AI trained from decades of culinary experimentation creates new dishes

    Artificial intelligence (its weak form) penetrates into science, art, military affairs, games and many other areas. Now this technology is beginning to be applied in such an unexpected industry as cooking. The pioneer was IBM, which took advantage of Watson's capabilities to create new recipes and improve existing ones.

    Experts who work on the project believe that if the test phase of the experiment goes well, then AI will have a significant impact on what we eat every day. The ideal is still very far away, but there are already some successes.

    For several years, IBM has partnered with McCormick & Company, specifically its food specialists, to study the culinary field. The combined team taught AI how to make dishes, and innext year, Watson recipes will arrive on store shelves. If this experiment succeeds, then experts from McCormick & Company will work with AI technologies starting in 2021.

    Representatives of the project say that with the help of Watson they want to create the best recipes in the world. Of course, all this has a commercial background - McCormick & Company will deliver the results of successful experiments to the market. Watson will develop not only dishes, but also seasonings for them, as well as various kinds of mixes.

    Slow and not very confident

    A McCormick & Company representative said that it takes weeks and months to develop and launch a new food product on the market.

    When a company creates a new product, dozens or even hundreds of experts - chemists, technologists, and culinary specialists - begin to work on it. They develop individual components that later become part of the whole. If necessary, the components of the product are replaced or completely removed until the final version approaches the ideal.

    IBM algorithms work in much the same way - the system conducts experiments and selects the best result. But Watson has a lot more different formulas in its memory than one person or even a group of experts can remember. AI can experiment by mixing and separating the ingredients, getting the most unexpected combinations.

    The developers believe that Watson can reduce the number of iterations required in the manufacturing process of the food product and generally significantly reduce the time required for development.

    AI works with a database that contains a huge number of different recipes, results of culinary experiments, descriptions of seasonings, etc.

    And something else

    The strength of AI is that it does not have an established opinion about what combinations of products should be for obtaining certain dishes. The main task of the system is to create what can be called tasty.

    Watson learns from McCormick & Company research data from decades of company operations. In addition, the algorithm receives information about the results of marketing research according to people's preferences regarding products, depending on factors such as culture, location, mood.

    Experts believe that all this must be taken into account. Next year, as mentioned above, Watson recipes, condiments, and sauces will go on sale. Many of them will simply be updated, without major changes, in order to improve taste.

    In the future, according to the developers, AI will completely change the world of cooking. Perhaps the AI ​​will pick up various new components, and the person will collect from this recipe preparations. Or the AI ​​will advise using chopped garlic cloves to prepare a specific dish instead of garlic powder - and it will be really tastier, and the size of the chopped elements will be set by the system.

    It may well be that in 20-30 years people will not understand the previous generations that cooked food without the help of AI.

    Also popular now: