Scientists develop new wine-tasting AI technology
Researchers at the Maryland-based National Institute of Standards and Technology (NIST) have created artificial intelligence capable of virtually “tasting” wine with more than 95% accuracy.
The new AI technology developed by U.S. scientists has passed a wine tasting test with great gusto, demonstrating a “more efficient but less fun” method of tasting wines, according to its manufacturers.
The team began training the virtual palate of the AI system, using 148 of the wines from a data set of 178 made from three grape varieties.
Each virtual wine had 13 characteristics to consider, including alcohol level, color, flavonoids, ash, alkalinity, and magnesium.
Each characteristic was assigned a value between 0 and 1 so that the network took into account when distinguishing one wine from the others.
Brian Hoskins, a NIST physicist, said of the experiment: “It’s a virtual wine tasting, but the tasting is done with an analytical team that is more efficient but less fun than tasting it yourself.”
Then the technology had to be tested. Tt was given a virtual wine tasting test with the complete dataset, which included 30 wines he had never seen before.
The system passed with 95.3% accuracy and only made two mistakes in the wines it had never seen before, which the researchers saw as a good sign.
“Getting 95.3% tells us this works,” NIST physicist Jabez McClelland said.
Wine tasting AI technology has not been designed as an artificial sommelier, but to demonstrate that similar devices could be expanded and used to build new AI systems.
These systems, which use less power than previous forms, could be used in applications such as “smart” clothing, miniature drones, or sensors that process data at the source.
An Australian electronic nose prototype has previously been developed to detect the difference between cheaper and more expensive whiskeys in less than four minutes. Read more about this here.