{"id":"https://openalex.org/W7128501944","doi":"https://doi.org/10.48550/arxiv.2602.08024","title":"FlashVID: Efficient Video Large Language Models via Training-free Tree-based Spatiotemporal Token Merging","display_name":"FlashVID: Efficient Video Large Language Models via Training-free Tree-based Spatiotemporal Token Merging","publication_year":2026,"publication_date":"2026-02-08","ids":{"openalex":"https://openalex.org/W7128501944","doi":"https://doi.org/10.48550/arxiv.2602.08024"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2602.08024","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2602.08024","pdf_url":"https://arxiv.org/pdf/2602.08024","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-nc-sa","license_id":"https://openalex.org/licenses/cc-by-nc-sa","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/2602.08024","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5125510940","display_name":"Ziyang Fan","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I204983213","display_name":"Harbin Institute of Technology","ror":"https://ror.org/01yqg2h08","country_code":"CN","type":"education","lineage":["https://openalex.org/I204983213"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Fan, Ziyang","raw_affiliation_strings":["Harbin Institute of Technology , Shenzhen"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Harbin Institute of Technology , Shenzhen","institution_ids":["https://openalex.org/I204983213"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5121785886","display_name":"Keyu Chen","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I204983213","display_name":"Harbin Institute of Technology","ror":"https://ror.org/01yqg2h08","country_code":"CN","type":"education","lineage":["https://openalex.org/I204983213"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Chen, Keyu","raw_affiliation_strings":["Harbin Institute of Technology , Shenzhen"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Harbin Institute of Technology , Shenzhen","institution_ids":["https://openalex.org/I204983213"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5116594387","display_name":"Ruilong Xing","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I204983213","display_name":"Harbin Institute of Technology","ror":"https://ror.org/01yqg2h08","country_code":"CN","type":"education","lineage":["https://openalex.org/I204983213"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Xing, Ruilong","raw_affiliation_strings":["Harbin Institute of Technology , Shenzhen"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Harbin Institute of Technology , Shenzhen","institution_ids":["https://openalex.org/I204983213"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5123149556","display_name":"Yulin Li","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I204983213","display_name":"Harbin Institute of Technology","ror":"https://ror.org/01yqg2h08","country_code":"CN","type":"education","lineage":["https://openalex.org/I204983213"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Li, Yulin","raw_affiliation_strings":["Harbin Institute of Technology , Shenzhen"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Harbin Institute of Technology , Shenzhen","institution_ids":["https://openalex.org/I204983213"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5125047830","display_name":"Li Jiang","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I4210116924","display_name":"Chinese University of Hong Kong, Shenzhen","ror":"https://ror.org/02d5ks197","country_code":"CN","type":"education","lineage":["https://openalex.org/I177725633","https://openalex.org/I180726961","https://openalex.org/I4210116924"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Jiang, Li","raw_affiliation_strings":["Shenzhen Loop Area Institute","The Chinese University of Hong Kong , Shenzhen"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Shenzhen Loop Area Institute","institution_ids":[]},{"raw_affiliation_string":"The Chinese University of Hong Kong , Shenzhen","institution_ids":["https://openalex.org/I4210116924"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5038086218","display_name":"Zhuotao Tian","orcid":"https://orcid.org/0000-0003-2698-6923","observed_orcids":["https://orcid.org/0000-0003-2698-6923"]},"institutions":[{"id":"https://openalex.org/I204983213","display_name":"Harbin Institute of Technology","ror":"https://ror.org/01yqg2h08","country_code":"CN","type":"education","lineage":["https://openalex.org/I204983213"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Tian, Zhuotao","raw_affiliation_strings":["Harbin Institute of Technology , Shenzhen","Shenzhen Loop Area Institute"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Harbin Institute of Technology , Shenzhen","institution_ids":["https://openalex.org/I204983213"]},{"raw_affiliation_string":"Shenzhen