Do you want to see Artificial Intelligence and machine learning in action for yourself? Something that was specifically pointed out as very hard for computers by the experts in the field just a few years ago? Content based automatic image classification is a fascinating window into what AI is already able to do and one that is available to all.
Look at the automatic classification of photos that is now available on Google Photos Things, Flickr’s Magic View. (You have to have your own account and look at your own photos to experience it.)
How can you manage tens of thousands of photos? You can’t, you don’t and you don’t have to. Your fancy AI helper can do it for you.
Birds, horses, monuments, concerts, lakes, palaces, hundreds of different concepts and objects are now recognized on photos automatically.
This is certainly bound to be a standard feature in next generation phones. Whether through the cloud, or on the device itself, the camera lenses of our machines, their eyes on the world, will progressively have a better and better understanding of *what they are seeing*!
It is fascinating to browse through the classification, genuinely useful for photo libraries with 100,000 photos like mine, and also extremely interesting in those cases where they get it wrong.
Soon the subtlety and precision, not only just the speed and volume, of image recognition by machines is going to be at superhuman levels. Machines now can beat humans in recognizing faces, for example.
With a concrete example we can see how different artificial general intelligence is going to be. Already a basic function such as vision is implemented in Google’s autonomous car in a totally different way than not by our eyes and brain: using an infrared laser beam, the machine scans 360 degrees with a very high spatial and temporal resolution all its surrounding. No blind spots, without batting an eyelash, without a missing a quick sequence of events.