Deceive D: Adaptive Pseudo Augmentation for GAN Training with Limited Data
arXiv:2111.06849
Abstract
Generative adversarial networks (GANs) typically require ample data for training in order to synthesize high-fidelity images. Recent studies have shown that training GANs with limited data remains formidable due to discriminator overfitting, the underlying cause that impedes the generator's convergence. This paper introduces a novel strategy called Adaptive Pseudo Augmentation (APA) to encourage healthy competition between the generator and the discriminator. As an alternative method to existing approaches that rely on standard data augmentations or model regularization, APA alleviates overfitting by employing the generator itself to augment the real data distribution with generated images, which deceives the discriminator adaptively. Extensive experiments demonstrate the effectiveness of APA in improving synthesis quality in the low-data regime. We provide a theoretical analysis to examine the convergence and rationality of our new training strategy. APA is simple and effective. It can be added seamlessly to powerful contemporary GANs, such as StyleGAN2, with negligible computational cost.
arXiv commentNeurIPS 2021. Code: https://github.com/EndlessSora/DeceiveD Project page: https://www.mmlab-ntu.com/project/apa/index.html
References in corpus (11)
- Conditional Generative Adversarial Nets
- Unsupervised Representation Learning with Deep Convolutional Generative Adversarial Networks
- GANs Trained by a Two Time-Scale Update Rule Converge to a Local Nash Equilibrium
- Large Scale GAN Training for High Fidelity Natural Image Synthesis
- Spectral Normalization for Generative Adversarial Networks
- Which Training Methods for GANs do actually Converge?
- Amortised MAP Inference for Image Super-resolution
- Image Augmentations for GAN Training
- Learning to Predict Layout-to-image Conditional Convolutions for Semantic Image Synthesis
- Real or Not Real, that is the Question
- Unsupervised multi-modal Styled Content Generation