Have you ever played Go? It is a boardgame composed of a 19 x 19 grid and black and white stones. Two players take turns placing stones on intersections with a goal to capture as much territory as possible. Confused? Yeah, so was I when I first tried playing it.
I first got interested in Go when I was reading a manga called Hikaru No Go. It made the game seem so interesting and complicated. It was hard not to go research about it and figure out how it played. Since then, I have been playing on and off, but I greatly enjoy the game. I would not say that I am a very good player, but I do know how to play. In reality, I really only have one friend whom I play against (and whom I demand a rematch!) and he tends to kick my ass. Nonetheless, I really enjoy the challenge the game brings compared to other boardgames which slide closer to luck or chance.
If I were to compare Go to another board game, it would be Chess. Both are games of skill, and the only luck is who gets to go first. They both involve deep logic and thinking ahead. They are also both complicated games to properly code computer AI for. Regarding chess, we have been able to create a machine that could defeat Chess master Garry Kasparov in 1996. When you think of it, that was over ten years ago. Imagine a new super computer with current technology and imagine how powerful the logic could be. Essentially, the computer must be able to see all possible scenarios that can occur and play accordingly. Thus, a faster machine means more possible computations within a reasonable time limit.
Now, Feng-hsiung Hsu, who was the architect and the principal designer of Deep Blue wants to repeat history again, but this time playing Go. While at first glance, it may not seem so complicated, Go is extremely more deep than Chess is. There are an absolute more number of possibilities which makes thinking ahead incredibly more complicated. Yet, even then, he suggests that within ten years, they will have a machine capable of beating the best Go players.
I was always fascinated by computer AI, including the challenges that you tend to come upon. If you are interested in this challenge, I highly reccomend this article. It explains in more detail why Go is such a complicated game to analyze and also the methods they plan on using to code the AI.
Side note: Ramsay… rematch!!
