Google’s latest and interesting X Lab project is about something that is performed by the internet public everyday. At Google’s Mountain View’s secret lab, they joined together 1,000 computers totaling 16,000 cores to form a neural network with over 1 billion connections, and sent it to YouTube looking for cats. Unlike the popular human time-sink, this was all in the name of science: specifically, simulating the human brain.
The neural machine was shown 10 million images from random videos, and it taught itself what our feline friends look like. This was not like other similar experiments, where some manual guidance and supervision is involved, this pseudo-brain of Google was given no such assistance.
The pseudo-brain searched for cats but it wasn’t just about cats. The broader aim was to check if computers can learn face detection without labeled images. After studying the large set of image-data, the results were positive and interesting, in addition to being able to develop concepts for human body parts (and of course cats). The experiment showed 15.8 percent accuracy in recognizing 20,000 object categories, which is a 70 percent rise as compared to previous studies, as claimed by the researchers. Complete details of the hows and whys will be presented at a forthcoming conference in Edinburgh.