Any home cook in need of inspiration can now get instant support from none other than IBM’s supercomputer Watson, thanks to a simple online app.
Chef Watson‘s suggestions are often unusual, sometimes outright bizarre, but always scientifically based. Welcome to cognitive cooking!
Meet the Chef
Who never dreamt of a meal-making robot? The Jetsons already did, back in the 1960s. Half a century later you’d still be hard put to find a machine that will do the actual cooking for you, but at least coming up with an answer to the dreaded “what’s for dinner?” has become a little easier.
The Chef Watson application is, according to IBM, “part of our mission to develop cognitive computing applications that can help people discover new ideas”.
The term cognitive computing typically refers to technology which mimics the information processing functions of the human brain, and helps improve human decision-making. Cognitive computing applications use natural language processing to analyze huge amounts of data, combine these data, and present the results as new combinations.
Chef Watson, for example, initially analyzed more than 30,000 recipes. Using natural language processing, it looked at which ingredients were typically used together and for which styles of cooking, in order to detect patterns and use the underlying logic to create new combinations.
It also learned about the particular characteristics of each individual ingredient, including chemical compounds, typical preparation techniques, and all sorts of human taste preferences.
In a second phase, the computer was fed more than 10,000 recipes from Bon Appétit Magazine‘s database, which is used as a basis for the new recipes it creates with different flavor combinations.
To use Bon Appétit’s own example, “a braise always requires an initial sear and low-and-slow cooking with liquid. Watson socked away that information, and added its knowledge of the chemical compounds that would make good ingredient pairings to inspire braises with brand-new flavors.”
Cooking with Chef Watson
To start using the application, you can either type in one or a few ingredients, select a type of dish (such as pasta or fruit dessert), choose a cooking style (such as Christmas, South American or vegan), or use any combination of these filters.
Armed with flavor profiles for each ingredient and its extensive knowledge of ingredient pairings, Chef Watson’s algorithm then suggests four ingredients that contain complementary flavors that should work well together. Don’t be surprised if you find coffee or beer on the list — apparently, both have lots of useful flavor compounds and are extremely versatile.
Chef Watson also churns out a list of unique, new recipes using said ingredients. If none of the recipes takes your fancy, you can refine the search by adding or removing criteria.
And the results?
Some of Chef Watson’s creations are indeed … creative. How about, for example, dried fig dip with soy beans and piquillo peppers? A Japanese whisky sandwich with chicken, cinnamon and caramelized olives? Or a cajun blue cheese burrito with shrimp paste and hyssop?
But overall Watson’s flavor combinations do work, according to Engadget, who took it upon themselves to try out some of the recipes included in the chef’s first cookbook, Cognitive Cooking with Chef Watson: Recipes for Innovation from IBM & the Institute of Culinary Education.
As for us, so far we can only vouch for Chef Watson’s red cabbage savory fritters (see below). They are surprisingly tasty indeed.
The application does insist that “Chef Watson really needs you to use your own creativity and judgment”. Indeed, it seems to work best when cooks use its suggestions as a starting point, taking inspiration and adjusting recipes based on their own cooking skills, tastes and available ingredients.
More than a gadget
While the application may seem like little more than a funny gadget for bored home cooks, it does serve a larger purpose: Learning how computers can enhance human creativity and decision-making. Chef Watson helps people gain new insights by discovering new connections, and this capability is applicable in almost any industry.
For example, in the same way Chef Watson can make connections between cooking ingredients and their chemical compounds, cognitive computing is used in the pharmaceutical industry to understand relationships between drugs and diseases, and to accelerate the discovery of new treatments or new applications for existing drugs.
In this way, Chef Watson demonstrates the potential for efficient human-machine collaboration. By combining a pinch of human intuition with a splash of the information-processing capability of machines, the result may be brand new solutions that neither man nor machine would have come up with on their own.
Chef Watson’s Red Cabbage Savory Fritters (4 servings)
1⁄2 oz crumbled blue cheese
2 beaten eggs
1⁄4 cup flour
1 1⁄2 tbsp vegetable oil
1⁄2 tsp kosher salt
1 calabrian chile
3 cup chopped, sliced red cabbage
Pulse egg, flour, blue cheese, and the kosher salt in a food processor to combine. Add tomatoes, red cabbage, and calabrian chiles; pulse 2-3 times.
Heat vegetable oil in a large nonstick skillet over medium heat. Working in batches, cook heaping tablespoonfuls of batter about 4 minutes per side; season fritters with kosher salt.
Serve with sour cream and mango wedges.