# Data Augmentation Revisited: Rethinking the Distribution Gap between Clean and Augmented Data

Source: [Data Augmentation Revisited: Rethinking the Distribution Gap between Clean and Augmented Data](<https://www.sota2.com/research/paper/68fa66ad1e209131cea11a5e>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2025-10-25T21:55:21.281Z

Authors: Zhuoxun He, Lingxi Xie, Xin Chen, Ya Zhang, Yanfeng Wang, Qi Tian

Published: 2019-09-19T08:36:45.000Z

## Abstract

Data augmentation has been widely applied as an effective methodology to improve generalization in particular when training deep neural networks. Recently, researchers proposed a few intensive data augmentation techniques, which indeed improved accuracy, yet we notice that these methods augment data have also caused a considerable gap between clean and augmented data. In this paper, we revisit this problem from an analytical perspective, for which we estimate the upper-bound of expected risk using two terms, namely, empirical risk and generalization error, respectively. We develop an understanding of data augmentation as regularization, which highlights the major features. As a result, data augmentation significantly reduces the generalization error, but meanwhile leads to a slightly higher empirical risk. On the assumption that data augmentation helps models converge to a better region, the model can benefit from a lower empirical risk achieved by a simple method, i.e., using less-augmented data to refine the model trained on fully-augmented data. Our approach achieves consistent accuracy gain on a few standard image classification benchmarks, and the gain transfers to object detection.

## Recorded benchmark results

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

## Original sources and implementation

[Original paper](<https://arxiv.org/abs/1909.09148>)

[PDF](<https://arxiv.org/pdf/1909.09148.pdf>)
