Like it or not, Google does not rest and while you put your content before the start of a new extension for Chrome, the people of Mountain View is working on an artificial brain of billions of neural connections using 16,000 computers. While still unable to go into space as the mind of the crew and then conspiring to kill all humans in it, now is working to identify cats on the web through a learning process that does not carry any human assistance.


Although Einstein said that human stupidity is one of the only certainties, our brain is the machine that every scientist dreams of imitating through some invention that forever revolutionize artificial intelligence. The calculation capacity, the unused potential and everything that makes the mystical brain has drawn attention to laboratory scientists Google X, apart from the invention of augmented reality devices and cars that drive themselves, have made research on neural networks and how machines can learn alone if they are connected to each other and the Internet. Thus, the artificial brain of Google begins to take shape.

The first test was done to the neural network consists of 16,000 processors Google more than a billion connections was to get online and expose it to 10 million thumbnail images found on YouTube videos. The neural network then began searching for what humans look for more: cats. According to information leaked, the neural network is taught herself to discover the cats in the pictures, and the result will be presented by researchers next week, but they are very satisfied as cats have been recognized efficiently in a list than 200,000 objects. It also reconfirmed the biological hypothesis that says that individual neurons are trained to detect important objects, even in the era of today are cats.



This learning model which is called "deep learning" (Learning in depth) is part of the latest in research of speech recognition by computers these days, and Microsoft also demonstrated to be aware and in the process of investigation. Despite not even dream of achieving the power of human cortex, the Google brain artificially proves to be more and more intelligent as it faces huge amounts of data. Is that since Google insisted that there was some support for the selection and identification of objects during the learning process and that this could have been neurologically successfully completed by repetition. That is, they never told "this is a cat", but the neural network invented the concept itself.

AI,artificial,Brain,deep learning,google,Technology