The future of food might not be in the kitchen. Thanks to research at IBM, the future of food could very well be controlled by your computer.
“In five years, computers will be able to construct never-before-heard-of recipes to delight palates – even those with health or dietary constraints – using foods’ molecular structure,” says Dr. Lav Varshney, a research scientist for IBM, in a new video for the company’s annual 5 in 5 campaign. This year’s campaign focuses on the five senses and a new era of computers based on cognitive systems.
Varshney chose to work with food because “It’s so visceral, it’s part of who we are. Everyone eats,” he said.
He and his six-person team, which includes a chef-turned-computer engineer, are working to design recipes that predict unexpected flavor combinations that taste delicious.
The goal of the program is to “design food that tastes good that is flavorful but is also healthy,” he said, “We’ve been using some ideas from culinary science and chemistry as well as psychology.”
The system looks to foods, such as strawberries, white wine or cooked apples, that are generally perceived as pleasant.
They use the flavor-pairing hypothesis to match flavors like black tea and bell peppers. The hypothesis is the idea that “in two ingredients that share a lot of flavor compounds, chemicals that make up the food, they typically go together in western cuisine.”
They hope to create more flavorful lunch meals for kids who generally skip their sandwich and opt for dessert.
“Various governments have imposed a lot of nutritional standards, school lunches have to meet these very stringent requirements and kids end up throwing away the food,” said Varshney.
It would also improve food options in hospitals and nursing homes as well as provide tasty recipes for those with food restrictions like food allergies.
But the system isn’t as simple as an app that would select random combinations. The team is also taking into consideration how the food looks. Creating recipes that have never been made before has required a bit of testing.
Varshney hopes that the machine will eventually be “Good enough that you can predict without testing.”