Privacy-Preserving Federated Autoencoder for ECG Anomaly Detection on Edge Devices

2026-06-09

AI Summary

This study addresses the dual challenges of privacy preservation and real-time anomaly detection in electrocardiogram (ECG) data across hospitals with non-independent and identically distributed (non-IID) distributions. To this end, the authors propose an end-to-end federated unsupervised detection system that uniquely integrates federated learning (FedAvg), (ε,δ)-differential privacy via DP-SGD under Rényi differential privacy, convolutional autoencoders (e.g., ConvAE), and INT8 post-training quantization for edge deployment on a Raspberry Pi 4. Experimental results demonstrate that the federated ConvAE achieves an AUROC of 0.782 under a clinically recommended privacy budget (ε=4). Furthermore, INT8 quantization halves model size and reduces inference latency by 44%, with negligible AUROC degradation (<0.12%). The work also establishes the statistical independence between differential privacy mechanisms and quantization-induced compression, offering theoretical grounding for joint optimization of privacy and efficiency.

Abstract

Continuous electrocardiography (ECG) monitoring could surface rhythm abnormalities before they escalate into cardiovascular events. However, a deployable system must satisfy three requirements simultaneously: legal-grade privacy (GDPR, HIPAA), real-time inference on constrained edge hardware, and detection quality under non-IID cross-hospital data. We design and evaluate an end-to-end federated system addressing all three for unsupervised 12-lead ECG anomaly detection on PTB-XL dataset, combining three autoencoder families (VanillaAE, ConvAE, VAE), Flower-based federated averaging (FedAvg) across ten simulated hospitals, client-side differentially private SGD (DP-SGD) with a Rényi-DP accountant, and 8-bit integer (INT8) post-training quantization with Raspberry Pi 4 benchmarking. Our main contributions are: an empirical characterization of how these mechanisms compose, practical DP-specific recommendations, and technical and security insights for a clinically sensitive setting. Federated learning matches or exceeds the centralized baseline across all architectures (ConvAE federated area under the ROC curve, AUROC, $0.782$), and an $\varepsilon$ sweep identifies $\varepsilon=4$ as the recommended clinical operating point. INT8 quantization roughly halves model size and cuts Pi 4 latency by up to $44%$ with $<0.12%$ AUROC loss. Crucially, DP and quantization penalties are empirically independent, so practitioners need not trade a strong privacy guarantee for a compact edge footprint. To our knowledge, this is the first system combining federated learning, formal $(\varepsilon,δ)$-DP, unsupervised reconstruction-based detection, and quantized AArch64 deployment.

Problem

• privacy-preserving
• federated learning
• ECG anomaly detection
• edge devices
• non-IID data

Innovation

• Federated Learning
• Differential Privacy
• ECG Anomaly Detection
• Model Quantization
• Edge Deployment