Surrogate Methods

This paper proposes a unified approach for designing stochastic optimization algorithms that robustly scale to the federated learning …