Research map / Federated learning and privacy / Personalized and heterogeneous FL
FedALA: Adaptive Local Aggregation for Personalized Federated Learning
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 |
|---|---|---|---|
| adaptively aggregate the downloaded global model and the previous local model with element-wise weights | FedALA · TINY, practical heterogeneous | test accuracy (%): 37.27$\pm$0.20 -> 40.54$\pm$0.02 | Helped |
| adaptively aggregate the downloaded global model and the previous local model with element-wise weights | FedALA · Cifar100, practical heterogeneous | test accuracy (%): 53.22$\pm$0.20 -> 55.92$\pm$0.03 | Helped |
| apply ALA to FedAvg local initialization | FedAvg · TINY, practical heterogeneous | test accuracy (%): 19.46$\pm$0.20 -> 40.54$\pm$0.17 | Helped |
| apply ALA to FedProx local initialization | FedProx · TINY, practical heterogeneous | test accuracy (%): 19.37$\pm$0.22 -> 40.53$\pm$0.26 | Helped |
| reuse the initial-stage aggregation weights without adapting them (ablation) | FedALA · TINY, practical heterogeneous | test accuracy (%): 40.54 -> 33.81 | Hurt |
| increase the ALA weight-learning sample from 5% to 80% of local data (ablation) | FedALA · TINY*, practical heterogeneous | test accuracy (%): 39.53 -> 41.94 | Helped |
| increase the ALA weight-learning sample from 80% to 100% of local data (ablation) | FedALA · TINY*, practical heterogeneous | test accuracy (%): 41.94 -> 42.11 | No clear effect |
| apply ALA to 2 higher ResNet-18 layers instead of 1 (ablation) | FedALA · TINY*, practical heterogeneous | test accuracy (%): 41.94 -> 42.47 | Helped |
| reduce ALA range from 4 layers to 1 in the four-layer CNN (ablation) | FedALA · TINY, practical heterogeneous | test accuracy (%): 39.75 -> 40.54 | Helped |
| increase ALA range from 1 layer to 3 in fastText (ablation) | FedALA · AG News | test accuracy (%): 96.52 -> 96.45 | No clear effect |
Closest papers by meaning
- FedQS: Optimizing Gradient and Model Aggregation for Semi-Asynchronous Federated Learning (NeurIPS 2025 · similarity 0.88)
- FedAWA: Adaptive Optimization of Aggregation Weights in Federated Learning Using Client Vectors (CVPR 2025 · similarity 0.88)
- An Aggregation-Free Federated Learning for Tackling Data Heterogeneity (CVPR 2024 · similarity 0.87)
- FedAS: Bridging Inconsistency in Personalized Federated Learning (CVPR 2024 · similarity 0.87)
- Adaptive Latent-Space Constraints in Personalized Federated Learning (NeurIPS 2025 · similarity 0.87)
- FedAdaVR: Adaptive Variance Reduction for Robust Federated Learning under Limited Client Participation (NeurIPS 2026 · 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.