kirlik@hms3 (Alex Kirlik) (03/03/89)
From article <11945@swan.ulowell.edu> , Sean Brunnock writes: From article <32125@gt-cmmsr.GATECH.EDU>, by kirlik@hms3 (Alex Kirlik): > !> Why should a net with only a few dozen neural units be !> successful at mimicking human behavior that is presumably !> the result of the activation of a tremendous number of !> neurons? ! I don't see why not: programs such as Doctor, Racter, and !Eliza are also successful at mimicking human behavior without !the need for nets at all. The point that I am trying to make !is that these programs simply mimic, they do not emulate the !human brain. !(More reasonable commentary) Yes Sean I agree. But this section of mine was taken a bit out of context. The point I was trying to make was the following: Some researchers explain the successes of neural-net models (in mimicking human behavior) in terms of their similarity with the structure and mode of processing in the brain. Now, if one wants to use such an explanation, I contend that the burden of proof is on that person to explain why their model with a small number of neural units behaves in the same way as the brain which uses a larger number of neurons to perform the same task. My question is not: Why can some unrealistic model behave as do humans ? (For there are countless models that mimic human behavior without using the "same" processes as used by the brain). Rather, my question is: How can we use the brain analogy to explain the successes of our models when we cannot specify the referential relation that holds between our theoretical constructs (units) and the world (the brain)? An analogy: Way back when someone (Bohr II) "discovers" that the fundamental constituents of all matter are these things called atoms. They have a dense center surrounded by orbiting objects. Now, that person uses this model to explain the planetary orbits around the sun. Let's say this model is successful in mimicking the behavior of the solar system. Can the modeler then explain the success of his/her model by pointing out the fact that the atom and the solar system are both matter? No. The solar system is made up of a large number of these atoms and, this is the relevant point, there is no reason to believe that a collection of atoms behaves in the same way as one atom. Similarly, I suggest that there is no reason to expect that a collection of neurons behaves in the same way as one neuron. Therefore, I contend that the brain analogy cannot explain the successes of neural net models in mimicking human behavior. Alex Kirlik UUCP: kirlik@chmsr.UUCP {backbones}!gatech!chmsr!kirlik INTERNET: kirlik@chmsr.gatech.edu