We all know very well that in recent years, DeepMind’s artificial intelligence (AI) has been defeating Go international champions, an old-age game of high complexity. However, according to the current reports, the tech giant Google’s AlphaGo AI can teach itself to master games like chess.
Google’s AI Masters Chess In Just 4 Hours! Defeats World’s Best Chess Program
This time it is not Go: the tech giant Google’s artificial intelligence learned alone to play chess and its Japanese version, called shogi. This smart AI without any human intervention managed to learn from other games, only with basic knowledge of the rules of board games.
AlphaGo Zero, the evolution of AlphaGo had already achieved this feat in the same way with Go, the Chinese game. The first algorithm, AlphaGo, needed to use a data set of more than 100,000 matches to learn the best moves.
Earlier we reported that a game – called Go, Weiqi or Baduk – consists of a board of 19 × 19 frames with white and black polka dots; the goal is to surround the opponent’s pieces. There are more than 10171 possible positions, against 1050 in chess.
The original AlphaGo used a data set of over 100,000 matches between two humans to learn the best tactics. Meanwhile, AlphaGo Zero was only programmed with Go’s basic rules, and it learned everything by itself.
The algorithm developed its playing abilities against itself. It started with random moves on the board, and each time it won, it would upgrade itself and play again. This process was repeated millions of times.
After three days, AlphaGo Zero was already able to completely defeat the used AlphaGo against former world champion Lee Sedol, where It defeated the champion every time (hundred matches) and won all the matches. And after forty days, it faced a more advanced version of the original AlphaGo, defeating it 90% of the time.
This time, AI is just called AlphaZero – it’s a more generic version of AlphaGo Zero, but powered by chess rules. And it reached a superhuman level of skill in all three board games in less than 24 hours.
It is worth noting that, between Go, chess and shogi, the Chinese game continues to be the most complicated: it has 10 to the power 171 possible positions, against 10 to the power 50 in chess. Still, the result shows AlphaZero’s “learning by reinforcement” capability, in which the algorithm learns without being taught how a task should be performed.
In just four hours, AlphaZero was able to defeat the world chess champion, which is another algorithm christened Stockfish. Just like in Go, the AI that created its own techniques to win.
When the AlphaGo Zero was put into practice, the strategies put together for Go have never been seen before. Since the algorithm learned from itself, and it overcame “the limitations of human knowledge,” as David Silver, senior programmer at AlphaGo Zero, said at a news conference.
However, as Engadget recalls, AlphaZero is not yet an algorithm that has the general purpose, which is the big challenge of AI today. You can not put it to zero StarCraft II, for example, since superhuman abilities are restricted to specific tasks.
In any case, it is expected that DeepMind, the creator of these algorithms, will continue to apply the results to other sectors. Last year, CEO of the company, Demis Hassabis, said that one-day AlphaGo Zero could be applied even in creating new medicines. A lot of good stuff comes around.
So, what do you think about this Super Human like AI (Artificial Intelligence)? Simply share all your views and thoughts in the comment section below.