Towards Distributed Coevolutionary GANs
This addresses training instability in GANs for researchers and practitioners in deep generative modeling, but it is incremental as it supplements existing methods rather than introducing a new paradigm.
The paper tackled the problem of training Generative Adversarial Networks (GANs), which are difficult to train due to issues like mode collapse and vanishing gradients, by investigating coevolution as a supplement to gradient-based methods, showing it is a promising framework for escaping these degenerate behaviors in experiments on a simple model.
Generative Adversarial Networks (GANs) have become one of the dominant methods for deep generative modeling. Despite their demonstrated success on multiple vision tasks, GANs are difficult to train and much research has been dedicated towards understanding and improving their gradient-based learning dynamics. Here, we investigate the use of coevolution, a class of black-box (gradient-free) co-optimization techniques and a powerful tool in evolutionary computing, as a supplement to gradient-based GAN training techniques. Experiments on a simple model that exhibits several of the GAN gradient-based dynamics (e.g., mode collapse, oscillatory behavior, and vanishing gradients) show that coevolution is a promising framework for escaping degenerate GAN training behaviors.