Skip to main content Skip to secondary navigation
Main content start

Dominik Rothenhaeusler (Stanford): Out-of-distribution generalization under random, dense distributional shifts

Event Details:

Thursday, April 25, 2024
5:00pm - 6:00pm PDT


Huang 305
475 Via Ortega
Stanford, CA 94305
United States

Dominik Rothenhaeusler (Stanford): Out-of-distribution generalization under random, dense distributional shifts

Abstract: Many existing approaches for estimating parameters in settings with distributional shifts operate under an invariance assumption. For example, under covariate shift, it is assumed that p(y|x) remains invariant. We refer to such distribution shifts as sparse, since they may be substantial but affect only a part of the system. In contrast, in various real-world settings, shifts might be dense. More specifically, these shifts may arise through many small and random changes in the population and environment. First, we will discuss empirical evidence for such random dense distributional shifts and explain why commonly used models for distribution shifts—including adversarial approaches—may not be appropriate under these conditions. Then, we will develop tools to infer parameters and make predictions for partially observed, shifted distributions. Finally, we will apply the framework to several real-world datasets and discuss diagnostics to evaluate the fit of the distributional uncertainty model.

This is joint work with Yujin Jeong, Ying Jin, and Ivy Zhang.

Bio: Dominik Rothenhaeusler is an Assistant Professor of Statistics at Stanford University. His research centers around causal inference, heterogeneous data, high-dimensional statistics and graphical models. He is particularly interested in inference in settings where traditional statistical measures of uncertainty fail. For example, due to changing circumstances or confounding the distribution of the data might change between data sets. Developing better methods to deal with distribution shift could increase replicability of feature selections and reliability of prediction mechanisms. He is the recipient of the David Cox Research Prize by the Royal Statistical Society. He completed his Ph.D. studies in mathematics and statistics at ETH Zurich. Prior to joining the Stanford faculty, he completed a postdoc at UC Berkeley.

Related Topics

Explore More Events

No events at this time. Please check back later.