maner@bgsuvax.UUCP (Walter Maner) (04/21/89)
From article <370@eurtrx.UUCP>, by hans@eurtrx.UUCP (Hans Schermer): > > > Can anyone out there give me a hand? > I am looking for philosophical papers, books or articles, with reactions > to connectionism as a model for the mind. The revised (paperback) edition of _Mind Over Machine_ by Hubert & Stuart Dreyfus would be a good place to begin. It's published by the Free Press, which is a division of Macmillan. The original (hardcover) edition of this book appeared before connectionism was a hot item, but the revised edition takes connectionism into account. -- CSNet : maner@research1.bgsu.edu | 419/372-8719 InterNet: maner@research1.bgsu.edu (129.1.1.2) | BGSU CS Dept UUCP : ... !osu-cis!bgsuvax!maner | Bowling Green, OH 43403 BITNet : MANER@BGSUOPIE
zqli@batcomputer.tn.cornell.edu (Zhenqin Li) (04/21/89)
A whole issue (Vol. XXVI, 1987) of "The Southern Journal of Philosophy" (published by the Philosophy Dept of Memphis State Univ, Memphis, TN 38152), is dedicated to philosophical discussions of Connectionism. Having not spent time on the subject, I can not make judgements. The lists of references there, however, seem to be extensive.
rapaport@sunybcs.uucp (William J. Rapaport) (04/21/89)
In article <370@eurtrx.UUCP> hans@eurtrx.UUCP (Hans Schermer) writes: > > >I am looking for philosophical papers, books or articles, with reactions >to connectionism as a model for the mind. Try: Horgan, Terence, & Tienson, John (eds.), _Connectionism and the Philosophy of Mind: Spindel Conference 1987_, in _Southern Journal of Philosophy_, Vol. 26 supplement (1987). Available from SJP, Dept. of Phil., Memphis State U., Memphis, TN 38152. William J. Rapaport Associate Professor of Computer Science Co-Director, Graduate Group in Cognitive Science Interim Director, Graduate Research Initiative in Cognitive and Linguistic Sciences Dept. of Computer Science||internet: rapaport@cs.buffalo.edu SUNY Buffalo ||bitnet: rapaport@sunybcs.bitnet Buffalo, NY 14260 ||uucp: {decvax,watmath,rutgers}!sunybcs!rapaport (716) 636-3193, 3180 ||fax: (716) 636-3464
myke@gatech.edu (Myke Rynolds) (04/21/89)
Hans Schermer writes: >I am looking for philosophical papers, books or articles, with reactions >to connectionism as a model for the mind. I think that BAMs (bi-direction associative memories) and it's conceptual parent, ART (adaptive resonance theory) give a profound critique of the connectionist models. Grossberg, the inventer of ART way back in '76, goes into great detail about how nothing anyone in the connectist school of thought has said is new, or even as powerful as what already exists! ART is proven to converge on any complexity of input, no connectionist model can claim this. They can learn only by limiting the complexity of the input, thus the failure of bp to deal with large and complex systems. For all its greater power, it is much much simpliar than these other models that cloud the issue with ad hoc hockus pockus. Grossberg's model is nothing more than matrix multiplication. You take a vector forward through a weight matrix, then take it backwards through it. When it resonates on the correct answer you're done. The most obvious way to get a weight matrix to satisfy this problem on a series of such vectors is to stack them in a matrix and do linear algebra. Walla! An article on BAMs can be found in a Byte from last year. BTW, Grossberg has three Ph.D's, two of which are in math and neurophysiology. Connectionists are generally psychologists and computer scientists who do not appreciate the deeper simplicity of math under the outer tremendous diversity. -- Myke Rynolds School of Information & Computer Science, Georgia Tech, Atlanta GA 30332 uucp: ...!{decvax,hplabs,ncar,purdue,rutgers}!gatech!myke Internet: myke@gatech.edu
jha@lfcs.ed.ac.uk (Jamie Andrews) (04/21/89)
In article <4034@bgsuvax.UUCP> maner@bgsuvax.UUCP (Walter Maner) writes: >The revised (paperback) edition of _Mind Over Machine_ by Hubert & Stuart >Dreyfus would be a good place to begin... >... The original (hardcover) edition of this >book appeared before connectionism was a hot item, but the revised edition >takes connectionism into account. You mean they added a couple of chapters simplistically trashing connectionism the way they simplistically trash the rest of AI? Terrific. --Jamie, who has been in a bad mood all day jha@lfcs.ed.ac.uk "Gonna melt them down for pills and soap"
kortge@Portia.Stanford.EDU (Chris Kortge) (04/21/89)
