Left, Construction takes place next to the Changi Airport control tower for Project Jewel in Singapore, August 17, 2016. REUTERS/Edgar Su. Right, L.S. Lowry. Industrial Landscape. 1955. Tate. Presented by the Trustees of the Chantrey Bequest 1956. © The estate of L.S. Lowry/ DACS 2016.
Creating a snapshot into the museum’s vast collection of British paintings, the Tate Britain now matches artificial intelligence (AI) with photo reportage to explore traditional British art. As part of the museum’s 2016 IK Prize—an annual award for a project seeking to enhance art gallery experience with digital technology—the Recognition project is anchored by a multifaceted algorithm, comparing the similarities between pieces of art with those of up-to-the-minute photojournalism from Reuters. The program will work over a three-month period to produce a virtual gallery of past and present works through four categorical processes to produce the matches: object recognition, facial recognition, composition recognition, and context recognition.The team behind Recognition is Fabrica, a communication research center based in Italy that works across digital storytelling disciplines. At the project launch, the team said, “We can’t wait to see what inspiring, insightful, humorous and thought-provoking relationships Recognition unearths between how the world is represented in British art and up-to-the-minute news.”Using art to demonstrate AI capabilities, visitors gain an inside look into how machines think, mimicking the human brain but with noticeable differences—where similar shapes and colors will pair a picture of L.S. Lowry’s Industrial Landscape (1955) with an image of Singapore’s Changi Airport.“AI is nowhere near as sophisticated as a human, whether they be an art expert or not,” says Tony Guillan, Tate IK Prize Producer. “The sophistication and flexibility of our brains can consider lots of different things simultaneously while cross-corresponding references against our memories, emotions, dispositions and personalities. AI can’t do that but what it’s trying to do here is take four simple criteria and blend them together in a way that the brain kind of works in. It’s a simulation of a level of human understanding.”The Recognition experience allows viewers to make their own matches between painting and photograph, inputting this data into the algorithm in hopes that it will learn from personal human experiences. Photojournalism from a Reuters live feed is used as it's a commonly accepted notion of providing a window into the world, thus creating a comparison of visual media both aesthetically and thematically. “What I love about Recognition is that it raises fundamental philosophical questions,” says Guillan. “Do we represent the world differently in painting than to the way we do now? Has the world been depicted in similar ways using different media? Is art relevant?”In collaboration with Tate and Microsoft, the team at Fabrica received a £15,000 prize and £90,000 production budget. Recognition is on display at Tate Britain through November. See more of the project online here.Related:Artificial Intelligence Controls These Surreal Virtual RealitiesHere's What Actually Goes into Creating Artificial IntelligenceGoogle Makes Learning Neural Networks Free
Advertisement
Advertisement