{"id":"https://openalex.org/W4415063166","doi":"https://doi.org/10.48550/arxiv.2503.21528","title":"Bayesian Pseudo Posterior Mechanism for Differentially Private Machine Learning","display_name":"Bayesian Pseudo Posterior Mechanism for Differentially Private Machine Learning","publication_year":2025,"publication_date":"2025-03-27","ids":{"openalex":"https://openalex.org/W4415063166","doi":"https://doi.org/10.48550/arxiv.2503.21528"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2503.21528","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2503.21528","pdf_url":"https://arxiv.org/pdf/2503.21528","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"},"parent_works":[],"lineage":["https://openalex.org/W4415063166"],"child_works_count":0,"type":"preprint","is_traditional":true,"indexed_in":["arxiv","datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/2503.21528","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5078399314","display_name":"Robert Chew","orcid":"https://orcid.org/0000-0002-6979-1766","observed_orcids":["https://orcid.org/0000-0002-6979-1766","https://orcid.org/0000-0001-9875-2371"]},"institutions":[{"id":"https://openalex.org/I180297670","display_name":"RTI International","ror":"https://ror.org/052tfza37","country_code":"US","type":"nonprofit","lineage":["https://openalex.org/I137902535","https://openalex.org/I170897317","https://openalex.org/I180297670"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Chew, Robert","raw_affiliation_strings":["Center for Data Science and AI RTI International Research Triangle Park , NC 27709 , USA"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Center for Data Science and AI RTI International Research Triangle Park , NC 27709 , USA","institution_ids":["https://openalex.org/I180297670"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5002518499","display_name":"Matthew R. Williams","orcid":"https://orcid.org/0000-0001-8894-1240","observed_orcids":["https://orcid.org/0000-0001-8894-1240"]},"institutions":[{"id":"https://openalex.org/I180297670","display_name":"RTI International","ror":"https://ror.org/052tfza37","country_code":"US","type":"nonprofit","lineage":["https://openalex.org/I137902535","https://openalex.org/I170897317","https://openalex.org/I180297670"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Williams, Matthew R.","raw_affiliation_strings":["Center for Official Statistics RTI International Washington , DC 20005 , USA"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Center for Official Statistics RTI International Washington , DC 20005 , USA","institution_ids":["https://openalex.org/I180297670"]}]},{"author_position":"middle","author":{"id":null,"display_name":"Segarra, Elan A.","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I180297670","display_name":"RTI International","ror":"https://ror.org/052tfza37","country_code":"US","type":"nonprofit","lineage":["https://openalex.org/I137902535","https://openalex.org/I170897317","https://openalex.org/I180297670"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Segarra, Elan A.","raw_affiliation_strings":["Center for Official Statistics RTI International Washington , DC 20005 , USA"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Center for Official Statistics RTI International Washington , DC 20005 , USA","institution_ids":["https://openalex.org/I180297670"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5067569497","display_name":"Alexander Preiss","orcid":"https://orcid.org/0000-0002-7966-527X","observed_orcids":["https://orcid.org/0000-0002-7966-527X"]},"institutions":[{"id":"https://openalex.org/I1290095210","display_name":"Bureau of Labor Statistics","ror":"https://ror.org/03z2d7y96","country_code":"US","type":"government","lineage":["https://openalex.org/I1290095210","https://openalex.org/I1312264882"]},{"id":"https://openalex.org/I180297670","display_name":"RTI International","ror":"https://ror.org/052tfza37","country_code":"US","type":"nonprofit","lineage":["https://openalex.org/I137902535","https://openalex.org/I170897317","https://openalex.org/I180297670"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Preiss, Alexander J.","raw_affiliation_strings":["Center for Data Science and AI RTI International Research Triangle Park , NC 27709 , USA","Office of Compensation and Working Conditions U.S. Bureau of Labor Statistics Hillcrest Heights , MD 20746 , USA"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Center for Data Science and AI RTI International Research Triangle Park , NC 27709 , USA","institution_ids":["https://openalex.org/I180297670"]},{"raw_affiliation_string":"Office of Compensation and Working Conditions U.S. Bureau of Labor Statistics Hillcrest Heights , MD 20746 , USA","institution_ids":["https://openalex.org/I1290095210"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5014098617","display_name":"Amanda Konet","orcid":"https://orcid.org/0000-0001-5295-6483","observed_orcids":["https://orcid.org/0000-0001-5295-6483"]},"institutions":[{"id":"https://openalex.org/I180297670","display_name":"RTI International","ror":"https://ror.org/052tfza37","country_code":"US","type":"nonprofit","lineage":["https://openalex.org/I137902535","https://openalex.org/I170897317","https://openalex.org/I180297670"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Konet, Amanda","raw_affiliation_strings":["Center for Data Science and AI RTI International Research Triangle Park , NC 27709 , USA"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Center for Data Science and AI RTI International Research Triangle Park , NC 27709 , USA","institution_ids":["https://openalex.org/I180297670"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5074507246","display_name":"Terrance D. Savitsky","orcid":"https://orcid.org/0000-0003-1843-3106","observed_orcids":["https://orcid.org/0000-0003-1843-3106"]},"institutions":[{"id":"https://openalex.org/I1290095210","display_name":"Bureau of Labor Statistics","ror":"https://ror.org/03z2d7y96","country_code":"US","type":"government","lineage":["https://openalex.org/I1290095210","https://openalex.org/I1312264882"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Savitsky, Terrance D.","raw_affiliation_strings":["Office of Survey Methods Research U.S. Bureau of Labor Statistics Hillcrest Heights , MD 20746 , USA"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Office of Survey Methods Research U.S. Bureau of Labor Statistics Hillcrest Heights , MD 20746 , USA","institution_ids":["https://openalex.org/I1290095210"]}]}],"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":1.0,"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":1.0,"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":0.0,"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/T11614","display_name":"Cloud Data Security Solutions","score":0.0,"subfield":{"id":"https://openalex.org/subfields/1710","display_name":"Information Systems"},"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":0.9959999918937683},{"id":"https://openalex.org/keywords/dp-sgd","display_name":"DP-SGD","score":0.7279999852180481},{"id":"https://openalex.org/keywords/class-imbalance","display_name":"class imbalance","score":0.7059999704360962},{"id":"https://openalex.org/keywords/disclosure-risk","display_name":"disclosure risk","score":0.7730000019073486},{"id":"https://openalex.org/keywords/text-classification","display_name":"text classification","score":0.6460000276565552},{"id":"https://openalex.org/keywords/official-statistics","display_name":"official statistics","score":0.5230000019073486},{"id":"https://openalex.org/keywords/privacy-utility-trade-off","display_name":"privacy-utility trade-off","score":0.8550000190734863}],"concepts":[{"id":"https://openalex.org/C23130292","wikidata":"https://www.wikidata.org/wiki/Q5275358","display_name":"Differential privacy","level":2,"score":0.7682999968528748},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.7681000232696533},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.7577999830245972},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6754999756813049},{"id":"https://openalex.org/C107673813","wikidata":"https://www.wikidata.org/wiki/Q812534","display_name":"Bayesian probability","level":2,"score":0.5838000178337097},{"id":"https://openalex.org/C57830394","wikidata":"https://www.wikidata.org/wiki/Q278079","display_name":"Posterior probability","level":3,"score":0.5375000238418579},{"id":"https://openalex.org/C2780451532","wikidata":"https://www.wikidata.org/wiki/Q759676","display_name":"Task (project management)","level":2,"score":0.5351999998092651},{"id":"https://openalex.org/C48044578","wikidata":"https://www.wikidata.org/wiki/Q727490","display_name":"Scalability","level":2,"score":0.5264999866485596},{"id":"https://openalex.org/C89611455","wikidata":"https://www.wikidata.org/wiki/Q6804646","display_name":"Mechanism (biology)","level":2,"score":0.4537999927997589},{"id":"https://openalex.org/C99221444","wikidata":"https://www.wikidata.org/wiki/Q1532069","display_name":"Private information retrieval","level":2,"score":0.38040000200271606},{"id":"https://openalex.org/C160234255","wikidata":"https://www.wikidata.org/wiki/Q812535","display_name":"Bayesian inference","level":3,"score":0.3375000059604645},{"id":"https://openalex.org/C75684735","wikidata":"https://www.wikidata.org/wiki/Q858810","display_name":"Big data","level":2,"score":0.3181000053882599},{"id":"https://openalex.org/C123201435","wikidata":"https://www.wikidata.org/wiki/Q456632","display_name":"Information