Author: Uri Blass
Date: 03:48:46 12/13/01
Go up one level in this thread
On December 13, 2001 at 05:33:48, Janosch Zwerensky wrote: > >>On the other hand there are heuristics the human players use that are of too >>high level at this time for computers. > >I, for one, am also not sure whether it would be worth the effort (from a purely >performance-oriented point of view) even *if* someone built, for example, >humanlike high-level planning into their chess engine. I suspect the >pattern-matching processes involved would be consuming a lot of computing power, >which would be lost for more standard procedures of that program. On top of >that, I guess the code would be harder to optimize than everything else in the >engine, so most likely the speed penalty of high-level planning could well be >even way worse than what one might expect from the (unknown) computational >complexity of the brain processes doing long-range planning in humans. > >Regards, >Janosch. I disagree here. If someone can find a way to teach programs to plan in similiar way to humans GMs they can be better than the best humans. Computers sre faster then GM's and I believe that they can be better in every thinking game that humans play. I believe that a GM who is a good programmer and understand what he knows can do the best program. The last part is the hardest part and I doubt if there is a human who understand what he knows. I certainly do not understand. Uri
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.