Kimia Ghobadi, Johns Hopkins University – One Size Doesn’t Fit All: An AI Approach to Healthier Eating

A.I. has been getting mixed press lately, but can it help us become healthier?

Kimia Ghobadi, John C. Malone assistant professor in the department of civil and systems engineering at Johns Hopkins University, considers algorithms and diets.

Kimia Ghobadi is a John C. Malone Assistant Professor in the Department of Civil and Systems Engineering at Johns Hopkins Whiting School of Engineering and an associate director of its Center for Systems Science and Engineering.

Her research focuses on using mathematical models, optimization techniques, and data analytics to solve problems in complex systems, particularly in healthcare and medical decision-making environments. She develops models and solution techniques in inverse optimization, mixed-integer programming, and online algorithms. Current projects include inverse optimization models for personalized diets and radiation therapy treatment plans, capacity management and resource allocation for healthcare systems, COVID-19 simulation and analytics, hospital scheduling and process efficiency, mobility in frail and elderly patients, and smart ICUs.

Before joining JHU in 2019, Ghobadi was a postdoctoral fellow at MIT Sloan School of Management and collaborated closely with the Massachusetts General Hospital on improving healthcare operations.

One Size Doesn’t Fit All: An AI Approach to Healthier Eating

Everyone knows that diets are hard. Changing how and what we eat—either for weight loss or to manage symptoms of chronic illnesses—is challenging. Some studies show that more than 80% of people who try to follow diets fall back into their old ways quickly.

One reason this happens is that one-size-fits-all nutrition plans don’t consider people’s actual food preferences, lifestyles, and habits.

Artificial intelligence can help. Using an AI process called inverse optimization, my team is customizing diet plans to account for people’s tastes and habits, increasing the chance that healthier approaches stick. We believe the key lies not in telling people what they should or shouldn’t be eating, but in helping them find the best achievable diet that considers what they enjoy eating, what is available to them, and more.

Unlike traditional optimization, which uses math to the model decision-making process and find the best possible decision, or machine learning, which replicates prior behavior, we start with both what they need to achieve and what the person is already doing, and create a realistic goal based on that.

Our approach combines the ideal–what the patients should be doing to lower their salt or sugar intake, for instance–with the practical–what they actually do in real life. The result is a series of individualized plans that begin with food choices that align closely with people’s daily habits, and gently guide them over time in a healthier direction.

We are currently developing an interactive website where patients can view and customize diets.

The ultimate goal is to enable data-driven diet personalization that helps people be successful in eating more healthfully.