A class of asymptotically stable algorithms for learning-rate adaptation.
Algorithmica, 22(1-2) pp. 198–210.
A stability criterion for learning is given. In the case of learning-rate adaptation of backpropagation, a class of asymptotically stable algorithms is presented and studied, including a convergence proof. Simulations demonstrate relevance and limitations.
||convex optimization; neural learning; learning-rate adaptation; line search; stability criterion
||Knowledge Media Institute
||Users 8580 not found.
||08 Oct 2008 13:13
||22 Mar 2016 16:47
|Share this page:
Actions (login may be required)