Researchers at the Google’s secretive X Laboratory have developed a neural network that claims to be able to identify cats and ‘learn’ what cats look like, enabling it to almost completely emulate a very small part of work done by human brain. Google, with the help of Stanford University researchers, has been working on this for quite some years now with it only bearing fruit recently.
Being one of the largest neural networks ever built, the ‘cat identifier’ consists of 16,000 computer processors. The system was fed with about 10 million images captured from random Youtube videos. The system then learned to identify cats from the images by learning the different characteristics of a cat.
The picture below shows an image of a cat that the system taught itself to recognize:
Google also said that the project has now been moved from the company’s X Laboratory to its main research area. The company expects the research to turn out fruitful in providing better image search, speech recognition, and machine language translation. However, clearly the consequences of this artificial semi-intelligence cannot be predicted.
Though it might not be a surprise to some, aside from potentially marking the beginning of when machines can think and learn by themselves, this shows the plethora of cat videos and images there are on the Internet.
Do you think that this is looking like something bad or do you believe that this is going to bring great advances to human lives? Share your thoughts with us in the comments below.
[via New York Times]