{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:KKGH2MXAYCLOJK5XFLZVGHOELD","short_pith_number":"pith:KKGH2MXA","schema_version":"1.0","canonical_sha256":"528c7d32e0c096e4abb72af3531dc458f5ac92185cd27c19220a516d1961fa73","source":{"kind":"arxiv","id":"2507.15532","version":2},"attestation_state":"computed","paper":{"title":"Data-Efficient Safe Policy Improvement Using Parametric Structure","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Guillermo A. P\\'erez, Kasper Engelen, Marnix Suilen","submitted_at":"2025-07-21T12:00:03Z","abstract_excerpt":"Safe policy improvement (SPI) is an offline reinforcement learning problem in which a new policy that reliably outperforms the behavior policy with high confidence needs to be computed using only a dataset and the behavior policy. Markov decision processes (MDPs) are the standard formalism for modeling environments in SPI. In many applications, additional information in the form of parametric dependencies between distributions in the transition dynamics is available. We make SPI more data-efficient by leveraging these dependencies through three contributions: (1) a parametric SPI algorithm tha"},"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":"2507.15532","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2025-07-21T12:00:03Z","cross_cats_sorted":[],"title_canon_sha256":"6c57621d524bd73e66112c0ef26ff30128049916b38c666b121aac2854ab0ab5","abstract_canon_sha256":"7fdb0522b46e11316ff952eea924142e23076b935bcd12cad244e8301cb24503"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:55:41.447338Z","signature_b64":"NZ3F9PvwCmNEKhvOOHejGHzJcmzZzXLml9WHI568BLkhYpGxhQVa6rLO7fqDFPCcc4nmaBNNGMyqH8FdP4rKAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"528c7d32e0c096e4abb72af3531dc458f5ac92185cd27c19220a516d1961fa73","last_reissued_at":"2026-07-05T11:55:41.446779Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:55:41.446779Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Data-Efficient Safe Policy Improvement Using Parametric Structure","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Guillermo A. P\\'erez, Kasper Engelen, Marnix Suilen","submitted_at":"2025-07-21T12:00:03Z","abstract_excerpt":"Safe policy improvement (SPI) is an offline reinforcement learning problem in which a new policy that reliably outperforms the behavior policy with high confidence needs to be computed using only a dataset and the behavior policy. Markov decision processes (MDPs) are the standard formalism for modeling environments in SPI. In many applications, additional information in the form of parametric dependencies between distributions in the transition dynamics is available. We make SPI more data-efficient by leveraging these dependencies through three contributions: (1) a parametric SPI algorithm tha"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2507.15532","kind":"arxiv","version":2},"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/2507.15532/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":"2507.15532","created_at":"2026-07-05T11:55:41.446840+00:00"},{"alias_kind":"arxiv_version","alias_value":"2507.15532v2","created_at":"2026-07-05T11:55:41.446840+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2507.15532","created_at":"2026-07-05T11:55:41.446840+00:00"},{"alias_kind":"pith_short_12","alias_value":"KKGH2MXAYCLO","created_at":"2026-07-05T11:55:41.446840+00:00"},{"alias_kind":"pith_short_16","alias_value":"KKGH2MXAYCLOJK5X","created_at":"2026-07-05T11:55:41.446840+00:00"},{"alias_kind":"pith_short_8","alias_value":"KKGH2MXA","created_at":"2026-07-05T11:55:41.446840+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.10293","citing_title":"Robust Probabilistic Shielding for Safe Offline Reinforcement Learning","ref_index":16,"is_internal_anchor":false},{"citing_arxiv_id":"2605.01339","citing_title":"Robust Parameter Learning for Uncertain MDPs","ref_index":21,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/KKGH2MXAYCLOJK5XFLZVGHOELD","json":"https://pith.science/pith/KKGH2MXAYCLOJK5XFLZVGHOELD.json","graph_json":"https://pith.science/api/pith-number/KKGH2MXAYCLOJK5XFLZVGHOELD/graph.json","events_json":"https://pith.science/api/pith-number/KKGH2MXAYCLOJK5XFLZVGHOELD/events.json","paper":"https://pith.science/paper/KKGH2MXA"},"agent_actions":{"view_html":"https://pith.science/pith/KKGH2MXAYCLOJK5XFLZVGHOELD","download_json":"https://pith.science/pith/KKGH2MXAYCLOJK5XFLZVGHOELD.json","view_paper":"https://pith.science/paper/KKGH2MXA","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2507.15532&json=true","fetch_graph":"https://pith.science/api/pith-number/KKGH2MXAYCLOJK5XFLZVGHOELD/graph.json","fetch_events":"https://pith.science/api/pith-number/KKGH2MXAYCLOJK5XFLZVGHOELD/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/KKGH2MXAYCLOJK5XFLZVGHOELD/action/timestamp_anchor","attest_storage":"https://pith.science/pith/KKGH2MXAYCLOJK5XFLZVGHOELD/action/storage_attestation","attest_author":"https://pith.science/pith/KKGH2MXAYCLOJK5XFLZVGHOELD/action/author_attestation","sign_citation":"https://pith.science/pith/KKGH2MXAYCLOJK5XFLZVGHOELD/action/citation_signature","submit_replication":"https://pith.science/pith/KKGH2MXAYCLOJK5XFLZVGHOELD/action/replication_record"}},"created_at":"2026-07-05T11:55:41.446840+00:00","updated_at":"2026-07-05T11:55:41.446840+00:00"}