Research map / Optimization / Distributed and compressed
FedSplit: an algorithmic framework for fast federated optimization
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Use FedSplit instead of FedGD | Federated least squares at κ = 10000 · Least squares, κ = 10000 | iterations to reach ε = 10^-3: 34000 to roughly 400 | Helped |
| Use 10 local gradient steps to approximate each FedSplit proximal update (ablation) | FedSplit · Logistic regression | optimality gap: below 10^{-6} | Helped |
| Use 1 or 5 local gradient steps to approximate each FedSplit proximal update (ablation) | FedSplit · Logistic regression | optimality gap | Hurt |
| Use 1, 5, or 10 local gradient steps instead of exact proximal updates (ablation) | FedSplit · Least squares | optimality gap | Hurt |
| Increase the problem condition number | FedGD and FedSplit · Least squares | iterations to reach ε = 10^-3 | Mixed |
| Use multiple local gradient epochs (e > 1) instead of one (ablation) | FedGD · Least squares | optimality gap | Hurt |
Closest papers by meaning
- Federated Optimization with Doubly Regularized Drift Correction (ICML 2024 · similarity 0.91)
- FedDR – Randomized Douglas-Rachford Splitting Algorithms for Nonconvex Federated Composite Optimization (NeurIPS 2021 · similarity 0.90)
- Stabilized Proximal-Point Methods for Federated Optimization (NeurIPS 2024 · similarity 0.90)
- Methods with Local Steps and Random Reshuffling for Generally Smooth Non-Convex Federated Optimization (ICLR 2025 · similarity 0.89)
- Decomposable Submodular Maximization in Federated Setting (ICML 2024 · similarity 0.88)
- Decentralized Sporadic Federated Learning: A Unified Algorithmic Framework with Convergence Guarantees (ICLR 2025 · similarity 0.88)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.