The Flavor Matrix: The Art and Science of Pairing Common Ingredients to Create Extraordinary Dishes

Book Preface

The idea for this book came to me while I was cooking with a computer.

The year was 2012, and my kitchen companion was Watson, IBM’s famous supercomputer. If you’re like most people, you first met Watson when it competed-and dominated-on the TV show Jeopardy! You might not have known that the com­puter’s talents extend beyond trivia. I certainly didn’t when I first met Watson-but I was in for a surprise. When IBM had approached the Institute of Culinary Educa­tion, where I work, to collaborate on a project that would mix computing and human creativity in a way that had never been attempted before, I was honored to be asked to participate, but I was also very skeptical. How could a trivia-playing computer teach me anything about cooking? After all, I had devoted my life to being at the cutting edge of the culinary world. Besides, I thought, cooking is a craft-an art, something that engages the senses and requires a human touch. What could a com­puter know about the way a ripe tomato yields to the touch, or what a steak sounds like as it sizzles to perfect doneness?

It wasn’t long before I realized that Watson was capable of much greater things than simply answering questions. Since proving its revolutionary computing abilities on Jeopardy!, Wat­son had turned its prodigious talents to managing care for cancer patients, understanding financial markets, helping law enforcement solve cold cases, even providing music recom­mendations. Now Watson was going to attempt to help sea­soned chefs become more creative in the kitchen. Jeopardy! had proven that Watson could answer any question posed to it, but what about the questions we don1 even know to ask?
The collaboration would work like this: Watson would create a list of ingredients that I or another of the chef-instructors at ICE would then use to create a dish with ingredient combi­nations that had rarely-if ever-been seen before. The foods Watson would select for us often seemed random at first lance, but would be anything but; rather, through a sweeping analysis of data pulled from academic articles, cookbooks, and other sources, Watson would find combinations of ingredients that it predicted would taste good when combined-but that did not commonly appear together in recipes.

I enjoyed cooking with Watson. We produced some interest­ing dishes: roast duck with tomato, sage, olives, and cherries; lobster with pork, saffron, basil, carrots, and balsamic vinegar; chicken with strawberries, mushrooms, and apple. These were combinations that I never would have come up with on my own, and our work together inspired me to reconsider what creativity in the kitchen actually looks like. But I knew I had only scratched the surface. Watson didn’t view ingredients the same way that I did as a chef. When I look at a tomato, I immediately think about all the ingredients I have tasted with tomato in the past, too often focusing on those that are familiar and comfortable like basil, cheese, or olives. Watson, on the other hand, thought about combining ingredients based only on their inherent flavors, with no notions of which foods conventionally go together. The computer seemed to be able to see the invisible filaments that bound different ingredients together.

For the moment, the ability to see these connections was the exclusive domain of scientists and supercomputers. But I wanted to learn to see them for myself, and to help others do the same. None of my students at ICE, let alone my family or friends, had access to a supercomputer that could help them think more creatively about flavor pairing as Watson had helped me do. How could I share this amazing experience with them? Luckily, there is a vast body of scientific data out there about flavor, just waiting to be explored. But this information isn1 part of the language of chefs-which is a shame, because the art and (yes) science of food pairing is in some ways the final frontier of modern cooking.