People have now accepted that they will never be as good at chess as they are robot, but now even the robots have to accept that they will never be as good as other robots!
A new artificial intelligence platform, known as AlphaZero, can learn to play Go, Chess and Shogi games without any human intervention. Using deep neural networks, AlphaZero quickly learned how to play each game "to be the best player in history".
AlphaZero was unveiled by DeepMind Technologies in a survey published in Science on 6 November. DeepMind, a British subsidiary of Alphabet, the parent company of Google, has been working with Go AI for many years. 2017, DeepMind withdrew the former AlphaGo champion, but continued to deal with AI. With AlphaZero, DeepMind's research has reached its zenith.
The program faced the best AI in the world in three board games:
- Stockfish, world chess champion AI
- elmo, winner of the 27 annual World Computer Shogi Championship 2017
- AlphaGo Zero, DeepMind's Go AI, the strongest Go player in history
In any case, AlphaZero only got the knowledge about the basic rules of the game. Before confronting the rest of the AI, he played millions of games against himself, starting from random tactics and trying to win, but slowly learned which strategies work best through a trial and error process called reinforcement learning.
The training and learning process took nine hours for chess, 12 hours for shogi and 13 days for Go using 5.000 TPUs (Tensor Processing Units). Only one TPU can process over 100 million photos per day on Google Photos. As soon as the learning process was completed, AlphaZero competed with the rest. And he crushed them!
What is unique in the study is the fact that algorithm learning was combined with a "search method" called Monte Carlo tree search (MCTS). This is how the Go AI programs determine which traffic should be done thereafter. The DeepMind team used the same chess and shogi system, showing for the first time that it could adapt to other composite games as well.
Perhaps more interesting for the chess players is the fact that AlphaZero, without having human hands, can and does its knowledge, applies strategic and innovative ideas that have never before seen. The aggressive style and extremely dynamic game surprised Chess grandmaster, Matthew Sadler, who spoke about this at DeepMind blog.
These unique strategies and capabilities make AlphaZero a great teaching tool for chess players.
The dominance of artificial intelligence on humans is quite common in the world of games, with robots winning in board games, complex video games like Dota 2 and of course Go.
Does that mean that AI is ready to literally beat us in every competing game ever invented? Fortunately no. Although the three games DeepMind is extremely sophisticated, they provide some advantages for AI as they include two players and all the information necessary for the next move is always visible.
So while they are definitely taking on the role of champion in games like chess, the robots will probably never beat us in Texas Hold 'Em.