Google taps chipmaker Movidius to add machine learning to phones

Google taps chipmaker Movidius to add automobile learning to phones

This site may earn affiliate commissions from the links on this page. Terms of utilise.

movidius-machine-intelligence

A large reason the electronic devices around the states are getting smarter, and hopefully more useful, is machine learning. Past building complex models of data, then training those models, tasks equally diverse as facial recognition, language translation, and autonomous driving can be accomplished. In addition to the need for huge amounts of compute power to railroad train those systems, they all tend to be both very power and processor intensive to run. That has kept nearly of them tethered to plug-in devices — similar the Kinect — or requiring large batteries, like those constitute in a car. For example, Nvidia'south Drive PX 2 trunk-mountable car computer volition require liquid cooling. For mobile devices that has meant a abiding connection to the cloud, with raw data sent upwards, analyzed at the information middle, and the results returned.

Vision-intensive tasks like object recognition are ideal for the Myriad VPU Google has been trying to change this dynamic with Project Tango, a mobile device that tin practise existent-fourth dimension mapping and some object tracking, while running off only a minor battery. To achieve that, information technology tapped a new kind of processor, the Video Processing Unit (VPU) scrap Myriad 1 from startup Movidius. By moving the processor-intensive tasks associated with estimator vision into a peculiarly designed bit, Myriad increased the functioning of, and decreased the ability requirements for, the vision-related functions of the Tango device. Movidius claims at least a cistron of 10 savings in ability, along with an eighty% reduction in both space and cost over competing technologies — all compelling stats when it comes to mobile device design.

Beyond Project Tango: Using Movidius for mobile auto intelligence

At present, Google has broadened its relationship with Movidius, announcing that it volition be using the company's newest and almost powerful VPU, the Myriad M2450, to help bring more intelligence to a wider array of mobile devices. The Myriad isn't limited to running vision-related applications, either. Google will use Movidius's software evolution environment to port its advanced neural ciphering engine to the flake, and so that a wide-diverseness of deep-learning-based algorithms can exist run in real time.

Movidius's development board includes its chip, sensors, and a reference camera, and is instrumented for power measurement Existence able to run deep-learning-enabled tasks locally volition reduce dependence on the cloud, thus reducing latency and privacy bug. For instance, your phone could recognize your friends in a photograph without you needing to upload information technology to the cloud. Remi El-Ouazzane, Movidius CEO, explains, "The challenge in embedding this engineering science into consumer devices boils down to the need for farthermost power efficiency, and this is where a deep synthesis between the underlying hardware compages and the neural compute comes in."

Unfortunately, in that location aren't any details yet on any new Google products that will use the Movidius fries (and there was no mention of them at the Lenovo and Google Projection Tango telephone annunciation), but given the importance of figurer vision and car learning to the future of mobile devices, I'm sure we'll be hearing more soon.

Now read: Artificial neural networks are changing the world. What are they?