The American technology leader Google announced the development of a new robot that outperforms humans in table tennis, by equipping it with artificial intelligence technologies.
Researchers at Google DeepMind, Google’s artificial intelligence research lab, have turned an ABB robot arm into a competitive table tennis player, equipping it with a 3D-printed paddle that allows it to move extremely efficiently and outperform its human competitors.
In the study published by the researchers, the robot’s arm is mounted on two linear bridges that allow it to move sideways, enhancing its ability to compete. To win, the researchers must ensure that the robot can make the best tactical decisions in record time.
The robot was equipped with a system consisting of two parts: low-level skill policies and a high-level controller. In training it, Google researchers used computer simulations to develop hitting skills, and then modified these skills using real-world data, allowing the robot to gradually improve its performance. This innovative training method reflects the great progress that artificial intelligence has made in understanding and analyzing sports.
Through a series of experiments, the bot achieved a 45% win rate in matches and 46% in solo games against beginners, while it recorded a 55% win rate against intermediate players.
In contrast, the device lost all of its matches against advanced players, indicating that the robot arm had already achieved average human play in the confrontations.
Technology companies are looking to explore technological innovations and advanced experiences to enhance their competitive capabilities by pushing the boundaries of artificial intelligence and machine learning, enabling them to improve their products and provide more efficient services.
