Verbal sentences are very natural to us humans, who know almost instinctively what you can and cannot do with your phone.

You can lift it, for example, open it or close it, but not eat it or drink it.

Robots on the other hand do not have such hires and training them in verbal formulas alone is not easy, struggling to learn how humans interact with the things of their world.

This is the problem that Brigham Young University (BYU) scientists are working on, employing an unconventional method, Wikipedia!

“When researchers let go of robots or artificial intelligence agents in unstructured environments, crazy things are being attempted by all reasoners,” said one scientist (Ben Murdoch), “common sense about what you can do with the objects are completely missing and we end up with robots that spend thousands of hours trying to eat a table. “

And so they developed a method that teaches artificial intelligence all the actions that can be applied to an object by cross-linking them with accepted verb-object combinations found in the electronic encyclopedia!

To test what robots learned from Wikipedia, scientists put them to play board-based (text-based) games with humans and saw that the Wikipedia solution “had improved computer performance in 12 out of 16 games”!

Artificial intelligence therefore learns man’s verbal energies so that he can serve him even more efficiently at some point.