Dr. Ganesh Bagler admits to never having been a foodie. “I come from a zone where historically famines have been the order of the day,” says Dr. Bagler who grew up in Solapur, Maharashtra. “You couldn’t be choosy about food. You ate what was on your plate, and be grateful for having food. That was the ethos inculcated in me,” says the New Delhi-based professor, who was in Bengaluru last week to deliver several lectures on his brainchild, Computational Gastronomy, an emerging data science that blends food and computing.
Over time, however—no doubt prompted by his deep research into this novel subject over the last 8-odd years—he has developed a deeper understanding of food and recipes. “It has been a heck of an exhilarating journey,” says Dr Bagler, the director of the Infosys Centre for Artificial Intelligence, IIIT-Delhi’s Complex Systems Lab, defined on its website as “the ground zero of computational gastronomy” and a place involved in “transforming food with artificial intelligence.”
Dr. Ganesh Bagler.
What is computational gastronomy?
So, what exactly is computational gastronomy? The simple technical definition goes like this. “Computation Gastronomy is a data science that blends food with computing to provide data-driven food innovations,” says Dr. Bagler, adding that this science looks at all aspects of food through the lens of data and computing.
But how does one funnel a multifaceted phenomenon like food into prosaic data sets? While we can all universally agree that learning how to cook and cultivate food has had a significant evolutionary effect on the human species as a whole, why we eat what we eat is a function of multiple, often intersecting, factors. Think traditions, culture, invasion, ancestry, availability, preference, health requirements, migration, war, etc.
Dr. Bagler agrees that food is a complex entity, a product of “history, evolution, geo-climatic factors, religion and other accidental factors including colonial rule, all of which played a role,” he says, adding that it is difficult to comprehend why we eat what we eat today. “However, collective data on various aspects of food allows us to understand this complexity by asking the simplest of questions, in an objective fashion,” he believes.
Food pairing and more
Take, for instance, the science around food pairing, a theory first presented by Michelin-starred chef Heston Blumenthal together with the flavour expert François Benzi. According to their hypothesis, “the more aromatic compounds two foods have in common, the better they pair together,” they claimed, adding that the flavour of a food was determined by volatile aromatic compounds rather than mouthfeel and taste. Caviar, for instance, pairs beautifully with white chocolate as does dark chocolate with blue cheese, strawberry with coriander, snails with beetroot and so on, as Blumenthal discovered.
“That was a proposition we tested out in our lab,” says Dr. Bagler, adding that this theory was broadly true. But there were also cultural idiosyncrasies, which did not follow this simple, rudimentary law of food pairing, he says.
“Spices are the molecular fulcrum of Indian cuisine,” he says, pointing out that if one vegetable is swapped with another in an Indian recipe, it only changes marginally. | Photo Credit: enviromantic
Indian cuisine, for instance, as Dr. Bagler and his team discovered using data sets, contains negative food pairings--dissimilarly-flavoured food ingredients—which are bound together by spices. “Spices are the molecular fulcrum of Indian cuisine,” he says, pointing out that if one vegetable is swapped with another in an Indian recipe, it only changes marginally. However, if one spice was replaced with another—a clove with cardamom or cinnamon—the change is more dramatic. “The unique placement of spices is very important in Indian cuisine,” he says, adding that though this was something chefs had always known about, the researchers had approached it from a data-driven rather than culinary perspective. “This was an objective, statistical way to show the relevance of spices,” he says of the finding that jettisoned him into the complex world of computational gastronomy.
Oriental indian spices. | Photo Credit: YelenaYemchuk
Data and more
It started with a discussion with his students around flavour networks, while Dr. Bagler was teaching network science at IIT Jodhpur. They were looking at a 2013 paper by Yong-Yeol Ahn and Sebastian Ahnert, which sought to “investigate regional variations in culinary culture by constructing a flavor network of food ingredients, based on shared flavor compounds, and comparing this network to recipe data,” as the abstract of the paper points out, and decided to “extend that work into Indian cuisine,” says Dr. Bagler, who is trained in physics, computer science, and computational biology.
Dr. Bagler and his team started by creating a data base of Indian recipes, culled from renowned Indian cookbook author Tarla Dalal’s website, going on to break down each recipe into ingredients and then analysing if the ingredients in a particular dish share flavour component, the findings of which were published in 2015 in a paper titled Spices form the basis of food pairing in Indian cuisine.
The positive reaction that this drew from the world had him continuing his research in the same. “I didn’t come from a culinary connoisseur’s perspective,” he admits. “But I realised that I had arrived at something deep and insightful,” says Dr. Bagler, who firmly believes that this new science offers immense societal value.
Use and misuse
So far Dr. Bagler’s lab has come up with a number of resources, including RecipeDB, a structured repository of recipes from around the globe, FlavorDB, a repository of flavour molecules found in food ingredients, DietRx which explores the health impact of dietary ingredients and Ratatouille, an app that helps generate novel new recipes using AI. All these resources are available for exploration for non-commercial purposes right now, while a start-up with a commercial model is being launched soon.
And what if a technology such as this one is misused by the processed food industry, which has been manipulating and overstimulating our taste buds for decades, catalysing an epidemic of obesity, diabetes and cardiac disorders among other things?
“Yes, fears are often expressed that this can be misused,” agrees Dr. Bagler. In his opinion, however, if an ethical food processing industry is given a choice between creating a product that is both enticing and nutritious and one that tastes good, without offering positive effects, and both are accepted by the customer, it is likely that the former will be chosen. “Understanding food, recipes and food products, traditional recipes from the lens of flavours, odours, and nutritional profiles will allow us to come up with potential food products, which are both nutritionally balanced and palatable.”
To know more about computational gastronomy log into https://cosylab.iiitd.edu.in/
COMMents
SHARE