attack success rate

Attack success rate refers to the proportion of adversarial attacks (deliberate attempts to mislead AI models) that successfully achieve the desired misclassification of inputs, important for assessing model robustness.

16 papers