{"id":"https://openalex.org/W7139089975","doi":"https://doi.org/10.48550/arxiv.2603.16001","title":"Mostly Text, Smart Visuals: Asymmetric Text-Visual Pruning for Large Vision-Language Models","display_name":"Mostly Text, Smart Visuals: Asymmetric Text-Visual Pruning for Large Vision-Language Models","publication_year":2026,"publication_date":"2026-03-16","ids":{"openalex":"https://openalex.org/W7139089975","doi":"https://doi.org/10.48550/arxiv.2603.16001"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2603.16001","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2603.16001","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.2603.16001","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5048483627","display_name":"Sijie Li","orcid":"https://orcid.org/0009-0003-5984-9121"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Li, Sijie","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5129954257","display_name":"Biao Qian","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Qian, Biao","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5129773941","display_name":"Jungong Han","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Han, Jungong","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":0.9100000262260437,"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":0.9100000262260437,"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":0.03009999915957451,"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":0.008100000210106373,"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/pruning","display_name":"Pruning","score":0.7980999946594238},{"id":"https://openalex.org/keywords/metric","display_name":"Metric (unit)","score":0.600600004196167},{"id":"https://openalex.org/keywords/sensitivity","display_name":"Sensitivity (control systems)","score":0.4796000123023987},{"id":"https://openalex.org/keywords/selection","display_name":"Selection (genetic algorithm)","score":0.43639999628067017},{"id":"https://openalex.org/keywords/process","display_name":"Process (computing)","score":0.4350000023841858},{"id":"https://openalex.org/keywords/focus","display_name":"Focus (optics)","score":0.4120999872684479},{"id":"https://openalex.org/keywords/visualization","display_name":"Visualization","score":0.38600000739097595},{"id":"https://openalex.org/keywords/abstraction","display_name":"Abstraction","score":0.36469998955726624}],"concepts":[{"id":"https://openalex.org/C108010975","wikidata":"https://www.wikidata.org/wiki/Q500094","display_name":"Pruning","level":2,"score":0.7980999946594238},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7551000118255615},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.6114000082015991},{"id":"https://openalex.org/C176217482","wikidata":"https://www.wikidata.org/wiki/Q860554","display_name":"Metric (unit)","level":2,"score":0.600600004196167},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.4814999997615814},{"id":"https://openalex.org/C21200559","wikidata":"https://www.wikidata.org/wiki/Q7451068","display_name":"Sensitivity (control systems)","level":2,"score":0.4796000123023987},{"id":"https://openalex.org/C81917197","wikidata":"https://www.wikidata.org/wiki/Q628760","display_name":"Selection (genetic algorithm)","level":2,"score":0.43639999628067017},{"id":"https://openalex.org/C98045186","wikidata":"https://www.wikidata.org/wiki/Q205663","display_name":"Process (computing)","level":2,"score":0.4350000023841858},{"id":"https://openalex.org/C192209626","wikidata":"https://www.wikidata.org/wiki/Q190909","display_name":"Focus (optics)","level":2,"score":0.4120999872684479},{"id":"https://openalex.org/C36464697","wikidata":"https://www.wikidata.org/wiki/Q451553","display_name":"Visualization","level":2,"score":0.38600000739097595},{"id":"https://openalex.org/C124304363","wikidata":"https://www.wikidata.org/wiki/Q673661","display_name":"Abstraction","level":2,"score":0.36469998955726624},{"id":"https://openalex.org/C165838908","wikidata":"https://www.wikidata.org/wiki/Q736777","display_name":"Calibration","level":2,"score":0.35679998993873596},{"id":"https://openalex.org/C205606062","wikidata":"https://www.wikidata.org/wiki/Q5249645","display_name":"Decoupling (probability)","level":2,"score":0.3508000075817108},{"id":"https://openalex.org/C2780586882","wikidata":"https://www.wikidata.org/wiki/Q7520643","display_name":"Simple (philosophy)","level":2,"score":0.3441999852657318},{"id":"https://openalex.org/C153180895","wikidata":"https://www.wikidata.org/wiki/Q7148389","display_name":"Pattern