arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2106.06624cs.LG
收藏

放松局部鲁棒性

Relaxing Local Robustness

Carnegie Mellon University(卡内基梅隆大学)

Klas Leino, Matt Fredrikson

摘要:

可证明的局部鲁棒性严格排除了小范数对抗性示例,作为解决深度学习中安全问题的一种手段,已经受到了极大的关注。然而,对于一些分类问题,局部鲁棒性不是一个自然的目标,即使在存在对手的情况下;例如,如果图像包含两类主题,则图像的正确标签可以被认为是任意的,因此在它们之间强制严格分离是不必要的。在这项工作中,我们介绍了两个宽松的安全属性的分类器,解决这个问题:(1)宽松的top-k鲁棒性,它作为top-k准确性的模拟;(2)亲和鲁棒性,它指定哪些标签集必须由鲁棒性裕度分离,以及哪些可以是$ε$ -接近$\ell_p$空间。我们展示了如何构建模型,可以有效地对每个放松的鲁棒性属性进行认证,并且相对于标准的梯度下降,训练的开销非常小。最后,我们通过实验证明,这些放松的变体的鲁棒性非常适合几个重要的分类问题,导致较低的拒绝率和较高的认证精度比可以获得认证时,“标准”的本地鲁棒性。

英文摘要:

Certifiable local robustness, which rigorously precludes small-norm adversarial examples, has received significant attention as a means of addressing security concerns in deep learning. However, for some classification problems, local robustness is not a natural objective, even in the presence of adversaries; for example, if an image contains two classes of subjects, the correct label for the image may be considered arbitrary between the two, and thus enforcing strict separation between them is unnecessary. In this work, we introduce two relaxed safety properties for classifiers that address this observation: (1) relaxed top-k robustness, which serves as the analogue of top-k accuracy; and (2) affinity robustness, which specifies which sets of labels must be separated by a robustness margin, and which can be $ε$-close in $\ell_p$ space. We show how to construct models that can be efficiently certified against each relaxed robustness property, and trained with very little overhead relative to standard gradient descent. Finally, we demonstrate experimentally that these relaxed variants of robustness are well-suited to several significant classification problems, leading to lower rejection rates and higher certified accuracies than can be obtained when certifying "standard" local robustness.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