Born in 1987, Chen earned a doctorate in computer science at Carnegie Mellon University in the U.S. She is an assistant professor in the Department of Computer Science and Information Engineering (CSIE) at National Taiwan University (NTU) in Taipei City and heads its Machine Intelligence and Understanding Laboratory. Last year, Chen received a research grant through the Ministry of Science and Technology’s (MOST) Einstein Program for projects led by doctorate holders aged 35 and under.
“When I was a student at Taipei First Girls’ High School, I enrolled in a programming course and advanced classes in math and natural sciences. Back then I wasn’t really fond of conducting experiments, though later I realized that this was because we were doing them in subjects such as chemistry using finicky equiapment like droppers. In computer science, you just need to click a button to run a test.
My college entrance exam scores weren’t good enough for admission to NTU’s Department of Electrical Engineering [EE], a top target for high schoolers specializing in science subjects like me and my classmates. Ultimately, that proved to be a blessing in disguise since I ended up studying in CSIE and my academic interest was always in programming. This department branched off from the hardware-oriented EE [in 1977] and is now challenging its elite status due to the rapid rise of software-focused industries. A new learning paradigm is emerging in the field of computer science. For one thing, a solid grasp of physics is essential for EE, but not really necessary for CSIE. In software development, math and logical thinking are far more important.
When I started studying speech recognition and dialogue systems technologies at universities in Taiwan and the U.S., there was limited interest in the field since smartphones had just been introduced and few people wanted to speak to a computer. Little progress was achieved until about five or six years ago, when academic and commercial institutions began making breakthroughs in areas like big data applications, deep learning and natural language processing. Even before the business potential of speech recognition became apparent, I was always fascinated by the discipline. I remember watching films as a kid that featured characters chatting with computers and giving them verbal commands. I found this kind of direct interaction inspiring and always believed it was possible to develop such systems.
(Photo by Huang Chung-hsin)
After finishing my doctorate, I worked at Microsoft Research in the U.S. for about six months before accepting a teaching post at NTU. You can earn more in the private sector, but you don’t have the freedom to determine your focus and follow your passions. In the academic world, you can contribute to pioneering projects and these may ultimately garner interest from businesses and achieve commercialization. So it’s often easier to move from academia to industry than the other way round.
The other reasons I chose to come back to Taiwan are the lifestyle here and the talent of local youngsters. U.S. students are brilliant and can express themselves very well. Their Taiwanese counterparts may be even brighter and are great at taking exams and writing papers, but in general they lack the ability to convey their views clearly and to the point. One of my major missions as a teacher is to help students learn how to share their ideas with clarity and confidence.
In 2018, I was awarded a grant under the MOST’s Einstein Program. My team, comprising 10 graduate students, is set to receive annual funding of nearly NT$5 million [US$162,340] for five years for our research project on knowledge acquisition models in language and dialogue processing. This artificial intelligence study isn’t about converting speech to text, which is already quite a mature technology. Rather, it concerns how to recognize the underlying meaning of speech-text inputs and respond accordingly by, for instance, suggesting restaurants in a neighborhood based on a user’s food preferences.
At the same time as leading this project, I’m also raising a 1-year-old. Being pregnant and looking after a child are formidable responsibilities. On this issue, I think that male managers and policymakers are not gender-aware enough and don’t take women’s needs into consideration. This is not so much a matter of discrimination as it’s a lack of understanding of the burdens of being a mother. I’ve suggested to the dean that female professors with young children be allowed to deliver fewer courses in a semester. Government policies such as the Einstein Program should also factor in related concerns. For instance, the age limit should be extended for a woman project leader if she previously put her career on hold to have children. Overall, I think that a greater female presence at the policymaking level would have a positive impact on many aspects of our society.”
—interview by Pat Gao
Write to Pat Gao at cjkao@mofa.gov.tw