{"id":"https://openalex.org/W7129080484","doi":"https://doi.org/10.48550/arxiv.2602.12943","title":"Neighborhood Blending: A Lightweight Inference-Time Defense Against Membership Inference Attacks","display_name":"Neighborhood Blending: A Lightweight Inference-Time Defense Against Membership Inference Attacks","publication_year":2026,"publication_date":"2026-02-13","ids":{"openalex":"https://openalex.org/W7129080484","doi":"https://doi.org/10.48550/arxiv.2602.12943"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2602.12943","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2602.12943","pdf_url":"https://arxiv.org/pdf/2602.12943","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,"listed_in":[],"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":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"volume":null,"type":"preprint","is_traditional":true,"indexed_in":["arxiv","datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/2602.12943","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5114353041","display_name":"Osama Zafar","orcid":"https://orcid.org/0009-0008-9621-6899","observed_orcids":["https://orcid.org/0009-0008-9621-6899"]},"institutions":[{"id":"https://openalex.org/I58956616","display_name":"Case Western Reserve University","ror":"https://ror.org/051fd9666","country_code":"US","type":"education","lineage":["https://openalex.org/I58956616"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Zafar, Osama","raw_affiliation_strings":["Case Western Reserve University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Case Western Reserve University","institution_ids":["https://openalex.org/I58956616"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5126149960","display_name":"Shaojie Zhan","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I12315562","display_name":"Texas Tech University","ror":"https://ror.org/0405mnx93","country_code":"US","type":"education","lineage":["https://openalex.org/I12315562"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Zhan, Shaojie","raw_affiliation_strings":["Texas Tech University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Texas Tech University","institution_ids":["https://openalex.org/I12315562"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5010253529","display_name":"Tianxi Ji","orcid":"https://orcid.org/0000-0002-9083-5452","observed_orcids":["https://orcid.org/0000-0002-9083-5452"]},"institutions":[{"id":"https://openalex.org/I12315562","display_name":"Texas Tech University","ror":"https://ror.org/0405mnx93","country_code":"US","type":"education","lineage":["https://openalex.org/I12315562"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Ji, Tianxi","raw_affiliation_strings":["Texas Tech University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Texas Tech University","institution_ids":["https://openalex.org/I12315562"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5028326739","display_name":"Erman Ayday","orcid":"https://orcid.org/0000-0003-3383-1081","observed_orcids":["https://orcid.org/0000-0003-3383-1081"]},"institutions":[{"id":"https://openalex.org/I58956616","display_name":"Case Western Reserve University","ror":"https://ror.org/051fd9666","country_code":"US","type":"education","lineage":["https://openalex.org/I58956616"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Ayday, Erman","raw_affiliation_strings":["Case Western Reserve University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Case Western Reserve University","institution_ids":["https://openalex.org/I58956616"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":2,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":true,"cited_by_count":1,"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":0.8144999742507935,"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":0.8144999742507935,"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/T11689","display_name":"Adversarial Robustness in Machine Learning","score":0.18479999899864197,"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/T11424","display_name":"Security and Verification in Computing","score":9.999999747378752e-05,"subfield":{"id":"https://openalex.org/subfields/1708","display_name":"Hardware and Architecture"},"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/membership-inference-attacks","display_name":"membership