The Food Pairing hypothesis says that aliments combine together well when they share some major flavour component, identified as a chemical constituent of the culinary product. The success of the food pairing is so that new matching of ingredients have been tried by avant-garde chefs. A well known example is the pairing of white chocolate and caviar, justified by the fact that these two products share trimethylamine, the organic compound that gives fish their fishy odor.
Flavour is a very complex sensation that involves many senses and chemical receptors in the body. As many sensations actually participate in defining what is the individual response to flavour, it is believed that the systematic study of the chemical components of food can shed light on the reason why some ingredients have more chances to fit together well in a recipe.
Actually, the compounds that are known to contribute to the flavour of culinary ingredients is determined by chemical analysis. The ingredients that are alike, following the food pairing hypothesis, are connected in a network that maps the relations between aliments. These results are used mainly as a source of inspiration for chefs, and other scholars studying food.
Nevertheless, the food pairing hypothesis is not widely acknowledged. Beside the bare chemical compounds, there are many ingredients whose main role may not be only flavouring but something else as well (e.g. eggs’ role to ensure mechanical stability or agar-agar’s role to ensure a dense appetizing texture). Moreover, the flavour of a dish owes as much to the mode of preparation as to the choice of particular ingredients. The cultural background also reveals its importance because traditional and local ingredients are the elements that forged the heritage of any kind of cuisine. This aspect seems to be put aside by blind followers of the food pairing hypothesis.
The research on food pairing however reveals the new interest in materials and techniques for cooking. The hypothesis starts from a solid ground, and participated to launch the recent scientific systematic approach to food studies, which must also search for evidence supporting (or refuting) any food pairing rule.
- Ahn Y., Sebastian E. Ahnert S., Bagrow J.P. & Barabási A-L. (2011), “Flavor network and the principles of food pairing”, Nature Scientific Reports 1, 196.
In 2012 the corean LG will commercialise the “Smart fridge”. This wonder of modern technology realizes the myth of the past decades: an intelligent refrigerator in the kitchen, able to solve all the housewife’s problems. The smart fridge is able to register the grocery via barcode scanning or voice recognition, and it offers online grocery shopping directly from the refrigerator’s LCD panel or smartphone. Based on the stored ingredients and dietary directives, the smart fridge can also select recipes to be cooked into its pal, the smart hoven.
Despite this huge amount of improvements, that basically match my first ideas on intelligent kitchen appliances, the smart fridge has received kind of a cold welcome. The main reason is that this technological monster is scaring. Nobody wants to open the door of his fridge with its user’s manual in hand. Talking to a fridge is even more creepy. It appears that the enthusiasm for new technology doesn’t always marry well usability.
The technology has to go close to the humans, not the other way around.
So, what are we missing here? Susie Steiner from the Guardian correctly states: “The only technology that will survive the furnace of the global market is intuitive technology.” In other words, the use of technology has to match the use that people already do of the tools they have, the so called artifacts. The natural way of adoption of new technologies passes through it’s integration with other existing artifacts and how, over time, it modifies their use.
This is also the reason why field research is so important: to understand how people work, and the use of artifacts in their context. Building an Artificial Intelligence is only the first step, but we believe that intelligence also means adaptability.
Case-based reasoning (CBR) is a problem solving paradigm based on reusing the solutions of similar past problems. Unlike other areas of AI, CBR does not solve a problem by using the knowledge embedded in it, but focuses more on a learning approach that reuse what has been learnt from previously experienced problem situations, called the cases.
This approach implies that CBR does not require an explicit domain model; implementation is reduced to identifying significant features of the problem that describe a case.
A new problem is then solved by retrieving relevant cases, and map them onto the current problem. A new possible solution is tested by executing it in a real (or simulated) environment, in order to eventually identify and correct flaws. Finally, the revised successful solution is retained as a case for further applications, enriching so the set of stored experiences.
Case-based reasoning have been applied to several successful industrial applications, and has found a major application area in health sciences. What interests us is that CBR community has produced a “Computer Cooking” contest (CCC) that takes place during the annual ICCBR conference, and of which we’re going to talk shortly.
