{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:KBERR6JH2OLYSOJUQIXIRXRHEX","short_pith_number":"pith:KBERR6JH","schema_version":"1.0","canonical_sha256":"504918f927d397893934822e88de2725fe2d43d05e13ff90e10ada4a52e0af37","source":{"kind":"arxiv","id":"2608.11560","version":1},"attestation_state":"computed","paper":{"title":"When Offline Evaluation Misleads: A Diagnostic Protocol for Reward and Policy Selection in Delayed-Feedback Contextual Bandits","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Sang Su Lee, Shishir Dash, Vijay Raghavan, Vineeth Loganathan","submitted_at":"2026-08-12T01:53:21Z","abstract_excerpt":"Personalizing marketing messages with contextual multi-armed bandits (CMABs) drives real business value, yet the objective that ultimately matters - a downstream conversion - is observed only weeks later, too late to drive online learning. Teams therefore train the bandit on a fast proxy reward, and separately must judge whether a contextual bandit is worth its complexity over sending one best message. Settling both decisions with the usual offline checks - a batch off-policy estimate, a marginal arm-discrimination test, a confidence interval - can mislead systematically under delayed feedback"},"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":"2608.11560","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2026-08-12T01:53:21Z","cross_cats_sorted":[],"title_canon_sha256":"a661e1f485d255b9653a50c4687e2cfd45f5f90b7fd7ca5292c61b10ed212725","abstract_canon_sha256":"6fa933b02e6dce4cb177ba8dff14d7a68272d30f841b95f98011b38390000508"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-08-13T00:22:11.329810Z","signature_b64":"kPZisF1nDua9iupbxdoa4YIhRZFMDEFarWavfZN4aOzI64VWY8yhg3XglaRkkR0o+gVlpOkTobPPnrOHfgGjBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"504918f927d397893934822e88de2725fe2d43d05e13ff90e10ada4a52e0af37","last_reissued_at":"2026-08-13T00:22:11.328269Z","signature_status":"signed_v1","first_computed_at":"2026-08-13T00:22:11.328269Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"When Offline Evaluation Misleads: A Diagnostic Protocol for Reward and Policy Selection in Delayed-Feedback Contextual Bandits","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Sang Su Lee, Shishir Dash, Vijay Raghavan, Vineeth Loganathan","submitted_at":"2026-08-12T01:53:21Z","abstract_excerpt":"Personalizing marketing messages with contextual multi-armed bandits (CMABs) drives real business value, yet the objective that ultimately matters - a downstream conversion - is observed only weeks later, too late to drive online learning. Teams therefore train the bandit on a fast proxy reward, and separately must judge whether a contextual bandit is worth its complexity over sending one best message. Settling both decisions with the usual offline checks - a batch off-policy estimate, a marginal arm-discrimination test, a confidence interval - can mislead systematically under delayed feedback"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2608.11560","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/2608.11560/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":"2608.11560","created_at":"2026-08-13T00:22:11.328576+00:00"},{"alias_kind":"arxiv_version","alias_value":"2608.11560v1","created_at":"2026-08-13T00:22:11.328576+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2608.11560","created_at":"2026-08-13T00:22:11.328576+00:00"},{"alias_kind":"pith_short_12","alias_value":"KBERR6JH2OLY","created_at":"2026-08-13T00:22:11.328576+00:00"},{"alias_kind":"pith_short_16","alias_value":"KBERR6JH2OLYSOJU","created_at":"2026-08-13T00:22:11.328576+00:00"},{"alias_kind":"pith_short_8","alias_value":"KBERR6JH","created_at":"2026-08-13T00:22:11.328576+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/KBERR6JH2OLYSOJUQIXIRXRHEX","json":"https://pith.science/pith/KBERR6JH2OLYSOJUQIXIRXRHEX.json","graph_json":"https://pith.science/api/pith-number/KBERR6JH2OLYSOJUQIXIRXRHEX/graph.json","events_json":"https://pith.science/api/pith-number/KBERR6JH2OLYSOJUQIXIRXRHEX/events.json","paper":"https://pith.science/paper/KBERR6JH"},"agent_actions":{"view_html":"https://pith.science/pith/KBERR6JH2OLYSOJUQIXIRXRHEX","download_json":"https://pith.science/pith/KBERR6JH2OLYSOJUQIXIRXRHEX.json","view_paper":"https://pith.science/paper/KBERR6JH","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2608.11560&json=true","fetch_graph":"https://pith.science/api/pith-number/KBERR6JH2OLYSOJUQIXIRXRHEX/graph.json","fetch_events":"https://pith.science/api/pith-number/KBERR6JH2OLYSOJUQIXIRXRHEX/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/KBERR6JH2OLYSOJUQIXIRXRHEX/action/timestamp_anchor","attest_storage":"https://pith.science/pith/KBERR6JH2OLYSOJUQIXIRXRHEX/action/storage_attestation","attest_author":"https://pith.science/pith/KBERR6JH2OLYSOJUQIXIRXRHEX/action/author_attestation","sign_citation":"https://pith.science/pith/KBERR6JH2OLYSOJUQIXIRXRHEX/action/citation_signature","submit_replication":"https://pith.science/pith/KBERR6JH2OLYSOJUQIXIRXRHEX/action/replication_record"}},"created_at":"2026-08-13T00:22:11.328576+00:00","updated_at":"2026-08-13T00:22:11.328576+00:00"}