{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:ENLRCP4UNQF2DLRK7MYERJ27DP","short_pith_number":"pith:ENLRCP4U","schema_version":"1.0","canonical_sha256":"2357113f946c0ba1ae2afb3048a75f1bec903805079437ddacccfb239920c677","source":{"kind":"arxiv","id":"2406.13936","version":2},"attestation_state":"computed","paper":{"title":"Communication-Efficient Adaptive Batch Size Strategies for Distributed Local Gradient Methods","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG","math.OC"],"primary_cat":"stat.ML","authors_text":"Chenwei Xu, Han Liu, Mladen Kolar, Tim Tsz-Kit Lau, Weijian Li","submitted_at":"2024-06-20T02:08:50Z","abstract_excerpt":"Modern deep neural networks often require distributed training with many workers due to their large size. As the number of workers increases, communication overheads become the main bottleneck in data-parallel minibatch stochastic gradient methods with per-iteration gradient synchronization. Local gradient methods like Local SGD reduce communication by only synchronizing model parameters and/or gradients after several local steps. Despite an understanding of their convergence and the importance of batch sizes for training efficiency and generalization, optimal batch sizes for local gradient me"},"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":"2406.13936","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"stat.ML","submitted_at":"2024-06-20T02:08:50Z","cross_cats_sorted":["cs.LG","math.OC"],"title_canon_sha256":"484af94cfbd6dac8c18c65c3c9f2d8800bbf821120be37f051428052e424829b","abstract_canon_sha256":"f412fa976f3ea334dbc04dd36222d77618e6d122fcbbbe6a690bad237a256636"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:31:46.922715Z","signature_b64":"29p8gCqcw+xzSxABvZ+aJHKIgaOVn4dO8lJom1DCZihzJ1j0/xp5G5EKvU63bnzuT5FFf8mwvihOqUzR3wuxAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"2357113f946c0ba1ae2afb3048a75f1bec903805079437ddacccfb239920c677","last_reissued_at":"2026-07-05T09:31:46.922242Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:31:46.922242Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Communication-Efficient Adaptive Batch Size Strategies for Distributed Local Gradient Methods","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG","math.OC"],"primary_cat":"stat.ML","authors_text":"Chenwei Xu, Han Liu, Mladen Kolar, Tim Tsz-Kit Lau, Weijian Li","submitted_at":"2024-06-20T02:08:50Z","abstract_excerpt":"Modern deep neural networks often require distributed training with many workers due to their large size. As the number of workers increases, communication overheads become the main bottleneck in data-parallel minibatch stochastic gradient methods with per-iteration gradient synchronization. Local gradient methods like Local SGD reduce communication by only synchronizing model parameters and/or gradients after several local steps. Despite an understanding of their convergence and the importance of batch sizes for training efficiency and generalization, optimal batch sizes for local gradient me"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2406.13936","kind":"arxiv","version":2},"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/2406.13936/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":"2406.13936","created_at":"2026-07-05T09:31:46.922296+00:00"},{"alias_kind":"arxiv_version","alias_value":"2406.13936v2","created_at":"2026-07-05T09:31:46.922296+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2406.13936","created_at":"2026-07-05T09:31:46.922296+00:00"},{"alias_kind":"pith_short_12","alias_value":"ENLRCP4UNQF2","created_at":"2026-07-05T09:31:46.922296+00:00"},{"alias_kind":"pith_short_16","alias_value":"ENLRCP4UNQF2DLRK","created_at":"2026-07-05T09:31:46.922296+00:00"},{"alias_kind":"pith_short_8","alias_value":"ENLRCP4U","created_at":"2026-07-05T09:31:46.922296+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2508.18182","citing_title":"AdLoCo: adaptive batching significantly improves communications efficiency and convergence for Large Language Models","ref_index":5,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/ENLRCP4UNQF2DLRK7MYERJ27DP","json":"https://pith.science/pith/ENLRCP4UNQF2DLRK7MYERJ27DP.json","graph_json":"https://pith.science/api/pith-number/ENLRCP4UNQF2DLRK7MYERJ27DP/graph.json","events_json":"https://pith.science/api/pith-number/ENLRCP4UNQF2DLRK7MYERJ27DP/events.json","paper":"https://pith.science/paper/ENLRCP4U"},"agent_actions":{"view_html":"https://pith.science/pith/ENLRCP4UNQF2DLRK7MYERJ27DP","download_json":"https://pith.science/pith/ENLRCP4UNQF2DLRK7MYERJ27DP.json","view_paper":"https://pith.science/paper/ENLRCP4U","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2406.13936&json=true","fetch_graph":"https://pith.science/api/pith-number/ENLRCP4UNQF2DLRK7MYERJ27DP/graph.json","fetch_events":"https://pith.science/api/pith-number/ENLRCP4UNQF2DLRK7MYERJ27DP/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/ENLRCP4UNQF2DLRK7MYERJ27DP/action/timestamp_anchor","attest_storage":"https://pith.science/pith/ENLRCP4UNQF2DLRK7MYERJ27DP/action/storage_attestation","attest_author":"https://pith.science/pith/ENLRCP4UNQF2DLRK7MYERJ27DP/action/author_attestation","sign_citation":"https://pith.science/pith/ENLRCP4UNQF2DLRK7MYERJ27DP/action/citation_signature","submit_replication":"https://pith.science/pith/ENLRCP4UNQF2DLRK7MYERJ27DP/action/replication_record"}},"created_at":"2026-07-05T09:31:46.922296+00:00","updated_at":"2026-07-05T09:31:46.922296+00:00"}