Tech News: This Sony Tech uses AI to predict what ingredients will pair together.
AI has moved on to gaming, autonomous driving, and other areas with varying degrees of success, and is now looking to make the switch to cooking. After Google’s AI hired a Great British Bake Off winner, Sony developed a deep learning system called FlavorGraph designed to combine ingredients like garlic, olives and milk.
Researchers from Sony and Korea University (KU) noted that chefs discovered how to combine ingredients by intuition, resulting in a gradual evolution for combinations such as cheese and tomato, pork and apple, and garlic and ginger. Many of these classic combinations were later explained by science when researchers realized that ingredients that share dominant aroma molecules often go well together. At the same time, other ingredients that combine well can have completely different chemical compositions.
To find out why the team looked at both the molecular information on the ingredients and how they were used in previous recipes. They then created the FlavorGraph database of flavor profiles such as bitter, fruity and sweet based on 1,561 flavor molecules. At the same time, they went through nearly a million recipes to see how the ingredients have been combined in the past.
The resulting data shows the chemical compounds common to foods such as wines and citrus fruits and how they affect their overall flavor and reveal which foods might pair well with specific wines or fruits. Some combinations are evident (biscuits and ice cream), others less so (white wine and condensed golden mushroom soup from Campbell). Researchers haven’t discovered anything special yet (caviar and white chocolate are one example), but FlavorGraph is just a starting point.
“As science advances and we get better and better representations of food, we should discover more and more intriguing combinations of ingredients, as well as new substitutes for unhealthy or unsustainable ingredients,” the team wrote.