{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2022:GCVKFP5SWJTKFCDDYL2I5KO6AJ","short_pith_number":"pith:GCVKFP5S","schema_version":"1.0","canonical_sha256":"30aaa2bfb2b266a28863c2f48ea9de024a42db040fab2e6cdd4c8b85f2ff2128","source":{"kind":"arxiv","id":"2202.07477","version":2},"attestation_state":"computed","paper":{"title":"Understanding DDPM Latent Codes Through Optimal Transport","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.LG","cs.NA","math.AP","math.NA"],"primary_cat":"stat.ML","authors_text":"Andrei Chertkov, Gleb Ryzhakov, Ivan Oseledets, Valentin Khrulkov","submitted_at":"2022-02-14T18:59:47Z","abstract_excerpt":"Diffusion models have recently outperformed alternative approaches to model the distribution of natural images, such as GANs. Such diffusion models allow for deterministic sampling via the probability flow ODE, giving rise to a latent space and an encoder map. While having important practical applications, such as estimation of the likelihood, the theoretical properties of this map are not yet fully understood. In the present work, we partially address this question for the popular case of the VP SDE (DDPM) approach. We show that, perhaps surprisingly, the DDPM encoder map coincides with the o"},"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":"2202.07477","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"stat.ML","submitted_at":"2022-02-14T18:59:47Z","cross_cats_sorted":["cs.AI","cs.LG","cs.NA","math.AP","math.NA"],"title_canon_sha256":"8cfe9cb736859612374c8e931ebac6c6079b3413900b3987a779c279eb9c299f","abstract_canon_sha256":"db4e21db00f44b3a9d0d6c0ce35f8a152ccc92c04bb4a415ae20aa60a8b58ed9"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T05:22:06.152723Z","signature_b64":"bVL6zmomAyX/UY0QPPYyDcUCGk/kXwbrlYfOzaRgFZm63PAwdHt269/EpEETM6hnRdYKtSxDK9skrgDBHjMBBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"30aaa2bfb2b266a28863c2f48ea9de024a42db040fab2e6cdd4c8b85f2ff2128","last_reissued_at":"2026-07-05T05:22:06.152287Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T05:22:06.152287Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Understanding DDPM Latent Codes Through Optimal Transport","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.LG","cs.NA","math.AP","math.NA"],"primary_cat":"stat.ML","authors_text":"Andrei Chertkov, Gleb Ryzhakov, Ivan Oseledets, Valentin Khrulkov","submitted_at":"2022-02-14T18:59:47Z","abstract_excerpt":"Diffusion models have recently outperformed alternative approaches to model the distribution of natural images, such as GANs. Such diffusion models allow for deterministic sampling via the probability flow ODE, giving rise to a latent space and an encoder map. While having important practical applications, such as estimation of the likelihood, the theoretical properties of this map are not yet fully understood. In the present work, we partially address this question for the popular case of the VP SDE (DDPM) approach. We show that, perhaps surprisingly, the DDPM encoder map coincides with the o"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2202.07477","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/2202.07477/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":"2202.07477","created_at":"2026-07-05T05:22:06.152355+00:00"},{"alias_kind":"arxiv_version","alias_value":"2202.07477v2","created_at":"2026-07-05T05:22:06.152355+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2202.07477","created_at":"2026-07-05T05:22:06.152355+00:00"},{"alias_kind":"pith_short_12","alias_value":"GCVKFP5SWJTK","created_at":"2026-07-05T05:22:06.152355+00:00"},{"alias_kind":"pith_short_16","alias_value":"GCVKFP5SWJTKFCDD","created_at":"2026-07-05T05:22:06.152355+00:00"},{"alias_kind":"pith_short_8","alias_value":"GCVKFP5S","created_at":"2026-07-05T05:22:06.152355+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":4,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.23489","citing_title":"MeshFlow: Mesh Generation with Equivariant Flow Matching","ref_index":117,"is_internal_anchor":false},{"citing_arxiv_id":"2305.14703","citing_title":"Generative diffusion learning for parametric partial differential equations","ref_index":10,"is_internal_anchor":false},{"citing_arxiv_id":"2209.14577","citing_title":"Rectified Flow: A Marginal Preserving Approach to Optimal Transport","ref_index":72,"is_internal_anchor":false},{"citing_arxiv_id":"2209.03003","citing_title":"Flow Straight and Fast: Learning to Generate and Transfer Data with Rectified Flow","ref_index":30,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/GCVKFP5SWJTKFCDDYL2I5KO6AJ","json":"https://pith.science/pith/GCVKFP5SWJTKFCDDYL2I5KO6AJ.json","graph_json":"https://pith.science/api/pith-number/GCVKFP5SWJTKFCDDYL2I5KO6AJ/graph.json","events_json":"https://pith.science/api/pith-number/GCVKFP5SWJTKFCDDYL2I5KO6AJ/events.json","paper":"https://pith.science/paper/GCVKFP5S"},"agent_actions":{"view_html":"https://pith.science/pith/GCVKFP5SWJTKFCDDYL2I5KO6AJ","download_json":"https://pith.science/pith/GCVKFP5SWJTKFCDDYL2I5KO6AJ.json","view_paper":"https://pith.science/paper/GCVKFP5S","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2202.07477&json=true","fetch_graph":"https://pith.science/api/pith-number/GCVKFP5SWJTKFCDDYL2I5KO6AJ/graph.json","fetch_events":"https://pith.science/api/pith-number/GCVKFP5SWJTKFCDDYL2I5KO6AJ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/GCVKFP5SWJTKFCDDYL2I5KO6AJ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/GCVKFP5SWJTKFCDDYL2I5KO6AJ/action/storage_attestation","attest_author":"https://pith.science/pith/GCVKFP5SWJTKFCDDYL2I5KO6AJ/action/author_attestation","sign_citation":"https://pith.science/pith/GCVKFP5SWJTKFCDDYL2I5KO6AJ/action/citation_signature","submit_replication":"https://pith.science/pith/GCVKFP5SWJTKFCDDYL2I5KO6AJ/action/replication_record"}},"created_at":"2026-07-05T05:22:06.152355+00:00","updated_at":"2026-07-05T05:22:06.152355+00:00"}