{"id":"https://openalex.org/W4406461867","doi":"https://doi.org/10.1109/slt61566.2024.10832333","title":"Label-Looping: Highly Efficient Decoding For Transducers","display_name":"Label-Looping: Highly Efficient Decoding For Transducers","publication_year":2024,"publication_date":"2024-12-02","ids":{"openalex":"https://openalex.org/W4406461867","doi":"https://doi.org/10.1109/slt61566.2024.10832333"},"language":"en","primary_location":{"id":"doi:10.1109/slt61566.2024.10832333","is_oa":false,"landing_page_url":"https://doi.org/10.1109/slt61566.2024.10832333","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2024 IEEE Spoken Language Technology Workshop (SLT)","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/A5103270162","display_name":"Vladimir Bataev","orcid":"https://orcid.org/0009-0005-7845-5042"},"institutions":[],"countries":[],"is_corresponding":true,"raw_author_name":"Vladimir Bataev","raw_affiliation_strings":["NVIDIA"],"affiliations":[{"raw_affiliation_string":"NVIDIA","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5034399968","display_name":"Hainan Xu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Hainan Xu","raw_affiliation_strings":["NVIDIA"],"affiliations":[{"raw_affiliation_string":"NVIDIA","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5007998626","display_name":"Daniel G\u00e1lvez","orcid":"https://orcid.org/0000-0001-6119-4310"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Daniel Galvez","raw_affiliation_strings":["NVIDIA"],"affiliations":[{"raw_affiliation_string":"NVIDIA","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5026088310","display_name":"Vitaly Lavrukhin","orcid":"https://orcid.org/0009-0006-7866-8301"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Vitaly Lavrukhin","raw_affiliation_strings":["NVIDIA"],"affiliations":[{"raw_affiliation_string":"NVIDIA","institution_ids":[]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5032957280","display_name":"Boris Ginsburg","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Boris Ginsburg","raw_affiliation_strings":["NVIDIA"],"affiliations":[{"raw_affiliation_string":"NVIDIA","institution_ids":[]}]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":5,"corresponding_author_ids":["https://openalex.org/A5103270162"],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":0.6891,"has_fulltext":false,"cited_by_count":2,"citation_normalized_percentile":{"value":0.77772102,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":{"min":95,"max":96},"biblio":{"volume":null,"issue":null,"first_page":"7","last_page":"13"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10201","display_name":"Speech Recognition and Synthesis","score":0.9957000017166138,"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":0.9957000017166138,"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/T11269","display_name":"Algorithms and Data Compression","score":0.9950000047683716,"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":0.9901000261306763,"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/decoding-methods","display_name":"Decoding methods","score":0.8047600984573364},{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.6804499626159668},{"id":"https://openalex.org/keywords/transducer","display_name":"Transducer","score":0.5143684148788452},{"id":"https://openalex.org/keywords/algorithm","display_name":"Algorithm","score":0.3023255169391632},{"id":"https://openalex.org/keywords/electrical-engineering","display_name":"Electrical engineering","score":0.1183408796787262},{"id":"https://openalex.org/keywords/engineering","display_name":"Engineering","score":0.1034812331199646}],"concepts":[{"id":"https://openalex.org/C57273362","wikidata":"https://www.wikidata.org/wiki/Q576722","display_name":"Decoding methods","level":2,"score":0.8047600984573364},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6804499626159668},{"id":"https://openalex.org/C56318395","wikidata":"https://www.wikidata.org/wiki/Q215928","display_name":"Transducer","level":2,"score":0.5143684148788452},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.3023255169391632},{"id":"https://openalex.org/C119599485","wikidata":"https://www.wikidata.org/wiki/Q43035","display_name":"Electrical engineering","level":1,"score":0.1183408796787262},{"id":"https://openalex.org/C127413603","wikidata":"https://www.wikidata.org/wiki/Q11023","display_name":"Engineering","level":0,"score":0.1034812331199646}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/slt61566.2024.10832333","is_oa":false,"landing_page_url":"https://doi.org/10.1109/slt61566.2024.10832333","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2024 IEEE Spoken Language Technology Workshop (SLT)","raw_type":"proceedings-article"}],"best_oa_location":null,"sustainable_development_goals":[{"id":"https://metadata.un.org/sdg/7","score":0.5600000023841858,"display_name":"Affordable and clean energy"}],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":23,"referenced_works":["https://openalex.org/W1494198834","https://openalex.org/W2127141656","https://openalex.org/W2962780374","https://openalex.org/W2962784628","https://openalex.org/W3005302685","https://openalex.org/W3015190365","https://openalex.org/W3097777922","https://openalex.org/W3167533889","https://openalex.org/W3198643121","https://openalex.org/W4295312788","https://openalex.org/W4375869165","https://openalex.org/W4375869197","https://openalex.org/W4385245566","https://openalex.org/W4388017359","https://openalex.org/W4391021542","https://openalex.org/W4402111732","https://openalex.org/W6638749077","https://openalex.org/W6739901393","https://openalex.org/W6766978945","https://openalex.org/W6767671539","https://openalex.org/W6775607357","https://openalex.org/W6851766456","https://openalex.org/W6929543497"],"related_works":["https://openalex.org/W4391375266","https://openalex.org/W2899084033","https://openalex.org/W2748952813","https://openalex.org/W2051487156","https://openalex.org/W2012283803","https://openalex.org/W4384820447","https://openalex.org/W2072454424","https://openalex.org/W2117438306","https://openalex.org/W2185942010","https://openalex.org/W2260725127"],"abstract_inverted_index":{"This":[0],"paper":[1],"introduces":[2],"a":[3,54],"highly":[4],"efficient":[5],"greedy":[6],"decoding":[7,78],"algorithm":[8,69,102],"for":[9,20,43],"Transducer-based":[10],"speech":[11],"recognition":[12],"models.":[13],"We":[14,115],"redesign":[15],"the":[16,29,36,44,67,121],"standard":[17],"nested-loop":[18],"design":[19],"RNN-T":[21],"decoding,":[22],"swapping":[23],"loops":[24],"over":[25,33,40],"frames":[26,41],"and":[27,105,112],"labels:":[28],"outer":[30],"loop":[31,38],"iterates":[32,39],"labels,":[34],"while":[35],"inner":[37],"searching":[42],"next":[45],"non-blank":[46],"symbol.":[47],"Additionally,":[48],"we":[49],"represent":[50],"partial":[51],"hypotheses":[52,62],"in":[53],"special":[55],"structure":[56],"using":[57,80],"CUDA":[58],"tensors,":[59],"supporting":[60],"parallelized":[61],"manipulations.":[63],"Experiments":[64],"show":[65],"that":[66],"label-looping":[68],"is":[70,103],"up":[71],"to":[72,96,119],"2.0X":[73],"faster":[74],"than":[75],"conventional":[76,110],"batched":[77],"when":[79],"batch":[81],"size":[82],"32.":[83],"It":[84],"can":[85,106],"be":[86],"further":[87],"combined":[88],"with":[89,108],"other":[90],"compiler":[91],"or":[92],"GPU":[93],"call-related":[94],"techniques":[95],"achieve":[97],"even":[98],"more":[99],"speedup.":[100],"Our":[101],"general-purpose":[104],"work":[107],"both":[109],"Transducers":[111],"Token-and-Duration":[113],"Transducers.":[114],"open-source":[116],"our":[117],"implementation":[118],"benefit":[120],"research":[122],"community.":[123]},"counts_by_year":[{"year":2025,"cited_by_count":2}],"updated_date":"2026-04-03T22:45:19.894376","created_date":"2025-10-10T00:00:00"}