Google's artificial intelligence model can predict odour like human beings
We know insects including mosquitos are drawn to humans by sensing the smell. Using the artificial intelligence model, scientists hope to predict these animal sense of smell to better respond to the deadly diseases transmitted by mosquitoes and ticks.
Google has built an artificial intelligence model with human like capability of predicting odour. The map developed by the team of Google AI links molecule structure to the aroma of substance and can even predict smell that is still unnoticeable by humans.
Why is it hard to plot smell?
Smells are sensed when molecules riding on the air stick to the sensory receptors present in the nose. However, it is more difficult to predict smell than colour. Because, unlike the human eye, which has only three sensory receptors for sensing the photons of red, green and blue colour, our nose has over 300 receptors.
Molecules vary in many more ways than photons do. It is challenging to draw an odour map to plot distinct smells, while a colour wheel to show different colours has existed for centuries.
What does the research say?
In 2019, with a deep learning algorithm, the scientists started exploring the interplay of molecular structure with the type of smell. A graph neural network (GNN) model was trained to learn numerous examples of particular molecules paired with the smell labels that they arouse, e.g., “beefy”, “floral”, or “minty”.
Now, scientists have achieved success in generating “Principal Odour Map” (POM) with properties of a sensory map. They tested the model on various parameters.
The scientists tried to find out if the model has learned to predict the odours of new molecules that humans have never smelled and also that differed from molecules used to train the GNN model. “This is an important test — many models perform well on data that looks similar to what the model has seen before, but break down when tested on novel cases,” researchers wrote in the Google post.
The model stood on the test and presented exceptional intelligence to predict odour from a molecule’s structure.
Then, the scientists tested if the map is versatile enough to predict odour perception in animals. They found the map could well predict the activity of sensory receptors, neurons, and behaviour in most animals.
This finding establishes olfaction is related to our natural world through the structure of metabolism and associates fundamental principles of biology. The map even applies to species separated by hundreds of millions of years of evolution.
Using this finding to tackle global health problem
The map of odour closely connected to perception and biology across the animal kingdom opens the scope of many advancements in medical sciences. We know insects including mosquitos are drawn to humans by sensing the smell. Using the POM, scientists hope to predict these animal olfaction to better respond to the deadly diseases transmitted by mosquitoes and ticks. “Less expensive, longer lasting, and safer repellents can reduce the worldwide incidence of diseases like malaria, potentially saving countless lives,” the researchers aim.
This research dealt with a range of questions about new smells and the molecules that generate them. It related smells back to their origins in evolution and nature. Now scientists believe that this method can leverage fresh solutions to obstacles in food and fragrance formulation, environmental quality monitoring, and the detection of human and animal diseases.
Personalise your news feed. Follow trending topics