Deep Learning without Poor Local Minima

Open in new window