论文标题
修剪对抗性稳健的神经网络，没有对抗性例子
Pruning Adversarially Robust Neural Networks without Adversarial Examples
论文作者
论文摘要
对抗修剪会在保持鲁棒性的同时压缩模型。当前的方法需要在修剪过程中访问对抗示例。这显着缩减了培训效率。此外，随着新的对抗性攻击和训练方法以快速的速度发展，需要对对抗性修剪方法进行相应的修改以跟上。在这项工作中，我们提出了一个新颖的框架，以修剪先前训练的强大神经网络，同时保持对抗性鲁棒性，而无需进一步产生对抗性例子。我们利用并发的自我鉴定和修剪来保留原始模型中的知识，并通过Hilbert-Schmidt Information Bottleneck将修剪模型正规化。我们全面评估了我们提出的框架，并在对MNIST，CIFAR-10和CIFAR-100数据集进行针对五次最先进的攻击的培训时，在对抗性鲁棒性和效率方面表现出了卓越的性能。代码可在https://github.com/neu-spiral/pwoa/上找到。
Adversarial pruning compresses models while preserving robustness. Current methods require access to adversarial examples during pruning. This significantly hampers training efficiency. Moreover, as new adversarial attacks and training methods develop at a rapid rate, adversarial pruning methods need to be modified accordingly to keep up. In this work, we propose a novel framework to prune a previously trained robust neural network while maintaining adversarial robustness, without further generating adversarial examples. We leverage concurrent self-distillation and pruning to preserve knowledge in the original model as well as regularizing the pruned model via the Hilbert-Schmidt Information Bottleneck. We comprehensively evaluate our proposed framework and show its superior performance in terms of both adversarial robustness and efficiency when pruning architectures trained on the MNIST, CIFAR-10, and CIFAR-100 datasets against five state-of-the-art attacks. Code is available at https://github.com/neu-spiral/PwoA/.
