{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:DMYZIDMHSZX7VGA3RRPDDVVIRN","short_pith_number":"pith:DMYZIDMH","schema_version":"1.0","canonical_sha256":"1b31940d87966ffa981b8c5e31d6a88b4e8908e7181e993f3d13d91a708b29e8","source":{"kind":"arxiv","id":"2606.06096","version":1},"attestation_state":"computed","paper":{"title":"OrderGrad: Optimizing Beyond the Mean with Order-Statistic Policy Gradient Estimation","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.LG","authors_text":"Kohsei Matsutani, Paavo Parmas, Shota Takashiro, Soichiro Nishimori, Takeshi Kojima, Yongmin Kim, Yusuke Iwasawa, Yutaka Matsuo","submitted_at":"2026-06-04T12:34:15Z","abstract_excerpt":"Policy-gradient methods usually optimize expected return, but many real world applications care about distributional properties of returns: tail risk, outlier robustness, or best-of-K discovery. We introduce OrderGrad, a family of likelihood-ratio and reparameterization gradient estimators for order-statistic objectives. OrderGrad optimizes finite-sample L-statistics, i.e., weighted averages of sorted rewards or costs, recovering objectives such as VaR, CVaR, trimmed means, medians, and top-m/best-of-K criteria by changing only the rank weights. For any fixed sample size and rank-weight vector"},"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":"2606.06096","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-06-04T12:34:15Z","cross_cats_sorted":["cs.AI","cs.CL"],"title_canon_sha256":"ada7f497879153bd9912be16ca1eae2314fd6feb81b7023afd32f9840c3f8498","abstract_canon_sha256":"8af8bf445927ffef654f5ae3e49aff2d9445c87057dc5a475506d5f30b2f40a0"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-05T01:15:33.194942Z","signature_b64":"asDLvYmZa4RVXtW6zdb4sJzWZuRe6UoIBbq93Vt5hSRn/RNXiaEp8v76TPbeugfSQ+2Xs6O+GicDyR72feokCQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"1b31940d87966ffa981b8c5e31d6a88b4e8908e7181e993f3d13d91a708b29e8","last_reissued_at":"2026-06-05T01:15:33.194206Z","signature_status":"signed_v1","first_computed_at":"2026-06-05T01:15:33.194206Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"OrderGrad: Optimizing Beyond the Mean with Order-Statistic Policy Gradient Estimation","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.LG","authors_text":"Kohsei Matsutani, Paavo Parmas, Shota Takashiro, Soichiro Nishimori, Takeshi Kojima, Yongmin Kim, Yusuke Iwasawa, Yutaka Matsuo","submitted_at":"2026-06-04T12:34:15Z","abstract_excerpt":"Policy-gradient methods usually optimize expected return, but many real world applications care about distributional properties of returns: tail risk, outlier robustness, or best-of-K discovery. We introduce OrderGrad, a family of likelihood-ratio and reparameterization gradient estimators for order-statistic objectives. OrderGrad optimizes finite-sample L-statistics, i.e., weighted averages of sorted rewards or costs, recovering objectives such as VaR, CVaR, trimmed means, medians, and top-m/best-of-K criteria by changing only the rank weights. For any fixed sample size and rank-weight vector"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.06096","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/2606.06096/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":"2606.06096","created_at":"2026-06-05T01:15:33.194324+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.06096v1","created_at":"2026-06-05T01:15:33.194324+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.06096","created_at":"2026-06-05T01:15:33.194324+00:00"},{"alias_kind":"pith_short_12","alias_value":"DMYZIDMHSZX7","created_at":"2026-06-05T01:15:33.194324+00:00"},{"alias_kind":"pith_short_16","alias_value":"DMYZIDMHSZX7VGA3","created_at":"2026-06-05T01:15:33.194324+00:00"},{"alias_kind":"pith_short_8","alias_value":"DMYZIDMH","created_at":"2026-06-05T01:15:33.194324+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/DMYZIDMHSZX7VGA3RRPDDVVIRN","json":"https://pith.science/pith/DMYZIDMHSZX7VGA3RRPDDVVIRN.json","graph_json":"https://pith.science/api/pith-number/DMYZIDMHSZX7VGA3RRPDDVVIRN/graph.json","events_json":"https://pith.science/api/pith-number/DMYZIDMHSZX7VGA3RRPDDVVIRN/events.json","paper":"https://pith.science/paper/DMYZIDMH"},"agent_actions":{"view_html":"https://pith.science/pith/DMYZIDMHSZX7VGA3RRPDDVVIRN","download_json":"https://pith.science/pith/DMYZIDMHSZX7VGA3RRPDDVVIRN.json","view_paper":"https://pith.science/paper/DMYZIDMH","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.06096&json=true","fetch_graph":"https://pith.science/api/pith-number/DMYZIDMHSZX7VGA3RRPDDVVIRN/graph.json","fetch_events":"https://pith.science/api/pith-number/DMYZIDMHSZX7VGA3RRPDDVVIRN/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/DMYZIDMHSZX7VGA3RRPDDVVIRN/action/timestamp_anchor","attest_storage":"https://pith.science/pith/DMYZIDMHSZX7VGA3RRPDDVVIRN/action/storage_attestation","attest_author":"https://pith.science/pith/DMYZIDMHSZX7VGA3RRPDDVVIRN/action/author_attestation","sign_citation":"https://pith.science/pith/DMYZIDMHSZX7VGA3RRPDDVVIRN/action/citation_signature","submit_replication":"https://pith.science/pith/DMYZIDMHSZX7VGA3RRPDDVVIRN/action/replication_record"}},"created_at":"2026-06-05T01:15:33.194324+00:00","updated_at":"2026-06-05T01:15:33.194324+00:00"}