{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2018:DDPADHPGCCJ3IT33NR37TQ26YD","short_pith_number":"pith:DDPADHPG","schema_version":"1.0","canonical_sha256":"18de019de61093b44f7b6c77f9c35ec0db31a80d6a06527c83a1d345084b275e","source":{"kind":"arxiv","id":"1805.09430","version":1},"attestation_state":"computed","paper":{"title":"A Two-Stage Subspace Trust Region Approach for Deep Neural Network Training","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Emilie Chouzenoux, Giovanni Chierchia, Jean-Christophe Pesquet, Viacheslav Dudar, Vladimir Semenov","submitted_at":"2018-05-23T21:12:48Z","abstract_excerpt":"In this paper, we develop a novel second-order method for training feed-forward neural nets. At each iteration, we construct a quadratic approximation to the cost function in a low-dimensional subspace. We minimize this approximation inside a trust region through a two-stage procedure: first inside the embedded positive curvature subspace, followed by a gradient descent step. This approach leads to a fast objective function decay, prevents convergence to saddle points, and alleviates the need for manually tuning parameters. We show the good performance of the proposed algorithm on benchmark da"},"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":"1805.09430","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2018-05-23T21:12:48Z","cross_cats_sorted":[],"title_canon_sha256":"48cf3eb79b8bd2b225a9777121c1a20c088b623aa9f10a22714e7c4f034bfdf8","abstract_canon_sha256":"fb6fb4e89d58de6d8787146767dbd48e62fc4beaa34fb314ec10e095e2633b80"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-18T00:15:03.970954Z","signature_b64":"l+E/AYsWkm4vB839V28oVY1G6mueDqGa3FxAj55hAbULyk/Tq63num8MkCx7bHMmqlakDIVCpLZFzQRgRDC/Ag==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"18de019de61093b44f7b6c77f9c35ec0db31a80d6a06527c83a1d345084b275e","last_reissued_at":"2026-05-18T00:15:03.970292Z","signature_status":"signed_v1","first_computed_at":"2026-05-18T00:15:03.970292Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"A Two-Stage Subspace Trust Region Approach for Deep Neural Network Training","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Emilie Chouzenoux, Giovanni Chierchia, Jean-Christophe Pesquet, Viacheslav Dudar, Vladimir Semenov","submitted_at":"2018-05-23T21:12:48Z","abstract_excerpt":"In this paper, we develop a novel second-order method for training feed-forward neural nets. At each iteration, we construct a quadratic approximation to the cost function in a low-dimensional subspace. We minimize this approximation inside a trust region through a two-stage procedure: first inside the embedded positive curvature subspace, followed by a gradient descent step. This approach leads to a fast objective function decay, prevents convergence to saddle points, and alleviates the need for manually tuning parameters. We show the good performance of the proposed algorithm on benchmark da"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1805.09430","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":"1805.09430","created_at":"2026-05-18T00:15:03.970399+00:00"},{"alias_kind":"arxiv_version","alias_value":"1805.09430v1","created_at":"2026-05-18T00:15:03.970399+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1805.09430","created_at":"2026-05-18T00:15:03.970399+00:00"},{"alias_kind":"pith_short_12","alias_value":"DDPADHPGCCJ3","created_at":"2026-05-18T12:32:19.392346+00:00"},{"alias_kind":"pith_short_16","alias_value":"DDPADHPGCCJ3IT33","created_at":"2026-05-18T12:32:19.392346+00:00"},{"alias_kind":"pith_short_8","alias_value":"DDPADHPG","created_at":"2026-05-18T12:32:19.392346+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/DDPADHPGCCJ3IT33NR37TQ26YD","json":"https://pith.science/pith/DDPADHPGCCJ3IT33NR37TQ26YD.json","graph_json":"https://pith.science/api/pith-number/DDPADHPGCCJ3IT33NR37TQ26YD/graph.json","events_json":"https://pith.science/api/pith-number/DDPADHPGCCJ3IT33NR37TQ26YD/events.json","paper":"https://pith.science/paper/DDPADHPG"},"agent_actions":{"view_html":"https://pith.science/pith/DDPADHPGCCJ3IT33NR37TQ26YD","download_json":"https://pith.science/pith/DDPADHPGCCJ3IT33NR37TQ26YD.json","view_paper":"https://pith.science/paper/DDPADHPG","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=1805.09430&json=true","fetch_graph":"https://pith.science/api/pith-number/DDPADHPGCCJ3IT33NR37TQ26YD/graph.json","fetch_events":"https://pith.science/api/pith-number/DDPADHPGCCJ3IT33NR37TQ26YD/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/DDPADHPGCCJ3IT33NR37TQ26YD/action/timestamp_anchor","attest_storage":"https://pith.science/pith/DDPADHPGCCJ3IT33NR37TQ26YD/action/storage_attestation","attest_author":"https://pith.science/pith/DDPADHPGCCJ3IT33NR37TQ26YD/action/author_attestation","sign_citation":"https://pith.science/pith/DDPADHPGCCJ3IT33NR37TQ26YD/action/citation_signature","submit_replication":"https://pith.science/pith/DDPADHPGCCJ3IT33NR37TQ26YD/action/replication_record"}},"created_at":"2026-05-18T00:15:03.970399+00:00","updated_at":"2026-05-18T00:15:03.970399+00:00"}