recognition (psychology)","level":2,"score":0.3433000147342682},{"id":"https://openalex.org/C48145219","wikidata":"https://www.wikidata.org/wiki/Q1335365","display_name":"Security token","level":2,"score":0.3382999897003174},{"id":"https://openalex.org/C2780226545","wikidata":"https://www.wikidata.org/wiki/Q6888030","display_name":"Modality (human\u2013computer interaction)","level":2,"score":0.3285999894142151},{"id":"https://openalex.org/C113174947","wikidata":"https://www.wikidata.org/wiki/Q2859736","display_name":"Tree (set theory)","level":2,"score":0.30059999227523804},{"id":"https://openalex.org/C2779903281","wikidata":"https://www.wikidata.org/wiki/Q6888026","display_name":"Modalities","level":2,"score":0.28110000491142273},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.2791999876499176},{"id":"https://openalex.org/C2776502983","wikidata":"https://www.wikidata.org/wiki/Q690182","display_name":"Contrast (vision)","level":2,"score":0.27810001373291016},{"id":"https://openalex.org/C2780898871","wikidata":"https://www.wikidata.org/wiki/Q860554","display_name":"Performance metric","level":2,"score":0.2621999979019165},{"id":"https://openalex.org/C44291984","wikidata":"https://www.wikidata.org/wiki/Q1074173","display_name":"Question answering","level":2,"score":0.25209999084472656}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2603.16001","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2603.16001","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.2603.16001","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2603.16001","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":[{"id":"https://metadata.un.org/sdg/4","display_name":"Quality Education","score":0.55147784948349}],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Network":[0],"pruning":[1,59,77,143],"is":[2,165],"an":[3],"effective":[4,125],"technique":[5],"for":[6,57,131,140],"enabling":[7],"lightweight":[8],"Large":[9],"Vision-Language":[10],"Models":[11],"(LVLMs),":[12],"which":[13,135],"primarily":[14],"incorporates":[15],"both":[16,150],"weights":[17],"and":[18,54,72,152,174],"activations":[19],"into":[20],"the":[21,49,68,76,87,102,106,137,146,200],"importance":[22,138],"metric.":[23],"However,":[24],"existing":[25],"efforts":[26],"typically":[27],"process":[28],"calibration":[29,163],"data":[30],"from":[31,149],"different":[32],"modalities":[33],"in":[34],"a":[35,43,122,162,175,183],"unified":[36],"manner,":[37],"overlooking":[38],"modality-specific":[39],"behaviors.":[40],"This":[41],"raises":[42],"critical":[44],"challenge:":[45],"how":[46],"to":[47,75,187],"address":[48],"divergent":[50],"behaviors":[51],"of":[52,60,70,177,202],"textual":[53,73,88,151,172],"visual":[55,71,103,107,153,178,190],"tokens":[56,74,148,173],"accurate":[58,141],"LVLMs.":[61],"To":[62],"this":[63],"end,":[64],"we":[65,120,181],"systematically":[66],"investigate":[67],"sensitivity":[69,100],"operation":[78],"by":[79,117,144,168],"decoupling":[80],"their":[81],"corresponding":[82],"weights,":[83],"revealing":[84],"that:":[85],"(i)":[86],"pathway":[89,108],"should":[90],"be":[91],"calibrated":[92],"via":[93],"text":[94],"tokens,":[95],"since":[96],"it":[97],"exhibits":[98,109],"higher":[99],"than":[101],"pathway;":[104],"(ii)":[105],"high":[110],"redundancy,":[111],"permitting":[112],"even":[113],"50%":[114],"sparsity.":[115],"Motivated":[116],"these":[118],"insights,":[119],"propose":[121],"simple":[123],"yet":[124],"Asymmetric":[126],"Text-Visual":[127],"Weight":[128],"Pruning":[129],"method":[130],"LVLMs,":[132],"dubbed":[133],"ATV-Pruning,":[134],"establishes":[136],"metric":[139],"weight":[142],"selecting":[145],"informative":[147],"pathways.":[154],"Specifically,":[155],"ATV-Pruning":[156,204],"integrates":[157],"two":[158],"primary":[159],"innovations:":[160],"first,":[161],"pool":[164],"adaptively":[166],"constructed":[167],"drawing":[169],"on":[170],"all":[171],"subset":[176],"tokens;":[179],"second,":[180],"devise":[182],"layer-adaptive":[184],"selection":[185],"strategy":[186],"yield":[188],"important":[189],"tokens.":[191],"Finally,":[192],"extensive":[193],"experiments":[194],"across":[195],"standard":[196],"multimodal":[197],"benchmarks":[198],"verify":[199],"superiority":[201],"our":[203],"over":[205],"state-of-the-art":[206],"methods.":[207]},"counts_by_year":[],"updated_date":"2026-07-01T06:00:48.157686","created_date":"2026-03-20T00:00:00"}