Sign In

Communications of the ACM

ACM TechNews

Machine Taught Itself to Solve Rubik's Cube Without Human Help, UC Irvine Researchers Say


playing Rubik's Cube

A computer has taught itself to solve a Rubik's Cube puzzle without human aid, according to University of California, Irvine (UCI) researchers. UCI's Stephen McAleer says this was achieved by presenting an advanced system with a solved Rubik's Cube and then letting the computer descramble it piecemeal.

An "autodidactic iteration" algorithm initially works in reverse to teach itself the moves that solve the puzzle. A second algorithm then enlists the trained neural networks to apply the moves learned by the first algorithm to solve the cube. McAleer notes the machine plays with the cube and learns how to descramble it from any starting point.

McAleer says the next challenge is to apply this research to biology, for example in protein folding, in which a protein structure assumes its functional configuration or conformation.

From Los Angeles Times
View Full Article

 

Abstracts Copyright © 2018 Information Inc., Bethesda, Maryland, USA


 

No entries found