Robots have just been given the ability to put pen to paper, thanks to a new algorithm that helps them write words just like humans.
The algorithm uses deep learning networks that analyze images of handwritten words or sketches in order to identify the pen strokes that created them. After that, the robot is able to recreate the movements it just learned.
Researchers believe that, in time, this type of robot will have significant input in communication with human coworkers to better understand their needs.
“Just by looking at a target image of a word or sketch, the robot can reproduce each stroke as one continuous action,” says Atsunobu Kotani, an undergraduate student at Brown University who led the algorithm’s development.
“That makes it hard for people to distinguish if it was written by the robot or actually written by a human.”
Another remarkable aspect is the robot’s accuracy when reproducing sketches. When he trained his deep learning algorithm, Kotani used a set of Japanese characters, and, to his surprise, the algorithm reproduced sketches with 93 percent accuracy.
“We would have been happy if it had only learned the Japanese characters,” Tellex says. “But once it started working on English, we were amazed. Then we decided to see how far we could take it.”
So, he drew a sketch of Mona Lisa on a dry erase board in Tellex’ lab and couldn’t believe when the robot copied him and made a lookalike.