Boosting the Performance of Decentralized Federated Learning via Catalyst Acceleration

TL;DR

This paper introduces DFedCata, an accelerated decentralized federated learning algorithm that improves convergence speed and generalization by addressing data heterogeneity with Catalyst acceleration techniques.

Contribution

The paper proposes a novel decentralized federated learning algorithm, DFedCata, combining Moreau envelope and Nesterov's extrapolation to enhance training efficiency and model performance.

Findings

Faster convergence on CIFAR datasets with non-iid data.

Improved generalization performance compared to existing methods.

Theoretical bounds on optimization and generalization errors.

Abstract

Decentralized Federated Learning has emerged as an alternative to centralized architectures due to its faster training, privacy preservation, and reduced communication overhead. In decentralized communication, the server aggregation phase in Centralized Federated Learning shifts to the client side, which means that clients connect with each other in a peer-to-peer manner. However, compared to the centralized mode, data heterogeneity in Decentralized Federated Learning will cause larger variances between aggregated models, which leads to slow convergence in training and poor generalization performance in tests. To address these issues, we introduce Catalyst Acceleration and propose an acceleration Decentralized Federated Learning algorithm called DFedCata. It consists of two main components: the Moreau envelope function, which primarily addresses parameter inconsistencies among clients caused by data heterogeneity, and Nesterov's extrapolation step, which accelerates the aggregation phase. Theoretically, we prove the optimization error bound and generalization error bound of the algorithm, providing a further understanding of the nature of the algorithm and the theoretical perspectives on the hyperparameter choice. Empirically, we demonstrate the advantages of the proposed algorithm in both convergence speed and generalization performance on CIFAR10/100 with various non-iid data distributions. Furthermore, we also experimentally verify the theoretical properties of DFedCata.