inference attacks","score":0.9919999837875366},{"id":"https://openalex.org/keywords/privacy-preserving-machine-learning","display_name":"privacy-preserving machine learning","score":0.7860000133514404},{"id":"https://openalex.org/keywords/training-data-privacy","display_name":"training data privacy","score":0.5059999823570251},{"id":"https://openalex.org/keywords/model-agnostic-defense","display_name":"model-agnostic defense","score":0.7009999752044678},{"id":"https://openalex.org/keywords/privacy-utility-trade-off","display_name":"privacy-utility trade-off","score":0.6650000214576721},{"id":"https://openalex.org/keywords/machine-learning-as-a-service","display_name":"machine learning as a service","score":0.7829999923706055}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7699999809265137},{"id":"https://openalex.org/C165696696","wikidata":"https://www.wikidata.org/wiki/Q11287","display_name":"Exploit","level":2,"score":0.7276999950408936},{"id":"https://openalex.org/C41065033","wikidata":"https://www.wikidata.org/wiki/Q2825412","display_name":"Adversary","level":2,"score":0.6812999844551086},{"id":"https://openalex.org/C37736160","wikidata":"https://www.wikidata.org/wiki/Q1801315","display_name":"Adversarial system","level":2,"score":0.6689000129699707},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.631600022315979},{"id":"https://openalex.org/C140547941","wikidata":"https://www.wikidata.org/wiki/Q7797194","display_name":"Threat model","level":2,"score":0.47450000047683716},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.45840001106262207},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.4404999911785126},{"id":"https://openalex.org/C38652104","wikidata":"https://www.wikidata.org/wiki/Q3510521","display_name":"Computer security","level":1,"score":0.43309998512268066},{"id":"https://openalex.org/C51632099","wikidata":"https://www.wikidata.org/wiki/Q3985153","display_name":"Training set","level":2,"score":0.4083000123500824},{"id":"https://openalex.org/C3017597292","wikidata":"https://www.wikidata.org/wiki/Q25052250","display_name":"Privacy protection","level":2,"score":0.38280001282691956},{"id":"https://openalex.org/C205711294","wikidata":"https://www.wikidata.org/wiki/Q176953","display_name":"Rendering (computer graphics)","level":2,"score":0.3801000118255615},{"id":"https://openalex.org/C77618280","wikidata":"https://www.wikidata.org/wiki/Q1155772","display_name":"Scheme (mathematics)","level":2,"score":0.37619999051094055},{"id":"https://openalex.org/C3770464","wikidata":"https://www.wikidata.org/wiki/Q775963","display_name":"Smoothing","level":2,"score":0.36410000920295715},{"id":"https://openalex.org/C177264268","wikidata":"https://www.wikidata.org/wiki/Q1514741","display_name":"Set (abstract data type)","level":2,"score":0.34689998626708984},{"id":"https://openalex.org/C2778712577","wikidata":"https://www.wikidata.org/wiki/Q3505966","display_name":"Retraining","level":2,"score":0.34049999713897705},{"id":"https://openalex.org/C123201435","wikidata":"https://www.wikidata.org/wiki/Q456632","display_name":"Information privacy","level":2,"score":0.31869998574256897},{"id":"https://openalex.org/C2780378061","wikidata":"https://www.wikidata.org/wiki/Q25351891","display_name":"Service (business)","level":2,"score":0.3174999952316284},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.29190000891685486},{"id":"https://openalex.org/C67186912","wikidata":"https://www.wikidata.org/wiki/Q367664","display_name":"Data modeling","level":2,"score":0.27480000257492065},{"id":"https://openalex.org/C23130292","wikidata":"https://www.wikidata.org/wiki/Q5275358","display_name":"Differential privacy","level":2,"score":0.2630999982357025},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.25780001282691956}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2602.12943","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2602.12943","pdf_url":"https://arxiv.org/pdf/2602.12943","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,"listed_in":[],"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":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},{"id":"doi:10.48550/arxiv.2602.12943","doi":"https://doi.org/10.48550/arxiv.2602.12943","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2602.12943","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,"listed_in":[],"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"}],"best_oa_location":{"id":"pmh:oai:arXiv.org:2602.12943","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2602.12943","pdf_url":"https://arxiv.org/pdf/2602.12943","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,"listed_in":[],"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":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"sustainable_development_goals":[],"study_designs":[],"awards":[],"funders":[],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7129080484.pdf","grobid_xml":"https://content.openalex.org/works/W7129080484.