{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:KT65RND5L5FMKFULJUFGALINQD","short_pith_number":"pith:KT65RND5","schema_version":"1.0","canonical_sha256":"54fdd8b47d5f4ac5168b4d0a602d0d80c03015b3d8dc047d0bf4f2bc95d55a3a","source":{"kind":"arxiv","id":"2507.09177","version":1},"attestation_state":"computed","paper":{"title":"Continual Reinforcement Learning by Planning with Online World Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","stat.ML"],"primary_cat":"cs.LG","authors_text":"Chao Du, Guoji Fu, Min Lin, Wee Sun Lee, Zichen Liu","submitted_at":"2025-07-12T07:52:31Z","abstract_excerpt":"Continual reinforcement learning (CRL) refers to a naturalistic setting where an agent needs to endlessly evolve, by trial and error, to solve multiple tasks that are presented sequentially. One of the largest obstacles to CRL is that the agent may forget how to solve previous tasks when learning a new task, known as catastrophic forgetting. In this paper, we propose to address this challenge by planning with online world models. Specifically, we learn a Follow-The-Leader shallow model online to capture the world dynamics, in which we plan using model predictive control to solve a set of tasks"},"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":"2507.09177","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-07-12T07:52:31Z","cross_cats_sorted":["cs.AI","stat.ML"],"title_canon_sha256":"3f76c68e2baac01c594e9a035d5e73ec3eebac53565ac08ece6327414feadec1","abstract_canon_sha256":"6c3b0298c0ea0200b5dcbf92eb949d981ec037fe6b1af39af6c3694ca65c0b77"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:36:05.054089Z","signature_b64":"xdb64QFYbSrx10M/7mw+PInIyYF8cUKyhBrgEmzsSx659E/EvrFpUuDXrcibJCgqaP7H5Xc2Q9t6DkKXK02+Bw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"54fdd8b47d5f4ac5168b4d0a602d0d80c03015b3d8dc047d0bf4f2bc95d55a3a","last_reissued_at":"2026-07-05T11:36:05.053691Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:36:05.053691Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Continual Reinforcement Learning by Planning with Online World Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","stat.ML"],"primary_cat":"cs.LG","authors_text":"Chao Du, Guoji Fu, Min Lin, Wee Sun Lee, Zichen Liu","submitted_at":"2025-07-12T07:52:31Z","abstract_excerpt":"Continual reinforcement learning (CRL) refers to a naturalistic setting where an agent needs to endlessly evolve, by trial and error, to solve multiple tasks that are presented sequentially. One of the largest obstacles to CRL is that the agent may forget how to solve previous tasks when learning a new task, known as catastrophic forgetting. In this paper, we propose to address this challenge by planning with online world models. Specifically, we learn a Follow-The-Leader shallow model online to capture the world dynamics, in which we plan using model predictive control to solve a set of tasks"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2507.09177","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/2507.09177/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":"2507.09177","created_at":"2026-07-05T11:36:05.053743+00:00"},{"alias_kind":"arxiv_version","alias_value":"2507.09177v1","created_at":"2026-07-05T11:36:05.053743+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2507.09177","created_at":"2026-07-05T11:36:05.053743+00:00"},{"alias_kind":"pith_short_12","alias_value":"KT65RND5L5FM","created_at":"2026-07-05T11:36:05.053743+00:00"},{"alias_kind":"pith_short_16","alias_value":"KT65RND5L5FMKFUL","created_at":"2026-07-05T11:36:05.053743+00:00"},{"alias_kind":"pith_short_8","alias_value":"KT65RND5","created_at":"2026-07-05T11:36:05.053743+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.03382","citing_title":"Local Guidance, Global Impact: Gaussian-Reshaped Trust Region Unlocks Behavior Transitions","ref_index":22,"is_internal_anchor":false},{"citing_arxiv_id":"2601.12538","citing_title":"Agentic Reasoning for Large Language Models","ref_index":186,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/KT65RND5L5FMKFULJUFGALINQD","json":"https://pith.science/pith/KT65RND5L5FMKFULJUFGALINQD.json","graph_json":"https://pith.science/api/pith-number/KT65RND5L5FMKFULJUFGALINQD/graph.json","events_json":"https://pith.science/api/pith-number/KT65RND5L5FMKFULJUFGALINQD/events.json","paper":"https://pith.science/paper/KT65RND5"},"agent_actions":{"view_html":"https://pith.science/pith/KT65RND5L5FMKFULJUFGALINQD","download_json":"https://pith.science/pith/KT65RND5L5FMKFULJUFGALINQD.json","view_paper":"https://pith.science/paper/KT65RND5","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2507.09177&json=true","fetch_graph":"https://pith.science/api/pith-number/KT65RND5L5FMKFULJUFGALINQD/graph.json","fetch_events":"https://pith.science/api/pith-number/KT65RND5L5FMKFULJUFGALINQD/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/KT65RND5L5FMKFULJUFGALINQD/action/timestamp_anchor","attest_storage":"https://pith.science/pith/KT65RND5L5FMKFULJUFGALINQD/action/storage_attestation","attest_author":"https://pith.science/pith/KT65RND5L5FMKFULJUFGALINQD/action/author_attestation","sign_citation":"https://pith.science/pith/KT65RND5L5FMKFULJUFGALINQD/action/citation_signature","submit_replication":"https://pith.science/pith/KT65RND5L5FMKFULJUFGALINQD/action/replication_record"}},"created_at":"2026-07-05T11:36:05.053743+00:00","updated_at":"2026-07-05T11:36:05.053743+00:00"}