Jovana Kondic seeks to model the irrational human brain
At a relatively young age, Jovana Kondic’s research interests have already brought her around the world.
A native of Novi Sad, Serbia, Kondic became a finalist in a Google Science Fair in her first year of high school, went on to make presentations at innovation fairs and conferences in Moscow and Yancheng, China, and attended a summer program at Yale University — all by the time she was a high school senior.
“It was there that I became enchanted with the vibrant campus life and community of the American universities and decided to give it a shot and apply,” says Kondic, who is now a senior at Princeton University.
With her sights firmly set on graduate school, Kondic recently took part in the Stanford Undergraduate Research Fellowship program — SURF — for a summer of intense study, working with Dorsa Sadigh, a professor of electrical engineering and of computer science, and doctoral candidate Andy Shih. The placement was perfect for Kondic — she has long been fascinated with the hypothesis that human decision-making is not entirely rational, and has an interest in developing computer models of often-irrational human behaviors within the context of artificial intelligence.
As a SURF student, Kondic worked with Sadigh and Shih to study how to improve the collaboration between multiple interacting intelligent agents.
“I’ve learned this has to do with our brain’s internal reward system that can favor cooperation over personal gain,” Kondic says. “Accurately modeling the seemingly irrational human behavior requires a deeper reflection on how the mind really works.”
Her SURF stint was Kondic’s second experience at Stanford. Last year, she spent two months working in the lab of Stanford bioengineering professor David Camarillo, participating in research on wearable sensors for quantifying traumatic brain injury, and getting listed as a co-author on a subsequent scientific paper. What impressed her most with that experience “was the inclusion, trust, and accountability I received from the group in spite of having been the youngest member bringing unarguably the least amount of skills to the table.”
Now, with a stint at SURF also under her belt, Kondic is focusing her curiosity on robotics and other systems that combine AI with computational models of cognition, with the idea of someday building socially adept robots and AI-enabled tiny machines (micro-electrical mechanical systems) for use in medical applications — minuscule smart machines that operate inside the human body.
The fact that SURF 2020 coincided with the travel restrictions imposed by COVID-19 affected Kondic in both obvious and subtle ways. Unlike her 2019 Camarillo Lab experience, her mentorship during SURF 2020 had to be virtual.
“Being an international student, unable to see my family while at the same time being suddenly separated from my [Princeton] campus community, I spent quite some time reexamining my impact, motivations and what gives my work purpose,” Kondic says. “What SURF showed me is that pursuing a research career can go hand-in-hand with the process of finding our place within the global efforts to deliver mindful social progress.”