Jan Kautz

Do Gradient Inversion Attacks Make Federated Learning Unsafe?

2023 · IEEE Transactions on Medical Imaging

Federated learning (FL) allows the collaborative training of AI models without needing to share raw data. This capability makes it especially interesting for healthcare applications where patient and data privacy is of utmost concern. However, …

The Best Defense is a Good Offense: Adversarial Augmentation against Adversarial Attacks

2023 · arXiv (Cornell University)

Many defenses against adversarial attacks (\eg robust classifiers, randomization, or image purification) use countermeasures put to work only after the attack has been crafted. We adopt a different perspective to introduce $A^5$ (Adversarial Augmentation Against …