What does synthetic intelligence (AI) should do with a 150-year-old lion statue in London’s Trafalgar Sq.? Good query. However Google’s by no means been one to let cognitive dissonance stand in its means. Right this moment in partnership with British designer Es Devlin, the search large took the wraps off “Please Feed the Lions,” an exhibit on show as a part of the London Design Pageant.
In the course of the competition, curious onlookers can “feed” the fifth lion in Trafalgar Sq. by typing a single phrase right into a Google Pixelbook, which can increase it right into a line of poetry projected contained in the lion’s mouth. As extra people contribute, the poem will “evolve” and (due to a intelligent little bit of nighttime projection mapping) wrap round each Nelson’s Column and the lion’s decrease physique.
A neural community is chargeable for the nouveau verses. Ross Goodwin, a self-described artistic technologist who’s a part of Google’s Artist and Machine Intelligence program on the search large’s Arts and Tradition division, educated it on 25 million phrases of 19th-century poetry and programmed it particularly by no means to repeat a line, in order to ensure a level of randomness.
Inspiration got here from a dialog with London Design Pageant Chairman Sir John Sorrell, Devlin said.
“[Sorrell] stated: ‘[Sir Edwin Henry] Landseer by no means needed these lions to look so passive; he proposed a way more animated stance, however Queen Victoria discovered it too surprising.’ The thought lodged in my thoughts. What if we may make investments the lion with a diversely crowd-sourced collective poetic voice?”
Poems will probably be revealed each day on the exhibit web site. Following the competition, which ends September 23, Google will digitize the sculpture as a part of a web-based exhibition on the Google Arts & Culture app.
Google’s public experiments with artwork and AI date again to 2016, when it launched DeepDream, a neural web that generates trippy canvases from pictures. April 2017 noticed the launch of Google’s AI Quick Draw, which faucets an algorithm to rework doodles into clip artwork. And in January, the Google Arts & Tradition app gained a characteristic that scans selfies for resemblances in a 70,000-strong library of work and portraits.