Adversarial machine learning, i.e., increasing the robustness of machine learning algorithms against so-called adversarial examples, is now an established field. Yet, newly proposed methods are evaluated and compared under unrealistic scenarios where costs for adversary and defender are not considered and either all samples or no samples are adversarially perturbed. We scrutinize these assumptions and propose the advanced adversarial classification game, which incorporates all relevant parameters of an adversary and a defender. Especially, we take into account economic factors on both sides and the fact that all so far proposed countermeasures against adversarial examples reduce accuracy on benign samples. Analyzing the scenario in detail, where both players have two pure strategies, we identify all best responses and conclude that in practical settings, the most influential factor might be the maximum amount of adversarial examples.
翻译:反对立机器学习,即提高机器学习算法对所谓的对抗性实例的稳健性,现已成为一个既定领域,然而,在不考虑对手和辩护人费用、不考虑所有样品或没有样品的不现实情况下,对新提议的方法进行评价和比较,我们仔细研究这些假设,并提出先进的对抗性分类游戏,其中包括对手和辩护人的所有相关参数。特别是,我们考虑到双方的经济因素,以及迄今为止所有针对对抗性例子提出的反措施都降低了良性样品的准确性。我们详细分析两种情况,即双方都有两种纯粹的战略,我们确定所有最佳的对策,并得出结论认为,在实际环境中,最有影响力的因素可能是对抗性例子的最大数量。