It is hard to imagine a script of the film.Instead of unraveling terror between futuristic war machines aware of themselves, perhaps everything will start with a simpler step: designing creative recipes.A few weeks ago, an intelligent machine, Watson, launched all its cognitive ability to invent innovative and sophisticated culinary products, and that knew they would delight diners, serving a multitude burritos of black chocolate with beanssoybean, EDAM cheese, meat and damask.Beyond the initial apocalyptic joke, a machine has been able to create.We do not talk about proposing rational solutions to a problem, but what could be the first steps of original and creative thought.
What makes a machine smart?Last Saturday, a computer program that was passed through a Ukrainian teenager named.When overcoming the Turingtest, many announced the victory of artificial intelligence.To imitate a boy's responses is an important challenge for a program, but the advances in artificial intelligence move by much more sophisticated and complex paths.For example, trying to make a machine demonstrate creativity.
A lobster lid with basil, crispy pork, saffron sauce and piquillo peppers.That dish, as peculiar and delicious according to those who have tried it, is the creation of a computer.Not anyone else, but the famous Watson, a supercomputer developed by IBM capable of winning irony and double senses - something so human - to embark on a popular lotevisive contest.That lobster, next to thousands of more recipes, has served to blur the border of what the machines cannot do: think creatively.
It is not a kitchen robot to dump ingredients inside to return the pudding ready to serve.On the contrary, he is the chief of cooking who invents suggestive dishes that no one had ever done before thanks to their knowledge about flavors, raw materials, chemical processes and human psychology.What Watson intends is not to imitate recognized chefs, but to innovate."We are not trying to overcome the kitchen Turing test, but trying to invent new recipes," one of the designers of this computer creativity system recently summed up.
The head of this IBM researchers team, Florian Pinel, explains that they landed in the kitchen looking for a field in which Watson's creativity could develop."To make a dish there are thousands of billion or trillions of possible combinations.Watson processes them all and only chooses the most interesting, "says Pinel, who in addition to Systems Engineer is Diploma in Culinary Arts.
The challenge was to get a human creativity, that is, to avoid limiting themselves to offering mixtures of random flavors and ingredients and creating solutions knowing that they were going to surprise.The first step to achieve that resolutive creativity was to learn to cook, know why we like certain flavors or what does thequicheunaquiche."We create a food database that contains around 30.000 wikia recipes.com, everything that wikipedia about gastronomy [Watson is able to understand what it reads], as well as information about the ingredients at the molecular level, nutritional information and studies on the perception of flavor in the human palate, ”lists the responsibleof this project, which has already published several works on the experiment, which was kept secret for a time.
Then, the most complex was to develop the algorithms that indicate the steps to follow to develop creative ideas: understand what is the problem to solve, accumulate knowledge to face it, start suggesting ideas and selecting those that work better at all levels.With all the information he has, Watson is able to make all kinds of combinations to obtain unexpected recipes from everything he knows.And from there the chocolate burritos, the kebab vietnamese of apple or the smoked cream and bacon, Pinel's favorite: "Although what I like the most is not a particular recipe, but you can try a new recipe each recipe, but you can try a new recipe, each recipe eachonce and never repeat the same twice, for the rest of his life ".
Calm, he still doesn't fly
Logically, Watson neither puts from the front nor grabs the pans.To put their ideas into practice, IBM researchers have been working with chefs from the New York Culinary Education Institute (ICE), one of the most important US cooking schools.In the first versions of the experiment, Watson just suggested combinations of ingredients based on their cultural, psychological and chemical knowledge, but without indicating neither amounts nor greater explanations in the section of the preparation of the dishes."At first we focus on the production of new combinations of tasty ingredients, because we have worked with chefs and we already know how to cook the dishes.But we are adding algorithms to generate proportions of more precise ingredients and instructions, "says Pinel.
Therefore, much of the challenge was in the hands of the chefs who had to interpret Watson's suggestions.This creative collaboration between the machine and humans is one of the main values that IBM defends of their experience: silicon and neurons working shoulder shoulder to innovate.What taste of the Boca had the chefs that had to do skewers for the computer?"There were proposals that seemed that there would be no way they could become a viable recipe.However, as we worked them we discovered that the dish was delicious, "summarizes James Biscione, director of Culinary Development of ICE and chief chef of the experiments.
"In some cases we have had to work very hard to achieve it, sometimes we had to perform eight or more different tests with the same set of ingredients before reaching a dish that we could serve with pride.But each system proposal has been a success, we have never failed, "says Briscion.As a cook, he assures that the most surprising recipe was the bacon music with peas, peppers, dill and Swiss cheese;The one he liked the most, the lobster lid;And the most successful among the public, a variant of the Catalan Xuixo with cocoa, black pepper, coconut milk and honey.
The science of ingredients
"It is a tool with an incalculable value as a source of inspiration and knowledge in the development of new dishes.Being able to access science that explains that the ingredients know well together is an incredible tool that could help every chef achieve new horizons in their kitchens, "Briscione argues.
IBM's intention is to develop Watson business lines in these areas, in which human creativity is improved with that of the machine.For this same reason, the Glaxosmithkline pharmaceutical.Therefore, doctors are also being helped to make diagnoses, since it is able to take thousands of factors that can influence the health of a person, much more than any doctor - except the Doctor House—.
"Beyond the food industry, we could use a similar approach to areas where we can establish a parallel with cooking recipes," says Pinel, "like industries that create new products from component combinations, such as perfumes, such as perfumes,and even for the preparation of recommended itineraries for travelers ".