Artificial Intelligence Can Solve Rubik’s Cube
Since its discovery by a Hungarian architect in 1974, the Rubik’s Cube has furrowed the brows of many who’ve tried to solve it, however, the 3-D logic puzzle is no match for an artificial intelligence system designed by researchers at the University of California, Irvine.
DeepCubeA, a deep reinforcement learning algorithm programmed by UCI computer scientists and mathematicians, can discover the answer in a fraction of a second, without any specific domain information or in-game coaching from humans. That is no simple activity considering that the cube has completion paths numbering within the billions however only one objective state, each of six sides displaying a solid color which apparently cannot be discovered by random moves.
For a study printed today in Nature Machine Intelligence, the researchers demonstrated that DeepCubeA solved 100 % of all test configurations, discovering the shortest path to the objective state about 60 % of the time. The algorithm additionally works on different combinatorial games such because the sliding tile puzzle, Lights Out and Sokoban.
“Artificial intelligence can defeat the world’s finest human chess and Go players, however some of the more difficult puzzles, such because the Rubik’s Cube, had not been solved by computer systems, so we thought they have been open for AI approaches,” mentioned senior author Pierre Baldi, UCI Distinguished Professor of computer science. “The solution to the Rubik’s Cube includes more symbolic, mathematical and abstract thinking, so a deep learning machine that may crack such a puzzle is getting closer to becoming a system that may think, reason, plan and make decisions.”