{"id":"https://openalex.org/W7135237850","doi":"https://doi.org/10.48550/arxiv.2603.12267","title":"EVATok: Adaptive Length Video Tokenization for Efficient Visual Autoregressive Generation","display_name":"EVATok: Adaptive Length Video Tokenization for Efficient Visual Autoregressive Generation","publication_year":2026,"publication_date":"2026-03-12","ids":{"openalex":"https://openalex.org/W7135237850","doi":"https://doi.org/10.48550/arxiv.2603.12267"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2603.12267","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2603.12267","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.2603.12267","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5128922843","display_name":"Tianwei Xiong","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Xiong, Tianwei","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5128980214","display_name":"Jun Hao Liew","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Liew, Jun Hao","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5128994491","display_name":"Zilong Huang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Huang, Zilong","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5129097643","display_name":"Zhijie Lin","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Lin, Zhijie","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5087024287","display_name":"Jiashi Feng","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Feng, Jiashi","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5128993587","display_name":"Xihui Liu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Liu, Xihui","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/T10775","display_name":"Generative Adversarial Networks and Image Synthesis","score":0.9120000004768372,"subfield":{"id":"https://openalex.org/subfields/1707","display_name":"Computer Vision and Pattern Recognition"},"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/T10775","display_name":"Generative Adversarial Networks and Image Synthesis","score":0.9120000004768372,"subfield":{"id":"https://openalex.org/subfields/1707","display_name":"Computer Vision and Pattern Recognition"},"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/T10036","display_name":"Advanced Neural Network Applications","score":0.013899999670684338,"subfield":{"id":"https://openalex.org/subfields/1707","display_name":"Computer Vision and Pattern Recognition"},"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/T10741","display_name":"Video Coding and Compression Technologies","score":0.005499999970197678,"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/security-token","display_name":"Security token","score":0.7901999950408936},{"id":"https://openalex.org/keywords/autoregressive-model","display_name":"Autoregressive model","score":0.5411999821662903},{"id":"https://openalex.org/keywords/encode","display_name":"ENCODE","score":0.44699999690055847},{"id":"https://openalex.org/keywords/video-quality","display_name":"Video quality","score":0.3910999894142151},{"id":"https://openalex.org/keywords/downstream","display_name":"Downstream (manufacturing)","score":0.3831000030040741},{"id":"https://openalex.org/keywords/video-tracking","display_name":"Video tracking","score":0.3506999909877777}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8353000283241272},{"id":"https://openalex.org/C48145219","wikidata":"https://www.wikidata.org/wiki/Q1335365","display_name":"Security token","level":2,"score":0.7901999950408936},{"id":"https://openalex.org/C159877910","wikidata":"https://www.wikidata.org/wiki/Q2202883","display_name":"Autoregressive model","level":2,"score":0.5411999821662903},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5389999747276306},{"id":"https://openalex.org/C79403827","wikidata":"https://www.wikidata.org/wiki/Q3988","display_name":"Real-time computing","level":1,"score":0.482699990272522},{"id":"https://openalex.org/C66746571","wikidata":"https://www.wikidata.org/wiki/Q1134833","display_name":"ENCODE","level":3,"score":0.44699999690055847},{"id":"https://openalex.org/C31972630","wikidata":"https://www.wikidata.org/wiki/Q844240","display_name":"Computer vision","level":1,"score":0.42910000681877136},{"id":"https://openalex.org/C103910844","wikidata":"https://www.wikidata.org/wiki/Q2631256","display_name":"Video