[From Bjoern Simonsen(2000.11,22.22:40 GMT+1)]
[From Rick Marken (2000.11.20.1940)]
Oh. And I think you might want to change the Comparator
function to Reference-Input_function (instead of
Reference+Input_function, which should be producing positive feedback).
Thank you for comments.
My thinking behind that part of the model is B:CP page 62 "The comparator is
a subtractor. The perceptual signal enters in the excitatory sense (minus
sign). The resulting "error signal" has a magnitude proportional to the
algebraic sum of these two neural currents - ..."
I made the Input_function negative to let it correspond to its excitatory
sense. Having done that I had to have the comparator function like
"Reference+Input_function". If I had the comparator function like
"Reference - Input_function" it would be producing positive
feedback.Wouldn't it?
I am still unsure how to make a simple negative feedback loop in my
software. But I'll get it. When I am satisfied I'll combine more and more
loops until I can have a simple model showing e.g. the human pupil contract
when luminous intensity is changing.
Also, please be careful about posting your models as attachments.
For some reason the diagram you posted took up nearly 600 KB in
the archives.
I will.
[From Bruce Abbott (2000.11.21.1700 EST)]
....................... At present I'm getting posts from
the CSG-net archives and don't want to download the current 600+ K file if
I
can avoid it.
You will not get any more 600+ K from me.
[from Jeff Vancouver (2000.11.22.1030)]
Why do you make the perception negative (via your input function)? This
goes to Mark's point about the plus in the comparator. I think you want
the
input function to be positive and the comparator to do the subtraction.
That way, you do not have negative neural signals and you have the standard
comparator.
See above
Also, I am not sure the point of the random upper levels auxiliary. It
ends
of an 8 anyway (unless input is 0).
My thinking behind that part of the model is still B:CP page 62. "The
"reference signal" is a neural current having some magnitude. It is assumed
to be generated elsewhere in the nervous system. ...."
As I understand PCT the Reference is stimulated from different
Output_functions from upper levels. Using random upper levels I try to show
that the reference is stimulated of different intensities coming from
different upper Output_functions. But this reference nevertheless gets the
strength 8. (just a selected number)
I have a problem when input is 0. Then the error will be 8. And when the
error is 8, the next error will be 0. And so on.
If the reference is 0 when input is 0, the error also will be 0.
I lack an explanation how PCT describe the negative feedback loop when input
is 0.
I still wish comments.
Finally, it is very abstract, what is it for?
For what it is worth.
I am describing a theory. That is about what I can't experience, but only
imagine. Theory is abstract.
Working with simulating models I discover how little I understand PCT. I run
into questions I have problems to answer. And sometimes I go to CSG net to
get help. I know these are simple questions but I'm not sure I yet can
answer them better myself. But I will.
Bjoern