During today's Google I/O 2016 keynote, the company focuses on its efforts in machine learning and AI, to help improve efforts in robotics, medicine and more
The machine learning segment was the last part of the keynote, which was presented by Google CEO Sundar Pichai. He stated that the company has open-sourced part of its machine learning platform, TensorFlow, as well as parts of its Cloud Platform APIs. This will allow for developers to access its computer vision APIs and specialized hardware.
Google has also started building customized machine learning hardware, called Tensor Processing Units (TPUs). That hardware was used to power AlphaGo, a version of the classic game Go. Pichai stated that most people thought an AI would not be able to play Go well for another decade but the new machine learning tools have allowed the AI to make creative moves in Go.
Robotics is another area where machine learning from Google is making strides, with robotic arms learning how to pick up specific objects without specific coded instructions. Finally, machine learning is being used to train computers to detect Diabetic Retinopathy in humans, which is one of the leading causes of blindlness.