A low-rank unlearning approach for Vision Transformers that avoids retaining the original training data.
Dec 1, 2024