Culinary Creativity Is An Algorithm, And Computers Are Learning It

Looks like computers are well on their way to learning creativity. Well, sort of. IBM's apparently figured out how to distill creativity down to programming algorithms, at least to some extent.  According to the organization, it's discovered that the best way to invent new recipes is to get help...from a supercomputer.

The company has begun serving what it calls Data-Driven Delicacies - dishes created by a cognitive computer system - out of a food truck that's been making the rounds at a number of industry events such as IBM Pulse and SXSW. For the moment, the truck's basically a promotional gimmick, IBM's way of saying "look what our computers can do!"  That isn't always going to be the case, however.

As time goes on and the technology driving IBM's supercomputer becomes more advanced, computer-generated food is going to become increasingly more common. See, while human intuition and creativity are great, they can be somewhat unreliable - if media distributors can nail down the algorithms behind what makes a particular work successful, they can use a computer system to duplicate the process of creation.

It's something they've been doing for quite some time already - I'm certain I've mentioned it before. Already, artificial intelligence has managed to duplicate literature, creating a novel that combines Leo Tolstoy with the style of famous Japanese author Haruki Murakami. It's written popular and beautiful music. It's even managed to cover the news, to some extent. .In short, artificial intelligence is distilling the science behind the arts into computer algorithms. The question, of course, is how far these developments are going to go - could computers eventually replace artists, writers, and master chefs?

Maybe. For the time being, the process is anything but perfect. How it works is deceptively simple: a computer creates a huge database of different recipes and components. With the oversight of researchers, a system can then parse these into full meals. Over time, the computers learn how best to string these components together. 

According to IBM researcher Lav Varshney, the system could potentially produce menus for individuals with even the strangest dietary requirements or the most unusual tastes. Already, he explains, it's manufacturing recipes that even some of the greatest chefs of our time failed to consider. See, one thing IBM's showing with this project is that human cooks - no matter how great - tend to have biases. They may, as a result, discount any of a wide number of particularly delectable combinations. 

One of the IBM food truck's current recipes, for example, The Swiss-Thai Asparagus Quiche" combines Swiss gruyère, Greek feta, and Thai curry spices. It's a recipe that most wouldn't even consider preparing, yet one which just so happens to be absolutely delicious. See, unlike a traditional chef - who's limited by their own experience - the computer can actually look at the flavor-related molecules in each ingredient, combining them to incredible effect.

Of course, as I'm sure we've already seen, these biases can extend beyond food preparation, too. Take House of Cards, a political thriller which just entered its second season on Netflix. What many might not know was that - though it was written entirely by humans - the concept was heavily informed by user data collected by Netflix. It knew that its viewers enjoyed Kevin Spacey, and a good proportion of those fans also like director David Fincher. Finally, the majority of those who were fans of both enjoyed political dramas.

So...taken together, what does all this mean?  

Robots might have distilled a certain portion of creativity down to the basic science behind it, but there's probably still something missing from robot-created works. Without injecting at least some sort of human element, artificial intelligence still isn't capable of completely replicating the creative process. Not yet, anyway- there's no saying it won't be in the near future. 

I'm not looking forward to that day, myself. 

Mar 5, 2014
by Anonymous

Interestingly, not

Interestingly, not everything that every great chef comes up with is palatable either. At the pinnacle it all reduces to experiment with jury decision. Sounds like IBM reached the pinnacle. I bet with a sufficient jury IBM could get 3 Michelen stars.