home *** CD-ROM | disk | FTP | other *** search
/ NetNews Usenet Archive 1993 #1 / NN_1993_1.iso / spool / comp / ai / 4744 < prev    next >
Encoding:
Internet Message Format  |  1993-01-05  |  3.1 KB
  1. Xref: sparky comp.ai:4744 comp.ai.neural-nets:4692 sci.cognitive:960 comp.theory.cell-automata:586 bionet.neuroscience:621 bionet.molbio.evolution:450 bionet.software:2364
  2. Newsgroups: comp.ai,comp.ai.neural-nets,sci.cognitive,comp.theory.cell-automata,bionet.neuroscience,bionet.molbio.evolution,bionet.software
  3. From: paulf@manor.demon.co.uk (Paul Fawcett)
  4. Path: sparky!uunet!pipex!demon!manor.demon.co.uk!paulf
  5. Subject: Biologically Plausible Dynamic Artificial Neural Networks
  6. Distribution: world
  7. Organization: UDI
  8. Reply-To: paulf@manor.demon.co.uk
  9. X-Mailer: Simple NEWS 1.90 (ka9q DIS 1.19)
  10. Lines: 59
  11. Date: Tue, 5 Jan 1993 05:53:57 +0000
  12. Message-ID: <726213237snz@manor.demon.co.uk>
  13. Sender: usenet@demon.co.uk
  14.  
  15.  
  16.           Biologically  Plausible  Dynamic Artificial Neural Networks.
  17.           -----------------------------------------------------------
  18.  
  19.           A   *Dynamic   Artificial   Neural   Network*   (DANN)   [1]
  20.           possesses   processing  elements  that  are  created  and/or
  21.           annihilated,  either in real time  or  as  some  part  of  a
  22.           development phase [2].
  23.  
  24.           Of    particular    interest    is    the   possibility   of
  25.           constructing   *biologically    plausible*    DANN's    that
  26.           models    developmental   neurobiological   strategies   for
  27.           establishing  and   modifying processing elements and  their
  28.           connections.
  29.  
  30.           Work  with  cellular  automata in modeling cell genesis  and
  31.           cell pattern  formation  could  be applicable to the  design
  32.           of  DANN topologies.  Likewise, biological features that are
  33.           determined by genetic  or  evolutionary  factors  [3]  would
  34.           also have a  role  to play.
  35.  
  36.           Putting  all  this  together  with  a view to constructing a
  37.           working DANN,  possessing cognitive/behavioral attributes of
  38.           a biological system is a tall order; the modeling of nervous
  39.           systems in simple organisms may be the  best  approach  when
  40.           dealing with a problem of such complexity [4].
  41.  
  42.           Any  comments,  opinions  or  references  in respect of  the
  43.           above assertions would be most welcome.
  44.  
  45.  
  46.           Many thanks
  47.  
  48.           Paul Fawcett.
  49.  
  50.           University of Westminster
  51.  
  52.  
  53.           References.
  54.  
  55.           1. Ross, M. D., et al  (1990);  Toward  Modeling  a  Dynamic
  56.              Biological   Neural   Network,  Mathl  Comput.  Modeling,
  57.              Vol 13 No.7, pp97-105.
  58.  
  59.           2. Lee, Tsu-Chang,(1991);  Structure  Level  Adaptation  for
  60.              Artificial  Neural  Networks, Kluwer Academic Publishers.
  61.  
  62.           3. Edleman,  Gerald,(1987);  Neural Darwinism the Theory of
  63.              Neural Group Selection, Basic Books.
  64.  
  65.           4. Beer, Randal, D,(1990); Intelligence as Adaptive Behavior
  66.              :  An   Experiment   in    Computational   Neuroethology.
  67.              Academic  Press.
  68.  
  69. -- 
  70. --------------------------------------------------------------------------
  71. Paul Fawcett                 |      Internet: paulf@manor.demon.co.uk
  72. London, UK.                  |                tenec@westminster.ac.uk
  73. --------------------------------------------------------------------------  
  74.