Author: Graham Laight
Date: 19:00:08 12/21/99
Go up one level in this thread
On December 18, 1999 at 00:57:01, Bruce Moreland wrote: >Intelligence is the ability to generalize. The "artificial" part doesn't have >to refer to computers but let's limit it at that. > >So the domain involves programs that can handle general cases which may be >unforeseen by the programmer. > Sorry, I can't agree that "Intelligence is the ability to generalise". This definition might apply to certain types of "learning" (e.g. fuzzy logic, data mining or neural networks) - but intelligence is not the learning - it's the doing. I would prefer to say that "Intelligence is the ability to bring some relevant knowledge to the situation". In the case of human chess players, this is probably done by "pattern recognition" of positions on the board. They remember what the important things to look for are for given patterns. In the case of computers, the knowledge is usually obtained by generating a "game tree", and then applying an "evaluation function" to the "leaves" of this "game tree". Either way, by my definition, they are being "intelligent" by bringing some knowledge to the current position. In the majority of cases, the computers are more intelligent than humans at chess - but there are a small minority of chess players who are more intelligent than most computers. Graham > >bruce
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.