WhiteMail
← 术语表

AI 与检测引擎

적대적 예제 (Adversarial Example)

定义

An input subtly perturbed to be misclassified by a model while appearing normal to humans. A core threat in ML security.

了解更多

同类别的其他术语