Adversarial and Natural Perturbations for General Robustness
arXiv:2010.01401
Abstract
In this paper we aim to explore the general robustness of neural network classifiers by utilizing adversarial as well as natural perturbations. Different from previous works which mainly focus on studying the robustness of neural networks against adversarial perturbations, we also evaluate their robustness on natural perturbations before and after robustification. After standardizing the comparison between adversarial and natural perturbations, we demonstrate that although adversarial training improves the performance of the networks against adversarial perturbations, it leads to drop in the performance for naturally perturbed samples besides clean samples. In contrast, natural perturbations like elastic deformations, occlusions and wave does not only improve the performance against natural perturbations, but also lead to improvement in the performance for the adversarial perturbations. Additionally they do not drop the accuracy on the clean images.
arXiv commentCurrently under review
References in corpus (8)
- Towards Deep Learning Models Resistant to Adversarial Attacks
- Theoretically Principled Trade-off between Robustness and Accuracy
- Robustness May Be at Odds with Accuracy
- Exploring the Landscape of Spatial Robustness
- Adversarial Examples Are a Natural Consequence of Test Error in Noise
- Testing Robustness Against Unforeseen Adversaries
- The Limitations of Adversarial Training and the Blind-Spot Attack
- A Large-scale Attribute Dataset for Zero-shot Learning