It is called Docent, has already been dubbed the “Spotify of art,” and will be launched to the public during the week of Art Basel, June 15-18. It is an app that allows collectors and art lovers to find little-known artists who are nevertheless able to satisfy their taste. An app, in short, for people to discover new contemporary art. The motto: Discover Art Beyond Your Echo Chamber.
Docent was born in France from the passion for art and the desire to make it accessible to everyone of its two founders, Helene Nguyen-Ban, a passionate collector and former gallery owner, and Mathieu Rosenbaum, a scientist and chair in applied mathematics at the École Polytechnique in Paris. Nguyen-Ban and Rosenbaum sensed the need for a new way of discovering and collecting contemporary art, so they put together a team of engineers, machine learning experts, data scientists and connoisseurs in the art market (a team of more than 20 professionals scattered between Paris, London, New York and Hong Kong) and, in the midst of the Covid pandemic, created Docent, with the idea of helping collectors face the challenges of the market: finding new artists, knowing what to buy and where to buy.
“When I was a total neophyte and I was obsessed with art,” Nguyen-Ban said in an interview, “I thought it would be wonderful to have a tool that could analyze everything that exists and come up with very specific suggestions in a safe and creative environment. That’s how the idea came about. However, I thought something like that was not possible because the art world was not sufficiently digitized. I changed my mind after the first lockdown-I thought that was the perfect time to present the product.”
Docent’s technology uses data and human experience to create personalized suggestions unique to each user. By retrieving the most reliable and accurate information about artists, Docent’s team has trained its algorithms under the supervision of a team of art experts, both internally and in collaboration with galleries and institutions, to ensure that Docent’s recommendations are relevant and engaging. “We recognize that a single work of art,” say the two founders, “is a node in a complex web of signifiers and attributes. Docent’s algorithms form associations between artworks based on multiple perspectives and attempt to map the social, cultural and aesthetic complexities that exist in and around an artwork.”
“The idea we had from the beginning,” Rosembaum adds, “is that the collector shows us what he likes and on that basis we make some suggestions. In the beginning it was all very naïve because we were trying to achieve this with very standardized algorithms that I used in the industry, and the suggestions that emerged were very naïve and uninteresting. For example, if a user indicated that they liked something blue, we would show something about blue. Basically we were suggesting copies of what the collector already had. So we developed what we call the ’supervised approach’: the Docent team worked on thousands of artworks, manually annotated the variables associated with these artworks (about 40 or 50 variables, some very basic like the main color, light, and so on, and some more complicated, like the emotional response the artist wanted to give), and based on this database we wrote an algorithm that is able to automatically annotate the artworks. In this way we turned objective variables--into the collector’s taste. However, it is not enough, because if you discuss with a collector there is much more than the visual aspect of a work of art: you may like a work because you like the artist, his environment, the reasons why he made a work. These are what we call ’contextual variables,’ the ones that are able to make a connection between the artist you already like and the one you might like.”
Using Computer Vision Technology, Docent’s team has developed a wide range of potential criteria for evaluating the aesthetic patterns and visual characteristics of a work of art, specifically designed for the nuances of contemporary art. In combination with Natural Language Processing (NLP), Docent’s algorithm reads millions of text-based pieces of information to decrypt and capture the themes and topics with which an artist might engage in his or her practice. Docent thus seeks to enhance individual knowledge of contemporary art by providing an opportunity to tap into a collective human intelligence, made more visible through technological advances. “Our unique algorithm,” the founders emphasize, “evolves as you evolve, allowing you to discover, learn and appreciate art from multiple perspectives.”
Moreover, according to the two founders, another advantage of Docent lies in its ability to support galleries, institutions and museums. “At Docent, we understand the importance of nurturing and supporting the evolving art landscape, rather than trying to change it,” say Nguyen-Ban and Rosenbaum. “Our goal remains clear: to help bridge the gap between the vast online world and the transformative power of the offline art experience.”
And finally, Docent also wants to be an app. democratic toward artists, granting equal opportunity to all: “We are aware of the power that algorithms and social media can have in shaping and defining our knowledge base today and, at times, in amplifying existing biases. At Docent, we have a clear ethos to attempt to eliminate bias and embrace diversity. While societies continue to systematically and institutionally address existing pervasive cultural and human biases, Docent recognizes the potential for machine learning to address these issues by presenting more than what is simply popular or advertised: our algorithm assigns each artist an identical probability of being recommended, based on a user’s preferences.”
Although the app has yet to be officially launched, it is already available for download, both for Android and iPhone.
Pictured are some screenshots of Docent.
Coming soon Docent, the Spotify of art: an app that helps you discover new artists |
Warning: the translation into English of the original Italian article was created using automatic tools. We undertake to review all articles, but we do not guarantee the total absence of inaccuracies in the translation due to the program. You can find the original by clicking on the ITA button. If you find any mistake,please contact us.