{"id":"https://openalex.org/W7129899482","doi":"https://doi.org/10.48550/arxiv.2602.14027","title":"Train Short, Inference Long: Training-free Horizon Extension for Autoregressive Video Generation","display_name":"Train Short, Inference Long: Training-free Horizon Extension for Autoregressive Video Generation","publication_year":2026,"publication_date":"2026-02-15","ids":{"openalex":"https://openalex.org/W7129899482","doi":"https://doi.org/10.48550/arxiv.2602.14027"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2602.14027","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2602.14027","pdf_url":"https://arxiv.org/pdf/2602.14027","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"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":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"volume":null,"type":"preprint","is_traditional":true,"indexed_in":["arxiv","datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/2602.14027","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5126201267","display_name":"Jia Li","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I4405256866","display_name":"ByteDance (China)","ror":"https://ror.org/04m8av638","country_code":"CN","type":"company","lineage":["https://openalex.org/I4405256866"]},{"id":"https://openalex.org/I889458895","display_name":"University of Hong Kong","ror":"https://ror.org/02zhqgq86","country_code":"HK","type":"education","lineage":["https://openalex.org/I889458895"]}],"countries":["CN","HK"],"is_corresponding":false,"raw_author_name":"Li, Jia","raw_affiliation_strings":["ByteDance","The University of Hong Kong"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"ByteDance","institution_ids":["https://openalex.org/I4405256866"]},{"raw_affiliation_string":"The University of Hong Kong","institution_ids":["https://openalex.org/I889458895"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5111595815","display_name":"Xiaoqin Fu","orcid":"https://orcid.org/0000-0001-7195-0765","observed_orcids":["https://orcid.org/0000-0001-7195-0765","https://orcid.org/0009-0005-2818-2373"]},"institutions":[{"id":"https://openalex.org/I19820366","display_name":"Chinese Academy of Sciences","ror":"https://ror.org/034t30j35","country_code":"CN","type":"government","lineage":["https://openalex.org/I19820366"]},{"id":"https://openalex.org/I4210156404","display_name":"Institute of Information Engineering","ror":"https://ror.org/04r53se39","country_code":"CN","type":"facility","lineage":["https://openalex.org/I19820366","https://openalex.org/I4210156404"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Fu, Xiaomeng","raw_affiliation_strings":["Institute of Information Engineering , Chinese Academy of Sciences"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Institute of Information Engineering , Chinese Academy of Sciences","institution_ids":["https://openalex.org/I19820366","https://openalex.org/I4210156404"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5126249289","display_name":"Xurui Peng","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I4405256866","display_name":"ByteDance (China)","ror":"https://ror.org/04m8av638","country_code":"CN","type":"company","lineage":["https://openalex.org/I4405256866"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Peng, Xurui","raw_affiliation_strings":["ByteDance"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"ByteDance","institution_ids":["https://openalex.org/I4405256866"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5123205050","display_name":"Weifeng Chen","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I4405256866","display_name":"ByteDance (China)","ror":"https://ror.org/04m8av638","country_code":"CN","type":"company","lineage":["https://openalex.org/I4405256866"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Chen, Weifeng","raw_affiliation_strings":["ByteDance"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"ByteDance","institution_ids":["https://openalex.org/I4405256866"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5123171784","display_name":"Youwei Zheng","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I4405256866","display_name":"ByteDance (China)","ror":"https://ror.org/04m8av638","country_code":"CN","type":"company","lineage":["https://openalex.org/I4405256866"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Zheng, Youwei","raw_affiliation_strings":["ByteDance"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"ByteDance","institution_ids":["https://openalex.org/I4405256866"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5031880799","display_name":"Tianyu