{"id":"https://openalex.org/W7168337708","doi":"https://doi.org/10.48550/arxiv.2607.12111","title":"PFAdapter: Hierarchical LoRA Decomposition for Personalized Federated MLLMs","display_name":"PFAdapter: Hierarchical LoRA Decomposition for Personalized Federated MLLMs","publication_year":2026,"publication_date":"2026-07-13","ids":{"openalex":"https://openalex.org/W7168337708","doi":"https://doi.org/10.48550/arxiv.2607.12111"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2607.12111","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2607.12111","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.2607.12111","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5140751939","display_name":"Jing Liu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Liu, Jing","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5140871140","display_name":"Kun Yang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yang, Kun","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5140806770","display_name":"Yan Wang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wang, Yan","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5140904909","display_name":"Dingkang Yang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yang, Dingkang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5140752600","display_name":"Xiaoshuai Hao","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Hao, Xiaoshuai","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5140873402","display_name":"Wei Zhang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhang, Wei","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5140956292","display_name":"Yang Liu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Liu, Yang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5140809435","display_name":"Wei Zhou","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhou, Wei","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/T11273","display_name":"Advanced Graph Neural Networks","score":0.36079999804496765,"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"}},"topics":[{"id":"https://openalex.org/T11273","display_name":"Advanced Graph Neural Networks","score":0.36079999804496765,"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/T10764","display_name":"Privacy-Preserving Technologies in Data","score":0.05570000037550926,"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/T13702","display_name":"Machine Learning in Healthcare","score":0.04540000110864639,"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"}}],"keywords":[{"id":"https://openalex.org/keywords/federated-learning","display_name":"Federated learning","score":0.4717999994754791},{"id":"https://openalex.org/keywords/key","display_name":"Key (lock)","score":0.44929999113082886},{"id":"https://openalex.org/keywords/personalization","display_name":"Personalization","score":0.4189999997615814},{"id":"https://openalex.org/keywords/adaptation","display_name":"Adaptation (eye)","score":0.41830000281333923},{"id":"https://openalex.org/keywords/locality","display_name":"Locality","score":0.4099000096321106},{"id":"https://openalex.org/keywords/software-deployment","display_name":"Software deployment","score":0.351500004529953},{"id":"https://openalex.org/keywords/edge-device","display_name":"Edge device","score":0.32749998569488525},{"id":"https://openalex.org/keywords/data-aggregator","display_name":"Data aggregator","score":0.32739999890327454}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8256000280380249},{"id":"https://openalex.org/C120314980","wikidata":"https://www.wikidata.org/wiki/Q180634","display_name":"Distributed computing","level":1,"score":0.5669000148773193},{"id":"https://openalex.org/C2992525071","wikidata":"https://www.wikidata.org/wiki/Q50818671","display_name":"Federated learning","level":2,"score":0.4717999994754791},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.44929999113082886},{"id":"https://openalex.org/C183003079","wikidata":"https://www.wikidata.org/wiki/Q1000371","display_name":"Personalization","level":2,"score":0.4189999997615814},{"id":"https://openalex.org/C139807058","wikidata":"https://www.wikidata.org/wiki/Q352374","display_name":"Adaptation (eye)","level":2,"score":0.41830000281333923},{"id":"https://openalex.org/C2779808786","wikidata":"https://www.wikidata.org/wiki/Q6664603","display_name":"Locality","level":2,"score":0.4099000096321106},{"id":"https://openalex.org/C105339364","wikidata":"https://www.wikidata.org/wiki/Q2297740","display_name":"Software deployment","level":2,"score":0.351500004529953},{"id":"https://openalex.org/C138236772","wikidata":"https://www.wikidata.org/wiki/Q25098575","display_name":"Edge device","level":3,"score":0.32749998569488525},{"id":"https://openalex.org/C82578977","wikidata":"https://www.wikidata.org/wiki/Q16773055","display_name":"Data aggregator","level":3,"score":0.32739999890327454},{"id":"https://openalex.org/C184337299","wikidata":"https://www.wikidata.org/wiki/Q1437428","display_name":"Semantics (computer science)","level":2,"score":0.31610000133514404},{"id":"https://openalex.org/C2776401178","wikidata":"https://www.wikidata.org/wiki/Q12050496","display_name":"Feature (linguistics)","level":2,"score":0.30720001459121704},{"id":"https://openalex.org/C162307627","wikidata":"https://www.wikidata.org/wiki/Q204833","display_name":"Enhanced