So Is Apple Brain: a Look through The Curtain of Secrecy

Apple has spent years integrating technology from of the artificial intelligence and machine learning (machine learning in English). According to a new interview with Steven Levy to several executives from Apple, the Cupertino company is in top form in this field. A claim that collides with what is saying in technology circles during the past months.
After the special emphasis that Google and Microsoft have been around this technology in their respective conferences, all bulbs are posed on Apple. While both competitors They advertised their progress to great fanfare, Apple remained silent. Thanks to this interview, we can get a better idea of what is hidden behind the curtain of secrecy of Apple.

“Apple is doing with one hand tied behind his back”

Everything you want to learning experts are automatic. But with his defense of the privacy of the user, Apple forces you to work with one hand behind his back. We can discuss if it is the right thing or not, but it has given Apple a reputation of not taking seriously the artificial intelligence experts- former employee anonymous Apple.
Excessive protection of the privacy of its users, coupled with the lack of a search engine like the one in Google, are arguments defending the late which is Apple on artificial intelligence and machine learning. And they do so outright. Without data, you can not train to the “brain” of Apple.
It is false that it must choose between the benefits of artificial intelligence and the privacy of users
The response from executives in the interview of Levy is compelling. Apple has managed to create its own technology without having to give up anything. According to Craig Federighi, senior Vice President of Apple software, There is no such commitment between privacy and artificial intelligence. We must not choose between one or the other.
Apple has built a “brain” within each of our devices. IOS 10, This brain will weigh around 200 MB. All sensitive data about the user, their habits and apps are stored on it. In addition to a general information not related to the user database. In this way, machine learning is performed in local, at each device, without any private data is stored in the cloud, or may be intercepted during transmission. Not even stored during the backup of iCloud.

Apple and iOS 10 differential privacy

A few months ago we talked about what was the differential privacy. It’s a technique developed at some universities that Apple will begin to implement in iOS 10. And it will do so on a massive scale. Hundreds of millions of iOS devices update from this September to the new mobile operating system from Apple. With this, Apple will have a broad base of users to feed their efforts in artificial intelligence and machine learning.
All without breaking his promise to paladin’s privacy in a world that has become currency of an infinite number of free services. In today’s technological industry, personal data are uploaded to servers to be crushed, analyzed and observed under the microscope. As stated in the interview Federighi, Apple wants to change this by means of differential privacy. This new technique is capable of obtaining patterns and trends at the same time which makes it mathematically impossible to find out anything about an individual user.
Rather than focusing on the technology, Apple looks for specific cases in which enhance the experience by using it
Little by little, users will enjoy more and more of the advantages of this technology applied in Apple products. Although the most obvious place that comes to mind is Siri, your applications go far beyond. The recommendations of QuickType keyboard that comes standard in iOS, facial recognition, objects and scenes in photos of iOS 10, write with Apple Pencil unless the screen detects the Palm of your hand, suggestions of places on maps, identification of unknown phone numbers or the own command “Hey Siri” are some of those examples.
And as Apple designs products in full way, is capable of much more. For years, artificial intelligence and machine learning they affect every step of the design of Apple products. For example, in the number and placement of microphones, the design of A series or audio processing software chips.
It is clear that Apple does not want to allow the technological narrative around these emerging technologies to flow against. To this end, the company is ready to show us a very intimate look in an earlier time would have been unthinkable. In any case, it’s a great interview for those who are interested in these topics.