Research map / Federated learning and privacy / Communication-efficient FL
Breaking the Capacity Bottleneck in Model-Heterogeneous Federated Learning via Gradual Model Restoration
Chengjie Ma, Seungeun Oh, Jihong Park, Seong-Lyun Kim
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 |
|---|---|---|---|
| gradually restore sub-model density (GMR) instead of fixed density (ablation) | FedGMR · High heterogeneity, Non-IID, four tasks | test accuracy over aggregation steps | Improved |
| use asynchronous coordination (ablation) | FedGMR · FEMNIST (High heterogeneity, Non-IID) | training speed | Improved |
| use asynchronous coordination (ablation) | FedGMR · ImageNet-100 and StackOverflow | accuracy | No clear effect |
| apply GMR to other MHFL methods (ablation) | HeteroFL/FedRolex/Fjord/Fiarse · FEMNIST | accuracy | Improved |
| pair GMR with MaskFedAvg aggregation instead of gradient averaging or FedAvg (ablation) | FedGMR · CIFAR-10 and ImageNet-100 | accuracy | Improved |
| add GMR under gradient averaging aggregation (ablation) | GA aggregation · CIFAR-10 and ImageNet-100 | accuracy | Hurt |
| use MaskFedAvg without GMR (fixed density) instead of gradient averaging (ablation) | MaskFedAvg · ResNet-18 / deeper models | accuracy | Hurt |
| use naive FedAvg aggregation with zeros for missing coordinates (ablation) | FedAvg aggregation · CIFAR-10 | accuracy: near-chance 10% | Hurt |
| FedGMR vs Fiarse (High heterogeneity IID) | FedGMR · ImageNet-100 | test accuracy: 54.00 -> 60.84 | Improved |
| FedGMR vs Fiarse (High heterogeneity Non-IID) | FedGMR · ImageNet-100 | test accuracy: 48.27 -> 58.01 | Improved |
| FedGMR vs FedRolex (High heterogeneity IID) | FedGMR · ImageNet-100 | test accuracy: 32.17 -> 60.84 | Improved |
| FedGMR vs FedRolex (High heterogeneity Non-IID) | FedGMR · ImageNet-100 | test accuracy: 15.26 -> 58.01 | Improved |
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.
- Fiarse — FIARSE: Model-Heterogeneous Federated Learning via Importance-Aware Submodel Extraction (NeurIPS 2024) · better on 4 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
- FedQS: Optimizing Gradient and Model Aggregation for Semi-Asynchronous Federated Learning (NeurIPS 2025 · similarity 0.90)
- FedSMU: Communication-Efficient and Generalization-Enhanced Federated Learning through Symbolic Model Updates (ICML 2025 · similarity 0.90)
- Federated Model Heterogeneous Matryoshka Representation Learning (NeurIPS 2024 · similarity 0.89)
- FedGPS: Statistical Rectification Against Data Heterogeneity in Federated Learning (NeurIPS 2025 · similarity 0.89)
- Low Precision Local Training is Enough for Federated Learning (NeurIPS 2024 · similarity 0.89)
- Overcoming Data and Model heterogeneities in Decentralized Federated Learning via Synthetic Anchors (ICML 2024 · similarity 0.89)
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.
