AI robot learns Jenga mastery
Researchers at the renowned Massachusetts Institute of Technology (MIT) in the USA have taught a robot with artificial intelligence to play Jenga. However, not to replace a human playmate but, as explained in
the scientific treatise, there’s more to it than that.
After starting up, the robot went straight to work and tried to push the Jenga bricks out of the stack and put them back on top. The two gripper arms measure the force needed to move the wooden brick, and the robot can decide whether the part can be successfully removed. After moving about 300 bricks, the robot and the AI had an understanding of the physics behind the game. “The robot builds clusters and then learns models for each of these clusters, rather than learning a model that captures absolutely everything that could happen,” says Nima Fazeli, lead author of the paper.
In a video, the robot proves that it has actually already mastered the game quite well. Watch the video until the end and you will see how the Jenga game looks through the “eyes” of the robot.
Behind such projects there is, of course, a deeper meaning than just Jenga. The complex tasks that the robot has to perform in combination with the controlling AI provide a basic understanding of physics in the system that can also be used for other things. Scenarios like this make development more tangible and understandable for the layman and steer the entire development in the field in a practical direction – which is certainly not a bad idea.
MIT Professor, Alberto Rodriguez, says games like Jenga require “the mastery of physical skills such as probing, pushing, pulling, placing and aligning parts.” The experience gained from the project should benefit the industry by making robots more sensitive and precise than they were before.