Unbiased Auxiliary Classifier GANs with MINE
This addresses instability and bias issues in conditional generative models for image generation, but it is incremental as it builds on prior methods like TAC-GAN.
The paper tackled the biased distribution problem in Auxiliary Classifier GANs (AC-GANs) by proposing UAC-GAN, which uses Mutual Information Neural Estimator (MINE) to estimate mutual information between generated data and labels, achieving better performance than AC-GAN and TAC-GAN on datasets like MNIST and CIFAR10.
Auxiliary Classifier GANs (AC-GANs) are widely used conditional generative models and are capable of generating high-quality images. Previous work has pointed out that AC-GAN learns a biased distribution. To remedy this, Twin Auxiliary Classifier GAN (TAC-GAN) introduces a twin classifier to the min-max game. However, it has been reported that using a twin auxiliary classifier may cause instability in training. To this end, we propose an Unbiased Auxiliary GANs (UAC-GAN) that utilizes the Mutual Information Neural Estimator (MINE) to estimate the mutual information between the generated data distribution and labels. To further improve the performance, we also propose a novel projection-based statistics network architecture for MINE. Experimental results on three datasets, including Mixture of Gaussian (MoG), MNIST and CIFAR10 datasets, show that our UAC-GAN performs better than AC-GAN and TAC-GAN. Code can be found on the project website.