Artificial intelligence: can now predict the taste and quality of beer

Belgian scientists have developed models of Artificial intelligence that can predict how consumers will rate a beer and what compounds can be added to improve it.

The research was published in Nature Communications and may revolutionize the way the food and beverage industry develops new products, according to the authors.

Comparing and ranking the flavor profiles of different beers is a challenge. There are plenty of guides on the market that describe drinks with generic terms like “fruity” and classify them based solely on a person’s taste. “This makes beer comparisons very biased and makes it difficult to predict the actual taste of a beer,” says Kevin Verstrepen, professor at KU Leuven and director of the VIB-KU Leuven Center for Microbiology and the Research Institute of Microbiology, in a statement. Leuven Beer.

“I wanted to have a more neutral and scientific description of the different beers in the world,” says Verstrepen.

250 types of beer in five years

The team began chemically analyzing the beers, carefully measuring the concentrations of hundreds of aromatic compounds. A trained panel of 15 people evaluated each beer based on a set of 50 criteria. “It was a truly herculean effort. We started the project with less than 100 beers and quickly realized that it would not be enough to capture the incredible beer diversity of Belgium, so we ended up analyzing 250 beers,” recalls Dr. Miguel Roncoroni, who led the chemical and panel analyses. tasting.

It took the team five years, but once they had chemical concentrations and detailed tasting reports for hundreds of different beers, they knew it would be possible to use AI to connect the two. A model could predict key aromas and the final appreciation score of a beer without the need for human tasting.

These results, in turn, were used to improve the flavor of an existing commercial Belgian beer by adding certain aromas predicted by the model to improve the quality of the beer. Indeed, the modified beer scored significantly better in the blind tastings.

“The flavor of beer is a complex mixture of aromatic compounds. It is impossible to predict how good a beer is by simply measuring one or several compounds. We really need the power of computers,” says Michiel Schreurs.

The study can be expanded to other food products, which could revolutionize the way new foods are made.

“Our biggest goal now is to make better non-alcoholic beer. Using our model, we have already managed to create a cocktail of natural aromatic compounds that mimic the taste and smell of alcohol without the risk of a hangover,” says Verstrepen.