A Multi-objective architecture search for generative adversarial networks

A Multi-objective architecture search for generative adversarial networks

연구 분야: Artificial Intelligence

학회: GECCO '20: Proceedings of the 2020 Genetic and Evolutionary Computation Conference Companion

This study aimed to search for better network architectures and their hyper-parameters for generative adversarial networks (GANs). To this end, we employ an evolutionary algorithm to search for network architectures and their hyper-parameters. We applied our method to several datasets and demonstrate that our method can discover promising network architectures together with appropriate hyper-parameters.