Towards Understanding Why Data Augmentation Improves Generalization

TL;DR

This paper proposes a unified theoretical framework explaining how data augmentation improves model generalization by promoting diverse feature learning and increasing training complexity through feature removal and mixing.

Contribution

It introduces a comprehensive theory that explains the mechanisms of data augmentation's benefits, unifying various techniques under a common understanding.

Findings

Data augmentation reduces reliance on single features.

Feature mixing increases training complexity and robustness.

Unified theory validated by experimental results.

Abstract

Data augmentation is a cornerstone technique in deep learning, widely used to improve model generalization. Traditional methods like random cropping and color jittering, as well as advanced techniques such as CutOut, Mixup, and CutMix, have achieved notable success across various domains. However, the mechanisms by which data augmentation improves generalization remain poorly understood, and existing theoretical analyses typically focus on individual techniques without a unified explanation. In this work, we present a unified theoretical framework that elucidates how data augmentation enhances generalization through two key effects: partial semantic feature removal and feature mixing. Partial semantic feature removal reduces the model's reliance on individual feature, promoting diverse feature learning and better generalization. Feature mixing, by scaling down original semantic features and introducing noise, increases training complexity, driving the model to develop more robust features. Advanced methods like CutMix integrate both effects, achieving complementary benefits. Our theoretical insights are further supported by experimental results, validating the effectiveness of this unified perspective.