An Artificial Intelligence to Recognize and Predict Smells
The Google AI team has developed artificial intelligence that seeks to recognize and replicate various odors.
The system is run by artificial intelligence (AI). The technology was developed and tested by Google, academic institutions, and the Monell Chemical Senses Center researchers. The researchers recently revealed the methodology’s findings and provided a description of the process in an article that was published in the journal Proceedings of the National Academy of Sciences.
Experts claim that scents are frequently referred to as the molecules that odours emit into the atmosphere. These compounds are then processed there by the “sensory receptors” in our noses. These receptors interact with the brain to help us recognise smells.
According to scientists, we use sensory maps to distinguish and quantify scents in a manner like how we distinguish and quantify colors. For instance, maps, like the traditional color wheel, can show a variety of hues. The maps show these color combinations and how they interact.
But historically, “useful maps for smell have been nonexistent,” the researchers claim. This is because the chemicals involved in smell are extremely complex. They are not only reliant on the photons necessary for color vision.
The researchers also noted that while the human eye only has
three color-sensing receptors, the human nose has more than three hundred.
The researchers developed a neural network model that seeks to
provide a “map” to identify chemicals connected to odors by building
on prior studies. A computer processing system designed to resemble the human
brain is known as a neural network.
The newly created technology is referred to as a “primary
odor map” by the Google AI team. The map’s major objective, according to
the researchers, is to “predict the odor qualities of compounds.”
The neural network models also taught the necessary molecular traits for making accurate smell predictions. The researchers
stated that they can “predict and find novel scents and the
molecules that cause them” thanks to the AI-powered map.
According to the study group, there are undoubtedly billions of
molecules that are aromatic but have not yet been detected as odorous.
The mapping method was demonstrated to be able to identify an odorous molecule
from its structure, and according to the researchers, “this guides us to
discover new classes of odorants.”
The researchers claimed to have gathered the largest-ever set of
information on odor descriptors for various compounds for the studies.
They employed human subjects to identify the odors of four hundred compounds using 55 different descriptions for the sake of comparison. According
to the research team, the neural network model consistently outperformed human
volunteers in odor recognition tests.
The group claimed that numerous industries, including those in
the food, fragrance, and healthcare sectors, may benefit from such a tool.
The instrument, according to the experts, can be used to combat
one of the largest health issues facing the globe today: diseases brought on by
insects. They claimed that the technique can gauge how well a chemical repels
mosquitoes from humans.
The scientists claimed to have found a number of novel compounds
that are at least as effective as DEET (DEET, also known by its chemical name N, N-diethyl-meta-toluamide, is the main component in many repellents. It is frequently used to ward off biting insects like ticks and mosquitoes), a substance frequently found in insect
repellents, at keeping mosquitoes away.
The researchers believe that this finding could lead to the development of repellents that are more affordable, long-lasting, and secure than those that contain DEET. They might also be employed “to lessen the global occurrence of diseases like malaria, ultimately preserving countless lives,” the scientists continued.
AI to Recognize and Predict Smells by Reviews & Views SSAAZS