In the span of a couple of years, AI has gone from science fiction to an indispensable technology, with nearly every industry scrambling to develop AI models customized to their unique environments and needs.
While such models are impossibly complex by nature, the tools for building and working with them are becoming easier to use. They are also enabling a host of potential research questions and experiments that were previously out of reach.
The John and Marcia Price College of Engineering is investing in those tools, but also the faculty skills necessary to make the most of them.
As such, it has now launched U-AI, an up-skilling program designed to quickly train faculty members from around the University who are interested in integrating AI into their research practices.
The program is part of Price Engineering’s larger investment into AI and machine learning, exemplified by the $194 million John and Marcia Price Computing and Engineering Building, slated to be completed in 2027.
The pilot program, developed and coordinated by Sneha Kasera, Associate Dean for Academic Affairs, and Vivek Srikumar, associate professor in the Kahlert School of Computing (KSOC), has just selected its first class of 31 faculty members. Their course of study will go far beyond how to engineer prompts for generative AI systems, such as OpenAI’s ChatGPT or Google’s Gemini.
Such systems are known as Large Language Models, but they are only one type of the many models that artificial intelligence researchers build. The thread connecting these diverse types is a computational technique known as “machine learning” — essentially, a way of training computers to find subtle patterns in enormous datasets.
Generative AI systems like ChatGPT are trained on text; they are designed to find patterns within natural language and recreate them in new ways. An AI system trained on the structure of a particular class of molecules, by contrast, might find patterns that represent a new material for solar panels, or a candidate for an anti-cancer drug.
“Many faculty want to start using these sorts of tools, but they don’t yet know what’s technically possible, or how to ask AI systems these questions without introducing bias or errors,” says Srikumar. “We’re aiming to give those faculty everything they need to get started in a matter of weeks, even if they have no prior experience at all.”
Beginning in September, participants will first take a four-week course on the basics of AI and machine learning taught by Srikumar, along with KSOC’s Alan Kuntz, Ana Marasović, and Daniel Brown. The cohort and instructors will then split into three concurrent specialization tracks, which will be determined by the research interests of their members.
Finally, participants will be partnered with faculty consultants who have common research interests; by the end of the calendar year, they will produce a grant application to fund a joint project.
The inaugural class was selected from nearly a hundred applicants.
“We want to build solid partnerships between the faculty who participate in this program, both as experts and as students,” says Kasera. “The size of the inaugural cohort will facilitate a deeper engagement between the AI experts in the Price College and the program participants.”