AI sommeliers recognise the taste and quality of beer better than human tasters

The third set of data was taken from the commercial sector, containing information about the prices of different beers, sales statistics, seasonal changes, etc. The AI was tasked with independently deriving average estimates, because it is impossible to know what exactly guided buyers in their choice in each specific case. As a result, after training on all types of data, the AI began to perceive beer as a complex and important product for people and to evaluate samples based on this approach. All the conclusions of the already trained AI matched or surpassed in accuracy the assessments from professional tasters.