FedDANE: A Federated Newton-Type Method
This is an incremental study for federated learning practitioners, highlighting practical limitations of a theoretical method.
The authors tackled the problem of federated learning by proposing FedDANE, a federated Newton-type method adapted from DANE, but found it underperformed baselines like FedAvg and FedProx in realistic settings due to low device participation and statistical heterogeneity.
Federated learning aims to jointly learn statistical models over massively distributed remote devices. In this work, we propose FedDANE, an optimization method that we adapt from DANE, a method for classical distributed optimization, to handle the practical constraints of federated learning. We provide convergence guarantees for this method when learning over both convex and non-convex functions. Despite encouraging theoretical results, we find that the method has underwhelming performance empirically. In particular, through empirical simulations on both synthetic and real-world datasets, FedDANE consistently underperforms baselines of FedAvg and FedProx in realistic federated settings. We identify low device participation and statistical device heterogeneity as two underlying causes of this underwhelming performance, and conclude by suggesting several directions of future work.