Researchers from Cornell University, New York and tech firm Intel have created a mammalian olfactory bulb (MOB), an Artificial Intelligence (AI) based on the olfactory system of mammals, to mimic the process of identifying and differentiating between different smells that are usually present as a mixture of compounds in the air.

The scientists designed a silicon-based electrical circuit based on the neural circuits in the brain which get activated when the brains receive and process odour. The chip is enhanced with algorithms that send an impulse through the circuit upon ‘detecting smells’. The algorithm is trained in data sets of 10 ‘smells’ such as ammonia and carbon monoxide which are characterised by the combination of 72 different chemical sensors. This chip has proven to be much more efficient with the help of fewer samples, as compared to conventional chips. The AI got almost 100 percent results while identifying eight of the smells, and almost 90 percent results for the remaining two. However, the AI is yet unable to detect the smell signal in a crowded environment with various other odours present. 

These chips, also called the neurotrophic chips, mimic the brain’s structure as much as possible to increase its learning efficiency. The chips currently in use are far less efficient as learners, their algorithms require massive data and energy; all of these drawbacks are eliminated by the MOB AI. While the chip is still in its nascent stage, as it evolves, the technology can be used in numerous ways - from bomb detection to pollution detector.