The ChemSenSim Team Honored in the DREAM Olfactory Mixtures Prediction Challenge
- Research
Published on September 16, 2026
–
Updated on September 16, 2026
Dates
on the September 7, 2026
A team from the Nice Institute of Chemistry (Université Côte d’Azur, CNRS) is among the winners of the international DREAM Olfactory Mixtures Prediction Challenge. Their work, published in PNAS in July 2026, paves the way for a better understanding and modeling of complex odors.
Predicting whether two odors are similar is a major scientific challenge. Unlike vision or hearing, olfaction relies on complex mechanisms that are difficult to model.
As part of the international DREAM challenge, the ChemSenSim team from the Nice Institute of Chemistry distinguished itself by tying for first place.
Its innovative method uses artificial intelligence to replicate the functioning of the human olfactory system. It combines graph neural networks and a so-called “Siamese” model to compare odor mixtures. Ultimately, these advances could transform the design of perfumes, flavorings, and olfactory substitutes.
This research, conducted by Matej Hladiš and Maxence Lalis during their doctoral studies, was published in the journal PNAS in July 2026
.
As part of the international DREAM challenge, the ChemSenSim team from the Nice Institute of Chemistry distinguished itself by tying for first place.
Its innovative method uses artificial intelligence to replicate the functioning of the human olfactory system. It combines graph neural networks and a so-called “Siamese” model to compare odor mixtures. Ultimately, these advances could transform the design of perfumes, flavorings, and olfactory substitutes.
This research, conducted by Matej Hladiš and Maxence Lalis during their doctoral studies, was published in the journal PNAS in July 2026
.
