Intel Corp. today revealed that researchers are using its neuromorphic chips to develop artificial skin for robots, in a project representing one of the first practical applications of the technology.

Intel, the leading maker of central processing units, is researching alternative chip architectures to help it maintain its long-term competitive advantage. Neuromorphic computing is one of the areas where the company is active. The term refers to an emerging class of chips that have transistors modeled after neurons to help them run artificial intelligence models faster. 

Intel’s research efforts have so far produced a 2 billion-transistor neuromorphic chip called Loihi. According to the company, it can solve some problems 1,000 times faster than a regular processor but uses 10,000 times less power. Loihi plays a leading role in the artificial skin project Intel detailed today. 

In the project, researchers affiliated with the National University of Singapore have developed “electronic skin” dubbed ACES that enables robots to detect when they come into contact with an object. ACES can also determine the shape, texture and hardness of a given object about 10 times faster than the blink of an eye. The idea is to provide machines such as robotic arms with more data about their environments to help them make better decisions.

The researchers behind ACES employed cloud-based Loihi chips to help process sensory readings from the artificial skin. Moreover, they used the opportunity to compare the technology with traditional chips.

Intel claims that the Loihi chips achieved accuracy of more than 92% in an experiment that involved using artificial skin and AI to read Braille letters, all while consuming 20 times less power than an unnamed “standard Von Neumann processor.” In another experiment, Loihi beat a “top-performing” graphics card while using 45 times less power.

“The work adds to a growing body of results showing that neuromorphic computing can deliver significant gains in latency and power consumption,” said Mike Davies, head of Intel’s Neuromorphic Computing Lab.

Artificial skin is far from the only use case Intel could potentially target with Loihi. There are entire categories of connected devices, from smart speakers to industrial sensors, that require the ability to run AI models locally on a relatively small onboard battery. A neuromorphic processor that performs the task with dozens of times less power than traditional silicon would have the potential to drive a lot of chip orders.

That’s particularly true as machine learning adoption continues to widen. As part of the trend, a growing number of phone makers are building AI chips into their handsets, and Intel has previously said it’s eventually hoping to make Loihi small enough to run inside a mobile device.

The data center is another market the company is eyeing. Intel is working on a rack-mounted implementation of Loihi and may have an opportunity to target enterprises looking to optimize on-premises AI infrastructure. There might even be interest from the major public cloud operators, given how key of a use case AI has become. 

Working with academia on initiatives such as the ACES artificial skin project allows Intel to gather valuable technical feedback as it fine-tunes Loihi. It’s also a way to produce real-world case studies with which to woo future customers.

Photo: Intel