{"id":"https://openalex.org/W7222852277","doi":null,"title":"Private Retrieval Augmented Generation via Random Projection","display_name":"Private Retrieval Augmented Generation via Random Projection","publication_year":2026,"publication_date":"2026-10-15","ids":{"openalex":"https://openalex.org/W7222852277"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2610.22297","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2610.22297","pdf_url":"https://arxiv.org/pdf/2610.22297","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"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/2610.22297","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5001337840","display_name":"Dixi Yao","orcid":"https://orcid.org/0009-0005-9875-1631","observed_orcids":["https://orcid.org/0009-0005-9875-1631"]},"institutions":[{"id":"https://openalex.org/I40347166","display_name":"University of Chicago","ror":"https://ror.org/024mw5h28","country_code":"US","type":"education","lineage":["https://openalex.org/I40347166"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Yao, Dixi","raw_affiliation_strings":["University of Chicago"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of Chicago","institution_ids":["https://openalex.org/I40347166"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5111050743","display_name":"Tahseen Rabbani","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I40347166","display_name":"University of Chicago","ror":"https://ror.org/024mw5h28","country_code":"US","type":"education","lineage":["https://openalex.org/I40347166"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Rabbani, Tahseen","raw_affiliation_strings":["University of Chicago"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of Chicago","institution_ids":["https://openalex.org/I40347166"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5158938613","display_name":"Tian Li","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I40347166","display_name":"University of Chicago","ror":"https://ror.org/024mw5h28","country_code":"US","type":"education","lineage":["https://openalex.org/I40347166"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Li, Tian","raw_affiliation_strings":["University of Chicago"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of Chicago","institution_ids":["https://openalex.org/I40347166"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":["https://openalex.org/I40347166"],"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":0.986299991607666,"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":0.986299991607666,"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":0.004999999888241291,"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/T10028","display_name":"Topic Modeling","score":0.0034000000450760126,"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/retrieval-augmented-generation","display_name":"retrieval-augmented generation","score":0.9879999756813049},{"id":"https://openalex.org/keywords/random-projections","display_name":"random projections","score":0.7070000171661377},{"id":"https://openalex.org/keywords/privacy-protection","display_name":"privacy preservation","score":0.5899999737739563},{"id":"https://openalex.org/keywords/data-extraction-attacks","display_name":"data extraction attacks","score":0.8600000143051147},{"id":"https://openalex.org/keywords/differential-privacy","display_name":"differential privacy","score":0.5519999861717224},{"id":"https://openalex.org/keywords/large-language-models","display_name":"large language models","score":0.8259999752044678},{"id":"https://openalex.org/keywords/query-embedding","display_name":"query embedding","score":0.4189999997615814},{"id":"https://openalex.org/keywords/semantic-similarity","display_name":"semantic similarity","score":0.5839999914169312}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8507000207901001},{"id":"https://openalex.org/C2777036070","wikidata":"https://www.wikidata.org/wiki/Q18393452","display_name":"Random projection","level":2,"score":0.7232999801635742},{"id":"https://openalex.org/C153083717","wikidata":"https://www.wikidata.org/wiki/Q6535263","display_name":"Leverage (statistics)","level":2,"score":0.7167999744415283},{"id":"https://openalex.org/C23130292","wikidata":"https://www.wikidata.org/wiki/Q5275358","display_name":"Differential privacy","level":2,"score":0.6646999716758728},{"id":"https://openalex.org/C57493831","wikidata":"https://www.wikidata.org/wiki/Q3134666","display_name":"Projection (relational algebra)","level":2,"score":0.5529000163078308},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.43700000643730164},{"id":"https://openalex.org/C123201435","wikidata":"https://www.wikidata.org/wiki/Q456632","display_name":"Information privacy","level":2,"score":0.4104999899864197},{"id":"https://openalex.org/C3017597292","wikidata":"https://www.wikidata.org/wiki/Q25052250","display_name":"Privacy protection","level":2,"score":0.3869999945163727},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.37229999899864197},{"id":"https://openalex.org/C103278499","wikidata":"https://www.wikidata.org/wiki/Q254465","display_name":"Similarity (geometry)","level":3,"score":0.36000001430511475},{"id":"https://openalex.org/C137822555","wikidata":"https://www.wikidata.org/wiki/Q2587068","display_name":"Information sensitivity","level":2,"score":0.35260000824928284},{"id":"https://openalex.org/C23123220","wikidata":"https://www.wikidata.org/wiki/Q816826","display_name":"Information