Machine learning algorithms can now solve a Rubik’s Cube without human help
According to CNET, creating a machine learning algorithm that can solve the Rubik’s Cube is relatively simple. However, addressing the Rubik’s Cube without human help may be an entirely different task. Stephen McAleer and his colleagues at the University of California believe that they have solved this problem and adopted a process called “autodidactic iteration.”
McAleer and his team called the process is “novel reinforcement learning algorithm that can teach itself how to solve the Rubik’s Cube with no human assistance.” They claim that this learning algorithm can be solved 100% within 30 steps. Randomly disrupts the Rubik’s Cube – this is comparable to human performance or human performance.
This type of algorithm differs from the algorithm that surpasses human performance in board games. These systems are “reinforced learning” systems. The system will see if each proposed step is an improvement after completing the puzzle solving.
You can read the full paper here.