IBM's Watson has yet another new gig, and it could lead to a novel way for big data to interact with personalized healthcare. IBM has partnered with Bon Appetit magazine to create Chef Watson, a web app that can create unique recipes based on a list of preferences. The app is going into a very limited beta (capped at 200 participants) today.
Watson is a computer program that can do big things with big data, parsing and combining large data sets in a way that seems almost human. That capacity brought Watson to victory against human opponents on Jeopardy! three years ago, and since then it has started to see applications in the medical field. At Memorial Sloan-Kettering Hospital, Watson used the same smarts to comb through the latest medical journals and help doctors sift through the large, ever-changing set of treatment options for cancer patients. And in its partnership with Welltok, IBM demonstrated that Watson could bring its data-based insights to the individual level, making personalized recommendations to a user based on their particular health needs.
Right now, Chef Watson isn't designed to create recipes that cater to a user's health needs, although it does have the capacity to exclude ingredients. Instead it's designed to work alongside chefs, to give inspiration for unexpected ingredients. The user inputs a style, a dish, and an ingredient they want to use (for example a Caribbean-style paella with eggplant). Watson comes up with a recipe that combines expected ingredients with ingredients that are off-the-wall, but, because they contain complementary flavor compounds, should still taste good. The Fourth of July-themed test run carried out by Bon Appetit included corn salad with nectarines, short ribs with Chinese mustard, and a berry cobbler with marjoram, for instance.
Having a computer that can mimic -- or even surpass -- the professional chef's instinct for flavor combinations could be huge for fitness and wellness, though. While mobile devices and apps have made huge strides in the last few years helping people track their exercise, there's still surprisingly little to help with the other side of the weight loss equation. It's not hard to imagine Watson adding filters like low-calorie, vegetarian, gluten-free, or diabetes-friendly into its existing algorithms. And then imagine integrating that into something like Welltok's CafeWell concierge, which already takes the user's dietary preferences into account to make restaurant recommendations.
IBM has already hinted at this future in a graphic created for SXSW in Austin, where an earlier iteration of Chef Watson was put to work developing the menu for an IBM food truck. "In the future, personalized web applications will offer recommendations based on our medical needs and flavor preferences," the company wrote.
If Chef Watson becomes Healthy Chef Watson, it wouldn't just be a program for generating recipes tuned to the user's specific dietary requirements. By churning out surprising, unexpected recipes like it does now, the app could help make cooking fun for people normally inclined to eat out, while catering to their particular needs in the process.
If the Quantified Self movement has taught us anything, it's that while tapping into one data stream is powerful, combining many data streams together is what yields real health insights. Chef Watson already knows flavors, and Watson-powered CafeWell Concierge already knows its user's dietary needs and preferences. Apps like MyFitnessPal know just how many calories their user needs as well. In an interview with Slate, Robyn Metcalfe, the director of the University of Texas’ Food Lab, suggested Watson could some day "modify a specialized diet with streaming data from, say, your Jawbone UP to compose a recipe that is ideal for you at that exact moment."
Perhaps what's most exciting about Watson's increasingly complex endeavors, is that people are veering away from the knee-jerk reaction of fear that a computer will replace their doctor/chef/personal trainer, and seeing Watson more as a tool. In the oncology ward as in the kitchen, Watson remembers and combines more pieces of information than a human could ever hope to, then turns it over to a human to make the best treatment decision -- or the best souffle -- based on that data.