Home Learn MIT Technology Review AI could make higher beer. Here’s how.

MIT Technology Review AI could make higher beer. Here’s how.

0
MIT Technology Review
AI could make higher beer. Here’s how.

Crafting a good-tasting beer is a difficult task. Big breweries select tons of of trained tasters from amongst their employees to check their recent products. But running such sensory tasting panels is dear, and perceptions of what tastes good could be highly subjective.  

What if artificial intelligence could help lighten the load? Latest AI models can accurately discover not only how highly consumers will rate a certain Belgian beer, but additionally what sorts of compounds brewers needs to be adding to make the beer taste higher, in line with research published in today.

These sorts of models could help food and drinks manufacturers develop recent products or tweak existing recipes to higher suit the tastes of consumers, which could help save quite a lot of money and time that may have gone into running trials. 

To coach their AI models, the researchers spent five years chemically analyzing 250 industrial beers, measuring each beer’s chemical properties and flavor compounds—which dictate the way it’ll taste. 

The researchers then combined these detailed analyses with a trained tasting panel’s assessments of the beers—including hop, yeast, and malt flavors—and 180,000 reviews of the identical beers taken from the favored online platform RateBeer, sampling scores for the beers’ taste, appearance, aroma, and overall quality.

This huge data set, which links chemical data with sensory features, was used to coach 10 machine-learning models to accurately predict a beer’s taste, smell, and mouthfeel and the way likely a consumer was to rate it highly. 

To match the models, they split the information right into a training set and a test set. Once a model was trained on the information inside the training set, they evaluated its ability to predict the test set.

The researchers found that each one the models were higher than the trained panel of human experts at predicting the rating a beer had received from RateBeer.

Through these models, the researchers were in a position to pinpoint specific compounds that contribute to consumer appreciation of a beer: people were more more likely to rate a beer highly if it contained these specific compounds. For instance, the models predicted that adding lactic acid, which is present in tart-tasting sour beers, could improve other forms of beers by making them taste brisker.

“We had the models analyze these beers after which asked them ‘How can we make these beers higher?’” says Kevin Verstrepen, a professor at KU Leuven and director of the VIB-KU Leuven Center for Microbiology, who worked on the project. “Then we went in and truly made those changes to the beers by adding flavor compounds. And lo and behold—once we did blind tastings, the beers became higher, and more generally appreciated.”

One exciting application of the research is that it may very well be used to make higher alcohol-free beers—a serious challenge for the beverage industry, he says. The researchers used the model’s predictions so as to add a mix of compounds to a nonalcoholic beer that human tasters rated significantly higher by way of body and sweetness than its previous incarnation.

This kind of machine-learning approach may be enormously useful in exploring food texture and nutrition and adapting ingredients to suit different populations, says Carolyn Ross, a professor of food science at Washington State University, who was not involved within the research. For instance, older people are likely to find complex combos of textures or ingredients less appealing, she says. 

“There’s a lot that we are able to explore there, especially after we’re different populations and attempting to give you specific products for them,” she says.

LEAVE A REPLY

Please enter your comment!
Please enter your name here