Chip giant Intel is in the middle of an experiment with a neuromorphic chip that attempts to resemble the way the human brain work. Imagine a situation where you are asked to guess the emotion of someone in a video clip. It will lead to neurons in your brain exchanging information in a flurry of electronic spikes. In a similar manner, researchers at Intel recently put a challenge to the prototype of their new chip. The chip, called Loihi, tried to solve the problem with thousands of spiking silicon neurons of its own. Like our brain’s neurons, they can adjust the connections between themselves to adapt to new tasks. The new chip consists of 128 computing cores and each core has 1,024 artificial neurons, giving the chip a total of more than 1,30,000 neurons and 130 million synaptic connections. The new design, which is named after a submarine volcano in Hawaii, still is far away from a human brain. The human brain is made up of more than 80 billion neurons. But it’s very different from a conventional processor. Based on the number of neurons, the Loihi chip is a little more complex than the simple lobster brain.
The company predicts that this approach could one day make cars, cameras and robots smarter without having to rely on an Internet connection to the cloud. The spiking nature of the simulated neurons will make the chip run more efficiently than a traditional chip design, Intel said. It will be able to interpret video using as little as one-thousandth of the energy of a conventional chip. Intel’s new AI processors learn locally inside of the machine they’re installed. Also, cutting the cord removes the need to wait for data to traverse the Internet and it also has some privacy benefits.
Intel has made two smaller prototypes and is now testing the full design. Loihi is still a research project and it will not be until November that the first full version of the chip will be fabricated. Some academic and research institutions will get to try it in 2018.
Loihi is Intel’s latest effort to turn the current vogue for AI into a new growth engine for the company. Last year, Intel acquired two start-ups working on chips to power machine learning in the cloud and for computer vision. Two years ago, it spent $16.7 billion, its largest acquisition ever, to acquire Altera, which builds the programmable chips that Microsoft uses. Last year, the company paid a reported $408 million buying Nervana, a company that was exploring a chip just for executing neural networks. Now, led by the Nervana team, Intel is developing a dedicated chip for training and executing neural networks. Rival chipmaker Nvidia currently dominates the AI market with its graphics processors. Officially, the company’s position is that improvements in traditional chips will continue well into the next decade.
But Intel is not the first company to design a chip using pointers from neuroscience. Tech giant IBM built two generations of its own neuromorphic processor. But that chip, unlike Intel’s, can’t learn from incoming data. The company has struck deals with two labs to build research systems with its chip, but not announced broad commercial availability. Whether or not Intel’s neuromorphic chip experiment ever becomes successful, it’s worthy to note that the unveiling highlights Intel’s interest in moving beyond the traditional central processing unit (or CPU) market, where the company is a market leader. Intel controls more than 90% of the data-centre market, making it by far the largest seller of traditional chips.
Intel is competing not only with chipmakers like Nvidia and Qualcomm, but also with companies like Google and Microsoft. Google is designing the second generation of its chips. Later this year, the company said, any business or developer that is a customer of its Cloud-computing service will be able to use the new chips to run its software. With this new chip, the hope is that there will be devices that will handle more complex tasks without having the need to call back to distant data centres.