":"https://orcid.org/0000-0002-1753-0732"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Han Lu","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5101403998","display_name":"Anshuman Tripathi","orcid":"https://orcid.org/0000-0002-4902-3719"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Anshuman Tripathi","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5050962927","display_name":"Qian Zhang","orcid":"https://orcid.org/0000-0003-3524-2109"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Qian Zhang","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5108537573","display_name":"Ha\u015fim Sak","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Hasim Sak","raw_affiliation_strings":[],"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":5,"corresponding_author_ids":["https://openalex.org/A5100344655"],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":1.3713,"has_fulltext":false,"cited_by_count":9,"citation_normalized_percentile":{"value":0.81501014,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":{"min":89,"max":97},"biblio":{"volume":null,"issue":null,"first_page":"3440","last_page":"3444"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T11309","display_name":"Music and Audio Processing","score":0.9983000159263611,"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"}},"topics":[{"id":"https://openalex.org/T11309","display_name":"Music and Audio Processing","score":0.9983000159263611,"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/T10201","display_name":"Speech Recognition and Synthesis","score":0.9979000091552734,"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":0.9940000176429749,"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":0.813640296459198},{"id":"https://openalex.org/keywords/viterbi-algorithm","display_name":"Viterbi algorithm","score":0.6616686582565308},{"id":"https://openalex.org/keywords/latency","display_name":"Latency (audio)","score":0.6496040225028992},{"id":"https://openalex.org/keywords/regularization","display_name":"Regularization (linguistics)","score":0.5413815975189209},{"id":"https://openalex.org/keywords/word-error-rate","display_name":"Word error rate","score":0.5389277935028076},{"id":"https://openalex.org/keywords/speech-recognition","display_name":"Speech recognition","score":0.535966157913208},{"id":"https://openalex.org/keywords/vocabulary","display_name":"Vocabulary","score":0.4967578053474426},{"id":"https://openalex.org/keywords/low-latency","display_name":"Low latency (capital markets)","score":0.4686547815799713},{"id":"https://openalex.org/keywords/word","display_name":"Word (group theory)","score":0.417802631855011},{"id":"https://openalex.org/keywords/algorithm","display_name":"Algorithm","score":0.369423508644104},{"id":"https://openalex.org/keywords/hidden-markov-model","display_name":"Hidden Markov model","score":0.32113227248191833},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.2983953356742859},{"id":"https://openalex.org/keywords/computer-network","display_name":"Computer network","score":0.11161667108535767},{"id":"https://openalex.org/keywords/mathematics","display_name":"Mathematics","score":0.10163167119026184},{"id":"https://openalex.org/keywords/telecommunications","display_name":"Telecommunications","score":0.08360815048217773}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.813640296459198},{"id":"https://openalex.org/C60582962","wikidata":"https://www.wikidata.org/wiki/Q83886","display_name":"Viterbi algorithm","level":3,"score":0.6616686582565308},{"id":"https://openalex.org/C82876162","wikidata":"https://www.wikidata.org/wiki/Q17096504","display_name":"Latency (audio)","level":2,"score":0.6496040225028992},{"id":"https://openalex.org/C2776135515","wikidata":"https://www.wikidata.org/wiki/Q17143721","display_name":"Regularization (linguistics)","level":2,"score":0.5413815975189209},{"id":"https://openalex.org/C40969351","wikidata":"https://www.wikidata.org/wiki/Q3516228","display_name":"Word error rate","level":2,"score":0.5389277935028076},{"id":"https://openalex.org/C28490314","wikidata":"https://www.wikidata.org/wiki/Q189436","display_name":"Speech recognition","level":1,"score":0.535966157913208},{"id":"https://openalex.org/C2777601683","wikidata":"https://www.wikidata.org/wiki/Q6499736","display_name":"Vocabulary","level":2,"score":0.4967578053474426},{"id":"https://openalex.org/C46637626","wikidata":"https://www.wikidata.org/wiki/Q6693015","display_name":"Low