Growing up in Shaoguan—a city of 2.8 million people in south China—Linguistics PhD student Jian Zhu was exposed to multiple languages, including Mandarin, Cantonese, and Hakka (a Chinese dialect). “I grew up speaking all of them,” says Jian.
After high school, Jian (a first-generation college student) attended the Beijing Foreign Studies University. There, more than 60 different languages were taught. The diversity of so many languages helped fuel Jian’s interest in linguistics, which had already been kindled by another factor: At the time, Jian recalls, “I was concerned about the preservation of Chinese dialects like Cantonese, whose presence was threatened by the promotion of Mandarin, the official language of China.”
To learn more about language preservation, Jian says he opened his first linguistic textbook and was “immediately intrigued.” He later applied for the U-M linguistics program, “which has one of the best phonetics laboratories in the U.S.,” says Jian, adding: “I was fortunate enough that I was able to be part of the phonetic research group here.”
Jian’s advisor in Linguistics is Pam Beddor and his co-advisor is David Jurgens in the School of Information. Jian is also pursuing a joint degree in Scientific Computing. In his research, Jian uses computational approaches to inform linguistic theories and facilitate linguistic research. One of his research directions is to develop machine learning models to automate some of the repetitive jobs done by language researchers.
“For example, I have developed a machine learning model to automatically track people's tongue movements in ultrasound images and another machine learning model to segment speech into phonetic units automatically,” explains Jian. “These jobs used to be done by humans and they took a very long time to complete.”
Jian has also briefly ventured into forensic linguistics. One of his papers explores using statistical methods to associate people's identities and their writing styles in social media.
As a computational linguist, Jian says he was fortunate that his work was not affected by the challenges of the pandemic during the last two years, “because I only need a computer to do research. On the good side, working from home does force me to improve my cooking by a lot!”
Looking ahead, beginning in the next academic year, Jian will join the Linguistics Department at the University of British Columbia as an assistant professor. He hopes to continue to advance our understanding of language by combining both machine learning and linguistic theories.