{"id":"https://openalex.org/W7140100370","doi":"https://doi.org/10.48550/arxiv.2603.19571","title":"CurveStream: Boosting Streaming Video Understanding in MLLMs via Curvature-Aware Hierarchical Visual Memory Management","display_name":"CurveStream: Boosting Streaming Video Understanding in MLLMs via Curvature-Aware Hierarchical Visual Memory Management","publication_year":2026,"publication_date":"2026-03-20","ids":{"openalex":"https://openalex.org/W7140100370","doi":"https://doi.org/10.48550/arxiv.2603.19571"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2603.19571","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2603.19571","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,"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.2603.19571","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5130402804","display_name":"Chao Wang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wang, Chao","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5130391531","display_name":"Xudong Tan","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Tan, Xudong","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5130373285","display_name":"Jianjian Cao","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Cao, Jianjian","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5130377229","display_name":"Kangcong Li","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Li, Kangcong","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5130329064","display_name":"Tao Chen","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Chen, Tao","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/T11714","display_name":"Multimodal Machine Learning Applications","score":0.8968999981880188,"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":0.8968999981880188,"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/T11307","display_name":"Domain Adaptation and Few-Shot Learning","score":0.01860000006854534,"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/T10775","display_name":"Generative Adversarial Networks and Image Synthesis","score":0.015200000256299973,"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/boosting","display_name":"Boosting (machine learning)","score":0.48910000920295715},{"id":"https://openalex.org/keywords/overlay","display_name":"Overlay","score":0.4763000011444092},{"id":"https://openalex.org/keywords/visualization","display_name":"Visualization","score":0.38510000705718994},{"id":"https://openalex.org/keywords/cache","display_name":"Cache","score":0.37299999594688416},{"id":"https://openalex.org/keywords/semantics","display_name":"Semantics (computer science)","score":0.3637000024318695},{"id":"https://openalex.org/keywords/key","display_name":"Key (lock)","score":0.35010001063346863},{"id":"https://openalex.org/keywords/security-token","display_name":"Security token","score":0.3456000089645386},{"id":"https://openalex.org/keywords/virtual-memory","display_name":"Virtual memory","score":0.34150001406669617},{"id":"https://openalex.org/keywords/semantic-feature","display_name":"Semantic feature","score":0.33970001339912415}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8434000015258789},{"id":"https://openalex.org/C46686674","wikidata":"https://www.wikidata.org/wiki/Q466303","display_name":"Boosting (machine learning)","level":2,"score":0.48910000920295715},{"id":"https://openalex.org/C136085584","wikidata":"https://www.wikidata.org/wiki/Q910289","display_name":"Overlay","level":2,"score":0.4763000011444092},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.4302000105381012},{"id":"https://openalex.org/C36464697","wikidata":"https://www.wikidata.org/wiki/Q451553","display_name":"Visualization","level":2,"score":0.38510000705718994},{"id":"https://openalex.org/C115537543","wikidata":"https://www.wikidata.org/wiki/Q165596","display_name":"Cache","level":2,"score":0.37299999594688416},{"id":"https://openalex.org/C184337299","wikidata":"https://www.wikidata.org/wiki/Q1437428","display_name":"Semantics (computer science)","level":2,"score":0.3637000024318695},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.35010001063346863},{"id":"https://openalex.org/C48145219","wikidata":"https://www.wikidata.org/wiki/Q1335365","display_name":"Security token","level":2,"score":0.3456000089645386},{"id":"https://openalex.org/C76399640","wikidata":"https://www.wikidata.org/wiki/Q189401","display_name":"Virtual memory","level":4,"score":0.34150001406669617},{"id":"https://openalex.org/C2781122975","wikidata":"https://www.wikidata.org/wiki/Q16928266","display_name":"Semantic feature","level":2,"score":0.33970001339912415},{"id":"https://openalex.org/C152124472","wikidata":"https://www.wikidata.org/wiki/Q1204361","display_name":"Redundancy (engineering)","level":2,"score":0.33149999380111694},{"id":"https://openalex.org/C160086991","wikidata":"https://www.wikidata.org/wiki/Q5939193","display_name":"Human visual system