{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:CLKGG7GH3AZD7NMO22XQLBJ7OR","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"a662487e501e5952911af607c6e10b4bdbd3b18692871938a12ba80531e99d48","cross_cats_sorted":["stat.ML"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-01-30T13:57:47Z","title_canon_sha256":"30ce4acde7984652dff459c97f283d90840781c4c1a41d0540b94b58f3c4c83c"},"schema_version":"1.0","source":{"id":"2601.22993","kind":"arxiv","version":4}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2601.22993","created_at":"2026-06-30T01:17:33Z"},{"alias_kind":"arxiv_version","alias_value":"2601.22993v4","created_at":"2026-06-30T01:17:33Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2601.22993","created_at":"2026-06-30T01:17:33Z"},{"alias_kind":"pith_short_12","alias_value":"CLKGG7GH3AZD","created_at":"2026-06-30T01:17:33Z"},{"alias_kind":"pith_short_16","alias_value":"CLKGG7GH3AZD7NMO","created_at":"2026-06-30T01:17:33Z"},{"alias_kind":"pith_short_8","alias_value":"CLKGG7GH","created_at":"2026-06-30T01:17:33Z"}],"graph_snapshots":[{"event_id":"sha256:3e0419dea08a18d9e91c6f7a34b72f1bf93c22f05e92e7e8622f70a4d02779a7","target":"graph","created_at":"2026-06-30T01:17:33Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":4,"items":[{"attestation":"unclaimed","claim_id":"C1","kind":"strongest_claim","source":"verdict.strongest_claim","status":"machine_extracted","text":"VaR-CPO achieves zero constraint violations during training in feasible environments, a property baseline methods fail to uphold, while providing worst-case bounds for policy improvement and constraint violation."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"The Cantelli inequality provides a sufficiently tight and conservative approximation to the true VaR constraint so that enforcing the bound still guarantees the original probabilistic constraint in practice."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"VaR-CPO approximates non-differentiable VaR constraints via Cantelli's inequality to enable safe, sample-efficient policy optimization with zero training violations in feasible environments."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"VaR-CPO approximates Value-at-Risk constraints via Cantelli's inequality to guarantee zero violations during training in feasible reinforcement learning environments."}],"snapshot_sha256":"5833deae7b8bf68db5033edbd4f3443af94e8f792a74de4f7a049cff80bd6916"},"formal_canon":{"evidence_count":2,"snapshot_sha256":"9040767e148b8c72e9fc099919b372291776c02fefffa41b70a4c425745a302e"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2601.22993/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"We introduce Canary, a risk-averse method designed to optimize Value-at-Risk (VaR) constrained reinforcement learning (RL) problems. We employ Cantelli's inequality to obtain a tractable, conservative and smooth bound on the VaR constraint based on the first two moments of the cost return. This yields a constraint estimator that remains stable with tight violation thresholds in dense cost regimes. Extending the trust-region framework of the Constrained Policy Optimization (CPO) method, we further provide worst-case bounds for both policy improvement and constraint violation during the training","authors_text":"Jan-Peter Calliess, Rohan Tangri","cross_cats":["stat.ML"],"headline":"VaR-CPO approximates Value-at-Risk constraints via Cantelli's inequality to guarantee zero violations during training in feasible reinforcement learning environments.","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-01-30T13:57:47Z","title":"Constrained Policy Optimization with Cantelli-Bounded Value-at-Risk"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2601.22993","kind":"arxiv","version":4},"verdict":{"created_at":"2026-05-16T09:35:17.857407Z","id":"7b84fecb-eea6-421d-90b3-a8385299ea5b","model_set":{"reader":"grok-4.3"},"one_line_summary":"VaR-CPO approximates non-differentiable VaR constraints via Cantelli's inequality to enable safe, sample-efficient policy optimization with zero training violations in feasible environments.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"VaR-CPO approximates Value-at-Risk constraints via Cantelli's inequality to guarantee zero violations during training in feasible reinforcement learning environments.","strongest_claim":"VaR-CPO achieves zero constraint violations during training in feasible environments, a property baseline methods fail to uphold, while providing worst-case bounds for policy improvement and constraint violation.","weakest_assumption":"The Cantelli inequality provides a sufficiently tight and conservative approximation to the true VaR constraint so that enforcing the bound still guarantees the original probabilistic constraint in practice."}},"verdict_id":"7b84fecb-eea6-421d-90b3-a8385299ea5b"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:7e3f335427f7f05d5c54ff0db69ff5bd325bd0d6da3d4d6ba3fc0058c614dbda","target":"record","created_at":"2026-06-30T01:17:33Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"a662487e501e5952911af607c6e10b4bdbd3b18692871938a12ba80531e99d48","cross_cats_sorted":["stat.ML"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-01-30T13:57:47Z","title_canon_sha256":"30ce4acde7984652dff459c97f283d90840781c4c1a41d0540b94b58f3c4c83c"},"schema_version":"1.0","source":{"id":"2601.22993","kind":"arxiv","version":4}},"canonical_sha256":"12d4637cc7d8323fb58ed6af05853f747e55a2b67d4353530956dd25adbf10a5","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"12d4637cc7d8323fb58ed6af05853f747e55a2b67d4353530956dd25adbf10a5","first_computed_at":"2026-06-30T01:17:33.333863Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-06-30T01:17:33.333863Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"zRk+JsQwltP0gDPfrzWW5Rck7jc6oyIh9IvsqCG5BMtjvwuYhucVBTnKWkna3PC9Z7oHB6cbwaJFsIXBy+2pDQ==","signature_status":"signed_v1","signed_at":"2026-06-30T01:17:33.334599Z","signed_message":"canonical_sha256_bytes"},"source_id":"2601.22993","source_kind":"arxiv","source_version":4}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:7e3f335427f7f05d5c54ff0db69ff5bd325bd0d6da3d4d6ba3fc0058c614dbda","sha256:3e0419dea08a18d9e91c6f7a34b72f1bf93c22f05e92e7e8622f70a4d02779a7"],"state_sha256":"0582e5cedfd91e75c46801a3d12474214a77c1c148430ede85d1e97975c56625"}