latency (capital markets)","level":2,"score":0.4686547815799713},{"id":"https://openalex.org/C90805587","wikidata":"https://www.wikidata.org/wiki/Q10944557","display_name":"Word (group theory)","level":2,"score":0.417802631855011},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.369423508644104},{"id":"https://openalex.org/C23224414","wikidata":"https://www.wikidata.org/wiki/Q176769","display_name":"Hidden Markov model","level":2,"score":0.32113227248191833},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.2983953356742859},{"id":"https://openalex.org/C31258907","wikidata":"https://www.wikidata.org/wiki/Q1301371","display_name":"Computer network","level":1,"score":0.11161667108535767},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.10163167119026184},{"id":"https://openalex.org/C76155785","wikidata":"https://www.wikidata.org/wiki/Q418","display_name":"Telecommunications","level":1,"score":0.08360815048217773},{"id":"https://openalex.org/C41895202","wikidata":"https://www.wikidata.org/wiki/Q8162","display_name":"Linguistics","level":1,"score":0.0},{"id":"https://openalex.org/C138885662","wikidata":"https://www.wikidata.org/wiki/Q5891","display_name":"Philosophy","level":0,"score":0.0},{"id":"https://openalex.org/C2524010","wikidata":"https://www.wikidata.org/wiki/Q8087","display_name":"Geometry","level":1,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.21437/interspeech.2021-322","is_oa":false,"landing_page_url":"https://doi.org/10.21437/interspeech.2021-322","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":[{"score":0.6000000238418579,"id":"https://metadata.un.org/sdg/4","display_name":"Quality Education"}],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W2136652457","https://openalex.org/W2169849734","https://openalex.org/W2116722627","https://openalex.org/W2129150969","https://openalex.org/W2236912844","https://openalex.org/W1975869217","https://openalex.org/W2401728283","https://openalex.org/W2383829109","https://openalex.org/W2379938888","https://openalex.org/W2386035178"],"abstract_inverted_index":{"Reducing":[0],"prediction":[1],"delay":[2,42,74,84,126,145],"for":[3],"streaming":[4],"end-to-end":[5],"ASR":[6,63],"models":[7],"with":[8,155],"minimal":[9],"performance":[10],"regression":[11],"is":[12,18,39],"a":[13,19,40,82],"challenging":[14],"problem.":[15],"Constrained":[16],"alignment":[17,91,96,112,118,132,154],"well-known":[20],"existing":[21,120],"approach":[22],"that":[23],"penalizes":[24],"predicted":[25],"word":[26,64,136],"boundaries":[27],"using":[28],"external":[29,95],"low-latency":[30],"acoustic":[31],"models.":[32,97],"On":[33],"the":[34,104,109,134],"contrary,":[35],"recently":[36],"proposed":[37],"FastEmit":[38,129,151],"sequence-level":[41],"regularization":[43],"scheme":[44],"encouraging":[45],"vocabulary":[46],"tokens":[47],"over":[48],"blanks":[49],"without":[50],"any":[51],"reference":[52],"alignments.":[53],"Although":[54],"all":[55],"these":[56,73],"schemes":[57],"are":[58],"successful":[59],"in":[60],"reducing":[61],"delay,":[62],"error":[65,137],"rate":[66],"(WER)":[67],"often":[68],"severely":[69],"degrades":[70],"after":[71],"applying":[72],"constraining":[75,85],"schemes.":[76],"In":[77],"this":[78],"paper,":[79],"we":[80],"propose":[81],"novel":[83],"method,":[86],"named":[87],"self":[88,117],"alignment.":[89],"Self":[90],"does":[92],"not":[93],"require":[94],"Instead,":[98],"it":[99],"utilizes":[100],"Viterbi":[101],"forced-alignments":[102],"from":[103],"trained":[105],"model":[106],"to":[107,128,150],"find":[108],"lower":[110],"latency":[111],"direction.":[113],"From":[114],"LibriSpeech":[115],"evaluation,":[116],"outperformed":[119],"schemes:":[121],"25%":[122,144],"and":[123,130,143,152],"56%":[124],"less":[125],"compared":[127,149],"constrained":[131,153],"at":[133],"similar":[135],"rate.":[138],"For":[139],"Voice":[140],"Search":[141],"evaluation,12%":[142],"reductions":[146],"were":[147],"achieved":[148],"more":[156],"than":[157],"2%":[158],"WER":[159],"improvements.":[160]},"counts_by_year":[{"year":2024,"cited_by_count":1},{"year":2023,"cited_by_count":4},{"year":2022,"cited_by_count":3},{"year":2021,"cited_by_count":1}],"updated_date":"2025-11-06T03:46:38.306776","created_date":"2025-10-10T00:00:00"}