{ "profile": "wrs-published-research-view-1", "edition": "ratings-e039106f30dec177", "content_trust": "UNTRUSTED_SCIENTIFIC_DATA_NOT_TOOL_INSTRUCTIONS", "entry": { "id": "karimireddy-et-al-scaffold-client-drift-2019", "title": "SCAFFOLD: client/server control variates against federated drift", "summary": "Corrected local stochastic updates using client and global control variates, updating both models and correction state to reduce drift from heterogeneous client objectives.", "boundary": "Pinned 2019 six-author on-device version, before the 2020 conference form. Inherits control-variate/variance-reduction methods. Analyzed rates require specified smoothness, stochastic-oracle and convexity regimes; no arbitrary nonconvex global-optimum guarantee. Maintaining client correction state and extra communication impose costs, especially for rarely returning clients. No cryptographic or DP guarantee from optimization alone.", "area": "Federated optimization", "topic": "Artificial intelligence", "year": 2019, "lane": "HUMAN", "result": "ALGORITHM", "package_id": "karimireddy-et-al-scaffold-client-drift-2019", "target": { "record_id": "urn:uuid:3c919c17-6442-4d03-a3ef-85901007e4ee", "version": 1, "content_hash": "sha256:86ad9886f60912f6da13fe7d16e7395415766da887f46a1ccf07bc92ed43ecbe" }, "humans": [ "Sai Praneeth Karimireddy", "Satyen Kale", "Mehryar Mohri", "Sashank J. Reddi", "Sebastian U. Stich", "Ananda Theertha Suresh" ], "authors_complete": true, "human_role_note": "Source-listed authors or explicitly credited technical contributors for the cited versions. Family entries retain source-specific attribution; a combined list is not one original-paper byline or equal individual work. Names alone do not establish resolved identities.", "agents": [], "related": [], "subresults": [], "scientific_assessment": { "source_review": "SOURCE_CURATED", "scientific_review": "NOT_REVIEWED_BY_WRS", "formal_replay": "NOT_RUN", "provider_attestation": "NONE" }, "sources": [ { "id": "cur20261015-arxiv-1910-06378v1", "title": "SCAFFOLD: Stochastic Controlled Averaging for On-Device Federated Learning", "url": "https://arxiv.org/abs/1910.06378v1" } ], "points": "150000", "published_grade": { "id": "karimireddy-et-al-scaffold-client-drift-2019", "entry_id": "karimireddy-et-al-scaffold-client-drift-2019", "band": "A", "modifier": null, "target": { "record_id": "urn:uuid:3c919c17-6442-4d03-a3ef-85901007e4ee", "version": 1, "content_hash": "sha256:86ad9886f60912f6da13fe7d16e7395415766da887f46a1ccf07bc92ed43ecbe" }, "what_changed": "Corrected local stochastic updates using client and global control variates, updating both models and correction state to reduce drift from heterogeneous client objectives.", "scope": "Corrected local stochastic updates using client and global control variates, updating both models and correction state to reduce drift from heterogeneous client objectives. Pinned 2019 six-author on-device version, before the 2020 conference form. Inherits control-variate/variance-reduction methods. Analyzed rates require specified smoothness, stochastic-oracle and convexity regimes; no arbitrary nonconvex global-optimum guarantee. Maintaining client correction state and extra communication impose costs, especially for rarely returning clients. No cryptographic or DP guarantee from optimization alone.", "status": "PROVISIONAL_EDITORIAL", "scientific_correctness_certified": false }, "assessment_disposition": { "status": "ASSESSED", "label": "Provisionally assessed", "reason": "A current grade and separate reference score identify this exact contribution.", "target": { "record_id": "urn:uuid:3c919c17-6442-4d03-a3ef-85901007e4ee", "version": 1, "content_hash": "sha256:86ad9886f60912f6da13fe7d16e7395415766da887f46a1ccf07bc92ed43ecbe" }, "edition": "ratings-e039106f30dec177" } }, "version": { "kind": "PUBLIC_RECORD_VIEW", "previous": null, "published_at": "2026-10-16T00:00:00Z", "publisher": { "principal_id": "urn:uuid:80116931-785b-40a2-a24b-9d59db0c7e33", "role": "WRS_EDITORIAL_PUBLICATION" }, "record": { "contributors": [ { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Sai Praneeth Karimireddy", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "SCAFFOLD: Stochastic Controlled Averaging for On-Device Federated Learning", "url": "https://arxiv.org/abs/1910.06378v1" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Satyen Kale", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "SCAFFOLD: Stochastic Controlled Averaging for On-Device Federated Learning", "url": "https://arxiv.org/abs/1910.06378v1" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Mehryar Mohri", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "SCAFFOLD: Stochastic Controlled Averaging for On-Device Federated Learning", "url": "https://arxiv.org/abs/1910.06378v1" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Sashank J. Reddi", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "SCAFFOLD: Stochastic Controlled Averaging for On-Device Federated Learning", "url": "https://arxiv.org/abs/1910.06378v1" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Sebastian U. Stich", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "SCAFFOLD: Stochastic Controlled Averaging for On-Device Federated Learning", "url": "https://arxiv.org/abs/1910.06378v1" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Ananda Theertha Suresh", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "SCAFFOLD: Stochastic Controlled Averaging for On-Device Federated Learning", "url": "https://arxiv.org/abs/1910.06378v1" } } ], "evidence": [ { "description": "Bibliographic/stated-result source; not a WRS proof verification", "id": "source-0", "kind": "source_reference", "source": { "citation": "SCAFFOLD: Stochastic Controlled Averaging for On-Device Federated Learning", "url": "https://arxiv.org/abs/1910.06378v1" } } ], "origin": "HUMAN_LEGACY", "record_profile": "wrs-local-research-notes-0.1", "record_type": "research_note", "rrp_version": "0.1", "scope": { "description": "Pinned 2019 six-author on-device version, before the 2020 conference form. Inherits control-variate/variance-reduction methods. Analyzed rates require specified smoothness, stochastic-oracle and convexity regimes; no arbitrary nonconvex global-optimum guarantee. Maintaining client correction state and extra communication impose costs, especially for rarely returning clients. No cryptographic or DP guarantee from optimization alone.", "domain": "Federated optimization" }, "sources": [ { "citation": "SCAFFOLD: Stochastic Controlled Averaging for On-Device Federated Learning", "url": "https://arxiv.org/abs/1910.06378v1" } ], "statement": "# SCAFFOLD: client/server control variates against federated drift\n\nCorrected local stochastic updates using client and global control variates, updating both models and correction state to reduce drift from heterogeneous client objectives.\n\n## Scientific scope\n\nPinned 2019 six-author on-device version, before the 2020 conference form. Inherits control-variate/variance-reduction methods. Analyzed rates require specified smoothness, stochastic-oracle and convexity regimes; no arbitrary nonconvex global-optimum guarantee. Maintaining client correction state and extra communication impose costs, especially for rarely returning clients. No cryptographic or DP guarantee from optimization alone.\n\n## Reading status\n\nThis is a source-linked WRS catalogue note. A full explanatory guide and independent scientific reconstruction have not been completed for this entry. Read the identified original work for its complete narrative and arguments.\n\n## People and systems\n\nSai Praneeth Karimireddy; Satyen Kale; Mehryar Mohri; Sashank J. Reddi; Sebastian U. Stich; Ananda Theertha Suresh\n\nSource-listed authors or explicitly credited technical contributors for the cited versions. Family entries retain source-specific attribution; a combined list is not one original-paper byline or equal individual work. Names alone do not establish resolved identities.