{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:2ACHD5Z4JCYB5L3OYBG45PMMSW","short_pith_number":"pith:2ACHD5Z4","schema_version":"1.0","canonical_sha256":"d00471f73c48b01eaf6ec04dcebd8c95bf7e8fd4fbccbe1c4f0bf27e9ba5a67a","source":{"kind":"arxiv","id":"2410.13770","version":2},"attestation_state":"computed","paper":{"title":"Probing the Latent Hierarchical Structure of Data via Diffusion Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cond-mat.dis-nn","cs.LG"],"primary_cat":"stat.ML","authors_text":"Alessandro Favero, Antonio Sclocchi, Matthieu Wyart, Noam Itzhak Levi","submitted_at":"2024-10-17T17:08:39Z","abstract_excerpt":"High-dimensional data must be highly structured to be learnable. Although the compositional and hierarchical nature of data is often put forward to explain learnability, quantitative measurements establishing these properties are scarce. Likewise, accessing the latent variables underlying such a data structure remains a challenge. In this work, we show that forward-backward experiments in diffusion-based models, where data is noised and then denoised to generate new samples, are a promising tool to probe the latent structure of data. We predict in simple hierarchical models that, in this proce"},"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":"2410.13770","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"stat.ML","submitted_at":"2024-10-17T17:08:39Z","cross_cats_sorted":["cond-mat.dis-nn","cs.LG"],"title_canon_sha256":"d2c5df9d23b7414fc7078c0cf0ab1a074bd741f2df506c48953dbe2ff5512df0","abstract_canon_sha256":"af4f3757b3fbc42bcb91eddee396ff5b01e84df3c4252430141134dc03658187"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:22:08.087111Z","signature_b64":"Kc40Ew9f9IgYUl7R72gW7wtDl2UEZAP7eg37zvDHIaqKK6FwXLDcU1MuRiKs5ER2kljjvQvtpgYstDyYUmscAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"d00471f73c48b01eaf6ec04dcebd8c95bf7e8fd4fbccbe1c4f0bf27e9ba5a67a","last_reissued_at":"2026-07-05T10:22:08.086582Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:22:08.086582Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Probing the Latent Hierarchical Structure of Data via Diffusion Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cond-mat.dis-nn","cs.LG"],"primary_cat":"stat.ML","authors_text":"Alessandro Favero, Antonio Sclocchi, Matthieu Wyart, Noam Itzhak Levi","submitted_at":"2024-10-17T17:08:39Z","abstract_excerpt":"High-dimensional data must be highly structured to be learnable. Although the compositional and hierarchical nature of data is often put forward to explain learnability, quantitative measurements establishing these properties are scarce. Likewise, accessing the latent variables underlying such a data structure remains a challenge. In this work, we show that forward-backward experiments in diffusion-based models, where data is noised and then denoised to generate new samples, are a promising tool to probe the latent structure of data. We predict in simple hierarchical models that, in this proce"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2410.13770","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/2410.13770/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":"2410.13770","created_at":"2026-07-05T10:22:08.086640+00:00"},{"alias_kind":"arxiv_version","alias_value":"2410.13770v2","created_at":"2026-07-05T10:22:08.086640+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2410.13770","created_at":"2026-07-05T10:22:08.086640+00:00"},{"alias_kind":"pith_short_12","alias_value":"2ACHD5Z4JCYB","created_at":"2026-07-05T10:22:08.086640+00:00"},{"alias_kind":"pith_short_16","alias_value":"2ACHD5Z4JCYB5L3O","created_at":"2026-07-05T10:22:08.086640+00:00"},{"alias_kind":"pith_short_8","alias_value":"2ACHD5Z4","created_at":"2026-07-05T10:22:08.086640+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":3,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.28486","citing_title":"Spectral phase transitions and trainability in neural network learning dynamics","ref_index":47,"is_internal_anchor":false},{"citing_arxiv_id":"2508.06614","citing_title":"Local Diffusion Models and Phases of Data Distributions","ref_index":23,"is_internal_anchor":false},{"citing_arxiv_id":"2604.02415","citing_title":"Generative models on phase space","ref_index":67,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/2ACHD5Z4JCYB5L3OYBG45PMMSW","json":"https://pith.science/pith/2ACHD5Z4JCYB5L3OYBG45PMMSW.json","graph_json":"https://pith.science/api/pith-number/2ACHD5Z4JCYB5L3OYBG45PMMSW/graph.json","events_json":"https://pith.science/api/pith-number/2ACHD5Z4JCYB5L3OYBG45PMMSW/events.json","paper":"https://pith.science/paper/2ACHD5Z4"},"agent_actions":{"view_html":"https://pith.science/pith/2ACHD5Z4JCYB5L3OYBG45PMMSW","download_json":"https://pith.science/pith/2ACHD5Z4JCYB5L3OYBG45PMMSW.json","view_paper":"https://pith.science/paper/2ACHD5Z4","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2410.13770&json=true","fetch_graph":"https://pith.science/api/pith-number/2ACHD5Z4JCYB5L3OYBG45PMMSW/graph.json","fetch_events":"https://pith.science/api/pith-number/2ACHD5Z4JCYB5L3OYBG45PMMSW/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/2ACHD5Z4JCYB5L3OYBG45PMMSW/action/timestamp_anchor","attest_storage":"https://pith.science/pith/2ACHD5Z4JCYB5L3OYBG45PMMSW/action/storage_attestation","attest_author":"https://pith.science/pith/2ACHD5Z4JCYB5L3OYBG45PMMSW/action/author_attestation","sign_citation":"https://pith.science/pith/2ACHD5Z4JCYB5L3OYBG45PMMSW/action/citation_signature","submit_replication":"https://pith.science/pith/2ACHD5Z4JCYB5L3OYBG45PMMSW/action/replication_record"}},"created_at":"2026-07-05T10:22:08.086640+00:00","updated_at":"2026-07-05T10:22:08.086640+00:00"}