Vincent Ryusuke Takahashi

Improving the Robustness/Accuracy Tradeoff Against Adversarial Attacks Using Information Bottleneck Distillation Through Dual Teachers

The proposed method improves classification accuracy on clean samples compared to the original IBD, while maintaining similar accuracy on adversarial samples, and is competitive with state-of-the-art approaches, including the recent dual-teacher distillation framework B-MTARD.