Boosting the Performance of Decentralized Federated Learning via Catalyst Acceleration
- 29 April 2026
- journal article
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Pattern Analysis and Machine Intelligence
- Vol. PP (01628828) , 1-12
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, computational cost, and generalization performance on CIFAR10/100 and Tiny-ImageNet with various non-iid data distributions. Moreover, extensive experiments are conducted to validate the theoretical properties of DFedCata, showing strong consistency between theory and empirical observations. Our code is available at https://github.com/zzylyxx/DFedCata.Keywords
All Related Versions
Funding Information
- National Natural Science Foundation of China (62576364)
- Basic and Applied Basic Research Foundation of Guangdong Province (2026B1515020071)
- Shenzhen Municipal Fundamental Research Program (JCYJ20241202124430041)
This publication has 29 references indexed in Scilit:
- Decentralized Directed Collaboration for Personalized Federated Learning2024
- Perada: Parameter-Efficient Federated Learning Personalization with Generalization Guarantees2024
- Decentralized Federated Learning: Fundamentals, State of the Art, Frameworks, Trends, and ChallengesIEEE Communications Surveys & Tutorials, 2023
- Enhancing Decentralized Federated Learning for Non-IID Data on Heterogeneous Devices2023
- Improving Generalization in Federated Learning by Seeking Flat Minima2022
- An Acceleration of Decentralized SGD Under General Assumptions with Low Stochastic Noise2021
- The Role of Network Topology for Distributed Machine Learning2019
- Minimizing finite sums with the stochastic average gradientMathematical Programming, 2016
- ImageNet: A large-scale hierarchical image database2009
- Collective dynamics of ‘small-world’ networksNature, 1998