Research map / Federated learning and privacy / Communication-efficient FL
Unified Analyses for Hierarchical Federated Learning: Topology Selection under Data Heterogeneity
Ziyi Zhou, Yipeng Li, Xinchen Lyu
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 |
|---|---|---|---|
| make inter-group distribution Non-IID (ablation) | Star-Ring HFL · CIFAR-10 ResNet-18 | test accuracy (%): 90.30 -> 88.22 | Hurt |
| make intra-group (client) distribution Non-IID (ablation) | Star-Ring HFL · CIFAR-10 ResNet-18 | test accuracy (%): 90.30 -> 89.55 | Hurt |
| make inter-group distribution Non-IID (ablation) | Star-Ring HFL · CINIC-10 ResNet-18 | test accuracy (%): 78.59 -> 72.53 | Hurt |
| make intra-group (client) distribution Non-IID (ablation) | Star-Ring HFL · CINIC-10 ResNet-18 | test accuracy (%): 78.59 -> 76.09 | Hurt |
| make inter-group distribution Non-IID (ablation) | Star-Ring HFL · Fashion-MNIST ResNet-10 | test accuracy (%): 92.59 -> 89.41 | Hurt |
| make intra-group (client) distribution Non-IID (ablation) | Star-Ring HFL · Fashion-MNIST ResNet-10 | test accuracy (%): 92.59 -> 92.33 | No clear effect |
| use ring instead of star at top tier (Ring-Ring vs Star-Ring) | HFL · CIFAR-10 ResNet-18 (inter Non-IID, intra Non-IID) | test accuracy (%): 87.40 -> 90.33 | Improved |
| use ring instead of star at top tier (Ring-Star vs Star-Star) | HFL · CIFAR-10 ResNet-18 (inter IID, intra IID) | test accuracy (%): 88.48 -> 90.40 | Improved |
| use ring top tier (Ring-Ring vs Star-Star) | HFL · SST-2 MLP (inter IID, intra IID) | test accuracy (%): 69.61 -> 81.42 | Improved |
| use more, smaller groups with Ring-Star topology (larger G) (ablation) | Ring-Star HFL · CIFAR-10 (G in 1, 5, 10, 20, 100) | test accuracy (%) | Improved |
| use fewer, larger groups with Star-Ring topology (small G) (ablation) | Star-Ring HFL · CIFAR-10 (G in 1, 5, 10, 20, 100) | test accuracy (%) | Improved |
| increase local steps K from 2 to 10 (ablation) | HFL topologies · CIFAR-10 Non-IID (K in 2, 5, 10) | test accuracy (%) | Improved |
Closest papers by meaning
- Heterogeneity-Guided Client Sampling: Towards Fast and Efficient Non-IID Federated Learning (NeurIPS 2024 · similarity 0.89)
- Hierarchical Federated Learning with Multi-Timescale Gradient Correction (NeurIPS 2024 · similarity 0.88)
- Beyond the Federation: Topology-aware Federated Learning for Generalization to Unseen Clients (ICML 2024 · similarity 0.87)
- Clustered Federated Learning via Gradient-based Partitioning (ICML 2024 · similarity 0.86)
- Enhancing Clustered Federated Learning: Integration of Strategies and Improved Methodologies (ICLR 2025 · similarity 0.86)
- SemiDFL: A Semi-Supervised Paradigm for Decentralized Federated Learning (AAAI 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.
