English

SYNAuG: Exploiting Synthetic Data for Data Imbalance Problems

Computer Vision and Pattern Recognition 2024-04-26 v3 Machine Learning

Abstract

Data imbalance in training data often leads to biased predictions from trained models, which in turn causes ethical and social issues. A straightforward solution is to carefully curate training data, but given the enormous scale of modern neural networks, this is prohibitively labor-intensive and thus impractical. Inspired by recent developments in generative models, this paper explores the potential of synthetic data to address the data imbalance problem. To be specific, our method, dubbed SYNAuG, leverages synthetic data to equalize the unbalanced distribution of training data. Our experiments demonstrate that, although a domain gap between real and synthetic data exists, training with SYNAuG followed by fine-tuning with a few real samples allows to achieve impressive performance on diverse tasks with different data imbalance issues, surpassing existing task-specific methods for the same purpose.

Keywords

Cite

@article{arxiv.2308.00994,
  title  = {SYNAuG: Exploiting Synthetic Data for Data Imbalance Problems},
  author = {Moon Ye-Bin and Nam Hyeon-Woo and Wonseok Choi and Nayeong Kim and Suha Kwak and Tae-Hyun Oh},
  journal= {arXiv preprint arXiv:2308.00994},
  year   = {2024}
}

Comments

The paper is under consideration at Pattern Recognition Letters

R2 v1 2026-06-28T11:46:13.096Z