论文标题
对抗性稳健的神经网络的重量协方顺序
Weight-Covariance Alignment for Adversarially Robust Neural Networks
论文作者
论文摘要
最近已证明将噪声注入其隐藏层的随机神经网络（SNN）可针对对抗性攻击实现强大的鲁棒性。但是，现有的SNN通常是出于启发性的，并且通常依靠对抗性训练，这在计算上是昂贵的。我们提出了一个新的SNN，可以在不依赖对抗性训练的情况下实现最先进的表现，并具有坚实的理论理由。具体而言，尽管现有的SNN注入了学习或手动调整的各向同性噪声，但我们的SNN学习了各向异性噪声分布，以优化对对抗性鲁棒性的学习理论。我们在许多流行的基准测试中评估了我们的方法，表明它可以应用于不同的体系结构，并为各种白色盒子和黑色盒子攻击提供了鲁棒性，同时与现有替代方案相比简单而快速地训练。
Stochastic Neural Networks (SNNs) that inject noise into their hidden layers have recently been shown to achieve strong robustness against adversarial attacks. However, existing SNNs are usually heuristically motivated, and often rely on adversarial training, which is computationally costly. We propose a new SNN that achieves state-of-the-art performance without relying on adversarial training, and enjoys solid theoretical justification. Specifically, while existing SNNs inject learned or hand-tuned isotropic noise, our SNN learns an anisotropic noise distribution to optimize a learning-theoretic bound on adversarial robustness. We evaluate our method on a number of popular benchmarks, show that it can be applied to different architectures, and that it provides robustness to a variety of white-box and black-box attacks, while being simple and fast to train compared to existing alternatives.
