Google Releases AI Project to Open Source Community
- Tuesday, November 10th, 2015
- Share this article:
The technology is being used by Google already to improve a number of functions in everything from Search to Gmail, and is the software that enables Google Translate to detect foreign words on street signs and translate them in real-time.
It recognises complex data structures in information, known as tensors, and passes them through an artificial neural network that resembles how the human brain deals with abstract concepts. The technology forms a key part of deep learning, the powerful AI tool used in many of Googles products.
“Just a couple of years ago, you couldnt talk to the Google app through the noise of a city sidewalk, or read a sign in Russian using Google Translate, or instantly find pictures of your Labradoodle in Google Photos,” said Sundar Pichai, CEO of Google in a blog post on the release. “Now, thanks to machine learning, you can do all those things pretty easily, and a lot more. But event with all the progress weve made with machine learning, it could still work much better.”
By releasing the technology as open source software, Google is hoping to spur a new wave of development and encourage the next evolution of machine learning and AI. The technology can run on a single smartphone, meaning that we could see a flood of new apps making use of the tool in exciting new ways soon.