THIS NEURAL NET MAPS MOLECULES TO AROMAS

Summary from the TLDR newsletter

Researchers have created a model that maps chemical structure to odor descriptors using a type of deep learning algorithm called a graph neural network. The model can successfully predict how humans would describe new smells. It could potentially be used to digitize smells. The Principal Odor Map is open source. A link to the repository for the project is available in the article.