Date: Mon 10 Oct 1988 16:03-EDT From: AIList Moderator Nick Papadakis Reply-To: AIList@AI.AI.MIT.EDU Us-Mail: MIT LCS, 545 Tech Square, Rm# NE43-504, Cambridge MA 02139 Phone: (617) 253-6524 Subject: AIList Digest V8 #103 To: AIList@AI.AI.MIT.EDU Status: RO AIList Digest Tuesday, 11 Oct 1988 Volume 8 : Issue 103 Philosophy: State and change/continuous actions Continuity and computability Belief and awareness Intelligence / Consciousness Test for Machines (Neural-Nets)??? The Grand Challenge is Foolish ---------------------------------------------------------------------- Date: 26 Sep 88 12:18:24 GMT From: steve@hubcap.UUCP ("Steve" Stevenson) Subject: Re: state and change/continuous actions >From a previous article, by smryan@garth.UUCP (Steven Ryan): > > Continuous systems are computably using calculus, but is this `effective > computation?' Calculus uses a number of existent theorems which prove some > point or set exists, but provide no method to effectively compute the value. Clearly numerical analysis emulates continuous systems. In the phil of math, this is, of course, an issue. For those denying reals but allowing the actual infinity of integers, NA is as good as the Tm. Not only are there existence theorems for point sets, but such theorems as the Peano Kernel Theorem are effective computations. At the point set level, one uses things called ``simple functions''. BTW, you're being too restrictive. There are many ``continuous'' systems which have a denumerable number of points of nondifferentiablity: there are several ways to handle this (e.g., measure theory). These are not ``calculus'' in the usual sense. Important applications are in diffusion and probability. So, is Riemann-Stiltjes the only true calculus? Nah. There's one per view. -- Steve (really "D. E.") Stevenson steve@hubcap.clemson.edu Department of Computer Science, (803)656-5880.mabell Clemson University, Clemson, SC 29634-1906 ------------------------------ Date: Mon, 3 Oct 88 13:49:33 GMT From: Mr Jack Campin Subject: continuity and computability smryan@garth.UUCP (Steven Ryan) wrote: >> "Insufficient attention has been paid to the problem of continuous >> actions." Now, a question that immediately comes to mind is "What problem?" > Continuous systems are computable using calculus, but is this `effective > computation?' Calculus uses a number of existence theorems which prove some > point or set exists, but provide no method to effectively compute the value. > It is not clear that all natural phenomena can be modelled on the discrete > and finite digital computer. If not, what computer could we use? I brought up this same point in the Usenet sci.logic newsgroup a short while ago. There is a precise sense in which analogue computers are more powerful than digital ones - i.e. there are continuous phenomena unsimulatable on a Turing machine. Most of the work on this has been done by Marian Pour-El and her coworkers. An early paper is "A computable ordinary differential equation which possesses no computable solution", Annals of Mathematical Logic, volume 17, 1979, pages 61-90. This result is a bit of a cheat (the way the equation is set up has little relation to anything in the physical world) but I believe later papers tighten it up somewhat (one uses the wave equation, which you'd expect to be a powerful computing device given that interferometers can calculate Fourier transforms in constant time). I haven't seen these later articles, though. -- Jack Campin, Computing Science Department, Glasgow University, 17 Lilybank Gardens, Glasgow G12 8QQ, SCOTLAND. 041 339 8855 x6045 wk 041 556 1878 ho ARPA: jack%cs.glasgow.ac.uk@nss.cs.ucl.ac.uk USENET: jack@glasgow.uucp JANET: jack@uk.ac.glasgow.cs PLINGnet: ...mcvax!ukc!cs.glasgow.ac.uk!jack ------------------------------ Date: 5 Oct 88 17:59:14 PDT From: "Joseph Y. Halpern" Subject: Belief and awareness In response to Fabrizio Sebastiani's question of Sept. 23 regarding further work on Fagin and my notion of "awareness", here is what I am aware of: (a) Kurt Konolige wrote a critique of the paper which appeared in the proceedings of the 1986 Conference on Theoretical Aspects of Reasoning About Knowledge; (b) Robert Hadley wrote a critique (and discussed other ways of dealing with the problem) which appeared as a Tech Report at Simon Fraser University (the exact reference can be found in the journal version of our paper, which appears in Artificial Intelligence, vol. 34, pp. 39-76); (c) Yoram Moses provided a model for polynomial time knowledge, which can be viewed as a notion of awareness; Yoram's paper appears in the proceedings of the 1988 Conference on Theoretical Aspects of Reasoning About Knowledge; (d) Mark Tuttle, Yoram Moses, and I have a paper in the 1988 Symposium on Theory of Computing which focuses on zero-knowledge protocols, but also extends Yoram's definitions to deal with learning. -- Joe Halpern ------------------------------ Date: 9 Oct 88 12:54:39 GMT From: TAURUS.BITNET!shani@ucbvax.berkeley.edu Subject: Re: Intelligence / Consciousness Test for Machines (Neural-Nets)??? In article <1141@usfvax2.EDU>, mician@usfvax2.BITNET writes: > > When can a machine be considered a conscious entity? > Oh no! not that again! ;-) Okay, I'll make it short and somewhat cyinc this time: The answer is: NEVER! You see andy, the only reason for you to assume that there is such a thing as a conscious entity at all, is that otherwise, YOU are not a conscious entity, and that probebly sounds nonsense to you (Actualy when saying that, I already take a dangerious step forward, assuming that YOU ARE... the only thing I can know is that I AM a conscious entity...). I hope that helps... O.S. ------------------------------ Date: 9 Oct 88 18:21:51 GMT From: uwslh!lishka@spool.cs.wisc.edu (Fish-Guts) Reply-to: uwslh!lishka@spool.cs.wisc.edu (Fish-Guts) Subject: Re: The Grand Challenge is Foolish In article JMC@SAIL.STANFORD.EDU writes: >[In reply to message sent Mon 26 Sep 1988 23:22-EDT.] >If John Nagle thinks that "The lesson of the last five years seems to >be that throwing money at AI is not enormously productive.", he is >also confusing science with engineering. It's like saying that the >lesson of the last five years of astronomy has been unproductive. >Progress in science is measured in longer periods than that. I don't think anyone could have said it better. If AI is going to progress at all, I think it will need quite a bit of time, for its goals seem to be fairly "grand." I think this definitely applies to research in Neural Nets and Connectionism: many people criticize this area, even though it has only really gotten going (again) in the past few years. There *have* been some really interesting discoveries due to AI; however, they have not been as amazing and earth-shattering as some would like. In my opinion, the great amount of hype in AI is what leads many people to say stuff such as "throwing money at AI is not enormously productive." If many scientists and companies would stop making their research or products out to be much more than they actually are, I feel that others reviewing the AI field would not be so critical. Many AI researchers and companies need to be much more "modest" in assessing their work; they should not make promises they cannot keep. After all, the goal of achieving true "artificial intelligence" (in the literal sense of the phrase) is not one that will occur in the next two, ten, fifty, one-hundred, or maybe even one-thousand years. .oO Chris Oo. -- Christopher Lishka ...!{rutgers|ucbvax|...}!uwvax!uwslh!lishka Wisconsin State Lab of Hygiene lishka%uwslh.uucp@cs.wisc.edu Immunology Section (608)262-1617 lishka@uwslh.uucp ---- "...Just because someone is shy and gets straight A's does not mean they won't put wads of gum in your arm pits." - Lynda Barry, "Ernie Pook's Commeek: Gum of Mystery" ------------------------------ End of AIList Digest ********************