{"id":"https://openalex.org/W7164022019","doi":"https://doi.org/10.48550/arxiv.2606.09234","title":"End-to-End Training for Discrete Token LLM based TTS System","display_name":"End-to-End Training for Discrete Token LLM based TTS System","publication_year":2026,"publication_date":"2026-06-08","ids":{"openalex":"https://openalex.org/W7164022019","doi":"https://doi.org/10.48550/arxiv.2606.09234"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2606.09234","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.09234","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"},"type":"preprint","indexed_in":["datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://doi.org/10.48550/arxiv.2606.09234","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5006453801","display_name":"Changfeng Gao","orcid":"https://orcid.org/0000-0002-8148-4099"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Gao, Changfeng","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5138206396","display_name":"Yong Ren","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Ren, Yong","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5138264981","display_name":"Jun Yuan","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yuan, Jun","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5138239326","display_name":"Ye Bai","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Bai, Ye","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5101967525","display_name":"Zhao You","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"You, Zhao","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5078353046","display_name":"Shidong Shang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Shang, ShiDong","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":0,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":false,"cited_by_count":0,"citation_normalized_percentile":null,"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":null,"last_page":null},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10201","display_name":"Speech Recognition and Synthesis","score":0.9294000267982483,"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":0.9294000267982483,"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":0.025299999862909317,"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/T10863","display_name":"Voice and Speech Disorders","score":0.010300000198185444,"subfield":{"id":"https://openalex.org/subfields/2737","display_name":"Physiology"},"field":{"id":"https://openalex.org/fields/27","display_name":"Medicine"},"domain":{"id":"https://openalex.org/domains/4","display_name":"Health Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/security-token","display_name":"Security token","score":0.7172999978065491},{"id":"https://openalex.org/keywords/pipeline","display_name":"Pipeline (software)","score":0.6711000204086304},{"id":"https://openalex.org/keywords/task","display_name":"Task (project management)","score":0.4918999969959259},{"id":"https://openalex.org/keywords/word","display_name":"Word (group theory)","score":0.4726000130176544},{"id":"https://openalex.org/keywords/language-model","display_name":"Language model","score":0.47130000591278076},{"id":"https://openalex.org/keywords/word-error-rate","display_name":"Word error rate","score":0.4066999852657318},{"id":"https://openalex.org/keywords/training","display_name":"Training (meteorology)","score":0.40619999170303345}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8324999809265137},{"id":"https://openalex.org/C48145219","wikidata":"https://www.wikidata.org/wiki/Q1335365","display_name":"Security token","level":2,"score":0.7172999978065491},{"id":"https://openalex.org/C43521106","wikidata":"https://www.wikidata.org/wiki/Q2165493","display_name":"Pipeline (software)","level":2,"score":0.6711000204086304},{"id":"https://openalex.org/C28490314","wikidata":"https://www.wikidata.org/wiki/Q189436","display_name":"Speech recognition","level":1,"score":0.6409000158309937},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.517799973487854},{"id":"https://openalex.org/C2780451532","wikidata":"https://www.wikidata.org/wiki/Q759676","display_name":"Task (project management)","level":2,"score":0.4918999969959259},{"id":"https://openalex.org/C90805587","wikidata":"https://www.wikidata.org/wiki/Q10944557","display_name":"Word (group