{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:RYSYDD7H3PBTUH3WWEAIBM5WOX","short_pith_number":"pith:RYSYDD7H","schema_version":"1.0","canonical_sha256":"8e25818fe7dbc33a1f76b10080b3b675e1e888194b8e601bf667900de72910aa","source":{"kind":"arxiv","id":"2412.18946","version":2},"attestation_state":"computed","paper":{"title":"Constraint-Adaptive Policy Switching for Offline Safe Reinforcement Learning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Alan Fern, Aryan Deshwal, Honghao Wei, Janardhan Rao Doppa, Yassine Chemingui","submitted_at":"2024-12-25T16:42:27Z","abstract_excerpt":"Offline safe reinforcement learning (OSRL) involves learning a decision-making policy to maximize rewards from a fixed batch of training data to satisfy pre-defined safety constraints. However, adapting to varying safety constraints during deployment without retraining remains an under-explored challenge. To address this challenge, we introduce constraint-adaptive policy switching (CAPS), a wrapper framework around existing offline RL algorithms. During training, CAPS uses offline data to learn multiple policies with a shared representation that optimize different reward and cost trade-offs. D"},"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":"2412.18946","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-12-25T16:42:27Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"0c9e252e09a6595ca1401c5b89c8c737b9e1590051a07a16dc8d140187996c47","abstract_canon_sha256":"8d85e4444c4ee042afe64ea3cf6c5b77ad1a36241b1430bf25736343d02cbad5"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:10:49.976645Z","signature_b64":"vsrVdXTmcujKWkId2FKY4hHgHM92A8BrXB/Mg8oBznwP2npw22lM1FtLaVBgT40uRkf2y9j4tMPK4JeKEeelCQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"8e25818fe7dbc33a1f76b10080b3b675e1e888194b8e601bf667900de72910aa","last_reissued_at":"2026-07-05T11:10:49.976097Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:10:49.976097Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Constraint-Adaptive Policy Switching for Offline Safe Reinforcement Learning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Alan Fern, Aryan Deshwal, Honghao Wei, Janardhan Rao Doppa, Yassine Chemingui","submitted_at":"2024-12-25T16:42:27Z","abstract_excerpt":"Offline safe reinforcement learning (OSRL) involves learning a decision-making policy to maximize rewards from a fixed batch of training data to satisfy pre-defined safety constraints. However, adapting to varying safety constraints during deployment without retraining remains an under-explored challenge. To address this challenge, we introduce constraint-adaptive policy switching (CAPS), a wrapper framework around existing offline RL algorithms. During training, CAPS uses offline data to learn multiple policies with a shared representation that optimize different reward and cost trade-offs. D"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2412.18946","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/2412.18946/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":"2412.18946","created_at":"2026-07-05T11:10:49.976162+00:00"},{"alias_kind":"arxiv_version","alias_value":"2412.18946v2","created_at":"2026-07-05T11:10:49.976162+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2412.18946","created_at":"2026-07-05T11:10:49.976162+00:00"},{"alias_kind":"pith_short_12","alias_value":"RYSYDD7H3PBT","created_at":"2026-07-05T11:10:49.976162+00:00"},{"alias_kind":"pith_short_16","alias_value":"RYSYDD7H3PBTUH3W","created_at":"2026-07-05T11:10:49.976162+00:00"},{"alias_kind":"pith_short_8","alias_value":"RYSYDD7H","created_at":"2026-07-05T11:10:49.976162+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/RYSYDD7H3PBTUH3WWEAIBM5WOX","json":"https://pith.science/pith/RYSYDD7H3PBTUH3WWEAIBM5WOX.json","graph_json":"https://pith.science/api/pith-number/RYSYDD7H3PBTUH3WWEAIBM5WOX/graph.json","events_json":"https://pith.science/api/pith-number/RYSYDD7H3PBTUH3WWEAIBM5WOX/events.json","paper":"https://pith.science/paper/RYSYDD7H"},"agent_actions":{"view_html":"https://pith.science/pith/RYSYDD7H3PBTUH3WWEAIBM5WOX","download_json":"https://pith.science/pith/RYSYDD7H3PBTUH3WWEAIBM5WOX.json","view_paper":"https://pith.science/paper/RYSYDD7H","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2412.18946&json=true","fetch_graph":"https://pith.science/api/pith-number/RYSYDD7H3PBTUH3WWEAIBM5WOX/graph.json","fetch_events":"https://pith.science/api/pith-number/RYSYDD7H3PBTUH3WWEAIBM5WOX/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/RYSYDD7H3PBTUH3WWEAIBM5WOX/action/timestamp_anchor","attest_storage":"https://pith.science/pith/RYSYDD7H3PBTUH3WWEAIBM5WOX/action/storage_attestation","attest_author":"https://pith.science/pith/RYSYDD7H3PBTUH3WWEAIBM5WOX/action/author_attestation","sign_citation":"https://pith.science/pith/RYSYDD7H3PBTUH3WWEAIBM5WOX/action/citation_signature","submit_replication":"https://pith.science/pith/RYSYDD7H3PBTUH3WWEAIBM5WOX/action/replication_record"}},"created_at":"2026-07-05T11:10:49.976162+00:00","updated_at":"2026-07-05T11:10:49.976162+00:00"}