Google simulates successful 'brain' artificially

Google and Stanford University have linked 1,000 computers, creating a network of more than 1 billion connections, simulating an artificial brain with the ability to learn the most basic concepts.

Andrew Ng demonstrates the cognitive ability of the artificial brain - (photo: Extremetech)

Dr Jeff Dean, head of the team from the secret laboratory of Google, joined two other colleagues from Stanford University, Andrew Ng and Le Quoc, to create a brain imaging system for infants. and able to perceive.

With 16,000 cores in 1,000 computers linked together, brain computer systems have been able to recognize the face of humans and cats with precision.

"We did not tell it (computer system) in the training process it was a cat. Basically, it was self-aware of the concept of a cat," said Dr. Jeff Dean, said the New York Times.

Previous face recognition programs are often programmed
Identification point, but Google's artificial brain is not - (Photo: Extremetech)

According to The Verge technology, Google engineers have not loaded any cat characteristics into the artificial brain to be as identifiable as they did in the past. Instead, they allow the system to "see" a large number of YouTube videos related to cats. Then, the artificial brain begins to detect familiar objects and map the image to that object.

Over time, the artificial brain will form the initial concepts of cats through data about shapes, sizes, colors ...

According to the New York Times, in the near future this artificial brain will be applied to Google's image search engine. Google will also improve algorithms for the artificial brain system to search for images, voice recognition and language translation.