A team at Johns Hopkins University and Stanford University have used video recordings to train a robot to perform surgical procedures. What’s more, the team claims that the robot can perform these tasks “with the skill of a human doctor.” The robot, officially known as the da Vinci Surgical System, was trained to perform three tasks essential to surgical procedures: manipulating a needle, lifting body tissue, and suturing. Researchers created a model that incorporates imitation learning with the “same machine learning architecture that underpins ChatGPT.” Instead of speaking to the machine learning architecture with text and words, researchers used kinematics, “a language that breaks down the angles of robotic motion into math.”
The learning model was fed videos recorded from wrist cameras during surgical procedures. The result? Not only was the robot able to perform to a human doctor’s standards, it was able to correct for obstacles and errors it hadn’t encountered yet.
“All we need is image input and then this AI system finds the right action,” said lead author Ji Woong “Brian” Kim. “We find that even with a few hundred demos, the model is able to learn the procedure and generalize new environments it hasn’t encountered.”
“The model is so good learning things we haven’t taught it,” added senior author Axel Krieger. “Like if it drops the needle, it will automatically pick it up and continue. This isn’t something I taught it do.”
See also: WATCH: Robotic arm learns how to clean by “watching” humans
