Suyun Liu: Working toward bias-free machines

Suyun Liu, PhD student, Industrial and Systems Engineering

It was mostly the optimization group within the Department of Industrial and Systems Engineering that brought Suyun Liu to Lehigh.

“Everyone is talking about machine learning today, and Lehigh’s optimization program is one of the top programs in the U.S.,” says Liu. But the other compelling draw? Her fellow Chinese students. Liu had reached out to them before accepting her PhD offer, and was immediately struck by how friendly and supportive they were. “The program itself was great for me, but the students were another big reason why I chose Lehigh.”

Liu is part of a group focusing on improving algorithm design for continuous optimization. Specifically, they’re looking at multi-objective optimization and its application to fairness in machine learning. For example, when a bank’s decision-making system collects data on loan applicants, it may produce results that are biased toward or against certain groups. Liu’s research aims to eliminate that bias while retaining the model’s accuracy.

Ultimately, she says, the team is helping lay the foundation for the machine learning community to apply their findings to any model.

Liu has spoken at conferences and seminars, and a paper she co-authored with her PhD advisor Luis Nunes Vicente was recently accepted by Annals of Operations Research. Two more papers were, at press time, under review.

She knows her experience at Lehigh has given her the tools she needs to get where she wants to go.

“This is a really practical area of optimization, and it’s one that a lot of industries use,” she says. “I know the work I’m doing is building a strong resume to get a good job in industry."

Next up: Shira Morosohk >>


This profile is part of Resolve Magazine's Soaring Together series.

Photography by Douglas Benedict/Academic Image