Towards Hardware Supported Domain Generalization in DNN-Based Edge Computing Devices for Health Monitoring

2024-06-24

IEEE Transactions on Biomedical Circuits and Systems

AI Summary

Domain generalization (DG) models for wearable ECG edge devices suffer from poor robustness against cross-sensor and cross-patient distribution shifts, while also failing to meet stringent resource constraints of embedded hardware. Method: We propose a lightweight DG framework featuring a learnable correction layer—enabling unknown-domain adaptation via fine-tuning only a single layer’s parameters, thereby drastically reducing computational and memory overhead. Crucially, this is the first work to co-optimize a DG algorithm with a dedicated low-power ECG accelerator. Results: Our approach reduces DG inference complexity by over 2.5×, improves average F1-score on unseen target domains by more than 20%, and enables real-time, robust ECG classification at the edge—achieving both high accuracy and high energy efficiency under strict hardware constraints.

Technology Category

• Machine Learning: Hardware-aware ML
• Application Domains: Internet of Things, Sensor Networks & Smart Cities
• Computer Vision: Adversarial Attacks & Robustness

Application Category

• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Systems and Infrastructure for Web, Mobile and WoT: Energy management for devices in mobile Web and WoT environments
• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation

Abstract

Deep neural network (DNN) models have shown remarkable success in many real-world scenarios, such as object detection and classification. Unfortunately, these models are not yet widely adopted in health monitoring due to exceptionally high requirements for model robustness and deployment in highly resource-constrained devices. In particular, the acquisition of biosignals, such as electrocardiogram (ECG), is subject to large variations between training and deployment, necessitating domain generalization (DG) for robust classification quality across sensors and patients. The continuous monitoring of ECG also requires the execution of DNN models in convenient wearable devices, which is achieved by specialized ECG accelerators with small form factor and ultra-low power consumption. However, combining DG capabilities with ECG accelerators remains a challenge. This article provides a comprehensive overview of ECG accelerators and DG methods and discusses the implication of the combination of both domains, such that multi-domain ECG monitoring is enabled with emerging algorithm-hardware co-optimized systems. Within this context, an approach based on correction layers is proposed to deploy DG capabilities on the edge. Here, the DNN fine-tuning for unknown domains is limited to a single layer, while the remaining DNN model remains unmodified. Thus, computational complexity (CC) for DG is reduced with minimal memory overhead compared to conventional fine-tuning of the whole DNN model. The DNN model-dependent CC is reduced by more than 2.5 $ imes$ compared to DNN fine-tuning at an average increase of F1 score by more than 20 % on the generalized target domain. In summary, this article provides a novel perspective on robust DNN classification on the edge for health monitoring applications.

Problem

• Addressing domain generalization in DNNs for health monitoring.
• Enabling robust ECG classification across sensors and patients.
• Reducing computational complexity in edge-based ECG accelerators.

Innovation

• Hardware-software co-optimized ECG accelerators
• Domain generalization via correction layers
• Reduced computational complexity with minimal memory overhead