Zhao","orcid":"https://orcid.org/0000-0003-3121-6042","observed_orcids":["https://orcid.org/0000-0003-3121-6042"]},"institutions":[{"id":"https://openalex.org/I4405256866","display_name":"ByteDance (China)","ror":"https://ror.org/04m8av638","country_code":"CN","type":"company","lineage":["https://openalex.org/I4405256866"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Zhao, Tianyu","raw_affiliation_strings":["ByteDance"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"ByteDance","institution_ids":["https://openalex.org/I4405256866"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5126257830","display_name":"Jiexi Wang","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I4405256866","display_name":"ByteDance (China)","ror":"https://ror.org/04m8av638","country_code":"CN","type":"company","lineage":["https://openalex.org/I4405256866"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Wang, Jiexi","raw_affiliation_strings":["ByteDance"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"ByteDance","institution_ids":["https://openalex.org/I4405256866"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5122054829","display_name":"Fangmin Chen","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I4405256866","display_name":"ByteDance (China)","ror":"https://ror.org/04m8av638","country_code":"CN","type":"company","lineage":["https://openalex.org/I4405256866"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Chen, Fangmin","raw_affiliation_strings":["ByteDance"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"ByteDance","institution_ids":["https://openalex.org/I4405256866"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5159946465","display_name":"Xing Wang","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I4405256866","display_name":"ByteDance (China)","ror":"https://ror.org/04m8av638","country_code":"CN","type":"company","lineage":["https://openalex.org/I4405256866"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Wang, Xing","raw_affiliation_strings":["ByteDance"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"ByteDance","institution_ids":["https://openalex.org/I4405256866"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5123484012","display_name":"Hayden Kwok-Hay So","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I889458895","display_name":"University of Hong Kong","ror":"https://ror.org/02zhqgq86","country_code":"HK","type":"education","lineage":["https://openalex.org/I889458895"]}],"countries":["HK"],"is_corresponding":false,"raw_author_name":"So, Hayden Kwok-Hay","raw_affiliation_strings":["The University of Hong Kong"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"The University of Hong Kong","institution_ids":["https://openalex.org/I889458895"]}]}],"institutions":[],"countries_distinct_count":2,"institutions_distinct_count":4,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":true,"cited_by_count":16,"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":1.0,"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":1.0,"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/T11714","display_name":"Multimodal Machine Learning Applications","score":0.0,"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/T11105","display_name":"Advanced Image Processing Techniques","score":0.0,"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"}}],"keywords":[{"id":"https://openalex.org/keywords/autoregressive-video-generation","display_name":"autoregressive video generation","score":0.8169999718666077},{"id":"https://openalex.org/keywords/video-diffusion-models","display_name":"video diffusion models","score":0.7960000038146973},{"id":"https://openalex.org/keywords/long-video-generation","display_name":"long video generation","score":0.6449999809265137},{"id":"https://openalex.org/keywords/training-free-inference","display_name":"training-free inference","score":0.6700000166893005},{"id":"https://openalex.org/keywords/spectral-bias","display_name":"spectral bias","score":0.902999997138977},{"id":"https://openalex.org/keywords/attention-sinks","display_name":"attention sinks","score":0.7070000171661377},{"id":"https://openalex.org/keywords/temporal-degradation","display_name":"temporal