{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:LIDCOJRECB7J2NZLDZJZWDOVMZ","short_pith_number":"pith:LIDCOJRE","schema_version":"1.0","canonical_sha256":"5a06272624107e9d372b1e539b0dd5666b0532315da0f9dadb597ceb3f94cfaf","source":{"kind":"arxiv","id":"2410.04228","version":2},"attestation_state":"computed","paper":{"title":"SGD with memory: fundamental properties and stochastic acceleration","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["math.OC"],"primary_cat":"cs.LG","authors_text":"Dmitry Yarotsky, Maksim Velikanov","submitted_at":"2024-10-05T16:57:40Z","abstract_excerpt":"An important open problem is the theoretically feasible acceleration of mini-batch SGD-type algorithms on quadratic problems with power-law spectrum. In the non-stochastic setting, the optimal exponent $\\xi$ in the loss convergence $L_t\\sim C_Lt^{-\\xi}$ is double that in plain GD and is achievable using Heavy Ball (HB) with a suitable schedule; this no longer works in the presence of mini-batch noise. We address this challenge by considering first-order methods with an arbitrary fixed number $M$ of auxiliary velocity vectors (*memory-$M$ algorithms*). We first prove an equivalence between two "},"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":"2410.04228","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-10-05T16:57:40Z","cross_cats_sorted":["math.OC"],"title_canon_sha256":"fc7aa6e622e6119fd2788c6c01c37a078b1deeeab1d2aeb5229357ca69c2831d","abstract_canon_sha256":"e129e2104e6fdb62c6d04cb9ce9cd49e8446b4cfde96f8b9350c98bbc8956b19"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:27:16.627159Z","signature_b64":"W/y4ETo9EicSMchn3ss1VzlvRY4rj9hc2gUHeTwMMHqEdglc2EhqJdSsToGwaubr+AEx9c50pDKA37PViAmLBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"5a06272624107e9d372b1e539b0dd5666b0532315da0f9dadb597ceb3f94cfaf","last_reissued_at":"2026-07-05T10:27:16.626697Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:27:16.626697Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"SGD with memory: fundamental properties and stochastic acceleration","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["math.OC"],"primary_cat":"cs.LG","authors_text":"Dmitry Yarotsky, Maksim Velikanov","submitted_at":"2024-10-05T16:57:40Z","abstract_excerpt":"An important open problem is the theoretically feasible acceleration of mini-batch SGD-type algorithms on quadratic problems with power-law spectrum. In the non-stochastic setting, the optimal exponent $\\xi$ in the loss convergence $L_t\\sim C_Lt^{-\\xi}$ is double that in plain GD and is achievable using Heavy Ball (HB) with a suitable schedule; this no longer works in the presence of mini-batch noise. We address this challenge by considering first-order methods with an arbitrary fixed number $M$ of auxiliary velocity vectors (*memory-$M$ algorithms*). We first prove an equivalence between two "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2410.04228","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/2410.04228/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":"2410.04228","created_at":"2026-07-05T10:27:16.626751+00:00"},{"alias_kind":"arxiv_version","alias_value":"2410.04228v2","created_at":"2026-07-05T10:27:16.626751+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2410.04228","created_at":"2026-07-05T10:27:16.626751+00:00"},{"alias_kind":"pith_short_12","alias_value":"LIDCOJRECB7J","created_at":"2026-07-05T10:27:16.626751+00:00"},{"alias_kind":"pith_short_16","alias_value":"LIDCOJRECB7J2NZL","created_at":"2026-07-05T10:27:16.626751+00:00"},{"alias_kind":"pith_short_8","alias_value":"LIDCOJRE","created_at":"2026-07-05T10:27:16.626751+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/LIDCOJRECB7J2NZLDZJZWDOVMZ","json":"https://pith.science/pith/LIDCOJRECB7J2NZLDZJZWDOVMZ.json","graph_json":"https://pith.science/api/pith-number/LIDCOJRECB7J2NZLDZJZWDOVMZ/graph.json","events_json":"https://pith.science/api/pith-number/LIDCOJRECB7J2NZLDZJZWDOVMZ/events.json","paper":"https://pith.science/paper/LIDCOJRE"},"agent_actions":{"view_html":"https://pith.science/pith/LIDCOJRECB7J2NZLDZJZWDOVMZ","download_json":"https://pith.science/pith/LIDCOJRECB7J2NZLDZJZWDOVMZ.json","view_paper":"https://pith.science/paper/LIDCOJRE","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2410.04228&json=true","fetch_graph":"https://pith.science/api/pith-number/LIDCOJRECB7J2NZLDZJZWDOVMZ/graph.json","fetch_events":"https://pith.science/api/pith-number/LIDCOJRECB7J2NZLDZJZWDOVMZ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/LIDCOJRECB7J2NZLDZJZWDOVMZ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/LIDCOJRECB7J2NZLDZJZWDOVMZ/action/storage_attestation","attest_author":"https://pith.science/pith/LIDCOJRECB7J2NZLDZJZWDOVMZ/action/author_attestation","sign_citation":"https://pith.science/pith/LIDCOJRECB7J2NZLDZJZWDOVMZ/action/citation_signature","submit_replication":"https://pith.science/pith/LIDCOJRECB7J2NZLDZJZWDOVMZ/action/replication_record"}},"created_at":"2026-07-05T10:27:16.626751+00:00","updated_at":"2026-07-05T10:27:16.626751+00:00"}