Rethinking Cost-sensitive Classification in Deep Learning via Adversarial Data Augmentation

Open in new window