{"id":"https://openalex.org/W7163162248","doi":"https://doi.org/10.48550/arxiv.2606.01790","title":"STaR-KV: Spatio-Temporal Adaptive Re-weighting for KV Cache Compression in GUI Vision-Language Models","display_name":"STaR-KV: Spatio-Temporal Adaptive Re-weighting for KV Cache Compression in GUI Vision-Language Models","publication_year":2026,"publication_date":"2026-06-01","ids":{"openalex":"https://openalex.org/W7163162248","doi":"https://doi.org/10.48550/arxiv.2606.01790"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2606.01790","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.01790","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.2606.01790","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5137653055","display_name":"Yuhang Han","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Han, Yuhang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5056290718","display_name":"Won-Hyuk Yang","orcid":"https://orcid.org/0009-0003-1918-8445"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yang, Wenzheng","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5137642200","display_name":"Yujie Chen","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Chen, Yujie","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5137637253","display_name":"Xiangqi Jin","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Jin, Xiangqi","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5137685955","display_name":"Yaojie Zhang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhang, Yaojie","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5137650956","display_name":"Siteng Huang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Huang, Siteng","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5137627245","display_name":"Linfeng Zhang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhang, Linfeng","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/T10054","display_name":"Parallel Computing and Optimization Techniques","score":0.24199999868869781,"subfield":{"id":"https://openalex.org/subfields/1708","display_name":"Hardware and Architecture"},"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/T10054","display_name":"Parallel Computing and Optimization Techniques","score":0.24199999868869781,"subfield":{"id":"https://openalex.org/subfields/1708","display_name":"Hardware and Architecture"},"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":0.12939999997615814,"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":0.11140000075101852,"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/cache","display_name":"Cache","score":0.73580002784729},{"id":"https://openalex.org/keywords/cpu-cache","display_name":"CPU cache","score":0.5605000257492065},{"id":"https://openalex.org/keywords/overhead","display_name":"Overhead (engineering)","score":0.5306000113487244},{"id":"https://openalex.org/keywords/compression","display_name":"Compression (physics)","score":0.5091999769210815},{"id":"https://openalex.org/keywords/data-compression","display_name":"Data compression","score":0.45739999413490295},{"id":"https://openalex.org/keywords/compression-ratio","display_name":"Compression ratio","score":0.3946000039577484},{"id":"https://openalex.org/keywords/stability","display_name":"Stability (learning theory)","score":0.37119999527931213}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7685999870300293},{"id":"https://openalex.org/C115537543","wikidata":"https://www.wikidata.org/wiki/Q165596","display_name":"Cache","level":2,"score":0.73580002784729},{"id":"https://openalex.org/C189783530","wikidata":"https://www.wikidata.org/wiki/Q352090","display_name":"CPU cache","level":3,"score":0.5605000257492065},{"id":"https://openalex.org/C2779960059","wikidata":"https://www.wikidata.org/wiki/Q7113681","display_name":"Overhead (engineering)","level":2,"score":0.5306000113487244},{"id":"https://openalex.org/C180016635","wikidata":"https://www.wikidata.org/wiki/Q2712821","display_name":"Compression (physics)","level":2,"score":0.5091999769210815},{"id":"https://openalex.org/C78548338","wikidata":"https://www.wikidata.org/wiki/Q2493","display_name":"Data compression","level":2,"score":0.45739999413490295},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.44339999556541443},{"id":"https://openalex.org/C9390403","wikidata":"https://www.wikidata.org/wiki/Q3966","display_name":"Computer hardware","level":1,"score":0.4293999969959259},{"id":"https://openalex.org/C25797200","wikidata":"https://www.wikidata.org/wiki/Q828137","display_name":"Compression ratio","level":3,"score":0.3946000039577484},{"id":"https://openalex.org/C112972136","wikidata":"https://www.wikidata.org/wiki/Q7595718","display_name":"Stability (learning