{"id":"https://openalex.org/W7131371254","doi":"https://doi.org/10.48550/arxiv.2602.19945","title":"DP-FedAdamW: An Efficient Optimizer for Differentially Private Federated Large Models","display_name":"DP-FedAdamW: An Efficient Optimizer for Differentially Private Federated Large Models","publication_year":2026,"publication_date":"2026-02-23","ids":{"openalex":"https://openalex.org/W7131371254","doi":"https://doi.org/10.48550/arxiv.2602.19945"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2602.19945","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2602.19945","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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"article"},"type":"preprint","indexed_in":["datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://doi.org/10.48550/arxiv.2602.19945","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5126791676","display_name":"Jin Liu","orcid":null},"institutions":[],"countries":[],"is_corresponding":true,"raw_author_name":"Liu, Jin","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5026667188","display_name":"Y. Miao","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Miao, Yinbin","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5023212284","display_name":"Ning Xi","orcid":"https://orcid.org/0000-0001-8276-5696"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Xi, Ning","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5048532598","display_name":"Junkang Liu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Liu, Junkang","raw_affiliation_strings":[],"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":4,"corresponding_author_ids":["https://openalex.org/A5126791676"],"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/T10764","display_name":"Privacy-Preserving Technologies in Data","score":0.9707000255584717,"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/T10764","display_name":"Privacy-Preserving Technologies in Data","score":0.9707000255584717,"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/T11307","display_name":"Domain Adaptation and Few-Shot Learning","score":0.00419999985024333,"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/T11612","display_name":"Stochastic Gradient Optimization Techniques","score":0.002899999963119626,"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/differential-privacy","display_name":"Differential privacy","score":0.8460999727249146},{"id":"https://openalex.org/keywords/robustness","display_name":"Robustness (evolution)","score":0.6909000277519226},{"id":"https://openalex.org/keywords/federated-learning","display_name":"Federated learning","score":0.6575999855995178},{"id":"https://openalex.org/keywords/estimator","display_name":"Estimator","score":0.6155999898910522},{"id":"https://openalex.org/keywords/convergence","display_name":"Convergence (economics)","score":0.4846000075340271},{"id":"https://openalex.org/keywords/variance","display_name":"Variance (accounting)","score":0.4715000092983246},{"id":"https://openalex.org/keywords/rate-of-convergence","display_name":"Rate of convergence","score":0.4643000066280365},{"id":"https://openalex.org/keywords/gradient-descent","display_name":"Gradient descent","score":0.4189999997615814},{"id":"https://openalex.org/keywords/information-privacy","display_name":"Information privacy","score":0.39579999446868896}],"concepts":[{"id":"https://openalex.org/C23130292","wikidata":"https://www.wikidata.org/wiki/Q5275358","display_name":"Differential privacy","level":2,"score":0.8460999727249146},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7383000254631042},{"id":"https://openalex.org/C63479239","wikidata":"https://www.wikidata.org/wiki/Q7353546","display_name":"Robustness (evolution)","level":3,"score":0.6909000277519226},{"id":"https://openalex.org/C2992525071","wikidata":"https://www.wikidata.org/wiki/Q50818671","display_name":"Federated learning","level":2,"score":0.6575999855995178},{"id":"https://openalex.org/C185429906","wikidata":"https://www.wikidata.org/wiki/Q1130160","display_name":"Estimator","level":2,"score":0.6155999898910522},{"id":"https://openalex.org/C2777303404","wikidata":"https://www.wikidata.org/wiki/Q759757","display_name":"Convergence (economics)","level":2,"score":0.4846000075340271},{"id":"https://openalex.org/C196083921","wikidata":"https://www.wikidata.org/wiki/Q7915758","display_name":"Variance (accounting)","level":2,"score":0.4715000092983246},{"id":"https://openalex.org/C57869625","wikidata":"https://www.wikidata.org/wiki/Q1783502","display_name":"Rate of convergence","level":3,"score":0.4643000066280365},{"id":"https://openalex.org/C153258448","wikidata":"https://www.wikidata.org/wiki/Q1199743","display_name":"Gradient descent","level":3,"score":0.4189999997615814},{"id":"https://openalex.org/C126255220","wikidata":"https://www.wikidata.org/wiki/Q141495","display_name":"Mathematical optimization","level":1,"score":0.4117000102996826},{"id":"https://openalex.org/C123201435","wikidata":"https://www.wikidata.org/wiki/Q456632","display_name":"Information