08 Feb Evolution of Generative Adversarial Neural Network
Since their introduction, machine learning algorithms have been applied to different problems in order to solve a wide range of problems.
In this paper, the authors of the Heriot-Watt University in collaboration with Google DeepMind, describe the evolution of a particular family of neural network known as Generative Adversarial Neural Network (GANN).
In particular, GANN allows to generate new unseen images showing landscape or face that do not exist in the reality. The results obtained with this technique are unbelievable, it’s almost impossible to discriminate between real and artificial generated images.
In Humanativa, we love to push forward the limit of our machine learning algorithms. We use GANN in order to improve the quality of our algorithms.
Info su: https://arxiv.org/pdf/1809.11096.pdf