Author: karen Dall Lynn
Date: 22:12:54 01/15/02
Go up one level in this thread
On January 15, 2002 at 13:35:38, Joshua Lee wrote: This is quite a complex - but interesting - question. I am not a programmer and I have no idea about what is going on within GUI's and engines; however, from a cognitive science point of view, I would say that no chess program can get close of having *playing insight*. An insight is a sudden understanding, the popping up - in a person's mind - of a sharp vision about a problem, that may come out from zero premisses or impose itself suddenly in the form of an output, appearantly without a previous input. No chess program can do that and I believe that developing hardware and software that could replicate insight - if possibly any - is something that really would take a lot of money, time, and human resources; also, it would require a next-step technology because for the best of my knowledge the extant highest techologies simply cannot do that. We have a very compelling and convincing simulation of a vivid expert in chess; but no real conceptual insight is burning from within. What happens is that when a GM lose for a program, he/she loses graciously and his/her thinking ways - no matter being a loser's ways - are full of insight. Search based chess programs are acknowledged as smart players but dumb cognitive agents. Following the lines of Hiarchs maybe, or more recent engines under the same inspiration, parts of chess knowledge may be built in as cutways to the right playing against massive extensive searching, which is prunned in the name of an ellegant decrease of speed with improvement of results. However, all of this is still purely sinthatic. No semantical move, no insight is really going on inside those tiny brains, no matter how many times we astonishingly lose for them. If you want to discuss more deeply this issue, search over the Internet - say - for "The Chinese Room Argument". Cyborg intelligence in chess is a variation of the question this argument addresses. Regards, Karen >What kind of knowledge is not in chessprograms? >Are there techniques that are too costly but are supposed to be better than >current methods?
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.