a02b5c28b1ebe56bf5513d16aa6ef59d08ee28a5,art/attacks/universal_perturbation.py,UniversalPerturbation,__init__,#UniversalPerturbation#Any#Any#Any#Any#Any#Any#Any#Any#,33

Before Change

        :type expectation: :class:`.ExpectationOverTransformations`
        
        super(UniversalPerturbation, self).__init__(classifier)
        kwargs = {"attacker": attacker,
                  "attacker_params": attacker_params,
                  "delta": delta,
                  "max_iter": max_iter,
                  "eps": eps,
                  "norm": norm,
                  "expectation": expectation
                  }
        self.set_params(**kwargs)

    def generate(self, x, **kwargs):
        

After Change

        
        :param classifier: A trained model.
        :type classifier: :class:`.Classifier`
        :param attacker: Adversarial attack name. Default is "deepfool". Supported names: "carlini", "carlini_inf",
                         "deepfool", "fgsm", "bim", "pgd", "margin", "ead", "newtonfool", "jsma", "vat".
        :type attacker: `str`
        :param attacker_params: Parameters specific to the adversarial attack.
        :type attacker_params: `dict`
        :param delta: desired accuracy
        :type delta: `float`
        :param max_iter: The maximum number of iterations for computing universal perturbation.

Instances