{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:Z5LNDR3MN7I2Z7D2IOXOZY7UFW","short_pith_number":"pith:Z5LNDR3M","schema_version":"1.0","canonical_sha256":"cf56d1c76c6fd1acfc7a43aeece3f42d997dc9ce61d2706ac16d2e66fa62c0f2","source":{"kind":"arxiv","id":"2506.13974","version":1},"attestation_state":"computed","paper":{"title":"Constant Stepsize Local GD for Logistic Regression: Acceleration by Instability","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Blake Woodworth, Michael Crawshaw, Mingrui Liu","submitted_at":"2025-06-16T20:29:00Z","abstract_excerpt":"Existing analysis of Local (Stochastic) Gradient Descent for heterogeneous objectives requires stepsizes $\\eta \\leq 1/K$ where $K$ is the communication interval, which ensures monotonic decrease of the objective. In contrast, we analyze Local Gradient Descent for logistic regression with separable, heterogeneous data using any stepsize $\\eta > 0$. With $R$ communication rounds and $M$ clients, we show convergence at a rate $\\mathcal{O}(1/\\eta K R)$ after an initial unstable phase lasting for $\\widetilde{\\mathcal{O}}(\\eta K M)$ rounds. This improves upon the existing $\\mathcal{O}(1/R)$ rate for"},"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":"2506.13974","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2025-06-16T20:29:00Z","cross_cats_sorted":[],"title_canon_sha256":"3b27544232d911bc66ebbabd2b0e908ff67fed2256bfcf5803696a2df76d0852","abstract_canon_sha256":"928c6c33b2d5c04d356a88b8f0809ab52169e0abc272008cd8b3925cabf6ac6e"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:22:46.501373Z","signature_b64":"FsJhHIvNzbWXCX5SC115X2Q8M2jyae3LfOETrdULDNdBy1s4/GCxLrcpqqrT/aXMbI338KM+COMaJ7L58KyTDA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"cf56d1c76c6fd1acfc7a43aeece3f42d997dc9ce61d2706ac16d2e66fa62c0f2","last_reissued_at":"2026-07-05T11:22:46.500895Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:22:46.500895Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Constant Stepsize Local GD for Logistic Regression: Acceleration by Instability","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Blake Woodworth, Michael Crawshaw, Mingrui Liu","submitted_at":"2025-06-16T20:29:00Z","abstract_excerpt":"Existing analysis of Local (Stochastic) Gradient Descent for heterogeneous objectives requires stepsizes $\\eta \\leq 1/K$ where $K$ is the communication interval, which ensures monotonic decrease of the objective. In contrast, we analyze Local Gradient Descent for logistic regression with separable, heterogeneous data using any stepsize $\\eta > 0$. With $R$ communication rounds and $M$ clients, we show convergence at a rate $\\mathcal{O}(1/\\eta K R)$ after an initial unstable phase lasting for $\\widetilde{\\mathcal{O}}(\\eta K M)$ rounds. This improves upon the existing $\\mathcal{O}(1/R)$ rate for"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2506.13974","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/2506.13974/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":"2506.13974","created_at":"2026-07-05T11:22:46.500952+00:00"},{"alias_kind":"arxiv_version","alias_value":"2506.13974v1","created_at":"2026-07-05T11:22:46.500952+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2506.13974","created_at":"2026-07-05T11:22:46.500952+00:00"},{"alias_kind":"pith_short_12","alias_value":"Z5LNDR3MN7I2","created_at":"2026-07-05T11:22:46.500952+00:00"},{"alias_kind":"pith_short_16","alias_value":"Z5LNDR3MN7I2Z7D2","created_at":"2026-07-05T11:22:46.500952+00:00"},{"alias_kind":"pith_short_8","alias_value":"Z5LNDR3M","created_at":"2026-07-05T11:22:46.500952+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.31386","citing_title":"Stepsize Hedging: an Alternative Mechanism for Accelerating Gradient Descent","ref_index":15,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/Z5LNDR3MN7I2Z7D2IOXOZY7UFW","json":"https://pith.science/pith/Z5LNDR3MN7I2Z7D2IOXOZY7UFW.json","graph_json":"https://pith.science/api/pith-number/Z5LNDR3MN7I2Z7D2IOXOZY7UFW/graph.json","events_json":"https://pith.science/api/pith-number/Z5LNDR3MN7I2Z7D2IOXOZY7UFW/events.json","paper":"https://pith.science/paper/Z5LNDR3M"},"agent_actions":{"view_html":"https://pith.science/pith/Z5LNDR3MN7I2Z7D2IOXOZY7UFW","download_json":"https://pith.science/pith/Z5LNDR3MN7I2Z7D2IOXOZY7UFW.json","view_paper":"https://pith.science/paper/Z5LNDR3M","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2506.13974&json=true","fetch_graph":"https://pith.science/api/pith-number/Z5LNDR3MN7I2Z7D2IOXOZY7UFW/graph.json","fetch_events":"https://pith.science/api/pith-number/Z5LNDR3MN7I2Z7D2IOXOZY7UFW/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/Z5LNDR3MN7I2Z7D2IOXOZY7UFW/action/timestamp_anchor","attest_storage":"https://pith.science/pith/Z5LNDR3MN7I2Z7D2IOXOZY7UFW/action/storage_attestation","attest_author":"https://pith.science/pith/Z5LNDR3MN7I2Z7D2IOXOZY7UFW/action/author_attestation","sign_citation":"https://pith.science/pith/Z5LNDR3MN7I2Z7D2IOXOZY7UFW/action/citation_signature","submit_replication":"https://pith.science/pith/Z5LNDR3MN7I2Z7D2IOXOZY7UFW/action/replication_record"}},"created_at":"2026-07-05T11:22:46.500952+00:00","updated_at":"2026-07-05T11:22:46.500952+00:00"}