quality","level":3,"score":0.3910999894142151},{"id":"https://openalex.org/C2776207758","wikidata":"https://www.wikidata.org/wiki/Q5303302","display_name":"Downstream (manufacturing)","level":2,"score":0.3831000030040741},{"id":"https://openalex.org/C28490314","wikidata":"https://www.wikidata.org/wiki/Q189436","display_name":"Speech recognition","level":1,"score":0.366100013256073},{"id":"https://openalex.org/C202474056","wikidata":"https://www.wikidata.org/wiki/Q1931635","display_name":"Video tracking","level":3,"score":0.3506999909877777},{"id":"https://openalex.org/C2779530757","wikidata":"https://www.wikidata.org/wiki/Q1207505","display_name":"Quality (philosophy)","level":2,"score":0.3264000117778778},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.32190001010894775},{"id":"https://openalex.org/C125411270","wikidata":"https://www.wikidata.org/wiki/Q18653","display_name":"Encoding (memory)","level":2,"score":0.31139999628067017},{"id":"https://openalex.org/C160633673","wikidata":"https://www.wikidata.org/wiki/Q355198","display_name":"Pixel","level":2,"score":0.29339998960494995},{"id":"https://openalex.org/C2776187449","wikidata":"https://www.wikidata.org/wiki/Q1513879","display_name":"Natural language generation","level":3,"score":0.2888999879360199},{"id":"https://openalex.org/C161765866","wikidata":"https://www.wikidata.org/wiki/Q184748","display_name":"Codec","level":2,"score":0.2736999988555908},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.2614000141620636}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2603.12267","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2603.12267","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.2603.12267","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2603.12267","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":[],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Autoregressive":[0],"(AR)":[1],"video":[2,7,34,85,129,143],"generative":[3],"models":[4],"rely":[5],"on":[6,50,110,154],"tokenizers":[8,35,105],"that":[9,106,118,141],"compress":[10],"pixels":[11],"into":[12],"discrete":[13],"token":[14,20,39,81,163],"sequences.":[15],"The":[16],"length":[17],"of":[18,44,98],"these":[19,99],"sequences":[21],"is":[22],"crucial":[23],"for":[24,83,95,128],"balancing":[25],"reconstruction":[26,130,149],"quality":[27,127],"against":[28],"downstream":[29,132],"generation":[30,153],"computational":[31],"cost.":[32],"Traditional":[33],"apply":[36],"a":[37,69],"uniform":[38],"assignment":[40],"across":[41],"temporal":[42],"blocks":[43],"different":[45],"videos,":[46],"often":[47],"wasting":[48],"tokens":[49],"simple,":[51],"static,":[52],"or":[53,59],"repetitive":[54],"segments":[55],"while":[56],"underserving":[57],"dynamic":[58],"complex":[60],"ones.":[61],"To":[62],"address":[63],"this":[64],"inefficiency,":[65],"we":[66],"introduce":[67],"$\\textbf{EVATok}$,":[68],"framework":[70,78],"to":[71,86,166],"produce":[72],"$\\textbf{E}$fficient":[73],"$\\textbf{V}$ideo":[74],"$\\textbf{A}$daptive":[75],"$\\textbf{Tok}$enizers.":[76],"Our":[77],"estimates":[79],"optimal":[80,100],"assignments":[82,112],"each":[84],"achieve":[87],"the":[88,111,167],"best":[89],"quality-cost":[90],"trade-off,":[91],"develops":[92],"lightweight":[93],"routers":[94],"fast":[96],"prediction":[97],"assignments,":[101],"and":[102,125,131,150,171],"trains":[103],"adaptive":[104],"encode":[107],"videos":[108],"based":[109],"predicted":[113],"by":[114,136],"routers.":[115],"We":[116],"demonstrate":[117],"EVATok":[119,146],"delivers":[120],"substantial":[121],"improvements":[122],"in":[123,161],"efficiency":[124],"overall":[126],"AR":[133],"generation.":[134],"Enhanced":[135],"our":[137,172],"advanced":[138],"training":[139],"recipe":[140],"integrates":[142],"semantic":[144],"encoders,":[145],"achieves":[147],"superior":[148],"state-of-the-art":[151,169],"class-to-video":[152],"UCF-101,":[155],"with":[156],"at":[157],"least":[158],"24.4%":[159],"savings":[160],"average":[162],"usage":[164],"compared":[165],"prior":[168],"LARP":[170],"fixed-length":[173],"baseline.":[174]},"counts_by_year":[],"updated_date":"2026-07-01T06:00:48.157686","created_date":"2026-03-14T00:00:00"}