Data Rates for GSM Evolution","level":2,"score":0.3018999993801117},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.29840001463890076},{"id":"https://openalex.org/C2778180026","wikidata":"https://www.wikidata.org/wiki/Q18378163","display_name":"Semantic heterogeneity","level":4,"score":0.29829999804496765},{"id":"https://openalex.org/C194498986","wikidata":"https://www.wikidata.org/wiki/Q680016","display_name":"Roaming","level":2,"score":0.2953999936580658},{"id":"https://openalex.org/C2778456923","wikidata":"https://www.wikidata.org/wiki/Q5337692","display_name":"Edge computing","level":3,"score":0.29339998960494995},{"id":"https://openalex.org/C193415008","wikidata":"https://www.wikidata.org/wiki/Q639681","display_name":"Network architecture","level":2,"score":0.2897999882698059},{"id":"https://openalex.org/C2778562939","wikidata":"https://www.wikidata.org/wiki/Q1298791","display_name":"Synchronization (alternating current)","level":3,"score":0.28769999742507935},{"id":"https://openalex.org/C2780598303","wikidata":"https://www.wikidata.org/wiki/Q65921492","display_name":"Flexibility (engineering)","level":2,"score":0.2639999985694885},{"id":"https://openalex.org/C80444323","wikidata":"https://www.wikidata.org/wiki/Q2878974","display_name":"Theoretical computer science","level":1,"score":0.2549000084400177},{"id":"https://openalex.org/C152124472","wikidata":"https://www.wikidata.org/wiki/Q1204361","display_name":"Redundancy (engineering)","level":2,"score":0.25220000743865967},{"id":"https://openalex.org/C2779582901","wikidata":"https://www.wikidata.org/wiki/Q21013010","display_name":"Distributed learning","level":2,"score":0.2506999969482422}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2607.12111","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2607.12111","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.2607.12111","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2607.12111","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":[],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Agentic":[0],"AI":[1,212],"systems":[2],"are":[3,111],"reshaping":[4],"communications":[5],"and":[6,79,104,108,126,166,180],"networking":[7],"by":[8,170],"deploying":[9],"autonomous":[10],"intelligent":[11],"agents":[12],"capable":[13],"of":[14],"collaborative":[15],"learning":[16],"while":[17,124],"maintaining":[18],"data":[19],"privacy":[20],"at":[21],"network":[22,26,56],"edges.":[23],"Within":[24],"distributed":[25],"environments,":[27],"Multimodal":[28],"Large":[29],"Language":[30],"Models":[31],"(MLLMs)":[32],"serve":[33],"as":[34],"cognitive":[35],"engines":[36],"for":[37,116,131,210],"edge":[38,200],"devices,":[39],"yet":[40],"federated":[41,59,161],"fine-tuning":[42],"faces":[43],"substantial":[44],"challenges":[45],"in":[46,76,214],"balancing":[47],"global":[48,114],"knowledge":[49],"aggregation":[50,153],"with":[51,71],"local":[52,164],"adaptation":[53],"under":[54],"heterogeneous":[55],"conditions.":[57],"Conventional":[58],"protocols":[60,154],"typically":[61],"rely":[62],"on":[63,138,175],"uniform":[64],"parameter":[65],"aggregation,":[66],"which":[67],"conflates":[68],"domain-invariant":[69],"features":[70],"client-specific":[72],"nuances,":[73],"thereby":[74],"resulting":[75],"suboptimal":[77],"personalization":[78],"excessive":[80],"communication":[81,168,216],"overhead.":[82],"To":[83],"address":[84],"these":[85,146],"challenges,":[86],"we":[87],"propose":[88],"PFAdapter,":[89],"a":[90],"communication-efficient":[91],"framework":[92,205],"introducing":[93],"hierarchical":[94],"LoRA":[95],"decomposition":[96],"to":[97,113,196],"explicitly":[98],"separate":[99],"adapter":[100],"parameters":[101],"into":[102],"global-shared":[103,157],"local-private":[105],"components.":[106],"Query":[107],"key":[109],"projections":[110,128],"assigned":[112],"synchronization":[115],"capturing":[117],"universal":[118],"multimodal":[119],"semantics":[120],"across":[121,159,198],"the":[122,139,160],"network,":[123,162],"value":[125],"output":[127],"remain":[129],"localized":[130],"edge-specific":[132],"adaptation.":[133],"Additionally,":[134],"orthogonality":[135],"regularization":[136],"based":[137],"Frobenius":[140],"norm":[141],"enforces":[142],"strict":[143],"separation":[144],"between":[145],"components,":[147],"preventing":[148],"redundant":[149],"feature":[150],"learning.":[151],"Selective":[152],"synchronize":[155],"only":[156],"components":[158],"preserving":[163],"expertise":[165],"reducing":[167],"costs":[169],"nearly":[171],"50%.":[172],"Extensive":[173],"experiments":[174],"VQA-RAD,":[176],"SLAKE,":[177],"Hateful":[178],"Memes,":[179],"CrisisMMD":[181],"datasets":[182],"demonstrate":[183],"that":[184],"PFAdapter":[185],"consistently":[186],"outperforms":[187],"state-of-the-art":[188],"baselines,":[189],"achieving":[190],"accuracy":[191],"improvements":[192],"ranging":[193],"from":[194],"2.4%":[195],"4.8%":[197],"diverse":[199],"intelligence":[201],"tasks.":[202],"Consequently,":[203],"our":[204],"establishes":[206],"an":[207],"efficient":[208],"solution":[209],"agentic":[211],"deployment":[213],"resource-constrained":[215],"networks.":[217]},"counts_by_year":[],"updated_date":"2026-07-28T07:46:37.118299","created_date":"2026-07-16T00:00:00"}