dg1v+@andrew.cmu.edu (David Greene) (08/24/90)
Excerpts from netnews.comp.ai.neural-nets: 22-Aug-90 Re: Observations on the Sta.. usenet news poster@nlm.n (1731) > km> We know that GAs can evolve true intelligence, because we've evolved to > km> our present human intellect. > I am not sure that this is strictly correct. Biological genetics evolved > a neural structure capable of being trained, but I am not aware of any > evidence for genetic type algorithms actually playing a role in biological > learning. I think some of the confusion is from associating the GA too directly with different levels of evolving human intelligence. The power of the GA is exploiting the adaptive sampling efficiency of selection and recombination. The "learning" is actually the process of evolving to solve the evaluation function -- in nature, that would be survival in the given environment. For a human, evolution is associated with developing the biological "tools" necessary to cope, such as a flexible mental structure. On a different level (different representation) John Holland also theorizes a GA type of processe in actual human learning, wherein cognitive rules compete for survival and reinforcement in memory ("Induction" Holland, Holyoke, Nisbett and Thagard) -David