{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:AOQRWIFLT6M2UUWI63HJEJCFZ6","short_pith_number":"pith:AOQRWIFL","schema_version":"1.0","canonical_sha256":"03a11b20ab9f99aa52c8f6ce922445cf971b0047d605186afcc497ba43bcd278","source":{"kind":"arxiv","id":"2110.15332","version":2},"attestation_state":"computed","paper":{"title":"Proximal Reinforcement Learning: Efficient Off-Policy Evaluation in Partially Observed Markov Decision Processes","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["math.OC","math.ST","stat.ML","stat.TH"],"primary_cat":"cs.LG","authors_text":"Andrew Bennett, Nathan Kallus","submitted_at":"2021-10-28T17:46:14Z","abstract_excerpt":"In applications of offline reinforcement learning to observational data, such as in healthcare or education, a general concern is that observed actions might be affected by unobserved factors, inducing confounding and biasing estimates derived under the assumption of a perfect Markov decision process (MDP) model. Here we tackle this by considering off-policy evaluation in a partially observed MDP (POMDP). Specifically, we consider estimating the value of a given target policy in a POMDP given trajectories with only partial state observations generated by a different and unknown policy that may"},"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":"2110.15332","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2021-10-28T17:46:14Z","cross_cats_sorted":["math.OC","math.ST","stat.ML","stat.TH"],"title_canon_sha256":"cb511ff778f046679a0a09e2aac845c3e069655e8824fc1871a4221ffbad4f02","abstract_canon_sha256":"6792b1fbd4a09c99957f0b88927e550174aa2f6c399e91c295685215e9ea2227"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T05:53:46.899729Z","signature_b64":"LwUnM+eCcJ62lmad11NMsHMQH4IiyUnMlmlsjkkwreVZt0FNY0NZnW+c3EUMbFu4yt4TZU0ojUZ4X+6EXGiOAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"03a11b20ab9f99aa52c8f6ce922445cf971b0047d605186afcc497ba43bcd278","last_reissued_at":"2026-07-05T05:53:46.899282Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T05:53:46.899282Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Proximal Reinforcement Learning: Efficient Off-Policy Evaluation in Partially Observed Markov Decision Processes","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["math.OC","math.ST","stat.ML","stat.TH"],"primary_cat":"cs.LG","authors_text":"Andrew Bennett, Nathan Kallus","submitted_at":"2021-10-28T17:46:14Z","abstract_excerpt":"In applications of offline reinforcement learning to observational data, such as in healthcare or education, a general concern is that observed actions might be affected by unobserved factors, inducing confounding and biasing estimates derived under the assumption of a perfect Markov decision process (MDP) model. Here we tackle this by considering off-policy evaluation in a partially observed MDP (POMDP). Specifically, we consider estimating the value of a given target policy in a POMDP given trajectories with only partial state observations generated by a different and unknown policy that may"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2110.15332","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/2110.15332/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":"2110.15332","created_at":"2026-07-05T05:53:46.899343+00:00"},{"alias_kind":"arxiv_version","alias_value":"2110.15332v2","created_at":"2026-07-05T05:53:46.899343+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2110.15332","created_at":"2026-07-05T05:53:46.899343+00:00"},{"alias_kind":"pith_short_12","alias_value":"AOQRWIFLT6M2","created_at":"2026-07-05T05:53:46.899343+00:00"},{"alias_kind":"pith_short_16","alias_value":"AOQRWIFLT6M2UUWI","created_at":"2026-07-05T05:53:46.899343+00:00"},{"alias_kind":"pith_short_8","alias_value":"AOQRWIFL","created_at":"2026-07-05T05:53:46.899343+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.20206","citing_title":"Off-Policy Evaluation for Missingness-Aware Policies in MDPs with Rewards Missing Not at Random","ref_index":63,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/AOQRWIFLT6M2UUWI63HJEJCFZ6","json":"https://pith.science/pith/AOQRWIFLT6M2UUWI63HJEJCFZ6.json","graph_json":"https://pith.science/api/pith-number/AOQRWIFLT6M2UUWI63HJEJCFZ6/graph.json","events_json":"https://pith.science/api/pith-number/AOQRWIFLT6M2UUWI63HJEJCFZ6/events.json","paper":"https://pith.science/paper/AOQRWIFL"},"agent_actions":{"view_html":"https://pith.science/pith/AOQRWIFLT6M2UUWI63HJEJCFZ6","download_json":"https://pith.science/pith/AOQRWIFLT6M2UUWI63HJEJCFZ6.json","view_paper":"https://pith.science/paper/AOQRWIFL","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2110.15332&json=true","fetch_graph":"https://pith.science/api/pith-number/AOQRWIFLT6M2UUWI63HJEJCFZ6/graph.json","fetch_events":"https://pith.science/api/pith-number/AOQRWIFLT6M2UUWI63HJEJCFZ6/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/AOQRWIFLT6M2UUWI63HJEJCFZ6/action/timestamp_anchor","attest_storage":"https://pith.science/pith/AOQRWIFLT6M2UUWI63HJEJCFZ6/action/storage_attestation","attest_author":"https://pith.science/pith/AOQRWIFLT6M2UUWI63HJEJCFZ6/action/author_attestation","sign_citation":"https://pith.science/pith/AOQRWIFLT6M2UUWI63HJEJCFZ6/action/citation_signature","submit_replication":"https://pith.science/pith/AOQRWIFLT6M2UUWI63HJEJCFZ6/action/replication_record"}},"created_at":"2026-07-05T05:53:46.899343+00:00","updated_at":"2026-07-05T05:53:46.899343+00:00"}