Olfaction — Machine Learning
Human Sensing and Neurotechnology
2024 — 2025
In collaboration with researchers across Yale, Cornell, KTH, and other institutions, I worked on modeling how people perceive the similarity between complex odor mixtures.
A semantic-based community model for high-fidelity tuning of olfactory mixture distances
This work introduces a machine-readable, quantitative metric for olfactory mixture similarity. It opens a path toward more precise digital olfaction, fragrance design, electronic noses, and olfactory foundation models.
Community challenge
1st overall
Our team, D2Smell, ranked first by average ranking across the challenge metrics. The broader effort brought together 26 international teams and developed a unified model from the strongest community contributions.