{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:UVBHDFTPR4F3LG6FSNYFOO4ULB","short_pith_number":"pith:UVBHDFTP","schema_version":"1.0","canonical_sha256":"a54271966f8f0bb59bc59370573b94587d4a10db7114de1661584d9f5af4ac0d","source":{"kind":"arxiv","id":"2412.13184","version":1},"attestation_state":"computed","paper":{"title":"Tilted Quantile Gradient Updates for Quantile-Constrained Reinforcement Learning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Chenglin Li, Guangchun Ruan, Hua Geng","submitted_at":"2024-12-17T18:58:00Z","abstract_excerpt":"Safe reinforcement learning (RL) is a popular and versatile paradigm to learn reward-maximizing policies with safety guarantees. Previous works tend to express the safety constraints in an expectation form due to the ease of implementation, but this turns out to be ineffective in maintaining safety constraints with high probability. To this end, we move to the quantile-constrained RL that enables a higher level of safety without any expectation-form approximations. We directly estimate the quantile gradients through sampling and provide the theoretical proofs of convergence. Then a tilted upda"},"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":"2412.13184","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-12-17T18:58:00Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"1fa35aa0ab5f2c63dc443c07ce5046074b82e39496533c135c63e5aaeef03d01","abstract_canon_sha256":"d9f25bb903152bae90c3b1f7cb64e026a6807917e99d330a047f1146d0ce92e2"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:50:41.435327Z","signature_b64":"dLIQ1MfPhRLGh3mqLzXaZzEMRDp823Vu2wH/ZW4VUHLYxtOVrHorSbS8m3ShB+Iv44wlGmC1nDFC6+LPNl4RCQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"a54271966f8f0bb59bc59370573b94587d4a10db7114de1661584d9f5af4ac0d","last_reissued_at":"2026-07-05T09:50:41.434822Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:50:41.434822Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Tilted Quantile Gradient Updates for Quantile-Constrained Reinforcement Learning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Chenglin Li, Guangchun Ruan, Hua Geng","submitted_at":"2024-12-17T18:58:00Z","abstract_excerpt":"Safe reinforcement learning (RL) is a popular and versatile paradigm to learn reward-maximizing policies with safety guarantees. Previous works tend to express the safety constraints in an expectation form due to the ease of implementation, but this turns out to be ineffective in maintaining safety constraints with high probability. To this end, we move to the quantile-constrained RL that enables a higher level of safety without any expectation-form approximations. We directly estimate the quantile gradients through sampling and provide the theoretical proofs of convergence. Then a tilted upda"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2412.13184","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/2412.13184/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":"2412.13184","created_at":"2026-07-05T09:50:41.434887+00:00"},{"alias_kind":"arxiv_version","alias_value":"2412.13184v1","created_at":"2026-07-05T09:50:41.434887+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2412.13184","created_at":"2026-07-05T09:50:41.434887+00:00"},{"alias_kind":"pith_short_12","alias_value":"UVBHDFTPR4F3","created_at":"2026-07-05T09:50:41.434887+00:00"},{"alias_kind":"pith_short_16","alias_value":"UVBHDFTPR4F3LG6F","created_at":"2026-07-05T09:50:41.434887+00:00"},{"alias_kind":"pith_short_8","alias_value":"UVBHDFTP","created_at":"2026-07-05T09:50:41.434887+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/UVBHDFTPR4F3LG6FSNYFOO4ULB","json":"https://pith.science/pith/UVBHDFTPR4F3LG6FSNYFOO4ULB.json","graph_json":"https://pith.science/api/pith-number/UVBHDFTPR4F3LG6FSNYFOO4ULB/graph.json","events_json":"https://pith.science/api/pith-number/UVBHDFTPR4F3LG6FSNYFOO4ULB/events.json","paper":"https://pith.science/paper/UVBHDFTP"},"agent_actions":{"view_html":"https://pith.science/pith/UVBHDFTPR4F3LG6FSNYFOO4ULB","download_json":"https://pith.science/pith/UVBHDFTPR4F3LG6FSNYFOO4ULB.json","view_paper":"https://pith.science/paper/UVBHDFTP","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2412.13184&json=true","fetch_graph":"https://pith.science/api/pith-number/UVBHDFTPR4F3LG6FSNYFOO4ULB/graph.json","fetch_events":"https://pith.science/api/pith-number/UVBHDFTPR4F3LG6FSNYFOO4ULB/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/UVBHDFTPR4F3LG6FSNYFOO4ULB/action/timestamp_anchor","attest_storage":"https://pith.science/pith/UVBHDFTPR4F3LG6FSNYFOO4ULB/action/storage_attestation","attest_author":"https://pith.science/pith/UVBHDFTPR4F3LG6FSNYFOO4ULB/action/author_attestation","sign_citation":"https://pith.science/pith/UVBHDFTPR4F3LG6FSNYFOO4ULB/action/citation_signature","submit_replication":"https://pith.science/pith/UVBHDFTPR4F3LG6FSNYFOO4ULB/action/replication_record"}},"created_at":"2026-07-05T09:50:41.434887+00:00","updated_at":"2026-07-05T09:50:41.434887+00:00"}