The Indian academic making the world look at flavours and food in a fresh way

If you are a foodie and love experimenting with flavours, how does a shrimp and egg salad sound for lunch? Or a tender juicy steak, just out of the oven, topped with a dollop of cheddar or Camembert, to be precise. And did you know mushrooms go smashingly well with nearly every common fruit, from apples to apricots and even coconut?

These food pairings may sound unappealing or at least unusual to most people, but Dr Ganesh Bagler of the Indraprastha Institute of Information Technology, New Delhi, says otherwise – and he has the data and research to show for it.

Can we encode the intelligence of a chef into a computer, or can a computer fool a chef into thinking a recipe is real?

Akshay Malhotra

The computational gastronomy expert has taken the food and drink industry by storm with his ground-breaking work on flavour molecules and its corresponding database, FlavorDB.

His laboratory has also developed DietRX, an archive of nearly 2,000 foods, their chemical and genetic compositions, and their effect on health, which can enable culinary and drug interventions. (Ayurvedic diets are a historically important example of the belief in healing via appropriate foods.)

“The power of data and food together is magic,” says Bagler.

Already, chefs such as Garima Arora of Restaurant Gaa in Bangkok, are using Bagler’s research to fuel their own food experiments. “What I find amazing about Bagler’s research is that his approach actually enables us to know exactly what makes up a cuisine – the things that make Indian cuisine Indian,” Arora, who is the first Indian woman with a Michelin star to her name, tells The National.

“Once we have that knowledge, we can truly get to the main taste of a cuisine, which will help us do away with the flavours and ingredients we don’t need.”

Bagler’s work is also critical to Arora’s Food Forward India, a non-profit initiative that aims to broaden the narrative around Indian food. “It fits into the framework by being a forward-thinking initiative, one that serves the purpose of codifying a cuisine, and identifying, quantitatively, its identity,” says project manager Matylda Grzelak.

Bagler, who is now considered the pioneer of computational gastronomy in India, credits curiosity for his success. Having studied various subjects from graduation through to postdoctoral studies – quantum mechanics, computer science, computational biology, computational neuroscience and molecular genetics – Bagler returned to India in 2010 following a stint at Berlin’s Max Planck Institute for Molecular Genetics.

He joined the CSIR-Institute of Himalayan Bioresource Technology at Palampur as a researcher, and worked on medicinal plants in the western Himalayas, and on diseases such as cancer and asthma. But it was not enough.

“I like to explain things; I’m a teacher,” he says. This led him to the prestigious Indian Institute of Technology Jodhpur and finally IIIT-Delhi.

How Heston Blumenthal kickstarted his work in food

Bagler's foray into gastronomy happened when he came across a 2011 paper that took off from British chef Heston Blumenthal’s food-pairing hypothesis: foods that share flavour molecules will taste better together than those that do not.

For example, chocolate and blue cheese taste great together because they share 73 flavours (Blumenthal’s interest was piqued when he paired white chocolate and caviar, and hit the right notes). This led to companies such as Foodpairing, which present thousands of combinations of ingredients for chefs to experiment with.

“Historically speaking, dishes have evolved over millennia from single-ingredient meals to complex ones,” says Bagler. “Cooking techniques and creative expression aside, why are some ingredients used together and others not?” This was one of the critical questions that led Bagler to expand his research. “Food science has been around, but it explored aspects such as the shelf life of foods or how to enhance sensory enjoyment. Now, people are looking at food from a data perspective.”

What Bagler did differently was focus on Indian food, which he found is different from other cuisines because of the spices. Breaking down a collection of the late, celebrated Indian chef Tarla Dalal’s recipes, Bagler realised that spices form the basis of food-pairing in Indian cuisine. Having divided various foods into 26 categories – vegetables, dairy, lentils, meats, etc – he saw that mixing up items across all other sections did not cause too much of a shift in flavour, but when the spices were shuffled, the taste changed entirely.

Spices are the molecular fulcrum of Indian food

Dr Ganesh Bagler

For example, you could replace spinach with fenugreek leaves in palak paneer and there would not be much change in the dish, but if you replaced turmeric with cinnamon, the very essence of the preparation alters. “Spices are the molecular fulcrum of Indian food,” says Bagler.

Flavour technology rather than food technology

In 2015, Bagler and his team of researchers sent this study to international science journals, which uploaded it to an open server where it was picked up by MIT Tech Review. This changed Bagler’s life. “I only understood the academic value of this work, not its futuristic value,” he says. “It took me a year to understand that this had led to the creation of a new field of study, and now, over the past five years, I’ve been developing the foundations of this area.”

From 2,543 of Dalal’s recipes to nearly 158,000 global recipes, Bagler’s database has expanded exponentially. Not only is the data free to access on various websites and apps, but the information is also provided in excruciating detail, from the scientific names of elements to a comprehensive flavour network, possible pairings and health benefits. Bagler is also due to launch RecipeDB at a conference – postponed amid the Covid-19 crisis – where a massive collection of structured recipes will be available for everyone from chefs and cooking enthusiasts, to restaurateurs, multinationals and scientific organisations to use freely.

“People know food technology, but they do not know about flavour technology,” says Deepika Nadiminti, a flavourist at Mane India, which develops flavours for the dairy, confectionery and drinks industries. “Bagler’s database is an all-in-one resource, where we can identify everything from flavour molecules to physical and chemical properties, and experiment easily,” she says.

While Bagler consults for institutions such as the Indian Institute of Hotel Management and Symbiosis School of Culinary Arts, as well as a range of multinationals, chefs also swear by his research, which has “significantly reduced time spent on developing new dishes”, says Akshay Malhotra, a chef, food consultant and former student of the Culinary Institute of America.

“FlavorDB will help us to understand the science behind Indian food, and it is only the beginning of how artificial intelligence will influence the food industry,” he says.

This aspect is also key to Bagler’s future experiments. “Can we encode the intelligence of a chef into a computer, or can a computer fool a chef into thinking a recipe is real?” says Malhotra.

Can human creativity, which is at the heart of cooking, be reproduced using AI? It remains to be seen. For now, Malhotra’s observations pertain to FlavorDB complementing chefs’ instincts about pairing ingredients.

As celebrity chefs Manjit Gill and Akshraj Jodha describe Bagler's work, he is successfully quantifying the knowledge that, until now, was only intuitively available to a cook – and everyone from chefs and diners to scientists will benefit from it.