So Google set out to create an Artificial Intelligence neural network made up of 16,000 computer processors which they then turned on to the internet with no direction or guidance to learn on its own. It went on the internet and like any internet citizen focused mostly on pictures & videos of cats on youtube.
The neural network taught itself to recognize cats, which, John Markoff of The New York Times reports, is actually no frivolous activity. This week the researchers will present the results of their work at a conference in Edinburgh. The Google scientists and programmers will note that while it is hardly news that the Internet is full of cat videos, the simulation nevertheless surprised them. It performed far better than any previous effort by roughly doubling its accuracy in recognizing objects in a challenging list of 20,000 distinct items.
Googles Artificial Intelligence main mission was to build high level features using large scale unsupervised training, basically look at things on the internet and create an image based on what it learned.