{"id":"https://openalex.org/W7154134062","doi":"https://doi.org/10.48550/arxiv.2604.08558","title":"WAND: Windowed Attention and Knowledge Distillation for Efficient Autoregressive Text-to-Speech Models","display_name":"WAND: Windowed Attention and Knowledge Distillation for Efficient Autoregressive Text-to-Speech Models","publication_year":2026,"publication_date":"2026-03-17","ids":{"openalex":"https://openalex.org/W7154134062","doi":"https://doi.org/10.48550/arxiv.2604.08558"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2604.08558","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2604.08558","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":null,"license_id":null,"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.2604.08558","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5133548446","display_name":"Hanna Lee","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Lee, Hanna","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5111922287","display_name":"Tan Dat Nguyen","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Nguyen, Tan Dat","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5133501560","display_name":"Jaehoon Kang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Kang, Jaehoon","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5133516512","display_name":"Kyuhong Shim","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Shim, Kyuhong","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":0.6351000070571899,"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":0.6351000070571899,"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/T10181","display_name":"Natural Language Processing Techniques","score":0.07800000160932541,"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/T10028","display_name":"Topic Modeling","score":0.07769999653100967,"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/autoregressive-model","display_name":"Autoregressive model","score":0.6334999799728394},{"id":"https://openalex.org/keywords/sequence","display_name":"Sequence (biology)","score":0.5339000225067139},{"id":"https://openalex.org/keywords/quality","display_name":"Quality (philosophy)","score":0.4765999913215637},{"id":"https://openalex.org/keywords/distillation","display_name":"Distillation","score":0.4523000121116638},{"id":"https://openalex.org/keywords/constant","display_name":"Constant (computer programming)","score":0.44830000400543213},{"id":"https://openalex.org/keywords/reduction","display_name":"Reduction (mathematics)","score":0.4399000108242035},{"id":"https://openalex.org/keywords/scale","display_name":"Scale (ratio)","score":0.4009999930858612}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7760999798774719},{"id":"https://openalex.org/C159877910","wikidata":"https://www.wikidata.org/wiki/Q2202883","display_name":"Autoregressive model","level":2,"score":0.6334999799728394},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5550000071525574},{"id":"https://openalex.org/C2778112365","wikidata":"https://www.wikidata.org/wiki/Q3511065","display_name":"Sequence (biology)","level":2,"score":0.5339000225067139},{"id":"https://openalex.org/C2779530757","wikidata":"https://www.wikidata.org/wiki/Q1207505","display_name":"Quality (philosophy)","level":2,"score":0.4765999913215637},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.4580000042915344},{"id":"https://openalex.org/C204030448","wikidata":"https://www.wikidata.org/wiki/Q101017","display_name":"Distillation","level":2,"score":0.4523000121116638},{"id":"https://openalex.org/C2777027219","wikidata":"https://www.wikidata.org/wiki/Q1284190","display_name":"Constant (computer programming)","level":2,"score":0.44830000400543213},{"id":"https://openalex.org/C111335779","wikidata":"https://www.wikidata.org/wiki/Q3454686","display_name":"Reduction (mathematics)","level":2,"score":0.4399000108242035},{"id":"https://openalex.org/C2778755073","wikidata":"https://www.wikidata.org/wiki/Q10858537","display_name":"Scale (ratio)","level":2,"score":0.4009999930858612},{"id":"https://openalex.org/C115537543","wikidata":"https://www.wikidata.org/wiki/Q165596","display_name":"Cache","level":2,"score":0.35010001063346863},{"id":"https://openalex.org/C195956108","wikidata":"https://www.wikidata.org/wiki/Q7268362","display_name":"Quadratic growth","level":2,"score":0.34389999508857727},{"id":"https://openalex.org/C40506919","wikidata":"https://www.wikidata.org/wiki/Q7452469","display_name":"Sequence learning","level":2,"score":0.27959999442100525},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.27730000019073486},{"id":"https://openalex.org/C189950617","wikidata":"https://www.wikidata.org/wiki/Q937228","display_name":"Property (philosophy)","level":2,"score":0.27219998836517334},{"id":"https://openalex.org/C196083921","wikidata":"https://www.wikidata.org/wiki/Q7915758","display_name":"Variance (accounting)","level":2,"score":0.27160000801086426},{"id":"https://openalex.org/C125411270","wikidata":"https://www.wikidata.org/wiki/Q18653","display_name":"Encoding (memory)","level":2,"score":0.26249998807907104}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2604.08558","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2604.08558","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":null,"license_id":null,"version":null,"is_accepted":false,"is_published":null,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"doi:10.48550/arxiv.2604.08558","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2604.08558","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":null,"license_id":null,"version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"},"sustainable_development_goals":[{"id":"https://metadata.un.org/sdg/4","display_name":"Quality Education","score":0.7438578009605408}],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Recent":[0],"decoder-only":[1],"autoregressive":[2],"text-to-speech":[3],"(AR-TTS)":[4],"models":[5,41],"produce":[6],"high-fidelity":[7,96],"speech,":[8],"but":[9],"their":[10],"memory":[11,48,121],"and":[12,32,47,63,123],"compute":[13],"costs":[14],"scale":[15],"quadratically":[16],"with":[17,44,99],"sequence":[18],"length":[19],"due":[20],"to":[21,42,94,117],"full":[22],"self-attention.":[23],"In":[24],"this":[25],"paper,":[26],"we":[27,73],"propose":[28],"WAND,":[29],"Windowed":[30],"Attention":[31],"Knowledge":[33],"Distillation,":[34],"a":[35,75,91],"framework":[36],"that":[37,79],"adapts":[38],"pretrained":[39],"AR-TTS":[40,107],"operate":[43],"constant":[45],"computational":[46],"complexity.":[49],"WAND":[50,109],"separates":[51],"the":[52,82,111],"attention":[53,59,66,83],"mechanism":[54],"into":[55],"two:":[56],"persistent":[57],"global":[58],"over":[60,67],"conditioning":[61],"tokens":[62],"local":[64],"sliding-window":[65],"generated":[68],"tokens.":[69],"To":[70],"stabilize":[71],"fine-tuning,":[72],"employ":[74],"curriculum":[76],"learning":[77],"strategy":[78],"progressively":[80],"tightens":[81],"window.":[84],"We":[85],"further":[86],"utilize":[87],"knowledge":[88],"distillation":[89],"from":[90],"full-attention":[92],"teacher":[93],"recover":[95],"synthesis":[97],"quality":[98,113],"high":[100],"data":[101],"efficiency.":[102],"Evaluated":[103],"on":[104],"three":[105],"modern":[106],"models,":[108],"preserves":[110],"original":[112],"while":[114],"achieving":[115],"up":[116],"66.2%":[118],"KV":[119],"cache":[120],"reduction":[122],"length-invariant,":[124],"near-constant":[125],"per-step":[126],"latency.":[127]},"counts_by_year":[],"updated_date":"2026-07-01T06:00:48.157686","created_date":"2026-04-14T00:00:00"}