In article <18496@gatech.edu> myke@gatech.UUCP (Myke Rynolds) writes: > >I think that BAMs (bi-direction associative memories) and it's conceptual >parent, ART (adaptive resonance theory) give a profound critique of the >connectionist models. Grossberg, the inventer of ART way back in '76, goes >into great detail about how nothing anyone in the connectist school of thought >has said is new, or even as powerful as what already exists! ART is proven >to converge on any complexity of input, no connectionist model can claim this. >They can learn only by limiting the complexity of the input, thus the failure >of bp to deal with large and complex systems. Hold on a second! Why is it, then, that people are using back-propagation learning on most practical applications? I agree that bp has trouble with large systems, but it's important to look at the *results* of the learning process, too. BP can learn distributed representations, which have advantages over strictly categorical ones, which is what ART learns. More importantly, since BP does supervised learning, its internal representation is automatically suited to the task at hand; ART is unsupervised, and thus it's categories are not necessarily useful for facilitating the required outputs. >For all its greater power, [ART] is much much simpliar than these other models >that cloud the issue with ad hoc hockus pockus. Grossberg's model is nothing >more than matrix multiplication. [...] Then why can't I understand his papers? (Don't answer that :-)) Most likely, it's because I'm a connectionist, and >Connectionists are generally psychologists and computer scientists who do not >appreciate the deeper simplicity of math under the outer tremendous diversity. Well, be patient with us, okay? Chris Kortge kortge@psych.stanford.edu
dhw@itivax.iti.org (David H. West) (04/22/89)
In article <370@eurtrx.UUCP> hans@eurtrx.UUCP (Hans Schermer) writes: > > >Can anyone out there give me a hand? >I am looking for philosophical papers, books or articles, with reactions >to connectionism as a model for the mind. Huh? Connectionism can't *be* a model for the mind. It might be a good way to *implement* certain models of the mind, or even a heuristic criterion for evaluating such models ("must map easily to hardware of this general nature"). But it doesn't relieve us of the task of coming up with the models separately. -David West dhw@itivax.iti.org {uunet,rutgers,ames}!sharkey!itivax!dhw COMPIS, Industrial Technology Institute, PO Box 1485, Ann Arbor, MI 48106
mbkennel@phoenix.Princeton.EDU (Matthew B. Kennel) (04/22/89)
In article <18496@gatech.edu> myke@gatech.UUCP (Myke Rynolds) writes: > >I think that BAMs (bi-direction associative memories) and it's conceptual >parent, ART (adaptive resonance theory) give a profound critique of the >connectionist models. Grossberg, the inventer of ART way back in '76, goes >into great detail about how nothing anyone in the connectist school of thought >has said is new, or even as powerful as what already exists! ART is proven >to converge on any complexity of input, no connectionist model can claim this. >They can learn only by limiting the complexity of the input, thus the failure >of bp to deal with large and complex systems. >For all its greater power, it is much much simpliar than these other models ^^^^^^^^^^^^^ >that cloud the issue with ad hoc hockus pockus. Grossberg's model is nothing >more than matrix multiplication. You take a vector forward through a weight ^^^^^^^^^^^^^^^^^^^^^^ >matrix, then take it backwards through it. When it resonates on the correct >answer you're done. The most obvious way to get a weight matrix to satisfy >this problem on a series of such vectors is to stack them in a matrix and >do linear algebra. Walla! ^^^^^^ Voila! That's exactly the point. For linear problems, than I have no doubt that classical algorithms (linear systems of equations) should work better than gradient descent (BP), with the whole shebang of nice rigorous results, but the whole point is that back-prop tries to learn general non-linear transformations that AREN'T matrix multiplications. For some kinds of associative memory something like ART may be fine, but associative memory isn't the whole story. It's generalization (i.e. high-dimensional interpolation) which is the the most interesting aspect of multi-layer perceptrons. Can something like a BAM network be more efficient than an "encoder" type of perceptron in terms of the number of connections? >An article on BAMs can be found in a Byte from last year. >BTW, Grossberg has three Ph.D's, two of which are in math and neurophysiology. >Connectionists are generally psychologists and computer scientists who do not >appreciate the deeper simplicity of math under the outer tremendous diversity. I've never been able to discern the deeper simplicity of math in any ART paper that I've seen (which is very few, I must admit); back-prop is >-- >Myke Rynolds >School of Information & Computer Science, Georgia Tech, Atlanta GA 30332 >uucp: ...!{decvax,hplabs,ncar,purdue,rutgers}!gatech!myke >Internet: myke@gatech.edu Matt Kennel mbkennel@phoenix.princeton.edu
