Neural network learns to create its own bizarre pictures of people
Eventually computers could be able to create photos based on description Computers can now create images of human faces based on what they think we look like, using artificial neural networks — and they've imagined us to be pretty horrifying. Artificial neural network is the technology that is responsible for the surreal images of Google’s DeepDream, which can ‘see’ objects and animals that aren’t really there. This artificial neural network is the same technology that is responsible for the surreal images of Google’s DeepDream, which can ‘see’ objects and animals that aren’t really there HOW TO TRAIN A NEURAL NETWORK Google trains an artificial neural network by showing it millions of training examples and gradually adjusting the network parameters until it gives the classifications the team want. The researchers hoped this would extract the concept of ‘smiling,’ and combine it with the concept of ‘man.’ It was a success; the neural network generated pictures of smiling men, all from its ‘imagination.’ The tests didn’t just focus on faces. The researchers hoped this would extract the concept of ‘smiling,’ and combine it with the concept of ‘man.’ It was a success; the neural network generated pictures of smiling men, all from its ‘imagination' The team asked the network to created different versions of the same scene, to prove that it wasn’t copying the training data.
Discover Related

Google Image results filled with AI images, users complain

Inside the ‘Black Box’ of a Neural Network

Digital artist Chris Rodley says artificial intelligence could spell death of the artist

Artificial Intelligence Is Killing the Uncanny Valley and Our Grasp on Reality

Google's Artificial Brain Is Pumping Out Trippy—And Pricey—Art

Facebook's New AI Can Paint, But Google's Knows How to Party
