The issue of area generalization considerations studying predictive fashions which can be sturdy to distribution shifts when deployed in new, beforehand unseen environments. Present strategies usually require labeled information from a number of coaching environments, limiting their applicability when labeled information are scarce. On this work, we examine area generalization in an anti-causal setting, the place the result causes the noticed covariates. Below this construction, setting perturbations that have an effect on the covariates don’t propagate to the result, which motivates regularizing the mannequin’s sensitivity to those perturbations. Crucially, estimating these perturbation instructions doesn’t require labels, enabling us to leverage unlabeled information from a number of environments. We suggest two strategies that penalize the mannequin’s sensitivity to variations within the imply and covariance of the covariates throughout environments, respectively, and show that these strategies have worst-case optimality ensures underneath sure courses of environments. Lastly, we display the empirical efficiency of our method on a managed bodily system and a physiological sign dataset.
- †Apple
- ‡ ETH Zürich