Loop Area Institute","institution_ids":[]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":2,"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/T11714","display_name":"Multimodal Machine Learning Applications","score":0.9613999724388123,"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/T11714","display_name":"Multimodal Machine Learning Applications","score":0.9613999724388123,"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":0.030400000512599945,"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/T10775","display_name":"Generative Adversarial Networks and Image Synthesis","score":0.0024999999441206455,"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/video-large-language-models","display_name":"video large language models","score":0.9980000257492065},{"id":"https://openalex.org/keywords/token-merging","display_name":"token merging","score":0.6759999990463257},{"id":"https://openalex.org/keywords/token-selection","display_name":"token selection","score":0.6449999809265137},{"id":"https://openalex.org/keywords/video-understanding","display_name":"video understanding","score":0.6710000038146973}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8258000016212463},{"id":"https://openalex.org/C48145219","wikidata":"https://www.wikidata.org/wiki/Q1335365","display_name":"Security token","level":2,"score":0.727400004863739},{"id":"https://openalex.org/C152124472","wikidata":"https://www.wikidata.org/wiki/Q1204361","display_name":"Redundancy (engineering)","level":2,"score":0.6690999865531921},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.5304999947547913},{"id":"https://openalex.org/C177148314","wikidata":"https://www.wikidata.org/wiki/Q170084","display_name":"Generalization","level":2,"score":0.4805000126361847},{"id":"https://openalex.org/C98045186","wikidata":"https://www.wikidata.org/wiki/Q205663","display_name":"Process (computing)","level":2,"score":0.4745999872684479},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.4745999872684479},{"id":"https://openalex.org/C81917197","wikidata":"https://www.wikidata.org/wiki/Q628760","display_name":"Selection (genetic algorithm)","level":2,"score":0.46650001406669617},{"id":"https://openalex.org/C126042441","wikidata":"https://www.wikidata.org/wiki/Q1324888","display_name":"Frame (networking)","level":2,"score":0.4140999913215637},{"id":"https://openalex.org/C31972630","wikidata":"https://www.wikidata.org/wiki/Q844240","display_name":"Computer vision","level":1,"score":0.38909998536109924},{"id":"https://openalex.org/C128840427","wikidata":"https://www.wikidata.org/wiki/Q1302174","display_name":"Motion compensation","level":2,"score":0.36660000681877136},{"id":"https://openalex.org/C79403827","wikidata":"https://www.wikidata.org/wiki/Q3988","display_name":"Real-time computing","level":1,"score":0.364300012588501},{"id":"https://openalex.org/C172849965","wikidata":"https://www.wikidata.org/wiki/Q3148875","display_name":"Reference frame","level":3,"score":0.358599990606308},{"id":"https://openalex.org/C104759252","wikidata":"https://www.wikidata.org/wiki/Q187120","display_name":"Time shifting","level":3,"score":0.34630000591278076},{"id":"https://openalex.org/C117896860","wikidata":"https://www.wikidata.org/wiki/Q11376","display_name":"Acceleration","level":2,"score":0.3400999903678894},{"id":"https://openalex.org/C2776760102","wikidata":"https://www.wikidata.org/wiki/Q5139990","display_name":"Code (set theory)","level":3,"score":0.29989999532699585},{"id":"https://openalex.org/C65483669","wikidata":"https://www.wikidata.org/wiki/Q3536669","display_name":"Video processing","level":2,"score":0.2912999987602234},{"id":"https://openalex.org/C202474056","wikidata":"https://www.wikidata.org/wiki/Q1931635","display_name":"Video tracking","level":3,"score":0.28439998626708984},{"id":"https://openalex.org/C45374587","wikidata":"https://www.wikidata.org/wiki/Q12525525","display_name":"Computation","level":2,"score":0.260699987411499},{"id":"https://openalex.org/C3261483","wikidata":"https://www.wikidata.org/wiki/Q119565","display_name":"Frame rate","level":2,"score":0.2606000006198883},{"id":"https://openalex.org/C179518139","wikidata":"https://www.wikidata.org/wiki/Q5140297","display_name":"Coding (social