Gordon Rams-AI! Scientists develop a robot CHEF that can recreate recipes by watching videos 

Gordon Ramsay better watch out, because there’s a new top chef in town – in the form of a robot.

The robo chef can learn how to make the perfect dish simply by watching cooking videos.

Researchers at the University of Cambridge have programmed a machine to prepare a meal by following how a human makes it.

Using advanced AI, the robot can determine from any frame which objects it is looking at – such as a vegetable, hand or knife – and how it is being used.

Over time, it’s then able to identify which ingredients work best together – even pointing out when the human chef might have used the wrong amount.

Gordon Ramsay better watch out, because there’s a new top chef in town – in the form of a robot

Robot chefs have been featured in science fiction for decades, but in reality, cooking is a challenging problem for a robot.

Several commercial companies have built prototype robot chefs, though none of these are currently commercially available, and they are far behind their human counterparts in terms of skills.

Study co-author Grzegorz Sochacki, a PHD student from Cambridge’s Bio-Inspired Robotics Laboratory, said: ‘We wanted to see if we could train a robot chef to learn in the same incremental way humans do – by identifying the ingredients and how they go together. in court. It’s amazing how much nuance the robot could detect.’

He added: ‘Our robot isn’t interested in the kind of food videos that go viral on social media – they’re just too hard to follow.

“But as these robot chefs get better and faster at identifying ingredients in food videos, maybe they can use sites like YouTube to learn a whole host of recipes.”

It is hoped that the findings – published in the journal IEEE Access – will make it easier and less expensive to deploy robot chefs.

The researchers first made a ‘cookbook’ of eight simple salad recipes and filmed themselves making it.

Using advanced AI, the robot can determine from any frame which objects it is looking at – such as a vegetable, hand or knife – and how it is being used. Over time, it is able to identify which ingredients work best together – and even indicate when the human chef may have used the wrong amount

The robot chef – who was trained on neural networks, which essentially mimic the workings of the human brain – then watched 16 of these videos.

By identifying the ingredients and the human chef’s actions, he was then able to correctly determine what was being prepared 93 percent of the time.

It could also detect minor variations in the recipe, such as a double portion.

The robot was also able to recognize a brand new ninth salad, add it to its cookbook and then create it.

The research was supported in part by Beko plc and the Engineering and Physical Sciences Research Council (EPSRC), part of UK Research and Innovation (UKRI).