As a top undergraduate student in her class at Tsinghua University in Beijing, Yanjun Qi had the opportunity to conduct research with a faculty member on methods for multimedia data mining. The experience changed her life.
An earlier homework assignment on computer recognition of written characters had whetted her interest in machine learning. Her mentored research project confirmed it.
“I found that I really enjoyed the process of developing mathematical techniques that could be used to identify patterns in the data,” Qi said.
Machine learning, a type of artificial intelligence, allows computers to handle new situations via analysis, observation and experience.
Over her career, Qi – now an assistant professor of computer science in the University of Virginia’s School of Engineering and Applied Science – has applied machine learning in a variety of contexts. Among other projects, she has developed algorithms to help researchers use blood samples to diagnose cancer and identify biomarkers of disease; devised a method to help physicians identify electronic medical records that fit specific criteria; and created a technique to identify the kinds of sentiments expressed in online reviews.
Exploring each new application becomes a way for her to create, test and refine machine learning techniques.
“My research is motivated by the problems posed by a specific application,” she said. “I view each experience as an opportunity to develop generic approaches that could be applied to other sets of data with similar properties.”
Nonetheless, Qi takes her choice of application seriously. “There are many data-driven problems you can tackle,” she said. “I want to work on projects that can make a difference.”
Qi is increasingly focusing on biomedical and health care applications. One reason she chose to come to U.Va. is the proximity of the School of Medicine and the opportunity to work with researchers there, both to advance scientific research and to help make health care more efficient. Qi was also attracted by the computer science department’s research strengths in privacy and security as well as sensors, which she feels are both critical to helping people enjoy healthier lives.
Qi has already embarked on a number of productive collaborations within the Engineering School. She has worked with Professor John Lach, chair of the Department of Electrical and Computer Engineering, post-doctoral fellow Jiaqi Gong and graduate student Philip Asare to analyze the potential of using gait monitoring as an indicator of health. They won the best paper award at the ICST BodyNets conference this past fall. She is exploring similar issues with computer science professor Jack Stankovic, in this case determining if online behavior can be used as a measure of health.
Qi has also become an affiliated faculty member with the Medical School’s Center for Public Health Genomics, which is dedicated to translating discoveries about genes and gene interactions into advances in health care and disease prevention. With Mazhar Adli, an assistant professor in the Department of Biochemistry and Molecular Genetics, she is exploring the use of machine learning to discern patterns at the genome and epigenome level related to human disease and cell development. She is also in discussion with Hui Li, an assistant professor of pathology, on ways to identify gene fusion events that can lead to cancer.
After earning her doctorate from Carnegie Mellon University, Qi joined NEC Labs America, drawn in part by the opportunity to work with such giants in her field as Vladimir Vapnik, before opting to return to academia. Qi discovered not only that she loves teaching, but also that it gives her the opportunity to encourage new generations of students to pursue their interests, much as her teachers had done for her when she was an undergraduate.
She enjoys the challenge. “There’s no one way to mentor students,” she said. “I adjust my approach depending on their character, personality and passion.”
-- by Charlie Feigenoff