[Comp-neuro] Re: Attractors, variability and noise

Malcolm Dean malcolmdean at gmail.com
Tue Aug 12 20:41:05 CEST 2008

On Tue, Aug 12, 2008 at 7:27 AM, Mario Negrello <mnegrello at gmail.com> wrote:

> Interesting point about dynamic attractors. You talk about the creation of
> new attractors for recognition. But is it not the case that a fixed network
> structure already has an implicit attractor structure? ...
> So, that's the idea with compositionality of attractors. It is amusing that
> we connectionists are driven back to explain stuff that old AI had to
> assume. Looking for mechanisms compositionality in neural networks leads to
> interesting conclusions about the 'meaning' of attractors.
> In a sense, one can think of RNNs lodging not only the attractors that
> enable compositional thougths, but also having attractor structures that
> induce some further implications. For that, in principle, one wouldn't need
> to generate new attractors, in the sense that the net struct is fixed in a
> certain time scale.... The querstion becomes: from the perspective of the
> organism, when are two attractors the same? In other words, when do two
> attractors have the same meaning?

One problem is to distinguish between observed mathematical phenomena and
biological phenomena. "Attraction" clearly invokes a gravitational model,
but no force is proposed, only tendencies and probabilities in formulae used
to describe a physical system. Thus, the approach is descriptive but has no
causal model, other than a form of Platonism.

Another problem is the imposition of Cartesian theology upon physical
events, as in the connection between an external stimulus and subsequent
neural reactions, and the search for "meaning." This has led to a
progression of rather odd neurological models which preserve the theology as
far as possible while progressively yielding to physical evidence (but only
as far as comfortably possible).

A formalism that does describe the idea in physically plausible terms is
Thermodynamics, which provides explanations of "attractions", noise, time
scales, the natural emergence of semantics, and the hierarchy of cognitive
activity alluded to by the term "compositionality."

Malcolm Dean
Research Affiliate, Human Complex Systems, UCLA
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://www.neuroinf.org/pipermail/comp-neuro/attachments/20080812/fe23b4dd/attachment.html

More information about the Comp-neuro mailing list