privacy","level":2,"score":0.30809998512268066},{"id":"https://openalex.org/C108583219","wikidata":"https://www.wikidata.org/wiki/Q197536","display_name":"Deep learning","level":2,"score":0.2797999978065491},{"id":"https://openalex.org/C177769412","wikidata":"https://www.wikidata.org/wiki/Q278090","display_name":"Prior probability","level":3,"score":0.2754000127315521},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.263700008392334},{"id":"https://openalex.org/C2781170535","wikidata":"https://www.wikidata.org/wiki/Q30587856","display_name":"Noisy data","level":2,"score":0.257099986076355},{"id":"https://openalex.org/C51632099","wikidata":"https://www.wikidata.org/wiki/Q3985153","display_name":"Training set","level":2,"score":0.2565999925136566},{"id":"https://openalex.org/C2776145971","wikidata":"https://www.wikidata.org/wiki/Q30673951","display_name":"Labeled data","level":2,"score":0.25529998540878296},{"id":"https://openalex.org/C207201462","wikidata":"https://www.wikidata.org/wiki/Q182505","display_name":"Bayes' theorem","level":3,"score":0.2531999945640564}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2503.21528","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2503.21528","pdf_url":"https://arxiv.org/pdf/2503.21528","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.2503.21528","doi":"https://doi.org/10.48550/arxiv.2503.21528","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2503.21528","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:2503.21528","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2503.21528","pdf_url":"https://arxiv.org/pdf/2503.21528","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":[{"score":0.4000000059604645,"display_name":"Reduced inequalities","id":"https://metadata.un.org/sdg/10"}],"study_designs":[],"awards":[],"funders":[{"id":"https://openalex.org/F4320306076","display_name":"National Science Foundation","ror":"https://ror.org/021nxhr62"}],"has_content":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4415063166.pdf","grobid_xml":"https://content.openalex.org/works/W4415063166.grobid-xml"},"referenced_works_count":9,"referenced_works":["https://openalex.org/W1985511977","https://openalex.org/W2473418344","https://openalex.org/W2792287754","https://openalex.org/W2965373594","https://openalex.org/W3033661152","https://openalex.org/W3207429447","https://openalex.org/W4385846436","https://openalex.org/W4399971973","https://openalex.org/W4404390133"],"related_works":["https://openalex.org/W7164034230","https://openalex.org/W2947693768","https://openalex.org/W2953449233","https://openalex.org/W3082856297","https://openalex.org/W4392539606","https://openalex.org/W6891891701","https://openalex.org/W2966439069","https://openalex.org/W4284893082","https://openalex.org/W4400064859","https://openalex.org/W4307003864"],"abstract_inverted_index":{"Differential":[0],"privacy":[1,20,145],"(DP)":[2],"is":[3,25],"becoming":[4],"increasingly":[5],"important":[6],"for":[7,17,64,142],"deployed":[8],"machine":[9,36],"learning":[10,37,66],"applications":[11],"because":[12],"it":[13],"provides":[14],"strong":[15],"guarantees":[16],"protecting":[18],"the":[19,86,114,138],"of":[21],"individuals":[22],"whose":[23],"data":[24],"used":[26,34],"to":[27,39,81],"train":[28],"models.":[29],"However,":[30],"DP":[31,62],"mechanisms":[32],"commonly":[33],"in":[35],"tend":[38],"struggle":[40],"on":[41,99],"many":[42],"real":[43],"world":[44],"distributions,":[45],"including":[46],"highly":[47,108],"imbalanced":[48,109],"or":[49],"small":[50],"labeled":[51],"training":[52],"sets.":[53],"In":[54],"this":[55],"work,":[56],"we":[57,96],"propose":[58],"a":[59,71,90,100,107,132,143],"new":[60],"scalable":[61],"mechanism":[63],"deep":[65],"models,":[67],"SWAG-PPM,":[68],"by":[69,113],"using":[70,106],"pseudo":[72],"posterior":[73],"distribution":[74],"that":[75,124],"downweights":[76],"by-record":[77],"likelihood":[78],"contributions":[79],"proportionally":[80],"their":[82],"disclosure":[83],"risks":[84],"as":[85],"randomized":[87],"mechanism.":[88],"As":[89],"motivating":[91],"example":[92],"from":[93],"official":[94],"statistics,":[95],"demonstrate":[97],"SWAG-PPM":[98,125],"workplace":[101],"injury":[102],"text":[103],"classification":[104],"task":[105],"public":[110],"dataset":[111],"published":[112],"U.S.":[115],"Occupational":[116],"Safety":[117],"and":[118],"Health":[119],"Administration":[120],"(OSHA).":[121],"We":[122],"find":[123],"exhibits":[126],"only":[127],"modest":[128],"utility":[129],"degradation":[130],"against":[131],"non-private":[133],"comparator":[134],"while":[135],"greatly":[136],"outperforming":[137],"industry":[139],"standard":[140],"DP-SGD":[141],"similar":[144],"budget.":[146]},"counts_by_year":[{"year":2026,"cited_by_count":1}],"updated_date":"2026-11-07T11:25:07.637180","created_date":"2025-10-11T00:00:00"}
