{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2022:AAHCCGZOUDULA2BKILZFZMTHYJ","short_pith_number":"pith:AAHCCGZO","schema_version":"1.0","canonical_sha256":"000e211b2ea0e8b0682a42f25cb267c241fb784d89ac9a55faf228dbc325a536","source":{"kind":"arxiv","id":"2212.12611","version":5},"attestation_state":"computed","paper":{"title":"Your diffusion model secretly knows the dimension of the data manifold","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["stat.ML"],"primary_cat":"cs.LG","authors_text":"Carola-Bibiane Sch\\\"onlieb, Georgios Batzolis, Jan Stanczuk, Teo Deveney","submitted_at":"2022-12-23T23:15:14Z","abstract_excerpt":"In this work, we propose a novel framework for estimating the dimension of the data manifold using a trained diffusion model. A diffusion model approximates the score function i.e. the gradient of the log density of a noise-corrupted version of the target distribution for varying levels of corruption. We prove that, if the data concentrates around a manifold embedded in the high-dimensional ambient space, then as the level of corruption decreases, the score function points towards the manifold, as this direction becomes the direction of maximal likelihood increase. Therefore, for small levels "},"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":"2212.12611","kind":"arxiv","version":5},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2022-12-23T23:15:14Z","cross_cats_sorted":["stat.ML"],"title_canon_sha256":"2a0a0b914bdbe711c4724c8fd5e0d28404ff36b5668e7d8866ca102aa3188813","abstract_canon_sha256":"e1a98565af50e583282b5629dc2c953a70bfaded982ab3a70205497d518164ab"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T06:13:54.343129Z","signature_b64":"B63h4XO8kC6DPRT9lN5BcLFHurolHk0iLvdgBXS/Qm+6LRT1RyrisWkFW216PMR8ZPIK4ysPGPCicp2af1b4CA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"000e211b2ea0e8b0682a42f25cb267c241fb784d89ac9a55faf228dbc325a536","last_reissued_at":"2026-07-05T06:13:54.342551Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T06:13:54.342551Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Your diffusion model secretly knows the dimension of the data manifold","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["stat.ML"],"primary_cat":"cs.LG","authors_text":"Carola-Bibiane Sch\\\"onlieb, Georgios Batzolis, Jan Stanczuk, Teo Deveney","submitted_at":"2022-12-23T23:15:14Z","abstract_excerpt":"In this work, we propose a novel framework for estimating the dimension of the data manifold using a trained diffusion model. A diffusion model approximates the score function i.e. the gradient of the log density of a noise-corrupted version of the target distribution for varying levels of corruption. We prove that, if the data concentrates around a manifold embedded in the high-dimensional ambient space, then as the level of corruption decreases, the score function points towards the manifold, as this direction becomes the direction of maximal likelihood increase. Therefore, for small levels "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2212.12611","kind":"arxiv","version":5},"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/2212.12611/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":"2212.12611","created_at":"2026-07-05T06:13:54.342626+00:00"},{"alias_kind":"arxiv_version","alias_value":"2212.12611v5","created_at":"2026-07-05T06:13:54.342626+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2212.12611","created_at":"2026-07-05T06:13:54.342626+00:00"},{"alias_kind":"pith_short_12","alias_value":"AAHCCGZOUDUL","created_at":"2026-07-05T06:13:54.342626+00:00"},{"alias_kind":"pith_short_16","alias_value":"AAHCCGZOUDULA2BK","created_at":"2026-07-05T06:13:54.342626+00:00"},{"alias_kind":"pith_short_8","alias_value":"AAHCCGZO","created_at":"2026-07-05T06:13:54.342626+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":3,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.14746","citing_title":"Style-CCL: Content-Preserving Style Transfer via Curriculum Continual Learning","ref_index":15,"is_internal_anchor":false},{"citing_arxiv_id":"2606.30370","citing_title":"MUSE: Unlocking Timestep as Native Task Steering for One-Step Dense Prediction","ref_index":39,"is_internal_anchor":false},{"citing_arxiv_id":"2604.25289","citing_title":"Exploring Time Conditioning in Diffusion Generative Models from Disjoint Noisy Data Manifolds","ref_index":74,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/AAHCCGZOUDULA2BKILZFZMTHYJ","json":"https://pith.science/pith/AAHCCGZOUDULA2BKILZFZMTHYJ.json","graph_json":"https://pith.science/api/pith-number/AAHCCGZOUDULA2BKILZFZMTHYJ/graph.json","events_json":"https://pith.science/api/pith-number/AAHCCGZOUDULA2BKILZFZMTHYJ/events.json","paper":"https://pith.science/paper/AAHCCGZO"},"agent_actions":{"view_html":"https://pith.science/pith/AAHCCGZOUDULA2BKILZFZMTHYJ","download_json":"https://pith.science/pith/AAHCCGZOUDULA2BKILZFZMTHYJ.json","view_paper":"https://pith.science/paper/AAHCCGZO","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2212.12611&json=true","fetch_graph":"https://pith.science/api/pith-number/AAHCCGZOUDULA2BKILZFZMTHYJ/graph.json","fetch_events":"https://pith.science/api/pith-number/AAHCCGZOUDULA2BKILZFZMTHYJ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/AAHCCGZOUDULA2BKILZFZMTHYJ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/AAHCCGZOUDULA2BKILZFZMTHYJ/action/storage_attestation","attest_author":"https://pith.science/pith/AAHCCGZOUDULA2BKILZFZMTHYJ/action/author_attestation","sign_citation":"https://pith.science/pith/AAHCCGZOUDULA2BKILZFZMTHYJ/action/citation_signature","submit_replication":"https://pith.science/pith/AAHCCGZOUDULA2BKILZFZMTHYJ/action/replication_record"}},"created_at":"2026-07-05T06:13:54.342626+00:00","updated_at":"2026-07-05T06:13:54.342626+00:00"}