TinyML4D: Scaling Embedded Machine Learning Education in the Developing World

Abstract

Embedded machine learning (ML) on low-power devices,also known as “TinyML,” enables intelligent applications onaccessible hardware and fosters collaboration across disci-plines to solve real-world problems. Its interdisciplinary andpractical nature makes embedded ML education appealing,but barriers remain that limit its accessibility, especially in de-veloping countries. Challenges include limited open-sourcesoftware, courseware, models, and datasets that can be usedwith globally accessible heterogeneous hardware. Our visionis that with concerted effort and partnerships between indus-try and academia, we can overcome such challenges and en-able embedded ML education to empower developers andresearchers worldwide to build locally relevant AI solutionson low-cost hardware, increasing diversity and sustainabil-ity in the field. Towards this aim, we document efforts madeby the TinyML4D community to scale embedded ML educa-tion globally through open-source curricula and introductoryworkshops co-created by international educators. We con-clude with calls to action to further develop modular and in-clusive resources and transform embedded ML into a trulyglobal gateway to embedded AI skills development.

—, S. Abdeljabar, —. (2024). “TinyML4D: Scaling Embedded Machine Learning Education in the Developing World.” Proceedings of the AAAI Symposium Series.
[Link] | [PDF]