Author: Fernando Villegas
Date: 14:16:48 01/25/98
Go up one level in this thread
On January 25, 1998 at 15:07:25, Komputer Korner wrote: >Don't get me wrong. I have suscribed to the ICCA journal for almost 10 >years >and I consider it a valuable source of computer chess news, but it's >scientific articles sometimes just do not measure up to a real >scientific >journal which is what the ICCA journal has always purported to be. In >the >latest issue there is a prime example of this, titled "Does Deep Blue >Use >Artificial Intelligence?" In that article R.E. Korf of Los Angeles >argues >that IBM (a $ 100 billion corporation) is wrong when it says that Deep >Blue >does not use artificial intelligence. Korf argues that Deep Blue is an >artificial intelligent machine!????? Without having access to the >source >code or hardware of Deep Blue and with only the games of the last match >with >Kasparov to guide him, Korf has the audacity to claim that Deep Blue has >intelligence. He is wrong for the following reasons. >1) If you pull the plug, the machine just sits there doing nothing. >2) The programmer can change the code at any time and make it play >stupidly. >3) Deep Blue can not procreate. >4) Even when Deep Blue is running, if a fire starts, the machine will >burn >up, thus it has no way of saving itself or even knowing that it needs >saving. >5) As a prime example of an expert system, it is limited only to the >rules >of chess. Change one rule and Deep Blue is obsolete until the DADDY >programmer changes it's codes. >6) All new conceptual ideas have to be fed to the program by its >programmerr >despite the fact that it DOES have an automatic evaluation tuner based >on >games it plays. >7) The match against Kasparov proved nothing except for the fact that it >is >possible to scare a world champion off his game by a lot of hype. >8) Korf's main reason for concluding that Deep Blue is Artificial >Intelligence is that it uses alpha beta minimax with a heuristic static >evaluation function. >9) He gives no proof or other arguments to back the previous argument >stated >in (8) above. All he does is state that "Chess was one of the original >AI >problems and remains a canonical task." That is tantamount to arguing >that >since all the world's astronomers in the 1400's said that the earth was >flat, that it indeed remains flat!!!!!!!!!!!!!!!!!!!!!!!! >10) He states that since Deep Blue defeated Kasparov , Deep Blue should >be >considered a product of artificial intelligence. Deep Blue is a product >of >human intelligence and nothing more. > >11) I would say that IBM is not naive in the field of artificial >intelligence. If they say their product is not a representative of >artificial intelligence who is a fellow named Korf to argue otherwise? >In >fact he presents no arguments at all, just statements that he puts out >as >"His truths". > >12) The ICCA referees have always been biased in favour of chess being a >subject of artificial intelligence so that their research grants >wouldn't >dry up. This article is another example of the board of referees looking >the >other way. Dear KK: Maybe Korf arguments are not very sound, but sometimes yours are not better. I cannot believe that is a sound argument against the existence of AI -in deep Blue or something erlse- the fact you can unplug the machine. Is like to say there is not human intelligence because you can shoot at the head of the guy thinking and surely he will not think anymore. The same with procreation. I know a couple of impotent guys pretty intelligents, to be sure.In fact, you have only showed what all of us knew, that Deep Blue is not a human being. OK, you are right in that, but the sheer concept of ARTIFICIAL intelligence let us know you are not going to find emotions, procreation, etc. I think these kind of confussion arise of a concious or unconcious equivalence made between human intelligence and intelligence at such. As I argued in CCR time ago, a more operative and useful definition of intelligence should put out of bussines any kind of description that openly or hiddenly makes reference at the kind of vehicle behind the intelligence, the kind of mechanisms that produces the intelligence, the kind of material of the processing unit, etc . On the contrary, if you define intelligence as a kind of behaviour in certain fields or task, one capable of solving problems and learning to a certaind degree, then you have intelligence and then, also, you can have artificial intelligence. The point of how intelligent is that intelligence, how wide its scope, that's another thing. Fernando
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.