Authors: Ian Goodfellow, Jean Pouget-Abadie, Mehdi Mirza, Bing Xu, David Warde-Farley, Sherjil Ozair, Aaron Courville, Yoshua Bengio
Generative adversarial networks are a kind of artificial intelligence algorithm designed to solve the generative modelling problem. The goal of a generative model is to study a collection of training examples and learn the probability distribution that generated them. Generative Adversarial Networks (GANs) are then able to generate more examples from the estimated probability distribution. Generative models based on deep learning are common, but GANs are among the most successful generative models (especially in their ability to generate realistic high-resolution images). GANs have been successfully applied to a wide variety of tasks (mostly in research settings) but continue to present unique challenges and research opportunities because they are based on game theory, while most other approaches to generative modelling are based on optimization.
GANs are a kind of generative model based on game theory.
They have had great practical success in generating realistic data, especially images. It is currently still difficult to train them. For GANs to become a more reliable technology, it will be necessary to design models, costs, or training algorithms for which it is possible to find good Nash equilibria consistently and quickly.
A Style-Based Generator Architecture for Generative Adversarial Networks
Tero Karras, Samuli Laine, Timo Aila
We propose an alternative generator architecture for generative adversarial networks, borrowing from style transfer literature. The new architecture leads to an automatically learned, unsupervised separation of high-level attributes (e.g., pose and identity when trained on human faces) and stochastic variation in the generated images (e.g., freckles, hair), and it enables intuitive, scale-specific control of the synthesis. The new generator improves the state-of-the-art of traditional distribution quality metrics, leads to demonstrably better interpolation properties, and disentangles the latent factors of variation. To quantify interpolation quality and disentanglement, we propose two new, automated methods that apply to any generator architecture. Finally, we introduce a new, highly varied and high-quality dataset of human faces.
GAN 2.0: NVIDIA’s Hyperrealistic Face Generator
The NVIDIA paper proposes an alternative generator architecture for GAN that draws insights from style transfer techniques. The system can learn and separate different aspects of an image unsupervised; and enables intuitive, scale-specific control of the synthesis.