Jia Di, department head and professor of computer science and computer engineering.

NSF awards $800,000 to support Edge AI app development

Spread the love

University Relations

Jia Di, head of department and professor of computer science and computer engineering.

Jia Di, department head and professor of computer science and computer engineering at the University of Arkansas, received an $800,000 award to support Edge AI application development. The award is part of a larger $6 million award from the National Science Foundation to six collaborating universities and several private sector partners in Alabama, Arkansas and North Dakota.

In their proposal to the NSF, the collaborating researchers noted that artificial intelligence, or AI, while highly effective in many real-world applications, requires access to the Internet as well as large, complex, and remote computers to make decisions and predictions. This can lead to long delays and increase privacy and security issues.

Edge AI is an emerging area of ​​AI development that avoids these issues by locating and analyzing data locally, whether through a camera, smartphone, or wearable device. The proposed work is designed to push the boundaries of new Edge AI technology.

The more limited goal of the project is to build a smart, wearable device for diabetics to monitor their blood sugar. The first virtue of this approach would be the elimination of blood tests. Instead, the wearable device will monitor the wearer’s breathing, collecting up-to-date data points on blood sugar levels. While the sensor won’t initially be as accurate as a blood draw, the AI ​​algorithm used by the device will still be able to make timely recommendations, such as “see a doctor.”

Di and his team are tasked with developing a new application-specific integrated circuit, or ASIC, that will be the main computing unit of the proposed wearable device. The ASIC will be able to implement different AI algorithms in a short period of time to reduce time to market, while maintaining long battery life thanks to innovations in asynchronous circuit design and system integration.

“I’m excited to be part of this multi-university team to leverage our expertise in asynchronous ASIC design to help with the development of this device,” Di said. “Given the diabetic population in these states [Alabama, Arkansas and North Dakota] and in the country, the research results of this project will have the potential to have substantial impacts on our society.

Beyond the goal of creating a wearable device, the research is expected to accelerate the development of Edge AI and increase US AI competitiveness. The award will also provide research training opportunities for advanced university students, in addition to training high school teachers to educate their own students in Edge AI principles to ground them in essential concepts as early as possible.

The National Science Foundation award falls under the EPSCoR Research Infrastructure Enhancement Program: Track-2 Focused EPSCoR Collaborations, which, according to their website, “supports cross-jurisdictional teams of EPSCoR researchers to conduct research in industries emerging countries, with the aim of promoting economic growth in their jurisdictions”.

Professor Di is the Rodger S. Kline Professor.

About the University of Arkansas: As Arkansas’ flagship institution, the U of A offers an internationally competitive education in more than 200 academic programs. Founded in 1871, the U of A contributes more than $2.2 billion to the Arkansas economy through the teaching of new knowledge and skills, entrepreneurship and employment development, discovery through research and creative activity while providing training for professional disciplines. The Carnegie Foundation ranks the U of A among the few American colleges and universities with the highest level of research activity. US News and World Report ranks the U of A among the top public universities in the nation. Learn how the U of A is working to build a better world at Arkansas Research News.

Leave a Comment

Your email address will not be published.