{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:MW6D4RKRMGEJE4VQBUDLJAUP2I","short_pith_number":"pith:MW6D4RKR","schema_version":"1.0","canonical_sha256":"65bc3e455161889272b00d06b4828fd2305ad011db86a00b8304e84a7fa132b3","source":{"kind":"arxiv","id":"2605.21938","version":1},"attestation_state":"computed","paper":{"title":"Optimal Guarantees for Auditing R\\'enyi Differentially Private Machine Learning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CR","cs.IT","math.IT"],"primary_cat":"cs.LG","authors_text":"Benjamin D. Kim, Daniel Alabi, Lav R. Varshney","submitted_at":"2026-05-21T03:18:09Z","abstract_excerpt":"We study black-box auditing for machine learning algorithms that claim R \\ 'enyi differential privacy (RDP) guarantees. We introduce an auditing framework, based on hypothesis testing, that directly estimates R\\'enyi divergence between neighboring executions using the Donsker-Varadhan (DV) variational estimator. Our analysis yields explicit and non-asymptotic confidence intervals for RDP auditing via class-restricted DV estimators, separating statistical estimation error from algorithmic privacy leakage. We prove matching minimax lower bounds showing that, up to logarithmic factors, our sample"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2605.21938","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-21T03:18:09Z","cross_cats_sorted":["cs.CR","cs.IT","math.IT"],"title_canon_sha256":"c0c45a62b1b37113e0855a325ad23a14c365712f1bd1407dd5e4a7c0bae0e1bc","abstract_canon_sha256":"c8cee2db7318d58502b8c734a5ad34a643d348916f21d80c485bacb1891d8965"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-22T01:04:15.875938Z","signature_b64":"x8iGl38Vt/l0hbh1bvJEMenKgWFvvAF3PGU+vbT4rAF+AcDzaXFTnLqBriCIGaVKaK5c6lFXn+Toz7R52QPmAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"65bc3e455161889272b00d06b4828fd2305ad011db86a00b8304e84a7fa132b3","last_reissued_at":"2026-05-22T01:04:15.875102Z","signature_status":"signed_v1","first_computed_at":"2026-05-22T01:04:15.875102Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Optimal Guarantees for Auditing R\\'enyi Differentially Private Machine Learning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CR","cs.IT","math.IT"],"primary_cat":"cs.LG","authors_text":"Benjamin D. Kim, Daniel Alabi, Lav R. Varshney","submitted_at":"2026-05-21T03:18:09Z","abstract_excerpt":"We study black-box auditing for machine learning algorithms that claim R \\ 'enyi differential privacy (RDP) guarantees. We introduce an auditing framework, based on hypothesis testing, that directly estimates R\\'enyi divergence between neighboring executions using the Donsker-Varadhan (DV) variational estimator. Our analysis yields explicit and non-asymptotic confidence intervals for RDP auditing via class-restricted DV estimators, separating statistical estimation error from algorithmic privacy leakage. We prove matching minimax lower bounds showing that, up to logarithmic factors, our sample"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.21938","kind":"arxiv","version":1},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2605.21938/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"aliases":[{"alias_kind":"arxiv","alias_value":"2605.21938","created_at":"2026-05-22T01:04:15.875242+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.21938v1","created_at":"2026-05-22T01:04:15.875242+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.21938","created_at":"2026-05-22T01:04:15.875242+00:00"},{"alias_kind":"pith_short_12","alias_value":"MW6D4RKRMGEJ","created_at":"2026-05-22T01:04:15.875242+00:00"},{"alias_kind":"pith_short_16","alias_value":"MW6D4RKRMGEJE4VQ","created_at":"2026-05-22T01:04:15.875242+00:00"},{"alias_kind":"pith_short_8","alias_value":"MW6D4RKR","created_at":"2026-05-22T01:04:15.875242+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/MW6D4RKRMGEJE4VQBUDLJAUP2I","json":"https://pith.science/pith/MW6D4RKRMGEJE4VQBUDLJAUP2I.json","graph_json":"https://pith.science/api/pith-number/MW6D4RKRMGEJE4VQBUDLJAUP2I/graph.json","events_json":"https://pith.science/api/pith-number/MW6D4RKRMGEJE4VQBUDLJAUP2I/events.json","paper":"https://pith.science/paper/MW6D4RKR"},"agent_actions":{"view_html":"https://pith.science/pith/MW6D4RKRMGEJE4VQBUDLJAUP2I","download_json":"https://pith.science/pith/MW6D4RKRMGEJE4VQBUDLJAUP2I.json","view_paper":"https://pith.science/paper/MW6D4RKR","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.21938&json=true","fetch_graph":"https://pith.science/api/pith-number/MW6D4RKRMGEJE4VQBUDLJAUP2I/graph.json","fetch_events":"https://pith.science/api/pith-number/MW6D4RKRMGEJE4VQBUDLJAUP2I/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/MW6D4RKRMGEJE4VQBUDLJAUP2I/action/timestamp_anchor","attest_storage":"https://pith.science/pith/MW6D4RKRMGEJE4VQBUDLJAUP2I/action/storage_attestation","attest_author":"https://pith.science/pith/MW6D4RKRMGEJE4VQBUDLJAUP2I/action/author_attestation","sign_citation":"https://pith.science/pith/MW6D4RKRMGEJE4VQBUDLJAUP2I/action/citation_signature","submit_replication":"https://pith.science/pith/MW6D4RKRMGEJE4VQBUDLJAUP2I/action/replication_record"}},"created_at":"2026-05-22T01:04:15.875242+00:00","updated_at":"2026-05-22T01:04:15.875242+00:00"}