This Japanese robot can now peel a banana
The robot’s two large mechanical arms are designed to mimic the movements of real human arms and hands, which means the robot’s “brain” must understand how human hands work. Unfortunately, programming dexterity into a robot is not as simple as writing a few lines of computer code. This is where AI deep learning – specifically, deep imitation learning technology – comes in. If you’re familiar with concepts like NVIDIA DLSS or Google DeepDream, you might already understand how artificial intelligence reuses existing data to create something new.
Similarly, AI can learn to model real human behavior by “observing” humans in real life. If you show an AI hundreds of videos of different people peeling bananas in different settings and with slightly different movements, you can paint a pretty big picture of what the process is supposed to look like. In theory, this can provide the AI’s deep imitation learning algorithm with enough direction to then complete each of the complex steps involved in holding the banana, pressing the tip, and unwrapping the fruit peel by skin. At least that’s how researchers at the University of Tokyo effectively taught the robot to peel bananas with a 57% success rate in over 13 hours.
Researchers hope that robots like this can make life easier in Japanese food processing facilities or even fill shortages of food service workers. However, it may take some time before these bots ship in droves.