{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:KV3KCS2XHLKSU57CV45YXHOGLF","short_pith_number":"pith:KV3KCS2X","schema_version":"1.0","canonical_sha256":"5576a14b573ad52a77e2af3b8b9dc6596a4e5f3e39f3123513beae59c9547452","source":{"kind":"arxiv","id":"2109.14119","version":2},"attestation_state":"computed","paper":{"title":"Stochastic Training is Not Necessary for Generalization","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["math.OC"],"primary_cat":"cs.LG","authors_text":"Jonas Geiping, Micah Goldblum, Michael Moeller, Phillip E. Pope, Tom Goldstein","submitted_at":"2021-09-29T00:50:00Z","abstract_excerpt":"It is widely believed that the implicit regularization of SGD is fundamental to the impressive generalization behavior we observe in neural networks. In this work, we demonstrate that non-stochastic full-batch training can achieve comparably strong performance to SGD on CIFAR-10 using modern architectures. To this end, we show that the implicit regularization of SGD can be completely replaced with explicit regularization even when comparing against a strong and well-researched baseline. Our observations indicate that the perceived difficulty of full-batch training may be the result of its opti"},"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":"2109.14119","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2021-09-29T00:50:00Z","cross_cats_sorted":["math.OC"],"title_canon_sha256":"4e55fee54d105b38e8ca7c9d86a700903077f61b3dd2a7fd88c7790be331f4ef","abstract_canon_sha256":"08eed64899fc4c5fa24c8dcb6c84d698c502e53e061b6aa0cb509f88ab403097"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T04:16:15.283406Z","signature_b64":"fQKBNDhTsyq0b4PBQYq5NQGuiQyOoVHtA90r8QzimFhkbKK9XdimAoGIUltvSEim7XJKP57s/ouB6M9d2Y8rBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"5576a14b573ad52a77e2af3b8b9dc6596a4e5f3e39f3123513beae59c9547452","last_reissued_at":"2026-07-05T04:16:15.282915Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T04:16:15.282915Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Stochastic Training is Not Necessary for Generalization","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["math.OC"],"primary_cat":"cs.LG","authors_text":"Jonas Geiping, Micah Goldblum, Michael Moeller, Phillip E. Pope, Tom Goldstein","submitted_at":"2021-09-29T00:50:00Z","abstract_excerpt":"It is widely believed that the implicit regularization of SGD is fundamental to the impressive generalization behavior we observe in neural networks. In this work, we demonstrate that non-stochastic full-batch training can achieve comparably strong performance to SGD on CIFAR-10 using modern architectures. To this end, we show that the implicit regularization of SGD can be completely replaced with explicit regularization even when comparing against a strong and well-researched baseline. Our observations indicate that the perceived difficulty of full-batch training may be the result of its opti"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2109.14119","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/2109.14119/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":"2109.14119","created_at":"2026-07-05T04:16:15.282972+00:00"},{"alias_kind":"arxiv_version","alias_value":"2109.14119v2","created_at":"2026-07-05T04:16:15.282972+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2109.14119","created_at":"2026-07-05T04:16:15.282972+00:00"},{"alias_kind":"pith_short_12","alias_value":"KV3KCS2XHLKS","created_at":"2026-07-05T04:16:15.282972+00:00"},{"alias_kind":"pith_short_16","alias_value":"KV3KCS2XHLKSU57C","created_at":"2026-07-05T04:16:15.282972+00:00"},{"alias_kind":"pith_short_8","alias_value":"KV3KCS2X","created_at":"2026-07-05T04:16:15.282972+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2509.01972","citing_title":"Knowledge distillation as a pathway toward next-generation intelligent ecohydrological modeling systems","ref_index":60,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/KV3KCS2XHLKSU57CV45YXHOGLF","json":"https://pith.science/pith/KV3KCS2XHLKSU57CV45YXHOGLF.json","graph_json":"https://pith.science/api/pith-number/KV3KCS2XHLKSU57CV45YXHOGLF/graph.json","events_json":"https://pith.science/api/pith-number/KV3KCS2XHLKSU57CV45YXHOGLF/events.json","paper":"https://pith.science/paper/KV3KCS2X"},"agent_actions":{"view_html":"https://pith.science/pith/KV3KCS2XHLKSU57CV45YXHOGLF","download_json":"https://pith.science/pith/KV3KCS2XHLKSU57CV45YXHOGLF.json","view_paper":"https://pith.science/paper/KV3KCS2X","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2109.14119&json=true","fetch_graph":"https://pith.science/api/pith-number/KV3KCS2XHLKSU57CV45YXHOGLF/graph.json","fetch_events":"https://pith.science/api/pith-number/KV3KCS2XHLKSU57CV45YXHOGLF/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/KV3KCS2XHLKSU57CV45YXHOGLF/action/timestamp_anchor","attest_storage":"https://pith.science/pith/KV3KCS2XHLKSU57CV45YXHOGLF/action/storage_attestation","attest_author":"https://pith.science/pith/KV3KCS2XHLKSU57CV45YXHOGLF/action/author_attestation","sign_citation":"https://pith.science/pith/KV3KCS2XHLKSU57CV45YXHOGLF/action/citation_signature","submit_replication":"https://pith.science/pith/KV3KCS2XHLKSU57CV45YXHOGLF/action/replication_record"}},"created_at":"2026-07-05T04:16:15.282972+00:00","updated_at":"2026-07-05T04:16:15.282972+00:00"}