<ul><li>Federated Learning (FL) is a framework for distributed machine learning that preserves privacy and enhances security.</li><li>Data heterogeneity poses a challenge for federated learning, as most methods overlook the adjustment of aggregation weights.</li><li>The proposed method, Federated learning with Adaptive Weight Aggregation (FedAWA), adjusts aggregation weights based on client vectors during the learning process.</li><li>FedAWA assigns higher weights to models whose updates align with the global optimization direction, improving stability and generalization of the global model.</li></ul>

FedAWA: Adaptive Optimization of Aggregation Weights in Federated Learning Using Client Vectors

Discover more