{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:POMY6D5J5HG7WRP5AIZALW2IUV","short_pith_number":"pith:POMY6D5J","schema_version":"1.0","canonical_sha256":"7b998f0fa9e9cdfb45fd023205db48a56f54d0728362afce2c2795547e4c3034","source":{"kind":"arxiv","id":"2412.16802","version":2},"attestation_state":"computed","paper":{"title":"Balls-and-Bins Sampling for DP-SGD","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CR","cs.DS","stat.ML"],"primary_cat":"cs.LG","authors_text":"Amer Sinha, Badih Ghazi, Charlie Harrison, Chiyuan Zhang, Ethan Leeman, Lynn Chua, Pasin Manurangsi, Pritish Kamath, Ravi Kumar","submitted_at":"2024-12-21T23:09:14Z","abstract_excerpt":"We introduce the Balls-and-Bins sampling for differentially private (DP) optimization methods such as DP-SGD. While it has been common practice to use some form of shuffling in DP-SGD implementations, privacy accounting algorithms have typically assumed that Poisson subsampling is used instead. Recent work by Chua et al. (ICML 2024), however, pointed out that shuffling based DP-SGD can have a much larger privacy cost in practical regimes of parameters. In this work we show that the Balls-and-Bins sampling achieves the \"best-of-both\" samplers, namely, the implementation of Balls-and-Bins sampli"},"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":"2412.16802","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-12-21T23:09:14Z","cross_cats_sorted":["cs.CR","cs.DS","stat.ML"],"title_canon_sha256":"f53341c1fd4951161924c5f7d6a18113a5ab7497f729e96705ab4bd99aa05247","abstract_canon_sha256":"29b30064bac8962dc79711e766f5987c24ef0e460e2fc6b2879cf8dbcf9b46c3"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:42:14.344601Z","signature_b64":"9W5imrJvdh13T6aXT1D+ZTOm2RWqfXOZyuMOHNimeBzjHjyeo0dheHYq6UTKrTaUoT0LkVdbt8eGjCsKa9TcCg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"7b998f0fa9e9cdfb45fd023205db48a56f54d0728362afce2c2795547e4c3034","last_reissued_at":"2026-07-05T10:42:14.344071Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:42:14.344071Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Balls-and-Bins Sampling for DP-SGD","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CR","cs.DS","stat.ML"],"primary_cat":"cs.LG","authors_text":"Amer Sinha, Badih Ghazi, Charlie Harrison, Chiyuan Zhang, Ethan Leeman, Lynn Chua, Pasin Manurangsi, Pritish Kamath, Ravi Kumar","submitted_at":"2024-12-21T23:09:14Z","abstract_excerpt":"We introduce the Balls-and-Bins sampling for differentially private (DP) optimization methods such as DP-SGD. While it has been common practice to use some form of shuffling in DP-SGD implementations, privacy accounting algorithms have typically assumed that Poisson subsampling is used instead. Recent work by Chua et al. (ICML 2024), however, pointed out that shuffling based DP-SGD can have a much larger privacy cost in practical regimes of parameters. In this work we show that the Balls-and-Bins sampling achieves the \"best-of-both\" samplers, namely, the implementation of Balls-and-Bins sampli"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2412.16802","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/2412.16802/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":"2412.16802","created_at":"2026-07-05T10:42:14.344135+00:00"},{"alias_kind":"arxiv_version","alias_value":"2412.16802v2","created_at":"2026-07-05T10:42:14.344135+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2412.16802","created_at":"2026-07-05T10:42:14.344135+00:00"},{"alias_kind":"pith_short_12","alias_value":"POMY6D5J5HG7","created_at":"2026-07-05T10:42:14.344135+00:00"},{"alias_kind":"pith_short_16","alias_value":"POMY6D5J5HG7WRP5","created_at":"2026-07-05T10:42:14.344135+00:00"},{"alias_kind":"pith_short_8","alias_value":"POMY6D5J","created_at":"2026-07-05T10:42:14.344135+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.07072","citing_title":"Less Random, More Private: What is the Optimal Subsampling Scheme for DP-SGD?","ref_index":4,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/POMY6D5J5HG7WRP5AIZALW2IUV","json":"https://pith.science/pith/POMY6D5J5HG7WRP5AIZALW2IUV.json","graph_json":"https://pith.science/api/pith-number/POMY6D5J5HG7WRP5AIZALW2IUV/graph.json","events_json":"https://pith.science/api/pith-number/POMY6D5J5HG7WRP5AIZALW2IUV/events.json","paper":"https://pith.science/paper/POMY6D5J"},"agent_actions":{"view_html":"https://pith.science/pith/POMY6D5J5HG7WRP5AIZALW2IUV","download_json":"https://pith.science/pith/POMY6D5J5HG7WRP5AIZALW2IUV.json","view_paper":"https://pith.science/paper/POMY6D5J","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2412.16802&json=true","fetch_graph":"https://pith.science/api/pith-number/POMY6D5J5HG7WRP5AIZALW2IUV/graph.json","fetch_events":"https://pith.science/api/pith-number/POMY6D5J5HG7WRP5AIZALW2IUV/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/POMY6D5J5HG7WRP5AIZALW2IUV/action/timestamp_anchor","attest_storage":"https://pith.science/pith/POMY6D5J5HG7WRP5AIZALW2IUV/action/storage_attestation","attest_author":"https://pith.science/pith/POMY6D5J5HG7WRP5AIZALW2IUV/action/author_attestation","sign_citation":"https://pith.science/pith/POMY6D5J5HG7WRP5AIZALW2IUV/action/citation_signature","submit_replication":"https://pith.science/pith/POMY6D5J5HG7WRP5AIZALW2IUV/action/replication_record"}},"created_at":"2026-07-05T10:42:14.344135+00:00","updated_at":"2026-07-05T10:42:14.344135+00:00"}