{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:SARQXGRJXZPHILWG2E5N4XKGO3","short_pith_number":"pith:SARQXGRJ","schema_version":"1.0","canonical_sha256":"90230b9a29be5e742ec6d13ade5d4676ec3c4ca8ece1cf01a8840ce081e5679c","source":{"kind":"arxiv","id":"2311.04898","version":2},"attestation_state":"computed","paper":{"title":"Two Complementary Perspectives to Continual Learning: Ask Not Only What to Optimize, But Also How","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CV","stat.ML"],"primary_cat":"cs.LG","authors_text":"Gido M. van de Ven, Timm Hess, Tinne Tuytelaars","submitted_at":"2023-11-08T18:57:19Z","abstract_excerpt":"Recent years have seen considerable progress in the continual training of deep neural networks, predominantly thanks to approaches that add replay or regularization terms to the loss function to approximate the joint loss over all tasks so far. However, we show that even with a perfect approximation to the joint loss, these approaches still suffer from temporary but substantial forgetting when starting to train on a new task. Motivated by this 'stability gap', we propose that continual learning strategies should focus not only on the optimization objective, but also on the way this objective i"},"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":"2311.04898","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2023-11-08T18:57:19Z","cross_cats_sorted":["cs.AI","cs.CV","stat.ML"],"title_canon_sha256":"4776b7032bc71002a7f76a8d862c0582531f5e89025b580c1305e16e26f43151","abstract_canon_sha256":"27be23f717a8ccf52078d908f2a804c4cade0d00caff0d26ef8b60988e5af322"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:29:13.363768Z","signature_b64":"ElhlqrD85ves5PsRX6bahFFyRLDFeb/FuXcoOD9yCX6cxwVdi4pZEtyv+21BoyINxAd0Xplb2DwXzJatbl/uDQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"90230b9a29be5e742ec6d13ade5d4676ec3c4ca8ece1cf01a8840ce081e5679c","last_reissued_at":"2026-07-05T09:29:13.363235Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:29:13.363235Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Two Complementary Perspectives to Continual Learning: Ask Not Only What to Optimize, But Also How","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CV","stat.ML"],"primary_cat":"cs.LG","authors_text":"Gido M. van de Ven, Timm Hess, Tinne Tuytelaars","submitted_at":"2023-11-08T18:57:19Z","abstract_excerpt":"Recent years have seen considerable progress in the continual training of deep neural networks, predominantly thanks to approaches that add replay or regularization terms to the loss function to approximate the joint loss over all tasks so far. However, we show that even with a perfect approximation to the joint loss, these approaches still suffer from temporary but substantial forgetting when starting to train on a new task. Motivated by this 'stability gap', we propose that continual learning strategies should focus not only on the optimization objective, but also on the way this objective i"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2311.04898","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/2311.04898/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":"2311.04898","created_at":"2026-07-05T09:29:13.363296+00:00"},{"alias_kind":"arxiv_version","alias_value":"2311.04898v2","created_at":"2026-07-05T09:29:13.363296+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2311.04898","created_at":"2026-07-05T09:29:13.363296+00:00"},{"alias_kind":"pith_short_12","alias_value":"SARQXGRJXZPH","created_at":"2026-07-05T09:29:13.363296+00:00"},{"alias_kind":"pith_short_16","alias_value":"SARQXGRJXZPHILWG","created_at":"2026-07-05T09:29:13.363296+00:00"},{"alias_kind":"pith_short_8","alias_value":"SARQXGRJ","created_at":"2026-07-05T09:29:13.363296+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2502.01866","citing_title":"Online Curvature-Aware Replay: Leveraging $\\mathbf{2^{nd}}$ Order Information for Online Continual Learning","ref_index":26,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/SARQXGRJXZPHILWG2E5N4XKGO3","json":"https://pith.science/pith/SARQXGRJXZPHILWG2E5N4XKGO3.json","graph_json":"https://pith.science/api/pith-number/SARQXGRJXZPHILWG2E5N4XKGO3/graph.json","events_json":"https://pith.science/api/pith-number/SARQXGRJXZPHILWG2E5N4XKGO3/events.json","paper":"https://pith.science/paper/SARQXGRJ"},"agent_actions":{"view_html":"https://pith.science/pith/SARQXGRJXZPHILWG2E5N4XKGO3","download_json":"https://pith.science/pith/SARQXGRJXZPHILWG2E5N4XKGO3.json","view_paper":"https://pith.science/paper/SARQXGRJ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2311.04898&json=true","fetch_graph":"https://pith.science/api/pith-number/SARQXGRJXZPHILWG2E5N4XKGO3/graph.json","fetch_events":"https://pith.science/api/pith-number/SARQXGRJXZPHILWG2E5N4XKGO3/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/SARQXGRJXZPHILWG2E5N4XKGO3/action/timestamp_anchor","attest_storage":"https://pith.science/pith/SARQXGRJXZPHILWG2E5N4XKGO3/action/storage_attestation","attest_author":"https://pith.science/pith/SARQXGRJXZPHILWG2E5N4XKGO3/action/author_attestation","sign_citation":"https://pith.science/pith/SARQXGRJXZPHILWG2E5N4XKGO3/action/citation_signature","submit_replication":"https://pith.science/pith/SARQXGRJXZPHILWG2E5N4XKGO3/action/replication_record"}},"created_at":"2026-07-05T09:29:13.363296+00:00","updated_at":"2026-07-05T09:29:13.363296+00:00"}