Alexandros Fakis

This paper provides the first comprehensive review of threats and defenses specifically targeting on-device AI inference, revealing a significant imbalance where certain attack types, like adversarial attacks, lack corresponding defense mechanisms.