Students Solve Real-World Problems with Summer AI Research

How do you build a future workforce of data science and artificial intelligence leaders? At WPI, one approach is to immerse undergraduates in an intense summer program—the Applied Artificial Intelligence for Advanced Applications research experience. With a three-year grant of $461,597 for a Research Experience for Undergraduates site from the National Science Foundation, WPI launched this 10-week research-intensive program in June 2024. Eleven students from eight institutions across the United States completed the program, successfully presenting their research findings on August first at WPI’s annual summer undergraduate research showcase. The students worked in teams of two or three to tackle diverse AI challenges, including the development of better algorithms for computer vision to the design of a model that can detect and remove biased language in medical curricula. “A summer program like this provides undergraduates with the opportunity to engage in cutting-edge AI research and experience firsthand its potential to solve critical societal problems, all of which inspires them to return to their colleges and universities and choose the course and project work that will best prepare them for graduate studies,” says Elke Rundensteiner, the William B. Smith Professor of Computer Science and founding head of WPI’s Data Science and Artificial Intelligence programs. “We challenge undergraduates with real problems, which require them to learn material that they have never before encountered. This experience pushes them academically and personally and helps them gain confidence.” Led by Rundensteiner as principal investigator and co-investigator Kelsey Briggs, director of Programs and Strategic Initiatives for Data Science, the summer experience built upon the university’s previous eight years of summer data science research programs for undergraduates. The program also reflected WPI’s long history of work on topics ranging from machine learning and deep learning to generative modeling. Students worked in teams mentored by WPI faculty advisors Fabricio Murai, Chun-Kit Ngan, Ziming Zhang, and Rundensteiner, as well as the faculty advisors’ graduate students, on authentic research topics. A diverse group of undergraduates was chosen for the program through an application process that included an essay. Those accepted received housing on campus, a stipend, and reimbursement for travel expenses. “I have long wanted to do research, and I’m especially passionate about studying Parkinson’s Disease and neurogenerative disorders,” says Olivia Liau, a member of the Class of 2027 at the University of Southern California who worked on a project that used machine learning to make predictions about genomic datasets based on gene expression. “Our project and our results were really cool. I learned so much about working with data.” In addition to conducting research, students met with industry speakers from the WPI Executive Advisory Board for Data Science and Artificial Intelligence to learn about career paths in startup companies, mid-size businesses, and major corporations. Some of the students are also preparing manuscripts of their research for publication in academic journals. Gavin Butts, Class of 2025 at Loyola Marymount University, worked with a team that trained the language model RoBERTa on about 8,000 excerpts from medical textbooks, slides, and documents so that the model could be used to remove biased language from medical curricula. He says he appreciated the opportunity to use AI to address a societal problem. “Many people are using AI for commercial purposes, to make money,” Butts says. “But there are also many of us who want to use AI to improve the world.”