Research map / Federated learning and privacy / Communication-efficient FL
Communication-Efficient Heterogeneous Federated Learning with Sparse Prototypes in Resource-Constrained Environments
Gyuejeong Lee, Daeyoung Choi
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 APS (prototype scaling by class sample count) to TinyProto-FT (ablation) | TinyProto-FT with CPS · Tiny ImageNet-200 | test accuracy (%): 22.73 -> 29.59 | Improved |
| add APS to TinyProto-FT (ablation) | FedTGP · CIFAR-100 | test accuracy (%): 45.30 -> 47.27 | Improved |
| add CPS (class-wise prototype sparsification) alone (ablation) | FedTGP · CIFAR-100 | test accuracy (%): 45.30 -> 43.51 | Hurt |
| add CPS alone (ablation) | FedTGP · CIFAR-10 | test accuracy (%): 87.62 -> 87.68 | No clear effect |
| add CPS on top of APS (ablation) | TinyProto-FT with APS · Tiny ImageNet-200 | test accuracy (%): 28.40 -> 29.59 | Improved |
| integrate CPS and APS into FedProto | FedProto · Flowers-102 | test accuracy (%): 39.85 -> 36.25 | Hurt |
| integrate CPS and APS into FedProto | FedProto · Tiny ImageNet-200 | test accuracy (%): 17.82 -> 20.06 | Improved |
| integrate CPS and APS into FedTGP | FedTGP · Tiny ImageNet-200 | test accuracy (%): 24.70 -> 29.59 | Improved |
| integrate CPS and APS into FedTGP | FedTGP · CIFAR-100 (alpha=0.01) | test accuracy (%): 70.99 -> 74.35 | Improved |
| use CPS sparsification instead of reducing feature dimension (ablation) | FedProto · CIFAR-10 / CIFAR-100 | test accuracy (%) at equal communication | Improved |
Closest papers by meaning
- SpaFL: Communication-Efficient Federated Learning With Sparse Models And Low Computational Overhead (NeurIPS 2024 · similarity 0.89)
- FedSMU: Communication-Efficient and Generalization-Enhanced Federated Learning through Symbolic Model Updates (ICML 2025 · similarity 0.89)
- SparsyFed: Sparse Adaptive Federated Learning (ICLR 2025 · similarity 0.89)
- Taming Cross-Domain Representation Variance in Federated Prototype Learning with Heterogeneous Data Domains (NeurIPS 2024 · similarity 0.89)
- Low Precision Local Training is Enough for Federated Learning (NeurIPS 2024 · similarity 0.89)
- Revisiting Ensembling in One-Shot Federated Learning (NeurIPS 2024 · similarity 0.88)
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.
