Patch augmentation: Towards efficient decision boundaries for neural networks
This work addresses the need for more efficient decision boundaries in neural networks for computer vision tasks, offering an incremental improvement over existing augmentation methods.
The paper tackles the problem of improving neural network accuracy and robustness by proposing patch augmentation, a data-independent technique that creates new images by superimposing patches from image pairs and linearly combining their labels, resulting in accuracy increases from 81% to 89% on CIFAR-10 and from 52% to 68% on CIFAR-100, along with enhanced robustness to adversarial attacks.
In this paper we propose a new augmentation technique, called patch augmentation, that, in our experiments, improves model accuracy and makes networks more robust to adversarial attacks. In brief, this data-independent approach creates new image data based on image/label pairs, where a patch from one of the two images in the pair is superimposed on to the other image, creating a new augmented sample. The new image's label is a linear combination of the image pair's corresponding labels. Initial experiments show a several percentage point increase in accuracy on CIFAR-10, from a baseline of approximately 81% to 89%. CIFAR-100 sees larger improvements still, from a baseline of 52% to 68% accuracy. Networks trained using patch augmentation are also more robust to adversarial attacks, which we demonstrate using the Fast Gradient Sign Method.