Searching Neural Architectures for Sensor Nodes on IoT Gateways

2025-05-29

AI Summary

Deploying privacy-sensitive, resource-constrained neural networks on edge devices in healthcare and industrial IoT (HIoT/IIoT) remains challenging due to stringent power, memory, and data-privacy constraints. Method: This paper introduces the first end-side neural architecture search (NAS) framework tailored for ultra-low-power gateways (e.g., Raspberry Pi Zero 2), operating entirely locally—without uploading raw sensor data—to preserve data sovereignty and privacy. It integrates an edge-aware search space, a lightweight NAS algorithm, quantization-aware training, and edge-specific deployment optimization to yield hardware-efficient, task-specialized models. Contribution/Results: Evaluated on the Visual Wake Words dataset, our framework achieves state-of-the-art accuracy while completing NAS in under 10 hours. It demonstrates, for the first time, the feasibility of efficiently constructing privacy-preserving AI models directly on milliwatt-class edge devices.

Technology Category

• Machine Learning: Learning on the Edge & Model Compression
• Search and Optimization: Learning to Search
• Philosophy and Ethics of AI: Privacy & Security

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Security and Privacy: Security and privacy of machine learning and AI applications
• Search and Retrieval-Augmented AI: Personalized, context-aware and across-device search

Abstract

This paper presents an automatic method for the design of Neural Networks (NNs) at the edge, enabling Machine Learning (ML) access even in privacy-sensitive Internet of Things (IoT) applications. The proposed method runs on IoT gateways and designs NNs for connected sensor nodes without sharing the collected data outside the local network, keeping the data in the site of collection. This approach has the potential to enable ML for Healthcare Internet of Things (HIoT) and Industrial Internet of Things (IIoT), designing hardware-friendly and custom NNs at the edge for personalized healthcare and advanced industrial services such as quality control, predictive maintenance, or fault diagnosis. By preventing data from being disclosed to cloud services, this method safeguards sensitive information, including industrial secrets and personal data. The outcomes of a thorough experimental session confirm that -- on the Visual Wake Words dataset -- the proposed approach can achieve state-of-the-art results by exploiting a search procedure that runs in less than 10 hours on the Raspberry Pi Zero 2.

Problem

• Automates neural network design for edge IoT gateways
• Enables local ML without sharing sensitive IoT data
• Optimizes hardware-friendly NNs for healthcare and industry

Innovation

• Automated edge-based Neural Network design
• Local data processing for privacy preservation
• Hardware-friendly custom NNs for IoT applications