Presented By: U-M Industrial & Operations Engineering
SEMINAR: "Unbiased Subdata Selection for Fair Classification: A Unified Framework and Scalable Algorithms" – Weijun Xie
The Departmental Seminar Series is open to all. U-M Industrial and Operations Engineering graduate students and faculty are especially encouraged to attend.
Title:
Unbiased Subdata Selection for Fair Classification: A Unified Framework and Scalable Algorithms
Abstract:
Fair classification concerns the issues of unintentional biases against the sensitive features (e.g., gender, race) in the conventional classification approaches. Due to the high nonconvexity of fairness measures, existing methods are often unable to model exact fairness, which can cause inferior fair classification outcomes. This paper fills the gap by developing a novel unified framework to jointly optimize accuracy and fairness. The proposed framework is versatile and can incorporate different fairness measures precisely as well as can be applicable to many classifiers, including deep classification models. Many classification models within this framework can be recast as mixed-integer convex programs, which can be solved effectively by off-the-shelf solvers when the instance sizes are moderate. We prove that in the proposed framework, when the classification outcomes are known, the resulting problem, termed “unbiased subdata selection,” is strongly polynomial-solvable and can be used to enhance the classification fairness by selecting more representative data points. This motivates us to develop an iterative refining strategy (IRS) to solve the large-scale instances, where we improve the classification accuracy and conduct the unbiased subdata selection in an alternating fashion. We numerically demonstrate that the proposed framework can consistently yield better fair classification outcomes than existing methods. This is a joint work of my Ph.D. student Qing Ye.
Bio:
Dr. Weijun Xie is an Assistant Professor of Industrial and Systems Engineering, Virginia Tech. He obtained his Ph.D. from Georgia Tech. His research interests are in theory and applications of stochastic, discrete, and convex optimization. Dr. Xie has won multiple awards including NSF Career Award, INFORMS Optimization Prize for Young Researchers, INFORMS Junior Faculty Interest Group Paper Competition, INFORMS George Nicholson Student Paper Competition. He currently serves as the Vice Chair of Optimization under Uncertainty at INFORMS Optimization Society.
Title:
Unbiased Subdata Selection for Fair Classification: A Unified Framework and Scalable Algorithms
Abstract:
Fair classification concerns the issues of unintentional biases against the sensitive features (e.g., gender, race) in the conventional classification approaches. Due to the high nonconvexity of fairness measures, existing methods are often unable to model exact fairness, which can cause inferior fair classification outcomes. This paper fills the gap by developing a novel unified framework to jointly optimize accuracy and fairness. The proposed framework is versatile and can incorporate different fairness measures precisely as well as can be applicable to many classifiers, including deep classification models. Many classification models within this framework can be recast as mixed-integer convex programs, which can be solved effectively by off-the-shelf solvers when the instance sizes are moderate. We prove that in the proposed framework, when the classification outcomes are known, the resulting problem, termed “unbiased subdata selection,” is strongly polynomial-solvable and can be used to enhance the classification fairness by selecting more representative data points. This motivates us to develop an iterative refining strategy (IRS) to solve the large-scale instances, where we improve the classification accuracy and conduct the unbiased subdata selection in an alternating fashion. We numerically demonstrate that the proposed framework can consistently yield better fair classification outcomes than existing methods. This is a joint work of my Ph.D. student Qing Ye.
Bio:
Dr. Weijun Xie is an Assistant Professor of Industrial and Systems Engineering, Virginia Tech. He obtained his Ph.D. from Georgia Tech. His research interests are in theory and applications of stochastic, discrete, and convex optimization. Dr. Xie has won multiple awards including NSF Career Award, INFORMS Optimization Prize for Young Researchers, INFORMS Junior Faculty Interest Group Paper Competition, INFORMS George Nicholson Student Paper Competition. He currently serves as the Vice Chair of Optimization under Uncertainty at INFORMS Optimization Society.
Related Links
Livestream Information
ZoomSeptember 16, 2021 (Thursday) 3:00pm
Meeting ID: 91950967806
Explore Similar Events
-
Loading Similar Events...