Developed by Hugo Liu, a graduate student at M.I.T.’s Media Lab, a novel “smart” cookbook recommends a recipe on the basis of the intuitive tastes and emotions commonly associated with the dish. From the NYT Magazine:
Synesthetic Recipes, a searchable computer database of 60,000 recipes, can’t actually read your mind, but it comes close. In the manner of a conventional cookbook, it is indexed according to 5,000 ingredients and 400 cooking procedures. But it can also be searched according to terms that range from the descriptive (“silky”) and the playful (“aha”) to the referential (“Popeye”) and the temperamental (“brooding”). Looking for something that’s “exotic” and “melodic” and “citrusy”? The cookbook suggests barbecued pork ribs with a currant glaze or jackfruit pudding.
The database takes its name from synesthesia – the blurring of sensations, as when you “see” sounds or “hear” colors. To create his web of food associations, Liu and a team of his fellow M.I.T. researchers first mined a variety of informational sources: food sites on the Web, the records of the culinary historian Barbara KetchamWheaton and a catalog of thousands of simple statements (like “butter tastes rich”) that were volunteered for an unrelated M.I.T. research project. Liu and his team then cross-referenced this information and combined it with a giant Web bank of recipes. Their task – which took about four years – essentially amounted to programming a computer with the knowledge that, for instance, a souffle is ethereal because it’s fluffy, that it’s fluffy because it’s made with well-aerated egg whites and that whipping egg whites aerates them.
Picture caption: “… a) … as keywords are typed, recipes suggestions fade in and out; b) avatars’ tastebuds are keyword-programmable; c) avatars react to browsing with love-neutral-hate and dialogs; d) recipe text is semantically highlighted to show the ‘essence’ of the query.”
More at the NYT Magazine…
If you would like to learn more about the project, read this .pdf presentation by Liu, et al.
Video…
(hat tip: WMMNA)