{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:DRTXG2MFRCARHRZ5COEMQTF4EV","short_pith_number":"pith:DRTXG2MF","schema_version":"1.0","canonical_sha256":"1c67736985888113c73d1388c84cbc2557cbf72089a10f9c3be8717ef8ce1ee0","source":{"kind":"arxiv","id":"2105.06351","version":2},"attestation_state":"computed","paper":{"title":"Convergence and Implicit Bias of Gradient Flow on Overparametrized Linear Networks","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Enrique Mallada, Hancheng Min, Rene Vidal, Salma Tarmoun","submitted_at":"2021-05-13T15:13:51Z","abstract_excerpt":"Neural networks trained via gradient descent with random initialization and without any regularization enjoy good generalization performance in practice despite being highly overparametrized. A promising direction to explain this phenomenon is to study how initialization and overparametrization affect convergence and implicit bias of training algorithms. In this paper, we present a novel analysis of single-hidden-layer linear networks trained under gradient flow, which connects initialization, optimization, and overparametrization. Firstly, we show that the squared loss converges exponentially"},"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":"2105.06351","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2021-05-13T15:13:51Z","cross_cats_sorted":[],"title_canon_sha256":"2266ef3d32c5a9a337113776a4b1bf96c4b17150a4f55ef8510def68b874a605","abstract_canon_sha256":"f98d0b11821d926c80f74e10268c9b247fa57bb7a1063e7f3529a3e86c43cfa6"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T04:23:20.382288Z","signature_b64":"DlTn8c6dd0D3VRTeAzzCMGH/k+8qggmb/2UbOmqTLCuN1mGvgJeWuf0iYWM75+GMrJKHN1vK8nGrSO5DJqvMCQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"1c67736985888113c73d1388c84cbc2557cbf72089a10f9c3be8717ef8ce1ee0","last_reissued_at":"2026-07-05T04:23:20.381703Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T04:23:20.381703Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Convergence and Implicit Bias of Gradient Flow on Overparametrized Linear Networks","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Enrique Mallada, Hancheng Min, Rene Vidal, Salma Tarmoun","submitted_at":"2021-05-13T15:13:51Z","abstract_excerpt":"Neural networks trained via gradient descent with random initialization and without any regularization enjoy good generalization performance in practice despite being highly overparametrized. A promising direction to explain this phenomenon is to study how initialization and overparametrization affect convergence and implicit bias of training algorithms. In this paper, we present a novel analysis of single-hidden-layer linear networks trained under gradient flow, which connects initialization, optimization, and overparametrization. Firstly, we show that the squared loss converges exponentially"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2105.06351","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/2105.06351/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":"2105.06351","created_at":"2026-07-05T04:23:20.381773+00:00"},{"alias_kind":"arxiv_version","alias_value":"2105.06351v2","created_at":"2026-07-05T04:23:20.381773+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2105.06351","created_at":"2026-07-05T04:23:20.381773+00:00"},{"alias_kind":"pith_short_12","alias_value":"DRTXG2MFRCAR","created_at":"2026-07-05T04:23:20.381773+00:00"},{"alias_kind":"pith_short_16","alias_value":"DRTXG2MFRCARHRZ5","created_at":"2026-07-05T04:23:20.381773+00:00"},{"alias_kind":"pith_short_8","alias_value":"DRTXG2MF","created_at":"2026-07-05T04:23:20.381773+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":4,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.01288","citing_title":"A Theory of Saddle Escape in Deep Nonlinear Networks","ref_index":31,"is_internal_anchor":false},{"citing_arxiv_id":"2605.05436","citing_title":"Estimating Implicit Regularization in Deep Learning","ref_index":27,"is_internal_anchor":false},{"citing_arxiv_id":"2605.01288","citing_title":"A Theory of Saddle Escape in Deep Nonlinear Networks","ref_index":31,"is_internal_anchor":false},{"citing_arxiv_id":"2605.01288","citing_title":"A Theory of Saddle Escape in Deep Nonlinear Networks","ref_index":31,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/DRTXG2MFRCARHRZ5COEMQTF4EV","json":"https://pith.science/pith/DRTXG2MFRCARHRZ5COEMQTF4EV.json","graph_json":"https://pith.science/api/pith-number/DRTXG2MFRCARHRZ5COEMQTF4EV/graph.json","events_json":"https://pith.science/api/pith-number/DRTXG2MFRCARHRZ5COEMQTF4EV/events.json","paper":"https://pith.science/paper/DRTXG2MF"},"agent_actions":{"view_html":"https://pith.science/pith/DRTXG2MFRCARHRZ5COEMQTF4EV","download_json":"https://pith.science/pith/DRTXG2MFRCARHRZ5COEMQTF4EV.json","view_paper":"https://pith.science/paper/DRTXG2MF","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2105.06351&json=true","fetch_graph":"https://pith.science/api/pith-number/DRTXG2MFRCARHRZ5COEMQTF4EV/graph.json","fetch_events":"https://pith.science/api/pith-number/DRTXG2MFRCARHRZ5COEMQTF4EV/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/DRTXG2MFRCARHRZ5COEMQTF4EV/action/timestamp_anchor","attest_storage":"https://pith.science/pith/DRTXG2MFRCARHRZ5COEMQTF4EV/action/storage_attestation","attest_author":"https://pith.science/pith/DRTXG2MFRCARHRZ5COEMQTF4EV/action/author_attestation","sign_citation":"https://pith.science/pith/DRTXG2MFRCARHRZ5COEMQTF4EV/action/citation_signature","submit_replication":"https://pith.science/pith/DRTXG2MFRCARHRZ5COEMQTF4EV/action/replication_record"}},"created_at":"2026-07-05T04:23:20.381773+00:00","updated_at":"2026-07-05T04:23:20.381773+00:00"}