{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:QAFZJGFFUSPQ7ASLTMB7NMXLKW","short_pith_number":"pith:QAFZJGFF","schema_version":"1.0","canonical_sha256":"800b9498a5a49f0f824b9b03f6b2eb55a103c07426f10d0f46c7603fbe5fa4bc","source":{"kind":"arxiv","id":"2405.20971","version":2},"attestation_state":"computed","paper":{"title":"Amortizing intractable inference in diffusion models for vision, language, and control","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CV"],"primary_cat":"cs.LG","authors_text":"Alexandre Adam, Emmanuel Bengio, Glen Berseth, Jarrid Rector-Brooks, Luca Scimeca, Luke Rowe, Marcin Sendera, Minsu Kim, Mohsin Hasan, Moksh Jain, Nikolay Malkin, Pablo Lemos, Sarthak Mittal, Siddarth Venkatraman, Yoshua Bengio","submitted_at":"2024-05-31T16:18:46Z","abstract_excerpt":"Diffusion models have emerged as effective distribution estimators in vision, language, and reinforcement learning, but their use as priors in downstream tasks poses an intractable posterior inference problem. This paper studies amortized sampling of the posterior over data, $\\mathbf{x}\\sim p^{\\rm post}(\\mathbf{x})\\propto p(\\mathbf{x})r(\\mathbf{x})$, in a model that consists of a diffusion generative model prior $p(\\mathbf{x})$ and a black-box constraint or likelihood function $r(\\mathbf{x})$. We state and prove the asymptotic correctness of a data-free learning objective, relative trajectory "},"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":"2405.20971","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-05-31T16:18:46Z","cross_cats_sorted":["cs.CV"],"title_canon_sha256":"cc9dc20afe3aef4f89ab832986976681daaec449e3586538be4f5b9dbb0bed55","abstract_canon_sha256":"9b00b5f7c3157672f03abc772bedb98c8a27a9216382877f646550988f21f2b3"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:59:47.951885Z","signature_b64":"u7UD8xMxgoSUSgVgwt+5Qqe3xX8vxpeSIBu529oaXFpssvWjgX985GmxekEiusynXdJhiW2bt/A/FtMokZ+rCQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"800b9498a5a49f0f824b9b03f6b2eb55a103c07426f10d0f46c7603fbe5fa4bc","last_reissued_at":"2026-07-05T09:59:47.951386Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:59:47.951386Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Amortizing intractable inference in diffusion models for vision, language, and control","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CV"],"primary_cat":"cs.LG","authors_text":"Alexandre Adam, Emmanuel Bengio, Glen Berseth, Jarrid Rector-Brooks, Luca Scimeca, Luke Rowe, Marcin Sendera, Minsu Kim, Mohsin Hasan, Moksh Jain, Nikolay Malkin, Pablo Lemos, Sarthak Mittal, Siddarth Venkatraman, Yoshua Bengio","submitted_at":"2024-05-31T16:18:46Z","abstract_excerpt":"Diffusion models have emerged as effective distribution estimators in vision, language, and reinforcement learning, but their use as priors in downstream tasks poses an intractable posterior inference problem. This paper studies amortized sampling of the posterior over data, $\\mathbf{x}\\sim p^{\\rm post}(\\mathbf{x})\\propto p(\\mathbf{x})r(\\mathbf{x})$, in a model that consists of a diffusion generative model prior $p(\\mathbf{x})$ and a black-box constraint or likelihood function $r(\\mathbf{x})$. We state and prove the asymptotic correctness of a data-free learning objective, relative trajectory "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2405.20971","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/2405.20971/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":"2405.20971","created_at":"2026-07-05T09:59:47.951451+00:00"},{"alias_kind":"arxiv_version","alias_value":"2405.20971v2","created_at":"2026-07-05T09:59:47.951451+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2405.20971","created_at":"2026-07-05T09:59:47.951451+00:00"},{"alias_kind":"pith_short_12","alias_value":"QAFZJGFFUSPQ","created_at":"2026-07-05T09:59:47.951451+00:00"},{"alias_kind":"pith_short_16","alias_value":"QAFZJGFFUSPQ7ASL","created_at":"2026-07-05T09:59:47.951451+00:00"},{"alias_kind":"pith_short_8","alias_value":"QAFZJGFF","created_at":"2026-07-05T09:59:47.951451+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2410.17891","citing_title":"Scaling Diffusion Language Models via Adaptation from Autoregressive Models","ref_index":187,"is_internal_anchor":false},{"citing_arxiv_id":"2604.24104","citing_title":"Factual and Edit-Sensitive Graph-to-Sequence Generation via Graph-Aware Adaptive Noising","ref_index":16,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/QAFZJGFFUSPQ7ASLTMB7NMXLKW","json":"https://pith.science/pith/QAFZJGFFUSPQ7ASLTMB7NMXLKW.json","graph_json":"https://pith.science/api/pith-number/QAFZJGFFUSPQ7ASLTMB7NMXLKW/graph.json","events_json":"https://pith.science/api/pith-number/QAFZJGFFUSPQ7ASLTMB7NMXLKW/events.json","paper":"https://pith.science/paper/QAFZJGFF"},"agent_actions":{"view_html":"https://pith.science/pith/QAFZJGFFUSPQ7ASLTMB7NMXLKW","download_json":"https://pith.science/pith/QAFZJGFFUSPQ7ASLTMB7NMXLKW.json","view_paper":"https://pith.science/paper/QAFZJGFF","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2405.20971&json=true","fetch_graph":"https://pith.science/api/pith-number/QAFZJGFFUSPQ7ASLTMB7NMXLKW/graph.json","fetch_events":"https://pith.science/api/pith-number/QAFZJGFFUSPQ7ASLTMB7NMXLKW/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/QAFZJGFFUSPQ7ASLTMB7NMXLKW/action/timestamp_anchor","attest_storage":"https://pith.science/pith/QAFZJGFFUSPQ7ASLTMB7NMXLKW/action/storage_attestation","attest_author":"https://pith.science/pith/QAFZJGFFUSPQ7ASLTMB7NMXLKW/action/author_attestation","sign_citation":"https://pith.science/pith/QAFZJGFFUSPQ7ASLTMB7NMXLKW/action/citation_signature","submit_replication":"https://pith.science/pith/QAFZJGFFUSPQ7ASLTMB7NMXLKW/action/replication_record"}},"created_at":"2026-07-05T09:59:47.951451+00:00","updated_at":"2026-07-05T09:59:47.951451+00:00"}