{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:BVZVM4KF34UFZWLTDVRXF7KZHY","short_pith_number":"pith:BVZVM4KF","schema_version":"1.0","canonical_sha256":"0d73567145df285cd9731d6372fd593e24f3fc1bb5570b30b7dcdfda1505b945","source":{"kind":"arxiv","id":"2301.07882","version":4},"attestation_state":"computed","paper":{"title":"Mathematical analysis of singularities in the diffusion model under the submanifold assumption","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.NA","math.NA","stat.ML"],"primary_cat":"cs.LG","authors_text":"Guillaume Bal, Yubin Lu, Zhongjian Wang","submitted_at":"2023-01-19T05:13:03Z","abstract_excerpt":"This paper concerns the mathematical analyses of the diffusion model in machine learning. The drift term of the backward sampling process is represented as a conditional expectation involving the data distribution and the forward diffusion. The training process aims to find such a drift function by minimizing the mean-squared residue related to the conditional expectation. Using small-time approximations of the Green's function of the forward diffusion, we show that the analytical mean drift function in DDPM and the score function in SGM asymptotically blow up in the final stages of the sampli"},"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":"2301.07882","kind":"arxiv","version":4},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2023-01-19T05:13:03Z","cross_cats_sorted":["cs.NA","math.NA","stat.ML"],"title_canon_sha256":"6a97e80462880229046a15531303e6a4c6d04335c1c9f22de725986272b20992","abstract_canon_sha256":"cb532498571dba8a266d84df2eb797d9469cfbce03d3805cd9a0843124681600"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:46:40.929058Z","signature_b64":"aXIgah8skeS6Ga5syO/pPCIPcqVncF/yq4YEz6Usgl0n59z7mINACCg4yOAoRYj0eBb5YQUd3k0/cLEy0xPrBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"0d73567145df285cd9731d6372fd593e24f3fc1bb5570b30b7dcdfda1505b945","last_reissued_at":"2026-07-05T09:46:40.927744Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:46:40.927744Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Mathematical analysis of singularities in the diffusion model under the submanifold assumption","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.NA","math.NA","stat.ML"],"primary_cat":"cs.LG","authors_text":"Guillaume Bal, Yubin Lu, Zhongjian Wang","submitted_at":"2023-01-19T05:13:03Z","abstract_excerpt":"This paper concerns the mathematical analyses of the diffusion model in machine learning. The drift term of the backward sampling process is represented as a conditional expectation involving the data distribution and the forward diffusion. The training process aims to find such a drift function by minimizing the mean-squared residue related to the conditional expectation. Using small-time approximations of the Green's function of the forward diffusion, we show that the analytical mean drift function in DDPM and the score function in SGM asymptotically blow up in the final stages of the sampli"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2301.07882","kind":"arxiv","version":4},"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/2301.07882/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":"2301.07882","created_at":"2026-07-05T09:46:40.927810+00:00"},{"alias_kind":"arxiv_version","alias_value":"2301.07882v4","created_at":"2026-07-05T09:46:40.927810+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2301.07882","created_at":"2026-07-05T09:46:40.927810+00:00"},{"alias_kind":"pith_short_12","alias_value":"BVZVM4KF34UF","created_at":"2026-07-05T09:46:40.927810+00:00"},{"alias_kind":"pith_short_16","alias_value":"BVZVM4KF34UFZWLT","created_at":"2026-07-05T09:46:40.927810+00:00"},{"alias_kind":"pith_short_8","alias_value":"BVZVM4KF","created_at":"2026-07-05T09:46:40.927810+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.15822","citing_title":"Intrinsic Wasserstein Rates for Score-Based Generative Models on Smooth Manifolds","ref_index":30,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/BVZVM4KF34UFZWLTDVRXF7KZHY","json":"https://pith.science/pith/BVZVM4KF34UFZWLTDVRXF7KZHY.json","graph_json":"https://pith.science/api/pith-number/BVZVM4KF34UFZWLTDVRXF7KZHY/graph.json","events_json":"https://pith.science/api/pith-number/BVZVM4KF34UFZWLTDVRXF7KZHY/events.json","paper":"https://pith.science/paper/BVZVM4KF"},"agent_actions":{"view_html":"https://pith.science/pith/BVZVM4KF34UFZWLTDVRXF7KZHY","download_json":"https://pith.science/pith/BVZVM4KF34UFZWLTDVRXF7KZHY.json","view_paper":"https://pith.science/paper/BVZVM4KF","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2301.07882&json=true","fetch_graph":"https://pith.science/api/pith-number/BVZVM4KF34UFZWLTDVRXF7KZHY/graph.json","fetch_events":"https://pith.science/api/pith-number/BVZVM4KF34UFZWLTDVRXF7KZHY/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/BVZVM4KF34UFZWLTDVRXF7KZHY/action/timestamp_anchor","attest_storage":"https://pith.science/pith/BVZVM4KF34UFZWLTDVRXF7KZHY/action/storage_attestation","attest_author":"https://pith.science/pith/BVZVM4KF34UFZWLTDVRXF7KZHY/action/author_attestation","sign_citation":"https://pith.science/pith/BVZVM4KF34UFZWLTDVRXF7KZHY/action/citation_signature","submit_replication":"https://pith.science/pith/BVZVM4KF34UFZWLTDVRXF7KZHY/action/replication_record"}},"created_at":"2026-07-05T09:46:40.927810+00:00","updated_at":"2026-07-05T09:46:40.927810+00:00"}