AlphaZero researchers, a possible "computing" program, provided fairly purely only in some of the most complex board games that are known, such as chess, shogi and Go, but also to teach themselves to play without prior knowledge, except for the rules of each game. "
The system was created by experts from DeepMind, a British company with the Alphabet, a Google main company.
The results, the researchers said, that it is an important step in the development of an AI player that can quickly learn and match any game.
Chess, shogi and go
As the IBM Deep Blue chess program has influenced Gary Kasparov's world-wide concert in 1997, AI games have become more advanced and can affect people in games that are increasingly complex.
Even in other straightforward strategic games, such as Shogi and Go, both much more complicated than chess, tools have also been compiled.
The AlphaZero author, developed by David Silver and his co-workers at DeepMind, is avoids the need for information derived from people and, instead, capable of learning chess, shogi and Take to & # 39; play again against each other so that they can survive each of the games
The system won the most advanced AI programs that were specialized in these three games after a few hours of self-learning.
As a result, researchers say, they must now turn into a new generation of games, such as multi-game video games, to address the next set of challenges for AI systems.
<! – – Download down a document linked to this news ->