GAN is an unsupervised learning algorithm that pits a discriminator and generator against one another so that they iteratively compete to enhance the overall model's ability to model/replicate a given dataset. So far I have only seen GAN applied in code to very large datasets that have hundreds of thousands of observations.

Not restricted to image construction, have there been any studies on how GAN performs (for a density estimation-sampling task or other) when it is only fed a couple thousand or just hundreds of sample observations?

If not, and given what we already know about what it is capable of for large sample sizes, what can we theoretically expect in terms of the training and performance of GAN for small datasets?


It's true that GANs generally produce good results when trained on a massive amount of data. However, CycleGAN has been around for a while and they produce realistic images when trained only on a few images (the open-sourced datasets used by the authors have about 1000 images on an average). That said, this (very recent) paper talks about differentiable augmentation as a novel technique for training data-efficient GANs. Their results include experiments with a fraction of data in CIFAR10 and CIFAR100 datasets. This paper from MIT considers evolutionary GAN training as a means to reduce the amount of data used for training. It would also help to look specifically into the papers that use GANs for medical image synthesis or translation problems because data is already scarce in the medical domain, hence people have found some interesting techniques to get around it.

As for the non-image-generation applications of GANs, WaveGAN is a model for synthesizing audio, spaceGAN for geospatial modeling, etc. This Reddit discussion gives many more examples of GANs in non-image applications.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.