Replies

Point being, evolutionary processes applied to computer science can and do create all sort of improvements and such by random processes... which was the original point you were trying to defeat in the first place.

However, I'll grant you this. It's impossible to properly re-create biological evolution in its entirity. However, certain details can be disproven/proven with the help of GAs... such as structures that supposedly have irreducible complexity.

And in certain ways, creating models help us to understand biological processes better. Orientation maps, for example, are neurons within V1 that have specific orientations. We can calculate orientation maps with good accuracy in monkeys, ferrets, etc., now--but they could not do so in the 70s. A computer program modelling the orientation map exhibited certain features of orientation maps that had not even been discovered yet!

Sigh... too late at night to be debating.

V1 = part of the visual cortex in the brain (present in all mammals, I think). V1 has neurons which react best to lines of a specific orientation, and it is possible to make an "orientation map" by mapping each neuron to a specific angle of orientation.

A computer model in the 1970s tried to emulate the process of the development of un-specific neurons to orientation-specific neurons, and in doing so, certain features in the end result were found that had not yet been discovered in genuine, biological orientation maps. That is a perfect example of how computer models help us to understand biological processes