Dear Colleagues,
You are invited to attend Quality, Statistics and Reliability (QSR) Section Webinar entitled "Guided flow generative models: A unified framework for robustness and inference", which will be given by Prof. Yao Xie (GaTech) on Wed. Oct. 15th between 1PM and 2PM EST.
We look forward to your participation in this informative event. Please feel free to forward this invitation to other interested colleagues or students. The Webinar Registration link is as follows:
Zoom Registration: https://uiowa.zoom.us/meeting/register/31VZJ7OWTlulhOWUFD5C4w
Abstract: Generative AI has achieved remarkable success in creating realistic data such as images and text, yet most advances focus on imitation-replicating what has been observed. In contrast, many problems in statistics, reliability, and engineering require models that can infer, extrapolate, or stress-test-that is, represent what has not yet been observed. In this talk, I will present a line of work that addresses these needs by treating generative models not as black-box imitators but as implicit representations of complex probability distributions that are otherwise intractable with standard statistical or optimization techniques. Our unifying perspective is guided flow modeling: we represent a target distribution Q as the pushforward of a reference distribution P through a continuous flow, implemented via a sequence of residual neural network blocks, and guide that flow using problem-specific formulations-minimax formulations for robustness, likelihoods or posterior consistency for Bayesian-inspired inference, equilibrium conditions for interacting systems, and samples from Q for transfer learning or domain adaptation. This framework yields a new class of generative models that are expressive, computationally efficient, and amenable to rigorous analysis.
Biographical Sketch: Yao Xie is the Coca-Cola Foundation Chair, Professor at Georgia Institute of Technology in the H. Milton Stewart School of Industrial and Systems Engineering, and Associate Director of the Machine Learning Center. From September 2017 until May 2023, she was the Harold R. and Mary Anne Nash Early Career Professor. She received her Ph.D. in Electrical Engineering (minor in Mathematics) from Stanford University in 2012 and was a Research Scientist at Duke University. Her research lies at the intersection of statistics, machine learning, and optimization in providing theoretical guarantees and developing computationally efficient and statistically powerful methods for problems motivated by real-world applications. She received the National Science Foundation (NSF) CAREER Award in 2017, the INFORMS Wagner Prize Finalist in 2021, the INFORMS Gaver Early Career Award for Excellence in Operations Research in 2022, and the CWS Woodroofe Award in 2024. She is currently an Associate Editor for IEEE Transactions on Information Theory, Journal of the American Statistical Association-Theory and Methods, the American Statistician, Operations Research, Annals of Applied Statistics, Sequential Analysis: Design Methods and Applications, INFORMS Journal on Data Science, an Area Chair of NeurIPS, ICML, and ICLR, and Senior Program Committee of AAAI.
------------------------------
Chao Wang
Associate Professor
University of Iowa
Iowa City, IA
------------------------------