Powell's dogleg trust-region steps with the quasi-Newton augmented Hessian for neural nonlinear least-squares learning.

Eiji Mizutani

Browse the full IJCNN paper archive.