{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2019:URXO6XUZXXQVSWOU3QHD5MPWPM","short_pith_number":"pith:URXO6XUZ","schema_version":"1.0","canonical_sha256":"a46eef5e99bde15959d4dc0e3eb1f67b1530a63afdf673b86a93933fe25b2341","source":{"kind":"arxiv","id":"1906.05392","version":2},"attestation_state":"computed","paper":{"title":"Generalization Guarantees for Neural Networks via Harnessing the Low-rank Structure of the Jacobian","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["math.OC","stat.ML"],"primary_cat":"cs.LG","authors_text":"Mahdi Soltanolkotabi, Mingchen Li, Samet Oymak, Zalan Fabian","submitted_at":"2019-06-12T21:39:06Z","abstract_excerpt":"Modern neural network architectures often generalize well despite containing many more parameters than the size of the training dataset. This paper explores the generalization capabilities of neural networks trained via gradient descent. We develop a data-dependent optimization and generalization theory which leverages the low-rank structure of the Jacobian matrix associated with the network. Our results help demystify why training and generalization is easier on clean and structured datasets and harder on noisy and unstructured datasets as well as how the network size affects the evolution of"},"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":"1906.05392","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2019-06-12T21:39:06Z","cross_cats_sorted":["math.OC","stat.ML"],"title_canon_sha256":"2d462bcdeb2bd22e5df79c543f1e61cd700801e41196ea4049292a6d1758b1bc","abstract_canon_sha256":"104b752f286783282044653e3c06409ed57fdfab63bc2a68d58e65836a9a427a"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-17T23:41:30.245371Z","signature_b64":"ptdosuaDf+k9/DmcvXCcDrA7jIvYcSARVKxZk0Wx9mi9iWBz9W+inxjZOG0WOyDexoantnffZ0vygmcWFcuyBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"a46eef5e99bde15959d4dc0e3eb1f67b1530a63afdf673b86a93933fe25b2341","last_reissued_at":"2026-05-17T23:41:30.244650Z","signature_status":"signed_v1","first_computed_at":"2026-05-17T23:41:30.244650Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Generalization Guarantees for Neural Networks via Harnessing the Low-rank Structure of the Jacobian","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["math.OC","stat.ML"],"primary_cat":"cs.LG","authors_text":"Mahdi Soltanolkotabi, Mingchen Li, Samet Oymak, Zalan Fabian","submitted_at":"2019-06-12T21:39:06Z","abstract_excerpt":"Modern neural network architectures often generalize well despite containing many more parameters than the size of the training dataset. This paper explores the generalization capabilities of neural networks trained via gradient descent. We develop a data-dependent optimization and generalization theory which leverages the low-rank structure of the Jacobian matrix associated with the network. Our results help demystify why training and generalization is easier on clean and structured datasets and harder on noisy and unstructured datasets as well as how the network size affects the evolution of"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1906.05392","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":""},"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":"1906.05392","created_at":"2026-05-17T23:41:30.244776+00:00"},{"alias_kind":"arxiv_version","alias_value":"1906.05392v2","created_at":"2026-05-17T23:41:30.244776+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1906.05392","created_at":"2026-05-17T23:41:30.244776+00:00"},{"alias_kind":"pith_short_12","alias_value":"URXO6XUZXXQV","created_at":"2026-05-18T12:33:30.264802+00:00"},{"alias_kind":"pith_short_16","alias_value":"URXO6XUZXXQVSWOU","created_at":"2026-05-18T12:33:30.264802+00:00"},{"alias_kind":"pith_short_8","alias_value":"URXO6XUZ","created_at":"2026-05-18T12:33:30.264802+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":4,"internal_anchor_count":3,"sample":[{"citing_arxiv_id":"2605.18598","citing_title":"Pointwise Generalization in Deep Neural Networks","ref_index":100,"is_internal_anchor":true},{"citing_arxiv_id":"2512.18248","citing_title":"On the Convergence Rate of LoRA Gradient Descent","ref_index":7,"is_internal_anchor":true},{"citing_arxiv_id":"2602.15823","citing_title":"CrispEdit: Low-Curvature Projections for Scalable Non-Destructive LLM Editing","ref_index":18,"is_internal_anchor":true},{"citing_arxiv_id":"2106.09685","citing_title":"LoRA: Low-Rank Adaptation of Large Language Models","ref_index":41,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/URXO6XUZXXQVSWOU3QHD5MPWPM","json":"https://pith.science/pith/URXO6XUZXXQVSWOU3QHD5MPWPM.json","graph_json":"https://pith.science/api/pith-number/URXO6XUZXXQVSWOU3QHD5MPWPM/graph.json","events_json":"https://pith.science/api/pith-number/URXO6XUZXXQVSWOU3QHD5MPWPM/events.json","paper":"https://pith.science/paper/URXO6XUZ"},"agent_actions":{"view_html":"https://pith.science/pith/URXO6XUZXXQVSWOU3QHD5MPWPM","download_json":"https://pith.science/pith/URXO6XUZXXQVSWOU3QHD5MPWPM.json","view_paper":"https://pith.science/paper/URXO6XUZ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=1906.05392&json=true","fetch_graph":"https://pith.science/api/pith-number/URXO6XUZXXQVSWOU3QHD5MPWPM/graph.json","fetch_events":"https://pith.science/api/pith-number/URXO6XUZXXQVSWOU3QHD5MPWPM/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/URXO6XUZXXQVSWOU3QHD5MPWPM/action/timestamp_anchor","attest_storage":"https://pith.science/pith/URXO6XUZXXQVSWOU3QHD5MPWPM/action/storage_attestation","attest_author":"https://pith.science/pith/URXO6XUZXXQVSWOU3QHD5MPWPM/action/author_attestation","sign_citation":"https://pith.science/pith/URXO6XUZXXQVSWOU3QHD5MPWPM/action/citation_signature","submit_replication":"https://pith.science/pith/URXO6XUZXXQVSWOU3QHD5MPWPM/action/replication_record"}},"created_at":"2026-05-17T23:41:30.244776+00:00","updated_at":"2026-05-17T23:41:30.244776+00:00"}