By Steve Abrams
One of my all-time favorite activities is barbequing on my deck (and I mean real barbecue, not grilling). My favorite dish to make is beef brisket, which, if you’ll allow me to be immodest, reaches a pinnacle of perfection after 12 hours in the smoker.
Yep, I love to eat and I love to cook and I love to experiment. I almost never use a recipe exactly as I find it.
That’s why I’m so happy that my group at IBM has joined with Bon Appétit one of the world’s most respected food-media brands, to enhance culinary creativity and discovery with cognitive computing.
Today, Bon Appétit’s editors published a package of stories about a web-based cognitive cooking application that we’re developing, called “Chef Watson with Bon Appétit.” This has evolved from the same technology that we debuted at SXSW in Austin, Texas, a couple of months ago, serving Watson’s culinary creations from a food truck. But now, in collaboration with Bon Appétit, we’re introducing a limited beta of an application built around this technology, tailored to the needs of avid home cooks.
And this is no ordinary application – we’re putting the power of cognitive computing into the hands of the home chef to inspire their creativity and help them discover new recipes that have never been imagined before.
If you want to be considered for participation in the pilot program, sign up here. We will be expanding access to the beta over the coming weeks and months.
You might wonder: Why is IBM, which caters primarily to large enterprises, involved in cooking? Our goal is to help people discover the potential in the new era of computing–the era of cognitive systems. These new systems ingest vast amounts of data, learn from their interactions with people and information, reason, make recommendations, and relate to humans in ways that are more natural to us. We believe that cognitive computing will transform professions, industries and, ultimately, society. To the layperson, these concepts can seem rather abstract. By illustrating cognitive computing with a system that generates new recipes, we can give everyone a ‘taste’ of cognitive, making it more tangible and meaningful.
To build this application, we taught Watson about cooking by giving it about 9,000 Bon Appétit recipes to read, letting it glean insights about ingredient pairings, cooking styles, and dishes. The application combines these insights with information about food chemistry, hedonic psychophysics (the psychology of what people find pleasant and unpleasant), and regional and ethnic cooking.
To generate a recipe, a chef inputs a handful of parameters, giving Watson guidance on what ingredients to use or avoid, as well as what styles of cooking seem interesting. Watson then generates millions of possible ingredient combinations, choosing a hundred variations that it thinks will surprise and delight the human palate. Watson even suggests ingredient amounts and cooking instructions. The idea is to inspire the chef to discover new dishes or ingredient combinations, amplifying their own culinary creativity.
The underlying technology is based on a system developed through more than two years of work by scientists at IBM Research. As part of this effort, IBM scientists collaborated with professional chefs at the Institute of Culinary Education (ICE) in New York, testing the limits of ingredient pairings and developing strategies to implement computational creativity in the culinary arts. When we showed this technology at SXSW hundreds of festivalgoers lined up at the Watson Food Truck to sample such creations as Austrian Chocolate Burrito, Belgian Bacon Pudding and Peruvian Potato Poutine.
A lot of people saw our original Watson system compete and win on the TV quiz show Jeopardy! That system was trained to absorb a large amount of information and answer all manner of questions. Today, Watson is hard at work as the ultimate assistant to our partners and clients in industries such as retail, healthcare and financial services. Chef Watson shows that cognitive systems can also help people be more creative. While the original Watson found answers that were already known, these new technologies help us discover brand new knowledge.
Cognitive computing applications designed to augment discovery and creativity have the potential make a big difference in a wide range of industries. They range from food and financial services, to chemicals and pharmaceuticals, to retailing and marketing.
We think food companies will use cognitive computing to create exciting new packaged products. And we anticipate that IBM business partners will create apps based on these technologies that enable individuals to dream up tasty dishes using what they already have in their refrigerator, or what’s available at the vegetable stand or fish market.
I can’t wait to see what our first round of pilot users creates using Chef Watson. My favorite cognitive creation so far is Kenyan Brussels Sprouts–crunchy sprouts spiced with cardamom on a sweet potato puree, but expect to see even more creative dishes in the coming months. Just imagine what Chef Watson could do with a smoker and a brisket of beef!
If you want to learn more about the new era of computing, read Smart Machines: IBM’s Watson and the Era of Cognitive Computing.
Credit for Steve Abrams photo: Jon Simon/Feature Photo Service for IBM