Blockchain-Enabled Privacy-Preserving Second-Order Federated Edge Learning in Personalized Healthcare

TL;DR

This paper tackles the privacy and personalization challenges of deploying federated learning in edge healthcare devices by introducing BFEL, a blockchain-enabled, second-order FEL framework built on optimized FedCurv. By leveraging the Fisher Information Matrix for curvature-based regularization, BFEL preserves task-relevant parameters across heterogeneous clients and reduces communication rounds, while an Ethereum-based secure aggregation layer provides auditability, verifiability, and privacy. The framework integrates edge clients, a blockchain layer, and cloud servers to support secure model updates, consent management, and policy enforcement, demonstrated on CNN/MLP benchmarks (MNIST, PathMnist, CIFAR-10). Experimental results show improved edge-client performance and robust convergence under non-iid data, with scalable, privacy-preserving operations enabled by Layer 2 blockchain technologies and cryptographic security, making BFEL suitable for wearable-enabled personalized healthcare at the edge.

Abstract

Federated learning (FL) is increasingly recognised for addressing security and privacy concerns in traditional cloud-centric machine learning (ML), particularly within personalised health monitoring such as wearable devices. By enabling global model training through localised policies, FL allows resource-constrained wearables to operate independently. However, conventional first-order FL approaches face several challenges in personalised model training due to the heterogeneous non-independent and identically distributed (non-iid) data by each individual's unique physiology and usage patterns. Recently, second-order FL approaches maintain the stability and consistency of non-iid datasets while improving personalised model training. This study proposes and develops a verifiable and auditable optimised second-order FL framework BFEL (blockchain enhanced federated edge learning) based on optimised FedCurv for personalised healthcare systems. FedCurv incorporates information about the importance of each parameter to each client's task (through fisher information matrix) which helps to preserve client-specific knowledge and reduce model drift during aggregation. Moreover, it minimizes communication rounds required to achieve a target precision convergence for each client device while effectively managing personalised training on non-iid and heterogeneous data. The incorporation of ethereum-based model aggregation ensures trust, verifiability, and auditability while public key encryption enhances privacy and security. Experimental results of federated CNNs and MLPs utilizing mnist, cifar-10, and PathMnist demonstrate framework's high efficiency, scalability, suitability for edge deployment on wearables, and significant reduction in communication cost.

Figures (11)

• Figure 1: Left: Weight divergence in FedAvg due to data heterogeneity. Right: FedCurv induces a regularization parameter to minimize divergence to update weights according to less critical parameters
• Figure 2: Hierarchical diagram depicting a sequential picture of complete network
• Figure 3: Edge Client Layer
• Figure 4: FedCurv for Federated Non-iid (a) CNN PathMnist (b) CNN Mnist (c) CNN Cifar (d) MLP PathMnist (e) MLP Mnist (f) MLP Cifar
• Figure 5: FedAvg for Federated (a) Non-iid Fed CNN PathMnist (b) Non-iid Fed CNN Mnist (c) Non-iid Fed CNN Cifar, (d) IId Fed CNN PathMnist (e) IId Fed CNN Mnist (f) IId Fed CNN Cifar