Are the roles of sommelier and winemaker in danger of extinction? According to the results of a study conducted at the University of Geneva in Switzerland and published a few days ago, the answer may be yes. The study revealed that an AI-based robot could identify different wines and qualities at a hundred percent level.
As part of the study, the machine was able to identify 80 red wines produced in 12 different vintages between 1990 and 2007. These wines came from seven different wineries in the Bordeaux area.
Sunset on vineyards in France, photo: Getty Images
As part of the study, the researchers operated a machine that knows how to break down the various compounds in wine into their various components. This technique identified about 30,<> factors that affect the final product of the wine in terms of taste and quality. "We wanted to see if there was a different chemical signature for each winery, that is, a different identity for each winery," explained Alexander Pozha, the study's leader.
Later, the researchers fed the algorithm 73 of the results given about the different types of wine. They also added to this the information about the year of wine production and the winery from which it comes from in the Bordeaux region. This process was performed 50 times, each time the algorithm received information about other wines. Now the AI has been asked to identify different types of wine on its own.
Red wine, photo: Getty Images
The results were amazing, AI technology identified 100 percent of the time the winery where the wine was made. "Such an astonishing result is something that very few humans are capable of doing," he noted. In addition to identifying the source of the wine, it also turned out that the algorithm can accurately guess the wine's year of production. In fact, it turned out that artificial intelligence could identify the winery where the wine was made, even if it received only 50 percent of the information collected about each wine in the process of breaking down its materials.
The artificial intelligence could not only identify the winery where the wine was produced, but also how to associate groups of wines with a specific region in which they were produced. The computer could clearly identify the wines produced on the right bank of the Garonne River and the wines produced on its left bank. "It turns out that the unique taste and sensation that wine evokes in the mouth does not depend on a small number of key molecules, but on the concentration of a large number of factors," he noted.
In summarizing their remarks, the researchers noted that according to the research carried out by artificial intelligence, there is great significance to the climate, geography, method of wine production and bacteria found in it. The AI also identified that there is a significant impact on the type of soil in which the vines grow and the chemicals they absorb from that soil.
"There's really information here that almost definitively proves that where vines are grown and how wine is made has a unique chemical signal that creates similarities and differentiations about the different types of wines," said researcher Barry Smith of the University of London's School of Advanced Studies.
Wrong? We'll fix it! If you find a mistake in the article, please share with us