Chef Watson has arrived and is ready to help you cook
By Richard Brandt | 4 minute read | January 1, 2016
Tired of making the same old meatloaf or baked chicken for dinner? The Chef Watson with Bon Appétit application is ready to help you spice up your culinary endeavors.
Chef Watson is part of IBM’s mission to develop cognitive computing applications that can help people discover new ideas, from creating surprising new recipes to improving medical research. Over the last year, as part of a beta phase, it has helped thousands of home chefs create delicious meals. On June 23, sporting a brand new interface, the program was opened up to home chefs everywhere.
Chef Watson, however, is more than just a slick cookbook (although a cookbook of recipes it helped create has been published, too). Rather, it works with chefs to make them better arbiters of ingredient combinations.
Wendy Hite, for example, who grew up cooking Cajun in Louisiana, wanted to build her culinary skills after she moved to Texas. Watson helped her expand her repertoire into Asian-style dishes featuring flavors like ginger and star anise. Chefs Hite and Watson even created some crawfish deviled eggs together. “Watson pushes you to be more creative,” says Hite.
Cognitive computing applications like Chef Watson process information more like a human than a computer. These applications can analyze huge amounts of data, using natural language processing technology to help decipher what the words and sentences contained within the data mean. (This technology helped an earlier version of Watson beat Jeopardy! champions a few years ago.) These programs also excel at recognizing patterns in large chunks of data, and get smarter over time, meaning they get better based on feedback.
Chef Watson, for example, read up on the chemical composition of hundreds of different ingredients and analyzed some 10,000 recipes from Bon Appétit. By combining that data and detecting certain patterns, Chef Watson has learned to suggest up to four different ingredients that blend together seamlessly. Four doesn’t sound like much, until you realize that adds up to about one quintillion possible combinations. Cooking Italian? Watson will run through the most typical ingredients used in that cuisine, plus the ingredients they’re likely to go well with, and help you make the tastiest choices.
“We’re trying to show the role that cognitive technology can play” in helping humans with complex tasks, explains Steven Abrams, distinguished engineer and director at Watson Life. “Together the human and computer [can] discover more interesting results.”
The program’s designers have steadily refined the program for home chefs. The newest version starts by letting the human partner pick an ingredient, dish or style of cuisine. It then presents four ingredients that work well together, but allows the home chef to change any of those by selecting one of several suggested alternatives. Next, Chef Watson presents a selection of recipes, based on dishes from Bon Appétit. Even then, the home chef can tweak the recipe by swapping out ingredients or dicing instead of slicing.
That collaboration, rather than the solo journey of a cook following a cookbook, is what gets people excited. “We’ve brought the ‘wow!’ moment to the surface,” says Jacqueline Martino, design lead for Chef Watson. “We’re really appealing to the adventurous home cook.”
Susan Walter Sink, an already knowledgeable home chef who does volunteer work with farms, schools and non-profits to develop healthy recipes, agrees. She uses Watson when she teaches kids about cooking in demonstrations at farmer’s markets. She loves showing them how they can switch ingredients to change a dish. “That’s the ‘Aha!’ moment for them,” she says. Watson’s biggest appeal, she adds, is “its potential to inspire people to get in the kitchen and cook.”
Sink likes to use fresh, farm-to-table ingredients, so her recipes have to change based on what’s in season. Some ingredients, such as broccoli or kale, are sweetest in the winter, so she uses Chef Watson to look for alternatives in the summer. “What’s neat about working with Watson is that I can take a recipe and push it to savory or sweet,” she says. “Watson has shown me how to form combinations that I would not have thought possible.”
Of course, Chef Watson is only one piece of IBM’s repertoire. Watson’s ability to sift through large data sets and match compounds may be applied in the future to anything from creating perfumes, to helping scientists create new metal alloys or plastics, to fighting cancer.
Hundreds of research papers are written each year exploring treatments for cancer and other diseases, but nobody can read all of them. Just as Chef Watson sifts through thousands of recipes to find patterns, a potential medical version of Watson might be able to go through research papers to find hidden patterns in diseases and their treatment protocols. Another possible application: sorting through thousands of legal briefs to find successful patterns, aiding lawyers in crafting their arguments during litigation.
But for now, Chef Watson is ready to help you get cooking.
Want to try Chef Watson out for yourself?