Skip to main navigation menu Skip to main content Skip to site footer

Electrotechnical and Computer Engineering

Vol. 39 No. 11 (2024): Proceedings of Faculty of Technical Sciences

IMPLEMENTING FEDERATED LEARNING FOR BINARY CLASSIFICATION SOLUTIONS

  • Andrija Stanišić
DOI:
https://doi.org/10.24867/29BE22Stanisic
Submitted
November 5, 2024
Published
2024-11-05

Abstract

This paper presents a solution that employs federated learning to address binary classification problems. The theoretical foundations of distributed systems and federated learning are primarily analyzed, followed by the implementation of a system for federated neural network training. The developed system enables the allocation and aggregation of model weights across multiple distributed nodes, thereby eliminating risks associated with centralized data storage. The paper includes a review of the employed technologies, a description of the federated system's structure, and an evaluation of the model. This approach significantly contributes to the understanding of federated learning and can serve as a foundation for further research and the development of more advanced distributed systems.

References

[1] George F Coulouris, Jean Dollimore, and Tim Kindberg. Distributed systems: concepts and design. pearson education, 2005.
[2] Brendan McMahan, Eider Moore, Daniel Ramage, Seth Hampson, and Blaise Aguery Arcas. Communication-efficient learning of deep networks from decentralized data. In Artificial intelligence and statistics, pages 1273--1282. PMLR, 2017.
[3] Maarten Van Steen and Andrew S Tanenbaum. Distributed systems. Maarten van Steen Leiden, The Netherlands, 2017.