{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:L7KCFHCAIIPZP2ZDHOHQMBEJFZ","short_pith_number":"pith:L7KCFHCA","schema_version":"1.0","canonical_sha256":"5fd4229c40421f97eb233b8f0604892e759c987b7fcd61dd875c4e422fd314bf","source":{"kind":"arxiv","id":"2606.04749","version":1},"attestation_state":"computed","paper":{"title":"COP-Q: Safety-First Reinforcement Learning for Robot Control via Cholesky-Ordered Projection","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.RO","authors_text":"Guopeng Li, Julian F. P. Kooij, Matthijs T. J. Spaan, Moritz A. Zanger","submitted_at":"2026-06-03T11:30:10Z","abstract_excerpt":"Safe robot control requires maximizing return while satisfying safety constraints. In off-policy safe reinforcement learning, reward and safety Q-values are commonly learned by separate critic ensembles, with uncertainty handled independently for each objective. This objective-wise treatment neglects inter-objective correlation and can lead to overly conservative value estimates, thereby reducing sample efficiency. To address this issue, we propose Cholesky-Ordered Projection Q-learning (COP-Q), a safety-first method that incorporates inter-objective covariance into vector-valued Q-value estim"},"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":"2606.04749","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.RO","submitted_at":"2026-06-03T11:30:10Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"61fa15bd4016b2111f056e4579fbe57ac53ee30e54be2d45af60c15640aa7537","abstract_canon_sha256":"4bd2f1615d8e6cce99b266588a083760c91847a4ab79cb7d4248d209b9cd6a02"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-04T01:09:28.418997Z","signature_b64":"Gmq0bNyjZWOZA2oVAz47j//6vzErsrexR6dOd33sF52RntZdXD9gwRyaZNz0JhADfGrqc0RPmlUayR+DMjj1Bg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"5fd4229c40421f97eb233b8f0604892e759c987b7fcd61dd875c4e422fd314bf","last_reissued_at":"2026-06-04T01:09:28.418270Z","signature_status":"signed_v1","first_computed_at":"2026-06-04T01:09:28.418270Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"COP-Q: Safety-First Reinforcement Learning for Robot Control via Cholesky-Ordered Projection","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.RO","authors_text":"Guopeng Li, Julian F. P. Kooij, Matthijs T. J. Spaan, Moritz A. Zanger","submitted_at":"2026-06-03T11:30:10Z","abstract_excerpt":"Safe robot control requires maximizing return while satisfying safety constraints. In off-policy safe reinforcement learning, reward and safety Q-values are commonly learned by separate critic ensembles, with uncertainty handled independently for each objective. This objective-wise treatment neglects inter-objective correlation and can lead to overly conservative value estimates, thereby reducing sample efficiency. To address this issue, we propose Cholesky-Ordered Projection Q-learning (COP-Q), a safety-first method that incorporates inter-objective covariance into vector-valued Q-value estim"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.04749","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/2606.04749/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":"2606.04749","created_at":"2026-06-04T01:09:28.418374+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.04749v1","created_at":"2026-06-04T01:09:28.418374+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.04749","created_at":"2026-06-04T01:09:28.418374+00:00"},{"alias_kind":"pith_short_12","alias_value":"L7KCFHCAIIPZ","created_at":"2026-06-04T01:09:28.418374+00:00"},{"alias_kind":"pith_short_16","alias_value":"L7KCFHCAIIPZP2ZD","created_at":"2026-06-04T01:09:28.418374+00:00"},{"alias_kind":"pith_short_8","alias_value":"L7KCFHCA","created_at":"2026-06-04T01:09:28.418374+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/L7KCFHCAIIPZP2ZDHOHQMBEJFZ","json":"https://pith.science/pith/L7KCFHCAIIPZP2ZDHOHQMBEJFZ.json","graph_json":"https://pith.science/api/pith-number/L7KCFHCAIIPZP2ZDHOHQMBEJFZ/graph.json","events_json":"https://pith.science/api/pith-number/L7KCFHCAIIPZP2ZDHOHQMBEJFZ/events.json","paper":"https://pith.science/paper/L7KCFHCA"},"agent_actions":{"view_html":"https://pith.science/pith/L7KCFHCAIIPZP2ZDHOHQMBEJFZ","download_json":"https://pith.science/pith/L7KCFHCAIIPZP2ZDHOHQMBEJFZ.json","view_paper":"https://pith.science/paper/L7KCFHCA","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.04749&json=true","fetch_graph":"https://pith.science/api/pith-number/L7KCFHCAIIPZP2ZDHOHQMBEJFZ/graph.json","fetch_events":"https://pith.science/api/pith-number/L7KCFHCAIIPZP2ZDHOHQMBEJFZ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/L7KCFHCAIIPZP2ZDHOHQMBEJFZ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/L7KCFHCAIIPZP2ZDHOHQMBEJFZ/action/storage_attestation","attest_author":"https://pith.science/pith/L7KCFHCAIIPZP2ZDHOHQMBEJFZ/action/author_attestation","sign_citation":"https://pith.science/pith/L7KCFHCAIIPZP2ZDHOHQMBEJFZ/action/citation_signature","submit_replication":"https://pith.science/pith/L7KCFHCAIIPZP2ZDHOHQMBEJFZ/action/replication_record"}},"created_at":"2026-06-04T01:09:28.418374+00:00","updated_at":"2026-06-04T01:09:28.418374+00:00"}