{"id":"https://openalex.org/W4408353417","doi":"https://doi.org/10.1109/icassp49660.2025.10887855","title":"Accelerating Codec-based Speech Synthesis with Multi-Token Prediction and Speculative Decoding","display_name":"Accelerating Codec-based Speech Synthesis with Multi-Token Prediction and Speculative Decoding","publication_year":2025,"publication_date":"2025-03-12","ids":{"openalex":"https://openalex.org/W4408353417","doi":"https://doi.org/10.1109/icassp49660.2025.10887855"},"language":"en","primary_location":{"id":"doi:10.1109/icassp49660.2025.10887855","is_oa":false,"landing_page_url":"https://doi.org/10.1109/icassp49660.2025.10887855","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"ICASSP 2025 - 2025 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/A5111922287","display_name":"Tan Dat Nguyen","orcid":null},"institutions":[{"id":"https://openalex.org/I157485424","display_name":"Korea Advanced Institute of Science and Technology","ror":"https://ror.org/05apxxy63","country_code":"KR","type":"education","lineage":["https://openalex.org/I157485424"]}],"countries":["KR"],"is_corresponding":true,"raw_author_name":"Tan Dat Nguyen","raw_affiliation_strings":["Korea Advanced Institute of Science and Technology,South Korea"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Korea Advanced Institute of Science and Technology,South Korea","institution_ids":["https://openalex.org/I157485424"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100377802","display_name":"Ji\u2010Hoon Kim","orcid":"https://orcid.org/0000-0002-5212-1686"},"institutions":[{"id":"https://openalex.org/I157485424","display_name":"Korea Advanced Institute of Science and Technology","ror":"https://ror.org/05apxxy63","country_code":"KR","type":"education","lineage":["https://openalex.org/I157485424"]}],"countries":["KR"],"is_corresponding":false,"raw_author_name":"Ji-Hoon Kim","raw_affiliation_strings":["Korea Advanced Institute of Science and Technology,South Korea"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Korea Advanced Institute of Science and Technology,South Korea","institution_ids":["https://openalex.org/I157485424"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5101725649","display_name":"Jeongsoo Choi","orcid":"https://orcid.org/0009-0005-6817-604X"},"institutions":[{"id":"https://openalex.org/I157485424","display_name":"Korea Advanced Institute of Science and Technology","ror":"https://ror.org/05apxxy63","country_code":"KR","type":"education","lineage":["https://openalex.org/I157485424"]}],"countries":["KR"],"is_corresponding":false,"raw_author_name":"Jeongsoo Choi","raw_affiliation_strings":["Korea Advanced Institute of Science and Technology,South Korea"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Korea Advanced Institute of Science and Technology,South Korea","institution_ids":["https://openalex.org/I157485424"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5051189765","display_name":"Shukjae Choi","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Shukjae Choi","raw_affiliation_strings":["42dot Inc,South Korea"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"42dot Inc,South Korea","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5101413982","display_name":"Jinseok Park","orcid":"https://orcid.org/0000-0002-0389-9707"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Jinseok Park","raw_affiliation_strings":["42dot Inc,South Korea"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"42dot Inc,South Korea","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5024990167","display_name":"Younglo Lee","orcid":"https://orcid.org/0000-0002-0725-3444"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Younglo Lee","raw_affiliation_strings":["42dot Inc,South Korea"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"42dot Inc,South Korea","institution_ids":[]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5038723822","display_name":"Joon Son Chung","orcid":"https://orcid.org/0000-0001-7741-7275"},"institutions":[{"id":"https://openalex.org/I157485424","display_name":"Korea Advanced Institute of Science and Technology","ror":"https://ror.org/05apxxy63","country_code":"KR","type":"education","lineage":["https://openalex.org/I157485424"]}],"countries":["KR"],"is_corresponding":false,"raw_author_name":"Joon