Research map / Federated learning and privacy / Communication-efficient FL
Resource-Adaptive Federated Learning with All-In-One Neural Composition
Yiqun Mei, Pengfei Guo, Mo Zhou, Vishal Patel
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 |
|---|---|---|---|
| use FLANC instead of HeteroFL | HeteroFL · CIFAR10 ResNet-18 dynamic IID, avg | Top-1 accuracy (%): 85.6 -> 87.8 | Improved |
| use FLANC instead of HeteroFL | HeteroFL · CIFAR10 ResNet-18 dynamic Non-IID, avg | Top-1 accuracy (%): 56.2 -> 69.3 | Improved |
| use FLANC instead of FjORD | FjORD · CIFAR10 ResNet-18 dynamic Non-IID, avg | Top-1 accuracy (%): 60.5 -> 69.3 | Improved |
| use FLANC instead of FjORD | FjORD · CIFAR100 ResNet-18 dynamic IID, avg | Top-1 accuracy (%): 54.8 -> 61.4 | Improved |
| use FLANC instead of HeteroFL | HeteroFL · Fashion-MNIST CNN dynamic IID, avg | Top-1 accuracy (%): 90.4 -> 91.1 | Improved |
| use FLANC instead of HeteroFL | HeteroFL · Shakespeare RNN dynamic IID, avg | perplexity: 5.03 -> 4.77 | Improved |
| use FLANC instead of pruning-based p-width models | HeteroFL/FjORD · CIFAR10 ResNet-18 avg | communication cost: 5.2M -> 2.0M | Improved |
| switch from IID to non-IID data partition | HeteroFL and FjORD · CIFAR10 | Top-1 accuracy (%): 85.6 -> 56.2 | Hurt |
| FjORD ordered dropout versus HeteroFL | FjORD · CIFAR10 ResNet-18 non-IID, x1.0 | Top-1 accuracy (%): 56.7 -> 65.7 | Mixed |
| set basis size R1 to 25%, 50%, 100% of smallest incoming channel (ablation) | FLANC · CIFAR100 IID dynamic | accuracy | Mixed |
| increase number of bases R2 from 25% to 50% to 75% width (ablation) | FLANC · CIFAR100 IID dynamic | accuracy | Improved |
| imbalanced device distribution (A 25%/75% with p=1, B p=0.25) | FLANC · CIFAR10 IID dynamic | accuracy | Improved |
Closest papers by meaning
- Towards Instance-adaptive Inference for Federated Learning (ICCV 2023 · similarity 0.89)
- ScaleFL: Resource-Adaptive Federated Learning With Heterogeneous Clients (CVPR 2023 · similarity 0.88)
- Federated Learning with Heterogeneous LLMs: Integrating Small Student Client Models with a Large Hungry Model (AAAI 2025 · similarity 0.87)
- Adaptive Federated Learning with Auto-Tuned Clients (ICLR 2024 · similarity 0.86)
- Resource-Aware Federated Self-Supervised Learning with Global Class Representations (NeurIPS 2024 · similarity 0.86)
- Efficient Federated Learning against Heterogeneous and Non-stationary Client Unavailability (NeurIPS 2024 · 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.
