eGAN: Unsupervised approach to class imbalance using transfer learning

TL;DR

This paper introduces eGAN, an unsupervised transfer learning method using GANs to address class imbalance in image classification without synthetic data augmentation, achieving notable improvements in F1-score.

Contribution

It is the first to use GANs without synthetic data augmentation for class imbalance, leveraging transfer learning with pre-trained models to improve classification performance.

Findings

Achieved 0.69 F1-score on CIFAR-10 with 1:2500 imbalance

Eliminates epistemic uncertainty in predictions

Allows thresholding for specificity or sensitivity

Abstract

Class imbalance is an inherent problem in many machine learning classification tasks. This often leads to trained models that are unusable for any practical purpose. In this study we explore an unsupervised approach to address these imbalances by leveraging transfer learning from pre-trained image classification models to encoder-based Generative Adversarial Network (eGAN). To the best of our knowledge, this is the first work to tackle this problem using GAN without needing to augment with synthesized fake images. In the proposed approach we use the discriminator network to output a negative or positive score. We classify as minority, test samples with negative scores and as majority those with positive scores. Our approach eliminates epistemic uncertainty in model predictions, as the P(minority) + P(majority) need not sum up to 1. The impact of transfer learning and combinations of different pre-trained image classification models at the generator and discriminator is also explored. Best result of 0.69 F1-score was obtained on CIFAR-10 classification task with imbalance ratio of 1:2500. Our approach also provides a mechanism of thresholding the specificity or sensitivity of our machine learning system. Keywords: Class imbalance, Transfer Learning, GAN, nash equilibrium