Robustmix: Improving Robustness by Regularizing the Frequency Bias of Deep Nets
This work addresses the robustness issue in deep learning for computer vision applications, offering an incremental improvement over existing methods.
The paper tackles the problem of deep networks being sensitive to perturbations by proposing Robustmix, a novel extension of Mixup that regularizes networks to classify based on lower-frequency spatial features, resulting in improved robustness on benchmarks like Imagenet-C and Stylized Imagenet, with a state-of-the-art mCE of 44.8 using EfficientNet-B8 and RandAugment, a reduction of 16 mCE compared to the baseline.
Deep networks have achieved impressive results on a range of well-curated benchmark datasets. Surprisingly, their performance remains sensitive to perturbations that have little effect on human performance. In this work, we propose a novel extension of Mixup called Robustmix that regularizes networks to classify based on lower-frequency spatial features. We show that this type of regularization improves robustness on a range of benchmarks such as Imagenet-C and Stylized Imagenet. It adds little computational overhead and, furthermore, does not require a priori knowledge of a large set of image transformations. We find that this approach further complements recent advances in model architecture and data augmentation, attaining a state-of-the-art mCE of 44.8 with an EfficientNet-B8 model and RandAugment, which is a reduction of 16 mCE compared to the baseline.