N.C. A&T Joins NCShare Partnership, Enhances Computing and Data Modeling Capacity

EAST GREENSBORO, N.C. (April 23, 2025) – North Carolina Agricultural and Technical State University has joined the North Carolina Shared Research and Education (NCShare) partnership, led by Duke University, which will enhance N.C. A&T researchers’ capacity to process large amounts of data and train artificial intelligence models faster.

“You need a lot of resources, especially when the researchers are working on millions of images and large video data files,” said Kaushik Roy, Ph.D., chair of A&T’s Department of Computer Science. “So, we need a lot of computing power, which is not possible using the regular desktop, CPU-based machine. A solution is a graphics processing unit, or GPU-based AI machines, to scale up our capability and improve our computing power and processing very quickly.”

Studying “deepfakes,” or AI-manipulated images, video or audio that purposely distort or misrepresent real people, are practical examples that could benefit from access to GPUs, since the datasets that create them are so large. The cost to any single institution of acquiring expensive deep-capacity GPUs is prohibitive, which increases the partnership’s value to all its members.

“With support from the College of Engineering, we have acquired state-of-the-art GPU computing systems for my cyber defense and AI lab,” said Roy. “While these resources have significantly enhanced our research capabilities, we have identified a growing need for additional high-performance computer infrastructure.”

For some of Roy’s cybersecurity-related research, team members have to create more realistic “fake,” or synthetic, data because of privacy reasons, which means even more data is generated leading to large file sizes.

Researchers apply different AI models to large datasets to get better, more accurate decisions. AI models need to be trained using large, diverse amounts of data.

“To address this, NCShare is creating a very powerful cluster of interconnected GPU machines that is scalable, and joining this partnership will help us dynamically allocate resources for other research groups across our campus,” said Roy. “It’s a great opportunity for our faculty and students so they can connect with it remotely and gain access that way.”

A&T researchers interested in using this for their projects will gain access to the cluster through Roy and Edmunson Effort, the department’s technology support analyst, and be allotted a specific timeframe to use the platform.

Roy is a co-principal investigator on the National Science Foundation grant, “NCShare Accelerating Impact - GPU-as-a-Service (AI-GaaS),” led by principal investigator Tracy Futhey (Duke University) and co-principals Charley Kneifel (Duke), Michael Baker (University of North Carolina at Chapel Hill) and Tracy Doaks (MCNC).

According to NCShare, “The platform will employ a novel architecture and shared-services model, adapting and extending mature techniques used at R1s (Research 1 institutions) that enable virtualization of powerful GPUs, and applying those same virtualization techniques within a shared, multi-institutional environment. This will create an affordable and scalable platform for AI and ML research and for advancing Foundational AI Research and Applied Research in AI.”

A&T recently announced a bachelor’s degree in AI, which will further enhance the university’s established leadership in AI research and education.

Additional partners involved in other NCShare-related initiatives include North Carolina Central University and Davidson College.