VVZ API is not affiliated with ETH Zurich. Data might be outdated or incorrect. Please view the official ETHZ Vorlesungsverzeichnis for binding information.
Student Seminar in Statistics: Distribution Generalization
Last Updated: 2026-07-21 00:35:21
Abstract
When learning how to predict a response variable from some covariates, we often assume that training and test distributions are equal. This may not be satisfied in practice, however. In this seminar, we study the problem of distribution generalization, which allows training and test distributions to differ. We consider statistical models, methods and algorithms, and theoretical results.
Objective
The participants of the seminar acquire knowledge about existing models and methods for distribution generalization, theoretical guarantees, and theoretical limitations.