Whether it’s burgers, tater tots or fried octopus balls, there are plenty of things robots can cook, and plenty of robots out there trying. But those robots are pretty rigid, following a specific set of criteria. The whole point is that they take over the repetitive task of cooking the same thing over and over and prepare it consistently.
But researchers at the University of Cambridge released a paper this week showing how they are teaching a robot to not just make an egg omelet for you, but also learn from how you like your eggs and adjust the cooking accordingly.
IEE Spectrum has the full story (and an interview with the paper’s authors), and explains how the robot is able to do this customized cooking:
The researchers employed a solution to this problem called batch Bayesian optimization. The more traditional sequential Bayesian optimization would take a human’s score of each omelet and use it to modify the cooking process for the next omelet, but this approach doesn’t work well because the human feedback is, as the researchers tell us, “often ambiguous and relative.” By running the optimization process only after all scores have been collected, the robot is able to explore many more combinations of variables, yielding a substantially better end result. Instead of the omelets gradually getting better as you go, you’ll instead be tasting them randomly, but you’ll end up with a much tastier omelet.
In this case, the robot is able to adjust a number of different variables such as the amount of salt/pepper, whisking time and cook time.
Adding this type of customization to a robot’s repertoire and being able to apply it to more kinds of food could help automated cooking move from the more mass, industrial-sized food prep of fast food and more into personal cooking. Having your own robo-chef that can cook eggs just as you like them while you get ready in the morning (or have dinner waiting for you when you get home), is an idea that would be welcome in a lot of homes.