{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:PVC3KZSSOBJDF7EHWBAX5PHHQB","short_pith_number":"pith:PVC3KZSS","schema_version":"1.0","canonical_sha256":"7d45b56652705232fc87b0417ebce7805833cf4aa052cf5324fd5d5077f00b07","source":{"kind":"arxiv","id":"2304.05907","version":1},"attestation_state":"computed","paper":{"title":"Diffusion models with location-scale noise","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.NA","math.NA"],"primary_cat":"cs.LG","authors_text":"Alexia Jolicoeur-Martineau, Ke Li, Kilian Fatras, Tal Kachman","submitted_at":"2023-04-12T15:24:33Z","abstract_excerpt":"Diffusion Models (DMs) are powerful generative models that add Gaussian noise to the data and learn to remove it. We wanted to determine which noise distribution (Gaussian or non-Gaussian) led to better generated data in DMs. Since DMs do not work by design with non-Gaussian noise, we built a framework that allows reversing a diffusion process with non-Gaussian location-scale noise. We use that framework to show that the Gaussian distribution performs the best over a wide range of other distributions (Laplace, Uniform, t, Generalized-Gaussian)."},"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":"2304.05907","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2023-04-12T15:24:33Z","cross_cats_sorted":["cs.AI","cs.NA","math.NA"],"title_canon_sha256":"e12a203f83953a6e886bd5292a6ded412d2bad494bb226fc5b750173166cb893","abstract_canon_sha256":"abd835dffa069f1125eeed6dc8fe0d0bc396d48d1b6479231a238fee34224109"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T06:00:25.548076Z","signature_b64":"YNz32L5snmP5LYIaq5cP4VDj97NqtNEirqPy2AebYZZjgvR7qzH6IrWz2K6XDVxdAe9gu8BEEg2WuFLYI6uSCA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"7d45b56652705232fc87b0417ebce7805833cf4aa052cf5324fd5d5077f00b07","last_reissued_at":"2026-07-05T06:00:25.547590Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T06:00:25.547590Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Diffusion models with location-scale noise","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.NA","math.NA"],"primary_cat":"cs.LG","authors_text":"Alexia Jolicoeur-Martineau, Ke Li, Kilian Fatras, Tal Kachman","submitted_at":"2023-04-12T15:24:33Z","abstract_excerpt":"Diffusion Models (DMs) are powerful generative models that add Gaussian noise to the data and learn to remove it. We wanted to determine which noise distribution (Gaussian or non-Gaussian) led to better generated data in DMs. Since DMs do not work by design with non-Gaussian noise, we built a framework that allows reversing a diffusion process with non-Gaussian location-scale noise. We use that framework to show that the Gaussian distribution performs the best over a wide range of other distributions (Laplace, Uniform, t, Generalized-Gaussian)."},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2304.05907","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/2304.05907/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":"2304.05907","created_at":"2026-07-05T06:00:25.547649+00:00"},{"alias_kind":"arxiv_version","alias_value":"2304.05907v1","created_at":"2026-07-05T06:00:25.547649+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2304.05907","created_at":"2026-07-05T06:00:25.547649+00:00"},{"alias_kind":"pith_short_12","alias_value":"PVC3KZSSOBJD","created_at":"2026-07-05T06:00:25.547649+00:00"},{"alias_kind":"pith_short_16","alias_value":"PVC3KZSSOBJDF7EH","created_at":"2026-07-05T06:00:25.547649+00:00"},{"alias_kind":"pith_short_8","alias_value":"PVC3KZSS","created_at":"2026-07-05T06:00:25.547649+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.19305","citing_title":"Mat\\'ern Noise for Triangulation-Agnostic Flow Matching on Meshes","ref_index":15,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/PVC3KZSSOBJDF7EHWBAX5PHHQB","json":"https://pith.science/pith/PVC3KZSSOBJDF7EHWBAX5PHHQB.json","graph_json":"https://pith.science/api/pith-number/PVC3KZSSOBJDF7EHWBAX5PHHQB/graph.json","events_json":"https://pith.science/api/pith-number/PVC3KZSSOBJDF7EHWBAX5PHHQB/events.json","paper":"https://pith.science/paper/PVC3KZSS"},"agent_actions":{"view_html":"https://pith.science/pith/PVC3KZSSOBJDF7EHWBAX5PHHQB","download_json":"https://pith.science/pith/PVC3KZSSOBJDF7EHWBAX5PHHQB.json","view_paper":"https://pith.science/paper/PVC3KZSS","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2304.05907&json=true","fetch_graph":"https://pith.science/api/pith-number/PVC3KZSSOBJDF7EHWBAX5PHHQB/graph.json","fetch_events":"https://pith.science/api/pith-number/PVC3KZSSOBJDF7EHWBAX5PHHQB/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/PVC3KZSSOBJDF7EHWBAX5PHHQB/action/timestamp_anchor","attest_storage":"https://pith.science/pith/PVC3KZSSOBJDF7EHWBAX5PHHQB/action/storage_attestation","attest_author":"https://pith.science/pith/PVC3KZSSOBJDF7EHWBAX5PHHQB/action/author_attestation","sign_citation":"https://pith.science/pith/PVC3KZSSOBJDF7EHWBAX5PHHQB/action/citation_signature","submit_replication":"https://pith.science/pith/PVC3KZSSOBJDF7EHWBAX5PHHQB/action/replication_record"}},"created_at":"2026-07-05T06:00:25.547649+00:00","updated_at":"2026-07-05T06:00:25.547649+00:00"}