Paper Contents
Abstract
Tiny Machine Learning (TinyML) is an emerging discipline that enables theexecution of machine learning inference on ultra-low-power,memory-constrained microcontrollers and edge devices. By shiftingintelligence to the device, TinyML reduces latency, improves privacy, andlowers bandwidth and energy costs associated with cloud-centredarchitectures. This paper reviews the technical architecture and toolchain ofTinyML, surveys prominent applications across domains, discusses modeloptimization and deployment strategies, and analyses challenges includingresource constraints, energy management, security vulnerabilities, andupdate mechanisms. We also highlight promising research directions such ason-device learning, federated approaches, neural architecture search tailoredfor microcontrollers, and hardwaresoftware co-design.
Copyright
Copyright © 2025 Patel Uday Bhadreshkumar. This is an open access article distributed under the Creative Commons Attribution License.
