QI Internship Opens Exchange in Artificial Intelligence, Culture between UC San Diego and Korea

This summer, 11 undergraduates from South Korea traveled to the UC San Diego Qualcomm Institute (QI) to learn techniques for applying artificial intelligence to elements of daily life. Hailing from Keimyung University and Kwangwoon University, these students were the latest to participate in QI’s Artificial Intelligence (AI) Development Project, a month-long exploration of AI in fields like energy, transportation and health.

“The QI AI Development Project is an example of a successful model that demonstrates how to collaborate internationally and respect diversity in an educational environment,” said Seokheon “Justin” Cho, the program’s principal investigator, a QI principal network architect and a professor in South Korea. 

Cho, who co-teaches the students alongside QI researcher Troy Bu, founded the program in 2013, both to pass on practical applications in a fast-growing field and to expose participants to a new culture. The program is active for six weeks from January to February and July to August, and has hosted students from ten universities across South Korea. 

At QI, the AI Development Project fits neatly into an ecosystem of research and mentorship programs in high-tech fields. Similar initiatives include nanotechnology programs for undergraduates and for teachers at the high school and middle school levels, and Neurodiversity in Tech, a summer internship program that helps prepare adults with autism spectrum disorder for careers in technology.

AI, from Credit Fraud to Public Health

The AI Development Project students, who were grouped into competing teams, began their summer with daily courses on crucial and fundamental AI algorithms, big data analysis, research papers and AI analytics platforms. Cho and Bu bolstered their learning experience through office hours and review sessions, and challenged each team to present their progress to their peers weekly.

The winning team, composed of Jiyong Park, Justin Song and Jeongsu Yoon, examined patterns of credit card fraud across the United States. Drawing on public repositories, they tried several tactics before finding the data that best fit their approach. 

Their results, which received marks for excellence in methodology, indicated that credit card fraud was highest in New York and Texas.

“The students showed active thinking in making creative and meaningful features to improve their model’s performance,” said Bu.

Other projects tackled problems of classification and public health and safety. Gyu Hyeok Kim and DongWook Kim (no relation) examined classification algorithms to distinguish between tweets from several celebrities with the first name “Justin.” Yeonju Park and Jiyeon Kim looked at rates of survival from breast cancer. Jiwoon Lee and Junhyeong Ryu tackled the development of a method for differentiating benign and malignant tumors from images.

‘Surprising’ Results

Among the participants, senior undergraduate KyungMin Oh stood apart in part for her interest in psychology. 

With fellow undergraduates DongWook Kwon and Yuntae Park, Oh examined the potential for a connection between psychology’s “Big Five” personality traits, including openness to experience, conscientiousness, extraversion, agreeableness and neuroticism, and risk of drug addiction.  

Building from previous studies, Oh, Kwon and Park hypothesized that individuals who scored high in traits like neuroticism, which influences a person’s tendency toward negative emotions, and overlaps with impulsive and sensation-seeking behaviors, would have a greater chance of abusing drugs. 

The interns collected public data on the age, gender, education and ethnicity of more than 1,000 individuals globally and cross-analyzed it with rates of consumption of drugs such as cannabis, amphetamines, ecstasy, heroin and LSD. Though they, too, had to try several different methods, they eventually charted their findings within a range of acceptable accuracy. There was one clear result: personality did not correlate strongly with drug use. 

“I was surprised,” said Oh. 

Though she remains unsure of whether she’ll continue in artificial intelligence, Oh says the internship did offer her a sense of satisfaction and a chance to explore a field that combines engineering, computer science and communications. Following graduation, she plans to pursue a career in clinical psychology, where she’ll be working to diagnose individuals living with depression and other mental health conditions.

With Cho’s leadership, the QI AI Development Project will continue to offer opportunities for students looking to improve their career prospects in artificial intelligence and gain enriching experiences in the U.S. To learn more about QI’s educational programs, see https://qi.ucsd.edu/education-outreach.php. To view a video with scenes from this year’s program, see below.