English

Synthesizing Informative Training Samples with GAN

Machine Learning 2022-12-22 v2 Computer Vision and Pattern Recognition

Abstract

Remarkable progress has been achieved in synthesizing photo-realistic images with generative adversarial networks (GANs). Recently, GANs are utilized as the training sample generator when obtaining or storing real training data is expensive even infeasible. However, traditional GANs generated images are not as informative as the real training samples when being used to train deep neural networks. In this paper, we propose a novel method to synthesize Informative Training samples with GAN (IT-GAN). Specifically, we freeze a pre-trained GAN model and learn the informative latent vectors that correspond to informative training samples. The synthesized images are required to preserve information for training deep neural networks rather than visual reality or fidelity. Experiments verify that the deep neural networks can learn faster and achieve better performance when being trained with our IT-GAN generated images. We also show that our method is a promising solution to dataset condensation problem.

Keywords

Cite

@article{arxiv.2204.07513,
  title  = {Synthesizing Informative Training Samples with GAN},
  author = {Bo Zhao and Hakan Bilen},
  journal= {arXiv preprint arXiv:2204.07513},
  year   = {2022}
}

Comments

NeurIPS 2022 Workshop on Synthetic Data for Empowering ML Research, https://openreview.net/forum?id=frAv0jtUMfS

R2 v1 2026-06-24T10:49:17.958Z