{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2019:ZWMJJRBERURJYEDDAIFQJ55FPG","short_pith_number":"pith:ZWMJJRBE","schema_version":"1.0","canonical_sha256":"cd9894c4248d229c1063020b04f7a579bb32ca1159d8ad0ac7c7966afa4ead75","source":{"kind":"arxiv","id":"1902.00247","version":2},"attestation_state":"computed","paper":{"title":"Sharp Analysis for Nonconvex SGD Escaping from Saddle Points","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CC","cs.LG"],"primary_cat":"math.OC","authors_text":"Cong Fang, Tong Zhang, Zhouchen Lin","submitted_at":"2019-02-01T09:35:27Z","abstract_excerpt":"In this paper, we give a sharp analysis for Stochastic Gradient Descent (SGD) and prove that SGD is able to efficiently escape from saddle points and find an $(\\epsilon, O(\\epsilon^{0.5}))$-approximate second-order stationary point in $\\tilde{O}(\\epsilon^{-3.5})$ stochastic gradient computations for generic nonconvex optimization problems, when the objective function satisfies gradient-Lipschitz, Hessian-Lipschitz, and dispersive noise assumptions. This result subverts the classical belief that SGD requires at least $O(\\epsilon^{-4})$ stochastic gradient computations for obtaining an $(\\epsilo"},"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":"1902.00247","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"math.OC","submitted_at":"2019-02-01T09:35:27Z","cross_cats_sorted":["cs.CC","cs.LG"],"title_canon_sha256":"55b0b75926211ceef396d23584e9f660a1799bce80fa2132b44ad55ae27223da","abstract_canon_sha256":"baa4554de493be485f7aa3ee6be7dabde73332e57ab3dacf46a7608e4ff21a43"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-17T23:44:21.738853Z","signature_b64":"EW3fdnJ6TOXD4H6fk52FI9v4XTv+lHF0hhwARwwy0s5ovFf7NAp94VPn/ulJdreR/00x92SgVvu629EYldzUDA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"cd9894c4248d229c1063020b04f7a579bb32ca1159d8ad0ac7c7966afa4ead75","last_reissued_at":"2026-05-17T23:44:21.738174Z","signature_status":"signed_v1","first_computed_at":"2026-05-17T23:44:21.738174Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Sharp Analysis for Nonconvex SGD Escaping from Saddle Points","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CC","cs.LG"],"primary_cat":"math.OC","authors_text":"Cong Fang, Tong Zhang, Zhouchen Lin","submitted_at":"2019-02-01T09:35:27Z","abstract_excerpt":"In this paper, we give a sharp analysis for Stochastic Gradient Descent (SGD) and prove that SGD is able to efficiently escape from saddle points and find an $(\\epsilon, O(\\epsilon^{0.5}))$-approximate second-order stationary point in $\\tilde{O}(\\epsilon^{-3.5})$ stochastic gradient computations for generic nonconvex optimization problems, when the objective function satisfies gradient-Lipschitz, Hessian-Lipschitz, and dispersive noise assumptions. This result subverts the classical belief that SGD requires at least $O(\\epsilon^{-4})$ stochastic gradient computations for obtaining an $(\\epsilo"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1902.00247","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":"1902.00247","created_at":"2026-05-17T23:44:21.738274+00:00"},{"alias_kind":"arxiv_version","alias_value":"1902.00247v2","created_at":"2026-05-17T23:44:21.738274+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1902.00247","created_at":"2026-05-17T23:44:21.738274+00:00"},{"alias_kind":"pith_short_12","alias_value":"ZWMJJRBERURJ","created_at":"2026-05-18T12:33:33.725879+00:00"},{"alias_kind":"pith_short_16","alias_value":"ZWMJJRBERURJYEDD","created_at":"2026-05-18T12:33:33.725879+00:00"},{"alias_kind":"pith_short_8","alias_value":"ZWMJJRBE","created_at":"2026-05-18T12:33:33.725879+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":3,"internal_anchor_count":3,"sample":[{"citing_arxiv_id":"1907.01848","citing_title":"Distributed Learning in Non-Convex Environments -- Part I: Agreement at a Linear Rate","ref_index":28,"is_internal_anchor":true},{"citing_arxiv_id":"1907.01849","citing_title":"Distributed Learning in Non-Convex Environments -- Part II: Polynomial Escape from Saddle-Points","ref_index":28,"is_internal_anchor":true},{"citing_arxiv_id":"2307.07030","citing_title":"Accelerated Gradient Methods for Nonconvex Optimization: Escape Trajectories From Strict Saddle Points and Convergence to Local Minima","ref_index":28,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/ZWMJJRBERURJYEDDAIFQJ55FPG","json":"https://pith.science/pith/ZWMJJRBERURJYEDDAIFQJ55FPG.json","graph_json":"https://pith.science/api/pith-number/ZWMJJRBERURJYEDDAIFQJ55FPG/graph.json","events_json":"https://pith.science/api/pith-number/ZWMJJRBERURJYEDDAIFQJ55FPG/events.json","paper":"https://pith.science/paper/ZWMJJRBE"},"agent_actions":{"view_html":"https://pith.science/pith/ZWMJJRBERURJYEDDAIFQJ55FPG","download_json":"https://pith.science/pith/ZWMJJRBERURJYEDDAIFQJ55FPG.json","view_paper":"https://pith.science/paper/ZWMJJRBE","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=1902.00247&json=true","fetch_graph":"https://pith.science/api/pith-number/ZWMJJRBERURJYEDDAIFQJ55FPG/graph.json","fetch_events":"https://pith.science/api/pith-number/ZWMJJRBERURJYEDDAIFQJ55FPG/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/ZWMJJRBERURJYEDDAIFQJ55FPG/action/timestamp_anchor","attest_storage":"https://pith.science/pith/ZWMJJRBERURJYEDDAIFQJ55FPG/action/storage_attestation","attest_author":"https://pith.science/pith/ZWMJJRBERURJYEDDAIFQJ55FPG/action/author_attestation","sign_citation":"https://pith.science/pith/ZWMJJRBERURJYEDDAIFQJ55FPG/action/citation_signature","submit_replication":"https://pith.science/pith/ZWMJJRBERURJYEDDAIFQJ55FPG/action/replication_record"}},"created_at":"2026-05-17T23:44:21.738274+00:00","updated_at":"2026-05-17T23:44:21.738274+00:00"}