Son Chung","raw_affiliation_strings":["Korea Advanced Institute of Science and Technology,South Korea"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Korea Advanced Institute of Science and Technology,South Korea","institution_ids":["https://openalex.org/I157485424"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":7,"corresponding_author_ids":["https://openalex.org/A5111922287"],"corresponding_institution_ids":["https://openalex.org/I157485424"],"apc_list":null,"apc_paid":null,"fwci":5.5808,"has_fulltext":false,"cited_by_count":3,"citation_normalized_percentile":{"value":0.95048179,"is_in_top_1_percent":false,"is_in_top_10_percent":true},"cited_by_percentile_year":{"min":98,"max":99},"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":0.9952999949455261,"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":0.9952999949455261,"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":0.982699990272522,"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/T12031","display_name":"Speech and dialogue systems","score":0.9599999785423279,"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/security-token","display_name":"Security token","score":0.8478970527648926},{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.8130907416343689},{"id":"https://openalex.org/keywords/decoding-methods","display_name":"Decoding methods","score":0.8071673512458801},{"id":"https://openalex.org/keywords/codec","display_name":"Codec","score":0.8062702417373657},{"id":"https://openalex.org/keywords/speech-recognition","display_name":"Speech recognition","score":0.5391862988471985},{"id":"https://openalex.org/keywords/adaptive-multi-rate-audio-codec","display_name":"Adaptive Multi-Rate audio codec","score":0.46304136514663696},{"id":"https://openalex.org/keywords/speech-coding","display_name":"Speech coding","score":0.46180716156959534},{"id":"https://openalex.org/keywords/speech-processing","display_name":"Speech processing","score":0.2383250594139099},{"id":"https://openalex.org/keywords/voice-activity-detection","display_name":"Voice activity detection","score":0.21579504013061523},{"id":"https://openalex.org/keywords/computer-network","display_name":"Computer network","score":0.18717169761657715},{"id":"https://openalex.org/keywords/computer-hardware","display_name":"Computer hardware","score":0.15510231256484985},{"id":"https://openalex.org/keywords/algorithm","display_name":"Algorithm","score":0.11946550011634827}],"concepts":[{"id":"https://openalex.org/C48145219","wikidata":"https://www.wikidata.org/wiki/Q1335365","display_name":"Security token","level":2,"score":0.8478970527648926},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8130907416343689},{"id":"https://openalex.org/C57273362","wikidata":"https://www.wikidata.org/wiki/Q576722","display_name":"Decoding methods","level":2,"score":0.8071673512458801},{"id":"https://openalex.org/C161765866","wikidata":"https://www.wikidata.org/wiki/Q184748","display_name":"Codec","level":2,"score":0.8062702417373657},{"id":"https://openalex.org/C28490314","wikidata":"https://www.wikidata.org/wiki/Q189436","display_name":"Speech recognition","level":1,"score":0.5391862988471985},{"id":"https://openalex.org/C177067256","wikidata":"https://www.wikidata.org/wiki/Q4676210","display_name":"Adaptive Multi-Rate audio codec","level":4,"score":0.46304136514663696},{"id":"https://openalex.org/C13895895","wikidata":"https://www.wikidata.org/wiki/Q3270773","display_name":"Speech coding","level":2,"score":0.46180716156959534},{"id":"https://openalex.org/C61328038","wikidata":"https://www.wikidata.org/wiki/Q3358061","display_name":"Speech processing","level":2,"score":0.2383250594139099},{"id":"https://openalex.org/C204201278","wikidata":"https://www.wikidata.org/wiki/Q1332614","display_name":"Voice activity detection","level":3,"score":0.21579504013061523},{"id":"https://openalex.org/C31258907","wikidata":"https://www.wikidata.org/wiki/Q1301371","display_name":"Computer network","level":1,"score":0.18717169761657715},{"id":"https://openalex.org/C9390403","wikidata":"https://www.wikidata.org/wiki/Q3966","display_name":"Computer