Research map / Optimization / Distributed and compressed
Federated Conditional Stochastic Optimization
Xidong Wu, Jianhui Sun, Zhengmian Hu, Junyi Li, Aidong Zhang, Heng Huang
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 |
|---|---|---|---|
| add momentum (FCSG-M instead of FCSG) when inner batch is small (ablation) | FCSG · Invariant logistic regression | test accuracy stability | Improved |
| use large inner batch m=100 (momentum vs none) (ablation) | FCSG vs FCSG-M · Invariant logistic regression | test accuracy | No clear effect |
| increase inner batch size m (1, 10, 100) as noise ratio grows (ablation) | FCSG / FCSG-M · Invariant logistic regression, sigma2/sigma1 in {1,1.5,2} | test accuracy | Improved |
| increase shots K from 1 to 5 (ablation) | FCSG family · Omniglot 5-way | training performance | Improved |
| use FCSG family instead of Local-SCGD/Local-SCGDM | Federated MAML · Omniglot 5-way-1-shot and 5-way-5-shot | convergence | Improved |
| FCSG instead of FedAvg (cross-entropy) | federated AUPRC · MNIST | AP: 0.9357 -> 0.9868 | Improved |
| FCSG instead of CODA+ | federated AUPRC · CIFAR-10 | AP: 0.6039 -> 0.7130 | Improved |
| add momentum (FCSG to FCSG-M) (ablation) | FCSG · CIFAR-10 AUPRC | AP: 0.7130 -> 0.7157 | Improved |
| add variance reduction (FCSG-M to Acc-FCSG-M) (ablation) | FCSG-M · CIFAR-10 AUPRC | AP: 0.7157 -> 0.7184 | Improved |
| add variance reduction (FCSG-M to Acc-FCSG-M) (ablation) | FCSG-M · MNIST AUPRC | AP: 0.9878 -> 0.9879 | No clear effect |
Closest papers by meaning
- Solving a Class of Non-Convex Minimax Optimization in Federated Learning (NeurIPS 2023 · similarity 0.88)
- Faster federated optimization under second-order similarity (ICLR 2023 · similarity 0.88)
- Federated Optimization with Doubly Regularized Drift Correction (ICML 2024 · similarity 0.87)
- Stabilized Proximal-Point Methods for Federated Optimization (NeurIPS 2024 · similarity 0.87)
- Methods with Local Steps and Random Reshuffling for Generally Smooth Non-Convex Federated Optimization (ICLR 2025 · similarity 0.86)
- Scaffold with Stochastic Gradients: New Analysis with Linear Speed-Up (ICML 2025 · similarity 0.86)
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.
