To date, the most popular federated learning algorithms use coordinate-wise averaging of the model parameters. We depart from this approach by differentiating client contributions according to the performance of local learning and its evolution. The technique is inspired from control theory and its classification performance is evaluated extensively in IID framework and compared with FedAvg.
@article{arxiv.2205.14236,
title = {FedControl: When Control Theory Meets Federated Learning},
author = {Adnan Ben Mansour and Gaia Carenini and Alexandre Duplessis and David Naccache},
journal= {arXiv preprint arXiv:2205.14236},
year = {2022}
}
Comments
arXiv admin note: substantial text overlap with arXiv:2205.10864