{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:VUM5VXBWMYCMSR6AKKT6ZXP445","short_pith_number":"pith:VUM5VXBW","schema_version":"1.0","canonical_sha256":"ad19dadc366604c947c052a7ecddfce76f518255ce151825ddf840db5805cc86","source":{"kind":"arxiv","id":"2504.04274","version":1},"attestation_state":"computed","paper":{"title":"Randomised Splitting Methods and Stochastic Gradient Descent","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.NA","math.NA","stat.ML"],"primary_cat":"math.OC","authors_text":"Luke Shaw, Peter A. Whalley","submitted_at":"2025-04-05T20:07:34Z","abstract_excerpt":"We explore an explicit link between stochastic gradient descent using common batching strategies and splitting methods for ordinary differential equations. From this perspective, we introduce a new minibatching strategy (called Symmetric Minibatching Strategy) for stochastic gradient optimisation which shows greatly reduced stochastic gradient bias (from $\\mathcal{O}(h^2)$ to $\\mathcal{O}(h^4)$ in the optimiser stepsize $h$), when combined with momentum-based optimisers. We justify why momentum is needed to obtain the improved performance using the theory of backward analysis for splitting int"},"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":"2504.04274","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"math.OC","submitted_at":"2025-04-05T20:07:34Z","cross_cats_sorted":["cs.NA","math.NA","stat.ML"],"title_canon_sha256":"2f789171b4a840c8fbe027e1b5470bddfed3e5c7681b7ea7c823bdd937e4c591","abstract_canon_sha256":"78a176a8345714d5dfe7ce42ff596ad07269062c12c67cf5591279fc07fd1578"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:45:21.661075Z","signature_b64":"dCsb9w1PI/AZ8mflR0plzufjseFF+ZYDM0GuY/pzom40POLY/xL6UWoteS6qc25peAfb93lhubEQ9dFyT73PDQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"ad19dadc366604c947c052a7ecddfce76f518255ce151825ddf840db5805cc86","last_reissued_at":"2026-07-05T10:45:21.660589Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:45:21.660589Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Randomised Splitting Methods and Stochastic Gradient Descent","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.NA","math.NA","stat.ML"],"primary_cat":"math.OC","authors_text":"Luke Shaw, Peter A. Whalley","submitted_at":"2025-04-05T20:07:34Z","abstract_excerpt":"We explore an explicit link between stochastic gradient descent using common batching strategies and splitting methods for ordinary differential equations. From this perspective, we introduce a new minibatching strategy (called Symmetric Minibatching Strategy) for stochastic gradient optimisation which shows greatly reduced stochastic gradient bias (from $\\mathcal{O}(h^2)$ to $\\mathcal{O}(h^4)$ in the optimiser stepsize $h$), when combined with momentum-based optimisers. We justify why momentum is needed to obtain the improved performance using the theory of backward analysis for splitting int"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2504.04274","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/2504.04274/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":"2504.04274","created_at":"2026-07-05T10:45:21.660647+00:00"},{"alias_kind":"arxiv_version","alias_value":"2504.04274v1","created_at":"2026-07-05T10:45:21.660647+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2504.04274","created_at":"2026-07-05T10:45:21.660647+00:00"},{"alias_kind":"pith_short_12","alias_value":"VUM5VXBWMYCM","created_at":"2026-07-05T10:45:21.660647+00:00"},{"alias_kind":"pith_short_16","alias_value":"VUM5VXBWMYCMSR6A","created_at":"2026-07-05T10:45:21.660647+00:00"},{"alias_kind":"pith_short_8","alias_value":"VUM5VXBW","created_at":"2026-07-05T10:45:21.660647+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.29554","citing_title":"Optimizer Memory Makes Shuffle Order a First-Order Source of Fine-Tuning Noise","ref_index":46,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/VUM5VXBWMYCMSR6AKKT6ZXP445","json":"https://pith.science/pith/VUM5VXBWMYCMSR6AKKT6ZXP445.json","graph_json":"https://pith.science/api/pith-number/VUM5VXBWMYCMSR6AKKT6ZXP445/graph.json","events_json":"https://pith.science/api/pith-number/VUM5VXBWMYCMSR6AKKT6ZXP445/events.json","paper":"https://pith.science/paper/VUM5VXBW"},"agent_actions":{"view_html":"https://pith.science/pith/VUM5VXBWMYCMSR6AKKT6ZXP445","download_json":"https://pith.science/pith/VUM5VXBWMYCMSR6AKKT6ZXP445.json","view_paper":"https://pith.science/paper/VUM5VXBW","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2504.04274&json=true","fetch_graph":"https://pith.science/api/pith-number/VUM5VXBWMYCMSR6AKKT6ZXP445/graph.json","fetch_events":"https://pith.science/api/pith-number/VUM5VXBWMYCMSR6AKKT6ZXP445/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/VUM5VXBWMYCMSR6AKKT6ZXP445/action/timestamp_anchor","attest_storage":"https://pith.science/pith/VUM5VXBWMYCMSR6AKKT6ZXP445/action/storage_attestation","attest_author":"https://pith.science/pith/VUM5VXBWMYCMSR6AKKT6ZXP445/action/author_attestation","sign_citation":"https://pith.science/pith/VUM5VXBWMYCMSR6AKKT6ZXP445/action/citation_signature","submit_replication":"https://pith.science/pith/VUM5VXBWMYCMSR6AKKT6ZXP445/action/replication_record"}},"created_at":"2026-07-05T10:45:21.660647+00:00","updated_at":"2026-07-05T10:45:21.660647+00:00"}