{"id":"https://openalex.org/W7213103011","doi":"https://doi.org/10.48550/arxiv.2609.12367","title":"Membership Inference via Pairwise Likelihood Ratios","display_name":"Membership Inference via Pairwise Likelihood Ratios","publication_year":2026,"publication_date":"2026-09-11","ids":{"openalex":"https://openalex.org/W7213103011","doi":"https://doi.org/10.48550/arxiv.2609.12367"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2609.12367","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2609.12367","pdf_url":"https://arxiv.org/pdf/2609.12367","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/2609.12367","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5153486799","display_name":"Shengjie Niu","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I14243506","display_name":"Hong Kong Polytechnic University","ror":"https://ror.org/0030zas98","country_code":"HK","type":"education","lineage":["https://openalex.org/I14243506"]}],"countries":["HK"],"is_corresponding":false,"raw_author_name":"Niu, Shengjie","raw_affiliation_strings":["The Hong Kong Polytechnic University , Hong Kong , China,"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"The Hong Kong Polytechnic University , Hong Kong , China,","institution_ids":["https://openalex.org/I14243506"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5113290360","display_name":"Zebin Yun","orcid":"https://orcid.org/0009-0005-7554-9529","observed_orcids":["https://orcid.org/0009-0005-7554-9529"]},"institutions":[{"id":"https://openalex.org/I14243506","display_name":"Hong Kong Polytechnic University","ror":"https://ror.org/0030zas98","country_code":"HK","type":"education","lineage":["https://openalex.org/I14243506"]},{"id":"https://openalex.org/I16391192","display_name":"Tel Aviv University","ror":"https://ror.org/04mhzgx49","country_code":"IL","type":"education","lineage":["https://openalex.org/I16391192"]}],"countries":["HK","IL"],"is_corresponding":false,"raw_author_name":"Yun, Zebin","raw_affiliation_strings":["Tel Aviv University , Israel ,","The Hong Kong Polytechnic University , Hong Kong , China,"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Tel Aviv University , Israel ,","institution_ids":["https://openalex.org/I16391192"]},{"raw_affiliation_string":"The Hong Kong Polytechnic University , Hong Kong , China,","institution_ids":["https://openalex.org/I14243506"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5051771646","display_name":"Yeheng Ge","orcid":"https://orcid.org/0009-0007-4982-0068","observed_orcids":["https://orcid.org/0009-0007-4982-0068"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Ge, Yeheng","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5151652576","display_name":"Jian Huang","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I14243506","display_name":"Hong Kong Polytechnic University","ror":"https://ror.org/0030zas98","country_code":"HK","type":"education","lineage":["https://openalex.org/I14243506"]}],"countries":["HK"],"is_corresponding":false,"raw_author_name":"Huang, Jian","raw_affiliation_strings":["The Hong Kong Polytechnic University , Hong Kong , China,"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"The Hong Kong Polytechnic University , Hong Kong , China,","institution_ids":["https://openalex.org/I14243506"]}]}],"institutions":[],"countries_distinct_count":2,"institutions_distinct_count":2,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":true,"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":0.9922000169754028,"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":0.9922000169754028,"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":0.0066999997943639755,"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/T11045","display_name":"Privacy, Security, and Data Protection","score":9.999999747378752e-05,"subfield":{"id":"https://openalex.org/subfields/3312","display_name":"Sociology and Political Science"},"field":{"id":"https://openalex.org/fields/33","display_name":"Social Sciences"},"domain":{"id":"https://openalex.org/domains/2","display_name":"Social Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/membership-inference-attacks","display_name":"membership inference attacks","score":0.9729999899864197},{"id":"https://openalex.org/keywords/privacy-auditing","display_name":"privacy