myke@gatech.edu (Myke Rynolds) (04/22/89)
At portia?! Hey, do you know Paul Gunnels? Chris Kortge writes: |Myke Rynolds writes: ||I think that BAMs (bi-direction associative memories) and it's conceptual ||parent, ART (adaptive resonance theory) give a profound critique of the ||connectionist models. Grossberg, the inventer of ART way back in '76, goes ||into great detail about how nothing anyone in the connectist school of thought ||has said is new, or even as powerful as what already exists! ART is proven ||to converge on any complexity of input, no connectionist model can claim this. ||They can learn only by limiting the complexity of the input, thus the failure ||of bp to deal with large and complex systems. || |Hold on a second! Why is it, then, that people are using |back-propagation learning on most practical applications? Good question. Maybe its fad? |I agree that |bp has trouble with large systems, but it's important to look at the |*results* of the learning process, too. BP can learn distributed |representations, which have advantages over strictly categorical ones, |which is what ART learns. False! ART learns internal reps. Both BP and ART generate their own internal reps (for no good reason in my opinion), but BAMs simply associate input vectors with output vectors. |More importantly, since BP does supervised |learning, its internal representation is automatically suited to the |task at hand; ART is unsupervised, and thus it's categories are not |necessarily useful for facilitating the required outputs. But unless the superviser is omniscient, it doesn't know when to stop being plastic to prevent memory washout. ART does not suffer from this. The lack of need for a superviser is not a weakness, it is a tremendous advantage! | ||For all its greater power, [ART] is much much simpliar than these other models ||that cloud the issue with ad hoc hockus pockus. Grossberg's model is nothing ||more than matrix multiplication. [...] | |Then why can't I understand his papers? (Don't answer that :-)) |Most likely, it's because I'm a connectionist, and Cuz the man is lost in his own little world. However, hes not being swept along by any mobs either. | ||Connectionists are generally psychologists and computer scientists who do not ||appreciate the deeper simplicity of math under the outer tremendous diversity. | |Well, be patient with us, okay? Ok, as long as y'all see the light soon! -- Myke Rynolds School of Information & Computer Science, Georgia Tech, Atlanta GA 30332 uucp: ...!{decvax,hplabs,ncar,purdue,rutgers}!gatech!myke Internet: myke@gatech.edu
aarons@syma.sussex.ac.uk (Aaron Sloman) (05/01/89)
hans@eurtrx.UUCP (Hans Schermer) writes: > Date: 20 Apr 89 10:40:46 GMT > > Can anyone out there give me a hand? > I am looking for philosophical papers, books or articles, with reactions > to connectionism as a model for the mind. > I would be interested in texts discussing representationalism, (sub)symbolic > representations, materialism, and other philosophical subjects that could be > influenced by a connectionist theory of the mind. > .....etc..... Here's my pennyworth. I am amazed when people try to produce philosophical arguments to show that connectionist models are superior, or inferior, to other kinds of AI models of mental processes. Instead of getting involved in these silly disputes, people should try to understand the rich multiplicity of function of the human mind and try to see what kinds of architectures might account for that multiplicity, and what kinds of mechanisms are capable of fitting in to those architectures in order to fulfil the roles required. For example the mechanisms required for low level vision are likely to be somewhat different from the mechanisms used in multiplying 395 by 11 in your head. Both are likely to be different from (though they may overlap with) the mechanisms involved in associative retrieval of stored memories on the basis of partial matches ("Suzie had a little goat" Yes? No? who had what