{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:J3A7IKJKHEUFOMXUPAGIIEDYET","short_pith_number":"pith:J3A7IKJK","schema_version":"1.0","canonical_sha256":"4ec1f4292a39285732f4780c84107824c99a4e742ada7d13c13b22d91c9488a0","source":{"kind":"arxiv","id":"2604.04342","version":2},"attestation_state":"computed","paper":{"title":"Generative models for decision-making under distributional shift","license":"http://creativecommons.org/licenses/by/4.0/","headline":"Generative models construct nominal, stressed, and conditional distributions for decisions under shift using transport maps and guided dynamics.","cross_cats":["stat.ML"],"primary_cat":"cs.LG","authors_text":"Xiuyuan Cheng, Yao Xie, Yunqin Zhu","submitted_at":"2026-04-06T01:35:13Z","abstract_excerpt":"Many data-driven decision problems are formulated using a nominal distribution estimated from historical data, while performance is ultimately determined by a deployment distribution that may be shifted, context-dependent, partially observed, or stress-induced. This tutorial presents modern generative models, particularly flow- and score-based methods, as mathematical tools for constructing decision-relevant distributions. From an operations research perspective, their primary value lies not in unconstrained sample synthesis but in representing and transforming distributions through transport "},"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":true},"canonical_record":{"source":{"id":"2604.04342","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-04-06T01:35:13Z","cross_cats_sorted":["stat.ML"],"title_canon_sha256":"f65cd7671df6949ddabd80f2af4a481e2f6b1e84fbb073c726694df3533639e8","abstract_canon_sha256":"c6b15e0abe93afeb2a9a0ad0bbb157002db00a27490974ca3c93813d11c33a6c"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-19T16:11:22.853307Z","signature_b64":"ms07GbFomaA3cZOU+R9RJrJQI6av+e/tgt8VHVdpH32mVkyhT8QbhVYP6vuQf6ZuY2ZfyWOuwRAH/r+0xzejCA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"4ec1f4292a39285732f4780c84107824c99a4e742ada7d13c13b22d91c9488a0","last_reissued_at":"2026-06-19T16:11:22.852952Z","signature_status":"signed_v1","first_computed_at":"2026-06-19T16:11:22.852952Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Generative models for decision-making under distributional shift","license":"http://creativecommons.org/licenses/by/4.0/","headline":"Generative models construct nominal, stressed, and conditional distributions for decisions under shift using transport maps and guided dynamics.","cross_cats":["stat.ML"],"primary_cat":"cs.LG","authors_text":"Xiuyuan Cheng, Yao Xie, Yunqin Zhu","submitted_at":"2026-04-06T01:35:13Z","abstract_excerpt":"Many data-driven decision problems are formulated using a nominal distribution estimated from historical data, while performance is ultimately determined by a deployment distribution that may be shifted, context-dependent, partially observed, or stress-induced. This tutorial presents modern generative models, particularly flow- and score-based methods, as mathematical tools for constructing decision-relevant distributions. From an operations research perspective, their primary value lies not in unconstrained sample synthesis but in representing and transforming distributions through transport "},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"Generative models can be used to learn nominal uncertainty, construct stressed or least-favorable distributions for robustness, and produce conditional or posterior distributions under side information and partial observation, within a unified framework based on pushforward maps, continuity, Fokker-Planck equations, Wasserstein geometry, and optimization in probability space.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That the mathematical tools (transport maps, score fields, guided stochastic dynamics) can be trained and deployed in a way that reliably produces decision-relevant distributions whose properties transfer to the actual deployment distribution under shift.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"Generative models via pushforward maps, Fokker-Planck equations, and Wasserstein geometry enable learning nominal uncertainty, stressed distributions for robustness, and conditional posteriors under distributional shift.