{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:JQXFHUUJA7A5OMWCUPFITYL2E3","short_pith_number":"pith:JQXFHUUJ","schema_version":"1.0","canonical_sha256":"4c2e53d28907c1d732c2a3ca89e17a26ea15cfff1e0014b7dc60b8f0d4c1daff","source":{"kind":"arxiv","id":"2404.01436","version":3},"attestation_state":"computed","paper":{"title":"Convergence Guarantees for RMSProp and Adam in Generalized-smooth Non-convex Optimization with Affine Noise Variance","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG","math.OC"],"primary_cat":"stat.ML","authors_text":"Qi Zhang, Shaofeng Zou, Yi Zhou","submitted_at":"2024-04-01T19:17:45Z","abstract_excerpt":"This paper provides the first tight convergence analyses for RMSProp and Adam in non-convex optimization under the most relaxed assumptions of coordinate-wise generalized smoothness and affine noise variance. We first analyze RMSProp, which is a special case of Adam with adaptive learning rates but without first-order momentum. Specifically, to solve the challenges due to dependence among adaptive update, unbounded gradient estimate and Lipschitz constant, we demonstrate that the first-order term in the descent lemma converges and its denominator is upper bounded by a function of gradient norm"},"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":"2404.01436","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"stat.ML","submitted_at":"2024-04-01T19:17:45Z","cross_cats_sorted":["cs.LG","math.OC"],"title_canon_sha256":"e6070b6a02eebeb96df0b2ee646da2c5b19fa5b2cf7a167c87c4021a0cddb87d","abstract_canon_sha256":"ab71ee4b5a3d05a1ec88c4726d95f8a1ef7e5234205c89843ec071376ec8c4c2"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:27:08.874462Z","signature_b64":"YWeOAmtqCCj0o3rQsC1sZU8LfjuOLIYd/NBxLY9Zwso1O7fo/CZf2yMMhy0JtAxEGN4gJx6RiI8yqFz9AQmDAQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"4c2e53d28907c1d732c2a3ca89e17a26ea15cfff1e0014b7dc60b8f0d4c1daff","last_reissued_at":"2026-07-05T10:27:08.873924Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:27:08.873924Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Convergence Guarantees for RMSProp and Adam in Generalized-smooth Non-convex Optimization with Affine Noise Variance","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG","math.OC"],"primary_cat":"stat.ML","authors_text":"Qi Zhang, Shaofeng Zou, Yi Zhou","submitted_at":"2024-04-01T19:17:45Z","abstract_excerpt":"This paper provides the first tight convergence analyses for RMSProp and Adam in non-convex optimization under the most relaxed assumptions of coordinate-wise generalized smoothness and affine noise variance. We first analyze RMSProp, which is a special case of Adam with adaptive learning rates but without first-order momentum. Specifically, to solve the challenges due to dependence among adaptive update, unbounded gradient estimate and Lipschitz constant, we demonstrate that the first-order term in the descent lemma converges and its denominator is upper bounded by a function of gradient norm"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2404.01436","kind":"arxiv","version":3},"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/2404.01436/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":"2404.01436","created_at":"2026-07-05T10:27:08.873985+00:00"},{"alias_kind":"arxiv_version","alias_value":"2404.01436v3","created_at":"2026-07-05T10:27:08.873985+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2404.01436","created_at":"2026-07-05T10:27:08.873985+00:00"},{"alias_kind":"pith_short_12","alias_value":"JQXFHUUJA7A5","created_at":"2026-07-05T10:27:08.873985+00:00"},{"alias_kind":"pith_short_16","alias_value":"JQXFHUUJA7A5OMWC","created_at":"2026-07-05T10:27:08.873985+00:00"},{"alias_kind":"pith_short_8","alias_value":"JQXFHUUJ","created_at":"2026-07-05T10:27:08.873985+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2509.02912","citing_title":"Stochastic versus Deterministic in Stochastic Gradient Descent","ref_index":16,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/JQXFHUUJA7A5OMWCUPFITYL2E3","json":"https://pith.science/pith/JQXFHUUJA7A5OMWCUPFITYL2E3.json","graph_json":"https://pith.science/api/pith-number/JQXFHUUJA7A5OMWCUPFITYL2E3/graph.json","events_json":"https://pith.science/api/pith-number/JQXFHUUJA7A5OMWCUPFITYL2E3/events.json","paper":"https://pith.science/paper/JQXFHUUJ"},"agent_actions":{"view_html":"https://pith.science/pith/JQXFHUUJA7A5OMWCUPFITYL2E3","download_json":"https://pith.science/pith/JQXFHUUJA7A5OMWCUPFITYL2E3.json","view_paper":"https://pith.science/paper/JQXFHUUJ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2404.01436&json=true","fetch_graph":"https://pith.science/api/pith-number/JQXFHUUJA7A5OMWCUPFITYL2E3/graph.json","fetch_events":"https://pith.science/api/pith-number/JQXFHUUJA7A5OMWCUPFITYL2E3/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/JQXFHUUJA7A5OMWCUPFITYL2E3/action/timestamp_anchor","attest_storage":"https://pith.science/pith/JQXFHUUJA7A5OMWCUPFITYL2E3/action/storage_attestation","attest_author":"https://pith.science/pith/JQXFHUUJA7A5OMWCUPFITYL2E3/action/author_attestation","sign_citation":"https://pith.science/pith/JQXFHUUJA7A5OMWCUPFITYL2E3/action/citation_signature","submit_replication":"https://pith.science/pith/JQXFHUUJA7A5OMWCUPFITYL2E3/action/replication_record"}},"created_at":"2026-07-05T10:27:08.873985+00:00","updated_at":"2026-07-05T10:27:08.873985+00:00"}