QUOTE(HomeSlice @ Mar 9 2010, 10:57 PM)

QUOTE(robcat2075 @ Mar 9 2010, 05:00 PM)

Can this thing learn to convert polygons to splines?
Theoretically, I suppose it could. But I have no idea how you would create the network to do that ... or how large a data set you would need before it actually starts making right choices ...
I'm guessing that converting polygons to splines isn't a fuzzy logic type of problem, which I believe are the usual situations that use neural networks (eg stock market trends, weather, military maneuvers, etc). There is, I would imagine, a limited and relatively small number of situations of an indeterminate nature, that would have to be addressed for converting polys to splines, and for those cases that are indeterminate, it would seem that the user could help in making those decisions, perhaps? I am not in any way, saying that I understand the true scope of the problem, nor that it would it be easy.
EDIT: I'm assuming you mean an easy-for-user back and forth poly/spline - as of course, there already are the plugins, which require user dinking of splines for 3, 5 pointers.
I too have always been fascinated by neural networks and artificial intelligence, but have never done anything with them. I became aware of them sometime back in the 80's?
I suspect that search engines (like google, bing) might be using some of the concepts for determining which links are the most successful for actually providing info that the searcher was after. Aside from who pays the most advertising dollars.
EDIT: another successful neural app perhaps: The ROOMBA vacuum! - true Fuzzy dust bunny logic!