RSG: A Simple but Effective Module for Learning Imbalanced Datasets

TLDR

Imbalanced datasets are common in practice and pose a significant challenge for training deep neural networks to generalize well on rare classes. This work proposes the Rare‑Class Sample Generator (RSG) to generate additional samples for infrequent classes during training, aiming to improve model performance on imbalanced data. RSG is a lightweight module that can be seamlessly integrated into any convolutional neural network, works with various loss functions, and is applied only during training, imposing no overhead at test time. Experiments show that RSG achieves competitive results on Imbalanced CIFAR and sets new state‑of‑the‑art performance on Places‑LT, ImageNet‑LT, and iNaturalist 2018. Source code is available at https://github.com/Jianf-Wang/RSG.

Abstract

Imbalanced datasets widely exist in practice and are a great challenge for training deep neural models with a good generalization on infrequent classes. In this work, we propose a new rare-class sample generator (RSG) to solve this problem. RSG aims to generate some new samples for rare classes during training, and it has in particular the following advantages: (1) it is convenient to use and highly versatile, because it can be easily integrated into any kind of convolutional neural network, and it works well when combined with different loss functions, and (2) it is only used during the training phase, and therefore, no additional burden is imposed on deep neural networks during the testing phase. In extensive experimental evaluations, we verify the effectiveness of RSG. Furthermore, by leveraging RSG, we obtain competitive results on Imbalanced CIFAR and new state-of-the-art results on Places-LT, ImageNet-LT, and iNaturalist 2018. The source code is available at https://github.com/Jianf-Wang/RSG.