This article is produced by NetEase intelligent studio (public number smartman163). Focus on AI and read the next big era!
[Netease Intelligence News Aug. 21] On the white wall of Christies Gallery in central London, there is a golden frame of a man in a black Puritan suit who looks hesitant - a painting that seems to be the work of an ancient master. But in its lower right corner there is an unexpected signature: a mathematical equation.
This is a painting by the French art organization Obvious called Edmond de Belamy, or, more accurately, an algorithm designed by Obvious.
The whole process is designed to minimize human involvement in the work. Obvious co-founder Gauthier Vignier (GauthierVernier) said. The company was co founded by two other 25 year old founders in April 2017. Since then, Obvious has created 11 works of art with the help of artificial intelligence by teaching the history of computer art and showing it how to make its own work.
The motto of the team is, art creation is not just a human exclusive product. Now, they are beginning to convince the art world that they are right. In February, Obvious sold his first work, LeComte de Belamy, to Paris collector Nicolas Laugero-Lasserre for 10,000 ($11.43 million). In October, they auctioned Edmond at Christies, New York, for the first time in the history of a large art auction in the hands of artificial intelligence.
Richard Lloyd, Christies international head of prints and reproductions, has arranged the auction, which he believes will spark discussion of the fundamental issues of art and creativity. Everyone has his own definition of a work of art, he said. I tend to think that the identity of the author is very important - its about the relationship with the viewer. But you can also say that art exists in the eyes of viewers. If people feel that it is full of feelings and inspiring, it is so. If it is rickety and quack, it is a duck.
Edmond de Belamy will be auctioned at Christies in New York in October and displayed at the London Gallery in July
Obvious created a whole series of Belamy works, including LeComte, LaComtesse, LeBaron and LaBaronne de Belamy - all of which were somewhat confusing, and the scrolling portraits looked like works from the 18th century. In order to complete these works, Obvious uses the GAN. GAN is an algorithm that was first created in 2014 by Ian Goodfellow, an American researcher on artificial intelligence. Goodfellow, now in San Francisco, introduced the algorithm in an influential paper at the University of Montreal. (The name of Obviouss first work, Belamy, is a rough translation of Goodfellows name, so named to honor him. First, Obviouss team needs to encode those networks to meet their own standards. Its like making a bicycle. If you miss a part, it wont work. Hugo Caselles-Dupre, a doctoral student in artificial intelligence at Obvious who does a lot of technical work, said.
The teams online art encyclopedia, WikiArt, collected 15,000 portraits spanning the 14th and 19th centuries and fed them into GAN algorithms. The GAN algorithm consists of two parts: generator and discriminator. The generator learned the rules of the portrait, for example, any character has two eyes and a nose. This process takes about two days. Then it will start creating new images based on these rules. At the same time, the discriminators job is to identify the image, guess which is from the data set real portrait, which is from the generator false portrait.
When the generator tries to deceive the discriminator, it will learn from every failure. When it successfully deceived the discriminator, the process is complete, and you have a new image. Its not a simple copy or patch-up of 15,000 existing images; Cass-Dupley says, Think of it as 1,5001. This is an original image. So its cool. They call their new art movement GAN doctrine.
Obvious isnt the only organization that does research on the art of artificial intelligence -- the Institute of Arts and Artificial Intelligence at Rutgers University in New Jersey has been experimenting with this since 2012. In 2017, researchers at the Institute said they had used artificial intelligence to create new styles of art that were different from any previously conceived style, raising questions about the limits of human imagination. In March, Robbie Barrat, a college student from West Virginia, was praised for publishing his own naked images using artificial intelligence on Twitter.
But Obvious is one of the first organizations to make substantive works of art. Pierre Fautrel, a third member of the team, says this makes them more convinced that what they are doing at the festival is worthwhile than the AI art they display on blogs or websites. Conservative art is more accepting of physical work, he said. They understand that we are not trying to deceive the art community. We want to become modern artists.
Although Obviouss work is uncertain whether it can be called art (Lloyd agrees), the team was surprised by the interest in it. One thing about our art is that no one will care nothing about it. Futre M said, people like it or hate it. But no one said they didnt care.
Like philosophical questions, the growing debate about artificial intelligence often raises concerns and doubts about unemployment. Some argue that artificial intelligence can create thousands of new, unique images by pressing a button, threatening the scarcity principle that gives value to artworks.
But Castle Dupley insists that they do not see artificial intelligence as a human alternative to mass-produced work. He likens todays artificial intelligence experiments to the emergence of photography in the mid-19th century, when miniature portrait artists lost their jobs. At that time people said that photography is not real art, and that photographers are like machines. He said, now, we have always believed that photography has become a real branch of art.
For Lloyd, the Belamy series of paintings, close to artifacts, gave her the first glimpse of a work that would one day become a common experience. Were all going to experience this cultural shock over and over again, in which we think were communicating or interacting with people, and then suddenly realize that were communicating with robots. Its aLight bulb momentthat were going to encounter all the time, whether on the Internet, on the phone, or in public space, Lloyd said. Obviouss work is just the beginning. (From: Time Compiler: Netease Intelligent Participation: Lebanon) Focus on Netease Intelligent Public Number (smartman 163), for you to interpret the AI field of major corporate events, new ideas and new applications. Source: NetEase intelligent editor: Ding Guang Sheng _NT1941
For Lloyd, the Belamy series of paintings, close to artifacts, gave her the first glimpse of a work that would one day become a common experience. Were all going to experience this cultural shock over and over again, in which we think were communicating or interacting with people, and then suddenly realize that were communicating with robots. Its aLight bulb momentthat were going to encounter all the time, whether on the Internet, on the phone, or in public space, Lloyd said. Obviouss work is just the beginning.
Focus on NetEase Smart Public Number (smartman 163) to explain the big AI events, new ideas and new applications.