Apple: we use machine learning to extend battery life

Apple uses machine learning to extend battery life ‘on all devices’. That is what the manufacturer claims. How the American maker of iPhones and Macbooks does this remains unclear for the time being.

Apple’s use of machine learning to extend device battery life is evidenced by a backstory from Backchannel journalist Steven Levy, who was briefed at headquarters about the manufacturer’s machine learning efforts. The story also states that Apple devices store up to 200MB of data to use for optimizations. Apple’s operating systems may limit the power consumption of devices, for example, by preventing little-used apps from running things that require power in the background or by turning off certain functions when the screen is off, something that Google does in Android with Doze.

Apple’s message with the story, according to the author, is that the company is at the forefront of artificial intelligence. In addition to extending battery life, the company uses machine learning to recognize digital assistant Siri’s speech, recognize faces in photos in iOS 10, determine whether users should switch from a weak Wi-Fi signal to a cellular network, and ignore unwanted calls. screen touches when using the Apple Pencil on the iPad Pro tablets.

The American manufacturer does not have a ‘machine learning’ team, but has specialists in various departments. In addition, Apple claims that it trains people internally to become specialists in this field, through training in relevant fields such as mathematics, statistics and cryptography.

It is the first time that Apple has spoken so strongly about machine learning, artificial intelligence and neural networks. The American company has in the past acquired companies in these fields, but unlike competitors such as Google and Microsoft, it has not been open about its own efforts in this area.

Cover photo of Steven Levy’s backstory on Backchannel