Scientists Succesfully Give Artificial Intelligence a Sense of Smell

It is evident that artificial intelligence is evolving at an accelerated pace as of late. Scientists have created a new set of algorithms that would allow AIs to determine the odor of particular molecules based on their chemical structure. In the future, this technology can aid food producers and fragrance makers to create new odors with that “perfect scent”.

Artificial Intelligence With a Sense of Smell

In a way, this development means AI has created a nose, bringing it one step closer to being equal to human beings. Although this development may put off technophobes, it goes to show scientists are exploring many different use cases for AI technology. That can only be seen as a good thing, as it will help improve our society as a whole.

When it comes to letting humans perform odor tests, there are less than two dozen adjectives we would use. During a recent experiment, 476 vials of pure odorants were being tested, rendering 19 different descriptions. That seems somewhat strange, considering no two samples were exactly alike. Then again, all test subjects rated every smell, creating a vast dataset of information.

Based on over 1 million data points, scientists went to work to create a new AI algorithm. Based on these findings, it is now possible for artificial intelligent beings to differentiate between different smells. To be more precise, AI-based solutions can define scents based on the chemical structure of individual molecules.

Earlier scientific research indicated humans have over 400 different odor receptors throughout bodies. However, no one knows for sure how they work and allow for distinguishing different smells. This is where artificial intelligence can play a big role, as the algorithms developed will help scientists gain a better understanding of how humans work. This latter part can be vital to the survival of our species in the future.

By putting these AI algorithms to the test, it became clear that there are some intriguing results. One team was capable of predicting how individual subjects would rate the odors. Another team scored best in how the average person would rate these particular smells. Both results validate the algorithm being used for this purpose, albeit some minor modifications may need to be made along the way.

In the end, the biggest “winners” of these experiments will be fragrance and flavor companies. Triggering that particular smell has proven to be difficult without the proper knowledge of how humans interpret them. This is only the first step towards giving artificial intelligence a “nose”. It will be intriguing to see how scientists improve upon these exciting findings.

If you liked this article, follow us on Twitter @themerklenews and make sure to subscribe to our newsletter to receive the latest bitcoin, cryptocurrency, and technology news.