Re:Neural Networks making a come-back?
by basman on Sunday May 01, 2011 @08:05PM
(#2504453)
That's an orthogonal distinction. The methods thus far have typically been parametric, in that there's a fixed network topology and the learning algorithm adjusts the (fixed set of) weights on the edges. There's no reason, though, why you couldn't have a nonparametric version that adaptively chose the number of hidden nodes in the networks and the connectivity structure.
[ Parent
| on HN ]
There are no comments on this comment yet.