Researchers at Facebook's Artificial Intelligence (AI) lab in Paris used algorithms called generative adversarial networks (GANs) to produce fashion items including handbags, jumpers, and T-shirts.
The algorithms forced two neural networks to compete against each other, one generating ideas and the other judging them; over millions of iterations, each network masters its skills until something is created that satisfies both the AIs and humans.
The Facebook team trained three GANs on about 4,000 images of existing fashion items created by humans. Two of these GANs were off-the-shelf programs, and the third, called StyleGAN, was constrained to the shapes it created so the final items would actually be wearable.
A total of 800 final products were shown to people for review, almost two-thirds of which were thought to have been created by humans.
From New Scientist
View Full Article - May Require Paid Subscription
Abstracts Copyright © 2018 Information Inc., Bethesda, Maryland, USA
No entries found