Researchers at UNT Engineering were recently awarded a three-year grant by the National Science Foundation through its Research Experiences for Undergraduates program to develop and operate 10-week intensive programs for students to learn about artificial intelligence. Over the next three summers, Hui Zhao and Mark Albert, assistant professors of computer science and engineering, will lead the program that will provide students opportunities for interdisciplinary experiences in advanced deep learning research.
“Deep learning is a big thing in AI, particularly in the last decade because it’s state-of-the-art in so many interesting machine learning problems,” Albert said.
Deep learning is involved in everything from computer vision to speech recognition in tools that are common to people adept in using their smartphones and smarthome devices. Many advances once thought uniquely human, such as evaluating the sentiment of spoken statements or even writing short stories are being automated with deep learning models. The REU will focus on accelerated deep learning through a hardware-software collaborative approach. Zhao does hardware-related research, and Albert’s research is in software. Together, they are collaborating to solve problems that can’t be solved independently.
“We’re starting to run into limitations where you wish you had more computing power,” Albert said. “What we’re doing is not just standard use of generic GPUs, but actually squeezing every last drop of computing power on the software side while also optimizing the hardware.”
In addition to providing undergraduates real-world research experience, one of the goals of the REU is to encourage students to participate in further scientific research and to familiarize them with and explore their interests in attending graduate school.
“We are trying to encourage more undergraduates to continue their academic careers,” Zhao said. “We hope they will later apply for UNT graduate programs. We want to help them develop research, communication and collaboration skills.”
Albert and Zhao are excited about preparing a workforce of young scientists. The program, which aims to provide a supportive and engaging learning environment, will recruit undergraduates from a diverse background who might not otherwise have opportunities for this type of research. Recruitment will target students typically underrepresented in STEM fields, including female students. They also will provide them support and guidance in applying for UNT’s graduate programs, such as the Master of Science in Artificial Intelligence, one of the few such programs nationwide.
Participants will learn fundamentals of machine learning and efficient deep neural network algorithm development, which plays an important role in security, surveillance, gaming, health care, defense and manufacturing.
“What’s extra special about these students is they’re not just getting deep learning experience, which in and of itself is really great, but they’re going to be getting the hardware, software and code design to get greater efficiency from more limited resources,” Albert said. “A traditional deep learning experience wouldn’t get you that.”
Albert, along with Ting Xiao, research assistant professor in computer science and engineering, also will be hosting a UNT AI summer research program for UNT students, which will run concurrently with the REU. The groups will collaborate. The programs this summer likely will be conducted remotely with future programs conducted in UNT labs.