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Generative models construct nominal, stressed, and conditional distributions for decisions under shift using transport maps and guided dynamics.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"6a8532b084197954d4520c582d2128644a04dd013b87f55c19a338c26b5ef8a6"},"source":{"id":"2604.04342","kind":"arxiv","version":2},"verdict":{"id":"6592cbfb-ce6e-4c96-bf8a-cf3a32741ab0","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-10T20:24:13.117908Z","strongest_claim":"Generative models can be used to learn nominal uncertainty, construct stressed or least-favorable distributions for robustness, and produce conditional or posterior distributions under side information and partial observation, within a unified framework based on pushforward maps, continuity, Fokker-Planck equations, Wasserstein geometry, and optimization in probability space.","one_line_summary":"Generative models via pushforward maps, Fokker-Planck equations, and Wasserstein geometry enable learning nominal uncertainty, stressed distributions for robustness, and conditional posteriors under distributional shift.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That the mathematical tools (transport maps, score fields, guided stochastic dynamics) can be trained and deployed in a way that reliably produces decision-relevant distributions whose properties transfer to the actual deployment distribution under shift.","pith_extraction_headline":"Generative models construct nominal, stressed, and conditional distributions for decisions under shift using transport maps and guided dynamics."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2604.04342/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":2,"snapshot_sha256":"6f36de3cc61172c565129c2cd6493e14f08b376e87cb1293db1cc55dce4e6494"},"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":"2604.04342","created_at":"2026-06-19T16:11:22.853010+00:00"},{"alias_kind":"arxiv_version","alias_value":"2604.04342v2","created_at":"2026-06-19T16:11:22.853010+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2604.04342","created_at":"2026-06-19T16:11:22.853010+00:00"},{"alias_kind":"pith_short_12","alias_value":"J3A7IKJKHEUF","created_at":"2026-06-19T16:11:22.853010+00:00"},{"alias_kind":"pith_short_16","alias_value":"J3A7IKJKHEUFOMXU","created_at":"2026-06-19T16:11:22.853010+00:00"},{"alias_kind":"pith_short_8","alias_value":"J3A7IKJK","created_at":"2026-06-19T16:11:22.853010+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2606.21773","citing_title":"Decision-Focused Learning: When and Why Traditional Prediction Models Fail","ref_index":10,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":2,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/J3A7IKJKHEUFOMXUPAGIIEDYET","json":"https://pith.science/pith/J3A7IKJKHEUFOMXUPAGIIEDYET.json","graph_json":"https://pith.science/api/pith-number/J3A7IKJKHEUFOMXUPAGIIEDYET/graph.json","events_json":"https://pith.science/api/pith-number/J3A7IKJKHEUFOMXUPAGIIEDYET/events.json","paper":"https://pith.science/paper/J3A7IKJK"},"agent_actions":{"view_html":"https://pith.science/pith/J3A7IKJKHEUFOMXUPAGIIEDYET","download_json":"https://pith.science/pith/J3A7IKJKHEUFOMXUPAGIIEDYET.json","view_paper":"https://pith.science/paper/J3A7IKJK","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2604.04342&json=true","fetch_graph":"https://pith.science/api/pith-number/J3A7IKJKHEUFOMXUPAGIIEDYET/graph.json","fetch_events":"https://pith.science/api/pith-number/J3A7IKJKHEUFOMXUPAGIIEDYET/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/J3A7IKJKHEUFOMXUPAGIIEDYET/action/timestamp_anchor","attest_storage":"https://pith.science/pith/J3A7IKJKHEUFOMXUPAGIIEDYET/action/storage_attestation","attest_author":"https://pith.science/pith/J3A7IKJKHEUFOMXUPAGIIEDYET/action/author_attestation","sign_citation":"https://pith.science/pith/J3A7IKJKHEUFOMXUPAGIIEDYET/action/citation_signature","submit_replication":"https://pith.science/pith/J3A7IKJKHEUFOMXUPAGIIEDYET/action/replication_record"}},"created_at":"2026-06-19T16:11:22.853010+00:00","updated_at":"2026-06-19T16:11:22.853010+00:00"}