Researchers at the University of Cambridge have trained a robotic ‘chef’ to recreate dishes simply by watching cooking videos.
As part of the experiment, the team programmed the robot with eight simple salad recipes made up of ‘chopped fruits and vegetables’. It was then monitored while it watched 16 recipe videos, eventually able to identify which recipe was being made and prepare the recipe itself.
The robot was even able to come up with a totally new ninth recipe, based on what it had learned.
'We wanted to see whether we could train a robot chef to learn in the same incremental way that humans can – by identifying the ingredients and how they go together in the dish,' said researcher Grzegorz Sochacki.
The robot correctly identified each recipe 93% of the time and was able to detect subtle variations between each one. For example, it was able to tell that if a human was holding a knife in one hand and a carrot in the other, it would then proceed to chop up the carrot. It was also able to recognise larger portions and normal human error.
While Sochacki does admit that the recipes ‘aren’t complex’, he goes on to explain that the experiment was ‘really effective at recognising, for example, that two chopped apples and two chopped carrots is the same recipe as three chopped apples and three chopped carrots.”
There are limits to the robot’s abilities, however. 'Our robot isn’t interested in the sorts of food videos that go viral on social media – they’re simply too hard to follow,” said Sochacki. 'But as these robot chefs get better and faster at identifying ingredients in food videos, they might be able to use sites like YouTube to learn a whole range of recipes.'
In other foodie news, find out everywhere Beyonce and Jay-Z have been eating while they're in London.