{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:FKVGBDRPW2VW3YZW7WRRXSOXPM","short_pith_number":"pith:FKVGBDRP","schema_version":"1.0","canonical_sha256":"2aaa608e2fb6ab6de336fda31bc9d77b2929caac791746463ea41796035349a2","source":{"kind":"arxiv","id":"2305.08062","version":2},"attestation_state":"computed","paper":{"title":"Off-Policy Evaluation for Large Action Spaces via Conjunct Effect Modeling","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"stat.ML","authors_text":"Qingyang Ren, Thorsten Joachims, Yuta Saito","submitted_at":"2023-05-14T04:16:40Z","abstract_excerpt":"We study off-policy evaluation (OPE) of contextual bandit policies for large discrete action spaces where conventional importance-weighting approaches suffer from excessive variance. To circumvent this variance issue, we propose a new estimator, called OffCEM, that is based on the conjunct effect model (CEM), a novel decomposition of the causal effect into a cluster effect and a residual effect. OffCEM applies importance weighting only to action clusters and addresses the residual causal effect through model-based reward estimation. We show that the proposed estimator is unbiased under a new c"},"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":"2305.08062","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"stat.ML","submitted_at":"2023-05-14T04:16:40Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"bc229fb6d6a4365cbe9bdfc62d8270a09346e1c452110484dcccc12a332606fc","abstract_canon_sha256":"90d6ed256772e7bb2e7fa2f1ae0c7d54362c62a47437e19e55f37308b3fdf45d"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T06:19:04.157293Z","signature_b64":"qtKaDbTUd3hyd9THfNdACgaS6MArJSntdGbI0iWaD9QWbwjbzIS4RhX9OTAEDXKfxY2kYAe2NGD8yONV8AOaCQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"2aaa608e2fb6ab6de336fda31bc9d77b2929caac791746463ea41796035349a2","last_reissued_at":"2026-07-05T06:19:04.156811Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T06:19:04.156811Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Off-Policy Evaluation for Large Action Spaces via Conjunct Effect Modeling","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"stat.ML","authors_text":"Qingyang Ren, Thorsten Joachims, Yuta Saito","submitted_at":"2023-05-14T04:16:40Z","abstract_excerpt":"We study off-policy evaluation (OPE) of contextual bandit policies for large discrete action spaces where conventional importance-weighting approaches suffer from excessive variance. To circumvent this variance issue, we propose a new estimator, called OffCEM, that is based on the conjunct effect model (CEM), a novel decomposition of the causal effect into a cluster effect and a residual effect. OffCEM applies importance weighting only to action clusters and addresses the residual causal effect through model-based reward estimation. We show that the proposed estimator is unbiased under a new c"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2305.08062","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/2305.08062/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":"2305.08062","created_at":"2026-07-05T06:19:04.156868+00:00"},{"alias_kind":"arxiv_version","alias_value":"2305.08062v2","created_at":"2026-07-05T06:19:04.156868+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2305.08062","created_at":"2026-07-05T06:19:04.156868+00:00"},{"alias_kind":"pith_short_12","alias_value":"FKVGBDRPW2VW","created_at":"2026-07-05T06:19:04.156868+00:00"},{"alias_kind":"pith_short_16","alias_value":"FKVGBDRPW2VW3YZW","created_at":"2026-07-05T06:19:04.156868+00:00"},{"alias_kind":"pith_short_8","alias_value":"FKVGBDRP","created_at":"2026-07-05T06:19:04.156868+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":3,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2603.18702","citing_title":"Off-Policy Learning with Limited Supply","ref_index":24,"is_internal_anchor":false},{"citing_arxiv_id":"2605.18509","citing_title":"Offline Contextual Bandits in the Presence of New Actions","ref_index":29,"is_internal_anchor":false},{"citing_arxiv_id":"2603.18702","citing_title":"Off-Policy Learning with Limited Supply","ref_index":24,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/FKVGBDRPW2VW3YZW7WRRXSOXPM","json":"https://pith.science/pith/FKVGBDRPW2VW3YZW7WRRXSOXPM.json","graph_json":"https://pith.science/api/pith-number/FKVGBDRPW2VW3YZW7WRRXSOXPM/graph.json","events_json":"https://pith.science/api/pith-number/FKVGBDRPW2VW3YZW7WRRXSOXPM/events.json","paper":"https://pith.science/paper/FKVGBDRP"},"agent_actions":{"view_html":"https://pith.science/pith/FKVGBDRPW2VW3YZW7WRRXSOXPM","download_json":"https://pith.science/pith/FKVGBDRPW2VW3YZW7WRRXSOXPM.json","view_paper":"https://pith.science/paper/FKVGBDRP","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2305.08062&json=true","fetch_graph":"https://pith.science/api/pith-number/FKVGBDRPW2VW3YZW7WRRXSOXPM/graph.json","fetch_events":"https://pith.science/api/pith-number/FKVGBDRPW2VW3YZW7WRRXSOXPM/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/FKVGBDRPW2VW3YZW7WRRXSOXPM/action/timestamp_anchor","attest_storage":"https://pith.science/pith/FKVGBDRPW2VW3YZW7WRRXSOXPM/action/storage_attestation","attest_author":"https://pith.science/pith/FKVGBDRPW2VW3YZW7WRRXSOXPM/action/author_attestation","sign_citation":"https://pith.science/pith/FKVGBDRPW2VW3YZW7WRRXSOXPM/action/citation_signature","submit_replication":"https://pith.science/pith/FKVGBDRPW2VW3YZW7WRRXSOXPM/action/replication_record"}},"created_at":"2026-07-05T06:19:04.156868+00:00","updated_at":"2026-07-05T06:19:04.156868+00:00"}