AI System Predicts Smells of Molecules, Opening Door to Digitizing Odors

Scientists have created an AI system that can predict what a molecule will smell like based on its chemical structure. The AI produced an "odor map" showing relationships between smells.
The AI was able to describe smells of new molecules, often better than humans in a test panel. This shows progress towards digitizing and synthesizing smells.
Smell is complex and personal, tied to memory and culture. So there are limitations to AI-predicted smell relationships.
The AI currently only looks at individual molecules. But most natural smells are blends of many chemicals, which is the next challenge.
This "odor map" is the first step towards potential future applications like smell-detection in machines and on-demand digitized smells.

popsci.com

Relevant topic timeline:

8/22/2023

AI Can Now Design Proteins That Behave Like Biological 'Transistors'

Scientists have used AI to design proteins with two different states, essentially creating biological transistors that can change their shape depending on inputs, opening up new possibilities for biotechnology and medical solutions.

8/23/2023

A map of every possible molecule could be possible with AI

Chemists are developing a chemical map of all possible molecules to accelerate the discovery process for drugs and materials, with the help of artificial intelligence to determine the properties and viability of these molecules.

8/31/2023

AI system does better than humans at identifying odors

A computer-trained scent model has outperformed humans in identifying odors and has created a Principal Odor Map, providing insights into how our brains and noses perceive smells.

9/1/2023

Could AI be the spark that ends the opioid epidemic?

Artificial intelligence (AI) is being explored as a potential solution to end the opioid epidemic, with innovations ranging from identifying at-risk individuals to detecting drug contamination and reducing overdoses, but concerns about discrimination and misinformation must be addressed.

9/2/2023

AI 'nose' predicts smells from molecular structures

Scientists have developed a machine learning model that can predict the odor profile of a molecule based on its structure, potentially revolutionizing the food, fragrance, and synthetic chemistry industries.

9/4/2023

Machine learning for chemistry: Basics and applications

A review published in Engineering explores the potential of machine learning (ML) in revolutionizing chemical research, providing insights into popular ML algorithms and their applications in chemistry.

9/13/2023

AI Accelerates Scientific Discovery Across Fields

AI is revolutionizing scientific research by accelerating drug discovery, predicting protein structures, improving weather forecasting, controlling nuclear fusion, automating laboratory work, and enhancing data analysis, allowing scientists to explore new frontiers and increase research productivity.

9/26/2023

New AI Algorithm Detects Signs of Life on Mars with 90% Accuracy

A newly developed method based on artificial intelligence (AI) is capable of detecting subtle differences in molecular patterns that indicate biological signals, even in samples hundreds of millions of years old, with 90% accuracy, potentially revolutionizing the search for signs of life on other planets.

10/1/2023

Perfumers Tap AI and Neuroscience to Engineer Emotional Responses Through Scent

Perfumiers are using AI and neuroscent research to develop fragrances that trigger specific emotional responses, with companies like L'Oréal and Puig using brain readings and biometric research to create scents that make consumers feel good.

Topics

Posted 9/11/2023, 3:00:00 PM