Safe Augmentation: Learning Task-Specific Transformations from Data

Re-assign community