Research map / Federated learning and privacy / Personalized and heterogeneous FL
Prototype Normalization: Optimizing Prototype Separation for Heterogeneous Federated Learning
Daeyoung Choi, Jihwan Shin, Gyuejeong Lee
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 |
|---|---|---|---|
| replace FedProto weighted averaging with ProtoNorm hyperspherical alignment plus upscaling | FedProto · CIFAR-100 practical alpha=0.1 | test accuracy (%): 41.31 -> 47.94 | Improved |
| ProtoNorm vs FedTGP | prototype-based FL · Tiny ImageNet pathological | test accuracy (%): 31.85 -> 40.24 | Improved |
| ProtoNorm vs FedKD on Tiny ImageNet practical | heterogeneous FL · Tiny ImageNet practical alpha=0.1 (large-scale models) | test accuracy (%): 28.87 -> 29.97 | Improved |
| ProtoNorm vs LG-FedAvg/FedKD on CIFAR-10 practical | heterogeneous FL · CIFAR-10 practical alpha=0.1 | test accuracy (%): 88.10 -> 88.95 | Improved |
| ProtoNorm vs FedTGP at 100 clients with 20% sampling | prototype-based FL · CIFAR-100 (100 clients, 20% sampling) | test accuracy (%): 39.20 -> 41.30 | Improved |
| ProtoNorm vs FedTGP across Dirichlet alpha 0.01 | prototype-based FL · CIFAR-100 (Dirichlet alpha=0.01) | test accuracy (%): 72.67 -> 72.95 | Improved |
| ProtoNorm vs FedTGP on ImageNet-1K 32x32 | prototype-based FL · ImageNet-1K 32x32 (10 clients) | test accuracy (%): 13.62 -> 21.52 | Improved |
| ProtoNorm vs FedUV in homogeneous setting | hyperspherical FL · CIFAR-100 (homogeneous ResNet-18) | test accuracy (%): 48.33 -> 52.67 | Improved |
| add prototype alignment and upscaling together (ablation) | FedProto · Flowers-102 (ablation) | test accuracy (%): 41.22 -> 57.52 | Improved |
| add prototype alignment alone (unit-norm global prototypes) (ablation) | FedProto · Flowers-102 (ablation) | test accuracy (%): 41.22 -> 37.32 | Hurt |
| add prototype alignment alone (ablation) | FedProto · Tiny ImageNet (ablation) | test accuracy (%): 17.88 -> 18.26 | No clear effect |
| add prototype upscaling alone (ablation) | FedProto · CIFAR-100 (ablation) | test accuracy (%): 41.31 -> 40.26 | No clear effect |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- FedUV — FedUV: Uniformity and Variance for Heterogeneous Federated Learning (CVPR 2024) · better on 1 of 1 shared result
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
- Taming Cross-Domain Representation Variance in Federated Prototype Learning with Heterogeneous Data Domains (NeurIPS 2024 · similarity 0.89)
- Beyond Federated Prototype Learning: Learnable Semantic Anchors with Hyperspherical Contrast for Domain-Skewed Data (AAAI 2025 · similarity 0.89)
- FedSA: A Unified Representation Learning via Semantic Anchors for Prototype-based Federated Learning (AAAI 2025 · similarity 0.88)
- FedTGP: Trainable Global Prototypes with Adaptive-Margin-Enhanced Contrastive Learning for Data and Model Heterogeneity in Federated Learning (AAAI 2024 · similarity 0.87)
- FedQS: Optimizing Gradient and Model Aggregation for Semi-Asynchronous Federated Learning (NeurIPS 2025 · similarity 0.86)
- Convergence Analysis of Split Federated Learning on Heterogeneous Data (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.
