The engineers at US-based tech giant Google, who are working on artificial intelligence, “i-S2RThey developed a robotic arm called ” 340 times rally got it done.

Google engineers, robotic arm Sim2Real They developed it using a training method called This method involved testing the artificial intelligence training in a real-world simulation environment. With this method, artificial intelligence training, which sometimes took years, could be completed in a few days or even hours. Here is this technique, google‘s development of a robotic arm that beats professionals.

Kensi can even execute her strategy!

The rule to win at table tennis, it’s a good strategy. Of course it’s also important to meet the incoming ball, but if you don’t tire the opponent, it won’t be easy to win the match. Here, Google’s artificial intelligence can develop this strategy itself. In addition, it is mathematical sensitivity it doesn’t matter.

Developing such a successful AI was not an easy task. Because the Sim2Real technique, human-robot interaction It wasn’t so good for him. It was not easy to simulate what a human can do. Google was aware of this and conducted a survey that ranged from unsuccessful engineers in table tennis to more successful engineers. Artificial intelligence in every new competitor gained new knowledge and healed them. On the last point, a robot has been developed that can rally 340 times.

340 rally, no world record

We have done a survey for our readers who are not very interested in table tennis. Guinness Book of RecordsLooking at the data in the world’s longest table tennis rally 11 hours 50 minutes 36 seconds We see it continue. 340 rallies in table tennis played with artificial intelligence developed by Google have a duration of 4 minutes. In other words, Google’s artificial intelligence is at a level of sophistication that could break records for now. does not have any.

You can watch Google’s artificial intelligence 340 rally below: