Whether you prefer a fruity Lambic or a complex Trappist, Belgian beers have long been known for their variety, quality and heritage. Now researchers say they’ve harnessed the power of artificial intelligence to make brews even better.
Professor Kevin Verstrepen, from KU Leuven University, who led the research, said AI could help tease apart the complex relationships involved in human odor perception.
“Beer – like most food products – contains hundreds of different aroma molecules that are picked up by our tongue and nose, and our brains then integrate them into one image. However, the compounds interact with each other, so how we perceive one also depends on the concentrations of the other,” he said.
Writing in the journal Nature Communications, Verstrepen and his colleagues report how they analyzed the chemical composition of 250 commercial Belgian beers in 22 different styles, including lagers, fruit beers, blond beers, West Flemish ales and non-alcoholic beers.
Properties examined included alcohol content, pH, sugar concentration and the presence and concentration of more than 200 different compounds involved in flavor – such as esters produced by yeasts and terpenoids from hops, both of which are involved help create fruity notes. .
A tasting panel of 16 participants tasted and rated each of the 250 beers on 50 different characteristics, such as hop flavor, sweetness and acidity – a process that took three years.
The researchers also collected 180,000 reviews of various beers from the online consumer rating platform RateBeer, finding that while the beers’ ratings were influenced by attributes such as price, meaning they differed from the tasting panel’s ratings, the ratings and comments regarding other characteristics such as bitterness, sweetness, alcohol and malt aroma – these corresponded well with those of the tasting panel.
“Small changes in chemical concentrations can have a big impact, especially when multiple components start to change,” says Verstrepen, adding that one surprise was that some substances traditionally known to be a turn-off can be positive if they are are present in lower concentrations. and occur in combination with other aroma compounds.
Using the various data sets, the team constructed models based on machine learning – a form of AI – to predict what a beer would taste like and its rating based on its composition.
They then used the results to improve an existing commercial beer, essentially fortifying it with substances identified by the models as important predictors of overall rating – such as lactic acid and glycerol.
The results from the tasting panel showed that the additions improved ratings for both alcoholic and non-alcoholic beers in areas including sweetness, body and overall rating.
While the models have limitations, including the fact that they were only developed using datasets based on high-quality commercial beers, Verstrepen said their biggest application could be in tweaking non-alcoholic beers to make them better .
But beer lovers need not worry that new technology could disrupt a rich heritage, with Verstrepen noting that brewers’ skills remain crucial.
“The AI models predict the chemical changes that can optimize a beer, but it is still up to the brewers to make that happen, based on the recipe and brewing methods,” he said.