Qianyi Rose Sun is on a musical mission to overcome computational bias. A master's student in the School of Music and president of the Women in Music Technology group, Sun recently completed the Bachelor's portion of the Dual Bachelor's and Master's program.

"People ask me all the time what music technology is. I tell them I work at the intersection of musicology, data science, informatics and artificial intelligence to change the way people think about, experience, and consume music," Sun said.

On a research project with Assistant Professor Claire Arthur, Sun realized that datasets from the US and Europe that teach computers to understand music were biased toward Western music.

With a background in Chinese traditional music, Sun saw the opportunity for better musical representation. "One day I want to see these models generate Chinese traditional music or music from other ethnic groups."

Not only that, she's also working on ways to make AI-generated models of music sound good. "A lot of generated music sounds really bad. It doesn't sound like real music," she said.

"In machine learning, the focus of generative models of music has been on improving the technical aspects of generated music. Think quantity and diversity of data, model architecture, and trends in the field. But a lot of the time, this comes at the expense of the quality of the music that is being generated."