Digital Description of Aroma

There’s a new artificial intelligence system called a graph neural network, that is capable of predicting what a compound will smell like to a person – rose, medicinal, earthy, and so on – based on the chemical features of odor molecules.

 

According to the researchers (Brian K Lee, Emily J Mayhew, Benjamin Sanchez-Lengeling, Jennifer N Wei, Alexander B Wiltschko, et al), the descriptions the AI generates – the AI can describe the aroma of roughly 5,000 odorants – are similar in language and often more precise than humans. The AI also analysed each odorant's chemical structure to determine the relationship between structure and aroma. 

 

The system identified around 250 correlations between specific patterns in a chemical's structure with a particular smell. The researchers combined these correlations into a principal odour map (POM) that the AI could consult when asked to predict a new molecule's scent. 

 

Eventually, the scientists from Google Research and the spinoff start-up Osmo believe the system could deepen our understanding of how the brain processes odors. And it could also lead to the design of new kinds of scents for commercial products. 

 

Note: The paper that describes the research, “A principal odor map unifies diverse tasks in olfactory perception” can be found in Science, August 31, 2023, Vol 381, Issue 6661, pp. 999-100.

 

The average human nose contains about 350 types of olfactory receptors, which can bind to a potentially enormous number of airborne molecules. These receptors then initiate neuronal signals that the brain later interprets a whiff of coffee, gasoline or perfume. Although scientists know how this process works in a broad sense, many of the details – such as the precise shape of odor receptors or how the system encodes these complex signals – still elude them.

 

Stuart Firestein, an olfactory neuroscientist at Columbia University, calls the model a “tour de force work of computational biology”. But as is typical for a lot of machine-learning-based studies, “it never gives you, in my opinion, much of a deeper sense of how things are working”, says Firestein, who was not involved in the paper. 

 

His critique stems from an inherent feature in the technology: such neural networks are generally not interpretable, meaning human researchers cannot get access to the reasoning a model uses to solve a problem.

 

What’s more, this model skips over the inscrutable workings of the nervous system, instead making direct connections between molecules and smells. Even so, Firestein and others describe it as a potentially useful tool with which to study the sense of smell and its fraught relationship with chemistry. 

 

To the researchers involved, the model represents a move toward a more precise, numbers-based means to describe the odor universe, which they hope could eventually bring this sense to the digital world.