This work describes and analyzes a set of state-of-the-art artificial intelligence (AI) hardware kits created for education and research that can be used in undergraduate AI labs. AI cloud-based computing devices and solutions like the Arduino-based Tiny Machine Learning kits or the mobile app by Edge Impulse, Raspberry Pi-based AIY Voice kits by Google, Quad-core Arm Cortex-A53 and Cortex-M4F-based Google Coral Dev Boards, as well as the more powerful Jetson AGX Xavier (512-core NVIDIA Ampere architecture GPU), and Jetson AGX Orin (2048-core NVIDIA Ampere architecture GPU) Developer kits, are compared using published characteristics and direct experiments. The comparison criteria used are (1) ease of setup and first use, (2) learning curve and required prior knowledge, (3) learning community support availability, (4) suitability for undergraduate learning, (5) computational speed, and (6) cost including both the hardware cost and the subscription services cost. Based on the results of the analysis the tested AI computing devices are ranked for use in various levels of undergraduate curricula. The goal is to provide the faculty interested in developing their own AI labs with some guidance in choosing appropriate AI hardware from an experimental perspective.
Are you a researcher? Would you like to cite this paper? Visit the ASEE document repository at peer.asee.org for more tools and easy citations.