IE 11 Not Supported

For optimal browsing, we recommend Chrome, Firefox or Safari browsers.

Case Study Finds UF, NVIDIA a Model Partnership for AI

The nonprofit Center for Data Innovation praises a partnership between the University of Florida and tech company NVIDIA as a model for other institutions to develop AI research and education.

University of Florida, sign, Gainesville
Combining hardware, software and funding from the technology company NVIDIA with supercomputer technology donated by the company’s co-founder Chris Malachowsky, the University of Florida announced plans in 2020 to become a first-of-its-kind “AI University,” with the goal of establishing an AI data center for AI-centric academic programming. According to a case study last month from the Center for Data Innovation, the initiative could serve as a model for how institutions can successfully create and sustain interdisciplinary research and education efforts that make use of the emerging technology.

In an email to Government Technology, UF Provost Joe Glover said the university has integrated AI across departments and course subjects for research and use cases, as well as to familiarize students with machine learning applications. He added that the university adopted a philosophy, “AI across the curriculum,” to equip students with related skills for future workforce demands.

“All colleges and departments have enthusiastically embraced this [initiative], and most are integrating AI into the curriculum. Not only does this give our students a leg up in the job market, but we believe it is a scalable model to answer the nation’s pressing need to develop a 21st-century AI-savvy workforce at scale. In short order, we will be graduating thousands of students into the workforce, and through partnerships with other universities in the state of Florida and the (Southeastern Conference), we are encouraging other universities to adopt a similar model to intensify the impact,” Glover wrote. “Over the past two years, we have hired an additional 100 faculty focused on AI and applications. We believe we are building the nation’s first AI university, and we could not do this without NVIDIA’s help and partnership.”

Hodan Omaar, a policy analyst with the nonprofit Center for Data Innovation, said the university serves as an example for how to incorporate AI across curricula, as well as how universities can pull together the funding and resources necessary for advanced technologies as they get more involved in machine-learning research, education and training efforts. In a case study describing lessons to take from UF’s partnership with industry to foster AI, Omaar singled out five:

  • focus on increasing AI computing capacity at universities
  • have clear and realistic goals based on institutional capacity
  • spread benefits of the partnership across disciplines to increase institutional buy-in at the university
  • encourage public agencies to stimulate demand for AI research and skills
  • create an assessment program with metrics for success

Omaar noted that the university leveraged industry partnerships, support across UF departments and funding from the state to get the initiative off the ground and regularly assess academic programming in AI.

According to the case study, the initiative began in October 2019 with UF alumnus Malachowsky offering the university $25 million toward building a powerful AI supercomputer. A university steering committee then approached his company, NVIDIA, to match the gift with $25 million worth of additional tech needs. From there, the university gathered $20 million in funding from across its 16 colleges, plus $110 million from the state of Florida to build a new data center and $15 million a year to hire and retain new faculty.

“The desire and the need for AI researchers, especially those in academia, wanting access to [supercomputer technology] is significantly higher than the supply. There’s been a lot of thinking, especially in the last year, of how to solve this,” Omaar said, adding that public universities like UF tend to have fewer resources at their disposal than private institutions involved in AI research like the Massachusetts Institute of Technology (MIT), Stanford University and Carnegie Mellon University.

“What this model between UF and NVIDIA is really solving is, ‘How do we close that gap? How do we ensure that some of those researchers [gain access]?’” she continued. “I think those are the goals that this UF and NVIDIA partnership really solve.”

Omaar said the university’s ability to clearly articulate its AI research and education goals to stakeholders was crucial to gaining the resources needed for the initiative. Her case study said the university’s main goals in its partnership with NVIDIA are to advance AI research at scale, support equal access to AI technology for underserved students and to improve AI workforce training and education to meet a growing demand among employers for applicants with AI and IT-related skill sets.

In the short time since HiPerGator AI became operational in 2021, the case study noted, UF has made several strides in research and AI workforce training efforts. In March, the university announced plans to create an Artificial Intelligence Academic Initiative Center to serve as its focal point for AI and data science education and research. The tech has also been used to train GatorTron, a natural language processing model developed by UF researchers that can analyze volumes of clinical data at the university’s academic health center.

Omaar said the university’s ability to garner support across departments and integrate AI skills across subjects are part of what makes its AI initiative unique compared to other machine learning research and education efforts.

“Partnerships around AI and universities tend to stick within computer science … the STEM subjects are where the funding is coming from and going to, but this one is really spread out across the entire university. The Business School is involved, the School of Medicine’s involved. Every single student is not only included, but also encouraged to use AI and think about AI,” she said. “What I’ve seen from that is that the research at the university has really broadened.”

Entering into the partnership’s second year, Omaar called its impact “quite impressive” so far.

“I think what makes this stand out is the size of the supercomputer itself and the scope of the initiative, and the fact that the university has done so much to leverage the system to take it outside of just the computer sciences,” she said.

Omaar noted that the university has also partnered with several HBCUs to help students traditionally underrepresented in science, technology, engineering and math (STEM)-related fields to gain access to supercomputer technology, and is working with the nonprofit Inclusive Engineering Consortium to increase the number of Black, Hispanic and Indigenous graduates with AI skills.

“It really speaks to one of the central goals of increasing access to these systems, not only across the country, but also to individuals that have historically been underrepresented in science and engineering,” she said.

Editor's note: An earlier version of this story interpreted Joe Glover's reference to the SEC as the Securities and Exchange Commission. It has been updated to reflect that he was referring to the Southeastern Conference, a regional athletic organization.
Brandon Paykamian is a staff writer for Government Technology. He has a bachelor's degree in journalism from East Tennessee State University and years of experience as a multimedia reporter, mainly focusing on public education and higher ed.