Qualcomm is gearing up to bring deep learning to future-gen smartphones, thanks to its new software development kit (SDK).
On May 2, the chip maker revealed that it was working on its first-ever deep learning SDK dubbed "Snapdragon Neural Processing Engine." It has been designed to take advantage of the diverse computing abilities of the Snapdragon 820 chipset. The SDK will run on top of the Zeroth Machine Intelligence Platform from Qualcomm.
Through the SDK, Qualcomm will be able to empower smartphones running on the snapdragon 820 processor to bring deep learning capabilities.
Deep learning is essentially a subcategory of machine learning. It is a means of teaching a machine how to perform a task instead of programming the computer to do it. Moreover, such a network has more comprehensive layers to help analyze the relationship that exists between variables.
Post the SDK's availability, it will become possible for smartphones such as the LG G5, HTC 10, or the Samsung Galaxy S, which runs on the Snapdragon 820 chipset, to get the deep learning abilities.
The Snapdragon Neural Processing Engine SDK will essentially have a run-time feature, which will take advantage of the chip's features to deep-learn things, such as recognition of a sound or even tracking an object.
"With the introduction of the new Snapdragon Neural Processing Engine SDK, we are making it possible for myriad sectors, including mobile, IoT and automotive to harnesses the power of Qualcomm Snapdragon 820 and make high-performance, power efficient on-device deep learning a reality," said Gary Brotman, Qualcomm's director of product management.
The SDK has immense possibilities as Brotman highlighted and could potentially be deployed in robots, autonomous drones and self-driving vehicles.
Qualcomm's Snapdragon Neural Processing Engine SDK is anticipated to become available later in 2016.
