Belgian Scientists Harness AI to Enhance Beer Flavour

Researchers from Belgium have employed artificial intelligence to enhance the flavour of beer.

They developed machine learning models to predict and improve the beverage’s characteristics.

The study’s lead, Professor Kevin Verstrepen, told The Guardian that AI aids in understanding the complex interactions of human receptors.

“Beer contains hundreds of different aromatic molecules that are detected by our senses. The brain then assembles them into a single picture. However, the compounds interact with each other, so the perception of one depends on the concentration of the others,” he added.

Verstrepen and his colleagues analysed the chemical composition of 250 commercial Belgian beers across 22 different styles.

The properties studied included alcohol content, pH, sugar concentration, and the presence of over 200 different compounds. Among them were esters produced by yeast and terpenes from hops, contributing to fruity notes.

The researchers also gathered 180,000 reviews of various beers from the online platform RateBeer. They found that ratings and comments on characteristics like bitterness, sweetness, strength, and malt aroma strongly correlated with tasters’ evaluations.

Sixteen panel members assessed each of the 250 beers on 50 different attributes. This process took three years.

Using the collected datasets, the team built machine learning models to predict taste and evaluate the beverages.

The researchers then applied the results to update the recipes of existing commercial brands. Tasting results showed that the neural network-induced changes significantly improved scores on certain flavour parameters.

In the summer of 2022, BarrelDAO launched sales of a limited edition NFT beer.