{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2019:ZUVZ4QCKYARHV5DJOS7KNAPC2P","short_pith_number":"pith:ZUVZ4QCK","schema_version":"1.0","canonical_sha256":"cd2b9e404ac0227af46974bea681e2d3d39e4fb5c19a006ffe6312b81193930c","source":{"kind":"arxiv","id":"1906.02353","version":1},"attestation_state":"computed","paper":{"title":"Efficient Subsampled Gauss-Newton and Natural Gradient Methods for Training Neural Networks","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["stat.ML"],"primary_cat":"cs.LG","authors_text":"Donald Goldfarb, Yi Ren","submitted_at":"2019-06-05T23:13:42Z","abstract_excerpt":"We present practical Levenberg-Marquardt variants of Gauss-Newton and natural gradient methods for solving non-convex optimization problems that arise in training deep neural networks involving enormous numbers of variables and huge data sets. Our methods use subsampled Gauss-Newton or Fisher information matrices and either subsampled gradient estimates (fully stochastic) or full gradients (semi-stochastic), which, in the latter case, we prove convergent to a stationary point. By using the Sherman-Morrison-Woodbury formula with automatic differentiation (backpropagation) we show how our method"},"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":"1906.02353","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2019-06-05T23:13:42Z","cross_cats_sorted":["stat.ML"],"title_canon_sha256":"546549736edb6b8b7211035dd6a9215efbe3ba485c963cf1214e7c874f67a43c","abstract_canon_sha256":"22f86f1ef82b45b418d4f81c32c7e098f69618fb16006c3deac89046d4ee9d2c"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-17T23:44:01.492461Z","signature_b64":"KfLkA9wv1TquaTSOz8zl5A+67L3bSkZLXNVVyNYQhBvt30HDM9m9HpnA1a/z/g3CQrDeta6QCNOo7OLMsdgVDQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"cd2b9e404ac0227af46974bea681e2d3d39e4fb5c19a006ffe6312b81193930c","last_reissued_at":"2026-05-17T23:44:01.491978Z","signature_status":"signed_v1","first_computed_at":"2026-05-17T23:44:01.491978Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Efficient Subsampled Gauss-Newton and Natural Gradient Methods for Training Neural Networks","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["stat.ML"],"primary_cat":"cs.LG","authors_text":"Donald Goldfarb, Yi Ren","submitted_at":"2019-06-05T23:13:42Z","abstract_excerpt":"We present practical Levenberg-Marquardt variants of Gauss-Newton and natural gradient methods for solving non-convex optimization problems that arise in training deep neural networks involving enormous numbers of variables and huge data sets. Our methods use subsampled Gauss-Newton or Fisher information matrices and either subsampled gradient estimates (fully stochastic) or full gradients (semi-stochastic), which, in the latter case, we prove convergent to a stationary point. By using the Sherman-Morrison-Woodbury formula with automatic differentiation (backpropagation) we show how our method"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1906.02353","kind":"arxiv","version":1},"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":"1906.02353","created_at":"2026-05-17T23:44:01.492055+00:00"},{"alias_kind":"arxiv_version","alias_value":"1906.02353v1","created_at":"2026-05-17T23:44:01.492055+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1906.02353","created_at":"2026-05-17T23:44:01.492055+00:00"},{"alias_kind":"pith_short_12","alias_value":"ZUVZ4QCKYARH","created_at":"2026-05-18T12:33:33.725879+00:00"},{"alias_kind":"pith_short_16","alias_value":"ZUVZ4QCKYARHV5DJ","created_at":"2026-05-18T12:33:33.725879+00:00"},{"alias_kind":"pith_short_8","alias_value":"ZUVZ4QCK","created_at":"2026-05-18T12:33:33.725879+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":6,"internal_anchor_count":4,"sample":[{"citing_arxiv_id":"2607.02292","citing_title":"One More Time: Revisiting Neural Quantum States from a Reinforcement Learning Perspective","ref_index":85,"is_internal_anchor":true},{"citing_arxiv_id":"2607.02292","citing_title":"One More Time: Revisiting Neural Quantum States from a Reinforcement Learning Perspective","ref_index":146,"is_internal_anchor":true},{"citing_arxiv_id":"2605.18591","citing_title":"Randomized Advantage Transformation (RAT): Computing Natural Policy Gradients via Direct Backpropagation","ref_index":19,"is_internal_anchor":true},{"citing_arxiv_id":"2601.03162","citing_title":"On the Convergence Behavior of Preconditioned Gradient Descent Toward the Rich Learning Regime","ref_index":17,"is_internal_anchor":true},{"citing_arxiv_id":"2605.06081","citing_title":"Fast Gauss-Newton for Multiclass Cross-Entropy","ref_index":30,"is_internal_anchor":false},{"citing_arxiv_id":"2604.22117","citing_title":"PermaFrost-Attack: Stealth Pretraining Seeding(SPS) for planting Logic Landmines During LLM Training","ref_index":59,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/ZUVZ4QCKYARHV5DJOS7KNAPC2P","json":"https://pith.science/pith/ZUVZ4QCKYARHV5DJOS7KNAPC2P.json","graph_json":"https://pith.science/api/pith-number/ZUVZ4QCKYARHV5DJOS7KNAPC2P/graph.json","events_json":"https://pith.science/api/pith-number/ZUVZ4QCKYARHV5DJOS7KNAPC2P/events.json","paper":"https://pith.science/paper/ZUVZ4QCK"},"agent_actions":{"view_html":"https://pith.science/pith/ZUVZ4QCKYARHV5DJOS7KNAPC2P","download_json":"https://pith.science/pith/ZUVZ4QCKYARHV5DJOS7KNAPC2P.json","view_paper":"https://pith.science/paper/ZUVZ4QCK","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=1906.02353&json=true","fetch_graph":"https://pith.science/api/pith-number/ZUVZ4QCKYARHV5DJOS7KNAPC2P/graph.json","fetch_events":"https://pith.science/api/pith-number/ZUVZ4QCKYARHV5DJOS7KNAPC2P/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/ZUVZ4QCKYARHV5DJOS7KNAPC2P/action/timestamp_anchor","attest_storage":"https://pith.science/pith/ZUVZ4QCKYARHV5DJOS7KNAPC2P/action/storage_attestation","attest_author":"https://pith.science/pith/ZUVZ4QCKYARHV5DJOS7KNAPC2P/action/author_attestation","sign_citation":"https://pith.science/pith/ZUVZ4QCKYARHV5DJOS7KNAPC2P/action/citation_signature","submit_replication":"https://pith.science/pith/ZUVZ4QCKYARHV5DJOS7KNAPC2P/action/replication_record"}},"created_at":"2026-05-17T23:44:01.492055+00:00","updated_at":"2026-05-17T23:44:01.492055+00:00"}