Current Issue
This Month's Print Issue

Follow Fast Company

We’ll come to you.

Company | Profile


Google has been improving its machine learning technology for years, and in 2015 it rolled out some impressive updates to its apps. In May 2015, the company announced that Google Photos was now searchable--type in "cat" and pictures of your pet will pop up--and in July 2015, it released real-time visual translations on Google Translate. Another development was unveiled in November 2015: Smart Reply, a plug-in for Inbox that reads emails and suggests short responses. None of the new developments are quite perfect yet, but all are the result of Google's artificial intelligence engine, TensorFlow, often considered one of the world's most advanced AI systems. TensorFlow processes complicated data through a neural network faster than any of Google's previous AI engines and can be run on thousands of computers, meaning it's scalable. Perhaps the most significant--and surprising--announcement regarding TensorFlow came when Google decided to open-source part of it, letting anyone try out its algorithms and, it hopes, improve them. The goal, according to the tech company, is to "accelerate research on machine learning"--which would benefit both Google and the data community as a whole.