The coprocessor will be able to run deep learning neural networks locally
Microsoft has revealed that the second generation of its HoloLens ‘mixed reality’ headset will incorporate a chip deigned to power artificial intelligence (AI) functions in the wearable device.
The chip will be a custom design by Microsoft and will function as a coprocessor to go alongside the custom multiprocessor unit featured in the current HoloLens known as a Holographic Processing Unit.
The coprocessor will be used for running deep learning neural networks withing the HoloLens 2 hardware, which are used to train smart software and AI algorithms in a fashion not dissimilar to the way brains dissect information.
HoloLens 2 will embrace AI
“The AI coprocessor is designed to work in the next version of HoloLens, running continuously, off the HoloLens battery. This is just one example of the new capabilities we are developing for HoloLens, and is the kind of thing you can do when you have the willingness and capacity to invest for the long term, as Microsoft has done throughout its history,” said Marc Pollefeys, director of science at Microsoft’s HoloLens division.
“And this is the kind of thinking you need if you’re going to develop mixed reality devices that are themselves intelligent. Mixed reality and artificial intelligence represent the future of computing, and we’re excited to be advancing this frontier.”
One of the advantages having a deep learning capabilities run locally on a device, is that it cuts out the latency encountered when trying to feed data to and fro between a device and a central system or cloud, allowing for smart software to act faster. It also means than in the case of a lost Internet or wireless connection the device’s smart capabilities will continue to function.
Microsoft has not revealed much more details into how an AI would be put to use within the second generation HoloLens, but it would not be too much of a leap of imagination to suggest that it will use the coprocessor’s capabilities to power apps and features that use image recognition and computer vision; two techniques that make heavy uses of deep learning neural networks.
While Microsoft is keen to embrace the potential of AI, it will be keen to avoid the pitfalls it has encountered with some of its AI work so far.
Want to display your knowledge on screen technologies? Take our quiz!