privacy","level":2,"score":0.39579999446868896},{"id":"https://openalex.org/C147764199","wikidata":"https://www.wikidata.org/wiki/Q6865248","display_name":"Minification","level":2,"score":0.3765999972820282},{"id":"https://openalex.org/C137822555","wikidata":"https://www.wikidata.org/wiki/Q2587068","display_name":"Information sensitivity","level":2,"score":0.375900000333786},{"id":"https://openalex.org/C99498987","wikidata":"https://www.wikidata.org/wiki/Q2210247","display_name":"Noise (video)","level":3,"score":0.3587000072002411},{"id":"https://openalex.org/C2780909371","wikidata":"https://www.wikidata.org/wiki/Q4801092","display_name":"Artificial noise","level":4,"score":0.3562000095844269},{"id":"https://openalex.org/C99221444","wikidata":"https://www.wikidata.org/wiki/Q1532069","display_name":"Private information retrieval","level":2,"score":0.33390000462532043},{"id":"https://openalex.org/C21200559","wikidata":"https://www.wikidata.org/wiki/Q7451068","display_name":"Sensitivity (control systems)","level":2,"score":0.3327000141143799},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.3109000027179718},{"id":"https://openalex.org/C2776760102","wikidata":"https://www.wikidata.org/wiki/Q5139990","display_name":"Code (set theory)","level":3,"score":0.30979999899864197},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.2994000017642975},{"id":"https://openalex.org/C206688291","wikidata":"https://www.wikidata.org/wiki/Q7617819","display_name":"Stochastic gradient descent","level":3,"score":0.29249998927116394},{"id":"https://openalex.org/C120936955","wikidata":"https://www.wikidata.org/wiki/Q2155640","display_name":"Empirical research","level":2,"score":0.27399998903274536},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.27379998564720154},{"id":"https://openalex.org/C74750220","wikidata":"https://www.wikidata.org/wiki/Q2662197","display_name":"Differential evolution","level":2,"score":0.2727000117301941},{"id":"https://openalex.org/C2986577269","wikidata":"https://www.wikidata.org/wiki/Q11306265","display_name":"Random noise","level":2,"score":0.2702000141143799},{"id":"https://openalex.org/C3018263672","wikidata":"https://www.wikidata.org/wiki/Q1296251","display_name":"Efficient algorithm","level":2,"score":0.2685999870300293},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.26570001244544983},{"id":"https://openalex.org/C160920958","wikidata":"https://www.wikidata.org/wiki/Q7662746","display_name":"Synthetic data","level":2,"score":0.2533999979496002}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2602.19945","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2602.19945","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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":null,"is_accepted":false,"is_published":null,"raw_source_name":null,"raw_type":"article"}],"best_oa_location":{"id":"doi:10.48550/arxiv.2602.19945","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2602.19945","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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"article"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Balancing":[0],"convergence":[1,119],"efficiency":[2],"and":[3,21,45,62,95,114,140,143],"robustness":[4],"under":[5,86],"Differential":[6],"Privacy":[7],"(DP)":[8],"is":[9,158],"a":[10,116],"central":[11],"challenge":[12],"in":[13,23,160],"Federated":[14],"Learning":[15],"(FL).":[16],"While":[17],"AdamW":[18,66,85],"accelerates":[19],"training":[20],"fine-tuning":[22],"large-scale":[24],"models,":[25],"we":[26,108],"find":[27],"that":[28],"directly":[29],"applying":[30],"it":[31],"to":[32,68,99,103],"Differentially":[33],"Private":[34],"FL":[35],"(DPFL)":[36],"suffers":[37],"from":[38],"three":[39],"major":[40],"issues:":[41],"(i)":[42],"data":[43],"heterogeneity":[44,123],"privacy":[46],"noise":[47],"jointly":[48],"amplify":[49,65],"the":[50,59,77,100,134,151],"variance":[51],"of":[52,136],"second-moment":[53,60,90,112],"estimator,":[54,61],"(ii)":[55],"DP":[56,64,87],"perturbations":[57],"bias":[58],"(iii)":[63],"sensitivity":[67],"local":[69,97],"overfitting,":[70],"worsening":[71],"client":[72,105],"drift.":[73,106],"We":[74],"propose":[75],"DP-FedAdamW,":[76],"first":[78],"AdamW-based":[79],"optimizer":[80],"for":[81],"DPFL.":[82],"It":[83],"restores":[84],"by":[88,154],"stabilizing":[89],"variance,":[91],"removing":[92],"DP-induced":[93],"bias,":[94],"aligning":[96],"updates":[98],"global":[101],"descent":[102],"curb":[104],"Theoretically,":[107],"establish":[109],"an":[110],"unbiased":[111],"estimator":[113],"prove":[115],"linearly":[117],"accelerated":[118],"rate":[120],"without":[121],"any":[122],"assumption,":[124],"while":[125],"providing":[126],"tighter":[127],"$(\\varepsilon,\u03b4)$-DP":[128],"guarantees.":[129],"Our":[130],"empirical":[131],"results":[132],"demonstrate":[133],"effectiveness":[135],"DP-FedAdamW":[137,149],"across":[138],"language":[139],"vision":[141],"Transformers":[142],"ResNet-18.":[144],"On":[145],"Tiny-ImageNet":[146],"(Swin-Base,":[147],"$\\varepsilon=1$),":[148],"outperforms":[150],"state-of-the-art":[152],"(SOTA)":[153],"5.83\\%.":[155],"The":[156],"code":[157],"available":[159],"Appendix.":[161]},"counts_by_year":[],"updated_date":"2026-02-26T06:34:08.959763","created_date":"2026-02-26T00:00:00"}