{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2018:FXYFWSKKEUZ3KZDLGOVK5BUU6K","short_pith_number":"pith:FXYFWSKK","schema_version":"1.0","canonical_sha256":"2df05b494a2533b5646b33aaae8694f2922ab206cd247ae00ad769d5bf0e428f","source":{"kind":"arxiv","id":"1807.06629","version":3},"attestation_state":"computed","paper":{"title":"Parallel Restarted SGD with Faster Convergence and Less Communication: Demystifying Why Model Averaging Works for Deep Learning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.DC","cs.LG"],"primary_cat":"math.OC","authors_text":"Hao Yu, Sen Yang, Shenghuo Zhu","submitted_at":"2018-07-17T19:14:17Z","abstract_excerpt":"In distributed training of deep neural networks, parallel mini-batch SGD is widely used to speed up the training process by using multiple workers. It uses multiple workers to sample local stochastic gradient in parallel, aggregates all gradients in a single server to obtain the average, and update each worker's local model using a SGD update with the averaged gradient. Ideally, parallel mini-batch SGD can achieve a linear speed-up of the training time (with respect to the number of workers) compared with SGD over a single worker. However, such linear scalability in practice is significantly l"},"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":"1807.06629","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"math.OC","submitted_at":"2018-07-17T19:14:17Z","cross_cats_sorted":["cs.DC","cs.LG"],"title_canon_sha256":"798e08eea1ad30da92c06d20a4542745d48af1f7b2c17ec9f6568a9e0f910f47","abstract_canon_sha256":"54faa7e68eb332a41dc71cb44c35b760d32534b5c849d6493a5589fb93183f69"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-18T00:00:35.021606Z","signature_b64":"mbOgb83auq82vxZb+Q1TkmoB4hoM4DEYeD0PjNjtXJb94vYufDX9Q555AwbVMlesUlcGkm7pIm0bMb0QaYD7DQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"2df05b494a2533b5646b33aaae8694f2922ab206cd247ae00ad769d5bf0e428f","last_reissued_at":"2026-05-18T00:00:35.021045Z","signature_status":"signed_v1","first_computed_at":"2026-05-18T00:00:35.021045Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Parallel Restarted SGD with Faster Convergence and Less Communication: Demystifying Why Model Averaging Works for Deep Learning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.DC","cs.LG"],"primary_cat":"math.OC","authors_text":"Hao Yu, Sen Yang, Shenghuo Zhu","submitted_at":"2018-07-17T19:14:17Z","abstract_excerpt":"In distributed training of deep neural networks, parallel mini-batch SGD is widely used to speed up the training process by using multiple workers. It uses multiple workers to sample local stochastic gradient in parallel, aggregates all gradients in a single server to obtain the average, and update each worker's local model using a SGD update with the averaged gradient. Ideally, parallel mini-batch SGD can achieve a linear speed-up of the training time (with respect to the number of workers) compared with SGD over a single worker. However, such linear scalability in practice is significantly l"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1807.06629","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":""},"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":"1807.06629","created_at":"2026-05-18T00:00:35.021116+00:00"},{"alias_kind":"arxiv_version","alias_value":"1807.06629v3","created_at":"2026-05-18T00:00:35.021116+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1807.06629","created_at":"2026-05-18T00:00:35.021116+00:00"},{"alias_kind":"pith_short_12","alias_value":"FXYFWSKKEUZ3","created_at":"2026-05-18T12:32:25.280505+00:00"},{"alias_kind":"pith_short_16","alias_value":"FXYFWSKKEUZ3KZDL","created_at":"2026-05-18T12:32:25.280505+00:00"},{"alias_kind":"pith_short_8","alias_value":"FXYFWSKK","created_at":"2026-05-18T12:32:25.280505+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/FXYFWSKKEUZ3KZDLGOVK5BUU6K","json":"https://pith.science/pith/FXYFWSKKEUZ3KZDLGOVK5BUU6K.json","graph_json":"https://pith.science/api/pith-number/FXYFWSKKEUZ3KZDLGOVK5BUU6K/graph.json","events_json":"https://pith.science/api/pith-number/FXYFWSKKEUZ3KZDLGOVK5BUU6K/events.json","paper":"https://pith.science/paper/FXYFWSKK"},"agent_actions":{"view_html":"https://pith.science/pith/FXYFWSKKEUZ3KZDLGOVK5BUU6K","download_json":"https://pith.science/pith/FXYFWSKKEUZ3KZDLGOVK5BUU6K.json","view_paper":"https://pith.science/paper/FXYFWSKK","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=1807.06629&json=true","fetch_graph":"https://pith.science/api/pith-number/FXYFWSKKEUZ3KZDLGOVK5BUU6K/graph.json","fetch_events":"https://pith.science/api/pith-number/FXYFWSKKEUZ3KZDLGOVK5BUU6K/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/FXYFWSKKEUZ3KZDLGOVK5BUU6K/action/timestamp_anchor","attest_storage":"https://pith.science/pith/FXYFWSKKEUZ3KZDLGOVK5BUU6K/action/storage_attestation","attest_author":"https://pith.science/pith/FXYFWSKKEUZ3KZDLGOVK5BUU6K/action/author_attestation","sign_citation":"https://pith.science/pith/FXYFWSKKEUZ3KZDLGOVK5BUU6K/action/citation_signature","submit_replication":"https://pith.science/pith/FXYFWSKKEUZ3KZDLGOVK5BUU6K/action/replication_record"}},"created_at":"2026-05-18T00:00:35.021116+00:00","updated_at":"2026-05-18T00:00:35.021116+00:00"}