model","level":3,"score":0.29919999837875366},{"id":"https://openalex.org/C178278151","wikidata":"https://www.wikidata.org/wiki/Q7936607","display_name":"Visual memory","level":3,"score":0.29510000348091125},{"id":"https://openalex.org/C197914299","wikidata":"https://www.wikidata.org/wiki/Q18650","display_name":"Semantic memory","level":3,"score":0.2948000133037567},{"id":"https://openalex.org/C105842133","wikidata":"https://www.wikidata.org/wiki/Q1899679","display_name":"Visual communication","level":2,"score":0.2937999963760376},{"id":"https://openalex.org/C2780878386","wikidata":"https://www.wikidata.org/wiki/Q1659648","display_name":"Visual language","level":2,"score":0.29179999232292175},{"id":"https://openalex.org/C19012869","wikidata":"https://www.wikidata.org/wiki/Q578372","display_name":"Response time","level":2,"score":0.28999999165534973},{"id":"https://openalex.org/C125411270","wikidata":"https://www.wikidata.org/wiki/Q18653","display_name":"Encoding (memory)","level":2,"score":0.2750999927520752},{"id":"https://openalex.org/C2776401178","wikidata":"https://www.wikidata.org/wiki/Q12050496","display_name":"Feature (linguistics)","level":2,"score":0.27090001106262207},{"id":"https://openalex.org/C188441871","wikidata":"https://www.wikidata.org/wiki/Q7554146","display_name":"Softmax function","level":3,"score":0.26820001006126404},{"id":"https://openalex.org/C2780226545","wikidata":"https://www.wikidata.org/wiki/Q6888030","display_name":"Modality (human\u2013computer interaction)","level":2,"score":0.2639000117778778},{"id":"https://openalex.org/C2781181686","wikidata":"https://www.wikidata.org/wiki/Q4226068","display_name":"Coherence (philosophical gambling strategy)","level":2,"score":0.2623000144958496}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2603.19571","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2603.19571","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,"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.2603.19571","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2603.19571","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,"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":[{"display_name":"Peace, Justice and strong institutions","id":"https://metadata.un.org/sdg/16","score":0.44786861538887024}],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Multimodal":[0],"Large":[1],"Language":[2],"Models":[3],"have":[4],"achieved":[5],"significant":[6],"success":[7],"in":[8],"offline":[9],"video":[10,185],"understanding,":[11],"yet":[12,72],"their":[13],"application":[14],"to":[15,31,134],"streaming":[16,184],"videos":[17],"is":[18,93],"severely":[19],"limited":[20],"by":[21,95],"the":[22,96],"linear":[23],"explosion":[24],"of":[25,165],"visual":[26,39,87],"tokens,":[27],"which":[28],"often":[29,60],"leads":[30],"Out-of-Memory":[32],"(OOM)":[33],"errors":[34],"or":[35,53],"catastrophic":[36],"forgetting.":[37],"Existing":[38],"retention":[40],"and":[41,69,127,140,172],"memory":[42,88,142],"management":[43,89],"methods":[44],"typically":[45],"rely":[46],"on":[47,114,170,174],"uniform":[48],"sampling,":[49],"low-level":[50],"physical":[51],"metrics,":[52],"passive":[54],"cache":[55],"eviction.":[56],"However,":[57],"these":[58,78],"strategies":[59],"lack":[61],"intrinsic":[62],"semantic":[63,74,111,121],"awareness,":[64],"potentially":[65],"disrupting":[66],"contextual":[67],"coherence":[68],"blurring":[70],"transient":[71],"critical":[73,109],"transitions.":[75,112],"To":[76],"address":[77],"limitations,":[79],"we":[80],"propose":[81],"CurveStream,":[82,159],"a":[83,124,145],"training-free,":[84],"curvature-aware":[85],"hierarchical":[86],"framework.":[90],"Our":[91],"approach":[92],"motivated":[94],"key":[97],"observation":[98],"that":[99,155],"high-curvature":[100],"regions":[101],"along":[102],"continuous":[103],"feature":[104],"trajectories":[105],"closely":[106],"align":[107],"with":[108],"global":[110],"Based":[113],"this":[115,156],"geometric":[116],"insight,":[117],"CurveStream":[118],"evaluates":[119],"real-time":[120],"intensity":[122],"via":[123],"Curvature":[125],"Score":[126],"integrates":[128],"an":[129],"online":[130],"K-Sigma":[131],"dynamic":[132],"threshold":[133],"adaptively":[135],"route":[136],"frames":[137],"into":[138],"clear":[139],"fuzzy":[141],"states":[143],"under":[144],"strict":[146],"token":[147],"budget.":[148],"Evaluations":[149],"across":[150],"diverse":[151],"temporal":[152],"scales":[153],"confirm":[154],"lightweight":[157],"framework,":[158],"consistently":[160],"yields":[161],"absolute":[162],"performance":[163],"gains":[164],"over":[166,176],"10%":[167],"(e.g.,":[168],"10.69%":[169],"StreamingBench":[171],"13.58%":[173],"OVOBench)":[175],"respective":[177],"baselines,":[178],"establishing":[179],"new":[180],"state-of-the-art":[181],"results":[182],"for":[183],"perception.The":[186],"code":[187],"will":[188],"be":[189],"released":[190],"at":[191],"https://github.com/streamingvideos/CurveStream.":[192]},"counts_by_year":[],"updated_date":"2026-07-28T07:46:37.118299","created_date":"2026-03-24T00:00:00"}