DeepNose: An Equivariant Convolutional Neural Network Predictive Of Human Olfactory Percepts
2024
DeepNose: A Neural Network for Predicting Smells
publication
Author Information
Author(s): Shuvaev Sergey, Tran Khue, Samoilova Khristina, Mascart Cyrille, Koulakov Alexei
Primary Institution: Cornell University
Hypothesis
Olfactory receptors can be viewed as 3D spatial filters that extract molecular features relevant to the olfactory system.
Conclusion
The DeepNose network can predict human olfactory percepts with high fidelity and identify molecular features responsible for odor quality.
Supporting Evidence
- The DeepNose network produced responses that are approximately invariant to the molecules' orientation.
- It offers high-fidelity perceptual predictions for different olfactory datasets.
- The architecture allows for inferring the perceptual quality of odor mixtures.
Takeaway
The DeepNose is like a smart computer that helps us understand how we smell things by looking at the shapes of molecules.
Methodology
A convolutional neural network was trained to predict human olfactory percepts from several semantic datasets.
Want to read the original?
Access the complete publication on the publisher's website