home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!zaphod.mps.ohio-state.edu!uwm.edu!bionet!raven.alaska.edu!news.u.washington.edu!milton.u.washington.edu!davisd
- From: davisd@milton.u.washington.edu (Daniel Davis)
- Newsgroups: comp.ai.neural-nets
- Subject: Re: Neural Nets and Brains
- Message-ID: <1992Jul24.235828.23813@u.washington.edu>
- Date: 24 Jul 92 23:58:28 GMT
- References: <arms.711935064@spedden>> <BILL.92Jul23224539@ca3.nsma.arizona.edu> <arms.711986585@spedden>
- Sender: news@u.washington.edu (USENET News System)
- Organization: University of Washington, Seattle
- Lines: 39
-
- In article <arms.711986585@spedden> arms@cs.UAlberta.CA (Bill Armstrong) writes:
- >
- >I don't fel quite comfortable with the way you phrased that. My point
- >was that BP nets use continuous signals and the brain doesn't -- an
- >obvious very significant difference. I was asking why people would
- >expect to understand the brain by studying a system (BP) that is
- >*different* at the most basic level of signalling.
- >
- >Don't you agree that if the brain works on 0-1 signals, then to study
- >the brain one could beneficially look at logical systems?
-
- There are many levels upon which one can model neuron function. When
- you choose to view it as 0-1 signals, you are stressing the action
- potentials propensity to spike to a maximum level. A neuron need not
- produce a spike, and instead the action potential can be viewed as a
- continuous function. If one wants finer detail, one can note that the
- action potentials are caused by transmission of neurotransmitters, and
- so one can talk about discrete systems again. (That is about as fine
- detail as I would care to go. No doubt one can talk quantum mechanical
- effects, if one is so inclined.)
-
- I personally do not believe that viewing neuron function in a binary
- fashion will be fruitful until one models the sequencing and rate of
- firing of action potentials. As for the non-applicability of
- continuous functions to neuron functioning, I would suggest that
- continuous valued artificial neurons can be viewed as modelling the
- rate at which different neurons fire action potentials. (The rate of
- fire is often what conveys information, not simply whether or not a
- neuron has fired.)
-
- Finally, it should be remembered that most of us are not in fact
- trying to model the brain, but are instead using insights from brain
- function (mainly the notion of distributed processing),
- _along_with_insights_gained_from_other_fields_, as design guides for
- information processing systems. Whether or not one can make analogies
- between one's chosen structure and a brain is often irrelevant.
-
- Buy Buy -- Dan Davis, Univ. of Wash., Dept. of EE,
- davisd@u.washington.edu
-