Understanding Approximate Fisher Information for
Fast Convergence of Natural Gradient Descent
in Wide Neural Networks
Ryo Karakida Kazuki Osawa
Artificial Intelligence Research Center Department of Computer Science
AIST, Japan Tokyo Institute of Technology, Japan
karakida.ryo@aist.go.jp oosawa.k.ad@m.titech.ac.jp
Abstract
Natural Gradient Descent (NGD) helps to accelerate the conver ...


雷达卡


京公网安备 11010802022788号







