{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2018:3W73LM64AYQQ46POITXWQ35ETA","short_pith_number":"pith:3W73LM64","schema_version":"1.0","canonical_sha256":"ddbfb5b3dc06210e79ee44ef686fa49807946f1fd4745cddf7b14bb5fc9cfc65","source":{"kind":"arxiv","id":"1808.05671","version":4},"attestation_state":"computed","paper":{"title":"On the Convergence of Adaptive Gradient Methods for Nonconvex Optimization","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["math.OC","stat.ML"],"primary_cat":"cs.LG","authors_text":"Dongruo Zhou, Jinghui Chen, Quanquan Gu, Yuan Cao, Ziyan Yang","submitted_at":"2018-08-16T20:25:28Z","abstract_excerpt":"Adaptive gradient methods are workhorses in deep learning. However, the convergence guarantees of adaptive gradient methods for nonconvex optimization have not been thoroughly studied. In this paper, we provide a fine-grained convergence analysis for a general class of adaptive gradient methods including AMSGrad, RMSProp and AdaGrad. For smooth nonconvex functions, we prove that adaptive gradient methods in expectation converge to a first-order stationary point. Our convergence rate is better than existing results for adaptive gradient methods in terms of dimension. In addition, we also prove "},"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":"1808.05671","kind":"arxiv","version":4},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2018-08-16T20:25:28Z","cross_cats_sorted":["math.OC","stat.ML"],"title_canon_sha256":"52a696e6bc19ce389dbf51ec5b820f88bd05b9591931e78e321affb8c1724708","abstract_canon_sha256":"ec41c1ecb42f150e461cd3c63218e16d64513016442ec0b90bd81903debe7114"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:34:46.095307Z","signature_b64":"q5sYNhj/9XZs2ayQLedwLX54tq6VtBrpu0vFgimYbzyTUSlfPGNAE02jlOk8u5C2FUdaIyXCFwpDcMvWZimqDQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"ddbfb5b3dc06210e79ee44ef686fa49807946f1fd4745cddf7b14bb5fc9cfc65","last_reissued_at":"2026-07-05T08:34:46.094826Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:34:46.094826Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"On the Convergence of Adaptive Gradient Methods for Nonconvex Optimization","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["math.OC","stat.ML"],"primary_cat":"cs.LG","authors_text":"Dongruo Zhou, Jinghui Chen, Quanquan Gu, Yuan Cao, Ziyan Yang","submitted_at":"2018-08-16T20:25:28Z","abstract_excerpt":"Adaptive gradient methods are workhorses in deep learning. However, the convergence guarantees of adaptive gradient methods for nonconvex optimization have not been thoroughly studied. In this paper, we provide a fine-grained convergence analysis for a general class of adaptive gradient methods including AMSGrad, RMSProp and AdaGrad. For smooth nonconvex functions, we prove that adaptive gradient methods in expectation converge to a first-order stationary point. Our convergence rate is better than existing results for adaptive gradient methods in terms of dimension. In addition, we also prove "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1808.05671","kind":"arxiv","version":4},"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/1808.05671/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":"1808.05671","created_at":"2026-07-05T08:34:46.094900+00:00"},{"alias_kind":"arxiv_version","alias_value":"1808.05671v4","created_at":"2026-07-05T08:34:46.094900+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1808.05671","created_at":"2026-07-05T08:34:46.094900+00:00"},{"alias_kind":"pith_short_12","alias_value":"3W73LM64AYQQ","created_at":"2026-07-05T08:34:46.094900+00:00"},{"alias_kind":"pith_short_16","alias_value":"3W73LM64AYQQ46PO","created_at":"2026-07-05T08:34:46.094900+00:00"},{"alias_kind":"pith_short_8","alias_value":"3W73LM64","created_at":"2026-07-05T08:34:46.094900+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":6,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.17526","citing_title":"MGUP: A Momentum-Gradient Alignment Update Policy for Stochastic Optimization","ref_index":28,"is_internal_anchor":false},{"citing_arxiv_id":"2606.09658","citing_title":"Muon Learns More Robust and Transferable Features than Adam","ref_index":45,"is_internal_anchor":false},{"citing_arxiv_id":"2606.04662","citing_title":"Why Muon Outperforms Adam: A Curvature Perspective","ref_index":200,"is_internal_anchor":false},{"citing_arxiv_id":"2003.00295","citing_title":"Adaptive Federated Optimization","ref_index":196,"is_internal_anchor":false},{"citing_arxiv_id":"2509.15816","citing_title":"On the Convergence of Muon and Beyond","ref_index":54,"is_internal_anchor":false},{"citing_arxiv_id":"2604.08742","citing_title":"Adam-HNAG: A Convergent Reformulation of Adam with Accelerated Rate","ref_index":8,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/3W73LM64AYQQ46POITXWQ35ETA","json":"https://pith.science/pith/3W73LM64AYQQ46POITXWQ35ETA.json","graph_json":"https://pith.science/api/pith-number/3W73LM64AYQQ46POITXWQ35ETA/graph.json","events_json":"https://pith.science/api/pith-number/3W73LM64AYQQ46POITXWQ35ETA/events.json","paper":"https://pith.science/paper/3W73LM64"},"agent_actions":{"view_html":"https://pith.science/pith/3W73LM64AYQQ46POITXWQ35ETA","download_json":"https://pith.science/pith/3W73LM64AYQQ46POITXWQ35ETA.json","view_paper":"https://pith.science/paper/3W73LM64","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=1808.05671&json=true","fetch_graph":"https://pith.science/api/pith-number/3W73LM64AYQQ46POITXWQ35ETA/graph.json","fetch_events":"https://pith.science/api/pith-number/3W73LM64AYQQ46POITXWQ35ETA/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/3W73LM64AYQQ46POITXWQ35ETA/action/timestamp_anchor","attest_storage":"https://pith.science/pith/3W73LM64AYQQ46POITXWQ35ETA/action/storage_attestation","attest_author":"https://pith.science/pith/3W73LM64AYQQ46POITXWQ35ETA/action/author_attestation","sign_citation":"https://pith.science/pith/3W73LM64AYQQ46POITXWQ35ETA/action/citation_signature","submit_replication":"https://pith.science/pith/3W73LM64AYQQ46POITXWQ35ETA/action/replication_record"}},"created_at":"2026-07-05T08:34:46.094900+00:00","updated_at":"2026-07-05T08:34:46.094900+00:00"}