The Rubik's Cube has been solved by a self-taught machine that required absolutely no human help, the Los Angeles Times noted.
Less than 6 percent of the human population can solve the 3-D combination puzzle, however, a group of University of California, Irvine researchers revealed a system that can teach itself to do so within 30 moves.
Recently, scientists created a robot that solved the Rubik's Cube in 0.38 seconds, but the development of a self-taught puzzle-solving machine pushes artificial intelligence to go beyond pattern recognition.
"At first I didn't think it was possible to solve the Rubik’s Cube without any human data or knowledge," said Stephen McAleer, a UCI doctoral student, according to the Los Angeles Times.
The trick lies in two algorithms, the first in which the machine teaches itself the moves to solve the puzzle by working backward, and the second in which the trained neural networks then use the learned moves to solve the puzzle.
At the basis of this latest development is research that teaches artificial intelligence to play the strategy game Go by using neural networks in conjunction with the state of-the-art Monte Carlo tree search programs.
McAleer said that, in order to solve the Rubik’s Cube, "this artificial intelligence has to reason symbolically," and has to "think about how it’s going to manipulate this mathematical structure," the Los Angeles Times reported.
IBM is also pushing limits when it comes to how artificial intelligence approaches problem solving.
Last week the tech giant revealed a computer that can successfully debate with humans and win these arguments.
The system, dubbed Project Debater, was trained in debating methods prior to being pitted against two human debaters, however, it did not receive any information about the debate until moments before.
Despite this, Project Debater was still able to nudge nine audience members toward its stance that the use of telemedicine should be increased in one of two debates.
© 2025 Newsmax. All rights reserved.