AI-Optimized Beers Score Higher in Taste Tests
-
Belgian researchers used machine learning to predict how changing a beer's composition would alter its taste. The model was trained on 180,000 online reviews and panel tastings.
-
The AI suggested modifications to commercially available beers. When tasted, the reformulated beers scored higher with the human panelists.
-
AI could help improve the taste of existing beers or make bad beers better.
-
The researchers also created non-alcoholic beers using the AI that tasted indistinguishable from regular beer.
-
The author notes that studying beer may not have been approved research topic in university, but wishes they had tried.