{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2019:ARKW4MTFSYOUH4I3HJEFCZ4NCN","short_pith_number":"pith:ARKW4MTF","schema_version":"1.0","canonical_sha256":"04556e3265961d43f11b3a4851678d136a071a675f82e1f30f7d1310fc1bb320","source":{"kind":"arxiv","id":"1908.07643","version":2},"attestation_state":"computed","paper":{"title":"AdaCliP: Adaptive Clipping for Private SGD","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CR","stat.ML"],"primary_cat":"cs.LG","authors_text":"Ananda Theertha Suresh, Felix X. Yu, Sanjiv Kumar, Sashank J. Reddi, Venkatadheeraj Pichapati","submitted_at":"2019-08-20T23:19:21Z","abstract_excerpt":"Privacy preserving machine learning algorithms are crucial for learning models over user data to protect sensitive information. Motivated by this, differentially private stochastic gradient descent (SGD) algorithms for training machine learning models have been proposed. At each step, these algorithms modify the gradients and add noise proportional to the sensitivity of the modified gradients. Under this framework, we propose AdaCliP, a theoretically motivated differentially private SGD algorithm that provably adds less noise compared to the previous methods, by using coordinate-wise adaptive "},"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":"1908.07643","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2019-08-20T23:19:21Z","cross_cats_sorted":["cs.CR","stat.ML"],"title_canon_sha256":"c3dfb390b98f4ade84761a7722b3bf45da6ee3b75be15249de664d5dc26651c8","abstract_canon_sha256":"ed20214a69929bff24019e0e5e94ab56fd82f65a8129fd11b2b498789072e4ed"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T00:14:33.413502Z","signature_b64":"l3AdKEXjkLS/99zJuhx8FXN/aDJ9czRkKFBUjTAGXwZ62heOxGo9XoFIRhlwX5XO/x5gYmk9zmGLS2smZjkcDA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"04556e3265961d43f11b3a4851678d136a071a675f82e1f30f7d1310fc1bb320","last_reissued_at":"2026-07-05T00:14:33.412954Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T00:14:33.412954Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"AdaCliP: Adaptive Clipping for Private SGD","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CR","stat.ML"],"primary_cat":"cs.LG","authors_text":"Ananda Theertha Suresh, Felix X. Yu, Sanjiv Kumar, Sashank J. Reddi, Venkatadheeraj Pichapati","submitted_at":"2019-08-20T23:19:21Z","abstract_excerpt":"Privacy preserving machine learning algorithms are crucial for learning models over user data to protect sensitive information. Motivated by this, differentially private stochastic gradient descent (SGD) algorithms for training machine learning models have been proposed. At each step, these algorithms modify the gradients and add noise proportional to the sensitivity of the modified gradients. Under this framework, we propose AdaCliP, a theoretically motivated differentially private SGD algorithm that provably adds less noise compared to the previous methods, by using coordinate-wise adaptive "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1908.07643","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/1908.07643/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":"1908.07643","created_at":"2026-07-05T00:14:33.413015+00:00"},{"alias_kind":"arxiv_version","alias_value":"1908.07643v2","created_at":"2026-07-05T00:14:33.413015+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1908.07643","created_at":"2026-07-05T00:14:33.413015+00:00"},{"alias_kind":"pith_short_12","alias_value":"ARKW4MTFSYOU","created_at":"2026-07-05T00:14:33.413015+00:00"},{"alias_kind":"pith_short_16","alias_value":"ARKW4MTFSYOUH4I3","created_at":"2026-07-05T00:14:33.413015+00:00"},{"alias_kind":"pith_short_8","alias_value":"ARKW4MTF","created_at":"2026-07-05T00:14:33.413015+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":5,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.05435","citing_title":"DP-MacAdam: Differentially Private Mechanism with Adaptive Clipping and Adaptive Momentum","ref_index":6,"is_internal_anchor":false},{"citing_arxiv_id":"2606.04384","citing_title":"Revisiting Privacy Amplification by Subsampling in Selective Release DPSGD","ref_index":31,"is_internal_anchor":false},{"citing_arxiv_id":"2606.29293","citing_title":"Private training in quantum machine learning","ref_index":7,"is_internal_anchor":false},{"citing_arxiv_id":"2602.22611","citing_title":"Mitigating Membership Inference in Intermediate Representations with Differentially Private Training","ref_index":9,"is_internal_anchor":false},{"citing_arxiv_id":"2605.01679","citing_title":"Class-Aware Adaptive Differential Privacy in Deep Learning for Sensor-Based Fall Detection","ref_index":25,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/ARKW4MTFSYOUH4I3HJEFCZ4NCN","json":"https://pith.science/pith/ARKW4MTFSYOUH4I3HJEFCZ4NCN.json","graph_json":"https://pith.science/api/pith-number/ARKW4MTFSYOUH4I3HJEFCZ4NCN/graph.json","events_json":"https://pith.science/api/pith-number/ARKW4MTFSYOUH4I3HJEFCZ4NCN/events.json","paper":"https://pith.science/paper/ARKW4MTF"},"agent_actions":{"view_html":"https://pith.science/pith/ARKW4MTFSYOUH4I3HJEFCZ4NCN","download_json":"https://pith.science/pith/ARKW4MTFSYOUH4I3HJEFCZ4NCN.json","view_paper":"https://pith.science/paper/ARKW4MTF","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=1908.07643&json=true","fetch_graph":"https://pith.science/api/pith-number/ARKW4MTFSYOUH4I3HJEFCZ4NCN/graph.json","fetch_events":"https://pith.science/api/pith-number/ARKW4MTFSYOUH4I3HJEFCZ4NCN/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/ARKW4MTFSYOUH4I3HJEFCZ4NCN/action/timestamp_anchor","attest_storage":"https://pith.science/pith/ARKW4MTFSYOUH4I3HJEFCZ4NCN/action/storage_attestation","attest_author":"https://pith.science/pith/ARKW4MTFSYOUH4I3HJEFCZ4NCN/action/author_attestation","sign_citation":"https://pith.science/pith/ARKW4MTFSYOUH4I3HJEFCZ4NCN/action/citation_signature","submit_replication":"https://pith.science/pith/ARKW4MTFSYOUH4I3HJEFCZ4NCN/action/replication_record"}},"created_at":"2026-07-05T00:14:33.413015+00:00","updated_at":"2026-07-05T00:14:33.413015+00:00"}