degradation","score":0.44699999690055847}],"concepts":[{"id":"https://openalex.org/C177769412","wikidata":"https://www.wikidata.org/wiki/Q278090","display_name":"Prior probability","level":3,"score":0.7276999950408936},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6973999738693237},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.683899998664856},{"id":"https://openalex.org/C132459708","wikidata":"https://www.wikidata.org/wiki/Q744069","display_name":"Extrapolation","level":2,"score":0.6807000041007996},{"id":"https://openalex.org/C159877910","wikidata":"https://www.wikidata.org/wiki/Q2202883","display_name":"Autoregressive model","level":2,"score":0.5755000114440918},{"id":"https://openalex.org/C48044578","wikidata":"https://www.wikidata.org/wiki/Q727490","display_name":"Scalability","level":2,"score":0.5274999737739563},{"id":"https://openalex.org/C99498987","wikidata":"https://www.wikidata.org/wiki/Q2210247","display_name":"Noise (video)","level":3,"score":0.4681999981403351},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.3788999915122986},{"id":"https://openalex.org/C140779682","wikidata":"https://www.wikidata.org/wiki/Q210868","display_name":"Sampling (signal processing)","level":3,"score":0.3668999969959259},{"id":"https://openalex.org/C2778755073","wikidata":"https://www.wikidata.org/wiki/Q10858537","display_name":"Scale (ratio)","level":2,"score":0.36649999022483826},{"id":"https://openalex.org/C153083717","wikidata":"https://www.wikidata.org/wiki/Q6535263","display_name":"Leverage (statistics)","level":2,"score":0.34040001034736633},{"id":"https://openalex.org/C2776459999","wikidata":"https://www.wikidata.org/wiki/Q2119376","display_name":"Fidelity","level":2,"score":0.3330000042915344},{"id":"https://openalex.org/C160234255","wikidata":"https://www.wikidata.org/wiki/Q812535","display_name":"Bayesian inference","level":3,"score":0.3172999918460846},{"id":"https://openalex.org/C113364801","wikidata":"https://www.wikidata.org/wiki/Q26674","display_name":"High fidelity","level":2,"score":0.31130000948905945},{"id":"https://openalex.org/C29265498","wikidata":"https://www.wikidata.org/wiki/Q7047719","display_name":"Noise measurement","level":3,"score":0.310699999332428},{"id":"https://openalex.org/C162269090","wikidata":"https://www.wikidata.org/wiki/Q1156047","display_name":"Rope","level":2,"score":0.30959999561309814},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.3073999881744385},{"id":"https://openalex.org/C89600930","wikidata":"https://www.wikidata.org/wiki/Q1423946","display_name":"Segmentation","level":2,"score":0.3059000074863434},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.3018999993801117},{"id":"https://openalex.org/C79403827","wikidata":"https://www.wikidata.org/wiki/Q3988","display_name":"Real-time computing","level":1,"score":0.2930999994277954},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.2870999872684479},{"id":"https://openalex.org/C52740198","wikidata":"https://www.wikidata.org/wiki/Q1539564","display_name":"Importance sampling","level":3,"score":0.2854999899864197},{"id":"https://openalex.org/C31972630","wikidata":"https://www.wikidata.org/wiki/Q844240","display_name":"Computer vision","level":1,"score":0.2824000120162964},{"id":"https://openalex.org/C2776252893","wikidata":"https://www.wikidata.org/wiki/Q1364836","display_name":"FLEX","level":2,"score":0.2822999954223633},{"id":"https://openalex.org/C2778598663","wikidata":"https://www.wikidata.org/wiki/Q1407599","display_name":"Video content analysis","level":4,"score":0.26429998874664307}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2602.14027","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2602.14027","pdf_url":"https://arxiv.org/pdf/2602.14027","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"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":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},{"id":"doi:10.48550/arxiv.2602.14027","doi":"https://doi.org/10.48550/arxiv.2602.14027","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2602.14027","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"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"}],"best_oa_location":{"id":"pmh:oai:arXiv.org:2602.14027","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2602.14027","pdf_url":"https://arxiv.org/pdf/2602.14027","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"