Asia Tech Wire (Aug 12) -- Google's DeepMind has announced that its R&D team has developed a table tennis robot that can reach the level of a human amateur in a match.
This is the first learning robot agent to reach the level of human amateurs in a table tennis match, and its main body is a six-axis robotic arm that can move back and forth, left and right, via a bottom slide.
In 29 matches against humans, the robot won 13, a 45% win rate.
The robot lost all matches against advanced players, but won all matches against beginners and 55% of matches against intermediates, DeepMind researchers said.