You know how you can look at a rainbow and name the colors? Or listen to music and say whether the note is high or low?
We can measure these things because we understand how light and sound work. But what about smell?
Until now, telling how something would smell just by looking at its chemical makeup has been almost impossible.
Professor Jane Parker from the University of Reading says, “We’re really good at understanding things like color and sound because we can measure them.
But smell? Not so much.
Sometimes two things can have the same structure but smell different. Or they can have different structures and smell the same! It’s confusing.”
That’s all changing thanks to an amazing breakthrough: scientists have created an ‘AI nose.’ This computer program uses machine learning, a kind of smart computer thinking, to predict how things will smell based on their molecules. And it’s not just a guess; it’s very accurate!
Why is this a big deal? Imagine you’re a person who creates new flavors for food or scents for perfumes.
Until now, you’ve had to rely mostly on your own nose to figure out what smells good or bad. It’s a lot of trial and error.
But what if a computer could tell you in advance whether a new scent will be a hit or a miss? That’s a game-changer!
This ‘AI nose’ can predict the smell of many different kinds of molecules, even those that have stumped experts before.
It can do this because it uses an ‘odor map.’ This map is like a big chart that shows what different smells are related to what kinds of molecules. And because the computer can learn, the more smells it studies, the better it gets at predictions.
Professor Parker collaborated with experts from the Monell Chemical Senses Center, Arizona State University, and a company called Osmo. Their work made sure that the computer program was really smelling what they thought it was smelling.
They said, “We tested a lot of samples. We wanted to make sure that if something smelled like butter, it was because it was supposed to smell like butter, not because something else got mixed in by accident. And once we cleared that up, the AI was really, really good at predicting smells.”
So what’s next? This computer nose could be huge for industries that make food, drinks, and perfumes.
But it’s not just them. Think bigger. What if we could use this technology to find out new smells we didn’t even know we liked?
Just like companies search for new medicines by testing lots of different molecules, we could do the same for new scents and flavors. We could explore thousands or even millions of new smells. As Professor Parker said, “This opens up a whole new world of smells we’ve never even thought of!”
In short, the ‘AI nose’ is not just a cool piece of tech; it’s a nose-opening experience that could change the way we think about smell forever!
So the next time you catch a whiff of something and wonder, “What’s that smell?” there might just be a computer program that can tell you.
The research was published in Science.
Follow us on Twitter for more articles about this topic.
Source: University of Reading.
