This post was originally published on this site.
LAWRENCE — Public high school students in Kansas and two other states will receive training in the cutting-edge field of artificial intelligence, learning to create both code that underpins AI and the microelectronics to run it — as part of the United States’ push to keep the lead in microchip manufacturing and AI software development.
Researchers at the University of Kansas, along with the University of Florida and the University of North Texas, will partner with regional high schools to engage about 500 students and 25 teachers in real-world projects to build interest in the technology as a career path. The work is enabled by a $1.4 million grant from the National Science Foundation. Of that, about $350,000 will come to KU.
The research at KU is headed by Tamzidul Hoque, assistant professor of electrical engineering & computer science. His team in Lawrence will partner with Shawnee Mission West High School in Overland Park, where computer science teacher Mark Lange will implement the curriculum.
A vital part of the training will allow students to run their code on Tiny Machine Learning (TinyML) devices — basic low-power machines that enable AI processing directly on hardware.
“This will be a small device performing AI tasks at the user end without connecting to the cloud,” Hoque said. “TinyML is one application that allows a large AI model to be converted into a smaller one that can run on a small device.”
These so-called “edge devices” process data with their own microelectronics rather than relying on a centralized cloud or data center.Â
“We want to demonstrate to students the wide range of edge AI applications available,” Hoque said. “By working with edge AI, they’ll not only learn about AI but also gain knowledge of microelectronics because it involves low-level hardware. Our curriculum addresses both of these important areas — microelectronics and AI.”
Hoque’s team at KU is developing the edge devices to be used by students in classes nationwide, work informed by his earlier NSF-funded research into training students in computing-hardware fundamentals though gamified learning.
The design of the edge devices will consider strapped budgets faced by many high schools, particularly in low-income communities, according to Hoque.
“We’re developing a hardware platform that includes microprocessors, various sensors and communication components,” he said. “We’ll collaborate with the University of Florida to develop the platform, with a key challenge being cost-effectiveness. While many existing platforms can be used for programming AI, they are not affordable. Our goal is to create a device costing less than $45, equipped with at least 10 different sensors, making it accessible even for high schools with limited resources.”
Part of the project involves measuring and honing effectiveness of the instruction. Hoque and his colleagues will focus the training on altruistic, community-centered projects so students understand how engineering helps people.
“When we try to motivate students about engineering, we often highlight high-paying salaries or the lucrative aspects of the jobs — but engineering is not only about those things, and many students may not feel motivated solely by them,” the KU researcher said. “Integrating the concept of altruism — how engineering can help their community — can be a stronger motivator. For example, developing an AI application for fire detection or supporting farmers through novel technologies gives students a sense of altruism and community support, inspiring them to pursue careers in those directions.”
Nonetheless, according to Hoque, the curriculum should provide access to high-paying jobs in AI and microelectronics for individual students. By developing this workforce, Kansas and other states in the project could succeed in drawing more high-tech companies as students qualify to specialize in the sector. To ensure this, the researchers have teamed with AI-industry partners to match workforce needs of those employers with the training.
“Our goal is to ensure the curriculum we develop is well aligned with the industry,” Hoque said. “We have an advisory board made up of industry members who provide feedback on whether the topics we have chosen are suitable for the field and whether learning these technical skills will help students secure jobs in the long run.”
Along these lines, the researchers will hold conferences where high school teachers in the project and industry partners will trade ideas on curriculum and teaching methods to ensure the training is industry focused.
The work at KU is enabled by the CHIPS and Science Act, passed by Congress in 2022, a law designed to support domestic production of semiconductors and strengthen national security.
“After COVID, we realized how dependent we are on external supply chains, prompting the government to provide significant incentives for developing domestic manufacturing facilities,” Hoque said. “This issue impacts not only consumers but also national security, as microelectronics used in mission-critical systems must be developed in secure facilities with no possibility of malicious alterations or security threats. For national security reasons, it’s essential to have domestic capabilities to design and fabricate our own microchips. But it’s not enough to develop these facilities — we also need people to work in them. Programs like this will motivate students to explore hardware and pursue careers in microelectronics.”