Adversarial Training Leads to Weighted Mean Curvature Flow

Adversarial training for binary classification can be connected to a weighted mean curvature flow equation, where the normal velocity of the evolving decision boundary is given by the mean curvature minus the gradient of the log-density of the data.