As generative artificial intelligence (AI) continues to advance and bring new possibilities, researchers have just developed an AI model that can independently create new custom fragrances based on user-defined descriptors.
Indeed, a research team at the Institute of Science Tokyo has created an AI model that relies on mass spectrometry profiles of essential oils and corresponding odor descriptors to generate essential oil blends for new aromas, according to a report published by Tech Xplore on April 23.
This way, the team has automated the process of designing new scents in perfume, food, and home product industries, which is currently time-consuming, labor-intensive, and reliant on the skill and expertise of specialized perfumers, requiring countless trial-and-error attempts to create a particular fragrance.
How generative AI designs custom fragrances
Specifically, the researchers, led by Professor Takamichi Nakamoto, have designed an AI model called Odor Generative Diffusion (OGDiffusion), which uses generative diffusion networks, a machine learning model that learns to create new content by reversing a noise process informed by existing data.
While these models are generally useful in creating images and text, the researchers have now succeeded in adapting the technology to create new aromas by analyzing the chemical profiles of 166 essential oils labeled with nine odor descriptors, like ‘sweet’ or ‘floral.’
Then, users specify desired fragrance characteristics and AI generates a corresponding chemical profile in line with those descriptors and calculates the mix of essential oils needed to recreate the scent using a mathematical method called non-negative least squares.
Commenting on the breakthrough, Professor Nakamoto explained:
“Our diffusion network uses patterns in mass spectrometry data of essential oils to generate new fragrance profiles in a fully automated, streamlined, and data-driven approach while maintaining high-quality data output. By eliminating human intervention and molecular synthesis from the process, we provide a fast, general, and efficient method for fragrance generation.”
Upon testing, the new AI model correctly matched people’s expectations, as the participants were consistently and reliably able to identify the correct odor. The results were equally compelling in various additional tests, indicating that the model generates clearly identifiable aroma profiles.
Elsewhere, the first AI-powered fragrance house called Generation was launched in New York last month by Osmo, the first company to digitize fragrance, and powered by olfactory intelligence (OI) – a technology that turns data into scent.
What do you think?
Join Techgaged on Telegram
Get first-access to daily trending tech stories, AI breakthroughs, and more, before it hits your feed.
