{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:VS44BEYEV4PPBYXOI7ZJILFYNN","short_pith_number":"pith:VS44BEYE","schema_version":"1.0","canonical_sha256":"acb9c09304af1ef0e2ee47f2942cb86b40ab408926fbe3a537209bddc4737e99","source":{"kind":"arxiv","id":"2605.14840","version":1},"attestation_state":"computed","paper":{"title":"In-Context Learning for Data-Driven Censored Inventory Control","license":"http://creativecommons.org/licenses/by/4.0/","headline":"ICGPS combines offline meta-trained generative models with online in-context autoregressive generation to bound Bayesian regret in decision-dependent censored inventory control and outperforms baselines under mismatch.","cross_cats":["math.OC","stat.ML"],"primary_cat":"cs.LG","authors_text":"Anh-Duy Pham, Richard Pibernik, Sohom Mukherjee, Yunbei Xu","submitted_at":"2026-05-14T13:45:20Z","abstract_excerpt":"We study inventory control with decision-dependent censoring, focusing on the censored or repeated newsvendor (R-NV), where each order quantity determines whether demand is fully observed or censored by sales. Existing approaches based on parametric Thompson sampling (TS) can be brittle under prior mismatch, while offline imputation methods need not transfer to online learning. Motivated by the predictive view of decision making, we combine these ideas by taking oracle actions on learned completions of latent demand. We propose in-context generative posterior sampling (ICGPS), which uses moder"},"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":"2605.14840","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-14T13:45:20Z","cross_cats_sorted":["math.OC","stat.ML"],"title_canon_sha256":"6d72e9ef78cd087f0c9fb14e1a40185b653dec7e737abcd5ee2b4ffc036c8849","abstract_canon_sha256":"5c1d7d50944aabb311f9e636ba38db31385a58fae82848ea78c2583d896f96c0"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-17T23:38:56.423112Z","signature_b64":"kY9/qwfhk8DZaA6vIq1FQFuu/a2c000DnaHysSytqM2PKN3o3goiY8HuZ+KJbTdFYMwgNSB2nraKAQL9ocMvDA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"acb9c09304af1ef0e2ee47f2942cb86b40ab408926fbe3a537209bddc4737e99","last_reissued_at":"2026-05-17T23:38:56.422398Z","signature_status":"signed_v1","first_computed_at":"2026-05-17T23:38:56.422398Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"In-Context Learning for Data-Driven Censored Inventory Control","license":"http://creativecommons.org/licenses/by/4.0/","headline":"ICGPS combines offline meta-trained generative models with online in-context autoregressive generation to bound Bayesian regret in decision-dependent censored inventory control and outperforms baselines under mismatch.","cross_cats":["math.OC","stat.ML"],"primary_cat":"cs.LG","authors_text":"Anh-Duy Pham, Richard Pibernik, Sohom Mukherjee, Yunbei Xu","submitted_at":"2026-05-14T13:45:20Z","abstract_excerpt":"We study inventory control with decision-dependent censoring, focusing on the censored or repeated newsvendor (R-NV), where each order quantity determines whether demand is fully observed or censored by sales. Existing approaches based on parametric Thompson sampling (TS) can be brittle under prior mismatch, while offline imputation methods need not transfer to online learning. Motivated by the predictive view of decision making, we combine these ideas by taking oracle actions on learned completions of latent demand. We propose in-context generative posterior sampling (ICGPS), which uses moder"},"claims":{"count":3,"items":[{"kind":"strongest_claim","text":"The Bayesian regret of ICGPS with a learned completion kernel is bounded by the Bayesian regret of a TS benchmark with the ideal completion kernel plus a deployment penalty scaling as √T times the square root of the completion mismatch; for R-NV this yields sublinear Bayesian regret by reduction to bandit convex optimization feedback.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"Under reasonable coverage and stability assumptions, the online completion mismatch is controlled by the offline censored predictive mismatch so that offline predictive quality transfers to online performance.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"ICGPS combines offline meta-trained generative models with online in-context autoregressive generation to bound Bayesian regret in decision-dependent censored inventory control and outperforms baselines under mismatch.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"}],"snapshot_sha256":"e4483846c159ea17de4188c7f9af05e9ba5c7e3423e4f8f880be97d3d9c6c5df"},"source":{"id":"2605.14840","kind":"arxiv","version":1},"verdict":{"id":"4fb5be43-ee8a-4ae9-9727-851337845dd6","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-15T03:17:30.720390Z","strongest_claim":"The Bayesian regret of ICGPS with a learned completion kernel is bounded by the Bayesian regret of a TS benchmark with the ideal completion kernel plus a deployment penalty scaling as √T times the square root of the completion mismatch; for R-NV this yields sublinear Bayesian regret by reduction to bandit convex optimization feedback.","one_line_summary":"ICGPS combines offline meta-trained generative models with online in-context autoregressive generation to bound Bayesian regret in decision-dependent censored inventory control and outperforms baselines under mismatch.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"Under reasonable coverage and stability assumptions, the online completion mismatch is controlled by the offline censored predictive mismatch so that offline predictive quality transfers to online performance.","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":"2605.14840","created_at":"2026-05-17T23:38:56.422508+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.14840v1","created_at":"2026-05-17T23:38:56.422508+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.14840","created_at":"2026-05-17T23:38:56.422508+00:00"},{"alias_kind":"pith_short_12","alias_value":"VS44BEYEV4PP","created_at":"2026-05-18T12:33:37.589309+00:00"},{"alias_kind":"pith_short_16","alias_value":"VS44BEYEV4PPBYXO","created_at":"2026-05-18T12:33:37.589309+00:00"},{"alias_kind":"pith_short_8","alias_value":"VS44BEYE","created_at":"2026-05-18T12:33:37.589309+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/VS44BEYEV4PPBYXOI7ZJILFYNN","json":"https://pith.science/pith/VS44BEYEV4PPBYXOI7ZJILFYNN.json","graph_json":"https://pith.science/api/pith-number/VS44BEYEV4PPBYXOI7ZJILFYNN/graph.json","events_json":"https://pith.science/api/pith-number/VS44BEYEV4PPBYXOI7ZJILFYNN/events.json","paper":"https://pith.science/paper/VS44BEYE"},"agent_actions":{"view_html":"https://pith.science/pith/VS44BEYEV4PPBYXOI7ZJILFYNN","download_json":"https://pith.science/pith/VS44BEYEV4PPBYXOI7ZJILFYNN.json","view_paper":"https://pith.science/paper/VS44BEYE","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.14840&json=true","fetch_graph":"https://pith.science/api/pith-number/VS44BEYEV4PPBYXOI7ZJILFYNN/graph.json","fetch_events":"https://pith.science/api/pith-number/VS44BEYEV4PPBYXOI7ZJILFYNN/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/VS44BEYEV4PPBYXOI7ZJILFYNN/action/timestamp_anchor","attest_storage":"https://pith.science/pith/VS44BEYEV4PPBYXOI7ZJILFYNN/action/storage_attestation","attest_author":"https://pith.science/pith/VS44BEYEV4PPBYXOI7ZJILFYNN/action/author_attestation","sign_citation":"https://pith.science/pith/VS44BEYEV4PPBYXOI7ZJILFYNN/action/citation_signature","submit_replication":"https://pith.science/pith/VS44BEYEV4PPBYXOI7ZJILFYNN/action/replication_record"}},"created_at":"2026-05-17T23:38:56.422508+00:00","updated_at":"2026-05-17T23:38:56.422508+00:00"}