Qualcomm AI Engine aims to reduce the time it takes your phone to perform tasks


Qualcomm has revealed a new artificial intelligence (AI) engine as part of a wave of MWC 2018 announcements today. While the engine is more of a background player, you'll likely see improvements on your mobile device wherever the engine is implemented.

The Qualcomm AI Engine isn't made up of new parts, Gary Brotman, Qualcomm's director of product management, said in a briefing. Rather, the engine itself is the new arrival. It brings together different hardware and software components to enable faster AI user experiences through on-device processing.  

The various parts now occupying one roof include hardware architectures, like the Adreno GPU, software components, such as the Snapdragon Neural Processing Engine and Android Neural Networks API, and frameworks, like TensorFlow and TensorFlow Lite.

To support the new AI Engine, devices will need to house a Snapdragon 845, 835, 820 or 660 chipset. No devices with the Snapdragon 845 are official yet, but the Samsung Galaxy S9 and Galaxy S9 Plus will be some of the first arrivals next week. The chip should also make its way to some of the year's best Android phones

Qualcomm notes that many of the manufacturers that already support parts of the new AI Engine, which include names like OnePlus and Motorola, are on board to utilize the full thing on future flagship devices running Snapdragon. 

Make it snappy

From a user standpoint, the new engine is said to cut down on the time it takes to do certain tasks on your phone, such as zeroing in on a subject when taking a photo. What's more, battery life doesn't suffer as much in the process.

Voice and natural language processing are other areas billed as having reduced latency. Conversation is more natural with on-device processing, Brotman said, and gone is the lag that creates the sensation of talking to a robot. 

Brotman put the improvements expected with the new AI Engine like this: "Anything to do with the time that it takes perform a specific task and the accuracy of the task are two key areas where the experience will improve for users."