{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:PSYHHRSUXPT2EDMBFHHCDHUQLC","short_pith_number":"pith:PSYHHRSU","schema_version":"1.0","canonical_sha256":"7cb073c654bbe7a20d8129ce219e905894f5d1a267d6d596c820d95cb3303d56","source":{"kind":"arxiv","id":"2605.26784","version":1},"attestation_state":"computed","paper":{"title":"Ratio-Variance Regularized Policy Optimization","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Dong Li, Fuchun Sun, Huaping Liu, Jianye Hao, Lei Lv, Shuo Han, Yihan Hu, Yu Luo","submitted_at":"2026-05-26T09:53:42Z","abstract_excerpt":"Standard on-policy reinforcement learning relies on heuristic clipping to enforce trust regions, but this mechanism imposes a severe cost by indiscriminately truncating high-return yet high-divergence updates. We demonstrate that explicitly constraining the policy ratio variance provides a principled local approximation to trust-region constraints, eliminating the need for binary hard clipping. By acting as a distributional ``soft brake'', this approach preserves critical gradient signals from novel discoveries while naturally down-weighting and enabling the reuse of stale, off-policy data. We"},"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":"2605.26784","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-26T09:53:42Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"04934deeb4ad9ca71966ea4dd7fc84a2a264bb2ee2e667f2a039fd8601c10999","abstract_canon_sha256":"2eb09ae68a96a5700b8ec7cfd1ebc3c04406bb50afe9e4cbe827fc8b78a5de89"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-27T01:06:12.533522Z","signature_b64":"UAzNUibJwklN1uhkUfCvgN7YY8VRtNWGXrMmuSP0yBPqtJbs4ANQKDKIvaGiXeAv4y2YRyBUNAsw+quqh2ZoDw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"7cb073c654bbe7a20d8129ce219e905894f5d1a267d6d596c820d95cb3303d56","last_reissued_at":"2026-05-27T01:06:12.532841Z","signature_status":"signed_v1","first_computed_at":"2026-05-27T01:06:12.532841Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Ratio-Variance Regularized Policy Optimization","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Dong Li, Fuchun Sun, Huaping Liu, Jianye Hao, Lei Lv, Shuo Han, Yihan Hu, Yu Luo","submitted_at":"2026-05-26T09:53:42Z","abstract_excerpt":"Standard on-policy reinforcement learning relies on heuristic clipping to enforce trust regions, but this mechanism imposes a severe cost by indiscriminately truncating high-return yet high-divergence updates. We demonstrate that explicitly constraining the policy ratio variance provides a principled local approximation to trust-region constraints, eliminating the need for binary hard clipping. By acting as a distributional ``soft brake'', this approach preserves critical gradient signals from novel discoveries while naturally down-weighting and enabling the reuse of stale, off-policy data. We"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.26784","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/2605.26784/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":"2605.26784","created_at":"2026-05-27T01:06:12.532932+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.26784v1","created_at":"2026-05-27T01:06:12.532932+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.26784","created_at":"2026-05-27T01:06:12.532932+00:00"},{"alias_kind":"pith_short_12","alias_value":"PSYHHRSUXPT2","created_at":"2026-05-27T01:06:12.532932+00:00"},{"alias_kind":"pith_short_16","alias_value":"PSYHHRSUXPT2EDMB","created_at":"2026-05-27T01:06:12.532932+00:00"},{"alias_kind":"pith_short_8","alias_value":"PSYHHRSU","created_at":"2026-05-27T01:06:12.532932+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/PSYHHRSUXPT2EDMBFHHCDHUQLC","json":"https://pith.science/pith/PSYHHRSUXPT2EDMBFHHCDHUQLC.json","graph_json":"https://pith.science/api/pith-number/PSYHHRSUXPT2EDMBFHHCDHUQLC/graph.json","events_json":"https://pith.science/api/pith-number/PSYHHRSUXPT2EDMBFHHCDHUQLC/events.json","paper":"https://pith.science/paper/PSYHHRSU"},"agent_actions":{"view_html":"https://pith.science/pith/PSYHHRSUXPT2EDMBFHHCDHUQLC","download_json":"https://pith.science/pith/PSYHHRSUXPT2EDMBFHHCDHUQLC.json","view_paper":"https://pith.science/paper/PSYHHRSU","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.26784&json=true","fetch_graph":"https://pith.science/api/pith-number/PSYHHRSUXPT2EDMBFHHCDHUQLC/graph.json","fetch_events":"https://pith.science/api/pith-number/PSYHHRSUXPT2EDMBFHHCDHUQLC/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/PSYHHRSUXPT2EDMBFHHCDHUQLC/action/timestamp_anchor","attest_storage":"https://pith.science/pith/PSYHHRSUXPT2EDMBFHHCDHUQLC/action/storage_attestation","attest_author":"https://pith.science/pith/PSYHHRSUXPT2EDMBFHHCDHUQLC/action/author_attestation","sign_citation":"https://pith.science/pith/PSYHHRSUXPT2EDMBFHHCDHUQLC/action/citation_signature","submit_replication":"https://pith.science/pith/PSYHHRSUXPT2EDMBFHHCDHUQLC/action/replication_record"}},"created_at":"2026-05-27T01:06:12.532932+00:00","updated_at":"2026-05-27T01:06:12.532932+00:00"}