A robot named Ace, developed by Sony AI, has achieved a significant milestone by defeating professional table tennis players. The robot utilizes a single arm with eight joints and nine camera eyes to track the ball and detect spin.
Researchers trained Ace using reinforcement learning, an AI method that allows the robot to learn from experience, mirroring the necessity of training for human athletes. This advancement highlights how AI can enhance robot speed and agility.
Experiments were conducted on a full-sized Olympic table tennis court under official rules. Sony AI's president, Michael Spranger, stated the goal was to create a robot that plays the game fairly, matching the skill and decision-making capabilities of a trained human athlete.
Sony's team continued to refine Ace, leading to improvements in speed, rally length, and aggressive play. The robot has demonstrated capabilities previously considered impossible for humans, pushing the boundaries of AI and robotics research.