Introduction to the Special Issue on tinyML
Theocharis Theocharides,Charlotte Frenkel,Lukas Cavigelli
DOI: https://doi.org/10.1145/3658375
2024-05-13
ACM Transactions on Embedded Computing Systems
Abstract:tinyML (Tiny Machine Learning) encapsulates and nurtures the fast-growing branch of ultra-low-power machine learning (ML) technologies and approaches dealing with machine intelligence at the very edge of the cloud. The fundamental goal of tinyML is to utilize cross-layer design approaches and deploy ML inference to ultra-low-power devices such as microcontrollers or custom-designed circuits that consume under or about a milliwatt of power and can last for months, if not years, with a single recharge of the battery. These integrated "tiny" ML applications require "full-stack" (hardware, system, software, and application) solutions including ML architectures, techniques, tools, benchmarks, and approaches capable of performing on-device analytics. A variety of sensing modalities (vision, audio, motion, environmental, human health monitoring, etc.) are used with extreme energy efficiency, typically in the single milliwatt (and below) power range, to enable machine intelligence right at the boundary of the physical and digital worlds. We see a new world with trillions of distributed intelligent devices enabled by energy-efficient ML technologies that sense, analyze, and autonomously act together to create a healthier and more sustainable environment for all! The tinyML ecosystem is fueled by (i) emerging commercial applications and new systems concepts on the horizon; (ii) significant progress on algorithms, networks, and models down to 100 kB and below; and (iii) current low-power applications in vision and audio that are already becoming mainstream and commercially available.
computer science, software engineering, hardware & architecture