A 19-year-old developed the code for the AI ​​portrait that sold for $ 432,000 at Christie's.

adminOctober 26, 2018

A woman looks at a work of art created by an algorithm of the French collective OBVIOUS that produces art using artificial intelligence, entitled "Portrait of Edmond de Belamy", in Christies in New York. (TIMOTHY A. CLARY / AFP / Getty Images) (Timothy A. Clary / AFP / Getty Images)

Robbie Barrat was 17 years old and bored in West Virginia when he started experimenting with artificial intelligence and art.

First, he coached a computer to write original rap songs from feeding it with 6,000 Kanye West letters. Then, he taught him how to make landscape paintings and nude portraits, feeding him with thousands of images taken from the Internet. He uploaded the code to GitHub, the shared code platform, so that others can download it and learn from it.

And many did, including the French art collective known as Obvious, whose AI portrait sold for $ 432,500 at Christie's on Thursday at an internationally held art auction. That's almost 45 times higher than Christie's original estimated sale price of $ 10,000, the auction house said in an article on its website.

"I really expected people to use [the code] As components for your own project. But I never thought that someone would sell it, just because it's not high quality work, "he told The Washington Post Barrat, who is now 19 years old and works in an IA research laboratory at Stanford University." It was a project I did in my free time when I was 17 years old.

Obviously, made up of three 25-year-old students, he made the portrait of AI using an existing algorithm and the Barrat code, among other things, as the group acknowledged in a detailed article in The Verge earlier this week.

Titled "Portrait of Edmond Belamy," the work of art was the subject of great attention this week for its unconventional nature in a major auction house such as Christie's, which was lined up in the same room as Warhol's work and Lichtenstein and announced it for its novelty. . The portrait emits a quality known in aesthetics as the strange valley: that strange sensation that one feels when one looks at something that looks like a human, but is not quite there. The man in the portrait does not have a nose. It has a spot on the mouth and its eyes resemble those of Frosty the Snowman. It seems, as indicated by your signature below on the right, to be the work of an algorithm.

But in spite of its apparent peculiarity, the relatively small artistic community of AI has described it as totally conventional, delayed by the advances of other artists in the art world of AI. Mario Klingemann, an AI artist based in Germany whom Obvious has cited as inspiration, told The Post he was surprised that it sold for hundreds of thousands of dollars. He said he believed for a second that "maybe this is just a practical joke among the oligarchs."

"It's a horrible art from an aesthetic point of view," he said. "You have to put some work to call it art. [The Obvious portrait] It is something that everyone can do. You can clone this [code] from GitHub, start your computer and start doing it. I do not know what art is about. You have to put your own handwriting, make your own brand with these tools. That requires some learning, work and finding something different to say. "

Obvious has never denied that his work has been based on innovations made by others. In fact, the name of the painting itself is a reference to the inventor of the algorithm that has made this whole field of AI art possible, a Google AI researcher scientist named Ian Goodfellow. "Belamy", the name of the fictional man in the portrait, is a play in "Bel ami", which means "good friend" in French.

Obvious could not be contacted immediately for comment on Thursday, but in an interview with Verge, Obvious's co-founder, Hugo Caselles-Dupré, confirmed that elements were borrowed from Barrat, who said he began to correspond on GitHub with the group. in October 2017. "If" we are only talking about the code, then there is not a large percentage that has been modified, "said Caselles-Dupré." But if you talk about working on the computer, making it work, there is a lot of effort there " .

Using Goodfellow's algorithm, making machines do AI art is like trying to present a false identification to a bartender at 10 p.m. a Saturday: if it's good enough and the waiter is busy enough, I could let it go.

It is the same with computer networks: a network of computers basically tries to deceive a second computer that the image it generated can happen as real art. It's like the Turing test, except that both participants are machines.

To train the first computer network to produce a portrait, for example, you have to feed hundreds or thousands of portrait images to learn what a good portrait is like. Then, the second computer, called a discriminator, plays the referee and decides if he passes the odor test. If it does not, as if a doorman confiscates a fake ID, the first computer has to "go back to the drawing board" and must work harder to produce compelling art, Klingemann said.

Once it is successful, the result is something usually strange and abstract, and perhaps even imaginative.

"[The computers] learn from scratch, "said Klingemann. "Initially, both parties do not know what criteria to look for, it is only with time, in this feedback loop, that they improve their abilities to detect differences, and over time, it becomes so good that even a human can no longer distinguish it from false. or real ".

Obvious told Christie that he fed the system with 15,000 painted portraits between the 14th and 20th centuries as part of his "training" for the computer.

In a statement after the sale of "Edmond Belamy," Obvious tilted his hats to Goodfellow and Barrat.

"We would like to thank the AI ​​community, especially for those who have been pioneers in the use of this new technology, including Ian Goodfellow, the creator of the GAN algorithm, who inspired the name of the Belamy Family series, and the artist Robbie Barrat , which has been a great influence for us. "It is an exciting time and our hope is that the focus on this sale will bring forward the incredible work that our predecessors and colleagues have been producing."

Source link