UNT students across disciplines are experimenting with AI in their research.

Cover illustration generated using Adobe Firefly

BY AMANDA LYONS

Ernest ‘Will’ Cubit (’22) is a student in UNT’s Master of Science in Artificial Intelligence program, which debuted as the first of its kind in Texas when it launched in Fall 2020. He’s combining his knowledge of AI and the materials science he learned as an undergraduate into one project developing an AI-based system that could discover new alloys for cars and airplanes. He uses three different AI programs he’s built to ingest past studies, find new relationships between elements and predict if the combination would work in the real world. “Physical products can’t advance until materials science does. Many of the products we use have been created or improved by a scientist, from the glass protecting our smart devices to the lithium ion batteries powering them,” Cubit says.

Medisetti Gowtham, a computer science master’s student, is helping the biological sciences department with an AI system that can create reliable food webs, a graphical representation of what-eats-what in an ecological community. Gowtham feeds past reports into an AI program, a second cleans up the data and a third program creates the webs. He expects the AIs to go through thousands of scientific papers by the end of the project. “The more data you give to a large language model, the more accurate it becomes,” Medisetti says.

Other students like Xiaoying Song are finding ways to make AI work for them in a supporting role. “For my research into online misbehavior, I usually need to write some code to process data,” says Song, an information science doctoral student. “Copilot is a program that will try to add in the rest of the code for me. It is very convenient for someone like me who is not good at coding.”

AI doesn’t have to be limited to data. Joshua ‘JD’ Fuller, a music sophomore, is working on a project allowing AI and humans to play music together. It’s an interactive installation where AI analyzes what a visitor plays on a drum set by listening through microphones and then plays its own sounds along with the person for a mindful human-AI jam session. “I use AI now in several other contexts, too, like audio mixing, sound analysis and generative coded music,” Fuller says. “I think AI is going to become more prevalent in computer music as more AI-based tools with creative uses appear.”