About this document
Fed Adp by hoang.nguyen2211093 is a document available to read on EtoBox.
The document presents the Federated Adaptive Weighting (FedAdp) algorithm aimed at improving the convergence speed of Federated Learning (FL) in the presence of non-IID data distributions among participating nodes. By adaptively assigning different weights to nodes based on their contribution to the global model, the FedAdp approach significantly reduces the number of communication rounds required for model convergence, outperforming the traditional Federated Averaging (FedAvg) method. Experimental results
- Author
- hoang.nguyen2211093
- Language
- EN