{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2017:IDAEN2HTE5ZMJWCPSAHCBFI7IT","short_pith_number":"pith:IDAEN2HT","schema_version":"1.0","canonical_sha256":"40c046e8f32772c4d84f900e20951f44fc914fb51ee9aeb5b8dbd7dcdc5f177e","source":{"kind":"arxiv","id":"1711.10456","version":1},"attestation_state":"computed","paper":{"title":"Accelerated Gradient Descent Escapes Saddle Points Faster than Gradient Descent","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["math.OC","stat.ML"],"primary_cat":"cs.LG","authors_text":"Chi Jin, Michael I. Jordan, Praneeth Netrapalli","submitted_at":"2017-11-28T18:38:35Z","abstract_excerpt":"Nesterov's accelerated gradient descent (AGD), an instance of the general family of \"momentum methods\", provably achieves faster convergence rate than gradient descent (GD) in the convex setting. However, whether these methods are superior to GD in the nonconvex setting remains open. This paper studies a simple variant of AGD, and shows that it escapes saddle points and finds a second-order stationary point in $\\tilde{O}(1/\\epsilon^{7/4})$ iterations, faster than the $\\tilde{O}(1/\\epsilon^{2})$ iterations required by GD. To the best of our knowledge, this is the first Hessian-free algorithm to"},"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":"1711.10456","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2017-11-28T18:38:35Z","cross_cats_sorted":["math.OC","stat.ML"],"title_canon_sha256":"025d612ac0ef93cec1c5e0360c2a51885044e99e7d125389181751ec80c9ba1d","abstract_canon_sha256":"e2bcd47dffb007535c542127fca360d9f9499560a0d6989c0583f4bd321e8f2f"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-18T00:29:22.079377Z","signature_b64":"YGjWPrC9CkKBkxQ/TU9D66oVw/A+rD7/+XjDwn+wJKp5xXSW0PM64R2FJApb9vX6mq7tmTsWWI8kEUQ3lfvxDQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"40c046e8f32772c4d84f900e20951f44fc914fb51ee9aeb5b8dbd7dcdc5f177e","last_reissued_at":"2026-05-18T00:29:22.078675Z","signature_status":"signed_v1","first_computed_at":"2026-05-18T00:29:22.078675Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Accelerated Gradient Descent Escapes Saddle Points Faster than Gradient Descent","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["math.OC","stat.ML"],"primary_cat":"cs.LG","authors_text":"Chi Jin, Michael I. Jordan, Praneeth Netrapalli","submitted_at":"2017-11-28T18:38:35Z","abstract_excerpt":"Nesterov's accelerated gradient descent (AGD), an instance of the general family of \"momentum methods\", provably achieves faster convergence rate than gradient descent (GD) in the convex setting. However, whether these methods are superior to GD in the nonconvex setting remains open. This paper studies a simple variant of AGD, and shows that it escapes saddle points and finds a second-order stationary point in $\\tilde{O}(1/\\epsilon^{7/4})$ iterations, faster than the $\\tilde{O}(1/\\epsilon^{2})$ iterations required by GD. To the best of our knowledge, this is the first Hessian-free algorithm to"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1711.10456","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":"1711.10456","created_at":"2026-05-18T00:29:22.078792+00:00"},{"alias_kind":"arxiv_version","alias_value":"1711.10456v1","created_at":"2026-05-18T00:29:22.078792+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1711.10456","created_at":"2026-05-18T00:29:22.078792+00:00"},{"alias_kind":"pith_short_12","alias_value":"IDAEN2HTE5ZM","created_at":"2026-05-18T12:31:21.493067+00:00"},{"alias_kind":"pith_short_16","alias_value":"IDAEN2HTE5ZMJWCP","created_at":"2026-05-18T12:31:21.493067+00:00"},{"alias_kind":"pith_short_8","alias_value":"IDAEN2HT","created_at":"2026-05-18T12:31:21.493067+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":3,"internal_anchor_count":2,"sample":[{"citing_arxiv_id":"1907.09697","citing_title":"Heavy-ball Algorithms Always Escape Saddle Points","ref_index":11,"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":41,"is_internal_anchor":true},{"citing_arxiv_id":"2604.24488","citing_title":"Scalable First-Order Interior Point Trust Region Algorithms for Linearly Constrained Optimization","ref_index":23,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/IDAEN2HTE5ZMJWCPSAHCBFI7IT","json":"https://pith.science/pith/IDAEN2HTE5ZMJWCPSAHCBFI7IT.json","graph_json":"https://pith.science/api/pith-number/IDAEN2HTE5ZMJWCPSAHCBFI7IT/graph.json","events_json":"https://pith.science/api/pith-number/IDAEN2HTE5ZMJWCPSAHCBFI7IT/events.json","paper":"https://pith.science/paper/IDAEN2HT"},"agent_actions":{"view_html":"https://pith.science/pith/IDAEN2HTE5ZMJWCPSAHCBFI7IT","download_json":"https://pith.science/pith/IDAEN2HTE5ZMJWCPSAHCBFI7IT.json","view_paper":"https://pith.science/paper/IDAEN2HT","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=1711.10456&json=true","fetch_graph":"https://pith.science/api/pith-number/IDAEN2HTE5ZMJWCPSAHCBFI7IT/graph.json","fetch_events":"https://pith.science/api/pith-number/IDAEN2HTE5ZMJWCPSAHCBFI7IT/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/IDAEN2HTE5ZMJWCPSAHCBFI7IT/action/timestamp_anchor","attest_storage":"https://pith.science/pith/IDAEN2HTE5ZMJWCPSAHCBFI7IT/action/storage_attestation","attest_author":"https://pith.science/pith/IDAEN2HTE5ZMJWCPSAHCBFI7IT/action/author_attestation","sign_citation":"https://pith.science/pith/IDAEN2HTE5ZMJWCPSAHCBFI7IT/action/citation_signature","submit_replication":"https://pith.science/pith/IDAEN2HTE5ZMJWCPSAHCBFI7IT/action/replication_record"}},"created_at":"2026-05-18T00:29:22.078792+00:00","updated_at":"2026-05-18T00:29:22.078792+00:00"}