Last year DeepMind's AlphaGo programme took on and beat two of the world's best players of the Chinese game of Go（围棋）-an unbelievable achievement seen as a milestone in the development of artificial intelligence (A). That programme was trained by first giving it vast amounts of data from amateur and professional games.
The new AlphaGo Zero began with a blank board and nothing but the rules of Go and set about playing against itself. Within three days it was so advanced that it took on the previous version, which had taken months to develop, and defeated it by a hundred games to zero.
DeepMind says that creating knowledge from first principles without learning from human expertise（专门技能）is an important step in developing artificial intelligence. The company's founder Demis Hassabis said this breakthrough could help in real world problems such as designing new drugs or discovering new materials.
In 1997, Garry Kasparov was defeated by Deep Blue, a computer program written by IBM, running on a supercomputer. This was the first time that a ruling world chess champion was defeated by a computer program in tournament conditions. Superficially, AlphaGo's win against Lee Sedol can be compared to Deep Blue's win against Gary Kasparov, which happened almost 20 years ago. So: what's the big deal? We have to understand the differences between chess and Go.
In chess, each player begins with 16 pieces of six different types. Each piece of type moves differently. The goal of the game is to seize the other player's king. Go starts with an empty board. The goal of the game is to gain as much territory（领地）as possible.
Although the rules of Go might appear simpler than those of chess, the complexity of Go is higher. Also, games usually last longer for lots of moves. A typical game in Go might last for 150 moves vs.80 in chess.
