DynGAN: A machine learning framework that detects collapsed samples in the generator by thresholding the outputs of the observable discriminator
Generative adversarial networks (GANs) are a popular tool for creating realistic data, but they often struggle with a problem called ...