Google explains its latest gains in computer vision

The annual ImageNet computer vision competition took place in August, and once again the winning techniques blew away past winners. John Markoff at the New York Times published a piece on competition and some of those improvements last month. On Friday, Google published a blog post highlighting the techniques — including, of course, deep learning networks — a team of its researchers used to achieve top score in a few different categories. Vision is only a piece of the puzzle for things like robots and driverless cars, but it’s a major component of nearer-term advances such as visual search engines.

Source: Google