Federated Accelerated Stochastic Gradient Descent
This work addresses communication efficiency in federated learning, offering a provable acceleration that is incremental but with specific performance gains.
The paper tackles the problem of slow convergence and high communication costs in distributed optimization by proposing Federated Accelerated Stochastic Gradient Descent (FedAc), which accelerates Federated Averaging, achieving a reduction in synchronization rounds from M to M^(1/3) for strongly convex and smooth functions.
We propose Federated Accelerated Stochastic Gradient Descent (FedAc), a principled acceleration of Federated Averaging (FedAvg, also known as Local SGD) for distributed optimization. FedAc is the first provable acceleration of FedAvg that improves convergence speed and communication efficiency on various types of convex functions. For example, for strongly convex and smooth functions, when using $M$ workers, the previous state-of-the-art FedAvg analysis can achieve a linear speedup in $M$ if given $M$ rounds of synchronization, whereas FedAc only requires $M^{\frac{1}{3}}$ rounds. Moreover, we prove stronger guarantees for FedAc when the objectives are third-order smooth. Our technique is based on a potential-based perturbed iterate analysis, a novel stability analysis of generalized accelerated SGD, and a strategic tradeoff between acceleration and stability.