IPMix: Label-Preserving Data Augmentation Method for Training Robust Classifiers

TL;DR

IPMix is a label-preserving data augmentation method that enhances neural network robustness against distribution shifts and adversarial attacks without sacrificing accuracy, by combining multi-level augmentation techniques.

Contribution

IPMix introduces a novel, simple, multi-level data augmentation approach that improves robustness across various benchmarks while maintaining high accuracy.

Findings

Outperforms state-of-the-art on CIFAR-C and ImageNet-C.

Improves robustness to adversarial perturbations.

Enhances calibration and anomaly detection.

Abstract

Data augmentation has been proven effective for training high-accuracy convolutional neural network classifiers by preventing overfitting. However, building deep neural networks in real-world scenarios requires not only high accuracy on clean data but also robustness when data distributions shift. While prior methods have proposed that there is a trade-off between accuracy and robustness, we propose IPMix, a simple data augmentation approach to improve robustness without hurting clean accuracy. IPMix integrates three levels of data augmentation (image-level, patch-level, and pixel-level) into a coherent and label-preserving technique to increase the diversity of training data with limited computational overhead. To further improve the robustness, IPMix introduces structural complexity at different levels to generate more diverse images and adopts the random mixing method for multi-scale information fusion. Experiments demonstrate that IPMix outperforms state-of-the-art corruption robustness on CIFAR-C and ImageNet-C. In addition, we show that IPMix also significantly improves the other safety measures, including robustness to adversarial perturbations, calibration, prediction consistency, and anomaly detection, achieving state-of-the-art or comparable results on several benchmarks, including ImageNet-R, ImageNet-A, and ImageNet-O.