 A perfumer is an artist who attempts to describe social motion and attitude through our sense of smell. Creating an original perfume is a time consuming process and can take an experienced individual months or years. The typical perfumer works in his lab and uses his knowledge of various oils and exotic scents to create, through trial and error, a blend of several dozen essences that will inspire a mood or feeling. The purpose of this proposal is to demonstrate that recently developed software in the field of artificial intelligence can reshape the way that fragrances are generated. To test this idea, we selected a number of oils that have a wide range of scents. The first step in our algorithm is to generate a random combination of oils from this set. After 10 combinations were prepared, we evaluated their aromas and selected the top three. A genetic algorithm then selectively mutated these top choices and generated a new set of 10 combinations. After six generations the result was a distinct, pleasing perfume. This project is obviously a blend of both art and science, and it allowed us to explore the art of making a perfume in a way that no one has ever done before.
Photograph: Emily Taylor |
