Developing new products or improving existing recipes to cater to consumer preferences can be a challenging task for food and drink manufacturers. However, utilizing AI models can significantly help in this regard. By training machine-learning models on comprehensive data sets that include both chemical composition and sensory features, manufacturers can accurately predict a product’s taste, aroma, mouthfeel, and overall quality. This can save a significant amount of time and money that would have otherwise been spent on conducting trials.
Over the course of five years, researchers analyzed 250 commercial beers to measure their chemical properties and flavor compounds. By combining these detailed chemical analyses with assessments from a trained tasting panel and reviews from an online platform, the researchers created a comprehensive data set. This data set was used to train 10 machine-learning models that were designed to accurately predict a beer’s taste, aroma, mouthfeel, and overall quality, as well as predict how likely consumers would be to rate it highly.
Incorporating a wide range of data sources allowed the researchers to develop sophisticated models that could help enhance the beer production process. By utilizing AI models in this way, food and drink manufacturers can improve their products while saving time and money in the process.
