Model Evasion

Adversarial Examples

Inputs specially designed to deceive a machine learning model, exploiting the model's vulnerabilities to cause incorrect predictions while remaining imperceptible to humans.

← Back