DeepFuture - Art & AI
Deep Future – Art & AI is a collective of friends, artists and future engineers from Ecole des Mines de Nancy working on using the latest Deep Learning and Artificial Intelligence models to explore and deploy the creative potential of artificial intelligence.
A certain Picasso evoked at the time: “Computer are useless. They can only give answers “. From our point of view, and as the Obvious collective, a French precursor collective in generative art, has so well pointed out, we find this reductive and think that a new piece of the history of art can be written through computers.
Our project consists in creating artworks by means of computer science, thanks to the use of AI. To realize these works of art, we will use a state-of-the-art creative process of machine learning involving the use of Convolutional Neural Networks (CNN) and Generative Adversarial Network (GAN). The CNN makes it possible to identify and refine, from a very large number of artists’ works, the different characteristics of works that are painted in a given style.
GANs, or antagonistic neural networks, work in pairs like two tennis players trying to beat their respective opponents. They are composed of a network of image-generating neurons that tries to produce works that have the same characteristics as those present in its training data set, and a second one, called a discriminating network, that trains itself to distinguish the false images created by its generating rival from the original images.
In this ambitious project, our objectives are to create unpredictable and random collections combining our sensibility for some great social and scientific subjects (the indeterminability of the future, space, sciences, …) and our innovative technology.
The Web graph Art
The Web Graph Art is a collection of 60 NFTs created by the Deep Future – Art&AI collective. Each piece has been algorithmically & randomly generated and algorithmically signed. Each piece is truly unique as there are an infinite number of possibilities through the use of uniform randomness.
In this project, we create the art by using the concept of a 2-dimensional tree to create unique art pieces. In computer science, a k-d tree (short for k-dimensional tree) is a space partitioning data structure for organising points in a k-dimensional space. K-d trees are a useful data structure for several applications, such as searches involving a multidimensional search key (e.g., range searches and nearest neighbour searches) and point cloud creation.
Imaginary Mansions by AI
The « Imaginary Mansion » is a collection of 8 NFTs created by the Deep Future – Art&AI collective. Each piece has been randomly generated and algorithmically signed by an AI algorithm. Each artwork of art is truly unique and reflects an AI’s thinking about a human idea. Starting with a simple thought of a dream house, as a child would do, the AI algorithm tries to recreate it based on its experience, as a human does.
The AI algorithm is based on the use of Generative Adversarial Networks, better known as GAN. It is a machine learning model in which two neural networks compete with each other to become more accurate in their predictions. GANs typically run unsupervised and use a cooperative zero-sum game framework to learn.