<ul><li>Anchor Regression proposes a method to enforce stability and robustness in predictive models.</li><li>The method relaxes the regularization in the optimization problem and interpolates between partialling out and instrumental variable estimation.</li><li>The solution of Anchor Regression optimizes worst-case risk under shift interventions on anchors.</li><li>The method increases the robustness of predictions to distribution shifts at the cost of reducing in-distribution generalization.</li></ul>

Anchor Regression: The Secret to Stable Predictions Across Shifting Data

Discover more