{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2019:4PB4CEUUDH4CAGYCQPB563VLSG","short_pith_number":"pith:4PB4CEUU","schema_version":"1.0","canonical_sha256":"e3c3c1129419f8201b0283c3df6eab91bd61c7d784f8b28ce73a2fb1e6d02766","source":{"kind":"arxiv","id":"1905.10395","version":5},"attestation_state":"computed","paper":{"title":"Leader Stochastic Gradient Descent for Distributed Training of Deep Learning Models: Extension","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.DC","math.OC","stat.ML"],"primary_cat":"cs.LG","authors_text":"Adrian Weller, Anna Choromanska, Donald Goldfarb, Francois Chalus, Wenbo Gao, Yunfei Teng","submitted_at":"2019-05-24T18:12:06Z","abstract_excerpt":"We consider distributed optimization under communication constraints for training deep learning models. We propose a new algorithm, whose parameter updates rely on two forces: a regular gradient step, and a corrective direction dictated by the currently best-performing worker (leader). Our method differs from the parameter-averaging scheme EASGD in a number of ways: (i) our objective formulation does not change the location of stationary points compared to the original optimization problem; (ii) we avoid convergence decelerations caused by pulling local workers descending to different local mi"},"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":"1905.10395","kind":"arxiv","version":5},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2019-05-24T18:12:06Z","cross_cats_sorted":["cs.DC","math.OC","stat.ML"],"title_canon_sha256":"e526c27cc6ff4fbb0d2500cbe162c22a5a6c0fd10bee5e76221e9b5fdb69bc42","abstract_canon_sha256":"f89e7ee44a6ce431bf38999bf55a5fcf7f299a0c99d3cafafc6676f488fb8633"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T04:18:25.980986Z","signature_b64":"7xGS8Q2ciNuQqCzx85HBYwauzsoUl6AOCxwp1YzO+2CE5LzcOGTkL+88Xvq0e0UjHUS9aVGoxzDiO9DD3YxwBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"e3c3c1129419f8201b0283c3df6eab91bd61c7d784f8b28ce73a2fb1e6d02766","last_reissued_at":"2026-07-05T04:18:25.980594Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T04:18:25.980594Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Leader Stochastic Gradient Descent for Distributed Training of Deep Learning Models: Extension","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.DC","math.OC","stat.ML"],"primary_cat":"cs.LG","authors_text":"Adrian Weller, Anna Choromanska, Donald Goldfarb, Francois Chalus, Wenbo Gao, Yunfei Teng","submitted_at":"2019-05-24T18:12:06Z","abstract_excerpt":"We consider distributed optimization under communication constraints for training deep learning models. We propose a new algorithm, whose parameter updates rely on two forces: a regular gradient step, and a corrective direction dictated by the currently best-performing worker (leader). Our method differs from the parameter-averaging scheme EASGD in a number of ways: (i) our objective formulation does not change the location of stationary points compared to the original optimization problem; (ii) we avoid convergence decelerations caused by pulling local workers descending to different local mi"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1905.10395","kind":"arxiv","version":5},"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/1905.10395/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":"1905.10395","created_at":"2026-07-05T04:18:25.980651+00:00"},{"alias_kind":"arxiv_version","alias_value":"1905.10395v5","created_at":"2026-07-05T04:18:25.980651+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1905.10395","created_at":"2026-07-05T04:18:25.980651+00:00"},{"alias_kind":"pith_short_12","alias_value":"4PB4CEUUDH4C","created_at":"2026-07-05T04:18:25.980651+00:00"},{"alias_kind":"pith_short_16","alias_value":"4PB4CEUUDH4CAGYC","created_at":"2026-07-05T04:18:25.980651+00:00"},{"alias_kind":"pith_short_8","alias_value":"4PB4CEUU","created_at":"2026-07-05T04:18:25.980651+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2509.03110","citing_title":"LSAM: Asynchronous Distributed Training with Landscape-Smoothed Sharpness-Aware Minimization","ref_index":49,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/4PB4CEUUDH4CAGYCQPB563VLSG","json":"https://pith.science/pith/4PB4CEUUDH4CAGYCQPB563VLSG.json","graph_json":"https://pith.science/api/pith-number/4PB4CEUUDH4CAGYCQPB563VLSG/graph.json","events_json":"https://pith.science/api/pith-number/4PB4CEUUDH4CAGYCQPB563VLSG/events.json","paper":"https://pith.science/paper/4PB4CEUU"},"agent_actions":{"view_html":"https://pith.science/pith/4PB4CEUUDH4CAGYCQPB563VLSG","download_json":"https://pith.science/pith/4PB4CEUUDH4CAGYCQPB563VLSG.json","view_paper":"https://pith.science/paper/4PB4CEUU","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=1905.10395&json=true","fetch_graph":"https://pith.science/api/pith-number/4PB4CEUUDH4CAGYCQPB563VLSG/graph.json","fetch_events":"https://pith.science/api/pith-number/4PB4CEUUDH4CAGYCQPB563VLSG/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/4PB4CEUUDH4CAGYCQPB563VLSG/action/timestamp_anchor","attest_storage":"https://pith.science/pith/4PB4CEUUDH4CAGYCQPB563VLSG/action/storage_attestation","attest_author":"https://pith.science/pith/4PB4CEUUDH4CAGYCQPB563VLSG/action/author_attestation","sign_citation":"https://pith.science/pith/4PB4CEUUDH4CAGYCQPB563VLSG/action/citation_signature","submit_replication":"https://pith.science/pith/4PB4CEUUDH4CAGYCQPB563VLSG/action/replication_record"}},"created_at":"2026-07-05T04:18:25.980651+00:00","updated_at":"2026-07-05T04:18:25.980651+00:00"}