A robotic arm integrated with technology developed by Google DeepMind Reaching amateur level in table tennis, They won 45% of their matches. Against 29 human players with different abilities. Against the beginners, the machine won every match.
The news was published in a scientific article. A video showing highlights from the match has also been released. Check it out below:
To develop the system, the researchers created a large dataset of the ball’s initial states, including information about its position, spin and speed. The AI was then trained through virtual simulations to learn different game techniques. The robotic arm, an ABB model IRB 1100, has 20 cameras to track an opponent’s racket.
The result is a major advance in this type of research, as ping pong is seen as a very important activity in the development and training of cognition and the results that robots can provide, as it requires a wide range of strategic skills, visual coordination and complex physical movements in real time.
The progress in this ping pong example represents progress in other contexts where robots can be used, performing tasks that involve skills in environments such as homes and warehouses.
Mapping variables
Development will continue to improve the machine's perception of the wide range of variables present in a real match, such as wind speed or even a change in trajectory due to dust on the table, for example.
Google DeepMind believes these limitations can be addressed in several ways, including developing predictive AI models designed to predict the ball’s trajectory as well as providing better collision detection algorithms.
Human players who participated in the test described the experience as: “Hazar” and “attractive”Regardless of the outcome of the match. Many have also expressed great interest in playing against the bot again in the future.