auditing","score":0.6919999718666077},{"id":"https://openalex.org/keywords/machine-learning-privacy","display_name":"machine learning privacy","score":0.8299999833106995},{"id":"https://openalex.org/keywords/model-output-statistics","display_name":"model output statistics","score":0.5379999876022339}],"concepts":[{"id":"https://openalex.org/C184898388","wikidata":"https://www.wikidata.org/wiki/Q1435712","display_name":"Pairwise comparison","level":2,"score":0.8047999739646912},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.6092000007629395},{"id":"https://openalex.org/C49344536","wikidata":"https://www.wikidata.org/wiki/Q726441","display_name":"Cauchy distribution","level":2,"score":0.5673999786376953},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.5058000087738037},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.48660001158714294},{"id":"https://openalex.org/C134261354","wikidata":"https://www.wikidata.org/wiki/Q938438","display_name":"Statistical inference","level":2,"score":0.4778999984264374},{"id":"https://openalex.org/C2908647359","wikidata":"https://www.wikidata.org/wiki/Q2625603","display_name":"Population","level":2,"score":0.446399986743927},{"id":"https://openalex.org/C89128539","wikidata":"https://www.wikidata.org/wiki/Q1949963","display_name":"Statistic","level":2,"score":0.42570000886917114},{"id":"https://openalex.org/C2776401178","wikidata":"https://www.wikidata.org/wiki/Q12050496","display_name":"Feature (linguistics)","level":2,"score":0.414000004529953},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.4065000116825104},{"id":"https://openalex.org/C105795698","wikidata":"https://www.wikidata.org/wiki/Q12483","display_name":"Statistics","level":1,"score":0.38530001044273376},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.38179999589920044},{"id":"https://openalex.org/C95167961","wikidata":"https://www.wikidata.org/wiki/Q4483495","display_name":"Fiducial inference","level":5,"score":0.3797999918460846},{"id":"https://openalex.org/C165838908","wikidata":"https://www.wikidata.org/wiki/Q736777","display_name":"Calibration","level":2,"score":0.37599998712539673},{"id":"https://openalex.org/C153180895","wikidata":"https://www.wikidata.org/wiki/Q7148389","display_name":"Pattern recognition (psychology)","level":2,"score":0.3589000105857849},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.34619998931884766},{"id":"https://openalex.org/C178197554","wikidata":"https://www.wikidata.org/wiki/Q1099110","display_name":"Sufficient statistic","level":2,"score":0.3456000089645386},{"id":"https://openalex.org/C41426520","wikidata":"https://www.wikidata.org/wiki/Q1192065","display_name":"Point estimation","level":2,"score":0.3440000116825104},{"id":"https://openalex.org/C4679612","wikidata":"https://www.wikidata.org/wiki/Q866298","display_name":"Aggregate (composite)","level":2,"score":0.33320000767707825},{"id":"https://openalex.org/C163716315","wikidata":"https://www.wikidata.org/wiki/Q901177","display_name":"Gaussian","level":2,"score":0.32690000534057617},{"id":"https://openalex.org/C28719098","wikidata":"https://www.wikidata.org/wiki/Q44946","display_name":"Point (geometry)","level":2,"score":0.3197000026702881},{"id":"https://openalex.org/C87007009","wikidata":"https://www.wikidata.org/wiki/Q210832","display_name":"Statistical hypothesis testing","level":2,"score":0.3068000078201294},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.289900004863739},{"id":"https://openalex.org/C177264268","wikidata":"https://www.wikidata.org/wiki/Q1514741","display_name":"Set (abstract data type)","level":2,"score":0.28690001368522644},{"id":"https://openalex.org/C99498987","wikidata":"https://www.wikidata.org/wiki/Q2210247","display_name":"Noise (video)","level":3,"score":0.2809999883174896},{"id":"https://openalex.org/C114289077","wikidata":"https://www.wikidata.org/wiki/Q3284399","display_name":"Statistical model","level":2,"score":0.2770000100135803},{"id":"https://openalex.org/C48372109","wikidata":"https://www.wikidata.org/wiki/Q3913","display_name":"Binary number","level":2,"score":0.27070000767707825},{"id":"https://openalex.org/C61326573","wikidata":"https://www.wikidata.org/wiki/Q1496376","display_name":"Gaussian