Author: Russell Reagan
Date: 10:33:12 10/19/02
Go up one level in this thread
On October 19, 2002 at 12:46:34, Dana Turnmire wrote: >How well would a computer program do against the top grandmasters if it had to >generate its own moves from start to finish without human intervention in the >opening play? Uri Blass said he didn't put much stock in openings books. I'm a >little surprised by that statment since it seems everyone puts such emphasis on >what opening book should be used etc. > >Humans have to learn and memorize the best lines for openings. Why should >humans play that part of the game for the computer? I understand the purpose >for opening books but to find out if computers are really superior to humans >shouldn't the computers have to learn and memorize the best opening moves just >as all humans have to do? This is kind of a "matter of opinion" issue IMO, but I don't think that opening books or endgame tablebases shouldn't be allowed. First, it's not clear whether or not they should be allowed in the first place. Some say yes, some say no. I think it's just another part of memory, just like humans have. The human chess players learned from other games and have their own "opening book" too. No top human chess player plays without his/her own opening book. It's just stored in a different medium. Secondly, how would you ever enforce such a rule? If you say, "can't query external files" then I can just dump my entire opening book into my executable, and I'm not breaking any rules. I don't see anything wrong with it in the first place, and I think it would be a nightmare trying to create rules and enforce those rules. I think we often mistake "artificial intelligence" for "intelligence simulated by a computer algorithm". If you could get a biological "machine" to play chess, now it's not so clear where "books" and "databases" start or end, and where the "thinking" takes place. Humans do the same thing that a computer chess program does, and the fact that it's on a computer and not some other medium shouldn't matter IMO. If a computer has a conversation with me, and all it does is look up sentence fragments from a database and piece them together, but it passes a turing test, that's AI, and it doesn't matter that it uses a database as opposed to some other method. Russell
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.