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W7168318436","https://openalex.org/W4392307024","https://openalex.org/W4386230986","https://openalex.org/W4311165761","https://openalex.org/W7167700889","https://openalex.org/W4315706478","https://openalex.org/W4412599059","https://openalex.org/W4416666727","https://openalex.org/W4206161712","https://openalex.org/W4386214387"],"abstract_inverted_index":{"In":[0,90],"recent":[1],"years,":[2],"the":[3,48,108,121,127],"widespread":[4],"adoption":[5],"of":[6,129,236],"Machine":[7],"Learning":[8],"as":[9,61,77],"a":[10,42,95,130,148,185,190],"Service":[11],"(MLaaS),":[12],"particularly":[13],"in":[14,47,68,234],"sensitive":[15],"environments,":[16],"has":[17],"raised":[18],"considerable":[19],"privacy":[20,55,196],"concerns.":[21],"Of":[22],"particular":[23],"importance":[24],"are":[25],"membership":[26],"inference":[27],"attacks":[28],"(MIAs),":[29],"which":[30,103],"exploit":[31],"behavioral":[32],"discrepancies":[33],"between":[34],"training":[35,50,138],"and":[36,65,154,173,207,229],"non-training":[37],"data":[38],"to":[39,157],"determine":[40],"whether":[41],"specific":[43],"record":[44],"was":[45],"included":[46],"model's":[49,122],"set,":[51],"thereby":[52],"presenting":[53],"significant":[54,112],"risks.":[56],"Although":[57],"existing":[58,224],"defenses,":[59],"such":[60,76],"adversarial":[62],"regularization,":[63],"DP-SGD,":[64],"MemGuard,":[66],"assist":[67],"mitigating":[69],"these":[70],"threats,":[71],"they":[72],"often":[73],"entail":[74],"trade-offs":[75],"compromising":[78],"utility,":[79],"increased":[80],"computational":[81,113],"requirements,":[82],"or":[83,110],"inconsistent":[84],"protection":[85],"against":[86],"diverse":[87,205],"attack":[88],"vectors.":[89],"this":[91],"paper,":[92],"we":[93,209],"introduce":[94],"novel":[96],"inference-time":[97],"defense":[98,213],"mechanism":[99],"called":[100],"Neighborhood":[101,165],"Blending,":[102],"mitigates":[104],"MIAs":[105],"without":[106,197],"retraining":[107],"model":[109,199,221],"incurring":[111],"overhead.":[114],"Our":[115],"approach":[116,187],"operates":[117],"post-training":[118],"by":[119],"smoothing":[120],"confidence":[123,150],"outputs":[124],"based":[125],"on":[126],"neighborhood":[128],"queried":[131],"sample.":[132],"By":[133],"averaging":[134],"predictions":[135],"from":[136],"similar":[137],"samples":[139],"selected":[140],"using":[141],"differentially":[142],"private":[143],"sampling,":[144],"our":[145,212],"method":[146],"establishes":[147],"consistent":[149],"pattern,":[151],"rendering":[152],"members":[153],"non-members":[155],"indistinguishable":[156],"an":[158,178],"adversary":[159],"while":[160,219],"maintaining":[161],"high":[162,175],"utility.":[163,200],"Significantly,":[164],"Blending":[166],"maintains":[167],"label":[168,171],"integrity":[169],"(zero":[170],"loss)":[172],"ensures":[174],"utility":[176,237],"through":[177],"adaptive,":[179],"\"pay-as-you-go\"":[180],"distortion":[181],"strategy.":[182],"It":[183],"is":[184],"model-agnostic":[186],"that":[188,194,211],"offers":[189],"practical,":[191],"lightweight":[192],"solution":[193],"enhances":[195],"sacrificing":[198],"Through":[201],"extensive":[202],"experiments":[203],"across":[204],"datasets":[206],"models,":[208],"demonstrate":[210],"significantly":[214],"reduces":[215],"MIA":[216],"success":[217],"rates":[218],"preserving":[220],"performance,":[222],"outperforming":[223],"post-hoc":[225],"defenses":[226],"like":[227,232],"MemGuard":[228],"training-time":[230],"techniques":[231],"DP-SGD":[233],"terms":[235],"retention.":[238]},"counts_by_year":[{"year":2026,"cited_by_count":1}],"updated_date":"2026-11-05T12:25:35.769452","created_date":"2026-02-17T00:00:00"}