حملات جعبه‌سفید

تعریف در منبع (انگلیسی)

Attacks assuming the attacker has full knowledge of the ML system, including training data, model architecture, and hyperparameters.