{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:GLH3PFKX7MS4C4KMXDWEFUFZQ2","short_pith_number":"pith:GLH3PFKX","schema_version":"1.0","canonical_sha256":"32cfb79557fb25c1714cb8ec42d0b986be6ef059fa73dcbcaac8cff7b73ecb67","source":{"kind":"arxiv","id":"2605.20866","version":1},"attestation_state":"computed","paper":{"title":"LOSCAR-SGD: Local SGD with Communication-Computation Overlap and Delay-Corrected Sparse Model Averaging","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.DC","math.OC","stat.ML"],"primary_cat":"cs.LG","authors_text":"Ammar Mahran, Artavazd Maranjyan, Peter Richt\\'arik, Yassine Maziane","submitted_at":"2026-05-20T08:01:45Z","abstract_excerpt":"Communication is a major bottleneck in distributed learning, especially in large-scale settings and in federated learning environments with slow links. Three standard ways to reduce this cost are communication compression, local training, and communication-computation overlap. Methods that combine these ingredients are used in practice and have been found to be effective for large-scale training, but there is little theory for methods that combine all three. We study a heterogeneous-compute setting in which different workers may take different numbers of local steps, and we propose LOSCAR-SGD,"},"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":"2605.20866","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-20T08:01:45Z","cross_cats_sorted":["cs.DC","math.OC","stat.ML"],"title_canon_sha256":"07a10def7becf06274cb014669ad09ab78ca6475b3ac25709332e677069a82f2","abstract_canon_sha256":"6e21448bdb68161fa2f770a22be5d4283a9e760f59f3e550ee2b340641c066d7"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-21T01:05:25.290686Z","signature_b64":"d9cy5EhM+uRb+G/U0cHARjcA+tn6v4aBJ4vbM/7zpVFc0khtHim87jDiDqMs17gjmVM145tVxSAzeOa8sGQtBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"32cfb79557fb25c1714cb8ec42d0b986be6ef059fa73dcbcaac8cff7b73ecb67","last_reissued_at":"2026-05-21T01:05:25.290114Z","signature_status":"signed_v1","first_computed_at":"2026-05-21T01:05:25.290114Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"LOSCAR-SGD: Local SGD with Communication-Computation Overlap and Delay-Corrected Sparse Model Averaging","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.DC","math.OC","stat.ML"],"primary_cat":"cs.LG","authors_text":"Ammar Mahran, Artavazd Maranjyan, Peter Richt\\'arik, Yassine Maziane","submitted_at":"2026-05-20T08:01:45Z","abstract_excerpt":"Communication is a major bottleneck in distributed learning, especially in large-scale settings and in federated learning environments with slow links. Three standard ways to reduce this cost are communication compression, local training, and communication-computation overlap. Methods that combine these ingredients are used in practice and have been found to be effective for large-scale training, but there is little theory for methods that combine all three. We study a heterogeneous-compute setting in which different workers may take different numbers of local steps, and we propose LOSCAR-SGD,"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.20866","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/2605.20866/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":"2605.20866","created_at":"2026-05-21T01:05:25.290217+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.20866v1","created_at":"2026-05-21T01:05:25.290217+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.20866","created_at":"2026-05-21T01:05:25.290217+00:00"},{"alias_kind":"pith_short_12","alias_value":"GLH3PFKX7MS4","created_at":"2026-05-21T01:05:25.290217+00:00"},{"alias_kind":"pith_short_16","alias_value":"GLH3PFKX7MS4C4KM","created_at":"2026-05-21T01:05:25.290217+00:00"},{"alias_kind":"pith_short_8","alias_value":"GLH3PFKX","created_at":"2026-05-21T01:05:25.290217+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/GLH3PFKX7MS4C4KMXDWEFUFZQ2","json":"https://pith.science/pith/GLH3PFKX7MS4C4KMXDWEFUFZQ2.json","graph_json":"https://pith.science/api/pith-number/GLH3PFKX7MS4C4KMXDWEFUFZQ2/graph.json","events_json":"https://pith.science/api/pith-number/GLH3PFKX7MS4C4KMXDWEFUFZQ2/events.json","paper":"https://pith.science/paper/GLH3PFKX"},"agent_actions":{"view_html":"https://pith.science/pith/GLH3PFKX7MS4C4KMXDWEFUFZQ2","download_json":"https://pith.science/pith/GLH3PFKX7MS4C4KMXDWEFUFZQ2.json","view_paper":"https://pith.science/paper/GLH3PFKX","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.20866&json=true","fetch_graph":"https://pith.science/api/pith-number/GLH3PFKX7MS4C4KMXDWEFUFZQ2/graph.json","fetch_events":"https://pith.science/api/pith-number/GLH3PFKX7MS4C4KMXDWEFUFZQ2/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/GLH3PFKX7MS4C4KMXDWEFUFZQ2/action/timestamp_anchor","attest_storage":"https://pith.science/pith/GLH3PFKX7MS4C4KMXDWEFUFZQ2/action/storage_attestation","attest_author":"https://pith.science/pith/GLH3PFKX7MS4C4KMXDWEFUFZQ2/action/author_attestation","sign_citation":"https://pith.science/pith/GLH3PFKX7MS4C4KMXDWEFUFZQ2/action/citation_signature","submit_replication":"https://pith.science/pith/GLH3PFKX7MS4C4KMXDWEFUFZQ2/action/replication_record"}},"created_at":"2026-05-21T01:05:25.290217+00:00","updated_at":"2026-05-21T01:05:25.290217+00:00"}