{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:EPXL6C77P2AI5ZHQ2R72EW4CX3","short_pith_number":"pith:EPXL6C77","schema_version":"1.0","canonical_sha256":"23eebf0bff7e808ee4f0d47fa25b82bed83bf33d94a0512690f3baa71ef508d5","source":{"kind":"arxiv","id":"2402.15650","version":3},"attestation_state":"computed","paper":{"title":"Uniformly Safe RL with Objective Suppression for Multi-Constraint Safety-Critical Applications","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Aleksandr Petiushko, Animesh Garg, Jonathan Booher, Khashayar Rohanimanesh, Wei Liu, Zihan Zhou","submitted_at":"2024-02-23T23:22:06Z","abstract_excerpt":"Safe reinforcement learning tasks are a challenging domain despite being very common in the real world. The widely adopted CMDP model constrains the risks in expectation, which makes room for dangerous behaviors in long-tail states. In safety-critical domains, such behaviors could lead to disastrous outcomes. To address this issue, we first describe the problem with a stronger Uniformly Constrained MDP (UCMDP) model where we impose constraints on all reachable states; we then propose Objective Suppression, a novel method that adaptively suppresses the task reward maximizing objectives accordin"},"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":"2402.15650","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-02-23T23:22:06Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"a101735719178879cad0e3cbcd4acfa1be5fa604632081ccfb486628fab10480","abstract_canon_sha256":"057b861b10d45b46a801c78e4e3f481ede1529fa7bb36f76d38f68897813c8b0"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:00:22.249300Z","signature_b64":"TfNA5NVdTWHN/kj+dDLYiIR0NkGut6lPDq4or0sFo7tSzQE4VpN1hyvEe3Ck/q2l9os1uTQtbRmJzpvGVJ0wDw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"23eebf0bff7e808ee4f0d47fa25b82bed83bf33d94a0512690f3baa71ef508d5","last_reissued_at":"2026-07-05T09:00:22.248805Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:00:22.248805Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Uniformly Safe RL with Objective Suppression for Multi-Constraint Safety-Critical Applications","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Aleksandr Petiushko, Animesh Garg, Jonathan Booher, Khashayar Rohanimanesh, Wei Liu, Zihan Zhou","submitted_at":"2024-02-23T23:22:06Z","abstract_excerpt":"Safe reinforcement learning tasks are a challenging domain despite being very common in the real world. The widely adopted CMDP model constrains the risks in expectation, which makes room for dangerous behaviors in long-tail states. In safety-critical domains, such behaviors could lead to disastrous outcomes. To address this issue, we first describe the problem with a stronger Uniformly Constrained MDP (UCMDP) model where we impose constraints on all reachable states; we then propose Objective Suppression, a novel method that adaptively suppresses the task reward maximizing objectives accordin"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2402.15650","kind":"arxiv","version":3},"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/2402.15650/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":"2402.15650","created_at":"2026-07-05T09:00:22.248865+00:00"},{"alias_kind":"arxiv_version","alias_value":"2402.15650v3","created_at":"2026-07-05T09:00:22.248865+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2402.15650","created_at":"2026-07-05T09:00:22.248865+00:00"},{"alias_kind":"pith_short_12","alias_value":"EPXL6C77P2AI","created_at":"2026-07-05T09:00:22.248865+00:00"},{"alias_kind":"pith_short_16","alias_value":"EPXL6C77P2AI5ZHQ","created_at":"2026-07-05T09:00:22.248865+00:00"},{"alias_kind":"pith_short_8","alias_value":"EPXL6C77","created_at":"2026-07-05T09:00:22.248865+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/EPXL6C77P2AI5ZHQ2R72EW4CX3","json":"https://pith.science/pith/EPXL6C77P2AI5ZHQ2R72EW4CX3.json","graph_json":"https://pith.science/api/pith-number/EPXL6C77P2AI5ZHQ2R72EW4CX3/graph.json","events_json":"https://pith.science/api/pith-number/EPXL6C77P2AI5ZHQ2R72EW4CX3/events.json","paper":"https://pith.science/paper/EPXL6C77"},"agent_actions":{"view_html":"https://pith.science/pith/EPXL6C77P2AI5ZHQ2R72EW4CX3","download_json":"https://pith.science/pith/EPXL6C77P2AI5ZHQ2R72EW4CX3.json","view_paper":"https://pith.science/paper/EPXL6C77","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2402.15650&json=true","fetch_graph":"https://pith.science/api/pith-number/EPXL6C77P2AI5ZHQ2R72EW4CX3/graph.json","fetch_events":"https://pith.science/api/pith-number/EPXL6C77P2AI5ZHQ2R72EW4CX3/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/EPXL6C77P2AI5ZHQ2R72EW4CX3/action/timestamp_anchor","attest_storage":"https://pith.science/pith/EPXL6C77P2AI5ZHQ2R72EW4CX3/action/storage_attestation","attest_author":"https://pith.science/pith/EPXL6C77P2AI5ZHQ2R72EW4CX3/action/author_attestation","sign_citation":"https://pith.science/pith/EPXL6C77P2AI5ZHQ2R72EW4CX3/action/citation_signature","submit_replication":"https://pith.science/pith/EPXL6C77P2AI5ZHQ2R72EW4CX3/action/replication_record"}},"created_at":"2026-07-05T09:00:22.248865+00:00","updated_at":"2026-07-05T09:00:22.248865+00:00"}