Recent work has shown that this type of attack, called a poisoning at- tack, enables adversaries to insert backdoors or trojans into the model, enabling ...
確定! 回上一頁