In article <1993Feb13.200558.19170 at news.acns.nwu.edu>,
mitchm at casbah.acns.nwu.edu (Mitchell Maltenfort) writes:
>>> Since this requires a "hunt-and-peck" approach, I suggested that he
> try using the real (i.e., experimental data) to try to train a backprop
> network to map force measurements to EMG (or vice-versa). Since he just wants
> to show whether or not a consistent relation exists, and a backprop learning
> algorithm should converge to the relation if one exists, the (in)ability of a
> backprop network to create a mapping should be a litmus test, right?
>> If not, could you tell me why? Thanks.
I'm not much of an expert, but isn't there an inherent danger in trying to
use the neural net in this kind of a testing bed? Let's suppose that you
fail to train the net. How will you be able to conclude whether the way
you set up the net or the training regimen (in terms of patterns or time
allowed) was sufficient to this problem? If you do succeed, how will you
be sure that the computational power available to the animal system isn't
being grossly exaggerated?
Isn't there a consistent relation between force and EMG though? Certainly
there isn't a 1:1 relationship, but one *IS* causal to the other.
Interested in learning this stuff myself.
mvcy at cornella.cit.cornell.edu