Adversarial example
Adversarial example: An input intentionally modified to induce model error, often with small perturbations under a chosen metric. White-box attacks use gradients; black-box attacks observe outputs or transfer perturbations. Empirical robustness is not a formal guarantee. This perspective connects the definition with practical behaviour in artificial intelligence systems.
In italiano: Adversarial ExampleLegge V · Overfitting, bias e collasso dei modelliWhite-box attacks use gradients; black-box attacks observe outputs or transfer perturbations. Empirical robustness is not a formal guarantee.