Fourier-Based Augmentations for Improved Robustness and Uncertainty CalibrationDownload PDF

09 Oct 2021, 14:49 (modified: 30 Nov 2021, 20:54)NeurIPS 2021 Workshop DistShift PosterReaders: Everyone
Keywords: augmentation strategies, distribution shifts, robustness, corruption, Fourier, attack, perturbation
TL;DR: We modify AugMix to target robustness to Fourier-basis noise, while boosting robustness to other common corruptions
Abstract: Diverse data augmentation strategies are a natural approach to improving robustness in computer vision models against unforeseen shifts in data distribution. However, the ability to tailor such strategies to inoculate a model against specific classes of corruptions or attacks---without incurring substantial losses in robustness against other classes of corruptions---remains elusive. In this work, we successfully harden a model against Fourier-based attacks, while producing superior-to-\texttt{AugMix} accuracy and calibration results on both the CIFAR-10-C and CIFAR-100-C datasets; classification error is reduced by over ten percentage points for some high-severity noise and digital-type corruptions. We achieve this by incorporating Fourier-basis perturbations in the \texttt{AugMix} image-augmentation framework. Thus we demonstrate that the \texttt{AugMix} framework can be tailored to effectively target particular distribution shifts, while boosting overall model robustness.
1 Reply