{"id":"https://openalex.org/W7163077331","doi":"https://doi.org/10.48550/arxiv.2605.31457","title":"VisionPulse: Dynamic Visual Sparsity for Efficient Multimodal Reasoning","display_name":"VisionPulse: Dynamic Visual Sparsity for Efficient Multimodal Reasoning","publication_year":2026,"publication_date":"2026-05-29","ids":{"openalex":"https://openalex.org/W7163077331","doi":"https://doi.org/10.48550/arxiv.2605.31457"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2605.31457","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.31457","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.2605.31457","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5137554902","display_name":"Hengbo Xu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Xu, Hengbo","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5137527807","display_name":"Shengjie Jin","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Jin, Shengjie","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5020661081","display_name":"Yanbiao Ma","orcid":"https://orcid.org/0000-0002-8472-1475"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Ma, Yanbiao","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5137603105","display_name":"Zhiwu Lu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Lu, Zhiwu","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":0.9624000191688538,"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":0.9624000191688538,"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/T12026","display_name":"Explainable Artificial Intelligence (XAI)","score":0.005499999970197678,"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/T10036","display_name":"Advanced Neural Network Applications","score":0.004900000058114529,"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/bottleneck","display_name":"Bottleneck","score":0.6481000185012817},{"id":"https://openalex.org/keywords/set","display_name":"Set (abstract data type)","score":0.5781999826431274},{"id":"https://openalex.org/keywords/context","display_name":"Context (archaeology)","score":0.570900022983551},{"id":"https://openalex.org/keywords/security-token","display_name":"Security token","score":0.5604000091552734},{"id":"https://openalex.org/keywords/decoding-methods","display_name":"Decoding methods","score":0.5364999771118164},{"id":"https://openalex.org/keywords/visual-reasoning","display_name":"Visual reasoning","score":0.44119998812675476},{"id":"https://openalex.org/keywords/pruning","display_name":"Pruning","score":0.43790000677108765},{"id":"https://openalex.org/keywords/human-visual-system-model","display_name":"Human visual system model","score":0.4043999910354614}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7670999765396118},{"id":"https://openalex.org/C2780513914","wikidata":"https://www.wikidata.org/wiki/Q18210350","display_name":"Bottleneck","level":2,"score":0.6481000185012817},{"id":"https://openalex.org/C177264268","wikidata":"https://www.wikidata.org/wiki/Q1514741","display_name":"Set (abstract data type)","level":2,"score":0.5781999826431274},{"id":"https://openalex.org/C2779343474","wikidata":"https://www.wikidata.org/wiki/Q3109175","display_name":"Context (archaeology)","level":2,"score":0.570900022983551},{"id":"https://openalex.org/C48145219","wikidata":"https://www.wikidata.org/wiki/Q1335365","display_name":"Security token","level":2,"score":0.5604000091552734},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5490999817848206},{"id":"https://openalex.org/C57273362","wikidata":"https://www.wikidata.org/wiki/Q576722","display_name":"Decoding methods","level":2,"score":0.5364999771118164},{"id":"https://openalex.org/C2777508537","wikidata":"https://www.wikidata.org/wiki/Q7936620","display_name":"Visual reasoning","level":2,"score":0.44119998812675476},{"id":"https://openalex.org/C108010975","wikidata":"https://www.wikidata.org/wiki/Q500094","display_name":"Pruning","level":2,"score":0.43790000677108765},{"id":"https://openalex.org/C160086991","wikidata":"https://www.wikidata.org/wiki/Q5939193","display_name":"Human visual system model","level":3,"score":0.4043999910354614},{"id":"https://openalex.org/C99740376","wikidata":"https://www.wikidata.org/wiki/Q17092520","display_name":"Interactive visual analysis","level":4,"score":0.39559999108314514},{"id":"https://openalex.org/C178278151","wikidata":"https://www.wikidata.org/wiki/Q7936607","display_name":"Visual