{"id":"https://openalex.org/W7163277576","doi":"https://doi.org/10.48550/arxiv.2606.01756","title":"EvoCut: Multi-Layer Evolution-Aware Visual Token Compression for Efficient Large Vision-Language Models","display_name":"EvoCut: Multi-Layer Evolution-Aware Visual Token Compression for Efficient Large Vision-Language Models","publication_year":2026,"publication_date":"2026-06-01","ids":{"openalex":"https://openalex.org/W7163277576","doi":"https://doi.org/10.48550/arxiv.2606.01756"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2606.01756","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.01756","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.2606.01756","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5137710812","display_name":"Hongyu Lu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Lu, Hongyu","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5137636510","display_name":"Feng Zhang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhang, Feng","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5026838971","display_name":"Wenwei Jin","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Jin, Wenwei","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5136426989","display_name":"Huanling Hu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Hu, Huanling","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5137620167","display_name":"Pengfei Zhang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhang, Pengfei","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5137629708","display_name":"Yao Hu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Hu, Yao","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5137654460","display_name":"Jiawei Li","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Li, Jiawei","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5136387284","display_name":"Shikai Jiang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Jiang, Shikai","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":0.7426000237464905,"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":0.7426000237464905,"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":0.06840000301599503,"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":0.0430000014603138,"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/security-token","display_name":"Security token","score":0.8640999794006348},{"id":"https://openalex.org/keywords/inference","display_name":"Inference","score":0.47350001335144043},{"id":"https://openalex.org/keywords/representation","display_name":"Representation (politics)","score":0.44449999928474426},{"id":"https://openalex.org/keywords/human-visual-system-model","display_name":"Human visual system model","score":0.37400001287460327},{"id":"https://openalex.org/keywords/data-compression","display_name":"Data compression","score":0.35929998755455017},{"id":"https://openalex.org/keywords/limit","display_name":"Limit (mathematics)","score":0.34529998898506165},{"id":"https://openalex.org/keywords/visualization","display_name":"Visualization","score":0.3276999890804291}],"concepts":[{"id":"https://openalex.org/C48145219","wikidata":"https://www.wikidata.org/wiki/Q1335365","display_name":"Security token","level":2,"score":0.8640999794006348},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7591999769210815},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5152999758720398},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.47350001335144043},{"id":"https://openalex.org/C2776359362","wikidata":"https://www.wikidata.org/wiki/Q2145286","display_name":"Representation (politics)","level":3,"score":0.44449999928474426},{"id":"https://openalex.org/C31972630","wikidata":"https://www.wikidata.org/wiki/Q844240","display_name":"Computer vision","level":1,"score":0.4099999964237213},{"id":"https://openalex.org/C160086991","wikidata":"https://www.wikidata.org/wiki/Q5939193","display_name":"Human visual system model","level":3,"score":0.37400001287460327},{"id":"https://openalex.org/C78548338","wikidata":"https://www.wikidata.org/wiki/Q2493","display_name":"Data compression","level":2,"score":0.35929998755455017},{"id":"https://openalex.org/C151201525","wikidata":"https://www.wikidata.org/wiki/Q177239","display_name":"Limit (mathematics)","level":2,"score":0.34529998898506165},{"id":"https://openalex.org/C36464697","wikidata":"https://www.wikidata.org/wiki/Q451553","display_name":"Visualization","level":2,"score":0.3276999890804291},{"id":"https://openalex.org/C2779668609","wikidata":"https://www.wikidata.org/wiki/Q623092","display_name":"Rapid serial visual presentation","level":3,"score":0.3172999918460846},{"id":"https://openalex.org/C178253425","wikidata":"https://www.wikidata.org/wiki/Q162668","display_name":"Visual perception","level":3,"score":0.3154999911785126},{"id":"https://openalex.org/C180016635","wikidata":"https://www.wikidata.org/wiki/Q2712821","display_name":"Compression (physics)","level":2,"score":0.3095000088214874},{"id":"https://openalex.org/C126042441","wikidata":"https://www.wikidata.org/wiki/Q1324888","display_name":"Frame (networking)","level":2,"score":0.2655999958515167},{"id":"https://openalex.org/C153180895","wikidata":"https://www.wikidata.org/wiki/Q7148389","display_name":"Pattern recognition (psychology)","level":2,"score":0.2603999972343445},{"id":"https://openalex.org/C80444323","wikidata":"https://www.wikidata.org/wiki/Q2878974","display_name":"Theoretical computer science","level":1,"score":0.25529998540878296},{"id":"https://openalex.org/C2781311116","wikidata":"https://www.wikidata.org/wiki/Q83306","display_name":"Group (periodic table)","level":2,"score":0.2540999948978424},{"id":"https://openalex.org/C13481523","wikidata":"https://www.wikidata.org/wiki/Q412438","display_name":"Image compression","level":4,"score":0.25360000133514404}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2606.01756","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.01756","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.2606.01756","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.01756","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":{"Large":[0],"vision-language":[1],"models":[2],"(LVLMs)":[3],"achieve":[4],"strong":[5],"performance":[6,67],"on":[7,112,147],"image":[8],"and":[9,65,82,120,162],"video":[10],"understanding":[11],"tasks,":[12],"but":[13],"their":[14],"inference":[15],"efficiency":[16,161],"is":[17],"constrained":[18],"by":[19,27],"the":[20,54,144,153],"large":[21],"number":[22],"of":[23,143,152],"visual":[24,32,50,78,122,145],"tokens":[25,51,85,100,146],"produced":[26],"vision":[28,55],"encoders.":[29],"Most":[30],"existing":[31],"token":[33,37,79,123,128],"compression":[34,124],"methods":[35],"estimate":[36],"importance":[38,63,129],"from":[39,106,130],"attention":[40],"scores":[41],"or":[42],"representation":[43],"properties":[44],"at":[45],"specific":[46],"layers,":[47],"overlooking":[48],"how":[49],"evolve":[52],"across":[53,91],"encoder.":[56],"Such":[57],"layer-specific":[58],"criteria":[59],"may":[60],"provide":[61],"incomplete":[62],"estimates":[64,127],"limit":[66],"preservation":[68],"after":[69],"compression.":[70],"To":[71],"address":[72],"this":[73,113],"issue,":[74],"we":[75,115],"analyze":[76],"layer-wise":[77],"evolution":[80,89,109,132],"directions":[81,90],"observe":[83],"that":[84,98,126,137],"form":[86],"multiple":[87],"group":[88,108],"vision-encoder":[92],"layers.":[93],"Our":[94],"analysis":[95],"further":[96],"shows":[97],"informative":[99],"tend":[101],"to":[102],"exhibit":[103],"persistent":[104],"deviations":[105],"common":[107],"directions.":[110],"Based":[111],"observation,":[114],"propose":[116],"EvoCut,":[117],"a":[118],"training-free":[119],"attention-free":[121],"method":[125],"multi-layer":[131],"deviation.":[133],"Experimental":[134],"results":[135],"show":[136],"EvoCut":[138],"can":[139],"retain":[140],"only":[141],"11.1\\%":[142],"LLaVA-1.5-7B":[148],"while":[149],"preserving":[150],"94.4\\%":[151],"average":[154],"performance,":[155],"demonstrating":[156],"its":[157],"effectiveness":[158],"in":[159],"balancing":[160],"accuracy.":[163]},"counts_by_year":[],"updated_date":"2026-07-01T06:00:48.157686","created_date":"2026-06-03T00:00:00"}