process","level":3,"score":0.2556000053882599}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2609.12367","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2609.12367","pdf_url":"https://arxiv.org/pdf/2609.12367","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.2609.12367","doi":"https://doi.org/10.48550/arxiv.2609.12367","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2609.12367","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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"pmh:oai:arXiv.org:2609.12367","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2609.12367","pdf_url":"https://arxiv.org/pdf/2609.12367","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":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7213103011.pdf","grobid_xml":"https://content.openalex.org/works/W7213103011.grobid-xml"},"referenced_works_count":6,"referenced_works":["https://openalex.org/W2807096445","https://openalex.org/W2964137095","https://openalex.org/W3028302264","https://openalex.org/W3211926155","https://openalex.org/W4394639701","https://openalex.org/W4414581747"],"related_works":["https://openalex.org/W4389471066","https://openalex.org/W4410609097","https://openalex.org/W7167900452","https://openalex.org/W4415311347","https://openalex.org/W7065963399","https://openalex.org/W7140302597","https://openalex.org/W4288057780","https://openalex.org/W7138916420","https://openalex.org/W4206079793","https://openalex.org/W6979350007"],"abstract_inverted_index":{"Membership":[0],"inference":[1,39],"attacks":[2],"(MIAs)":[3],"are":[4],"the":[5,10,33,43,48,97,105,130,146,152,177,186],"standard":[6],"tool":[7],"for":[8,138,212],"auditing":[9],"privacy":[11,214],"risks":[12],"of":[13],"machine":[14],"learning":[15],"models.":[16],"Given":[17],"a":[18,83,88,165],"query":[19,131],"point,":[20],"an":[21],"MIA":[22,81],"aims":[23],"to":[24,31,65,164],"determine":[25],"whether":[26],"that":[27,86,155,171],"point":[28,132],"was":[29],"used":[30,137],"train":[32],"target":[34],"model.":[35],"In":[36],"practice,":[37],"such":[38,51],"must":[40],"rely":[41],"on":[42],"statistical":[44,71,198],"signals":[45,109,144],"exposed":[46],"by":[47,182],"model's":[49],"outputs,":[50],"as":[52],"confidence":[53],"scores,":[54],"logits,":[55],"and":[56,68,96,110,117,133,140,209],"intermediate":[57],"feature":[58],"representations.":[59],"However,":[60],"existing":[61],"methods":[62],"often":[63],"fail":[64],"efficiently":[66],"summarize":[67],"combine":[69],"these":[70,142],"signals.":[72],"To":[73],"address":[74],"this":[75],"limitation,":[76],"we":[77],"propose":[78],"Pairwise":[79],"Likelihood":[80],"(PL-MIA),":[82],"unified":[84],"method":[85],"combines":[87],"Gaussian":[89],"likelihood-ratio":[90],"(GLR)":[91],"statistic":[92],"with":[93],"population":[94,115],"calibration":[95,116],"Cauchy":[98,118,147],"combination":[99,119,148],"test.":[100,149],"We":[101,123],"characterize":[102],"theoretically":[103],"how":[104,197],"GLR":[106],"retains":[107],"variance-contraction":[108],"establish":[111],"conditions":[112],"under":[113],"which":[114],"improve":[120],"attack":[121],"power.":[122],"obtain":[124],"$p$-values":[125],"from":[126],"pairwise":[127,160],"comparisons":[128],"between":[129],"reference":[134],"points":[135],"not":[136],"training,":[139],"aggregate":[141],"continuous":[143],"using":[145],"This":[150],"preserves":[151],"evidence":[153,211],"strength":[154],"is":[156,162],"discarded":[157],"when":[158],"each":[159],"comparison":[161],"reduced":[163],"binary":[166],"vote.":[167],"Extensive":[168],"experiments":[169],"demonstrate":[170,196],"PL-MIA":[172],"outperforms":[173],"strong":[174],"baselines,":[175],"improving":[176],"true":[178],"positive":[179],"rate":[180],"(TPR)":[181],"over":[183],"25\\%":[184],"in":[185],"critical":[187],"low-false-positive":[188],"regime,":[189],"corroborating":[190],"our":[191],"theoretical":[192],"findings.":[193],"These":[194],"results":[195],"principles":[199],"can":[200],"turn":[201],"noisy":[202],"model":[203],"outputs":[204],"into":[205],"more":[206],"powerful,":[207],"calibrated,":[208],"reproducible":[210],"membership":[213],"auditing.":[215]},"counts_by_year":[],"updated_date":"2026-11-07T11:25:07.637180","created_date":"2026-09-15T00:00:00"}