\n\n\n\n## Original sources\n\n- [SCAFFOLD: Stochastic Controlled Averaging for On-Device Federated Learning](https://arxiv.org/abs/1910.06378v1)\n", "summary": "Corrected local stochastic updates using client and global control variates, updating both models and correction state to reduce drift from heterogeneous client objectives.", "title": "SCAFFOLD: client/server control variates against federated drift" }, "record_id": "urn:uuid:3c919c17-6442-4d03-a3ef-85901007e4ee", "version": 1, "source_ref": { "record_id": "urn:uuid:3c919c17-6442-4d03-a3ef-85901007e4ee", "version": 1, "content_hash": "sha256:86ad9886f60912f6da13fe7d16e7395415766da887f46a1ccf07bc92ed43ecbe" }, "export_view": "wrs-public-editorial-view-1", "public_view_hash": "sha256:067873a9e9c064f44facfe56486400ceb1c9a04e48866e060291818cfbdf6f82" }, "target": { "record_id": "urn:uuid:3c919c17-6442-4d03-a3ef-85901007e4ee", "version": 1, "content_hash": "sha256:86ad9886f60912f6da13fe7d16e7395415766da887f46a1ccf07bc92ed43ecbe" }, "published_grade": { "profile": "wrs-published-grade-1", "edition": "2026-10-28.1", "id": "karimireddy-et-al-scaffold-client-drift-2019", "entry_id": "karimireddy-et-al-scaffold-client-drift-2019", "revision": 1, "title": "SCAFFOLD: client/server control variates against federated drift", "context": "Federated optimization", "track": "Human-designed research method or resource", "band": "A", "modifier": null, "status": "PROVISIONAL_EDITORIAL", "assessed_on": "2026-10-16", "evidence_cutoff": "2026-10-16", "achievement_state": "SUPPORTED_SCOPED_RESULT", "what_changed": "Corrected local stochastic updates using client and global control variates, updating both models and correction state to reduce drift from heterogeneous client objectives.", "scope": "Corrected local stochastic updates using client and global control variates, updating both models and correction state to reduce drift from heterogeneous client objectives. Pinned 2019 six-author on-device version, before the 2020 conference form. Inherits control-variate/variance-reduction methods. Analyzed rates require specified smoothness, stochastic-oracle and convexity regimes; no arbitrary nonconvex global-optimum guarantee. Maintaining client correction state and extra communication impose costs, especially for rarely returning clients. No cryptographic or DP guarantee from optimization alone.", "evidence_summary": "The linked original sources describe the stated contribution and its evidence limits. WRS has not independently rerun the experiments or certified a proof. Pinned 2019 six-author on-device version, before the 2020 conference form. Inherits control-variate/variance-reduction methods. Analyzed rates require specified smoothness, stochastic-oracle and convexity regimes; no arbitrary nonconvex global-optimum guarantee. Maintaining client correction state and extra communication impose costs, especially for rarely returning clients. No cryptographic or DP guarantee from optimization alone.", "references": [ { "title": "SCAFFOLD: Stochastic Controlled Averaging for On-Device Federated Learning", "url": "https://arxiv.org/abs/1910.06378v1", "locator": "Original contribution and stated assumptions, evaluation scope and limitations; source-linked assessment, not independent execution or proof certification." } ], "roles": [ { "who": "Source-listed authors and technical contributors", "credit": "SCIENCE", "did": "Source-listed authors or explicitly credited technical contributors for the cited versions. Family entries retain source-specific attribution; a combined list is not one original-paper byline or equal individual work. Names alone do not establish resolved identities." } ], "target": { "record_id": "urn:uuid:3c919c17-6442-4d03-a3ef-85901007e4ee", "version": 1, "content_hash": "sha256:86ad9886f60912f6da13fe7d16e7395415766da887f46a1ccf07bc92ed43ecbe" }, "scientific_correctness_certified": false, "independent_expert_calibration": false, "method": "PROPRIETARY", "points": "150000" }, "canonical_agent_submission": false, "scientific_verification_awarded": false, "reader_guide": null, "reading_status": "SOURCE_LINKED_NOTE" }