Qualcomm last year began launching its first chip for Android phones that support the GPU drivers that can be upgraded to optimize performance, so now doing similar things for AI-devices AI and engine learning. Droid-Life shows that during Google I / O, Google and Qualcomm have announced an updated Neural Network Fire Driver, representing a new model that will be launched with Android 12.
While the Fire NN drivers are usually updated together with large OS updates, now the company says they can slide quickly through the Google Play service. Even better, updates will be available for longer chipsets and several Android versions.
In the presentation of I / O about progress in engine learning, Google developers say Fire NN can improve performance as if the phone has two additional CPU cores, while using less power and creating less heat. Qualcomm points to Google Assistant and Google Maps as an example of an application that uses AI processing on the device and can benefit from tweaks. In the video, they also show how the power of the machine learning tools on devices such as direct shows or automatic background replacement.
For developers, the benefit is to have one API with routine updates that they can target in several versions, and if other chipset manufacturers join this program, the reach will extend more. For users, we must be able to get a smaller application that runs better and does not drain the battery to do things like voice or face recognition, while also not uploading your data back to the server, impacting your privacy.