hughc@spinifex.eecs.unsw.oz (Hugh Clapin) (10/05/89)
I'm presently getting my act together for my honours thesis (in philosophy), and want to explore the comparisons between representation in NNs and in traditional symbol systems. Also of interest is a comparsion between how NNs use the knowledge they represent and how a Von Neumann computer uses the knowledge it represents (at the binary level where the work is actually done). I recognise that this is a pretty wide and unspecified area to waffle about, but if anyone can point me to discussions which might have something to do with representation, and particularly comparing NNs and Von Neumann architectures, I'd appreciate it. I'll post a summary if I get enough responses thanks, Hugh Clapin
knareddy@umn-d-ub.D.UMN.EDU (krishna nareddy) (10/07/89)
> I'm presently getting my act together for my honours thesis (in philosophy), > and want to explore the comparisons between representation in NNs and in > traditional symbol systems. Also of interest is a comparsion between how > > thanks, > Hugh Clapin Here is a tech. report that may address an aspect of your question. "Connectionism and Information Processing Processing Abstractions", B.Chandrasekharan, A.Goel, and D.Allemang, 88-BC-CONPROAB, published by the Ohio State Univ., Lab for AI. Address is : Computer and Info. science, Research center, 2036 Neil Avenue Mall, Columbus, Ohio 43210-1277 (614)292-5813 Thank you, krishna, Univ. of Minnesota - Duluth PS : Unable to e-mail to your address. Hence I posted response here.
zador-anthony@CS.YALE.EDU (anthony zador) (10/07/89)
In article <485@spinifex.eecs.unsw.oz> hughc@spinifex.eecs.unsw.oz (Hugh Clapin) writes: >I'm presently getting my act together for my honours thesis (in philosophy), >and want to explore the comparisons between representation in NNs and in >traditional symbol systems. Also of interest is a comparsion between how >NNs use the knowledge they represent and how a Von Neumann computer uses Mail bounced. Here is one reference: Connections and Symbols, ed. S. Pinker ~1988 It has articles by Fodor & Pylyshyn, Pinker, and several others condemning NNs in favor of symbol processing. It is one of the best thought out critical accts. It also has many very good references. (If the title is a little off and you cant find it, I can dig up the full reference). Tony Zador
bill@wayback.unm.edu (william horne) (10/07/89)
In article <1678@cs.yale.edu> zador-anthony@CS.YALE.EDU (anthony zador) writes: >In article <485@spinifex.eecs.unsw.oz> hughc@spinifex.eecs.unsw.oz (Hugh Clapin) writes: >>I'm presently getting my act together for my honours thesis (in philosophy), >>and want to explore the comparisons between representation in NNs and in >>traditional symbol systems. Also of interest is a comparsion between how >>NNs use the knowledge they represent and how a Von Neumann computer uses > >Connections and Symbols, ed. S. Pinker ~1988 > These are generally critical articles of the NN approach. The best pro- article I saw was in Brain and Behavioral Sciences 1988 by Paul Smolensky, called "On the Proper Treatment of Connectionism", 11, 1-74. BBS is an "Open Peer" journal. For this particular article there were at least thirty responses. This is a very good article, although it is pretty difficult to get through (its long, and packed with information). Basically, Smolensky is trying to characterize connectionist models as using "subsymbolic" representations, which are quite different from symbolic representations in traditional symbol systems. These subsymbols somehow describe micro-features of cognition as opposed to macro-features. In essence, it is a reductionist approach. One of the issues he brings up (actually in his response to the peer responses) is the classification of various approaches to cognition researchers have takens with connectionist systems: Eliminativist, Limitivist, Revisionist, or Implementationalist. Which are basically opinions as to whether connectionism is just as doomed as traditional AI, something completely new and different, or just the same old concepts in a different architecture. I think that there are new concepts being offered that are unique to connectionism, and that the consequences (if connectionism is successful) is that we would see a paradigm shift away from logic towards pattern recognition as the dominant mechanism of cognition. What do people out there think? -bill
dg1v+@andrew.cmu.edu (David Greene) (10/07/89)
> Excerpts from netnews.comp.ai.neural-nets: 6-Oct-89 Re: Representational > NN william horne@wayback.un (2038) > I think that there are new concepts being offered that > are unique to connectionism, and that the consequences (if connectionism > is successful) is that we would see a paradigm shift away from logic > towards pattern recognition as the dominant mechanism of cognition. What do people out there think? I'm curious as to what the necessity of a "dominant mechanism" is... From the cog. psych literature, there seems to be favorable evidence of a "dual process" approach in cognition (such as by Posner or Kellog). This approach suggests that there are two levels, one being our conscious effort represented by "hypothesis testing" and the other being a sub-conscious process of "frequency sampling". Hypothesis testing fits into the frame of symbolic and logic processing while frequency sampling is associated with pattern-recognition approaches. Each area has its positive and negative qualities, but both perform essential functions in the human learning process. Why the frequent debate as to which is the "true path"? Although for a particular focus one might be provably superior, neither seems dominant, much less definitive. In fact, it would suggest that both approaches need to interact to form an effective "artificial intelligence". That interaction in the brain itself seems to be a ripe area of research. -David -------------------------------------------------------------------- David Perry Greene || ARPA: dg1v@andrew.cmu.edu GSIA /Robotics || dpg@isl1.ri.cmu.edu Carnegie Mellon Univ. || BITNET: dg1v%andrew@vb.cc.cmu.edu Pittsburgh, PA 15213 || UUCP: !harvard!andrew.cmu.edu!dg1v -------------------------------------------------------------------- "You're welcome to use my opinions, just don't get them all wrinkled."