April 5, 2021

Renos Zabounidis: UMass Amherst Rising Researcher

Student combines three majors to improve artificial intelligence
Renos Zabounidis ’22, computer science, mathematics, BDIC, Commonwealth Honors College

“It’s one thing to solve a problem and another thing to help someone else figure out how to solve a problem. In math and in research in general, it’s not just about the cold hard numbers, it’s about intuition.”

Having taken advanced classes at UMass Amherst, worked in campus labs, and benefited from summer internships, Renos Zabounidis ’22 has settled on a bold research quest: to fuse principles from cognitive psychology and statistical machine learning to better understand the nature of intelligence.

Renos plans to continue his research into the human side of artificial intelligence as a PhD student and eventually as a university professor. Now in his junior year, he is excelling in three majors: computer science, mathematics, and a Bachelor’s Degree with Individual Concentration (BDIC) in computational and cognitive science.

Renos dug into undergraduate research early, working in the Advanced Healthcare and Human Analytics Lab (AHHA) during his first semester at UMass. There, he and another undergraduate created a mobile application for the monitoring and management of movement disorders such as Parkinson’s Disease.

Through that lab work, Renos discovered he was more interested in theoretical than applied research and, after thriving in a master’s level machine learning class, moved on to work in the UMass Information Fusion Lab with Madalina Fiterau, assistant professor of computer science. Renos’s work there was unparalleled and speedy. “He is exceptionally gifted, demonstrating enthusiasm about doing research and diligence in completing projects,” Fiterau says.

Last summer, Renos participated in a robotics institute at Carnegie Mellon University. There, drawing on his two years of UMass research and advanced courses, he created an algorithm, called “Introspection,” that applies an interdisciplinary approach to imbuing artificial intelligence with more human qualities.

“Humans can’t fully understand what’s in another person’s brain and, furthermore, people change their beliefs over time,” he explains. “In order for AI to be helpful, it has to understand what a person is thinking as well as what they should be thinking and are not. The intent of the algorithm is to make inferences about what humans feel, their confidence level, and what they might do at any given moment.” Renos will continue work on the algorithm this summer.

“This field of research is what I’m passionate about,” he says. “There’s always more to be found.”