Author: Bob Durrett
Date: 07:46:58 11/02/02
Go up one level in this thread
On November 02, 2002 at 06:36:04, Peter Berger wrote: >On November 02, 2002 at 00:40:45, Uri Blass wrote: > >>On November 02, 2002 at 00:06:08, Robert Hyatt wrote: >> >>>On November 01, 2002 at 22:52:14, Bob Durrett wrote: >>> >>>>On October 31, 2002 at 20:01:10, Robert Hyatt wrote: >>>> >>>>>On October 31, 2002 at 17:00:19, Bob Durrett wrote: >>>>> >>>>>> >>>>>>Solving the general problem of emulating the chess play of "humanity" might be a >>>>>>prohibitively difficult task. >>>>>> >>>>> >>>>>This has been the "holy grail" of AI since its early days. But the problem is, >>>>>in 25 words or less "we have no idea how a person does what he does when playing >>>>>chess (or anything else for that matter), which makes it _impossible_ to emulate >>>>>what we don't understand." >>>> >>>>Well, Bob H., emulating the chess play of a human is not exactly what the AI >>>>people want to do, is it. They wish to make a carbon copy of a human in all >>>>it's gory details. >>>> >>>>Many orders of magnitude different, I would say. >>>> >>>>Bob D. >>> >>>They really want to emulate human thought processes related to chess, >>>at least for the computer chess/AI purists. But until we know how the >>>human does what he does, emulation is futile, to paraphrase the borg. >>> >>>:) >> >>We do not need to know exactly what humans do to try to emulate them. >> >>If the target is to predict human moves then programs can calculate statistics >>about the success of different algorithms in predicting human moves and choose >>the algorithm with the best results. >> >>Uri > > >First you have to decide which kind of human player you want to emulate - a >strong one or a weaker one ? Both is interesting - let's take a human IM or GM >player first. > >You can take a collection of master games and tune your program to emulate to >predict the maximum percentage of human moves, or you can compair different >programs and have a look which program comes closest. > >Will this program play most human-like? > >I don't think so. The problem is not the average move ( computers and human >masters are already difficult to identify when you only look at the majority of >moves), but the one, two or three "special" moves in a game. The moves where >everyone would be sure it is a computer playing - take Fritz's Bf8 in game 2 >against Kramnik for example. As long as you get one or two moves a games like >this one you won't "deceive" anyone. In fact this is also a way to detect >cheaters on chessservers. Maybe the "ultimate test" of how human-like a chess engine is would be to let it play as a human at ICC. If the "Cheater Cops" at ICC could not detect the non-human nature of the chess engine, then the engine could be declared "human." : ) : ) : ) : ) : ) : ) : ) : ) : ) : ) : ) : ) Bob D. > >There could be another program that reproduces less human moves but also makes >less "computer moves" - I am convinced it will look much more like a human and >be a better emulation. > >Regards, >Peter
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.