Research map / Federated learning and privacy / Communication-efficient FL
HeteroFL: Computation and Communication Efficient Federated Learning for Heterogeneous Clients
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 |
|---|---|---|---|
| Train heterogeneous clients with nested-width subnetworks and aggregate shared parameter subsets | HeteroFL · CIFAR10 IID | accuracy: 69.76 -> 91.19 | Helped |
| Use mixed computation levels a and e instead of all clients at level a | HeteroFL · CIFAR10 IID | accuracy: 91.19 -> 90.29 | Hurt |
| Use mixed computation levels a and e instead of all clients at level a | HeteroFL · CIFAR10 IID | Parameters: 9.6 M -> 4.8 M | Helped |
| Use sBN instead of InstanceNorm (ablation) | Model a · CIFAR10 IID | accuracy: 87.7 -> 91.7 | Helped |
| Use sBN instead of InstanceNorm (ablation) | Model e · CIFAR10 IID | accuracy: 66.4 -> 77.0 | Helped |
| Add Masked CrossEntropy (ablation) | HeteroFL a-e with sBN · CIFAR10 non-IID Global | accuracy: 41.3 -> 60.4 | Helped |
| Add Masked CrossEntropy (ablation) | HeteroFL a-e with sBN · MNIST non-IID Local | accuracy: 99.2 -> 99.8 | Helped |
| Add Scaler (ablation) | HeteroFL a-e with sBN · CIFAR10 IID | accuracy: 90.1 -> 90.1 | No clear effect |
Closest papers by meaning
- Federated Learning with Heterogeneous LLMs: Integrating Small Student Client Models with a Large Hungry Model (AAAI 2025 · similarity 0.92)
- One-Shot Heterogeneous Federated Learning with Local Model-Guided Diffusion Models (ICML 2025 · similarity 0.91)
- Low Precision Local Training is Enough for Federated Learning (NeurIPS 2024 · similarity 0.91)
- Federated Model Heterogeneous Matryoshka Representation Learning (NeurIPS 2024 · similarity 0.90)
- FedLMT: Tackling System Heterogeneity of Federated Learning via Low-Rank Model Training with Theoretical Guarantees (ICML 2024 · similarity 0.90)
- FedLPS: Heterogeneous Federated Learning for Multiple Tasks with Local Parameter Sharing (AAAI 2024 · similarity 0.90)
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.