{"id":"https://openalex.org/W7154449882","doi":"https://doi.org/10.48550/arxiv.2604.12767","title":"CLASP: Class-Adaptive Layer Fusion and Dual-Stage Pruning for Multimodal Large Language Models","display_name":"CLASP: Class-Adaptive Layer Fusion and Dual-Stage Pruning for Multimodal Large Language Models","publication_year":2026,"publication_date":"2026-04-14","ids":{"openalex":"https://openalex.org/W7154449882","doi":"https://doi.org/10.48550/arxiv.2604.12767"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2604.12767","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2604.12767","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.2604.12767","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5111214495","display_name":"Yunkai Dang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Dang, Yunkai","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5006110323","display_name":"Yizhu Jiang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Jiang, Yizhu","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5133699000","display_name":"Yifan Jiang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Jiang, Yifan","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5133715522","display_name":"Qi Fan","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Fan, Qi","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5133682433","display_name":"Yinghuan Shi","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Shi, Yinghuan","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5133635399","display_name":"Wenbin Li","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Li, Wenbin","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5133635591","display_name":"Yang Gao","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Gao, Yang","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":0.9545000195503235,"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":0.9545000195503235,"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":0.012400000356137753,"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":0.005200000014156103,"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":0.78329998254776},{"id":"https://openalex.org/keywords/redundancy","display_name":"Redundancy (engineering)","score":0.5755000114440918},{"id":"https://openalex.org/keywords/pruning","display_name":"Pruning","score":0.5526999831199646},{"id":"https://openalex.org/keywords/feature","display_name":"Feature (linguistics)","score":0.5515999794006348},{"id":"https://openalex.org/keywords/language-model","display_name":"Language model","score":0.4659999907016754},{"id":"https://openalex.org/keywords/relevance","display_name":"Relevance (law)","score":0.40619999170303345},{"id":"https://openalex.org/keywords/layer","display_name":"Layer (electronics)","score":0.38960000872612},{"id":"https://openalex.org/keywords/code","display_name":"Code (set theory)","score":0.37860000133514404}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8091999888420105},{"id":"https://openalex.org/C48145219","wikidata":"https://www.wikidata.org/wiki/Q1335365","display_name":"Security token","level":2,"score":0.78329998254776},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5771999955177307},{"id":"https://openalex.org/C152124472","wikidata":"https://www.wikidata.org/wiki/Q1204361","display_name":"Redundancy (engineering)","level":2,"score":0.5755000114440918},{"id":"https://openalex.org/C108010975","wikidata":"https://www.wikidata.org/wiki/Q500094","display_name":"Pruning","level":2,"score":0.5526999831199646},{"id":"https://openalex.org/C2776401178","wikidata":"https://www.wikidata.org/wiki/Q12050496","display_name":"Feature (linguistics)","level":2,"score":0.5515999794006348},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.4659999907016754},{"id":"https://openalex.org/C158154518","wikidata":"https://www.wikidata.org/wiki/Q7310970","display_name":"Relevance (law)","level":2,"score":0.40619999170303345},{"id":"https://openalex.org/C2779227376","wikidata":"https://www.wikidata.org/wiki/Q6505497","display_name":"Layer (electronics)","level":2,"score":0.38960000872612},{"id":"https://openalex.org/C2776760102","wikidata":"https://www.wikidata.org/wiki/Q5139990","display_name":"Code (set theory)","level":3,"score":0.37860000133514404},{"id":"https://openalex.org/C2779960059","wikidata":"https://www.wikidata.org/wiki/Q7113681","display_name":"Overhead (engineering)","level":2,"score":0.3684000074863434},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.3549000024795532},{"id":"https://openalex.org/C158525013","wikidata":"https://www.wikidata.org/wiki/Q2593739","display_name":"Fusion","level":2,"score":0.32600000500679016},{"id":"https://openalex.org/C153180895","wikidata":"https://www.wikidata.org/wiki/Q7148389","display_name":"Pattern recognition (psychology)","level":2,"score":0.3172000050544739},{"id":"https://openalex.org/C111335779","wikidata":"https://www.wikidata.org/wiki/Q3454686","display_name":"Reduction (mathematics)","level":2,"score":0.302700012922287},{"id":"https://openalex.org/C52622490","wikidata":"https://www.wikidata.org/wiki/Q1026626","display_name":"Feature extraction","level":2,"score":0.2978000044822693},{"id":"https://openalex.org/C142575187","wikidata":"https://www.wikidata.org/wiki/Q3358290","display_name":"Pyramid (geometry)","level":2,"score":0.2973000109195709},{"id":"https://openalex.org/C36464697","wikidata":"https://www.wikidata.org/wiki/Q451553","display_name":"Visualization","level":2,"score":0.2930000126361847},{"id":"https://openalex.org/C31972630","wikidata":"https://www.wikidata.org/wiki/Q844240","display_name":"Computer vision","level":1,"score":0.2768999934196472},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.2689000070095062},{"id":"https://openalex.org/C2776359362","wikidata":"https://www.wikidata.org/wiki/Q2145286","display_name":"Representation (politics)","level":3,"score":0.2687999904155731},{"id":"https://openalex.org/C204323151","wikidata":"https://www.wikidata.org/wiki/Q905424","display_name":"Range (aeronautics)","level":2,"score":0.26339998841285706},{"id":"https://openalex.org/C5339829","wikidata":"https://www.wikidata.org/wiki/Q1425977","display_name":"Machine vision","level":2,"score":0.2554999887943268}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2604.12767","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2604.12767","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.2604.12767","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2604.12767","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":{"Multimodal":[0],"Large":[1],"Language":[2],"Models":[3],"(MLLMs)":[4],"suffer":[5],"from":[6],"substantial":[7],"computational":[8],"overhead":[9],"due":[10],"to":[11],"the":[12,83],"high":[13],"redundancy":[14],"in":[15],"visual":[16,70,113],"token":[17,54,84,114],"sequences.":[18],"Existing":[19],"approaches":[20],"typically":[21],"address":[22],"this":[23],"issue":[24],"using":[25],"single-layer":[26],"Vision":[27],"Transformer":[28],"(ViT)":[29],"features":[30],"and":[31,62,92,106,133],"static":[32],"pruning":[33,131],"strategies.":[34],"However,":[35],"such":[36],"fixed":[37],"configurations":[38],"are":[39],"often":[40],"brittle":[41],"under":[42],"diverse":[43],"instructions.":[44],"To":[45],"overcome":[46],"these":[47],"limitations,":[48],"we":[49],"propose":[50],"CLASP,":[51],"a":[52,126],"plug-and-play":[53],"reduction":[55],"framework":[56],"based":[57],"on":[58],"class-adaptive":[59,99],"layer":[60],"fusion":[61,105],"dual-stage":[63,80],"pruning.":[64],"Specifically,":[65],"CLASP":[66,101,120],"first":[67],"constructs":[68],"category-specific":[69],"representations":[71],"through":[72],"multi-layer":[73],"vision":[74],"feature":[75,104],"fusion.":[76],"It":[77],"then":[78],"performs":[79],"pruning,":[81,100],"allocating":[82],"budget":[85,107],"between":[86],"attention-salient":[87],"pivot":[88],"tokens":[89,95],"for":[90,96],"relevance":[91],"redundancy-aware":[93],"completion":[94],"coverage.":[97],"Through":[98],"enables":[102],"prompt-conditioned":[103],"allocation,":[108],"allowing":[109],"aggressive":[110],"yet":[111],"robust":[112],"reduction.":[115],"Extensive":[116],"experiments":[117],"demonstrate":[118],"that":[119],"consistently":[121],"outperforms":[122],"existing":[123],"methods":[124],"across":[125],"wide":[127],"range":[128],"of":[129],"benchmarks,":[130],"ratios,":[132],"MLLM":[134],"architectures.":[135],"Code":[136],"will":[137],"be":[138],"available":[139],"at":[140],"https://github.com/Yunkaidang/CLASP.":[141]},"counts_by_year":[],"updated_date":"2026-07-01T06:00:48.157686","created_date":"2026-04-16T00:00:00"}