retrieval","level":1,"score":0.34869998693466187},{"id":"https://openalex.org/C130318100","wikidata":"https://www.wikidata.org/wiki/Q2268914","display_name":"Semantic similarity","level":2,"score":0.33899998664855957},{"id":"https://openalex.org/C169258074","wikidata":"https://www.wikidata.org/wiki/Q245748","display_name":"Random forest","level":2,"score":0.3305000066757202},{"id":"https://openalex.org/C12725497","wikidata":"https://www.wikidata.org/wiki/Q810247","display_name":"Baseline (sea)","level":2,"score":0.3244999945163727},{"id":"https://openalex.org/C80444323","wikidata":"https://www.wikidata.org/wiki/Q2878974","display_name":"Theoretical computer science","level":1,"score":0.3147999942302704},{"id":"https://openalex.org/C195807954","wikidata":"https://www.wikidata.org/wiki/Q1662562","display_name":"Information extraction","level":2,"score":0.29739999771118164},{"id":"https://openalex.org/C99221444","wikidata":"https://www.wikidata.org/wiki/Q1532069","display_name":"Private information retrieval","level":2,"score":0.2883000075817108},{"id":"https://openalex.org/C184337299","wikidata":"https://www.wikidata.org/wiki/Q1437428","display_name":"Semantics (computer science)","level":2,"score":0.27230000495910645},{"id":"https://openalex.org/C67186912","wikidata":"https://www.wikidata.org/wiki/Q367664","display_name":"Data modeling","level":2,"score":0.27140000462532043},{"id":"https://openalex.org/C65557600","wikidata":"https://www.wikidata.org/wiki/Q7249451","display_name":"Projection method","level":3,"score":0.2685000002384186},{"id":"https://openalex.org/C2776517306","wikidata":"https://www.wikidata.org/wiki/Q29017317","display_name":"Similarity measure","level":2,"score":0.26109999418258667},{"id":"https://openalex.org/C143271835","wikidata":"https://www.wikidata.org/wiki/Q254515","display_name":"Similitude","level":2,"score":0.26089999079704285},{"id":"https://openalex.org/C2780586882","wikidata":"https://www.wikidata.org/wiki/Q7520643","display_name":"Simple (philosophy)","level":2,"score":0.25209999084472656},{"id":"https://openalex.org/C153180895","wikidata":"https://www.wikidata.org/wiki/Q7148389","display_name":"Pattern recognition (psychology)","level":2,"score":0.25040000677108765}],"mesh":[],"locations_count":1,"locations":[{"id":"pmh:oai:arXiv.org:2610.22297","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2610.22297","pdf_url":"https://arxiv.org/pdf/2610.22297","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"}],"best_oa_location":{"id":"pmh:oai:arXiv.org:2610.22297","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2610.22297","pdf_url":"https://arxiv.org/pdf/2610.22297","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/W7222852277.pdf","grobid_xml":"https://content.openalex.org/works/W7222852277.grobid-xml"},"referenced_works_count":14,"referenced_works":["https://openalex.org/W4379538556","https://openalex.org/W4404652639","https://openalex.org/W4405173318","https://openalex.org/W4414594371","https://openalex.org/W4414881629","https://openalex.org/W4415347810","https://openalex.org/W4415428429","https://openalex.org/W4415981883","https://openalex.org/W4416189337","https://openalex.org/W4416280686","https://openalex.org/W6891858134","https://openalex.org/W7125491773","https://openalex.org/W7160273362","https://openalex.org/W7160659693"],"related_works":["https://openalex.org/W7159109849","https://openalex.org/W4392425374","https://openalex.org/W4408354281","https://openalex.org/W7169610995","https://openalex.org/W4408977202","https://openalex.org/W7130617151","https://openalex.org/W4399986892","https://openalex.org/W4416189337","https://openalex.org/W7162786102","https://openalex.org/W4405173318"],"abstract_inverted_index":{"Retrieval-Augmented":[0],"Generation":[1],"(RAG)":[2],"enhances":[3],"the":[4,28,48,116],"capabilities":[5],"of":[6],"large":[7],"language":[8],"models":[9],"(LLMs)":[10],"by":[11,23],"querying":[12],"external":[13],"structured":[14],"knowledge.":[15],"However,":[16],"it":[17],"can":[18,123],"also":[19],"introduce":[20],"privacy":[21,39,101,108],"risks":[22],"leaking":[24],"sensitive":[25],"information":[26],"from":[27],"retrieval":[29,89],"datastore.":[30],"We":[31],"propose":[32],"a":[33,126],"simple":[34],"method":[35,59],"to":[36,51,94],"preserve":[37],"datastore":[38,53],"in":[40],"RAG":[41,77,133],"systems":[42],"via":[43],"random":[44,121],"projection.":[45],"By":[46],"applying":[47],"same":[49],"projection":[50,122],"both":[52],"embeddings":[54],"and":[55,66,79,90,128],"query":[56],"embeddings,":[57],"our":[58],"provably":[60],"preserves":[61],"semantic":[62],"similarity":[63],"between":[64],"queries":[65],"retrieved":[67],"items":[68],"while":[69,104],"substantially":[70],"mitigating":[71],"data":[72],"extraction":[73],"attacks.":[74],"Across":[75],"multiple":[76],"architectures":[78],"datasets,":[80],"we":[81],"show":[82],"that":[83,120],"this":[84],"lightweight":[85],"approach":[86],"achieves":[87],"superior":[88],"generation":[91],"performance":[92],"compared":[93],"prior":[95],"methods":[96],"which":[97],"leverage":[98],"formal":[99],"differential":[100],"(DP)":[102],"guarantees,":[103],"exhibiting":[105],"comparable":[106],"empirical":[107],"under":[109],"strong":[110],"attack":[111],"models.":[112],"Our":[113],"results":[114],"for":[115,131],"first":[117],"time":[118],"suggest":[119],"serve":[124],"as":[125],"competitive":[127],"practical":[129],"baseline":[130],"privacy-preserving":[132],"systems.":[134]},"counts_by_year":[],"updated_date":"2026-11-07T11:25:07.637180","created_date":"2026-10-18T00:00:00"}