memory","level":3,"score":0.3885999917984009},{"id":"https://openalex.org/C36464697","wikidata":"https://www.wikidata.org/wiki/Q451553","display_name":"Visualization","level":2,"score":0.3659000098705292},{"id":"https://openalex.org/C2779960059","wikidata":"https://www.wikidata.org/wiki/Q7113681","display_name":"Overhead (engineering)","level":2,"score":0.36169999837875366},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.34619998931884766},{"id":"https://openalex.org/C59732488","wikidata":"https://www.wikidata.org/wiki/Q2528440","display_name":"Visual analytics","level":3,"score":0.3398999869823456},{"id":"https://openalex.org/C60008888","wikidata":"https://www.wikidata.org/wiki/Q6031013","display_name":"Information bottleneck method","level":3,"score":0.3352999985218048},{"id":"https://openalex.org/C2780103172","wikidata":"https://www.wikidata.org/wiki/Q1309721","display_name":"Visual Objects","level":3,"score":0.3262999951839447},{"id":"https://openalex.org/C125411270","wikidata":"https://www.wikidata.org/wiki/Q18653","display_name":"Encoding (memory)","level":2,"score":0.296999990940094},{"id":"https://openalex.org/C31972630","wikidata":"https://www.wikidata.org/wiki/Q844240","display_name":"Computer vision","level":1,"score":0.2935999929904938},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.29330000281333923},{"id":"https://openalex.org/C153180895","wikidata":"https://www.wikidata.org/wiki/Q7148389","display_name":"Pattern recognition (psychology)","level":2,"score":0.28110000491142273},{"id":"https://openalex.org/C178253425","wikidata":"https://www.wikidata.org/wiki/Q162668","display_name":"Visual perception","level":3,"score":0.272599995136261},{"id":"https://openalex.org/C183322885","wikidata":"https://www.wikidata.org/wiki/Q17007702","display_name":"Context model","level":3,"score":0.2637999951839447}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2605.31457","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.31457","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.2605.31457","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.31457","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":{"With":[0],"the":[1,28,60,78,84,111,130],"rapid":[2],"advancement":[3],"of":[4,50,165],"large":[5],"multimodal":[6],"models":[7],"(LMMs),":[8],"inference-time":[9],"overhead":[10],"has":[11],"become":[12],"a":[13,47,69,94,104],"key":[14],"bottleneck":[15,71],"for":[16],"real-world":[17],"deployment.":[18],"Existing":[19],"methods":[20],"typically":[21],"prune":[22],"visual":[23,30,41,51,74,96,106,124,139,146,151,166],"tokens":[24,52,133,167],"at":[25,55],"prefill,":[26],"assuming":[27],"required":[29],"evidence":[31,42],"remains":[32],"static":[33],"during":[34,100,141],"reasoning.":[35,65,101],"However,":[36],"we":[37,67,91],"empirically":[38],"show":[39,159],"that":[40,160],"is":[43,53],"strongly":[44],"step-dependent:":[45],"only":[46,129,162],"sparse":[48],"subset":[49],"critical":[54,61,132],"each":[56],"decoding":[57],"step,":[58],"and":[59,127],"set":[62],"evolves":[63],"across":[64],"Furthermore,":[66],"identify":[68],"coupled":[70],"where":[72],"redundant":[73,145],"context":[75,147],"can":[76],"steer":[77],"model":[79],"toward":[80],"query-irrelevant":[81],"regions,":[82],"lengthening":[83],"reasoning":[85,154,171],"trace.":[86],"Guided":[87],"by":[88,115,174],"these":[89],"insights,":[90],"propose":[92],"VisionPulse,":[93],"step-wise":[95,112],"token":[97,125],"pruning":[98],"framework":[99],"VisionPulse":[102,143,161],"computes":[103],"lightweight":[105],"attention":[107],"mass":[108],"to":[109],"estimate":[110],"retention":[113],"budget":[114],"exploiting":[116],"its":[117],"strong":[118],"positive":[119],"correlation":[120],"with":[121,170],"LMMs'":[122],"effective":[123],"usage":[126],"retain":[128],"most":[131],"under":[134],"this":[135],"budget.":[136],"By":[137],"enforcing":[138],"sparsity":[140],"reasoning,":[142],"filters":[144],"while":[148,176],"preserving":[149],"relevant":[150],"evidence,":[152],"shortening":[153],"traces":[155,172],"naturally.":[156],"Extensive":[157],"experiments":[158],"retains":[163],"5%":[164],"per":[168],"step":[169],"shortened":[173],"11.2%,":[175],"keeping":[177],"accuracy":[178],"almost":[179],"unchanged.":[180]},"counts_by_year":[],"updated_date":"2026-07-01T06:00:48.157686","created_date":"2026-06-02T00:00:00"}