{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:35EVBZPFGVY56T63MRRU4IWJZO","short_pith_number":"pith:35EVBZPF","schema_version":"1.0","canonical_sha256":"df4950e5e53571df4fdb64634e22c9cbbc07d7731dde9ecc1e8ac07f7d1fbe11","source":{"kind":"arxiv","id":"2606.12733","version":1},"attestation_state":"computed","paper":{"title":"Let's Ask Gauss: Improved One-Run Privacy Auditing","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Adya Agrawal, Jaspal Singh, Malik Magdon-Ismail, Vassilis Zikas, Yu Wei","submitted_at":"2026-06-10T22:44:47Z","abstract_excerpt":"Privacy auditing provides an important safeguard by estimating the actual information leaked by a model, thus ensuring that theoretical privacy guarantees hold in practice. We study empirical privacy auditing for differentially private (DP) machine learning, focusing on efficient one-run methods for mechanisms such as DP-SGD. Prior one-run approaches threshold training examples or \"canaries\" into binary membership guesses, which discards useful information. We show that, in the white-box DP-SGD setting, canary-aligned signals naturally form a sequence of random variables whose normalized sum i"},"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":"2606.12733","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-06-10T22:44:47Z","cross_cats_sorted":[],"title_canon_sha256":"306718beabe7faf8f0e86eadad84300a3a97402f7c4bb169f3bcf30aa668d9a9","abstract_canon_sha256":"5602de6ea857c924c78dd6227ba5d073d4ac8403244e03781e45dc40b24437ac"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-12T01:08:48.002051Z","signature_b64":"guLhZxl2r0tPryCul85j8CipZMp2H57VUqTYrQ+eNQZQbSqXJXdTgaBPRerRTS7JQEY5urhvzFyzQwHYoQAoDA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"df4950e5e53571df4fdb64634e22c9cbbc07d7731dde9ecc1e8ac07f7d1fbe11","last_reissued_at":"2026-06-12T01:08:48.001547Z","signature_status":"signed_v1","first_computed_at":"2026-06-12T01:08:48.001547Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Let's Ask Gauss: Improved One-Run Privacy Auditing","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Adya Agrawal, Jaspal Singh, Malik Magdon-Ismail, Vassilis Zikas, Yu Wei","submitted_at":"2026-06-10T22:44:47Z","abstract_excerpt":"Privacy auditing provides an important safeguard by estimating the actual information leaked by a model, thus ensuring that theoretical privacy guarantees hold in practice. We study empirical privacy auditing for differentially private (DP) machine learning, focusing on efficient one-run methods for mechanisms such as DP-SGD. Prior one-run approaches threshold training examples or \"canaries\" into binary membership guesses, which discards useful information. We show that, in the white-box DP-SGD setting, canary-aligned signals naturally form a sequence of random variables whose normalized sum i"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.12733","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/2606.12733/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":"2606.12733","created_at":"2026-06-12T01:08:48.001647+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.12733v1","created_at":"2026-06-12T01:08:48.001647+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.12733","created_at":"2026-06-12T01:08:48.001647+00:00"},{"alias_kind":"pith_short_12","alias_value":"35EVBZPFGVY5","created_at":"2026-06-12T01:08:48.001647+00:00"},{"alias_kind":"pith_short_16","alias_value":"35EVBZPFGVY56T63","created_at":"2026-06-12T01:08:48.001647+00:00"},{"alias_kind":"pith_short_8","alias_value":"35EVBZPF","created_at":"2026-06-12T01:08:48.001647+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/35EVBZPFGVY56T63MRRU4IWJZO","json":"https://pith.science/pith/35EVBZPFGVY56T63MRRU4IWJZO.json","graph_json":"https://pith.science/api/pith-number/35EVBZPFGVY56T63MRRU4IWJZO/graph.json","events_json":"https://pith.science/api/pith-number/35EVBZPFGVY56T63MRRU4IWJZO/events.json","paper":"https://pith.science/paper/35EVBZPF"},"agent_actions":{"view_html":"https://pith.science/pith/35EVBZPFGVY56T63MRRU4IWJZO","download_json":"https://pith.science/pith/35EVBZPFGVY56T63MRRU4IWJZO.json","view_paper":"https://pith.science/paper/35EVBZPF","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.12733&json=true","fetch_graph":"https://pith.science/api/pith-number/35EVBZPFGVY56T63MRRU4IWJZO/graph.json","fetch_events":"https://pith.science/api/pith-number/35EVBZPFGVY56T63MRRU4IWJZO/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/35EVBZPFGVY56T63MRRU4IWJZO/action/timestamp_anchor","attest_storage":"https://pith.science/pith/35EVBZPFGVY56T63MRRU4IWJZO/action/storage_attestation","attest_author":"https://pith.science/pith/35EVBZPFGVY56T63MRRU4IWJZO/action/author_attestation","sign_citation":"https://pith.science/pith/35EVBZPFGVY56T63MRRU4IWJZO/action/citation_signature","submit_replication":"https://pith.science/pith/35EVBZPFGVY56T63MRRU4IWJZO/action/replication_record"}},"created_at":"2026-06-12T01:08:48.001647+00:00","updated_at":"2026-06-12T01:08:48.001647+00:00"}