{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2022:6GKZG5CSKMGFLNHCSTHERKWV4Y","short_pith_number":"pith:6GKZG5CS","schema_version":"1.0","canonical_sha256":"f195937452530c55b4e294ce48aad5e61ef4dc6e431f23975888077362970a63","source":{"kind":"arxiv","id":"2204.07172","version":4},"attestation_state":"computed","paper":{"title":"Diagnosing and Fixing Manifold Overfitting in Deep Generative Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.LG","stat.ME"],"primary_cat":"stat.ML","authors_text":"Anthony L. Caterini, Brendan Leigh Ross, Gabriel Loaiza-Ganem, Jesse C. Cresswell","submitted_at":"2022-04-14T18:00:03Z","abstract_excerpt":"Likelihood-based, or explicit, deep generative models use neural networks to construct flexible high-dimensional densities. This formulation directly contradicts the manifold hypothesis, which states that observed data lies on a low-dimensional manifold embedded in high-dimensional ambient space. In this paper we investigate the pathologies of maximum-likelihood training in the presence of this dimensionality mismatch. We formally prove that degenerate optima are achieved wherein the manifold itself is learned but not the distribution on it, a phenomenon we call manifold overfitting. We propos"},"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":"2204.07172","kind":"arxiv","version":4},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"stat.ML","submitted_at":"2022-04-14T18:00:03Z","cross_cats_sorted":["cs.AI","cs.LG","stat.ME"],"title_canon_sha256":"40e7632738f6819b3722f4f5e4246b31dedd53613a23234d7363fb7a93dcb1d7","abstract_canon_sha256":"8f9fe4b9b45b7fd5aa4ea2e0148b692bda7750358a79f227ed5db2e3b3ff2f95"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T05:20:10.969183Z","signature_b64":"EmJGuU/lgwwpe85EXk2xTY0aejFnoMYMXaLOGqHwGwOSUSm7ZVFs1qqmVvG76gsUHWiN8S/Nv1ndfVDg7F8YAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"f195937452530c55b4e294ce48aad5e61ef4dc6e431f23975888077362970a63","last_reissued_at":"2026-07-05T05:20:10.968736Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T05:20:10.968736Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Diagnosing and Fixing Manifold Overfitting in Deep Generative Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.LG","stat.ME"],"primary_cat":"stat.ML","authors_text":"Anthony L. Caterini, Brendan Leigh Ross, Gabriel Loaiza-Ganem, Jesse C. Cresswell","submitted_at":"2022-04-14T18:00:03Z","abstract_excerpt":"Likelihood-based, or explicit, deep generative models use neural networks to construct flexible high-dimensional densities. This formulation directly contradicts the manifold hypothesis, which states that observed data lies on a low-dimensional manifold embedded in high-dimensional ambient space. In this paper we investigate the pathologies of maximum-likelihood training in the presence of this dimensionality mismatch. We formally prove that degenerate optima are achieved wherein the manifold itself is learned but not the distribution on it, a phenomenon we call manifold overfitting. We propos"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2204.07172","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/2204.07172/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":"2204.07172","created_at":"2026-07-05T05:20:10.968801+00:00"},{"alias_kind":"arxiv_version","alias_value":"2204.07172v4","created_at":"2026-07-05T05:20:10.968801+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2204.07172","created_at":"2026-07-05T05:20:10.968801+00:00"},{"alias_kind":"pith_short_12","alias_value":"6GKZG5CSKMGF","created_at":"2026-07-05T05:20:10.968801+00:00"},{"alias_kind":"pith_short_16","alias_value":"6GKZG5CSKMGFLNHC","created_at":"2026-07-05T05:20:10.968801+00:00"},{"alias_kind":"pith_short_8","alias_value":"6GKZG5CS","created_at":"2026-07-05T05:20:10.968801+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2506.12187","citing_title":"Characterizing Neural Manifolds' Properties and Curvatures using Normalizing Flows","ref_index":70,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/6GKZG5CSKMGFLNHCSTHERKWV4Y","json":"https://pith.science/pith/6GKZG5CSKMGFLNHCSTHERKWV4Y.json","graph_json":"https://pith.science/api/pith-number/6GKZG5CSKMGFLNHCSTHERKWV4Y/graph.json","events_json":"https://pith.science/api/pith-number/6GKZG5CSKMGFLNHCSTHERKWV4Y/events.json","paper":"https://pith.science/paper/6GKZG5CS"},"agent_actions":{"view_html":"https://pith.science/pith/6GKZG5CSKMGFLNHCSTHERKWV4Y","download_json":"https://pith.science/pith/6GKZG5CSKMGFLNHCSTHERKWV4Y.json","view_paper":"https://pith.science/paper/6GKZG5CS","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2204.07172&json=true","fetch_graph":"https://pith.science/api/pith-number/6GKZG5CSKMGFLNHCSTHERKWV4Y/graph.json","fetch_events":"https://pith.science/api/pith-number/6GKZG5CSKMGFLNHCSTHERKWV4Y/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/6GKZG5CSKMGFLNHCSTHERKWV4Y/action/timestamp_anchor","attest_storage":"https://pith.science/pith/6GKZG5CSKMGFLNHCSTHERKWV4Y/action/storage_attestation","attest_author":"https://pith.science/pith/6GKZG5CSKMGFLNHCSTHERKWV4Y/action/author_attestation","sign_citation":"https://pith.science/pith/6GKZG5CSKMGFLNHCSTHERKWV4Y/action/citation_signature","submit_replication":"https://pith.science/pith/6GKZG5CSKMGFLNHCSTHERKWV4Y/action/replication_record"}},"created_at":"2026-07-05T05:20:10.968801+00:00","updated_at":"2026-07-05T05:20:10.968801+00:00"}