{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2022:F57V2647MQAXC64ZUX477WR5EN","short_pith_number":"pith:F57V2647","schema_version":"1.0","canonical_sha256":"2f7f5d7b9f6401717b99a5f9ffda3d2373cf2339e42e8bdd581432c874968f2f","source":{"kind":"arxiv","id":"2211.06530","version":2},"attestation_state":"computed","paper":{"title":"Multi-Epoch Matrix Factorization Mechanisms for Private Machine Learning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CR","cs.DS","stat.ML"],"primary_cat":"cs.LG","authors_text":"Abhradeep Thakurta, Christopher A. Choquette-Choo, H. Brendan McMahan, Keith Rush","submitted_at":"2022-11-12T00:41:11Z","abstract_excerpt":"We introduce new differentially private (DP) mechanisms for gradient-based machine learning (ML) with multiple passes (epochs) over a dataset, substantially improving the achievable privacy-utility-computation tradeoffs. We formalize the problem of DP mechanisms for adaptive streams with multiple participations and introduce a non-trivial extension of online matrix factorization DP mechanisms to our setting. This includes establishing the necessary theory for sensitivity calculations and efficient computation of optimal matrices. For some applications like $>\\!\\! 10,000$ SGD steps, applying th"},"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":"2211.06530","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2022-11-12T00:41:11Z","cross_cats_sorted":["cs.CR","cs.DS","stat.ML"],"title_canon_sha256":"2af91a1b61e43fe0707212a28513916fa9e5740d5f3ae0dc2c5fe4f9eb34cade","abstract_canon_sha256":"7ab490e1a7ea5121346713e60a0d5ca0a8683be4ff6e02b02092993b5f8956b7"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T06:19:00.463429Z","signature_b64":"s2VtdTDCvjduszddfd6NsmkUmRmfHNKI3XZZCItcmrKjCwTj2/rpZK69yyt5t7pIeb8QgNqZ86wo10EibehWBA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"2f7f5d7b9f6401717b99a5f9ffda3d2373cf2339e42e8bdd581432c874968f2f","last_reissued_at":"2026-07-05T06:19:00.462980Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T06:19:00.462980Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Multi-Epoch Matrix Factorization Mechanisms for Private Machine Learning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CR","cs.DS","stat.ML"],"primary_cat":"cs.LG","authors_text":"Abhradeep Thakurta, Christopher A. Choquette-Choo, H. Brendan McMahan, Keith Rush","submitted_at":"2022-11-12T00:41:11Z","abstract_excerpt":"We introduce new differentially private (DP) mechanisms for gradient-based machine learning (ML) with multiple passes (epochs) over a dataset, substantially improving the achievable privacy-utility-computation tradeoffs. We formalize the problem of DP mechanisms for adaptive streams with multiple participations and introduce a non-trivial extension of online matrix factorization DP mechanisms to our setting. This includes establishing the necessary theory for sensitivity calculations and efficient computation of optimal matrices. For some applications like $>\\!\\! 10,000$ SGD steps, applying th"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2211.06530","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/2211.06530/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":"2211.06530","created_at":"2026-07-05T06:19:00.463035+00:00"},{"alias_kind":"arxiv_version","alias_value":"2211.06530v2","created_at":"2026-07-05T06:19:00.463035+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2211.06530","created_at":"2026-07-05T06:19:00.463035+00:00"},{"alias_kind":"pith_short_12","alias_value":"F57V2647MQAX","created_at":"2026-07-05T06:19:00.463035+00:00"},{"alias_kind":"pith_short_16","alias_value":"F57V2647MQAXC64Z","created_at":"2026-07-05T06:19:00.463035+00:00"},{"alias_kind":"pith_short_8","alias_value":"F57V2647","created_at":"2026-07-05T06:19:00.463035+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.24879","citing_title":"Efficient DP-SGD for LLMs with Randomized Clipping","ref_index":17,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/F57V2647MQAXC64ZUX477WR5EN","json":"https://pith.science/pith/F57V2647MQAXC64ZUX477WR5EN.json","graph_json":"https://pith.science/api/pith-number/F57V2647MQAXC64ZUX477WR5EN/graph.json","events_json":"https://pith.science/api/pith-number/F57V2647MQAXC64ZUX477WR5EN/events.json","paper":"https://pith.science/paper/F57V2647"},"agent_actions":{"view_html":"https://pith.science/pith/F57V2647MQAXC64ZUX477WR5EN","download_json":"https://pith.science/pith/F57V2647MQAXC64ZUX477WR5EN.json","view_paper":"https://pith.science/paper/F57V2647","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2211.06530&json=true","fetch_graph":"https://pith.science/api/pith-number/F57V2647MQAXC64ZUX477WR5EN/graph.json","fetch_events":"https://pith.science/api/pith-number/F57V2647MQAXC64ZUX477WR5EN/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/F57V2647MQAXC64ZUX477WR5EN/action/timestamp_anchor","attest_storage":"https://pith.science/pith/F57V2647MQAXC64ZUX477WR5EN/action/storage_attestation","attest_author":"https://pith.science/pith/F57V2647MQAXC64ZUX477WR5EN/action/author_attestation","sign_citation":"https://pith.science/pith/F57V2647MQAXC64ZUX477WR5EN/action/citation_signature","submit_replication":"https://pith.science/pith/F57V2647MQAXC64ZUX477WR5EN/action/replication_record"}},"created_at":"2026-07-05T06:19:00.463035+00:00","updated_at":"2026-07-05T06:19:00.463035+00:00"}