Please join us online for a talk by Dr. She will present 鈥�Statistical Inference for Evolving Time Series: Change Detection and Conformal Prediction鈥� on Wednesday, October 7, at 1:30 pm Central Time. For access to this seminar, contact series administrator Cierra Streeter.
Yao Xie, PhD, is the Coca-Cola Foundation Chair and Professor in the H. Milton Stewart School of Industrial and Systems Engineering at the Georgia Institute of Technology, where she also serves as Associate Director of the (ML@GT). She received her Ph.D. in Electrical Engineering with a minor in Mathematics from Stanford University in 2012 and was previously a Research Scientist at Duke University. Her research lies at the intersection of statistics, machine learning, and optimization, with a focus on statistically sound and computationally efficient methods for high-dimensional, sequential, and spatio-temporal data. She is a member of the 2026 cohort of the National Academies鈥� New Voices in Sciences, Engineering, and Medicine program and the IEEE Information Theory Society Distinguished Lecturer for 2026 to 2027. Her honors include the NSF CAREER Award, the INFORMS Gaver Early Career Award for Excellence in Operations Research, and the C.W.S. Woodroofe Award. She serves as an Associate Editor for multiple journals, including Operations Research, IEEE Transactions on Information Theory, and the Journal of the American Statistical Association (Theory and Methods), and as an Area Chair for NeurIPS, ICML, and ICLR.
Abstract:
Modern time-series data are often observed sequentially in settings where dependence and distributional change are the rule rather than the exception. This talk considers two statistical inference problems for such evolving systems. First, I will discuss sequential change-point detection, with an emphasis on nonparametric methods for detecting unknown distributional shifts quickly while controlling false alarms. In particular, I will describe an online kernel CUSUM approach that extends classical CUSUM beyond parametric likelihood models. Second, I will turn to conformal prediction for time series, where temporal dependence challenges the classical exchangeability assumption underlying conformal inference. I will discuss methods for constructing prediction sets with reliable uncertainty quantification under temporal dependence and distributional change. Together, these two lines of work illustrate how classical statistical ideas can be adapted to provide reliable detection and prediction when data arrive over time and the underlying system does not remain fixed.