Employees of the design agency AKQA from the USA presented a new sport, the development of which was carried out by a neural network. The rules for a new team ball game called Speedgate were created by an algorithm based on a neural network that learned textual data about 400 sports. Ultimately, the system generated about 1000 new rules for various sports. Further processing of information was carried out by the authors of the project, who tried to test the games invented by artificial intelligence.
Speedgate is played by two teams, each with six players. The action takes place on a 55-meter rectangular field, at the beginning, middle and end of which there are gates. The gameplay begins with the fact that a member of one of the teams knocks the ball through the central gate. After that, the attackers' task is to score the ball into the opponent's goal as many times as possible, avoiding hitting the goal in the middle of the field. Players are prohibited from crossing the border of the area where the central gate is installed. Otherwise, a violation is counted and the ball goes to the other team. An ordinary rugby ball acts as a sports equipment. One of the rules of the game says that the ball must move every three seconds, so the participants in the competition have to constantly be in motion. One full match consists of three halves of 7 minutes, between which two-minute breaks are included. If a draw is recorded in regular time, then three extra halves of 3 minutes each are assigned.
In addition, the developers have created the official logo of the new game. It was generated by a neural network that had previously studied 10 logos of different sports teams. Negotiations are currently underway to create the first sports league based on Speedgate.
Source: 3dnews.ru