Google releases web app Keen that collects topics through machine learning

Google has released a new application that allows users to search for content on specific topics through machine learning. Users enter a topic and then see photos, videos, and articles about it.

Keen can be used on, or via the Android app. Keen is a kind of social network based on different topics and interests, a bit like Pinterest.

Users must sign in with a Google account. Then they can create ‘keens’, personalized sections. After entering a topic, the network automatically provides additional suggestions based on the first search term.

Articles, photos and videos on the subject are then put together in a ‘keen’. They can save it, but also add a link to it themselves. Of course, the keens themselves can also be shared with others, and users can receive a weekly email with additions.

The network distinguishes itself by working with Google’s own machine learning algorithms, the company writes in a blog post. “The more you store and organize in your ‘keens’, the better the recommendations get,” says developer CJ Adams.

Keen was founded by Google’s Area 120 division, a department that creates experimental new features. The Area 120 programmers also collaborated with Google’s People and AI Research or PAIR department. Given the silent release, it seems to be an experimental new network for the time being.