Identifying and attacking the saddle point problem in high-dimensional non-convex optimization

Open in new window

The document is on a web server that cannot be reached. Please try again later.