Research map / Federated learning and privacy / Communication-efficient FL
Resolving the Tug-of-War: A Separation of Communication and Learning in Federated Learning
Junyi Li, 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 |
|---|---|---|---|
| FedSep adaptive sub-model extraction instead of FedRolex rolling extraction | ResNet-18 model-heterogeneous FL · CIFAR-100 high heterogeneity | test accuracy (%): 56.57 -> 58.16 | Improved |
| FedSep instead of FedRolex | ResNet-18 model-heterogeneous FL · CIFAR-100 low heterogeneity | test accuracy (%): 58.73 -> 61.41 | Improved |
| FedSep instead of FedRolex | ResNet-18 model-heterogeneous FL · CIFAR-10 high heterogeneity | test accuracy (%): 69.44 -> 71.13 | Improved |
| FedSep instead of FedRolex | ResNet-18 model-heterogeneous FL · CIFAR-10 low heterogeneity | test accuracy (%): 84.45 -> 84.61 | No clear effect |
| FedSep instead of Flash | ResNet-18 model-heterogeneous FL · CIFAR-10 low heterogeneity | test accuracy (%): 84.61 -> 84.61 | No clear effect |
| FedSep (data-dependent mask) instead of HeteroFL fixed extraction | ResNet-18 model-heterogeneous FL · CIFAR-10 low heterogeneity | test accuracy (%): 73.19 -> 84.61 | Improved |
| FedSep instead of FedDST | ResNet-18 model-heterogeneous FL · CIFAR-100 low heterogeneity | test accuracy (%): 54.97 -> 61.41 | Improved |
| FedSep (no public data) instead of KD-based Fed-ET | ResNet-18 model-heterogeneous FL · CIFAR-10 high heterogeneity | test accuracy (%): 78.66 -> 71.13 | Hurt |
| FedSep instead of KD-based Fed-ET | ResNet-18 model-heterogeneous FL · CIFAR-100 high heterogeneity | test accuracy (%): 35.78 -> 58.16 | Improved |
| FedSep instead of KD-based Fed-ET | ResNet-18 model-heterogeneous FL · CIFAR-10 low heterogeneity | test accuracy (%): 81.13 -> 84.61 | Improved |
| FedSep instead of Federated Dropout | ResNet-18 model-heterogeneous FL · CIFAR-10 high heterogeneity | test accuracy (%): 46.64 -> 71.13 | Improved |
| Heterogeneous submodels with FedSep versus homogeneous largest model | ResNet-18 FedSep · CIFAR-100 high heterogeneity | test accuracy (%): 60.89 -> 58.16 | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- Flash — Flash: Concept Drift Adaptation in Federated Learning (ICML 2023) · better on 3 of 4 shared results
- FedRolex — FedRolex: Model-Heterogeneous Federated Learning with Rolling Sub-Model Extraction (NeurIPS 2022) · better on 4 of 4 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Achieving Dimension-Free Communication in Federated Learning via Zeroth-Order Optimization (ICLR 2025 · similarity 0.89)
- Bridging Model Heterogeneity in Federated Learning via Uncertainty-based Asymmetrical Reciprocity Learning (ICML 2024 · similarity 0.87)
- FedLMT: Tackling System Heterogeneity of Federated Learning via Low-Rank Model Training with Theoretical Guarantees (ICML 2024 · similarity 0.87)
- Low Precision Local Training is Enough for Federated Learning (NeurIPS 2024 · similarity 0.87)
- FedSMU: Communication-Efficient and Generalization-Enhanced Federated Learning through Symbolic Model Updates (ICML 2025 · similarity 0.87)
- A Bayesian Approach for Personalized Federated Learning in Heterogeneous Settings (NeurIPS 2024 · similarity 0.87)
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.
