Evasion Attacks & Defenses

FGSM, PGD, adversarial training, input sanitization, and detection — defending inference-time attacks.

Part of AI Security on neo-ai.