Artificial intelligence has exploded on the art scene in recent years, with everyone from artists to technical giants experimenting with the new tools that technology provides. While the generative adversarial networks (GANs) running like Google Google's BigGAN can create spectacularly weird images, they require a great deal of human interaction and guidance. Not so with the AICAN system developed by Professor Ahmed Elgammal and his team at Rutgers University's AI & Art Lab. It is an almost autonomous system trained on the 500-year-old host of Western artistic aesthetics that produces its own interpretations of these classic styles. And now it's hosting its first solo gallery show in NYC.
AICAN stands for "Creative Intelligent Adversarial Network Artificial Intelligence" while simultaneously using the same adversarial network architecture as GAN, they engage them in various ways. Adversarial networks work with two sets of nodes: one set generates images based on the visual training data set it was provided while the second set judges how close the generated image is to the actual images from the training data. AICAN strives for different goals. "At one end it tries to learn the aesthetics of existing artworks," wrote Elgammal in an October FastCo article. "On the other hand it will be punished if, when creating your own work, it emulates too close to an established style." That is, AICAN tries to create unique ̵
And, unlike GANs, AICAN is not trained on a certain set of visuals – say chihuahuas, blueberry muffins or American cubists from the 20th century. Instead, the AICAN contains the aesthetics of Western art history as it crawls through databases, absorbs examples of everything – landscapes, portraits, abstractions, but without focus on specific genres or subjects. If the piece was made in western style between the 15th and 20th century, AICAN will eventually analyze it. So far, the system has found more than 100,000 examples. Interestingly, this learning method is a departure from the laboratory's previous research on teaching AI to classify various historical art movements.
Elgammal notes that this training style more closely mimics the method used by human artists. "An artist has the ability to relate to existing art and … innovate. A great artist is one who truly melts in art history, blends on what happened earlier in art but generates his own artistic style," he told Engadget. "It's really what we were trying to do with AICAN – how can we look at art history and digest older art movements, learn from these aesthetics but create things that aren't in those [training] files." It can also name the art it creates with titles on works it has already learned.
In order to regulate the unique properties of the generated works, Elgammal's team must first quantify "uniqueness". The team cited "the most common definition of creativity, which emphasizes the originality of the product, along with its lasting influence", wrote Elgammal in an article in 2015. The team then "showed that the problem of quantification of creativity could be reduced to a variation of network centrality problems", The same class of algorithms that Google uses to show you the most relevant results for your search. Tested the quantification system on more than 1700 paintings, AICAN generally chose what is widely regarded masterpieces: ratings Edvard Munch's Scream and Picasso's ladies of Avignon are much higher in creativity than their peer works, for example, but boasted Da Vincent's Mona Lisa
Until then, you can check out AICAN's first solo gallery show, "Faceless Portraits Transcending Time," at HG Contemporary in New York City. This show will feature two series of images – a surreal, the other abstract – g enerated from the Renaissance your works.
"For the abstract portraits n, I chose images that were abstracted from the facial functions than probably enough in known figures. I used titles as portraits of a king and portrait of a queen to reflect generic conventions, "Elgammal wrote in a new post." For the surrealistic collection I chose images that intrigue the perception and ask questions about the subject, knowing that inspiration and aesthetics all come only from portraits and photographs of people, as well as skulls, nothing else. "
The performance runs February 13 to March 5, 2019.