Is the public ready to accept the infinite glance of our own technology? That question springs up nearly everywhere I look these days, from the land rush in “deep learning” and AI companies (here, here, here) to the cultural stir that accompanied Spike Jonze’ Her. The relentless flow of Snowden NSA revelations, commercial data breaches, and our culture’s ongoing battle over personal data further frame the question.
But no single development made me sit up and ponder as much as the recent news that Google’s using neural networks to decode images of street addresses. On its face, the story isn’t that big a deal: Through its Street View program, Google collects a vast set of images, including pictures of actual addresses. This address data is very useful to Google, as the piece notes: “The company uses the images to read house numbers and match them to their geolocation. This physically locates the position of each building in its database.”
In the past, Google has used teams of humans to “read” its street address images – in essence, to render images into actionable data. But using neural network technology, the company has trained computers to extract that data automatically – and with a level of accuracy that meets or beats human operators.Not to mention, it’s a hell of a lot faster, cheaper, and scaleable.
Read More