Artificial Neural Nets
e8627164 at fbma.tuwien.ac.at
Tue Apr 25 12:45:40 EST 1995
: >I think the flow is from natural NN's to artificial NN's but never back.
: >This means that the artificial neuronal networks are modeled after there
: >natural counterpart but not for giving feedback on the research done there.
: >Same with genetic algorithms.
: I would hope NOT. Isolation from research results from artificial
: neuron networks would only slow down research results from natural
: neuronal networks. The world is not naturally divided into subjects
: that we study. The problem is that there is so much to learn that no
: one can know it all. Watch the PBS series on Connections and you
: quickly realize that you want to keep your data base wide to alert
: yourselves to possibilities in your own field. Ron Blue x011 at lehigh.edu
I did not say that it is bad if there is flow back. But my impression was
There is one more problem with the flow back to understand natural systems:
NNN's gave the Basic idea for the ANN algorithms. But then they took off
and implemented a lot of things which have no counterpart in nature. This
makes it sometimes hard to reflect back. Some of these 'inventions' are
motivated by optimazation (ironicaly since natural systems work better).
When you add more degrees of freedom (e.g. layers) maybe there are
additonal effects which make such 'shortcuts' superfluos.
Anyway, I still have the impression that theories of complexity,
collectiv effects and so forth do more for understanding NNN's then ANN's
do right now. I hope that ANN's catch up.
Please correct me if you think it is different
More information about the Neur-sci