Home > Web > Google mixes art and machine learning to create…

Google mixes art and machine learning to create new interactive experiments

Google has been injecting machine learning into some of its products lately, and now it’s mixing it with art.

The company’s Arts & Culture team in Paris was curious to see what “creative coders” could do with machine learning, and the results are three experimental works that introduce new ways to interact with and find art — X Degrees of Separation, t-SNE Map, and Tags.

X Degrees of Separation is a project aimed at finding what visual similarities a computer algorithm uses to connect a sculpture and a painting. The website lets you drag and drop works of art to find sculptures or paintings that are visually connected. For example, we dropped in Van Gogh’s self-portrait and Claude Monet’s “Water Lillies.” The tool provided three new photos, but the median was a woman standing in a river.

google-arts

Next up is t-SNE Map, which creates a “3D landscape” housing a vast number of artworks spanning centuries. But the art isn’t randomly placed — the computer places artworks it believes are visually linked closer together. You can zoom in up close and interact with each work of art to get more information. It’s a large collection, so it may take some time to load the clusters of photos.

And the final experiment is Tags, and like the name suggests, a computer analyzes various artwork to tag them with specific categories. The Tags website lets you sift through these categories and see what works of art fall under each one. These tags can range from “azure” and “skin,” to “fiction” and “map.”

grass

The three tools are the first to be a part of the new Google Arts & Culture Experiments, which the company says is an “online space where you can see and play with the experimental projects that we have built.” It’s one of Google’s many labs experimenting with various technologies such as artificial intelligence and machine learning.