US researchers recently developed an AI system that can recommend recipes that are tailored to the preferences and nutritional needs of individual users.
“Our work focuses on personalized food recommendations. In particular, when a user request is made in natural language, we want to get the best matches in a recipe data set,” said Mohammed J. Zaki, one of the researchers who developed the pFoodReQ system.
Developed at the Rensselaer Polytechnic Institute and IBM Research in New York, the system is designed to help people find healthy recipes that meet both their nutritional needs and inclinations, while preventing them from containing certain components that can be harmful to certain health problems Conditions like carbohydrates and sugary foods for diabetic patients and some types of fruits or nuts for allergy sufferers.
“The key idea is that for the same query the answer should actually be different for different users. This is a very challenging task, especially in terms of identifying the implicit constraints that are actually relevant to the query,” Zaki told TechXplore Website.
The pFoodReQ dataset contains over 67 million records as well as graphical representations of the relationships between these recipes and the ingredients required for them, as well as data on the properties of the ingredients, the nutritional content and the nutritional value.