{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:SL447NJTJIU2DFYRQLIJYXFARC","short_pith_number":"pith:SL447NJT","schema_version":"1.0","canonical_sha256":"92f9cfb5334a29a1971182d09c5ca088a718101108dd2d1e7630da6c8172d053","source":{"kind":"arxiv","id":"2509.12958","version":2},"attestation_state":"computed","paper":{"title":"Forget What's Sensitive, Remember What Matters: Token-Level Differential Privacy in Memory Sculpting for Continual Learning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Bihao Zhan, Hang Yan, Jie Zhou, Junsong Li, Liang He, Qianjun Pan, Qin Chen, Shilian Chen, Wen Wu, Xingjiao Wu, Xin Li, Yutao Yang","submitted_at":"2025-09-16T11:01:59Z","abstract_excerpt":"Continual Learning (CL) models, while adept at sequential knowledge acquisition, face significant and often overlooked privacy challenges due to accumulating diverse information. Traditional privacy methods, like a uniform Differential Privacy (DP) budget, indiscriminately protect all data, leading to substantial model utility degradation and hindering CL deployment in privacy-sensitive areas. To overcome this, we propose a privacy-enhanced continual learning (PeCL) framework that forgets what's sensitive and remembers what matters. Our approach first introduces a token-level dynamic Different"},"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":"2509.12958","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2025-09-16T11:01:59Z","cross_cats_sorted":[],"title_canon_sha256":"5a673e03ca4beede9b5b02d43715e59652095c1cdf7a7e988b2018214a904aa5","abstract_canon_sha256":"ae2b023409ca97433fa47292bc60518e90eb17efb2f6e908e4a7ecb555a38e91"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-25T02:01:07.165142Z","signature_b64":"8PZqrIekcyIH6adnujzq1Bi+YbjWKP2JUkz8dR7TbXHdynBDAAB9wDpiPQNe1HaTajsRf7gAMw0RsOx9ZwfCDA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"92f9cfb5334a29a1971182d09c5ca088a718101108dd2d1e7630da6c8172d053","last_reissued_at":"2026-05-25T02:01:07.164300Z","signature_status":"signed_v1","first_computed_at":"2026-05-25T02:01:07.164300Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Forget What's Sensitive, Remember What Matters: Token-Level Differential Privacy in Memory Sculpting for Continual Learning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Bihao Zhan, Hang Yan, Jie Zhou, Junsong Li, Liang He, Qianjun Pan, Qin Chen, Shilian Chen, Wen Wu, Xingjiao Wu, Xin Li, Yutao Yang","submitted_at":"2025-09-16T11:01:59Z","abstract_excerpt":"Continual Learning (CL) models, while adept at sequential knowledge acquisition, face significant and often overlooked privacy challenges due to accumulating diverse information. Traditional privacy methods, like a uniform Differential Privacy (DP) budget, indiscriminately protect all data, leading to substantial model utility degradation and hindering CL deployment in privacy-sensitive areas. To overcome this, we propose a privacy-enhanced continual learning (PeCL) framework that forgets what's sensitive and remembers what matters. Our approach first introduces a token-level dynamic Different"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2509.12958","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/2509.12958/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":"2509.12958","created_at":"2026-05-25T02:01:07.164429+00:00"},{"alias_kind":"arxiv_version","alias_value":"2509.12958v2","created_at":"2026-05-25T02:01:07.164429+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2509.12958","created_at":"2026-05-25T02:01:07.164429+00:00"},{"alias_kind":"pith_short_12","alias_value":"SL447NJTJIU2","created_at":"2026-05-25T02:01:07.164429+00:00"},{"alias_kind":"pith_short_16","alias_value":"SL447NJTJIU2DFYR","created_at":"2026-05-25T02:01:07.164429+00:00"},{"alias_kind":"pith_short_8","alias_value":"SL447NJT","created_at":"2026-05-25T02:01:07.164429+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/SL447NJTJIU2DFYRQLIJYXFARC","json":"https://pith.science/pith/SL447NJTJIU2DFYRQLIJYXFARC.json","graph_json":"https://pith.science/api/pith-number/SL447NJTJIU2DFYRQLIJYXFARC/graph.json","events_json":"https://pith.science/api/pith-number/SL447NJTJIU2DFYRQLIJYXFARC/events.json","paper":"https://pith.science/paper/SL447NJT"},"agent_actions":{"view_html":"https://pith.science/pith/SL447NJTJIU2DFYRQLIJYXFARC","download_json":"https://pith.science/pith/SL447NJTJIU2DFYRQLIJYXFARC.json","view_paper":"https://pith.science/paper/SL447NJT","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2509.12958&json=true","fetch_graph":"https://pith.science/api/pith-number/SL447NJTJIU2DFYRQLIJYXFARC/graph.json","fetch_events":"https://pith.science/api/pith-number/SL447NJTJIU2DFYRQLIJYXFARC/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/SL447NJTJIU2DFYRQLIJYXFARC/action/timestamp_anchor","attest_storage":"https://pith.science/pith/SL447NJTJIU2DFYRQLIJYXFARC/action/storage_attestation","attest_author":"https://pith.science/pith/SL447NJTJIU2DFYRQLIJYXFARC/action/author_attestation","sign_citation":"https://pith.science/pith/SL447NJTJIU2DFYRQLIJYXFARC/action/citation_signature","submit_replication":"https://pith.science/pith/SL447NJTJIU2DFYRQLIJYXFARC/action/replication_record"}},"created_at":"2026-05-25T02:01:07.164429+00:00","updated_at":"2026-05-25T02:01:07.164429+00:00"}