An IT business exhibition may seem like an unlikely place for a chef to be. But there I was at CeBIT 2014, held from November 12-14 in Bangalore. I had been invited by IBM to showcase Cognitive Cooking and the web based Chef Watson App.
This app, created through collaboration between IBM and Bon Appetit, has been designed to meet the challenge that most people face — of discovering delightful new recipes or simply deciding what to make for dinner with the ingredients they have at home.
Chef Watson is a first-ever cognitive cooking beta app, a web-based app that is powered by IBM Watson technology and Bon Appetit’s culinary expertise and about 9,000 recipes, enabling users to design and cook recipes they might never have thought of. At CeBIT, the aim was to bring together Chef Watson and the expertise of a chef.
At the event, I selected 5 recipes created by Watson, modified them to suit Indian tastes and preferences and served them in mono portions. The dishes includes Turkish Style Bruschetta; Swiss-Thai Asparagus Quiche; Vietnamese Chicken and Apple Kebab; Caribbean-Style Chicken and Chips; Coconut and Plantain Espuma. The reactions of the crowd I must say was overwhelming and cognitive cooking became the most talked about activity at the event.
For the Special lunch attended by 120 CXOs of various companies we also served a duet of chocolate for dessert — Italian Chocolate Cake using Chef Watson’s recipe and Cryo Chocolate Ganache Powder with Saffron Soil & Mint Soil created by me.
This app is ‘cognitive’ because Watson can reason about flavour the same way a human use their palate. The system understands natural language, reading and analyzing tens of thousands of existing recipes to learn about ingredient pairings, dish composition, and cooking steps. It then cross references these with data on the flavour compounds found in ingredients, and the psychology of people’s likes and dislikes to model how the human palate might respond to different combinations of flavours.
The ingredients it selects are based on what it has learned might pair well based on the recipes it has read, as well as what it knows about the underlying food chemistry – specifically the flavour compounds of each ingredient. There are a huge number of possible combinations and Watson generates thousands that it thinks will work well, sorts them, and then presents the user with 100 options.
For me, it was an engaging opportunity to see technology play a role in creating new dishes and imagine the possibilities ahead. We live and cook in interesting times to be sure.