Support Vector Based Anomaly Detection in Federated Learning

Anomaly detection plays a crucial role in various domains. However, traditional centralized approaches often encounter challenges related to data privacy. In this context, Federated Learning emerges as a promising solution. This work introduces two algorithms that leverage Support Vector Machines for anomaly detection in a federated setting. In comparison with the Neural Networks typically used in this field, these algorithms emerge as potential alternatives, as they can operate with small datasets and incur lower computational costs. The algorithms are tested in various configurations, yielding promising initial results Specifically, we attain comparable results to the centralized counterpart when the distributed system simulates a centralized setting. A trade-off emerges between split bias and client fraction, indicating that higher client fractions are necessary for optimal performance in scenarios with high bias.