Training-Free Robust Multimodal Learning via Sample-Wise Jacobian Regularization

Re-assign community