I remember reading about Tom Ray's Tierra and how amazing the results it obtained were. However as far as I know, the only "evolution" based computer technology we use now is ML, which is very limited in its abilities. Why haven't we used artificial evolution perhaps similar to Tierra for practical purposes? Is it too hard? Do we lack the technology to make it practical so far or have we decided we don't need to do it, even if we had the technology?
Add a comment
|
1 Answer
$\begingroup$
$\endgroup$
Because we don't know how to make it work well to solve practical problems we care about.
If your'e asking about evolutionary algorithms in general: As a rough rule of thumb, evolutionary algorithms haven't turned out to be all that wonderful. As a starting heuristic, they're... usually OK. But, practical experience is that other methods often seem to perform better.
If you're asking about Tierra specifically: Tierra was a demo applied to a very specific, artificial situation. It doesn't seem to have generalized to allow us to solve many practical problems we care about. Also, I'm not sure, but it looks like Tierra might have been intended more to try to understand evolution, more than to apply evolution to solving practical computing problems.
