Last week, experts from Deepmind published a study on teaching miniature humanoid robots complex movement skills and strategic understanding through deep training with AI reinforcement. The study outlines how the team trained the robots to play one-on-one football, first practicing skills alone and later combining them with match settings.
The report includes a 27-second video demonstration of a robot being pushed by a project employee several times until it falls to the floor and manages to get back up to score a goal. The video caused controversy on Twitter with users expressing their displeasure at the treatment of the robot.
Despite the backlash, the researchers have highlighted that the modifications made to the robots’ equipment and the regulator of their behavior during training has led to safe and effective movements while maintaining their dynamic and mobile abilities.
The demonstration of the robots’ football skills is a significant development in the field of robotics and AI. It shows the potential of these technologies to create robots capable of complex movements and strategic thinking. The use of reinforcement learning has played a crucial role in training the robots and could pave the way for further advancements in the field.