Thenweusethevariantsubgraphs,i.e., complements of invariant subgraphs, to infer the latent environment labels.

In this paper, we show this is not the case by disentangling the notions of expressivity and learnability.

We develop a new primitive for stochastic optimization: a low-bias, low-cost estimator of the minimizerx? of any Lipschitz strongly-convex function.

Wepostulate thatthevector ofFGW distances to a set of template graphs has a strong discriminative power, which is then fed to anon-linear classifier forfinalpredictions.