UF creates a computer program that can taste and help farmers

It’s a computer that can taste, essentially. The program is used to determine which variety of fruit has the best flavor.

GAINESVILLE, Fla. — The robots are taking over. We joked about the smell of a vision, but we could get closer to a computer that can taste.

The focus, however, is for farmers and consumers.

Researchers at the University of Florida’s Institute of Food and Agricultural Sciences have created a computer that can taste, a certain way. The “artificial intelligence geek” can break down food into its chemical compounds and tell you if it will taste good.

“Taste is a very individual and particular opinion,” said Dr. Marcia Resende. “For me, I like high-sugar fruits in general.”

Resende is an assistant professor and plant breeder at the University of Florida. He was an integral part of the creation of the computer model.

“What we wanted was to develop a method that allowed for objective scoring or identification to help plant breeders identify the tastiest variety,” he said.

The program can “taste” by analyzing the chemical compounds in fruits and determining their flavor.

Resende says that in general, plant breeders need to taste varieties in the field or have a panel of 50 to 100 people to taste. But all this is subjective and slow.

Their computer model can filter 70 chemical compounds to determine if consumers, like you, will like it. He says they can also test hundreds of varieties per day.

They use the UF supercomputer to run the program. Resende says it takes a few hours to get the data, but if you try to run the program on your computer, it could take a week.

“Flavour is associated with the consumption of fruits and vegetables. The better a strain tastes, the more people tend to return to it and consume it again,” Resende said. “So we hope this will increase overall fruit and vegetable consumption in the long term.

Aside from the volatile chemicals in fruit, Resende says aroma plays a major role in evaluating a fruit’s flavor.

They are currently using this technology on blueberries, strawberries, corn, tomatoes and citrus fruits as part of UF’s IFAS plant breeding program.

Gordon K. Morehouse