The Expressivity and Training of Deep Neural Networks: toward the Edge of Chaos?

Open in new window