Belgian beer is renowned for its diverse flavors and rich history, from fruity lambics to complex Trappist ales. Now, researchers are using artificial intelligence to enhance the taste of beer.
Professor Kevin Verstreppen led the research at the University of Leuven, exploring how AI can help understand the complex relationships in human scent perception.
In a study published in Nature Communications, Versträpen and his team analyzed the chemical composition of 250 Belgian beers across 22 styles. They looked at properties like alcohol content, sugar concentration, and over 200 flavor compounds.
A tasting panel rated these beers on various characteristics over three years. The researchers also analyzed consumer reviews from RateBeer, finding differences in ratings based on factors like price.
Small changes in chemical concentrations can greatly impact flavor. Using machine learning, the team created a model to predict beer taste and ratings based on composition.
By adding specific substances identified by the model, commercial beers saw improved ratings for sweetness and overall quality. This could benefit both alcoholic and non-alcoholic beers.
Despite the potential of AI, Versträpen emphasizes the importance of the brewer’s craft in maintaining beer tradition. The model aids in optimizing beer, but it’s the brewer who brings the recipe to life.
Source: www.theguardian.com
