Evaluating the Impact of Geometric and Statistical Skews on Out-of-Distribution Generalization Performance
Abstract
Out-of-distribution (OOD) or domain generalization is the problem of generalizing to unseen distributions. Recent work suggests that the marginal difficulty of generalizing to OOD over in-distribution data (OOD-ID generalization gap) is due to spurious correlations, which arise due to statistical and geometric skews, and can be addressed by careful data augmentation and class balancing. We observe that after constructing a dataset where we remove all conceivable sources of spurious correlation between interpretable factors, classifiers still fail to close the OOD-ID generalization gap.
Cite
Text
Lynch et al. "Evaluating the Impact of Geometric and Statistical Skews on Out-of-Distribution Generalization Performance." NeurIPS 2022 Workshops: CML4Impact, 2022.Markdown
[Lynch et al. "Evaluating the Impact of Geometric and Statistical Skews on Out-of-Distribution Generalization Performance." NeurIPS 2022 Workshops: CML4Impact, 2022.](https://mlanthology.org/neuripsw/2022/lynch2022neuripsw-evaluating/)BibTeX
@inproceedings{lynch2022neuripsw-evaluating,
title = {{Evaluating the Impact of Geometric and Statistical Skews on Out-of-Distribution Generalization Performance}},
author = {Lynch, Aengus and Kaddour, Jean and Silva, Ricardo},
booktitle = {NeurIPS 2022 Workshops: CML4Impact},
year = {2022},
url = {https://mlanthology.org/neuripsw/2022/lynch2022neuripsw-evaluating/}
}