Have you ever solved the Rubik’s Cube? For those who don’t know what Rubik’s Cube is, here is a brief about it.
The Rubik’s cube (please see the image below) was invented in 1974 by Erno Rubik, a Hungarian architect. However, it was never meant to be a toy. Erno just wanted a working model to help explain three-dimensional geometry.
The cube, made up of nine coloured squares on each side, can be rearranged in 43 quintillion different ways. That is 43,000,000,000,000,000,000.
Across the world, several “speed cubing” competitions are held, where people try to solve the cube as quickly as possible. This is no simple task. Why? Because the cube has completion paths numbering in the billions but only one goal state — each of six sides displaying a solid color — which apparently can’t be found through random moves. The current world record for solving the cube is held by Yusheng Du, who can solve it at an amazing 3.47 seconds.
That’s good enough. Eh!
Interestingly I came across this article on artificial intelligence by University of California, Irvine. They have created DeepCubeA. It is an artificial intelligence system created by researchers at the University.
It is a ‘a deep reinforcement learning algorithm programmed by UC Irvine computer scientists and mathematicians that can find the solution in a fraction of a second without any specific domain knowledge or in-game coaching from humans’. Moreover, this AI takes about 20 moves, most of the time solving Rubik’s Cube in the minimum number of steps.
The research is published in Nature Machine Intelligence. The researchers in the study demonstrated that DeepCubeA solved 100 percent of all test configurations, finding the shortest path to the goal state for Rubik’s cube about 60 percent of the time.
Therefore, the algorithm also works on other combinatorial games such as the sliding tile puzzle, Lights Out and Sokoban.
Senior author Pierre Baldi, UC Irvine distinguished professor of computer science is the man behind this AI. Baldi is involved in projects that research to build the next generation of AI systems.
People across the world are touched by artificial intelligence every day through apps such as Siri and Alexa. Besides, recommendation engines working behind the scenes of their favourite online services prompt suggestions and links that are based on your search history.
(This was an interesting article for me and I thought of sharing it with my friends at the campus.)