Organization: University of Toledo, Computer Services
Lines: 29
In article <1993Jan6.182726.29899@cs.brown.edu>, ra@cs.brown.edu (Ronny Ashar) writes:
> Hi
>
> I've been using standard backprop and on-line training, i.e., the algorithm can sample a single randomly generated example at a time. Backprop performs well
> when given about 50,000 such examples( thats a small subset of the domain).
>
> However, I would prefer a more robust algorithm. I was looking at Fahlman's
> Cascade Correlation. My impression is that Cascor needs epoch training only;
> it could be modified to work in stochastic mode, but, in that case, it will
> end up creating huge nets with redundant units. Is that correct?
>
> Have people been using Cascade Correlation in stochastic mode? I'll appreciate
> getting comments/suggestions.
>
> Ronny
>
>
--
In the proceedings of COGANN '92 held in Baltimore, there are
two articles on Genetic Cascade Correlation. It's not exactly
purely stochastic but you might find it interesting. It just
happens to be what I am working on for my thesis too.
