artzi@cpsvax.cps.msu.edu (Ytshak Artzi - CPS) (02/13/89)
I read a very interesting paper by I. Aleksander, Department of Computing, Imperial College of Science and Tech., London: "Logical Connectionist Systems". He shows a neuron model called PLN (Probabilistic Logic Node). If, by chance anyone else has read this paper (or knows the subject) maybe they can answer some questions (or comments): 1. Is there any research being made on the relationship between Neural Nets and Probabilistic Automata ? 2. Is there anyone else making research on PLNs ? 3. Has anyone tried to simulate PLN (in software) ? (I do) The above mentioned article can be found in "Neural Comnputers", NATO ASI Series, Series F:Comp and Sys Sci, vol. 41 --------------------------------------------------------------------- Itzik Artzi, Comp. Sci., Michigan State University artzi@cpsvax.cps.msu.edu
ian@shire (Ian Parberry) (02/14/89)
In <1799@cps3xx.UUCP> artzi@cpsvax.cps.msu.edu (Ytshak Artzi-CPS) writes: > > 1. Is there any research being made on the relationship between > Neural Nets and Probabilistic Automata ? As far as computation with probabilistic neural networks goes (as opposed to learning), some elementary observations have been made in: Parberry and Schnitger, "Relating Boltzmann Machines to Conventional Models of Computation'', Neural Networks, Vol. 2., pp. 59-79, 1989. I don't know whether the details carry over to your model, but you might want to check it out. For some background on the relationship between complexity theory (for those unfamiliar with the term, it can be viewed as an outgrowth of automata theory concerned with resource- bounded computation) and neural networks you might want to consult: Parberry and Schnitger, "Parallel Computation with Threshold Functions", Journal of Computer and System Sciences, Vol. 36, No. 3, 1988. Parberry, "A Primer on the Complexity Theory of Neural Networks", in "A Sourcebook on Formal Techniques in Artificial Intelligence", R. B. Banerji (Ed.), Elsevier, To Appear in 1989 (hopefully). I apologize for appearing to have written a commercial for myself. All of the references on the subject known to me are mentioned in my articles above. Call by reference is more efficient than call by value. :-) ------------------------------------------------------------------------------- Ian Parberry "The bureaucracy is expanding to meet the needs of an expanding bureaucracy" ian@psuvax1.cs.psu.edu ian@psuvax1.BITNET ian@psuvax1.UUCP (814) 863-3600 Dept of Comp Sci, 333 Whitmore Lab, Penn State Univ, University Park, Pa 16802
jp@dce.ie (John Pelan) (10/26/90)
I am currently engaged in a final year project investigating the use of (pattern recognition) NN's for processing data from sensor arrays. The hardware will be based on Random Access Memory chips forming so called probabilistic logic nodes (extension of WISARD / CONNEX). If anyone has any hints or ideas regarding this I would welcome any E-Mail to; 87070057@vax1.dcu.ie. (Note not from where this was sent!) Thanks. -- John Pelan (jp@dce.ie) -------------------------------------------------------------------- Splice the main() brace { | Traditional nautical programmers saying. --------------------------------------------------------------------