Coevolution of Generative Adversarial Networks

Abstract

Generative adversarial networks (GAN) present state-of-the-art results in the generation of samples following the distribution of the input dataset. However, GANs are difficult to train, and several aspects of the model should be previously designed by hand. Neuroevolution is a well-known technique used to provide the automatic design of network architectures which was recently expanded to deep neural networks. COEGAN is a model that uses neuroevolution and coevolution in the GAN training algorithm to provide a more stable training method and the automatic design of neural network architectures. COEGAN makes use of the adversarial aspect of the GAN components to implement coevolutionary strategies in the training algorithm. Our proposal was evaluated in the Fashion-MNIST and MNIST dataset. We compare our results with a baseline based on DCGAN and also with results from a random search algorithm. We show that our method is able to discover efficient architectures in the Fashion-MNIST and MNIST datasets. The results also suggest that COEGAN can be used as a training algorithm for GANs to avoid common issues, such as the mode collapse problem.

Exploring the evolution of GANs through quality diversity

“…We describe in this section our proposal to use a Quality Diversity (QD) algorithm to evolve GANs. This new model is based on the original COEGAN proposal [4,5], adapted to be guided by a different evolutionary algorithm. Thus, first we introduce the fundamentals of the COEGAN model.…”

“…Table 1 lists the parameters used in our experiments. These parameters were selected based on preliminary experiments and the experiments presented in [4,5]. The number of generations used in 1 Code available at https://github.com/vfcosta/qd-coegan.…”

“…This metric is capable of quantifying the quality and diversity of the generative model. The use of evolutionary algorithms to train and evolve GANs was recently proposed [1,4,5,8,27,28]. The solutions present a diverse set of strategies to not only overcome common GAN problems but also to provide better quality on the produced samples.…”

“…In COEGAN, the weights of parents are transferred to the children whenever it is possible [4,5]. This mechanism of transference ensures that the information achieved on training in previous generations is kept during the whole evolution process.…”

“…Coevolutionary Generative Adversarial Networks (COEGAN) [4,5] is a solution inspired on NEAT [25] and DeepNEAT [18] that applies a coevolution model to evolve GANs. Experimental results show that the model provides a more reliable GAN training when compared to regular GANs in equivalent scenarios.…”