RE: Simple neural models

From: Chris Lofting (ddiamond@ozemail.com.au)
Date: Wed Jul 26 2000 - 04:51:48 BST

  • Next message: Joe E. Dees: "RE: Simple neural models"

    Received: by alpheratz.cpm.aca.mmu.ac.uk id EAA06300 (8.6.9/5.3[ref pg@gmsl.co.uk] for cpm.aca.mmu.ac.uk from fmb-majordomo@mmu.ac.uk); Wed, 26 Jul 2000 04:36:12 +0100
    From: "Chris Lofting" <ddiamond@ozemail.com.au>
    To: <memetics@mmu.ac.uk>
    Subject: RE: Simple neural models
    Date: Wed, 26 Jul 2000 13:51:48 +1000
    Message-ID: <LPBBICPHCJJBPJGHGMCIOEKCCHAA.ddiamond@ozemail.com.au>
    Content-Type: text/plain; charset="iso-8859-1"
    Content-Transfer-Encoding: 7bit
    X-Priority: 3 (Normal)
    X-MSMail-Priority: Normal
    X-Mailer: Microsoft Outlook IMO, Build 9.0.2416 (9.0.2910.0)
    X-MimeOLE: Produced By Microsoft MimeOLE V5.00.2314.1300
    In-Reply-To: <4.3.1.0.20000725165641.00c30c70@pop3.htcomp.net>
    Importance: Normal
    Sender: fmb-majordomo@mmu.ac.uk
    Precedence: bulk
    Reply-To: memetics@mmu.ac.uk
    

    > -----Original Message-----
    > From: fmb-majordomo@mmu.ac.uk [mailto:fmb-majordomo@mmu.ac.uk]On Behalf
    > Of Mark M. Mills
    > Sent: Wednesday, 26 July 2000 8:39
    > To: memetics@mmu.ac.uk
    > Subject: RE: Simple neural models
    >
    >
    > Derek,
    >
    > At 11:06 AM 7/25/00 +0200, you wrote:
    >
    > >Derek:
    > >Yes, I can follow that. Either neurotransmitter is released or it isn't.
    > >Fair enough.
    > >
    > >Mark:
    > >there is an inherent
    > >memory storage system involved. Knowing the 'charge state' at
    > one moment,
    > >implies knowing the previous state. Voila!, memetic memory.
    > >
    > >Derek:
    > >Here you lose me. Why does the binary state of a nerve cell imply any
    > >inherent memory, or memory of any kind?
    >

    take an XOR gate and feed the output back onto *one* of the inputs and you
    have a cyclic system and so a memory mechanism in the form of an idefinite
    memory. Holland nicely summarises this in Holland, J.H., (2000) "Emergence :
    From Chaos to Order" OUP. You want more 'meat'? get a text on neural nets
    etc.

    Note that an XOR gate picks up DIFFERENCE rather than SAMENESS and as I have
    noted elsewhere our senses are highly tuned to detecting difference and
    habituating to sameness.

    From a neurological perspective create a neuron with two inputs and an
    output with only one input 'up' and the other 'down' to fire. take a loop
    back to one of the inputs and hey presto, indefinite memory. The two inputs
    can be more, just that the general A XOR B is maintained at the soma firing
    position.

    To create the XOR you initially need to use feedforward processes, your
    basic neural network of in-hidden-out. This system is good for pattern
    recognition where you use weights to get the system to learn. Use this
    system to create an XOR system and within that use the feedback loop.

    Thus feedforward creates the conditions within which you can create an XOR.
    Sameness sets the context for processing and remembering difference.

    It follows that synchronisation of a set of neurons within a network can
    lead to a group memory and so on up the hierarchy...

    Best,

    Chris.
    ------------------
    Chris Lofting
    websites:
    http://www.eisa.net.au/~lofting
    http://www.ozemail.com.au/~ddiamond

    ===============================================================
    This was distributed via the memetics list associated with the
    Journal of Memetics - Evolutionary Models of Information Transmission
    For information about the journal and the list (e.g. unsubscribing)
    see: http://www.cpm.mmu.ac.uk/jom-emit



    This archive was generated by hypermail 2b29 : Wed Jul 26 2000 - 04:37:06 BST