How can beer taste and popularity be improved using AI analytics? In a large-scale Belgian study, AI was used on 250 beers.

In no other country do people take the taste of beer as seriously as in Belgium. Researchers from KU Leuven and the Flemish Institute of Biotechnology had an AI model taste 250 beers. The aim of the research was to use an analysis of chemical and non-chemical data to predict how to predict and improve the taste of a beer. The research was published in the journal Nature.

The researchers created a detailed description of the chemical composition of each beer. They then also collected the opinions of a test panel of fifteen beer lovers and 180,000 internet reviews. Special AI models were then used for all this data.

About taste can not be argued

Brewing beer is a chemical process, but chemical composition is just one of the factors that determine whether a beer tastes good. Conditions such as the temperature at which the beer is stored and served or how long the beer is in the cellar affect the taste. Of course, taste is also a personal matter. This sum makes it difficult for brewers to predict how well a new beer will perform.

It turned out that the AI ​​models were pretty good at predicting whether the test panel would like a beer. The next step was to investigate whether the models could also make suggestions for improving the taste of a beer. They added some flavors that the AI ​​thought would make the beer taste better. It turns out: The test panel rated the “new” beer better.

Kevin Verstrepen, a researcher at KU Leuven, believes the research can be a good starting point for making better non-alcoholic beer. “With our model, we have already managed to create a cocktail of natural flavorings that imitates the taste and smell of alcohol without the risk of a hangover,” he told VRT NWS.