2 ms·
Now I want to see DeepMind apply AlphaZero to Rubik's solving.
by fastball 3y ago
Now I want to see DeepMind apply AlphaZero to Rubik's solving.
- nsgodshall 3y agoRubick's cube's are "solved". Traditional computer programs can determine the most efficient solves already.
- fastball 3y agoRight, but per the article, Korf’s algorithm takes hours and Thistlethwaite’s algorithm doesn't necessarily find the optimal solution. I wonder if a self-learning AlphaZero approach could result in finding optimal solutions in seconds.