Generation 550
----
Artificial Sempervivum grown from 256-dimension z-space Generative Adversarial Network
----
The [GAN](https://arxiv.org/abs/1406.2661) is trained for 12,300 generations on the paintings of Le Besque. Using image processing neural networks inspired by biological methods, it produces pictures from a 256-dimensional latent-space to try to match the underlying probability distribution across all of the original paintings.