THE COLLEGE HILL INDEPENDENT


Landscaping Fine Arts

The promise of Art.sy's Genome Project

by by Olivia-Jené Fagon

illustration by by Diane Zhou

 

 

In 2011, the launch of Google Art Project brought the public instant access to seven billion pixels worth of artistic masterpieces in high resolution, creating considerable anxiety over whether the site would permanently derail museum-going. It didn’t. Discussions concerning any attempt to wed digitized cultural material and online access have since moved past whether online art viewing can approximate the experience of seeing a work of art in person—it can’t. But the drive to find a way to bring the art world online is not slowing down, and it may have found a new way in.

After almost two years of invite-only beta-testing, the online fine art image database Art.sy (‘sy’ indicates its Syrian domain name) went public this past month with 17,000 individual art works and 3,000 artists, and an algorithm-backed reference system called the Art Genome Project that sorts and connects all of its images. The site functions along the lines of Big Data—the sector of technology that synthesizes the data produced by our actions online to then predict what we’ll buy, what we’ll like, and where we’ll make our next online move, the algorithmic engine behind sites like Netflix and Amazon.

Art.sy founder Carter Cleveland, a 24-year-old computer science engineer with an artistic bent, heads Art.sy’s cross-curricular team, comprised of art historians who workshop and evaluate the various genes, an engineering and technology department, and a set of gallery liaisons who maintain and develop the company’s commercial and institutional working relationships. Art.sy’s minimalist design makes for a well-crafted, easy to use interface that showcases the high-resolution quality of their images. The “For Sale” labels on the majority of the art makes it clear that Art.sy isn’t necessarily an art-for-all-people initiative like Etsy or 20 x 200, which both make a point of eliminating any commercial gallery presence. Rather, Art.sy partners with 75 galleries and 50 museums and institutions, enabling it to house works by lesser known contemporary artists alongside formative figures ranging from Jackson Pollock and Warhol to old masters like Rembrandt and Francisco de Goya—all within the network of its genome system.

Created in the image of Pandora, The Art Genome Project isolates a work of art’s characteristics in order to define and relate it to other works. Rather then using tags, Art.sy isolates different genes. The site currently has 800+ genes which include techniques (i.e. multiple exposure), concepts (i.e. Globalization), art historical movements (i.e. Impressionism), time periods (i.e. early 19th century), content (i.e. female subject), geographical regions (i.e. Chinese) as well as basic formal elements of work. The genes range from purely descriptive (color, subject-matter, medium) to more nuanced (“Contemporary Traces of Memory”). The Art Genome Project’s Tumblr described the site’s genes as “polyhierarchical,” indicating that works fall under several different genes and so have multiple and weighted relationships to one another.

The project uses computer software to identify basic visual aspects of works, but ultimately the genes are applied individually by team members. “We learned that the data matters much more than the math.” Daniel Doubrovkine, Art.sy’s lead engineer, told The New York Times. “How are you going to pick something that shows ‘warmth’ with a machine? We’re not.” Art.sy’s art historians refine the genome system by determining to what degree each work corresponds to a gene on a scale of one to 100. When you think about an Art.sy employee determining how high a piece of work should rate on the “Contemporary Traces of Memory” gene’s scale, it’s obvious that the site is expanding and refining the possibilities of algorithmic analytics.

//

 

The fact that Art.sy is trying to sell the work it features suggests that the algorithmic-function of its genome system isn’t simply about creating a road map for fine art but also about finding the right work that a user might consider acquiring. This expands possibilities, but when the promise of algorithmic technology fails the results are jarring and unmistakable. Think about the laughable customer recommendations of Netflix (“Because you enjoyed Season 4 of The Chappelle Show, you may also enjoy “Transformers: The Movie”) or the transparency of Facebook sidebar advertisements. Yet Pandora is able to deliver song after song that I enjoy based on my initial song or artist choice, so why wouldn’t a similar process work for fine arts? According to Douglas Nickel, professor of art history at Brown University, “The problem may be in searching itself.” Art.sy is offering a personalized journey through contemporary art, but “there just isn’t a guide book yet to what good, collectable, contemporary art is.”

The Art Genome Project’s Tumblr states specifically that the reference system is meant to “provide the structure for related art search... that opens up seemingly infinite pathways,” indicating that Art.sy is not promising the “right answer” to a search in the way that traditional search engines like Google’s strive to. Rather, Art.sy gives options.

 

//

 

“When you go into a museum you know that the educational information—the wall text, curatorial statements, docent tours—is coming from a place that’s impartial” explained Nickel, “but if you go into a commercial gallery and the gallerist approaches you and says ‘Look at John Doe’s work. He is going to be the next big thing, that’s educational too, but you know it’s motivated education. It’s not disinterested.” Art.sy is a blend of the two, making artwork available to the general public on the one hand while maintaining a commercial interest. That the art on Art.sy either is from artists who have gallery representation or from established art institutions also presupposes the fact that some artists are left out. In this sense, while the random searches of Art.sy gesture toward a democratizing of artists’ exposure, the site is not as potentially democratic as sites like Tumblr or Pinterest, both of which allow users to amass images without the network of museums, art-historians and the gallery. Art.sy comes from and is part of this network.

