The British chip design firm ARM came up with the processors used in virtually all the world’s smartphones. Now it plans to add the hardware that will let them run artificial-intelligence algorithms, too.
ARM announced today that it has created its first dedicated machine-learning chips, which are meant for use in mobile and smart-home devices. The company says it’s sharing the plans with its hardware partners, including smartphone chipmaker Qualcomm, and expects to see devices packing the hardware by early 2019.
Currently, most small or portable devices that use machine learning lack the horsepower to run AI algorithms, so they enlist the help of big servers in the cloud. But enabling mobile devices to run their own AI software is attractive. It can speed things up, cutting the lag inherent in sending information back and forth. It will allow hardware to run offline. And it pleases privacy advocates, who are comforted by the idea of data remaining on the device.
Jem Davies, who leads the machine-learning group at ARM, says the company spent a long time getting the chips to run AI software efficiently. “We analyze compute workloads, work out which bits are taking the time and the power, and look to see if we can improve on our existing processors,” he explains. The new chips use less power than the company’s other designs to perform the kinds of linear-algebra calculations that underpin modern artificial intelligence. They’re also better at moving data in and out of memory.
Of course, ARM isn’t alone in building mobile AI chips. The iPhone X, for example, contains a “neural engine” as part of its main chipset, which Apple created to run artificial neural networks for things like images and speech processing. Huawei’s Mate 10 smartphone contains a similar, homegrown chip that it calls a neural processing unit. The Pixel 2 handset has a chipset designed by Google to help it crunch imaging and machine-learning problems.
But ARM has an impressive track record of designing energy-efficient processors for mobile applications, and manufacturers are used to using its chips in their devices. Despite the competition, its new AI brains are likely to appear in plenty of devices next year.