then?) Then there is our ability to store and retrieve intricate detail exactly, as when we memorize a long poem or a piano sonata. Different again must be the mechanisms by which new motives (desires, fears, wishes, and the like) are generated (by physical needs, by perceiving something in the environment, by thinking about past events or future possibilities etc). These motives, in turn, interact in many intricate ways with other motives, beliefs, percepts, personality traits, etc. Some, but not all, motives(desires) become intentions. ("Yes, I will try to get ...." or "That is very tempting, but I mustn't..."). Planning processes sometimes arise out of intentions ("Now, how can I get that box open. Perhaps I can borrow a crow-bar from Jim, though I'll have to offer him something in return, he's so mean...Now where can I find him. His wife will know..."). But sometimes intentions directly interact with percepts to generate behaviour controlled by tight feedback loops (like bringing your car to a gentle stop just at the traffic lights). Some kinds of abilities seem to encompass a finite or fixed dimensional range of possibilities (e.g. the set of ways of moving your arm so that your forefinger moves quickly in a smooth path from touching one thing to touching another?) whereas other abilities involve a kind of generative competence that implies unbounded complexity, at least in principle, (e.g. the set of algebraic expressions you can evaluate). There are very many different kinds of learning, training, development, improvement. Some kinds of actions can be achieved perfectly once you know what to do (long division). Others require training or tuning of low level mechanisms, in ways that are very hard to understand (coaxing a beautiful tone out of a violin). Some things are inaccessible to consciousness normally yet can become accessible after appropriate training, such as the use of grammatical categories in producing or understanding language. (One kind of philosophical training is concerned with this kind of heightened awareness. Compare learning phonetics.) We can do some things in parallel (walking and talking, listening and looking, enjoying a meal and a view, or seeing the different ballet dancers that form an intricate and changing pattern), yet others are difficult or impossible, like reciting two poems in your head at once. Some things are easily reversed (sing a high note and swoop down to a low note - then do it in reverse) but others not (recite a poem then say it backwards). Some kinds of mental processes are transformed by alcohol and other drugs, and some not. E.g. alcohol (in relatively small doses) may alter what you will agree to do, but it probably won't change the semantic interpretation you give to "The cat sat on the mat". There are many far more detailed requirements for explanatory mechanisms. It seems to me absurd to argue over whether either connectionist models or conventionalist AI models provide better theories of the nature of mind when it is patently clear both are still miles away from accounting for more than highly simplified versions of tiny fragments of human ability. Instead of silly squabbles we need to work both top-down (collecting requirements for adequate models and explanations), and bottom up (trying to investigate different kinds of mechanisms and finding out what can and cannot be achieved by putting them together in different ways). It seems very likely that the final story (if we ever find it) will involve many different kinds of mechanisms put together in a complex variety of ways. Attempts to do it all using one kind of technique (Production systems, Logic, PDP mechanisms) will then just look silly. Aaron Sloman, School of Cognitive and Computing Sciences, Univ of Sussex, Brighton, BN1 9QN, England INTERNET: aarons%uk.ac.sussex.cogs@nsfnet-relay.ac.uk aarons%uk.ac.sussex.cogs%nsfnet-relay.ac.uk@relay.cs.net JANET aarons@cogs.sussex.ac.uk BITNET: aarons%uk.ac.sussex.cogs@uk.ac or aarons%uk.ac.sussex.cogs%ukacrl.bitnet@cunyvm.cuny.edu UUCP: ...mcvax!ukc!cogs!aarons or aarons@cogs.uucp
gall@yunexus.UUCP (Norman R. Gall) (05/03/89)
All of this discussion on 'models of mind' already presupposes that
the question of 'what mechanisms underlie the workings of the mind?'
is a coherent one. I would like to ask: precisely what question could
possibly be answered by the 'discovery' of the actual mechanism by
which 'the mind' operates, and just what makes us think that 'the
mind' or psychological predicates 'refer' to mental processes?
