Global Optimality Beyond Two Layers: Training Deep ReLU Networks via Convex Programs

Open in new window