The Google Play Store uses machine learning techniques from DeepMind to recommend apps to users by using a number of AI systems.
DeepMind is a company that specializes in artificial intelligence and is owned by Google’s parent company, Alphabet. The technology produced by the company is now leveraged to create Play Store app recommendations, as outlined in a recent blog post.
Through the AI systems, DeepMind says that the Play Store “supports one of the largest recommendation systems in the world.”
Google is able to consider users’ past preferences like their downloads and installs to deliver a personalized experience. DeepMind has created systems that are able to determine the relevance of an app for a specific user.
For instance, if a user installs a travel app, the Play Store will then recommend a translation app. The algorithm knows not to recommend more travel apps because the user already has one, and instead is able to provide suggestions that go along with the initial download.
The system has a generator that analyzes more than a million apps and then retrieves the most suitable ones. It then predicts whether the user would be interested in it. Lastly, the model puts the predictions into an optimization model that provides the most suitable apps to the user.
The Play Store and DeepMind teams have worked together to implement and test a number of phases to create a final product.