On the Generalization Effects of Linear Transformations in Data Augmentation

By Sen Wu, Hongyang Zhang, Gregory Valiant, Christopher R'e. 2020

93 citations recorded by the source corpus.

Data augmentation is a powerful technique to improve performance in applications such as image and text classification tasks. Yet, there is little rigorous understanding of why and how various augmentations work. In this work, we consider a family of linear transformations and study their effects on the ridge estimator in an over-parametrized linear regression setting. First, we show that transformations which preserve the labels of the data can improve estimation by enlarging the span of the training data. Second, we show that transformations which mix data can improve estimation by playing a regularization effect. Finally, we validate our theoretical insights on MNIST. Based on the insights, we propose an augmentation scheme that searches over the space of transformations by how uncertain the model is about the transformed data. We validate our proposed scheme on image and text datasets. For example, our method outperforms RandAugment by 1.24% on CIFAR-100 using Wide-ResNet-28-10. Furthermore, we achieve comparable accuracy to the SoTA Adversarial AutoAugment on CIFAR datasets.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

1. Jonghyun Choi
2. Jung-Woo Ha
3. Jane Lee
4. Heesu Kim
5. Wanxiang Che
6. Bohan Li
7. Y. Yoo
8. Edgar Dobriban

In Computer Vision and Pattern Recognition

1. Jonghyun Choi
2. Jung-Woo Ha
3. Jane Lee
4. Heesu Kim
5. Wanxiang Che
6. Bohan Li
7. Y. Yoo
8. Edgar Dobriban

Citations in the corpus

• Rainbow Memory: Continual Learning with a Memory of Diverse Samples (2021, methodology): “In particular, label mixing-based data augmentation, such as Cut-Mix [49], projects the input samples into a more complex dimension by mixing the image-label of multiple data samples randomly and has reported notable successes in various recognition tasks [47, 26].”
• Data Augmentation Approaches in Natural Language Processing: A Survey (2021, background): “The few related works are of DA in the image domain, considering data augmentation as encoding a priori knowledge about data or task invariance [140], variance reduction [141] or regularization methods [142].”
• A Group-Theoretic Framework for Data Augmentation (2019, background): “For example, a line of work proposes to add random or adversarial noise when training [11, 50, 25, 72, 8, 48, 71, 36] and argue that this can lead to a form of regularization [61, 33, 34, 17, 69].”