{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:KDXIWPXUE64RE2XA62FBR57QBG","short_pith_number":"pith:KDXIWPXU","schema_version":"1.0","canonical_sha256":"50ee8b3ef427b9126ae0f68a18f7f009876ce8e7ff8c216fd30a0c4b11dace1d","source":{"kind":"arxiv","id":"2602.17284","version":2},"attestation_state":"computed","paper":{"title":"Efficient privacy loss accounting for subsampling and random allocation","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Moshe Shenfeld, Vitaly Feldman","submitted_at":"2026-02-19T11:44:25Z","abstract_excerpt":"We consider the privacy amplification properties of a sampling scheme in which a user's data isused in $k$ steps chosen randomly and uniformly from a sequence (or set) of $t$ steps. This sampling scheme has been recently applied in the context of differentially private optimization (Chua et al., 2024a; Choquette-Choo et al., 2025) and communication-efficient high-dimensional private aggregation (Asi et al., 2026), where it was shown to have utility advantages over the standard Poisson sampling. Theoretical analyses of this sampling scheme (Feldman & Shenfeld, 2025; Dong et al., 2025) lead to b"},"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":"2602.17284","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.LG","submitted_at":"2026-02-19T11:44:25Z","cross_cats_sorted":[],"title_canon_sha256":"e623951d02b6073ef04a41bb285bed36c7daed4b2c187110b6bbc731e71c6870","abstract_canon_sha256":"89eb8216439b8a235aebad472b01e568153aeff3ac20fe7502b9a40a1ce90567"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-07T02:19:47.264492Z","signature_b64":"Mz418pt0FVxcIP3TA1XP4kRhn+LqwmsuaGmOuiR5FL6tB7K3LCQHMoxkudFqb/mB7WBx8Zga6vANvV0JaW/aBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"50ee8b3ef427b9126ae0f68a18f7f009876ce8e7ff8c216fd30a0c4b11dace1d","last_reissued_at":"2026-07-07T02:19:47.263494Z","signature_status":"signed_v1","first_computed_at":"2026-07-07T02:19:47.263494Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Efficient privacy loss accounting for subsampling and random allocation","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Moshe Shenfeld, Vitaly Feldman","submitted_at":"2026-02-19T11:44:25Z","abstract_excerpt":"We consider the privacy amplification properties of a sampling scheme in which a user's data isused in $k$ steps chosen randomly and uniformly from a sequence (or set) of $t$ steps. This sampling scheme has been recently applied in the context of differentially private optimization (Chua et al., 2024a; Choquette-Choo et al., 2025) and communication-efficient high-dimensional private aggregation (Asi et al., 2026), where it was shown to have utility advantages over the standard Poisson sampling. Theoretical analyses of this sampling scheme (Feldman & Shenfeld, 2025; Dong et al., 2025) lead to b"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2602.17284","kind":"arxiv","version":2},"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/2602.17284/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":"2602.17284","created_at":"2026-07-07T02:19:47.263633+00:00"},{"alias_kind":"arxiv_version","alias_value":"2602.17284v2","created_at":"2026-07-07T02:19:47.263633+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2602.17284","created_at":"2026-07-07T02:19:47.263633+00:00"},{"alias_kind":"pith_short_12","alias_value":"KDXIWPXUE64R","created_at":"2026-07-07T02:19:47.263633+00:00"},{"alias_kind":"pith_short_16","alias_value":"KDXIWPXUE64RE2XA","created_at":"2026-07-07T02:19:47.263633+00:00"},{"alias_kind":"pith_short_8","alias_value":"KDXIWPXU","created_at":"2026-07-07T02:19:47.263633+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":2,"sample":[{"citing_arxiv_id":"2605.21780","citing_title":"Provable Robustness against Backdoor Attacks via the Primal-Dual Perspective on Differential Privacy","ref_index":21,"is_internal_anchor":true},{"citing_arxiv_id":"2605.07072","citing_title":"Less Random, More Private: What is the Optimal Subsampling Scheme for DP-SGD?","ref_index":13,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/KDXIWPXUE64RE2XA62FBR57QBG","json":"https://pith.science/pith/KDXIWPXUE64RE2XA62FBR57QBG.json","graph_json":"https://pith.science/api/pith-number/KDXIWPXUE64RE2XA62FBR57QBG/graph.json","events_json":"https://pith.science/api/pith-number/KDXIWPXUE64RE2XA62FBR57QBG/events.json","paper":"https://pith.science/paper/KDXIWPXU"},"agent_actions":{"view_html":"https://pith.science/pith/KDXIWPXUE64RE2XA62FBR57QBG","download_json":"https://pith.science/pith/KDXIWPXUE64RE2XA62FBR57QBG.json","view_paper":"https://pith.science/paper/KDXIWPXU","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2602.17284&json=true","fetch_graph":"https://pith.science/api/pith-number/KDXIWPXUE64RE2XA62FBR57QBG/graph.json","fetch_events":"https://pith.science/api/pith-number/KDXIWPXUE64RE2XA62FBR57QBG/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/KDXIWPXUE64RE2XA62FBR57QBG/action/timestamp_anchor","attest_storage":"https://pith.science/pith/KDXIWPXUE64RE2XA62FBR57QBG/action/storage_attestation","attest_author":"https://pith.science/pith/KDXIWPXUE64RE2XA62FBR57QBG/action/author_attestation","sign_citation":"https://pith.science/pith/KDXIWPXUE64RE2XA62FBR57QBG/action/citation_signature","submit_replication":"https://pith.science/pith/KDXIWPXUE64RE2XA62FBR57QBG/action/replication_record"}},"created_at":"2026-07-07T02:19:47.263633+00:00","updated_at":"2026-07-07T02:19:47.263633+00:00"}