“Adversarially robust classifiers have interpretable gradients and feature representations [0]. The problem seems to be that the standard networks capture all the statistics there is, including surface statistical regularities and noise. It can be mitigated, though. [0]: https://arxiv.org/abs/1805.12152”
Robustness May Be at Odds with Accuracy
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.
候选层:至少 2 个不同的人、在 2 个不同月份提到过,还没到主书架的 3 人 3 个月。没有做立场标注,所以只写“被提及”。
HN 书评
推荐在前,同类按时间 · 每篇最多保留 20 条,统计用全部提及“We took a look at adversarial examples for linear classifiers (and in general, we looked at properties that adversarial training induces) here: https://arxiv.org/abs/1805.12152 For $\ell_\infty$ adversarial examples on linear classifiers we found that adversarial training forces a tradeoff between the $\ell_1$ norm of the weights (which is directly associ…”
“In the light of [0], it seems that standard (not adversarially trained) NNs learn each and every feature which happens to correlate with a class label in a training set (including noise). So moving along adversarial gradient looks like noise. Take my interpretation with a (huge) grain of salt. [0] https://arxiv.org/abs/1805.12152”