Google technical scientists have discovered that a series of computers is learning on its own without having to be directly programmed by humans. The secretive ‘X Lab’, funded by Google, which is most well known for innovating cars which can pilot themselves and eye glasses which augment reality for the wearer, recently began a project with the goal of creating a computerized neural network. The processed involved connecting almost sixteen thousand computers together and allowing the phalanx of computers to perform their own searches on various topics via the Internet.
Shockingly, without any outside programming by humans, the network of computes has learned how to identify and recognize what a cat looks like and its basic physiology. Although this may not seem to be a very impressive feat the Google team stated that their expectations for the computer network were much lower; instead, the computer neural network has doubled its accuracy rate in identifying various objects from a list of over twenty thousand items!
As any Internet socialite knows, YouTube and Google image searches have an inordinate amount of cat results. The computer neural network was uploaded randomized images from the Internet (which simply included many images of cats) which the neural network then identified as ‘cats’ on its own. There are biological theories that declare that the human brain is trained to detect patterns and specific objects and the fact that this computerized neural network came to this same result supports those theories. Google researcher Jeff Dean commented, “We never told it during the training, ‘this is a cat.’ It basically invented the concept of a cat.” As the technology of computerized science falls and the demand for results increases there are sure to be more innovations and discoveries in the field of artificial intelligence in the near future.