sciences)","level":2,"score":0.25589999556541443}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2602.08024","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2602.08024","pdf_url":"https://arxiv.org/pdf/2602.08024","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-nc-sa","license_id":"https://openalex.org/licenses/cc-by-nc-sa","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},{"id":"doi:10.48550/arxiv.2602.08024","doi":"https://doi.org/10.48550/arxiv.2602.08024","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2602.08024","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":null,"license_id":null,"version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"pmh:oai:arXiv.org:2602.08024","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2602.08024","pdf_url":"https://arxiv.org/pdf/2602.08024","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-nc-sa","license_id":"https://openalex.org/licenses/cc-by-nc-sa","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"sustainable_development_goals":[],"study_designs":[],"awards":[{"id":"https://openalex.org/G8463459816","display_name":null,"funder_award_id":"2025A1515011546","funder_id":"https://openalex.org/F4320337111","funder_display_name":"Basic and Applied Basic Research Foundation of Guangdong Province"}],"funders":[{"id":"https://openalex.org/F4320337111","display_name":"Basic and Applied Basic Research Foundation of Guangdong Province","ror":null}],"has_content":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7128501944.pdf","grobid_xml":"https://content.openalex.org/works/W7128501944.grobid-xml"},"referenced_works_count":19,"referenced_works":["https://openalex.org/W4384918448","https://openalex.org/W4390438326","https://openalex.org/W4394780994","https://openalex.org/W4400141325","https://openalex.org/W4401306886","https://openalex.org/W4401660192","https://openalex.org/W4403853618","https://openalex.org/W4404304985","https://openalex.org/W4405655184","https://openalex.org/W4405903187","https://openalex.org/W4407794264","https://openalex.org/W4414281281","https://openalex.org/W4414939355","https://openalex.org/W4416097713","https://openalex.org/W4416780398","https://openalex.org/W6948369276","https://openalex.org/W7075993604","https://openalex.org/W7077911270","https://openalex.org/W7160117124"],"related_works":["https://openalex.org/W7214988040","https://openalex.org/W4416288898","https://openalex.org/W7220665341","https://openalex.org/W4417283153","https://openalex.org/W7221575788","https://openalex.org/W4416037134","https://openalex.org/W7215114085","https://openalex.org/W4413158337","https://openalex.org/W4416036547","https://openalex.org/W4406073438"],"abstract_inverted_index":{"Although":[0],"Video":[1],"Large":[2],"Language":[3],"Models":[4],"(VLLMs)":[5],"have":[6],"shown":[7],"remarkable":[8],"capabilities":[9],"in":[10,58,176,183],"video":[11,106,129,169,177],"understanding,":[12],"they":[13],"are":[14,54],"required":[15],"to":[16,45,56,69,98,180],"process":[17],"high":[18],"volumes":[19],"of":[20,73,137,145,151,154,187],"visual":[21,52,146],"tokens,":[22,147],"causing":[23],"significant":[24],"computational":[25,192],"inefficiency.":[26],"Existing":[27],"VLLMs":[28,126],"acceleration":[29,85],"frameworks":[30],"usually":[31],"compress":[32],"spatial":[33,59],"and":[34,63,93,135,163],"temporal":[35],"redundancy":[36,118],"independently,":[37],"which":[38,171],"overlooks":[39],"the":[40,70,100,133,152,190],"spatiotemporal":[41,47,117],"relationships,":[42],"thereby":[43],"leading":[44],"suboptimal":[46],"compression.":[48],"The":[49],"highly":[50],"correlated":[51],"features":[53],"likely":[55],"change":[57],"position,":[60],"scale,":[61],"orientation,":[62],"other":[64],"attributes":[65],"over":[66],"time":[67],"due":[68],"dynamic":[71],"nature":[72],"video.":[74],"Building":[75],"on":[76,123],"this":[77],"insight,":[78],"we":[79],"introduce":[80],"FlashVID,":[81],"a":[82,161,173,184],"training-free":[83,162],"inference":[84],"framework":[86],"for":[87,104,115,166],"VLLMs.":[88],"Specifically,":[89],"FlashVID":[90,148,157],"utilizes":[91],"Attention":[92],"Diversity-based":[94],"Token":[95,112],"Selection":[96],"(ADTS)":[97],"select":[99],"most":[101],"representative":[102,125],"tokens":[103],"basic":[105],"representation,":[107],"then":[108],"applies":[109],"Tree-based":[110],"Spatiotemporal":[111],"Merging":[113],"(TSTM)":[114],"fine-grained":[116],"elimination.":[119],"Extensive":[120],"experiments":[121],"conducted":[122],"three":[124],"across":[127],"five":[128],"understanding":[130],"benchmarks":[131],"demonstrate":[132],"effectiveness":[134],"generalization":[136],"our":[138],"method.":[139],"Notably,":[140],"by":[141],"retaining":[142],"only":[143],"10%":[144],"preserves":[149],"99.1%":[150],"performance":[153],"LLaVA-OneVision.":[155],"Consequently,":[156],"can":[158],"serve":[159],"as":[160],"plug-and-play":[164],"module":[165],"extending":[167],"long":[168],"frames,":[170],"enables":[172],"10x":[174],"increase":[175],"frame":[178],"input":[179],"Qwen2.5-VL,":[181],"resulting":[182],"relative":[185],"improvement":[186],"8.6%":[188],"within":[189],"same":[191],"budget.":[193],"Code":[194],"is":[195],"available":[196],"at":[197],"https://github.com/Fanziyang-v/FlashVID.":[198]},"counts_by_year":[{"year":2026,"cited_by_count":16}],"updated_date":"2026-11-07T11:25:07.637180","created_date":"2026-02-11T00:00:00"}