Pinterest is an online pin board of visual inspiration founded in 2009 by Ben Silbermann, Evan Sharp, and Paul Sciarra. Covering everything from travel to home renovation projects, it is a unique social platform that gives its users a place to express what they want in the future. Every idea is represented by a “pin” that includes an image, a description, and a link back to the image’s source online where they can learn more about the idea. Pins can then be further organized into "boards," which sort ideas based on categories. Every month, Pinterest processes hundreds of millions of image searches in its bid to help users find things they’re most interested in. It is increasingly turning to machine learning to surface content that resembles objects its users have already pinned—so if someone pins a picture of, say, a midcentury table, it will automatically recommend that they look at pictures of other furniture from the same era.
Beauty is one of Pinterest’s most popular categories, but members often found it challenging to find results that specifically complemented their skin tones. To better serve diversity, the visual discovery engine leveraged AI and user recommendations to develop an algorithm for people to search beauty images by skin tone ranges and hair colors. The company also generated more training data to correct the machine learning algorithm by passing them through a human evaluation platform. The result is an inclusive filter that sifts through billions of images ranging from wedding hair tutorials to everyday makeup looks.