Facebook and Google are working on implementing the concept of “deep learning.” Basically, it’s teaching machines to recognize the underlying patterns and structures of every type of information, to create true artificial intelligence. This maps nicely onto one of the major goals of library instruction: To illuminate the deep structure of information in order to facilitate transfer. The fact that Google and Facebook are using this to build robust AI is interesting–it’s the process of being able to take something and apply it to a new context that signifies that our brains are, well, human and up to this extremely complex task.
Deep-learning software attempts to mimic the activity in layers of neurons in the neocortex, the wrinkly 80 percent of the brain where thinking occurs. The software learns, in a very real sense, to recognize patterns in digital representations of sounds, images, and other data.
There’s a great article in MIT Technology Review, if you’d like to read more.