theory)","level":2,"score":0.37119999527931213},{"id":"https://openalex.org/C113775141","wikidata":"https://www.wikidata.org/wiki/Q428691","display_name":"Computer engineering","level":1,"score":0.3578000068664551},{"id":"https://openalex.org/C38556500","wikidata":"https://www.wikidata.org/wiki/Q13404475","display_name":"Cache algorithms","level":4,"score":0.34130001068115234},{"id":"https://openalex.org/C2776760102","wikidata":"https://www.wikidata.org/wiki/Q5139990","display_name":"Code (set theory)","level":3,"score":0.3310999870300293},{"id":"https://openalex.org/C136085584","wikidata":"https://www.wikidata.org/wiki/Q910289","display_name":"Overlay","level":2,"score":0.3165999948978424},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.3057999908924103},{"id":"https://openalex.org/C149635348","wikidata":"https://www.wikidata.org/wiki/Q193040","display_name":"Embedded system","level":1,"score":0.30320000648498535},{"id":"https://openalex.org/C13481523","wikidata":"https://www.wikidata.org/wiki/Q412438","display_name":"Image compression","level":4,"score":0.3001999855041504},{"id":"https://openalex.org/C113843644","wikidata":"https://www.wikidata.org/wiki/Q901882","display_name":"Interface (matter)","level":4,"score":0.29409998655319214},{"id":"https://openalex.org/C79403827","wikidata":"https://www.wikidata.org/wiki/Q3988","display_name":"Real-time computing","level":1,"score":0.2922999858856201},{"id":"https://openalex.org/C37789001","wikidata":"https://www.wikidata.org/wiki/Q782543","display_name":"Graphical user interface","level":2,"score":0.28380000591278076},{"id":"https://openalex.org/C3826847","wikidata":"https://www.wikidata.org/wiki/Q188768","display_name":"FLOPS","level":2,"score":0.2660999894142151}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2606.01790","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.01790","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.2606.01790","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.01790","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":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Vision-language-model-based":[0],"graphical":[1],"user":[2],"interface":[3],"(GUI)":[4],"agents":[5],"have":[6],"shown":[7],"broad":[8],"automation":[9],"capabilities,":[10],"yet":[11],"deployment":[12],"is":[13,194],"bottlenecked":[14],"by":[15,123,185],"a":[16,59,66,98,106,129,189],"key-value":[17],"(KV)":[18],"cache":[19,109,136],"that":[20,112,133,147],"grows":[21],"linearly":[22],"with":[23,174],"interaction":[24],"steps.":[25],"For":[26],"instance,":[27],"UI-TARS-1.5-7B":[28],"consumes":[29],"76":[30],"GB":[31,45],"of":[32,42],"GPU":[33,183],"memory":[34,184],"on":[35],"merely":[36],"five":[37],"screenshots,":[38],"approaching":[39],"the":[40,71,83,91,150,159],"capacity":[41],"mainstream":[43],"80":[44],"accelerators.":[46],"Existing":[47],"KV":[48,108,165],"compression":[49,110,166],"methods":[50,167],"share":[51],"two":[52],"structural":[53],"assumptions:":[54],"aggregating":[55],"visual-token":[56],"importance":[57,115],"into":[58],"single":[60],"shared":[61],"saliency":[62],"map,":[63],"and":[64,86,142,180],"applying":[65],"fixed":[67],"top-B":[68],"cutoff":[69],"to":[70],"fused":[72],"score":[73,92,151],"distribution.":[74,152],"Pilot":[75],"measurements":[76],"refute":[77],"both:":[78],"spatial":[79,125],"specialization":[80],"lives":[81],"at":[82,171,188,196],"attention-subspace":[84],"level":[85],"migrates":[87],"across":[88],"layers,":[89],"while":[90],"distribution":[93],"drifts":[94],"in":[95],"shape":[96],"along":[97,116],"trajectory.":[99],"We":[100],"propose":[101],"STaR-KV":[102,157],"(Spatio-Temporal":[103],"Adaptive":[104],"Re-weighting),":[105],"training-free":[107],"framework":[111],"calibrates":[113],"token":[114],"three":[117],"axes:":[118],"(i)":[119],"subspace-aware":[120],"scoring":[121],"driven":[122],"online":[124],"mutual":[126],"information;":[127],"(ii)":[128],"temporal":[130],"stability":[131],"discount":[132],"suppresses":[134],"redundant":[135],"entries":[137],"from":[138],"persistently":[139],"attended":[140],"subspaces;":[141],"(iii)":[143],"an":[144],"entropy-derived":[145],"temperature":[146],"adaptively":[148],"reshapes":[149],"Across":[153],"four":[154],"GUI":[155],"benchmarks,":[156],"achieves":[158],"strongest":[160],"average":[161],"accuracy":[162],"among":[163],"state-of-the-art":[164],"(e.g.,":[168],"GUIKV,":[169],"SnapKV)":[170],"matched":[172],"budgets,":[173],"no":[175],"compression-stage":[176],"FLOPs":[177],"overhead":[178],"(-0.07%)":[179],"cutting":[181],"peak":[182],"nearly":[186],"40%":[187],"20%":[190],"KV-cache":[191],"budget.":[192],"Code":[193],"available":[195],"https://github.com/kawhiiiileo/STaR-KV.":[197]},"counts_by_year":[],"updated_date":"2026-07-01T06:00:48.157686","created_date":"2026-06-03T00:00:00"}
