Al is going for Go

摘要

Google just upped the stakes. The firm's UK-based artificial intelligence subsidiary, DeepMind, has developed the first program capable of beating a human professional player at the ancient Chinese game of Go. With this long-standing Al challenge felled, the system could tackle complex real-world problems. Go is played on a 19 by 19 board, which allows for 10~(171) valid layouts, versus roughly 10~(50) possible configurations of an 8 by 8 chess board. As such, DeepMind's AlphaGo program has to work harder than IBM's Deep Blue, which beat chess champion Garry Kasparov in 1997.