Linear unit-tests for invariance discovery
This work addresses the need for standardized benchmarks in out-of-distribution generalization for researchers, though it is incremental as it focuses on evaluation rather than new methods.
The authors proposed six linear low-dimensional unit tests to evaluate out-of-distribution generalization algorithms, finding that none of three recent methods passed all tests.
There is an increasing interest in algorithms to learn invariant correlations across training environments. A big share of the current proposals find theoretical support in the causality literature but, how useful are they in practice? The purpose of this note is to propose six linear low-dimensional problems -- unit tests -- to evaluate different types of out-of-distribution generalization in a precise manner. Following initial experiments, none of the three recently proposed alternatives passes all tests. By providing the code to automatically replicate all the results in this manuscript (https://www.github.com/facebookresearch/InvarianceUnitTests), we hope that our unit tests become a standard steppingstone for researchers in out-of-distribution generalization.