Timezone: »

 
Evaluating the Impact of Geometric and Statistical Skews on Out-Of-Distribution Generalization Performance
Aengus Lynch · Jean Kaddour · Ricardo Silva

@
in Causal Machine Learning for Real-World Impact »
Event URL: https://openreview.net/forum?id=ZweYSQ0pXmF »

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.

Author Information

Aengus Lynch (University College London, University of London)
Jean Kaddour (University College London)
Ricardo Silva (University College London)

More from the Same Authors