{"id":"https://openalex.org/W7166687504","doi":"https://doi.org/10.48550/arxiv.2606.30081","title":"Discard the Dross and Select the Essential: Pre-query Sample Selection for Black-box Membership Inference Attacks","display_name":"Discard the Dross and Select the Essential: Pre-query Sample Selection for Black-box Membership Inference Attacks","publication_year":2026,"publication_date":"2026-06-29","ids":{"openalex":"https://openalex.org/W7166687504","doi":"https://doi.org/10.48550/arxiv.2606.30081"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2606.30081","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.30081","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.2606.30081","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5085817734","display_name":"Dan Zhao","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhao, Dongdong","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5028521210","display_name":"Jinrong Hu","orcid":"https://orcid.org/0000-0001-7732-8141"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Hu, Jinrong","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5139681585","display_name":"Changtian Song","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Song, Changtian","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5139694211","display_name":"Jian Chen","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Chen, Jian","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5139672290","display_name":"Hongmin Wang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wang, Hongmin","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5108924109","display_name":"Baogang Song","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Song, Baogang","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/T11689","display_name":"Adversarial Robustness in Machine Learning","score":0.5945000052452087,"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/T11689","display_name":"Adversarial Robustness in Machine Learning","score":0.5945000052452087,"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":0.1527000069618225,"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/T11273","display_name":"Advanced Graph Neural Networks","score":0.04859999939799309,"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/ranking","display_name":"Ranking (information retrieval)","score":0.7325999736785889},{"id":"https://openalex.org/keywords/selection","display_name":"Selection (genetic algorithm)","score":0.6520000100135803},{"id":"https://openalex.org/keywords/inference","display_name":"Inference","score":0.6104000210762024},{"id":"https://openalex.org/keywords/sample","display_name":"Sample (material)","score":0.5134000182151794},{"id":"https://openalex.org/keywords/key","display_name":"Key (lock)","score":0.3910999894142151}],"concepts":[{"id":"https://openalex.org/C189430467","wikidata":"https://www.wikidata.org/wiki/Q7293293","display_name":"Ranking (information retrieval)","level":2,"score":0.7325999736785889},{"id":"https://openalex.org/C81917197","wikidata":"https://www.wikidata.org/wiki/Q628760","display_name":"Selection (genetic algorithm)","level":2,"score":0.6520000100135803},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.6310999989509583},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.6104000210762024},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.5831999778747559},{"id":"https://openalex.org/C198531522","wikidata":"https://www.wikidata.org/wiki/Q485146","display_name":"Sample (material)","level":2,"score":0.5134000182151794},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.3910999894142151},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.304500013589859},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.3025999963283539},{"id":"https://openalex.org/C2779343474","wikidata":"https://www.wikidata.org/wiki/Q3109175","display_name":"Context (archaeology)","level":2,"score":0.2824999988079071},{"id":"https://openalex.org/C148483581","wikidata":"https://www.wikidata.org/wiki/Q446488","display_name":"Feature selection","level":2,"score":0.25529998540878296}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2606.30081","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.30081","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.2606.30081","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.30081","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/10","score":0.6500011086463928,"display_name":"Reduced inequalities"}],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Black-box":[0],"membership":[1,19,111,154],"inference":[2],"attacks":[3],"(MIAs)":[4],"rely":[5],"on":[6,164],"target-model":[7],"queries":[8,114],"to":[9,108],"infer":[10],"whether":[11],"candidate":[12,27,64,101,146],"samples":[13,28,65,102,117,147],"were":[14],"used":[15],"for":[16,56,69,123,205],"training.":[17],"However,":[18],"signals":[20],"are":[21,66],"highly":[22],"non-uniform":[23],"across":[24],"samples:":[25],"some":[26],"support":[29,109],"strong":[30],"member/non-member":[31],"separability,":[32],"whereas":[33],"many":[34],"others":[35],"provide":[36],"little":[37],"useful":[38],"signal.":[39],"Consequently,":[40],"indiscriminate":[41],"querying":[42,68],"can":[43,85,193],"incur":[44],"substantial":[45],"query":[46,130,203],"cost":[47,131],"and":[48,103,118,132,167,199],"increase":[49],"query-induced":[50,133],"exposure,":[51],"with":[52,88,169,178],"limited":[53],"marginal":[54],"benefit":[55],"inference.":[57],"This":[58],"raises":[59],"a":[60,79,105,188],"key":[61],"question:":[62],"which":[63,84,144],"worth":[67],"black-box":[70,126,172],"MIAs?":[71],"To":[72],"address":[73],"this":[74],"question,":[75],"we":[76,139],"propose":[77,140],"PSS-MIA,":[78],"pre-query":[80],"sample":[81],"selection":[82],"framework":[83],"be":[86],"embedded":[87],"any":[89],"existing":[90,125],"MIA":[91,173],"methods.":[92,185],"PSS-MIA":[93,177,192],"proceeds":[94],"in":[95],"two":[96],"stages:":[97],"it":[98],"first":[99,137],"ranks":[100,145],"selects":[104],"subset":[106],"expected":[107],"stronger":[110],"inference,":[112],"then":[113],"the":[115,120,136,150,202,206],"selected":[116],"uses":[119],"returned":[121],"outputs":[122],"an":[124],"MIA,":[127],"thereby":[128],"reducing":[129],"exposure.":[134],"In":[135],"stage,":[138],"Loss-Gap":[141],"Ranking":[142],"(LGR),":[143],"by":[148],"estimating":[149],"strength":[151],"of":[152,201],"their":[153],"signal":[155],"using":[156],"loss":[157],"gaps":[158],"computed":[159],"from":[160],"reference":[161],"models.":[162],"Experiments":[163],"CIFAR-10,":[165],"CIFAR-100,":[166],"CINIC-10":[168],"five":[170],"representative":[171],"methods":[174],"demonstrate":[175],"that":[176],"LGR":[179],"consistently":[180],"outperforms":[181],"all":[182],"other":[183],"compared":[184],"Moreover,":[186],"under":[187],"0.1%":[189],"FPR":[190],"constraint,":[191],"save":[194],"at":[195],"least":[196],"83.1%,":[197],"60.6%,":[198],"80.4%":[200],"budget":[204],"three":[207],"datasets,":[208],"respectively.":[209]},"counts_by_year":[],"updated_date":"2026-07-01T06:29:00.853634","created_date":"2026-07-01T00:00:00"}