Intel is making chipsets learn how to smell like human beings

The chip can currently identify 10 different smells, for which the researchers had to record the response of 72 chemical sensors sitting in a wind tunnel.

Intel's logo is pictured during preparations at the CeBit computer fair. (Reuters)

Chipmaker Intel in collaboration with researchers from Cornell University has found a way to make processors learn how to smell just like humans. The new algorithm that is published online, is said to be based on the same biological signals that let humans smell food for instance, reports ZDNet. This mathematical algorithm has been made by the Intel's neuromorphic computing group and was implemented in the company's Loihi neuromorphic computing chip - a 14-nm, 130,000-"neuron", 130 million-"synapse" processor.

As a refresher, the Loihi neuromorphic computing chip was unveiled back in 2017 and was reportedly six years in the making. It is specifically made to solve complex AI and deep learning problems. It is described as a 'self learning' chipset that mimics the human brain and is developed by the Intel Labs division.

The chip can currently identify 10 different smells, for which the researchers had to record the response of 72 chemical sensors sitting in a wind tunnel. These 10 scents were circulated through it and the data was later added in the Loihi chip. The report adds that Loihi can not just recognise smells but can also identify them with background interference.

"My friends at Cornell study the biological olfactory system in animals and measure the electrical activity in their brains as they smell odors. On the basis of these circuit diagrams and electrical pulses, we derived a set of algorithms and configured them on silicon, specifically our Loihi test chip," said Nabil Imam, senior research scientist in Intel's neuromorphic computing groups.

This development comes after company's last year's announcement of 8 million-neuron neuromorphic system made of 64 Loihi research chips.