theory)","level":2,"score":0.4726000130176544},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.47130000591278076},{"id":"https://openalex.org/C40969351","wikidata":"https://www.wikidata.org/wiki/Q3516228","display_name":"Word error rate","level":2,"score":0.4066999852657318},{"id":"https://openalex.org/C2777211547","wikidata":"https://www.wikidata.org/wiki/Q17141490","display_name":"Training (meteorology)","level":2,"score":0.40619999170303345},{"id":"https://openalex.org/C2780719617","wikidata":"https://www.wikidata.org/wiki/Q1030752","display_name":"Salient","level":2,"score":0.3935000002384186},{"id":"https://openalex.org/C18555067","wikidata":"https://www.wikidata.org/wiki/Q8375051","display_name":"Joint (building)","level":2,"score":0.3560999929904938},{"id":"https://openalex.org/C159877910","wikidata":"https://www.wikidata.org/wiki/Q2202883","display_name":"Autoregressive model","level":2,"score":0.34779998660087585},{"id":"https://openalex.org/C2778572836","wikidata":"https://www.wikidata.org/wiki/Q380933","display_name":"Space (punctuation)","level":2,"score":0.3093999922275543},{"id":"https://openalex.org/C51632099","wikidata":"https://www.wikidata.org/wiki/Q3985153","display_name":"Training set","level":2,"score":0.3061000108718872},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.2782000005245209},{"id":"https://openalex.org/C61328038","wikidata":"https://www.wikidata.org/wiki/Q3358061","display_name":"Speech processing","level":2,"score":0.2669000029563904},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.25049999356269836}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2606.09234","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.09234","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":null,"is_accepted":false,"is_published":null,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"doi:10.48550/arxiv.2606.09234","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.09234","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Recent":[0],"state-of-the-art":[1],"(SOTA)":[2],"text-to-speech":[3],"(TTS)":[4],"systems":[5,188],"typically":[6],"adopt":[7],"a":[8,13,23,39,154,163,168,190],"cascaded":[9,145],"pipeline":[10],"consisting":[11],"of":[12,49,159],"speech":[14,51,93],"tokenizer,":[15,52],"an":[16,57],"autoregressive":[17],"large":[18],"language":[19],"model":[20,60],"(LLM),":[21],"and":[22,56,81,99,118,122,128,161,171],"diffusion":[24],"based":[25],"flow-matching":[26],"(FM)":[27],"model,":[28,55],"with":[29,167,189],"these":[30],"components":[31],"trained":[32],"independently.":[33],"In":[34],"this":[35],"paper,":[36],"we":[37,63],"propose":[38],"fully":[40],"end-to-end":[41],"(E2E)":[42],"optimization":[43,181],"framework":[44,142],"that":[45,103,139,178],"unifies":[46],"the":[47,50,67,91,113,130,148],"training":[48,89,193],"LLM,":[53,80],"FM":[54,121,173],"additional":[58],"reward":[59],"(RM).":[61],"Specifically,":[62],"first":[64],"jointly":[65],"optimize":[66,112],"tokenizer":[68],"using":[69,115],"multi-task":[70],"objectives":[71],"derived":[72],"from":[73],"reconstruction":[74,117],"for":[75,79,85,184],"FM,":[76],"next-token":[77],"prediction":[78],"multi":[82],"recognition":[83,119],"task":[84],"RM.":[86],"This":[87],"joint":[88],"encourages":[90],"discrete":[92],"token":[94],"space":[95],"to":[96,107],"capture":[97],"acoustically":[98],"semantically":[100],"salient":[101],"information":[102],"is":[104,182],"better":[105],"tailored":[106],"TTS.":[108],"We":[109],"then":[110],"further":[111],"LLM":[114,131,170],"downstream":[116],"by":[120],"RM,":[123],"which":[124],"reduces":[125],"inference-time":[126],"mismatch":[127],"steers":[129],"toward":[132],"more":[133],"preferred":[134],"generations.":[135],"Experimental":[136],"results":[137,176],"show":[138],"our":[140,151],"E2E":[141,180],"consistently":[143],"outperforms":[144],"baselines.":[146],"On":[147],"Seed-TTS-Eval":[149],"benchmark,":[150],"system":[152],"achieves":[153],"word":[155],"error":[156],"rate":[157],"(WER)":[158],"0.78%":[160],"1.56%,":[162],"new":[164],"SOTA":[165],"result":[166],"0.6B-parameter":[169],"0.5B-parameter":[172],"model.":[174],"These":[175],"validate":[177],"holistic":[179],"critical":[183],"improving":[185],"discrete-token-based":[186],"TTS":[187],"much":[191],"simpler":[192],"pipeline.":[194]},"counts_by_year":[],"updated_date":"2026-07-01T06:00:48.157686","created_date":"2026-06-10T00:00:00"}