{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:644ZUBDV7PIDX7RHQ4GWBVW6GO","short_pith_number":"pith:644ZUBDV","schema_version":"1.0","canonical_sha256":"f7399a0475fbd03bfe27870d60d6de33aca6482985a0ad5dbc232237820c6e4d","source":{"kind":"arxiv","id":"2507.12091","version":1},"attestation_state":"computed","paper":{"title":"Improved Analysis for Sign-based Methods with Momentum Updates","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"math.OC","authors_text":"Dingzhi Yu, Lijun Zhang, Sifan Yang, Wei Jiang, Wenhao Yang","submitted_at":"2025-07-16T09:54:08Z","abstract_excerpt":"In this paper, we present enhanced analysis for sign-based optimization algorithms with momentum updates. Traditional sign-based methods, under the separable smoothness assumption, guarantee a convergence rate of $\\mathcal{O}(T^{-1/4})$, but they either require large batch sizes or assume unimodal symmetric stochastic noise. To address these limitations, we demonstrate that signSGD with momentum can achieve the same convergence rate using constant batch sizes without additional assumptions. Our analysis, under the standard $l_2$-smoothness condition, improves upon the result of the prior momen"},"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":"2507.12091","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"math.OC","submitted_at":"2025-07-16T09:54:08Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"3ddf90b02a5b0c7a56b0b219a519b2e98882356ad3adf214ed82b0383366acc0","abstract_canon_sha256":"2ac1b6a69f2a970cecff2ac53fb07ed36028c7220df8a8bd47ff3f9c30d921a7"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:38:14.290055Z","signature_b64":"MwIFHtAKrlebsuGj++mR6v0T8RgGSzLOSmuvGqfsHkIivfQZcDzTCEWmo1fncorDrDWtiuY2dHAA7qC76TawCA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"f7399a0475fbd03bfe27870d60d6de33aca6482985a0ad5dbc232237820c6e4d","last_reissued_at":"2026-07-05T11:38:14.289655Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:38:14.289655Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Improved Analysis for Sign-based Methods with Momentum Updates","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"math.OC","authors_text":"Dingzhi Yu, Lijun Zhang, Sifan Yang, Wei Jiang, Wenhao Yang","submitted_at":"2025-07-16T09:54:08Z","abstract_excerpt":"In this paper, we present enhanced analysis for sign-based optimization algorithms with momentum updates. Traditional sign-based methods, under the separable smoothness assumption, guarantee a convergence rate of $\\mathcal{O}(T^{-1/4})$, but they either require large batch sizes or assume unimodal symmetric stochastic noise. To address these limitations, we demonstrate that signSGD with momentum can achieve the same convergence rate using constant batch sizes without additional assumptions. Our analysis, under the standard $l_2$-smoothness condition, improves upon the result of the prior momen"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2507.12091","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":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2507.12091/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":"2507.12091","created_at":"2026-07-05T11:38:14.289707+00:00"},{"alias_kind":"arxiv_version","alias_value":"2507.12091v1","created_at":"2026-07-05T11:38:14.289707+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2507.12091","created_at":"2026-07-05T11:38:14.289707+00:00"},{"alias_kind":"pith_short_12","alias_value":"644ZUBDV7PID","created_at":"2026-07-05T11:38:14.289707+00:00"},{"alias_kind":"pith_short_16","alias_value":"644ZUBDV7PIDX7RH","created_at":"2026-07-05T11:38:14.289707+00:00"},{"alias_kind":"pith_short_8","alias_value":"644ZUBDV","created_at":"2026-07-05T11:38:14.289707+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":3,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.23676","citing_title":"Open Problem: Is AdamW Effective Under Heavy-Tailed Noise?","ref_index":2,"is_internal_anchor":false},{"citing_arxiv_id":"2605.06615","citing_title":"When and Why SignSGD Outperforms SGD: A Theoretical Study Based on $\\ell_1$-norm Lower Bounds","ref_index":16,"is_internal_anchor":false},{"citing_arxiv_id":"2604.15416","citing_title":"StoSignSGD: Unbiased Structural Stochasticity Fixes SignSGD for Training Large Language Models","ref_index":20,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/644ZUBDV7PIDX7RHQ4GWBVW6GO","json":"https://pith.science/pith/644ZUBDV7PIDX7RHQ4GWBVW6GO.json","graph_json":"https://pith.science/api/pith-number/644ZUBDV7PIDX7RHQ4GWBVW6GO/graph.json","events_json":"https://pith.science/api/pith-number/644ZUBDV7PIDX7RHQ4GWBVW6GO/events.json","paper":"https://pith.science/paper/644ZUBDV"},"agent_actions":{"view_html":"https://pith.science/pith/644ZUBDV7PIDX7RHQ4GWBVW6GO","download_json":"https://pith.science/pith/644ZUBDV7PIDX7RHQ4GWBVW6GO.json","view_paper":"https://pith.science/paper/644ZUBDV","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2507.12091&json=true","fetch_graph":"https://pith.science/api/pith-number/644ZUBDV7PIDX7RHQ4GWBVW6GO/graph.json","fetch_events":"https://pith.science/api/pith-number/644ZUBDV7PIDX7RHQ4GWBVW6GO/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/644ZUBDV7PIDX7RHQ4GWBVW6GO/action/timestamp_anchor","attest_storage":"https://pith.science/pith/644ZUBDV7PIDX7RHQ4GWBVW6GO/action/storage_attestation","attest_author":"https://pith.science/pith/644ZUBDV7PIDX7RHQ4GWBVW6GO/action/author_attestation","sign_citation":"https://pith.science/pith/644ZUBDV7PIDX7RHQ4GWBVW6GO/action/citation_signature","submit_replication":"https://pith.science/pith/644ZUBDV7PIDX7RHQ4GWBVW6GO/action/replication_record"}},"created_at":"2026-07-05T11:38:14.289707+00:00","updated_at":"2026-07-05T11:38:14.289707+00:00"}