Using mixup as regularization and tuning hyper-parameters for ResNets

TL;DR

This paper enhances ResNet50 by integrating mixup data augmentation for regularization and hyper-parameter tuning, aiming to improve image classification performance, especially with limited data, while emphasizing training efficiency.

Contribution

It introduces the use of mixup augmentation specifically for ResNets and provides insights into hyper-parameter tuning to boost performance.

Findings

Improved accuracy of ResNet50 with mixup augmentation

Enhanced training stability and efficiency

Better generalization on limited data

Abstract

While novel computer vision architectures are gaining traction, the impact of model architectures is often related to changes or exploring in training methods. Identity mapping-based architectures ResNets and DenseNets have promised path-breaking results in the image classification task and are go-to methods for even now if the data given is fairly limited. Considering the ease of training with limited resources this work revisits the ResNets and improves the ResNet50 \cite{resnets} by using mixup data-augmentation as regularization and tuning the hyper-parameters.