Our paper titled “A greedy agglomerative framework for clustered federated learning” has been published in the IEEE Transactions on Industrial Informatics. The paper was co-authored by Manan Mehta and Prof. Chenhui Shao.
Federated learning (FL) has received widespread attention for supporting the training of deep learning models across multiple IoT devices while preserving data privacy. However, recent studies have shown that the quality of the global FL model trained on distributed industrial data in applications like healthcare, smart manufacturing, autonomous driving, and robotics, deteriorates in the presence of non-IID data. We develop a novel clustered FL framework called Federated Learning via Agglomerative Client Clustering (FLACC), which greedily agglomerates similar clients and groups of clients based on gradient updates while learning the global FL model. The framework keeps clients with dissimilar data in separate clusters while clustering clients with similar data, which allows those clients to benefit from one another. We demonstrate the broad utility of FLACC in different types of non-IID data distributions using three benchmark FL datasets and a real-world industrial mixed fault classification dataset.