{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:73LN5XCCXZOBEGWCCHTJ4QQX4W","short_pith_number":"pith:73LN5XCC","schema_version":"1.0","canonical_sha256":"fed6dedc42be5c121ac211e69e4217e5836d53b55e8840807d7b9842a499b1b1","source":{"kind":"arxiv","id":"2603.01761","version":2},"attestation_state":"computed","paper":{"title":"Position: Modular Memory is the Key to Continual Learning Agents","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Ameya Prabhu, Andrew D. Bagdanov, Antonio Carta, Barbara Hammer, Christopher Kanan, Darshan Patil, Dhireesha Kudithipudi, Elisa Ricci, Gido M. van de Ven, Jonghyun Choi, Joost Van De Weijer, Jorge Mendez-Mendez, Laurent Charlin, Liyuan Wang, Lucas Caccia, Malte Schwerin, Martin Mundt, Rahaf Aljundi, Timm Hess, Tinne Tuytelaars, Tyler L. Hayes, Vaggelis Dorovatas, Vincenzo Lomonaco, Xialei Liu","submitted_at":"2026-03-02T11:40:05Z","abstract_excerpt":"Foundation models have transformed machine learning through large-scale pretraining and increased test-time compute. Despite surpassing human performance in several domains, these models remain fundamentally limited in continuous operation, experience accumulation, and personalization, capabilities that are central to adaptive intelligence. While continual learning research has long targeted these goals, its historical focus on in-weight learning (IWL), i.e., updating a single model's parameters to absorb new knowledge, has rendered catastrophic forgetting a persistent challenge. Our position "},"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":"2603.01761","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-03-02T11:40:05Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"74318b32e943a525611d33373cdcdd3c706ce1a291abfa2ca3037e1edc24eacf","abstract_canon_sha256":"46cfc8480bf3d1a7f1f588deda90901db9164f23d27a23a64354ec2059f7d170"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-19T16:09:55.723469Z","signature_b64":"w81F/Xc6dqWzQRzqqyyGrMfzTvwRBM0pN36cffpXH/8EM/5nanSdgiNVceEn3WStp0i59Cd1va+vlJ0k5DAaAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"fed6dedc42be5c121ac211e69e4217e5836d53b55e8840807d7b9842a499b1b1","last_reissued_at":"2026-06-19T16:09:55.723051Z","signature_status":"signed_v1","first_computed_at":"2026-06-19T16:09:55.723051Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Position: Modular Memory is the Key to Continual Learning Agents","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Ameya Prabhu, Andrew D. Bagdanov, Antonio Carta, Barbara Hammer, Christopher Kanan, Darshan Patil, Dhireesha Kudithipudi, Elisa Ricci, Gido M. van de Ven, Jonghyun Choi, Joost Van De Weijer, Jorge Mendez-Mendez, Laurent Charlin, Liyuan Wang, Lucas Caccia, Malte Schwerin, Martin Mundt, Rahaf Aljundi, Timm Hess, Tinne Tuytelaars, Tyler L. Hayes, Vaggelis Dorovatas, Vincenzo Lomonaco, Xialei Liu","submitted_at":"2026-03-02T11:40:05Z","abstract_excerpt":"Foundation models have transformed machine learning through large-scale pretraining and increased test-time compute. Despite surpassing human performance in several domains, these models remain fundamentally limited in continuous operation, experience accumulation, and personalization, capabilities that are central to adaptive intelligence. While continual learning research has long targeted these goals, its historical focus on in-weight learning (IWL), i.e., updating a single model's parameters to absorb new knowledge, has rendered catastrophic forgetting a persistent challenge. Our position "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2603.01761","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/2603.01761/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":"2603.01761","created_at":"2026-06-19T16:09:55.723111+00:00"},{"alias_kind":"arxiv_version","alias_value":"2603.01761v2","created_at":"2026-06-19T16:09:55.723111+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2603.01761","created_at":"2026-06-19T16:09:55.723111+00:00"},{"alias_kind":"pith_short_12","alias_value":"73LN5XCCXZOB","created_at":"2026-06-19T16:09:55.723111+00:00"},{"alias_kind":"pith_short_16","alias_value":"73LN5XCCXZOBEGWC","created_at":"2026-06-19T16:09:55.723111+00:00"},{"alias_kind":"pith_short_8","alias_value":"73LN5XCC","created_at":"2026-06-19T16:09:55.723111+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/73LN5XCCXZOBEGWCCHTJ4QQX4W","json":"https://pith.science/pith/73LN5XCCXZOBEGWCCHTJ4QQX4W.json","graph_json":"https://pith.science/api/pith-number/73LN5XCCXZOBEGWCCHTJ4QQX4W/graph.json","events_json":"https://pith.science/api/pith-number/73LN5XCCXZOBEGWCCHTJ4QQX4W/events.json","paper":"https://pith.science/paper/73LN5XCC"},"agent_actions":{"view_html":"https://pith.science/pith/73LN5XCCXZOBEGWCCHTJ4QQX4W","download_json":"https://pith.science/pith/73LN5XCCXZOBEGWCCHTJ4QQX4W.json","view_paper":"https://pith.science/paper/73LN5XCC","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2603.01761&json=true","fetch_graph":"https://pith.science/api/pith-number/73LN5XCCXZOBEGWCCHTJ4QQX4W/graph.json","fetch_events":"https://pith.science/api/pith-number/73LN5XCCXZOBEGWCCHTJ4QQX4W/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/73LN5XCCXZOBEGWCCHTJ4QQX4W/action/timestamp_anchor","attest_storage":"https://pith.science/pith/73LN5XCCXZOBEGWCCHTJ4QQX4W/action/storage_attestation","attest_author":"https://pith.science/pith/73LN5XCCXZOBEGWCCHTJ4QQX4W/action/author_attestation","sign_citation":"https://pith.science/pith/73LN5XCCXZOBEGWCCHTJ4QQX4W/action/citation_signature","submit_replication":"https://pith.science/pith/73LN5XCCXZOBEGWCCHTJ4QQX4W/action/replication_record"}},"created_at":"2026-06-19T16:09:55.723111+00:00","updated_at":"2026-06-19T16:09:55.723111+00:00"}