Enhancing Federated Learning Through Secure Cluster-Weighted Client Aggregation
Kanishka Ranaweera, Azadeh Ghari Neiat, Xiao Liu, Bipasha Kashyap, Pubudu N. Pathirana
TL;DR
This work tackles fairness and security in federated learning under non-IID data and adversarial threats by introducing ClusterGuardFL, a framework that dynamically clusters clients based on dissimilarity scores and dataset sizes and then weights updates via reconciliation confidence scores using a Softmax mechanism. The approach combines Earth Movers Distance to quantify global-local divergence, K-means clustering to organize participants, and a per-point confidence metric to guide aggregation, with a global aggregation protocol designed to resist data and model poisoning. The paper provides a convergence analysis under convexity and PL conditions, and validates the method empirically on MNIST, Fashion-MNIST, and CIFAR-10 across i.i.d. and non-i.i.d. settings with 20% malicious clients, showing improved robustness and accuracy in adversarial scenarios. The results suggest that secure cluster-weighted aggregation can improve FL robustness and fairness in real-world IoT-like deployments, with potential for integration with additional security layers in future work.
Abstract
Federated learning (FL) has emerged as a promising paradigm in machine learning, enabling collaborative model training across decentralized devices without the need for raw data sharing. In FL, a global model is trained iteratively on local datasets residing on individual devices, each contributing to the model's improvement. However, the heterogeneous nature of these local datasets, stemming from diverse user behaviours, device capabilities, and data distributions, poses a significant challenge. The inherent heterogeneity in federated learning gives rise to various issues, including model performance discrepancies, convergence challenges, and potential privacy concerns. As the global model progresses through rounds of training, the disparities in local data quality and quantity can impede the overall effectiveness of federated learning systems. Moreover, maintaining fairness and privacy across diverse user groups becomes a paramount concern. To address this issue, this paper introduces a novel FL framework, ClusterGuardFL, that employs dissimilarity scores, k-means clustering, and reconciliation confidence scores to dynamically assign weights to client updates. The dissimilarity scores between global and local models guide the formation of clusters, with cluster size influencing the weight allocation. Within each cluster, a reconciliation confidence score is calculated for individual data points, and a softmax layer generates customized weights for clients. These weights are utilized in the aggregation process, enhancing the model's robustness and privacy. Experimental results demonstrate the efficacy of the proposed approach in achieving improved model performance in diverse datasets.