Elements of TinyML on Constrained Resource Hardware

Abstract

The next phase of intelligent computing could be entirely reliant on the Internet of Things (IoT). The IoT is critical in changing industries into smarter entities capable of providing high-quality services and products. The widespread adoption of IoT devices raises numerous issues concerning the privacy and security of data gathered and retained by these services. This concern increases exponentially when such data is generated by healthcare applications. To develop genuinely intelligent devices, data must be transferred to the cloud for processing due to the computationally costly nature of current Neural Network implementations. Tiny Machine Learning (TinyML) is a new technology that has been presented by the scientific community as a means of developing autonomous and secure devices that can gather, process, and provide output without transferring data to remote third party organizations. This work presents three distinct TinyML applications to cope with the aforementioned issues and open the road for intelligent machines that provide tailored results to their users. © 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.