Optimal Layer Selection for Latent Data Augmentation

TL;DR

This paper investigates optimal layers for applying data augmentation within neural networks and introduces AdaLASE, an adaptive method that automatically selects layers for augmentation, improving classification accuracy.

Contribution

The study systematically analyzes layer selection for feature augmentation and proposes AdaLASE, an automatic, gradient-based method for optimal layer choice during training.

Findings

AdaLASE effectively adjusts augmentation layers during training.

Applying DA to specific layers improves test accuracy.

AdaLASE outperforms fixed-layer augmentation methods.

Abstract

While data augmentation (DA) is generally applied to input data, several studies have reported that applying DA to hidden layers in neural networks, i.e., feature augmentation, can improve performance. However, in previous studies, the layers to which DA is applied have not been carefully considered, often being applied randomly and uniformly or only to a specific layer, leaving room for arbitrariness. Thus, in this study, we investigated the trends of suitable layers for applying DA in various experimental configurations, e.g., training from scratch, transfer learning, various dataset settings, and different models. In addition, to adjust the suitable layers for DA automatically, we propose the adaptive layer selection (AdaLASE) method, which updates the ratio to perform DA for each layer based on the gradient descent method during training. The experimental results obtained on several image classification datasets indicate that the proposed AdaLASE method altered the ratio as expected and achieved high overall test accuracy.