It is from such authorities that come general consensus on the art historical canon of ‘great art.’ However, by placing lesser known artists in the same visual company as mainstream, recognized works, even works that are considered masterpieces, the site is offering its users an interesting vetting process. I don’t get the sense that the site is assuming the average Art.sy user operates on the logic that if they can’t have a Rembrandt, for example, they’ll take a work that is ‘visually similar’ but assembling these works together seems to be a way of giving more contemporary works a sense of ‘value-by-association.’ A contemporary work found by way of a Rembrandt takes on some of the value of the masterpiece with which it is associated, while also locating them a midst a larger art historical context. Yet looking at the images that fall under the broad black and white photography gene, the images of works by Man Ray or Moriyama stand out to me, but for Nickels, when grouped all together they undergo a certain leveling of distinction. “That’s what this does,” Nickel said, gesturing toward the expanse of black and white images. “When they’re put like this they all kind of look the same.”

Art.sy’s automated guidance gives its user an overview, a lay of the fine art’s land, and a sense of direction. Moving through the site you have an idea of where you’ve been, how to move forward, and how to go back, a positioning constituted by individual art works or artists. Art.sy gives a visualized and mappable six degrees of separation between the seemingly most unrelated artworks so a 17th century Rembrandt portrait could be connected formally and conceptually to a sculpture made in 2009 of discarded telephone cables by Daniel Conagar. That often times the connections between works are as basic as the ‘use of repetition’ to something more substantial like works that deal with ‘identity politics’ causes Nickel to wonder, “When do these genes stop being meaningful?”

//

Academic pushback to the genome project centers around the potentially reductive nature of the genes—a slightly ironic accusation considering so much of art history operates through categorization: established classifications and specific terminology. The interplay between the user’s open-ended searching and the various combinations of genes each artwork can fall under has the ability to connect art works that would never share the same space in a gallery, in a museum, or in an art history textbook.

The danger with any system of classification is that it can very quickly breed essentialism. Combined with ethnic or social bias, certain artists historically have become trapped by academic labels like ‘Feminist’ or ‘Primitive.’ When asked whether the site’s genome system has the potential to unsettle the classifications and categories of traditional art history, Nickel countered, “If you want to be disruptive to art history, then randomize it.” To some extent, the seemingly accidental results that the site’s gene-filtered searches delivers does exactly that, but is it disruptive in a productive way?

Reflecting on the advent of Feminist art in the late 50s, Nickel describes the value of differentiation in art history. “You may get the most insight into what feminist art was in the fifties by looking at a Pollock, considered the exact opposite of feminist art, but no one would ever ‘categorize’ his work as Feminist.” Because the genomes work by identifying similarities or influences, they can’t (not yet) point to perhaps a more complex discourse surrounding works in a way that, according to Nickel, would be meaningful. I did however note that Grace Hartigan, an abstract expressionist painter and disciple of Pollock who was labeled a pioneer of feminist art though resented that her work was considered in relation to her gender, was included under Pollock’s “related artists” list.

//

On Art.sy, I searched for El Anatsui, a renowned Ghanaian sculptor, who is most well known for his metal fabric hangings. His work has been included in an exhibition of African tapestries at the Metropolitan Museum of Art and a show of contemporary African art at the Haywood Gallery in London entitled “Africa REMIX” and he recently had a solo retrospective at the National Museum of African Art in Washington, DC.

Art.sy doesn’t currently have any of his work but it offers a list of suggested artists and works, including American-born John Chamberlain’s scrap metal automobile-like sculptures, Louise Nevelson’s sculptural pieces created from dismembered furniture, and Bruce Conner’s collage sculptures. All four artists’ assemblage sculptural works are made from cast-off and found materials, playing on a certain raw materiality that the artists manipulate. El Anatsui is predominantly labeled an ‘African’ artist, and his work may not generally be considered or viewed in the company of these artists, while Art.sy genetics included him within the practice of assemblage.

But Anatsui’s work does relate to his Ghanaian heritage, and there must be a gene for that, too, though not yet. As another example, Chinese artist Cai Guo Ziang’s Art.sy profile lists Ai Wei Wei and Maya Lin as related artists, though the content of all three artists’ works do not share a direct correlation outside of the artists’ shared nationality. By finding ‘genetic’ links between artworks and artists that have been art historically relegated to distinct categories, the site can rethink these potentially limiting categories, showcasing works and artists without embedded value judgments and cultural bias. Though it also runs the risk of falling into the same trap of reductive groupings.

Ultimately if you judge Art.sy solely as a commercial enterprise intended to find the next acquisition for a collector, then the site may not deliver. Their desire, according to their blog, to “map the serendipitous” suggests the site is not necessarily promising users that, nor is that their aim. Even in the tenuousness of some of the gene’s relations to works, it offers a considered method of accidental but guided searches and encounters with artists and works you haven’t seen. It is in these serendipitous but mappable moments that Art.sy’s Genome Project proves truly innovative.

OLIVIA-JENE FAGON B’13 is unsearchable.