鲁棒性可能与准确率相冲突
Robustness May Be at Odds with Accuracy
Dimitris Tsipras, Shibani Santurkar, Logan Engstrom, Alexander Turner, Aleksander Madry
AI 速读 基于摘要
作者论证对抗鲁棒性与标准泛化之间存在内在张力，训练鲁棒模型可能导致标准准确率下降，但鲁棒模型学到的特征表示更符合人类感知。
stat.ML · cs.CV · cs.LG · cs.NE
查看原文摘要
We show that there may exist an inherent tension between the goal of adversarial robustness and that of standard generalization. Specifically, training robust models may not only be more resource-consuming, but also lead to a reduction of standard accuracy. We demonstrate that this trade-off between the standard accuracy of a model and its robustness to adversarial perturbations provably exists in a fairly simple and natural setting. These findings also corroborate a similar phenomenon observed empirically in more complex settings. Further, we argue that this phenomenon is a consequence of robust classifiers learning fundamentally different feature representations than standard classifiers. These differences, in particular, seem to result in unexpected benefits: the representations learned by robust models tend to align better with salient data characteristics and human perception.