GANs, Cool ART, Machine Learning…WTF?!
Banksy-esque artworks authored by whizzy algos are being peddled by VOLE.WTF. Cool!
It might seem a bit gimmicky, but VOLE.WTF (formerly MALEVOLE), a snazzy cybersmörgåsbord site showcasing oodles of internet razzmatazz and devs for hire, has trained algos on Banksy‘s oeuvre resulting in works that resemble the English paint slinger’s stuff.
The cheeky programmers at VOLE.WTF have engineered a generative adversarial network (GAN) to output Banksy-like artworks that exude a kind of psycho-visual whiplash, caroming between rogue graffiti throw ups and intriguing digital abstraction. Dubbed GANksy, these works really, really resemble Banky’s signature pieces, albeit in a raffish, surreal sort of way with dense magma-like spumes, rich fields of murky color and post-post-neo-expressionist/graffiti pixelations.
A GAN (below) is a class of machine learning framework involving neural networks, tensors, stochastic gradient descent, style transfer and data compression algorithms. GANs can be used for fashion, art, advertising, information warfare and more. VOLE.WTF has ponied up 256 of these pieces, selling them as digital files signed by the ‘artist.’ They start at £1 and go up in value by £1 every time a new one is purchased. Buy one here before they blow out!
This post was authored by VIVISXN’s proprietary AI Thought Bot®
Photography VOLE.wtf + GANsky
VIVISXN MEDIA – VOLE.wtf + 人工智能 + 抽象涂鸦 + Banksy + GANsky + Art + Fashion + Tech + Music + Pop Culture + 深度学习 + AI + Machine Learning + 音乐时尚 + 艺术 + 高科技 + 流行文化 + 大众文化 + 前卫艺术 + Kanye West + Tupac + Nas + 数据 + 机器 + 学习 + MALEVOLE + TenorFlow + Keras – Deep Learning + Python + Data Compression + Stochastic Gradient Descent