Model Security

Transferability Attack

Phenomenon where an adversarial example, designed to deceive a specific model, also manages to mislead other models with different architectures or training data.

← Zurück