Now, yes, I know that there is a very deep-rooted tradition in
psychology (generally carried on by cognitive science) that treats
psychological verbs as referring to actual mental processes, but
evidence do we have for treating them as such? Intuition? Empirical
evidence that is unmitigated? Careful philosophical scrutiny of the
very concepts these psychological verbs deal with?
--
York University Department of Philosophy Toronto, Ontario, Canada
"Don't, _for_heaven's_sake_, be afraid of talking nonsense! But you
must pay attention to your nonsense." -- L. Wittgenstein
_____________________________________________________________________________coggins@coggins.cs.unc.edu (Dr. James Coggins) (05/03/89)
> Can anyone out there give me a hand? > I am looking for philosophical papers, books or articles, with reactions > to connectionism as a model for the mind. > I would be interested in texts discussing representationalism, (sub)symbolic > representations, materialism, and other philosophical subjects that could be > influenced by a connectionist theory of the mind. > .....etc..... My assessment of the neural net area is as follows: (consider these Six Theses nailed to the church door) 1. NNs are a parallel implementation technique that shows promise for making perceptual processes run in real time. 2. There is nothing in the NN work that is fundamentally new except as a fast implementation. Their ability to learn incrementally from a series of samples nice but not new. The way they learn and make decisions is decades old and first arose in communication theory, then was further developed in statistical pattern recognition. 3. The claims that NNs are fundamentally new are founded on ignorance of statistical pattern recognition or on simplistic views of the nature of statistical pattern recognition. I have heard supposedly competent people working in NNs claim that statistical pattern recognition is based on assumptions of Gaussian distributions which are not required in NNs, therefore NNs are fundamentally different. This is ridiculous. Statistical pattern recognition is not bound to Gaussians, and NNs do, most assuredly, incorporate distributional assumptions in their decision criteria. 4. A more cynical view that I do not fully embrace says that the main function of "Neural Networks" is as a label for money. It is a flag you wave to attract money dispensed by people who are interested in the engineering of real-time perceptual processing and who are ignorant of statistical pattern recognition and therefore the lack of substance of the neural net field. 5. Neural nets raise lots of engineering questions but little science. Much of the excitement they have raised is based on uncritical acceptance of "neat" demos and ignorance. As such, the area resembles a religion more than a science. 6. The "popularity" of neural net research is a consequence of the miserable mathematical backgrounds of computer science students (and some professors!). You don't need to know any math to be a hacker, but you have to know math and statistics to work in statistical pattern recognition. Thus, generations of computer science students are susceptible to hoodwinking by neat demos based on simple mathematical and statistical techniques that incorporate some engineering hacks that can be tweaked forever. They'll think they are accomplishing something by their endless tweaking because they don't know enough math and statistics to tell what's really going on. --------------------------------------------------------------------- Dr. James M. Coggins coggins@cs.unc.edu Computer Science Department A neuromorphic minimum distance classifier! UNC-Chapel Hill Big freaking hairy deal. Chapel Hill, NC 27599-3175 -Garfield the Cat and NASA Center of Excellence in Space Data and Information Science ---------------------------------------------------------------------
hkhenson@cup.portal.com (H Keith Henson) (05/20/89)
This may be entirely redundant to these groups, but two books which strongly support Aaron Sloman's views (aarons@cogs.sussex.ac.uk) are _The Social Brain_ by Micheal Gazzaniga, and (of course) _Society of Mind_ by Marvin Minsky. If there are others, or articles, I would appreciate email or postings. Keith Henson (hkhenson@cup.portal.com)