Google AI presents a sensory map generated by machine learning called “Principal Odor Map” (POM) that allows the prediction of odors from invisible molecules
The molecules disperse in the air, travel to our nostrils and bind to receptors in our nose to create odors. There are potentially billions of chemicals that can create an odor, making it difficult to categorize or predict which molecules create certain odors. Researchers have long believed that sensational maps would help find a solution to this problem. However, odor mapping is a difficult problem because molecules vary more than photons; data collection requires physical proximity between smell and smell, and the human nose has over 300 sensory receptors for smell.
A new Google search introduces a “master odor map” (POM), which locates the point in a high-dimensional space corresponding to the vector representation of each odor molecule in the model’s integration space. The POM works like a sensory map because pairs of perceptually similar odors are located very close to each other in the POM. POM can also find and anticipate new compounds that give off odors.
The team investigated the model’s ability to accurately predict the aromas of molecules that had never been sniffed before and that were very different from those used to train the model. To do this, they have compiled the most comprehensive database of odor descriptors for unknown compounds. Monell Center experts formed panels to assign fragrance ratings to 400 different compounds using 55 unique labels (such as “mint”) that were chosen to be comprehensive without being overly specific. Their findings show that the model’s predictions aligned more with the consensus than those of the average panelist. That is, the performance of the model in predicting odor from the structure of the molecule was well above average.
On other human olfactory tasks, including odor intensity and similarity detection, POM also demonstrated high-level performance. As a result, POM can be used to make educated guesses about the smell of billions of undiscovered chemicals, which has important implications for the flavor and fragrance industries.
Following the success of the master odor map in predicting human olfactory perception, the team investigated its potential for predicting olfactory perception and underlying brain activity in animals. According to the results, the map accurately predicted sensory receptors, neurons, and behavioral activity in most animals studied by olfactory neuroscientists. This includes rodents, flies and birds.
Researchers believe that one of the main reasons humans have a sense of smell is to be able to tell the difference between, say, a ripe apple and a rotten apple. They compiled information on the metabolic reactions of a wide range of species in all kingdoms of life, and the structure of the map turned out to be very similar to that of metabolism. The POM manages to encapsulate some of the most fundamental principles of biology while demonstrating the connection between olfaction and the organization of metabolism and the natural world.
Additionally, the team retrained the POM to combat one of humanity’s most serious problems: the epidemic of mosquito and tick-borne diseases. To do this, they incorporated data from two additional sources into their original model:
- An obscure set of experiments conducted by the USDA on human volunteers from the 1930s and recently discovered by Google Books
- A new set of data has been collected at TropIQ using their high throughput laboratory mosquito test. The effectiveness of a certain chemical in repelling mosquitoes is measured in both databases.
The resulting model allows a virtual screen over large regions of chemical space, predicting mosquito repellency of virtually any molecule. The team tested their model with completely new compounds and identified more than a dozen with repellency at least as strong as DEET, the main ingredient in most insect repellents.
In the future, the team hopes their work will advance understanding of how to formulate and flavor foods, monitor environmental quality, and identify human and animal diseases.
This Article is written as a research summary article by Marktechpost Staff based on the research paper 'A Principal Odor Map Unifies Diverse Tasks in Human Olfactory Perception'. All Credit For This Research Goes To Researchers on This Project. Check out the paper and reference article. Please Don't Forget To Join Our ML Subreddit
Tanushree Shenwai is an intern consultant at MarktechPost. She is currently pursuing her B.Tech from Indian Institute of Technology (IIT), Bhubaneswar. She is a data science enthusiast and has a keen interest in the scope of application of artificial intelligence in various fields. She is passionate about exploring new technological advancements and applying them to real life.