Demystifying Block-Cyclic Sampling for Federated Learning Using MNIST
Arnold S, Fietta D (2026)
Publication Type: Conference contribution
Publication year: 2026
Publisher: Institute of Electrical and Electronics Engineers Inc.
Pages Range: 236-241
Conference Proceedings Title: 2025 3rd International Conference on Federated Learning Technologies and Applications, FLTA 2025
Event location: Dubrovnik
ISBN: 9798331556709
DOI: 10.1109/FLTA67013.2025.11336576
Abstract
Federated Learning (FL) enables statistical models to be built on user-generated data without compromising data security and user privacy. For this reason, FL is well suited for ondevice learning from mobile devices where data is abundant and highly privatized. Constrained by the temporal availability of geodistributed mobile devices, only a subset of devices is accessible to participate in the iterative aggregation of model parameters. In this study, we take a step toward better understanding the effect of non-independent data distributions arising from block-cyclic sampling. Using controlled experiments on visual classification, we measure the effects of block-cyclic sampling (both standalone and in combination with non-balanced block distributions). Specifically, we measure the alterations induced by block-cyclic sampling from the perspective of accuracy, fairness, and convergence rate. Experimental results indicate robustness to cycling over a two-block structure, e.g., due to time zones. In contrast, drawing data samples dependently from a multiblock structure significantly degrades the performance and rate of convergence. Moreover, we find that this performance degeneration is further aggravated by unbalanced block distributions to a point that can no longer be naively compensated by higher communication and more frequent synchronization.
Authors with CRIS profile
How to cite
APA:
Arnold, S., & Fietta, D. (2026). Demystifying Block-Cyclic Sampling for Federated Learning Using MNIST. In Feras M. Awaysheh, Sadi Alawadi (Eds.), 2025 3rd International Conference on Federated Learning Technologies and Applications, FLTA 2025 (pp. 236-241). Dubrovnik, HR: Institute of Electrical and Electronics Engineers Inc..
MLA:
Arnold, Stefan, and Dilara Fietta. "Demystifying Block-Cyclic Sampling for Federated Learning Using MNIST." Proceedings of the 3rd IEEE International Conference on Federated Learning Technologies and Applications, FLTA 2025, Dubrovnik Ed. Feras M. Awaysheh, Sadi Alawadi, Institute of Electrical and Electronics Engineers Inc., 2026. 236-241.
BibTeX: Download