[Firein1945[1945সালেআগুন Gustav Klimt claims three of the most controversial paintings. Commissioned for the University of Vienna in 1894, “The Faculty Paintings” – as they became known – was different from the previous work of the Austrian symbolist. As soon as they were presented, the critics were shocked by their dramatic departure from the aesthetics of the time. University professors immediately rejected them and withdrew from the Clement project. After that, the works found their way into other collections. During World War II, they were housed in a castle north of Vienna to keep them safe, but the castle burned down and the images probably went with it. All that remains today are some of the black-and-white photographs and writings of that time. Yet I look at them.
Well, not the paintings themselves. Franz Smola, a Climate expert and Emil Walner, a machine learning researcher, spent six months combining their skills to revive Climate’s lost work. It’s a laborious process, starting with those black-and-white photos and then incorporating artificial intelligence and score scores about the painter’s art, in an attempt to recreate what those lost images might look like. What Smola and Wallner are showing me is the result এবং and they’re amazed at the impressive technical images created by AI.
Let’s make one thing clear: no one is saying that this AI is bringing back the real work of Klimt. “It’s not a process of recreating the original colors, it’s a process of repainting the photographs,” Smolla quickly notes. “The medium of photography is already an abstraction from real work.” What machine learning is doing hints at something that has been believed to be lost for decades.
Smola and Wallner find it enjoyable, but not everyone supports AI filling in the blanks. The idea of machine learning to recreate lost or damaged works is as controversial as faculty painting. Industry conservator Ben Fino-Radin said, “My main concern about the ethical level of machine learning use in the context of conservation is, simply because of the sheer amount of ethical and moral issues.” Afflicted The field of machine learning. “
To be sure, the use of technology raises thorny questions to revive the work of human artistry. Even if there is a perfect AI that Klimt can determine which color or brushstroke to use, no algorithm can create authentic intent. This has been debated for centuries. In 1936, before Klimt’s paintings were destroyed, essayist Walter Benjamin argued against mechanical replication, even in photographs, stating that “even the most perfect reproduction of a work of art lacks an element: its presence in time and space, its uniqueness. ” This, Benjamin wrote The work of art in the age of mechanical reproduction, Whom he employed. “Ava“For many art lovers, the idea of reproducing that elusive component of a computer is absurd, if not absolutely impossible.
And yet, there is still much to learn from what AI can do. Faculty paintings played a key role in Klimt’s development as an artist, an important bridge between his more traditional earlier paintings and later, more original works. But what they looked like in full color remains a mystery. Smola and Welner were trying to solve that puzzle. Their project, Organized by Google Arts & Culture, Was not about perfect reproduction; It was about giving a hint of what was missing.
To do this, Walner developed and trained a three-part algorithm. First, the algorithm was given nearly a million images of art from the Google Arts and Culture database. It has helped it to understand objects, works of art and composition. Later, it was specifically schooled on Clemott’s paintings. “It tends to be biased towards its color and its motifs over time,” Walner explains. And finally, AI was given color clues on certain parts of the paintings. But there is no mention of color in the paintings, where did these formulas come from? Even Climato expert Smola was amazed at how detailed the writings of the time were. Since the paintings were considered so dirty and bizarre, critics tended to describe them in length, according to the artist’s choice of color, he said. “You could call it a tragedy of history,” said Simon Raine, the project’s program manager. “The paintings created a scandal and the rejection put us in a better position to recover them because there was a lot of documentation. And these kinds of data points, if fed into algorithms, make these images probably more accurate versions of what they were like at the time.“
The key to that accuracy lies in combining algorithms with Smaller skills. His research shows that Klimt’s work has strong patterns and continuity during this period. Studying the existing paintings before and after the painting of the faculty gives an indication of the color and motifs of the repetition of his work at that time. Even the surprises that Smolla and Walner faced have been confirmed by historical evidence. When Clemot first showed his paintings, critics noted his use of red, which was rare in the artist’s palette at the time. But Three eras of women, Painted right after the faculty painting, boldly using a red, a smola believes the same color that caused a stir when first seen in the faculty paintings. The writings of that time made a fuss about the tragic green sky in another faculty painting. Combining these texts with Smaller’s knowledge of the special palette of greens of Clement while feeding on the algorithm, made it one of the first amazing images outside of AI.