Comments on: The Game of Go http://thegreendestiny.wordpress.com/2008/10/01/the-game-of-go/ AI, Games and Computer Science Tue, 27 Oct 2009 15:00:10 +0000 http://wordpress.com/ hourly 1 By: Computer Go (….continued) « The Green Destiny http://thegreendestiny.wordpress.com/2008/10/01/the-game-of-go/#comment-61 Computer Go (….continued) « The Green Destiny Tue, 27 Oct 2009 14:22:40 +0000 http://thegreendestiny.wordpress.com/?p=78#comment-61 [...] Stone, Deep Blue, explanation, Feng-Hsuing hsu, Game AI, GNU Go, Go, MoGo, Monte Carlo, UCT The last post on Computer Go (about a year back) was a brief description of the game, and why it is very [...] [...] Stone, Deep Blue, explanation, Feng-Hsuing hsu, Game AI, GNU Go, Go, MoGo, Monte Carlo, UCT The last post on Computer Go (about a year back) was a brief description of the game, and why it is very [...]

]]> By: Sanjeev Chandran http://thegreendestiny.wordpress.com/2008/10/01/the-game-of-go/#comment-53 Sanjeev Chandran Mon, 26 Jan 2009 09:54:41 +0000 http://thegreendestiny.wordpress.com/?p=78#comment-53 [Edited] The comparison with the number of particles in the Universe was meant as pop-science note. The direct implication was that a brute force min-max approach to solving the game would be futile. ( Although there is a school who believe its possible ) Therefore, an indirect representation such as those used in GAs would be a good option. I found a <a href="http://nn.cs.utexas.edu/?stanley:gecco04" rel="nofollow">paper</a> on a Roving Eye ANN developed using NeuroEvolution that plays smaller boards. I don't how well it would scale though. [Edited]
The comparison with the number of particles in the Universe was meant as pop-science note. The direct implication was that a brute force min-max approach to solving the game would be futile. ( Although there is a school who believe its possible ) Therefore, an indirect representation such as those used in GAs would be a good option. I found a paper on a Roving Eye ANN developed using NeuroEvolution that plays smaller boards. I don’t how well it would scale though.

]]> By: Deltaway http://thegreendestiny.wordpress.com/2008/10/01/the-game-of-go/#comment-51 Deltaway Sat, 24 Jan 2009 20:39:30 +0000 http://thegreendestiny.wordpress.com/?p=78#comment-51 The number 10^700 is the complexity of the game of Go as measured by the number of possible non-redundant games. The number of particles in the universe does not impose a limitation that would require this number to be scaled down because while the rules of the game need to be explained with a certain minimum number of bits, the game possibilities are theoretical and do not have to be recorded in order to define the game according to its rules. Because of this, the number 10^700 should not be considered an overestimate merely because the number of particles in the universe is less. In fact, genetic algorithms in particular use their rules as an efficient way to find an ideal solution in a relatively complex possibility space. A simpler definition of the problem facilitates the acquisition of an answer. The number 10^700 is the complexity of the game of Go as measured by the number of possible non-redundant games. The number of particles in the universe does not impose a limitation that would require this number to be scaled down because while the rules of the game need to be explained with a certain minimum number of bits, the game possibilities are theoretical and do not have to be recorded in order to define the game according to its rules. Because of this, the number 10^700 should not be considered an overestimate merely because the number of particles in the universe is less. In fact, genetic algorithms in particular use their rules as an efficient way to find an ideal solution in a relatively complex possibility space. A simpler definition of the problem facilitates the acquisition of an answer.

]]> By: Takuan http://thegreendestiny.wordpress.com/2008/10/01/the-game-of-go/#comment-46 Takuan Mon, 10 Nov 2008 03:22:12 +0000 http://thegreendestiny.wordpress.com/?p=78#comment-46 I look forward to seeing more articles about Go. It's been an interest of mine for nearly two decades now, though I have to admit that I'm not very good at it, haha. I've been intensely interested in the progress of Go software for that entire period, and it often seems to me as if creating a strong Go program is one of the few pathologically difficult problems remaining in software development. I may be overstating that somewhat, but you have to admit that even the 'best of the best' in Go software is ridiculously naive compared to human players, and unlike Chess and other games it's very difficult to identify and weight potential patterns. I've subscribed (having found your blog through a link on AIGameDev), and hope to see more on this topic in the future :) I look forward to seeing more articles about Go. It’s been an interest of mine for nearly two decades now, though I have to admit that I’m not very good at it, haha.

I’ve been intensely interested in the progress of Go software for that entire period, and it often seems to me as if creating a strong Go program is one of the few pathologically difficult problems remaining in software development. I may be overstating that somewhat, but you have to admit that even the ‘best of the best’ in Go software is ridiculously naive compared to human players, and unlike Chess and other games it’s very difficult to identify and weight potential patterns.

I’ve subscribed (having found your blog through a link on AIGameDev), and hope to see more on this topic in the future :)

]]>