AI that identities the odours of chemicals

Being able to digitally build chemicals that are pleasant smelling opens up a huge range of possibilities in the fragrance, food and household goods industries

Spotted: Two researchers from the University of California Riverside Department of Molecular, Cell and Systems Biology have created an algorithm capable of identifying the odour of a chemical. The machine learning system examined many thousands of combinations of compounds in order to identify what a chemical smells like to humans.

Being able to digitally build chemicals that are pleasant smelling to humans opens up a huge range of possible applications in the fragrance, food and household goods industries. Good smelling scents could be built from scratch, using previously unknown combinations, which also allows for the inclusion of new ingredients. Harsh smelling and dangerous chemicals are replaceable by naturally occurring materials and compounds, and vast improvements to resourcing are also possible. The need to use rare, hard to cultivate plants could be eliminated with the digital versions instead.

AI startup Sensorygen licensed the technology and a patent is pending. Retail, gaming and myriad other areas of work and play may find uses for the discovery, particularly as multi-sensory experiences continue to grow in popularity. Artificial intelligence is being used more and more and in a range of ways. Some of Springwise’s recent spottings of AI at work include the creation of bespoke, sustainable jeans and a genderless voice for digital assistants.

Explore more: Health and Wellbeing Innovations | Computing & Tech Innovations