hardware","level":1,"score":0.15510231256484985},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.11946550011634827}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/icassp49660.2025.10887855","is_oa":false,"landing_page_url":"https://doi.org/10.1109/icassp49660.2025.10887855","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"ICASSP 2025 - 2025 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":47,"referenced_works":["https://openalex.org/W2519091744","https://openalex.org/W2726515241","https://openalex.org/W2972359262","https://openalex.org/W3196468212","https://openalex.org/W3215615641","https://openalex.org/W4225956675","https://openalex.org/W4243463929","https://openalex.org/W4285605725","https://openalex.org/W4307323391","https://openalex.org/W4387968164","https://openalex.org/W4389524500","https://openalex.org/W4391833199","https://openalex.org/W4392903704","https://openalex.org/W4392904136","https://openalex.org/W4402672020","https://openalex.org/W4404740148","https://openalex.org/W4406417959","https://openalex.org/W4408355207","https://openalex.org/W4409362462","https://openalex.org/W4409362648","https://openalex.org/W6749555683","https://openalex.org/W6763832098","https://openalex.org/W6771467084","https://openalex.org/W6772383348","https://openalex.org/W6795261426","https://openalex.org/W6796464841","https://openalex.org/W6796730497","https://openalex.org/W6847386241","https://openalex.org/W6848735303","https://openalex.org/W6849530321","https://openalex.org/W6853165267","https://openalex.org/W6853188576","https://openalex.org/W6853249747","https://openalex.org/W6853530687","https://openalex.org/W6853998256","https://openalex.org/W6855885476","https://openalex.org/W6856434366","https://openalex.org/W6859859934","https://openalex.org/W6860434696","https://openalex.org/W6862144568","https://openalex.org/W6864712218","https://openalex.org/W6866506555","https://openalex.org/W6868809805","https://openalex.org/W6869393727","https://openalex.org/W6869585042","https://openalex.org/W6870225445","https://openalex.org/W6871686990"],"related_works":["https://openalex.org/W4252424250","https://openalex.org/W2370747337","https://openalex.org/W2374470469","https://openalex.org/W2129913000","https://openalex.org/W2899318368","https://openalex.org/W2363660481","https://openalex.org/W2352448339","https://openalex.org/W2159899134","https://openalex.org/W2787189739","https://openalex.org/W3095556352"],"abstract_inverted_index":{"The":[0],"goal":[1],"of":[2,50,69,90,115,132],"this":[3],"paper":[4],"is":[5,42,110],"to":[6,15,43,85,106,117,120],"accelerate":[7],"codec-based":[8],"speech":[9,16,133],"synthesis":[10,64],"systems":[11],"with":[12,123],"minimum":[13],"sacrifice":[14],"quality.":[17],"We":[18],"propose":[19],"an":[20],"enhanced":[21],"inference":[22,34,48],"method":[23],"that":[24,80,102],"allows":[25],"for":[26],"flexible":[27],"trade-offs":[28],"between":[29],"speed":[30],"and":[31],"quality":[32,125],"during":[33],"without":[35],"requiring":[36],"additional":[37],"training.":[38],"Our":[39],"core":[40],"idea":[41],"predict":[44,107],"multiple":[45,55],"tokens":[46,92],"per":[47],"step":[49],"the":[51,67,87,103],"AR":[52],"module":[53],"using":[54],"prediction":[56],"heads,":[57],"resulting":[58],"in":[59,63,130],"a":[60,75,82,113],"linear":[61],"reduction":[62],"time":[65,104],"as":[66],"number":[68],"heads":[70],"increases.":[71],"Furthermore,":[72],"we":[73,100],"introduce":[74],"novel":[76],"speculative":[77],"decoding":[78,95],"technique":[79],"utilises":[81],"Viterbi-based":[83],"algorithm":[84],"select":[86],"optimal":[88],"sequence":[89],"generated":[91],"at":[93],"each":[94,108],"step.":[96],"In":[97],"our":[98],"experiments,":[99],"demonstrate":[101],"required":[105],"token":[109],"reduced":[111],"by":[112],"factor":[114],"4":[116],"5":[118],"compared":[119],"baseline":[121],"models,":[122],"minimal":[124],"trade-off":[126],"or":[127],"even":[128],"improvement":[129],"terms":[131],"intelligibility.":[134],"Audio":[135],"samples":[136],"are":[137],"available":[138],"at:":[139],"mm.kaist.ac.kr/projects/mtp.":[140]},"counts_by_year":[{"year":2026,"cited_by_count":3}],"updated_date":"2026-05-28T09:10:13.091523","created_date":"2025-10-10T00:00:00"}