Research map / Federated learning and privacy / Personalized and heterogeneous FL
Tackling Data Heterogeneity in Federated Learning with Class Prototypes
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 |
|---|---|---|---|
| Use FedNH's fixed, uniformly initialized prototypes and semantic prototype updates instead of FedROD | Federated image classification · Cifar100, Dir(0.3) | PM(V) accuracy: 28.53 ± 1.27 -> 38.25 ± 1.23 | Helped |
| Use FedNH's fixed, uniformly initialized prototypes and semantic prototype updates instead of FedROD | Federated image classification · Cifar100, Dir(1.0) | GM accuracy: 35.20 ± 0.19 -> 43.19 ± 0.24 | Helped |
| Use FedNH's fixed, uniformly initialized prototypes and semantic prototype updates instead of FedROD | Federated image classification · Cifar10, Dir(0.3) | GM accuracy: 72.31 ± 0.16 -> 69.01 ± 2.51 | Hurt |
| Use FedNH's fixed, uniformly initialized prototypes and semantic prototype updates instead of FedROD | Federated image classification · TinyImageNet, Dir(1.0) | PM(L) accuracy: 38.43 ± 1.76 -> 38.25 ± 2.14 | Hurt |
| Increase client participation ratio from 0.1 to 0.3 | FedNH · Cifar100, Dir(0.3) | GM accuracy: 41.34 ± 0.25 -> 42.13 ± 0.17 | Helped |
| Set prototype smoothing parameter rho to 0.1 rather than higher tested values | FedNH · Sensitivity analysis | relative performance gain | Hurt |
Closest papers by meaning
- FedGPS: Statistical Rectification Against Data Heterogeneity in Federated Learning (NeurIPS 2025 · similarity 0.90)
- Adaptive Latent-Space Constraints in Personalized Federated Learning (NeurIPS 2025 · similarity 0.90)
- Tackling Intertwined Data and Device Heterogeneities in Federated Learning with Unlimited Staleness (AAAI 2025 · similarity 0.90)
- Tackling Feature-Classifier Mismatch in Federated Learning via Prompt-Driven Feature Transformation (NeurIPS 2025 · similarity 0.89)
- Fine-Tuning Personalization in Federated Learning to Mitigate Adversarial Clients (NeurIPS 2024 · similarity 0.89)
- Understanding the Statistical Accuracy-Communication Trade-off in Personalized Federated Learning with Minimax Guarantees (ICML 2025 · similarity 0.89)
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.