Belgian researchers: Artificial intelligence should make beer taste better

class = “sc-cffd1e67-0 iQNQmc”>

According to Belgian researchers in the journal “Nature Communications”, the new method could help food manufacturers meet specific consumer demands more efficiently and cost-effectively.

Researchers led by Kevin Verstrepen of the Catholic University of Leuven explained that predicting which new food flavors consumers will like is often complicated. This is mainly due to the presence of many taste-active chemicals in foods. There are also interactions and complex, reinforcing or reducing effects on taste perception. For example, sweetness and bitterness were masking each other.

The use of trained tasters is common, but this causes high costs. On the other hand, online rating databases are prone to errors because factors such as a product’s price or current cult status are also reflected there.

Scientists have now recorded more than 200 chemical properties of 250 Belgian beers from 22 different beer styles, including Blond, Tripel and Lager. These were linked to descriptive sensory profile data on hops, malt and yeast aromas, off-odors and spices from a trained tasting panel of 16 people, as well as data from more than 180,000 consumer reviews from an online beer ratings database.

The resulting dataset was then used to train and test ten machine learning models that predict consumer liking and disliking. The effectiveness of the most powerful AI approach was tested by applying predictions to modify alcoholic and non-alcoholic commercial beer. In tastings, AI beers were said to receive a better overall score from testers.

The study confirms that the concentration of flavor compounds does not always correlate with perception; this suggests complex interactions that traditional approaches often overlook.

The research team says commonly used models are still immature. More comprehensive data sets are crucial for further improvements. It should also be clear that models can only recognize correlations, not causal relationships.

Still, the team believes AI could be a basis for developing new, personalized foods that taste great.

(SDA)

Source : Blick