What if every cuisine has its own grammar? And could artificial intelligence learn to speak it?
Professor Ganesh Bagler of IIIT-Delhi studies food through data, algorithms and an emerging field called computational gastronomy. His research explores the patterns hidden in recipes, the ways ingredients are combined and what computers might discover about centuries of culinary knowledge.
But behind the databases is a personal story: a childhood shaped by food scarcity, his mother's cooking, and a scientist who found his way into the kitchen through curiosity rather than culinary ambition.
In this conversation, we explore how AI learns from recipes, why culinary traditions have recognisable patterns, and what gets lost when food is reduced to data. Ganesh introduces projects including RecipeDB, FlavorDB and Ratatouille, an AI-based recipe-generation tool.
We also discuss the limits of AI-generated recipes, the knowledge contained in a sushi chef's hands and the difference between reading a recipe and actually cooking it.
Then Nyx, Holger's AI cooking collaborator, joins the conversation to test an idea about culinary traditions and creative freedom.
Can a machine learn the rules of cooking without ever experiencing the taste of food?
Guest: Professor Ganesh Bagler, IIIT-Delhi
Host: Holger Nauheimer
About the AI voice: This episode features Nyx, Holger's AI collaborator, speaking with a synthetic voice. The human voices in this episode have not been synthetically generated or replaced.
Links & references
Ganesh Bagler: https://iiitd.ac.in/bagler
Complex Systems Laboratory: https://cosylab.iiitd.edu.in/
RecipeDB: https://cosylab.iiitd.edu.in/recipedb2/
FlavorDB: https://cosylab.iiitd.edu.in/flavordb2/
Ratatouille: https://cosylab.iiitd.edu.in/ratatouille2/
I Have Never Tasted a Tomato explores how AI is changing the way we cook — and what cooking reveals about human creativity, memory and taste.