{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2018:LWORRY3PZ6JPJHRD76N7O6MJDU","short_pith_number":"pith:LWORRY3P","schema_version":"1.0","canonical_sha256":"5d9d18e36fcf92f49e23ff9bf779891d2414981d2d7cbef05f3843a8705610f9","source":{"kind":"arxiv","id":"1807.01066","version":2},"attestation_state":"computed","paper":{"title":"Behaviour Policy Estimation in Off-Policy Policy Evaluation: Calibration Matters","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["stat.ML"],"primary_cat":"cs.LG","authors_text":"Aldo Faisal, Aniruddh Raghu, Emma Brunskill, Finale Doshi-Velez, Matthieu Komorowski, Omer Gottesman, Yao Liu","submitted_at":"2018-07-03T10:20:22Z","abstract_excerpt":"In this work, we consider the problem of estimating a behaviour policy for use in Off-Policy Policy Evaluation (OPE) when the true behaviour policy is unknown. Via a series of empirical studies, we demonstrate how accurate OPE is strongly dependent on the calibration of estimated behaviour policy models: how precisely the behaviour policy is estimated from data. We show how powerful parametric models such as neural networks can result in highly uncalibrated behaviour policy models on a real-world medical dataset, and illustrate how a simple, non-parametric, k-nearest neighbours model produces "},"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":"1807.01066","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2018-07-03T10:20:22Z","cross_cats_sorted":["stat.ML"],"title_canon_sha256":"4281d9b28b1109e4f09957f47cdb89aa5a8ffe1d98f75d8da798ad24d45f1f51","abstract_canon_sha256":"903ec605818b206981d372095bcde710b88d061335110856d2743d2ba75ad60a"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-18T00:11:10.940652Z","signature_b64":"Ix6StcNP0FaQuS+pOe16H5i8jVO+AxyebWOs8vt88eH2ZxltyF8qBv6C36gaZdH9z4Wr4P/j7zecw9SU2U5vBA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"5d9d18e36fcf92f49e23ff9bf779891d2414981d2d7cbef05f3843a8705610f9","last_reissued_at":"2026-05-18T00:11:10.939895Z","signature_status":"signed_v1","first_computed_at":"2026-05-18T00:11:10.939895Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Behaviour Policy Estimation in Off-Policy Policy Evaluation: Calibration Matters","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["stat.ML"],"primary_cat":"cs.LG","authors_text":"Aldo Faisal, Aniruddh Raghu, Emma Brunskill, Finale Doshi-Velez, Matthieu Komorowski, Omer Gottesman, Yao Liu","submitted_at":"2018-07-03T10:20:22Z","abstract_excerpt":"In this work, we consider the problem of estimating a behaviour policy for use in Off-Policy Policy Evaluation (OPE) when the true behaviour policy is unknown. Via a series of empirical studies, we demonstrate how accurate OPE is strongly dependent on the calibration of estimated behaviour policy models: how precisely the behaviour policy is estimated from data. We show how powerful parametric models such as neural networks can result in highly uncalibrated behaviour policy models on a real-world medical dataset, and illustrate how a simple, non-parametric, k-nearest neighbours model produces "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1807.01066","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":""},"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":"1807.01066","created_at":"2026-05-18T00:11:10.940001+00:00"},{"alias_kind":"arxiv_version","alias_value":"1807.01066v2","created_at":"2026-05-18T00:11:10.940001+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1807.01066","created_at":"2026-05-18T00:11:10.940001+00:00"},{"alias_kind":"pith_short_12","alias_value":"LWORRY3PZ6JP","created_at":"2026-05-18T12:32:37.024351+00:00"},{"alias_kind":"pith_short_16","alias_value":"LWORRY3PZ6JPJHRD","created_at":"2026-05-18T12:32:37.024351+00:00"},{"alias_kind":"pith_short_8","alias_value":"LWORRY3P","created_at":"2026-05-18T12:32:37.024351+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2005.01643","citing_title":"Offline Reinforcement Learning: Tutorial, Review, and Perspectives on Open Problems","ref_index":194,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/LWORRY3PZ6JPJHRD76N7O6MJDU","json":"https://pith.science/pith/LWORRY3PZ6JPJHRD76N7O6MJDU.json","graph_json":"https://pith.science/api/pith-number/LWORRY3PZ6JPJHRD76N7O6MJDU/graph.json","events_json":"https://pith.science/api/pith-number/LWORRY3PZ6JPJHRD76N7O6MJDU/events.json","paper":"https://pith.science/paper/LWORRY3P"},"agent_actions":{"view_html":"https://pith.science/pith/LWORRY3PZ6JPJHRD76N7O6MJDU","download_json":"https://pith.science/pith/LWORRY3PZ6JPJHRD76N7O6MJDU.json","view_paper":"https://pith.science/paper/LWORRY3P","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=1807.01066&json=true","fetch_graph":"https://pith.science/api/pith-number/LWORRY3PZ6JPJHRD76N7O6MJDU/graph.json","fetch_events":"https://pith.science/api/pith-number/LWORRY3PZ6JPJHRD76N7O6MJDU/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/LWORRY3PZ6JPJHRD76N7O6MJDU/action/timestamp_anchor","attest_storage":"https://pith.science/pith/LWORRY3PZ6JPJHRD76N7O6MJDU/action/storage_attestation","attest_author":"https://pith.science/pith/LWORRY3PZ6JPJHRD76N7O6MJDU/action/author_attestation","sign_citation":"https://pith.science/pith/LWORRY3PZ6JPJHRD76N7O6MJDU/action/citation_signature","submit_replication":"https://pith.science/pith/LWORRY3PZ6JPJHRD76N7O6MJDU/action/replication_record"}},"created_at":"2026-05-18T00:11:10.940001+00:00","updated_at":"2026-05-18T00:11:10.940001+00:00"}