Communication-Efficient Learning of Deep Networks from Decentralized Data, Brendan McMahan, Eider Moore, Daniel Ramage, Seth Hampson, Blaise Aguera y Arcas, 2017Proceedings of the 20th International Conference on Artificial Intelligence and Statistics, Vol. 54 (PMLR) - The foundational paper introducing Federated Averaging (FedAvg), outlining its design to address communication bottlenecks and statistical heterogeneity in federated settings.
SCAFFOLD: Stochastic Controlled Averaging for Federated Learning, Sai Praneeth Karimireddy, Satyen Kale, Mehryar Mohri, Sashank Reddi, Sebastian Stich, Ananda Theertha Suresh, 2020Proceedings of the 37th International Conference on Machine Learning, Vol. 119 (PMLR) - Presents SCAFFOLD, an algorithm that uses control variates to correct for client drift due to statistical heterogeneity, offering improved convergence guarantees and practical performance.