Author: Matthew Hull
Date: 08:41:06 08/06/02
Go up one level in this thread
On August 06, 2002 at 10:41:11, Robert Hyatt wrote: >On August 05, 2002 at 16:28:39, Matthew Hull wrote: > >>On August 05, 2002 at 11:45:28, Robert Hyatt wrote: >> >>>On August 05, 2002 at 11:10:55, Terry McCracken wrote: >>> >>>>Do computers make decisions? >>>>If so, what is your definition of a "computer decision" and how it relates and >>>>differs from human decisions? >>>> >>>>Please cite examples. This can be from chess to any area of so-called "machine >>>>intelligence", please give _your_ answers, as well as information that can be >>>>obtained on the net. >>>> >>>>Your help with these answers will be greatly appreciated! >>>> >>>>Thanks in Advance. >>>> >>>>Regards, >>>> Terry McCracken >>> >>> >>>A couple of points. >>> >>>First, _yes_ a computer makes decisions. For example, you can use an >>>external A/D converter to measure two temperatures in a steam plant and make >>>a decision as to which burner should be turned up or down based on those >>>measurements. >>> >>>Second, does a computer make decisions like _we_ do? Impossible to say. IE >>>can you _prove_ that the human mind doesn't rely on anything other than pure >>>binary values? Nobody has to date, so that is an open question. Wouldn't it >>>be funny if we one day find out that at the elementary level, everything we do >>>is on/off? :) >>> >>>Perhaps one of the best examples of "making a decision" is in computer chess, >>>where the computer has to choose between N moves and pick just one. That is >>>_clearly_ a decision... >> >> >>We say that computers make decisions but that is really just for the sake of >>convenience when talking about processes. That computers decide anything is >>purely notional. >> >>We wouldn't say that a colander is deciding to let the water drain out of it >>while not permitting the pasta to pass through the holes. Neither would we say >>that bean sorters, gravel sorters, or hollerith card sorters are making >>decisions. Their functions are completely and totally deterministic. > >OK... but at least for parallel search, even that is not true. IE my program >is not "deterministic" in any way when using multiple cpus. .. although I >don't see what that has to do with whether or not it is "making a decision." > >I'm afraid that one day we are going to discover that the human mind is not >nearly as remarkable as we give it credit for being. > > >> >>We would say that the _designers_ are really making all the decisions here! > >Then your parents/teachers are making all _your_ decisions? After all, they >teach you good vs bad, etc... > > > >> >>Some will say that we don't know that that's not all there is to the "mind" >>ultimately. I would say that Penrose has shown that definitely, Turing machines >>_dont_ work like our minds! Those proofs falsify the hypothesis. > >How can you prove computers don't work like our minds when you can't even >state precisely how our minds work? That seems like an impossibility... > > > > >> >>I've read many critics of Penrose on this issue, but they NEVER provide any math >>for their arguments. The math is the whole game. > > >Neither does _he_ provide any math relating exactly to how the human brain >functions... so proving something is different becomes just an opinion, not >a "fact"... But he did provide proofs that certain problems cannot be traversed algorithmically? Is this not the point? If it is proven that there are not algorithmic solutions to important types of problems, then Turing machines can't solve them, ever, period. Yet "minds" _do_ solve them with ease. Therefore, "minds" are not Turing machines. You don't have to have the "math" of the human brain to reach that conclusion. The Turing machine is trapped in an algorithmic universe from which there is no escape. If the proofs that he provided stand, then his point is made. It seems to me that to dispute his point and still be credible, you MUST dispute the proofs. Anyway, that was my understanding of his argument. I've yet to see anybody challenge his central thesis in any serious manner. One suspects there's alot of intellectual capital invested in Strong AI for which this kind of news is not welcome. BTW, chess, as we all know, _is_ a class of problem that _can_ be solved algorithmically. A fact for which I'm sure we can all agree is quite delightful.:-) Best regards, Matt > > > >> >>Regards, >>Matt
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.