Research map / Federated learning and privacy / Personalized and heterogeneous FL
Homophily-Heterogeneity Gradient Surgery for Federated Graph Learning
Sujia Huang, Lele Fu, Shunxin Xiao, Xiaoya Zhang, Chunyan Xu, Tong Zhang, Bo Huang, Zhen Cui
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 |
|---|---|---|---|
| add inter-group residual projection gradient exchange to isolated groups (ablation) | FedGCM · CiteSeer | accuracy: 72.4 -> 74.6 | Improved |
| add inter-group residual projection gradient exchange to isolated groups (ablation) | FedGCM · Chameleon | accuracy: 66.5 -> 68.6 | Improved |
| add inter-group gradient exchange to isolated groups (ablation) | FedGCM · Cora | accuracy: 86.0 -> 86.1 | No clear effect |
| add inter-group gradient exchange to isolated groups (ablation) | FedGCM · Squirrel | accuracy: 51.1 -> 51.2 | No clear effect |
| add inter-group gradient exchange to isolated groups (ablation) | FedGCM · Penn94 | accuracy: 75.6 -> 76.6 | Improved |
| add projection instead of direct summation across groups (ablation) | FedGCM · Squirrel | accuracy: 41.6 -> 51.2 | Improved |
| add projection instead of direct summation across groups (ablation) | FedGCM · Amazon-ratings | accuracy: 39.3 -> 53.4 | Improved |
| add projection instead of direct summation across groups (ablation) | FedGCM · Cora | accuracy: 81.6 -> 86.1 | Improved |
| direct summation without projection compared to isolation (ablation) | FedGCM variant · CiteSeer | accuracy: 72.4 -> 73.5 | No clear effect |
| add homophily-based grouping and group broadcast to FedAvg (ablation) | FedAvg · six datasets | accuracy | Improved |
| use RPGrad instead of PCGrad (ablation) | FedGCM · six datasets | accuracy | Improved |
| use group-oriented broadcast instead of uniform broadcast (ablation) | FedGCM · six datasets | accuracy | Improved |
Closest papers by meaning
- FedSPA: Generalizable Federated Graph Learning under Homophily Heterogeneity (CVPR 2025 · similarity 0.93)
- Less is More: Federated Graph Learning with Alleviating Topology Heterogeneity from A Causal Perspective (ICML 2025 · similarity 0.90)
- FedSSP: Federated Graph Learning with Spectral Knowledge and Personalized Preference (NeurIPS 2024 · similarity 0.89)
- Federated Graph Condensation with Information Bottleneck Principles (AAAI 2025 · similarity 0.88)
- Modeling Inter-Intra Heterogeneity for Graph Federated Learning (AAAI 2025 · similarity 0.88)
- FedGOG: Federated Graph Out-of-Distribution Generalization with Diffusion Data Exploration and Latent Embedding Decorrelation (AAAI 2025 · 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.
