Author: Dann Corbit
Date: 20:18:04 11/04/02
Go up one level in this thread
On November 04, 2002 at 22:00:35, Bob Durrett wrote: >On November 04, 2002 at 21:33:52, Dann Corbit wrote: > >>On November 04, 2002 at 20:22:44, Bob Durrett wrote: >> >>> >>> >>>I guess people think trying to program self-awareness into their chess engines >>>is getting dangerously close to trying to create life. That is clearly taboo >>>for chess programmers. >> >>Not necessarily taboo. Just beyond what is possible right now. In 20 years or >>so, we will be able to create computers smarter than we are. >> >>>So how about just adding code to do "self-monitoring" and "self-adjustment" >>>during and between games? That wouldn't be taboo, would it? It could almost be >>>called "learning." You don't have to be alive to learn, do you? >> >>Lots of computer programs learn. >>There is TD-Lambda learning >>There is book learning >>There is position learning >> >>Probably some other kinds too. >> >>The real problem is the compute power. The reason we don't have computer >>programs that learn in the way that humans do is that we don't have the >>horsepower to do it. Hence, some other method than a neural net with feedback >>is needed. > >Dann: > >Do you, then, see "doing self-monitoring and self-adjustment during and between >games" as being identically the same thing as "learning"? It is learning. However, cognition is not involved. >Could no "self-monitoring" and/or "self-adjustment" be done which is not already >covered by the cases you cited above? There will always be things that have not been implemented yet until an omniscient chess program has been written.
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.