{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:5P2TZRROD6I4UG7ULZTYCA3PMT","short_pith_number":"pith:5P2TZRRO","schema_version":"1.0","canonical_sha256":"ebf53cc62e1f91ca1bf45e6781036f64e3f557d0d4ff4adfc2bd8e82d0fd1aec","source":{"kind":"arxiv","id":"2110.05223","version":1},"attestation_state":"computed","paper":{"title":"Continual Learning with Differential Privacy","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":["cs.CR"],"primary_cat":"cs.LG","authors_text":"My T. Thai, NHatHai Phan, Phung Lai, Pradnya Desai","submitted_at":"2021-10-11T12:39:55Z","abstract_excerpt":"In this paper, we focus on preserving differential privacy (DP) in continual learning (CL), in which we train ML models to learn a sequence of new tasks while memorizing previous tasks. We first introduce a notion of continual adjacent databases to bound the sensitivity of any data record participating in the training process of CL. Based upon that, we develop a new DP-preserving algorithm for CL with a data sampling strategy to quantify the privacy risk of training data in the well-known Averaged Gradient Episodic Memory (A-GEM) approach by applying a moments accountant. Our algorithm provide"},"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":"2110.05223","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.LG","submitted_at":"2021-10-11T12:39:55Z","cross_cats_sorted":["cs.CR"],"title_canon_sha256":"cbbc93371cc642277939f5de7271d7a841319706883a4f4965ce5f1e531f7b4f","abstract_canon_sha256":"3cf194d085891d9dfe1fcae7cae07e7573a1576e7cfd2399dd7ea4f80b3c0a0e"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T03:21:26.703086Z","signature_b64":"51XhBivdMQ4PgYHZmtpskFKd6cSno7bDfwBSGsAtbDultqJ82enxifrQah2+vqf3uhOejyO06NYgsJQ4EdSQDg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"ebf53cc62e1f91ca1bf45e6781036f64e3f557d0d4ff4adfc2bd8e82d0fd1aec","last_reissued_at":"2026-07-05T03:21:26.702723Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T03:21:26.702723Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Continual Learning with Differential Privacy","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":["cs.CR"],"primary_cat":"cs.LG","authors_text":"My T. Thai, NHatHai Phan, Phung Lai, Pradnya Desai","submitted_at":"2021-10-11T12:39:55Z","abstract_excerpt":"In this paper, we focus on preserving differential privacy (DP) in continual learning (CL), in which we train ML models to learn a sequence of new tasks while memorizing previous tasks. We first introduce a notion of continual adjacent databases to bound the sensitivity of any data record participating in the training process of CL. Based upon that, we develop a new DP-preserving algorithm for CL with a data sampling strategy to quantify the privacy risk of training data in the well-known Averaged Gradient Episodic Memory (A-GEM) approach by applying a moments accountant. Our algorithm provide"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2110.05223","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/2110.05223/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":"2110.05223","created_at":"2026-07-05T03:21:26.702780+00:00"},{"alias_kind":"arxiv_version","alias_value":"2110.05223v1","created_at":"2026-07-05T03:21:26.702780+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2110.05223","created_at":"2026-07-05T03:21:26.702780+00:00"},{"alias_kind":"pith_short_12","alias_value":"5P2TZRROD6I4","created_at":"2026-07-05T03:21:26.702780+00:00"},{"alias_kind":"pith_short_16","alias_value":"5P2TZRROD6I4UG7U","created_at":"2026-07-05T03:21:26.702780+00:00"},{"alias_kind":"pith_short_8","alias_value":"5P2TZRRO","created_at":"2026-07-05T03:21:26.702780+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2505.17664","citing_title":"What is the role of memorization in Continual Learning?","ref_index":17,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/5P2TZRROD6I4UG7ULZTYCA3PMT","json":"https://pith.science/pith/5P2TZRROD6I4UG7ULZTYCA3PMT.json","graph_json":"https://pith.science/api/pith-number/5P2TZRROD6I4UG7ULZTYCA3PMT/graph.json","events_json":"https://pith.science/api/pith-number/5P2TZRROD6I4UG7ULZTYCA3PMT/events.json","paper":"https://pith.science/paper/5P2TZRRO"},"agent_actions":{"view_html":"https://pith.science/pith/5P2TZRROD6I4UG7ULZTYCA3PMT","download_json":"https://pith.science/pith/5P2TZRROD6I4UG7ULZTYCA3PMT.json","view_paper":"https://pith.science/paper/5P2TZRRO","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2110.05223&json=true","fetch_graph":"https://pith.science/api/pith-number/5P2TZRROD6I4UG7ULZTYCA3PMT/graph.json","fetch_events":"https://pith.science/api/pith-number/5P2TZRROD6I4UG7ULZTYCA3PMT/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/5P2TZRROD6I4UG7ULZTYCA3PMT/action/timestamp_anchor","attest_storage":"https://pith.science/pith/5P2TZRROD6I4UG7ULZTYCA3PMT/action/storage_attestation","attest_author":"https://pith.science/pith/5P2TZRROD6I4UG7ULZTYCA3PMT/action/author_attestation","sign_citation":"https://pith.science/pith/5P2TZRROD6I4UG7ULZTYCA3PMT/action/citation_signature","submit_replication":"https://pith.science/pith/5P2TZRROD6I4UG7ULZTYCA3PMT/action/replication_record"}},"created_at":"2026-07-05T03:21:26.702780+00:00","updated_at":"2026-07-05T03:21:26.702780+00:00"}