On Thursday February 4th 2016, researchers at Massachusetts Institute of Technology, or MIT, announced a new type of computer chip that acts as an AI or artificial intelligence. This AI has deep learning capabilities which could potentially recognize objects. The chip is known as the Eyeriss. The team of researchers at Massachusetts Institute of Technology are spear head by Vivienne Sze from the department of electrical engineering and computer science.
Eyeriss has the potential of recognizing faces without being connected to an internet source. Currently as it stands, the team of researchers are trying to find a way to make the Eyeriss more portable. In doing so, this AI chip might be seen in self-driving cars, robots, wearable technologies, and smart phone devices. Another place where one might see the AI chip is in IoT devices.
In the past, deep learning capabilities usually required a high amount of computer processing in the past. However, the Eyeriss chip will have a power usage which is more than tenfold that of older deep learning functions. In the past, deep learning capabilities require an internet connection where the AI would constantly upload raw data over the internet while using very strong GPU servers.
Alex Brokaw, a journalist at The Verge explains a little more about the deep learning. “The first is that if your mobile device can’t find an internet connection, then it can’t carry out deep learning tasks.” He then uses Apple’s Siri as an example. “Siri, for example, needs lots of processing power to understand speech, which is why it won't work unless it can reach Apple's servers over the web.”
On the other hand, Eyeriss will be using remote servers that do not require internet access. There has been no release date as of yet for the Eyeriss.