Aamodt A. and Plaza E.,(1994) “Case-based reasoning: Foundational issues, methodo- logical variations, and system approaches”, Artificial Intelligence Communications, 7 (1): 39-59
Begum, S.; M. U Ahmed, P. Funk, Ning Xiong, M. Folke. (2011) “Case-Based Reasoning Systems in the Health Sciences: A Survey of Recent Trends and Developments”. IEEE Transactions on Systems, Man, and Cybernetics, Part C: Applications and Reviews 41 (4): 421–434.
Alimentaria 2012 is one of the most important food and drinks trade shows in the world. In this culinary context, BCNVanguardia is an international conference about new trends in technology applied to gastronomy where I’ve been invited to give a talk to explain my vision to an audience of operators in food and drink manufacturing, distribution and trade. Not an easy task, in particular because it was my very first popular scientific talk. But it has been a success!
I noticed a couple of things. In Alimentaria – Restaurama, most of the exhibitors focused on two aspects: the simplicity of food preservation, and cooking those products without complication. An overall aspect was to maintain the quality along the whole process of preservation-use.
I focused my talk on few example of AI disciplines that could be applied to gastronomy and that make use of these optimization aspects that the exhibitors were looking for: scheduling, and recommender systems. Seeing examples helped a lot in understanding how the advances of AI can be applied to real kitchen issues, in order to solve them. I suppose that everyone reflected their own problems in the applications I depicted. I believe in the importance of mapping the theory into real applications: for this reason the next posts will resume a little bit the theory and advances of AI in these couple of fields.
Using Artificial Intelligence in the Kitchen
A lecture about the opportunities offered by applying Artificial Intelligence techniques to optimize resource and working time by Alexandre Albore (professor at Universitat Pompeu Fabra).
Every new activity needs some fieldwork, because not all the valuable information is located in a library or in a web page. This need of direct information comes from the fact that I am moving from research in computer science, strictly bound to academia, to the world of cuisine, which has been considered as essentially practical up to now. Many chefs will tell you that the tour de main, a knack learnt through experience, is central in their job, and it cannot be grasped easily, both by machines or humans.
One day in the Suite 7, a family-owned restaurant in Raval (Barcelona), has been a first contact with the food industry, and a rich harvest of data. Two main points in organising a restaurant called my attention: grocery shopping and the elaboration of the menu.
In such a restaurant, groceries are bought on a daily basis, depending on the needs and on what is available in the marketplace. This means that seasonal ingredients are very easily integrated in the menu. This also means that there is a very quick turnover, and basically no latency in the fridge. The menu is thus make of two parts: one fixed set of dishes, and the “chef’s suggestions of the day” that depend on what came out from the shopping.
Fresh and seasonal ingredients have they cons, in this case: no long term strategy can be established on shopping in relation to menu and user’s demand. And optimisation is something where AI can play its role.
Alícia is a private non-profit research centre devoted to technological innovation in cuisine, to the improvement of eating habits and to the evaluation of the food and gastronomic heritage. Alícia has been founded under the supervision of the chef Ferran Adrià, and hosts nowadays researchers from the fields of chemistry, dietetics (the study of nutrition as it relates to health), medecine, and of course, chefs.
Living in Catalonia and working on applying AI research to innovate in cooking, makes of Fondació Alícia a mandatory stop. We found there a warm welcome, a lot of expertise, an a lot of interest in stepping on these virgin fields together. Researchers working at Alícia are the ideal interlocutors both to discuss ideas and to gather inspiration for new projects.
Professional cuisine has different needs that private customers. Among the topics discussed, two thoughts were kind of unexpected to me. First, the absence of an on-line repository for recipes that can be used as a reference. Second, the difficulty to deal with sensors at the stove. I wonder if these are common issues for chefs world-wile.
Carme Ruscalleda, gave a talk at ESAB about the importance of raw materials in the Mediterranean cuisine. Carme Ruscalleda is one of the very few women with 3 Michelin stars. She talked about her approach to cooking, and how she does select raw materials and integrate seasonal ingredients in her menus. We have to say that Carmen Ruscalleda is daughter and sister of farmers, thus she has a privilegiate link with agriculture products.
Interesting enough are the distribution problems that the majority of local farmers have, generally unable to compete with the big distribution chains, notwithstanding the better quality and lower prices they offer.
Another question that is trotting in my mind is how fast a restaurant, a canteen, or even a chain is able to integrate these seasonal ingredients in their menu. Both quality and earnings should be improved by such a choice: seasonal ingredients are an opportunity. However, this clear optimisation step is far from being straightforward for the majority of the culinary industry.