Google’s X labs have given the world a few technological innovations and the latest thing that has come from the mostly-secret research lab is a neural network that has learned how to recognise cats.
Google scientists built the leaning network out of 16,000 processors and it promptly started teaching itself about cat videos. The scientists fed the neural network random YouTube thumbnails from a pool of 10 billion YouTube videos and, says Google’s Jeff Dean, “We never told it during the training, ‘This is a cat. It basically invented the concept of a cat. We probably have other ones that are side views of cats.”
Scientist Andrew Y. Ng, speaking to the New York Times, said, “The idea is that instead of having teams of researchers trying to find out how to find edges, you instead throw a ton of data at the algorithm and you let the data speak and have the software automatically learn from the data.”
The researchers think that they may have created something – in machine form – akin to what happens in in a human’s visual cortex but also maintain that they probably haven’t hit on machines that can teach themselves. Yet.