In what may be a huge step for the intersection of Artificial Intelligence and neuroscience, researchers at technology company Intel and Cornell University recently demonstrated the ability of a neuromorphic computing chip to learn and then recognize the odour of certain hazardous chemicals.

The researchers used a neural algorithm based on our brain’s olfactory circuits to train the Loihi research chip to sniff out the scents of 10 hazardous chemicals, including ammonia and methane. In order to do so, the team fed Loihi a dataset consisting of the activity of 72 chemical sensors in response to these smells and configured the circuit diagram of biological olfaction on Loihi, according to a news release. The chip quickly learnt the neural representation of each smell and recognized each odour, even in the presence of significant background interference. The findings were published recently in the journal Nature Machine Intelligence.

Such neuromorphic chips could be used to build intelligent “electronic nose systems" that could then be used in robots to detect hazardous materials or even for environmental monitoring. They could also be used for medical purposes—certain diseases, for instance, emit a particular odour.

My Reads Logout