The Rubik’s cube, a 3D colour combination puzzle, was invented in 1974 by Hungarian sculptor, Erno Rubik and has to this day been a pain in the neck to people across the globe. But this difficulty has got nothing on DeepCubeA, an AI algorithm that solved it and believe it or not, the time it took was less than the one you spent to read the last word on this statement.
The number of hyper-intelligent machines seems to keep growing over time especially since this comes days after another report of a poker-genius algorithm left the game’s world champions in dismay. This week, however, the University of California based in Irvine announced that the artificial intelligence system solved the puzzle in just a second.
This milestone comes as a thorough beating to the human world record by more than two seconds. DeepCubeA, a reinforcement-learning algorithm was programmed by scientists and mathematicians from the university and immediately given the ultimate test. The study saw the algorithm given 10 billion different combinations of the puzzle with a target to decode all of them within 30 moves.
Mind you, all this was done before the algorithm was fed with any prior knowledge of the game or coaching from any of its human handlers as the report says. “Artificial intelligence can defeat the world’s best human chess and Go players, but some of the more difficult puzzles, such as the Rubik’s Cube, had not been solved by computers, so we thought they were open for AI approaches,” said senior author Professor Pierre Baldi in a statement released by the university.
It is, however, safe to note that this is not the first algorithm that solved the puzzle as all this credit goes to min2phase, a system developed at the Massachusetts Institute of Technology (MIT) that solved the puzzle three times faster. But unlike DeepCubeA, the system didn’t use a neural network that mimics how the human brain works – or machine learning techniques and was programmed just to solve the puzzle.
All this achievement as a complement to the project created in 2018 that solved the same puzzle in 0.38 seconds. Creating a system that teaches itself to complete the challenges may then just be seen as the first step to making AI a real-life problem solver, an idea that the general tech industry will be glad to embrace in years to come.
And in the words of Prof. Baldi, “How do we create advanced AI that is smarter, more robust and capable of reasoning, understanding and planning? This work is a step toward this hefty goal.”