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":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"sustainable_development_goals":[],"study_designs":[],"awards":[],"funders":[],"has_content":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7129899482.pdf","grobid_xml":"https://content.openalex.org/works/W7129899482.grobid-xml"},"referenced_works_count":24,"referenced_works":["https://openalex.org/W4226057201","https://openalex.org/W4303441850","https://openalex.org/W4322718191","https://openalex.org/W4382603041","https://openalex.org/W4383994075","https://openalex.org/W4386435778","https://openalex.org/W4387294617","https://openalex.org/W4387929853","https://openalex.org/W4392090070","https://openalex.org/W4403580037","https://openalex.org/W4407359557","https://openalex.org/W4409092337","https://openalex.org/W4414939355","https://openalex.org/W4415163123","https://openalex.org/W4415337140","https://openalex.org/W4417136857","https://openalex.org/W4417163885","https://openalex.org/W4417287217","https://openalex.org/W4417514206","https://openalex.org/W6891877074","https://openalex.org/W6910899322","https://openalex.org/W7083704664","https://openalex.org/W7083825679","https://openalex.org/W7125377354"],"related_works":["https://openalex.org/W4414836149","https://openalex.org/W7201831071","https://openalex.org/W4414928226","https://openalex.org/W6910899322","https://openalex.org/W6929393341","https://openalex.org/W7161936546","https://openalex.org/W4399911525","https://openalex.org/W7167613792","https://openalex.org/W7116162452","https://openalex.org/W4415800771"],"abstract_inverted_index":{"Autoregressive":[0],"video":[1,12,183],"diffusion":[2],"models":[3,135,175],"have":[4],"emerged":[5],"as":[6,177],"a":[7,70,155,186],"scalable":[8],"paradigm":[9],"for":[10,165],"long":[11],"generation.":[13],"However,":[14],"they":[15],"often":[16],"suffer":[17],"from":[18,43],"severe":[19],"extrapolation":[20,138],"failure,":[21],"where":[22],"rapid":[23],"error":[24],"accumulation":[25],"leads":[26],"to":[27,89,99,112,121],"significant":[28],"temporal":[29,102],"degradation":[30],"when":[31],"extending":[32],"beyond":[33],"training":[34,80],"horizons.":[35],"We":[36],"identify":[37],"that":[38,74,130],"this":[39],"failure":[40],"primarily":[41],"stems":[42],"the":[44,52,76,143,171],"spectral":[45],"bias":[46],"of":[47,54,145,174],"3D":[48],"positional":[49],"embeddings":[50],"and":[51,81,117,141,181],"lack":[53],"dynamic":[55,115,182],"priors":[56,116],"in":[57],"noise":[58],"sampling.":[59],"To":[60],"address":[61],"these":[62],"issues,":[63],"we":[64],"propose":[65],"FLEX":[66,84,131,158],"(Frequency-aware":[67],"Length":[68],"EXtension),":[69],"training-free":[71],"inference-time":[72],"framework":[73],"bridges":[75],"gap":[77],"between":[78],"short-term":[79],"long-term":[82],"inference.":[83],"introduces":[85],"Frequency-aware":[86],"RoPE":[87],"Modulation":[88],"adaptively":[90],"interpolate":[91],"under-trained":[92],"low-frequency":[93],"components":[94],"while":[95],"extrapolating":[96],"high-frequency":[97,114],"ones":[98],"preserve":[100],"multi-scale":[101],"discriminability.":[103],"This":[104],"is":[105,191],"integrated":[106],"with":[107],"Antiphase":[108],"Noise":[109],"Sampling":[110],"(ANS)":[111],"inject":[113],"Inference-only":[118],"Attention":[119],"Sink":[120],"anchor":[122],"global":[123],"structure.":[124],"Extensive":[125],"evaluations":[126],"on":[127],"VBench":[128],"demonstrate":[129],"significantly":[132],"outperforms":[133],"state-of-the-art":[134],"at":[136,149,185,193],"6x":[137],"(30s":[139],"duration)":[140],"matches":[142],"performance":[144],"long-video":[146],"fine-tuned":[147],"baselines":[148],"12x":[150],"scale":[151],"(60s":[152],"duration).":[153],"As":[154],"plug-and-play":[156],"augmentation,":[157],"seamlessly":[159],"integrates":[160],"into":[161],"existing":[162],"inference":[163],"pipelines":[164],"horizon":[166],"extension.":[167],"It":[168],"effectively":[169],"pushes":[170],"generation":[172],"limits":[173],"such":[176],"LongLive,":[178],"supporting":[179],"consistent":[180],"synthesis":[184],"4-minute":[187],"scale.":[188],"Project":[189],"page":[190],"available":[192],"https://ga-lee.github.io/FLEX_demo.":[194]},"counts_by_year":[{"year":2026,"cited_by_count":15},{"year":2025,"cited_by_count":1}],"updated_date":"2026-11-07T11:25:07.637180","created_date":"2026-02-18T00:00:00"}