统计与数据科学系系列学术报告之四百三十八期

时间：2024年7月4日（周四）10:00-11:00

地点：史带楼301室

主持人：复旦大学管理学院统计与数据科学系郁文教授

报告人：Xinwei Shen, Ph.D Seminar for Statistics, ETH Zürich

题目：Engression: Extrapolation for Nonlinear Regression?

摘要：Extrapolation is crucial in many statistical and machine learning applications, as it is common to encounter test data outside the training support. However, extrapolation is a considerable challenge for nonlinear models. Conventional models typically struggle in this regard: while tree ensembles provide a constant prediction beyond the support, neural network predictions tend to become uncontrollable. This work aims at providing a nonlinear regression methodology whose reliability does not break down immediately at the boundary of the training support. Our primary contribution is a new method called ‘engression’ which, at its core, is a distributional regression technique for pre-additive noise models, where the noise is added to the covariates before applying a nonlinear transformation. Our experimental results indicate that this model is typically suitable for many real data sets. We show that engression can successfully perform extrapolation under some assumptions such as a strictly monotone function class, whereas traditional regression approaches such as least-squares regression and quantile regression fall short under the same assumptions. We establish the advantages of engression over existing approaches in terms of extrapolation, showing that engression consistently provides a meaningful improvement. Our empirical results, from both simulated and real data, validate these findings, highlighting the effectiveness of the engression method.

个人简介：Xinwei Shen is a postdoctoral researcher at the Seminar for Statistics at ETH Zürich working with professors Peter Bühlmann and Nicolai Meinshausen. She is an incoming assistant professor at the Department of Statistics at the University of Washington. Previously, she obtained her PhD at HKUST supervised by professor Tong Zhang, and a Bachelor of Science degree at Fudan University. Her research interests lie at the interface of statistics and machine learning. Her current research focuses on distributional learning, causality, robustness, as well as climate applications.