This is not an easy task:
The artificial intelligence model was trained on 5,000 odor molecules from a catalog of fragrances to recognize which aromatic compositions match verbal descriptions.
Based on the data obtained, scientists developed a scent map that, based georgia phone number resource on molecular composition, classified aromas into categories: fruity, musky, meaty, and floral.
To test how the machine learning software worked, it was fed 400 odor molecules synthesized by scientists but not found in nature. The model was then asked to predict what the resulting compounds would smell like.
To assess the accuracy of the artificial intelligence, 15 volunteers sniffed and described each odor.
It turned out that the model's guesses were more accurate than those of the average taster in most cases.
Why all this? Good question. Osmo offers several exceptional applications:
The most obvious is the $60 billion global perfume market. Osmo's technology could help companies replicate scents that are typically produced by industrially harvesting rare plants or harming animals.
Mosquito-borne diseases like malaria and dengue fever kill over 440,000 people each year. Osmo tested 400 scents and selected over 10 that were as effective as or better than the best repellents.
The company's long-term goal is to learn how to identify diseases by smell.
Osmo could eventually become a kind of “Shazam for smells,” allowing users to recognize and store smells on their smartphones.
Well, all that remains is to wait for someone to figure out how to taste these mouth-watering recipe photos from our feed.
